CELLULAR HEMOGLOBIN-BASED OXYGEN CARRIERS AS POTENTIAL ARTIFICIAL BLOOD SUBSTITUTES
A Dissertation
Submitted to the Graduate School of the University of Notre Dame in Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy
by
Dian Respati Arifin, B.S., M.S.
___________________________ Dr. Andre F. Palmer, Director
Graduate Program in Chemical and Biomecular Engineering Notre Dame, Indiana May 2005
CELLULAR HEMOGLOBIN-BASED OXYGEN CARRIERS AS POTENTIAL ARTIFICIAL BLOOD SUBSTITUTES
Abstract by Dian Respati Arifin
The objective of this study was to develop an effective and reliable cellular hemoglobin-based blood substitute (HBOC). Cellular HBOCs were created by encapsulating bovine hemoglobin (Hb) into the aqueous cores of liposomes, poly(ethylene glycol) (PEG) conjugated liposomes and polymersomes. In order to evaluate the potential of the dispersions as cellular HBOCs, the following physical properties were measured: vesicle size distribution, Hb encapsulation efficiency, oxygen binding property (as indicated by P50 and cooperativity coefficient), and encapsulated methemoglobin level. The oxygen binding properties of unmodified liposomes encapsulated Hb (ULEHs), PEGylated LEHs (PEG-LEHs) and polymersomes encapsulated Hb (PEHs) were comparable to human RBCs, indicating that these vesicles displayed good potential
Dian Respati Arifin as cellular HBOCs. The physical integrity of ULEH dispersions in phosphate buffered saline at physiological pH and temperature was unstable post-production, due to osmosis of water into and out of the liposome core, which implies that ULEHs will be osmotically fragile in the blood stream. This potential problem was solved by grafting PEG molecules onto the liposome surface, thereby strengthening the liposome bilayers. However, the Hb encapsulation efficiency of PEG-LEH dispersions was low. The intravascular circulation, biocompatibility and colloidal state of PEG-LEH dispersions were limited due to the limited PEG surface coverage and molecular weight that can be stably conjugated onto the liposome surface. In contrast, PEH dispersions possessed higher Hb encapsulation efficiencies compared to ULEH, PEG-LEH, PEGylated actin-containing LEH, lipogel particle and nanoscale hydrogel particle dispersions loaded with Hb. Unlike LEH dispersions, encapsulation of Hb into polymersomes did not enhance Hb oxidation. Polymersomes possessed superior PEG shielding ability compared to PEG-liposomes, due to 100% PEG surface coverage with longer PEG brushes. Since polymersomes can be synthesized with thicker hydrophobic membranes compared to liposomes, polymersomes are mechanically stronger than liposomes. Simulation of in-vivo oxygen transport in a capillary and to surrounding tissues, demonstrated that PEH dispersions could be engineered for applications from routine surgery to treatment of trauma. We concluded that PEH dispersions were readily prepared and exhibited good potential as a cellular HBOC, while offering superior physical properties, which may alleviate the limitations encountered with current designs of cellular HBOCs.
To my parents, Rudolph and Astrid, and my brother, Iskandar, for their endless love and support
ii
CONTENTS
TABLES ………………………………………………………………………….. vi
FIGURES …………………………………………………………………………
viii
ACKNOWLEDGMENTS ………………………………………………………
x
CHAPTER 1: INTRODUCTION ………………………………………………... 1.1 Motivation ……………………………………………………………………. 1.2 Artificial Blood Substitutes …………………………………………………... 1.3 Perfluorocarbon-Based Oxygen Carriers …………………………………….. 1.4 Hemoglobin-Based Oxygen Carriers ………………………………………… 1.4.1 Acellular Hemoglobin-Based Oxygen Carriers …………………..... 1.4.2 Cellular Hemoglobin-Based Oxygen Carriers ……………………... 1.5 Objective and Scope of Research ……………………………………………. 1.5.1 Design Criteria of Cellular Hemoglobin-Based Oxygen Carriers ..... 1.5.2 Outline of Dissertation ……………………………………………...
1 1 3 3 5 7 9 12 12 14
CHAPTER 2: MATERIALS AND METHODS ………………………………… 2.1 Preparation of Hemoglobin Stock Solution ………………………………...... 2.1.1 Purification of Bovine Hemoglobin ………………………………... 2.1.2 Assay of Hemoglobin and Methemoglobin Concentrations ……….. 2.2 Preparation of Unmodified Liposome, PEG-Liposome and Polymersome Encapsulated Hemoglobin Dispersions ……………………………………… 2.2.1 Materials …………………………………………………………… 2.2.2 Preparation Procedure ……………………………………………… 2.3 Vesicle Size Distribution …………………………………………………….. 2.3.1 Introduction ………………………………………………………… 2.3.2 Experimental Procedure ……………………………………………. 2.3.3 Theoretical Background ……………………………………………. iii
18 18 18 19 20 20 22 24 24 26 28
2.3.3.1 Flow field-flow fractionation theory ……………………... 28 2.3.3.2 Light scattering theory …………………………………… 2.3.3.3 Differential interferometric refractometer ……………….. 2.4 Hemoglobin Encapsulation Efficiency ………………………………………. 2.4.1 Introduction ………………………………………………………… 2.4.2 Experimental Procedure ……………………………………………. 30 34 35 35 36
2.5 Oxygen Binding Properties …………………………………………………... 39 2.6 Encapsulated Methemoglobin Level …………………………………………. 42
CHAPTER 3: UNMODIFIED LIPOSOME ENCAPSULATED HEMOGLOBIN DISPERSIONS …………………………………………………………………... 43 3.1 Introduction …………………………………………………………………... 3.2.1 Average Radius …………………………………………………….. 43 44 3.2 Results and Discussion ……………………………………………………….. 44 3.2.2 The Effect of Osmotic Pressure on the Size of Unmodified Liposomes Encapsulated Hemoglobin …………………………….. 48 3.2.3 Size Distribution …………………………………………………… 3.2.4 Hemoglobin Encapsulation Efficiency …………………………….. 50 57
3.2.5 Oxygen Binding Properties ………………………………………… 61 3.2.6 Encapsulated Methemoglobin Level ……………………………….. 63 3.3 Concluding Remarks …………………………………………………………. 63
CHAPTER 4: POLY(ETHYLENE GLYCOL) CONJUGATED LIPOSOME ENCAPSULATED HEMOGLOBIN DISPERSIONS …………………………... 4.1 Introduction …………………………………………………………………... 4.1.1 Background ………………………………………………………… 4.1.2 Research Overview ………………………………………………… 4.2.1 Average Radius …………………………………………………….. 4.2.2 Size Distribution …………………………………………………… 4.2.3 Hemoglobin Encapsulation Efficiency …………………………….. 4.2.4 Oxygen Binding Properties and Encapsulated Methemoglobin Level ………………………………………………………………. iv
65 65 65 67 68 71 75 79
4.2 Results and Discussion ……………………………………………………….. 68
4.3 Concluding Remarks …………………………………………………………. 82 CHAPTER 5: POLYMERSOME ENCAPSULATED HEMOGLOBIN DISPERSIONS ……………………………………………………………..……. 5.1 Introduction ……………………………………………………………..……. 5.1.1 Background ………………………………………………………… 5.1.2 Research Overview ………………………………………………… 5.2 Results and Discussion ………………………………………………............. 5.2.1 Size Distribution and Average Radius ……………………………... 5.2.2 Hemoglobin Encapsulation Efficiency ……………………………..
84 84 84 87 89 89 100
5.2.3 Oxygen Binding Properties ………………………………………… 103 5.2.4 Encapsulated Methemoglobin Level ……………………………….. 105 5.3 Concluding Remarks …………………………………………………………. 106 CHAPTER 6: SIMULATION OF IN VIVO OXYGEN TRANSPORT IN A CAPILLARY AND TO THE SURROUNDING TISSUE UTILIZING AN OXYGEN CARRIER ……………………………………………………………. 6.1 Introduction …………………………………………………………………... 6.2 Theoretical Background ……………………………………………………… 6.3 Results ………………………………………………………………………... 6.4 Discussion …………………………………………………………………….
107 107 108 121 129
6.5 Concluding Remarks …………………………………………………………. 133
CHAPTER 7: CONLUSIONS AND PROPOSED FUTURE STUDIES ………... 135 7.1 Conclusions …………………………………………………………………... 7.2 Proposed Future Studies ……………………………………………………... 135 138
REFERENCES ………………………………………………………………........ 146
APPENDIX A: MATLAB CODE FOR SIMULATION OF OXYGEN TRANSPORT …………………………………………………………………….
159
v
TABLES
3.1
NUMBER- (Rn), WEIGHT- (Rw), Z-AVERAGED (Rz) RADII AND POLYDISPERSITY INDICES OF ULEH AND CONTROL DISPERSIONS EXTRUDED IN PB AND PBS………………………... 46 OSMOTIC PRESSURE GRADIENTS (∆P) BETWEEN THE AQUEOUS CORE OF ULEHS AND THE EXTERIOR ENVIRONMENT (PBS) ………………………………………………... 49 ENCAPSULATION EFFICIENCY (E%), WEIGHT RATIO OF ENTRAPPED HB TO PHOSPHOLIPIDS (R), P50, COOPERATIVITY COEFFICIENT (n), AND ENCAPSULATED METHB LEVEL OF PBLEHS ……………………………………………………………………. 58 ENCAPSULATION EFFICIENCY (E%), WEIGHT RATIO OF ENTRAPPED HB TO PHOSPHOLIPIDS (R), P50, COOPERATIVITY COEFFICIENT (n), AND ENCAPSULATED METHB LEVEL OF PBS-LEHS ……………………………………………………………… 59
3.2
3.3
3.4
4.1
NUMBER- (Rn), WEIGHT- (Rw), Z-AVERAGED RADII (Rz) AND POLYDISPERSITY INDICES OF PEG-LEH AND PEG-CONTROL DISPERSIONS EXTRUDED IN PB …………………………………… 69 NUMBER- (Rn), WEIGHT- (Rw), Z-AVERAGED RADII (Rz) AND POLYDISPERSITY INDICES OF PEG-LEH AND PEG-CONTROL DISPERSIONS EXTRUDED IN PBS ………………………………….. 70 ENCAPSULATION EFFICIENCY (E%), WEIGHT RATIO OF ENTRAPPED HB TO PHOSPHOLIPIDS (R), P50, COOPERATIVITY COEFFICIENT (n), AND ENCAPSULATED METHB LEVEL OF PEG-LEHS EXTRUDED IN PB ……………………………………….. 77 ENCAPSULATION EFFICIENCY (E%), WEIGHT RATIO OF ENTRAPPED HB TO PHOSPHOLIPIDS (R), P50, COOPERATIVITY COEFFICIENT (n), AND ENCAPSULATED METHB LEVEL OF PEG-LEHS EXTRUDED IN PBS ……………………………………… 78 vi
4.2
4.3
4.4
5.1
NUMBER-AVERAGED MOLECULAR WEIGHT (Mn), PEG LENGTH, HYDROPHILIC MASS FRACTION (fhydrophilic), AND HYDROPHOBIC MEMBRANE THICKNESS (d) OF PBD-PEO DIBLOCK COPOLYMERS USED IN THIS STUDY ………………… 89 NUMBER- (Rn), WEIGHT- (Rw), AND Z-AVERAGED (Rz) RADII AND SIZE DISTRIBUTION WIDTHS OF PEH DISPERSIONS …….. 93 NUMBER- (Rn), WEIGHT- (Rw), AND Z-AVERAGED (Rz) RADII AND SIZE DISTRIBUTION WIDTHS OF CONTROL (EMPTY POLYERSOME) DISPERSIONS ………………………………………. 94 ENCAPSULATION EFFCIENCY (E%), WEIGHT RATIO OF ENTRAPPED HB TO COPOLYMER (R), P50, COOPERATIVITY COEFFICIENT (n) AND ENCAPSULATED METHB LEVEL OF PEH DISPERSIONS ……………………………………………………. 101
5.2
5.3
5.4
6.1
TOTAL AMOUNT OF OXYGEN RELEASED IN PANCREAS AND BRAIN TISSUES (µmol/s) EXPOSED TO VARIOUS OXYGEN CARRIERS WITH DIFFERENT OXYGEN BINDING PROPERTIES (P50 and COOPERATIVITY COEFFICIENT (n)), AND OXYGENATION SCENARIOS (pO2,inlet) ……………………………... 128 TOTAL AMOUNT OF OXYGEN RELEASED IN MUSCLE TISSUE (µmol/s) EXPOSED TO VARIOUS OXYGEN CARRIERS WITH DIFFERENT OXYGEN BINDING PROPERTIES (P50 and COOPERATIVITY COEFFICIENT (n)), AND OXYGENATION SCENARIOS (pO2,inlet) …………………………………………………. 128 TYPE OF PBD BLOCK, NUMBER-AVERAGED MOLECULAR WEIGHT (Mn), PEG LENGTH, HYDROPHILIC MASS FRACTION (fhydrophilic), AND HYDROPHOBIC MEMBRANE THICKNESS (d) OF PBD-PEO DIBLOCK COPOLYMERS ………………………………… 139
6.2
7.1
vii
FIGURES Staggered chromatograms of the DIR and 90o light scattering detector signals for ULEH and PEG-LEH dispersions …………………………... 27 Staggered chromatograms of the DIR and 90o light scattering detector signals for PEH dispersions ……………………………………………... 28 Staggered chromatograms of DIR signals of the Hb reference and unencapsulated Hb of ULEH and PEG-LEH dispersions ………………. 38 Staggered chromatograms of DIR signals of the Hb reference and unencapsulated Hb of PEH dispersions ………………………………… 39
2.1
2.2
2.3
2.4
3.1 3.2 3.3 3.4
Size distributions of ULEHs extruded in PB …………………………… Size distributions of ULEHs extruded in PBS ………………………….. Size distributions of unmodified controls, which are empty liposomes ...
51 52 53
Oxygen dissociation curves of bovine Hb, and ULEHs prepared using an initial Hb concentration of 200 mg/mL and extruded through 400 nm pore diameter membranes in PB and PBS ……………………………… 62
4.1
Size distributions of PEG-LEH dispersions extruded with an initial Hb concentration of 300 mg/mL ……………………………………………. 73 Size distributions of PEG-control dispersions ………………………….. 74
4.2 4.3
Oxygen dissociation curves of bovine Hb, PEG-LEHs grafted with 2000 Da PEG and extruded in PB and PBS, and PEG-LEHs grafted with 550 Da PEG and extruded in PB and PBS …………………………………... 81
5.1 5.2
Differential and cumulative size distributions of PEH dispersions ……... 91 Differential and cumulative size distributions of control (empty viii
polymersome) dispersions ………………………………………………. 92 5.3 Oxygen dissociation curves of bovine Hb and PEH dispersions extruded through 200 nm pore diameter membranes ……………………………... 104
6.1 6.2
Geometry of the Krogh tissue cylinder model …………………………..
109
Oxygen dissociation curves of cellular oxygen carriers with P50 and cooperativity coefficients of 38 mmHg and 2.9, 28 mmHg and 2.4, and 17 mmHg and 2, respectively …………………………………………… 121 Oxygen partial pressure profiles (pO2) in a capillary and the surrounding Krogh tissue region for the case where no oxygen carrier is present in the blood ………………………………………………………………… 122 Oxygen partial pressure profiles (pO2) in a capillary and the surrounding Krogh tissue region simulated using the P50 and cooperativity coefficient of human RBCs (26 mmHg and 2.3, respectively) …………. 123 Oxygen partial pressure profiles (pO2) in a capillary and the surrounding Krogh tissue region simulated using the highest P50 and cooperativity coefficient observed for ULEH/PEG-LEH/PEH oxygen carriers (38 mmHg and 2.9, respectively) …………………………………………... 124 Oxygen partial pressure profiles (pO2) in a capillary and the surrounding Krogh tissue region simulated using the average P50 and cooperativity coefficient observed for ULEH/PEG-LEH/PEH oxygen carriers (28 mmHg and 2.4, respectively) …………………………………………… 125 Oxygen partial pressure profiles (pO2) in a capillary and the surrounding Krogh tissue region simulated using the lowest P50 and cooperativity coefficient observed for ULEH/PEG-LEH/PEH oxygen carriers (17 mmHg and 2, respectively) ……………………………………………... 126 Oxygen partial pressure profiles (pO2) in a capillary and the surrounding Krogh tissue region simulated using a high oxygen affinity oxygen carrier (P50 and cooperativity coefficient of 7 mmHg and 2, respectively) …………………………………………………………….. 127
6.3
6.4
6.5
6.6
6.7
6.8
7.1
Circulation half-life of PEH dispersions in rats (t1/2) versus length of PEG block ………………………………………………………………. 140
ix
ACKNOWLEDGMENTS
I would like to thank many people who have contributed to my graduate study and life in Notre Dame: My advisor, Professor Andre Palmer, for his insightful guidance and advice throughout the course of my research. My colleagues, Julie Eike, Jason Gordon, Shuliang Li, Michael Dimino, Jaquanda Patton, Jesse Sullivan and Sharon Gundersen, for their help in the experiments and for being good friends. Dennis Birdsell and Rian Galloway from the Center for Environmental Science and Technology (University of Notre Dame, South Bend, IN, USA) for the use of the Center’s facilities. My class-mates and friends for making my life in Notre Dame a wonderful experience. And last but not the least, Arun Ramachandran, for his continuous support during my study and for his insightful discussion in completing this dissertation. The work presented in this dissertation was supported by the National Science Foundation (Arlington, VA, USA), grant number BES-0196432.
x
CHAPTER 1
INTRODUCTION
1.1 Motivation The need for human blood for blood transfusions is steadily increasing. However, donated human blood is riddled with many problems, such as, limited availability, short storage lifetime, possibility of transmission of infectious diseases by blood-borne pathogens, allergic reactions, and problems associated with cross-matching different blood types [1-5]. In the U.S. alone, a deficit of 3 million units of blood is projected by 2030 [6]. This deficit was predicted based on the projected population growth, and the projected increased need for blood in cases of routine surgery and trauma. However, this projected deficit does not take into account the more acute need for blood in cases of mass civilian casualties for example: natural disasters, terrorist attacks, and wars. While the demand for blood is increasing at a rate of 1% per year, U.S. blood donations are decreasing at an annual rate of 1% [7]. Assuming uniform medical practices world-wide and applying the U.S. transfusion rate of 0.05 units per capita per year, a total international demand for 300 millions units of blood per year can be projected. However, the number of units actually transfused is speculated to be as low as 90 million. If this projection is accurate, then there exists an international shortage of blood of more than
1
200 millions units per year [8]. Although an exact estimate is hard to obtain, blood shortages in developing countries are more evident [1, 4]. For example, one-quarter of maternal deaths from pregnancy-related causes in developing countries are associated with blood loss [9]. Moreover, blood has a storage lifetime of 15 days in a 2-3oC environment, which can be extended to 42 days using adenine-acid citrate dextrose and adenine-citrate phosphate dextrose storage solutions [10]. The collection and storage of blood, which requires refrigeration systems and screening procedures, is expensive. Based on the current estimate that 14 million units of blood are annually used in the U.S. [6] at a cost of $100-$225 per unit [7, 11], the cost of blood transfusions is estimated to range between $1.4-3 billion per year in the U.S. Before the development of reliable blood screening procedures, the probability of contracting hepatitis C and HIV were approximately 1:11 and 1:2500 [12], respectively. Currently in developed countries, the probability of contracting hepatitis C is 1:1 million [13], and 1:676,000 for HIV [14]. However, blood screening is largely not available or poorly performed in less developed countries. Currently, an estimated 6 million units of global blood donations are not tested for HIV, hepatitis and syphilis, primarily in developing countries where the infected population is large [1]. Even transfusion of screened donor blood is hardly risk-free. Although the data is still ambiguous, blood transfusion may exert an immunosuppressive effect on the recipient, making them more susceptible to infection [15]. Blood transfusion may also lead to various metabolic conditions, such as, hyperkalemia, hypocalcemia and alkalosis [11]. The aforementioned concerns emphasize the increasing need for the development of a reliable and safe artificial blood substitute.
2
1.2 Artificial Blood Substitutes Artificial blood substitutes possess several advantages over the use of donor blood for blood transfusions. Artificial blood substitutes can be designed to have no antigenic blood groups on their surface, and thus can be administered to individuals possessing any type of blood group [5]. They can be readily mass-produced with guaranteed sterility, therefore eliminating the possibility of transmittal of infections and the need for blood screening as in the case of donor blood [4]. They can be designed to have longer storage lifetimes, while requiring simpler storage facilities compared to donor blood [2, 5]. In addition, artificial blood substitutes are also expected to be cheaper to produce compared to the costs associated with screening and storing donor blood [2, 5, 12]. The first step in the development of an artificial blood substitute is to concentrate on the major function of blood, which is to transport oxygen to tissues. Due to the poor solubility of oxygen in blood plasma, this function is carried out by hemoglobin (Hb), a protein encapsulated inside red blood cells (RBCs), which is capable of binding and releasing oxygen [16, 17]. Currently, artificial oxygen carriers are being developed based on two types of molecules that are capable of transporting oxygen namely: perfluorocarbons (PFCs) and Hb [2-4].
1.3 Perfluorocarbon-Based Oxygen Carriers PFCs are synthetic, inert fluorinated hydrocarbons that are capable of storing large amounts of dissolved oxygen [3, 4, 18]. At a given dose of PFC, a linear relationship exists between the amount of oxygen dissolved in the fluid phase of PFC dispersions and the partial pressure of oxygen (pO2). Therefore, high concentrations of 3
dissolved oxygen in PFC-based oxygen carriers are available for diffusion into tissues only at high pO2s. Hence, high-inspired-oxygen concentrations are required for physiological oxygen delivery from PFC dispersions [4, 12]. Moreover, PFCs are immiscible in water, and need to be emulsified in order to be administered to patients [4, 18]. Fluosol DA, a commercial PFC preparation created by Green Cross Corporation (Osaka, Japan), was approved by the FDA for human use in 1989 as an adjunct for oxygen delivery during angioplasty. However, patients treated with Fluosol DA in clinical trials were plagued with adverse physiological reactions. Fluosol DA became obsolete with new developments in elucidating blood flow mechanisms for angioplasty, and was removed from the market in 1994 [9, 15, 19]. Second generation PFC products include OxyfluorTM (HemaGen/PFC, Waltham, MA, USA), OxygentTM (Alliance Pharmaceutical Corporation, San Diego, CA, USA), and PerftoranTM (SPC-Perftoran, Moscow, Russia). In phase I clinical trials of OxyfluorTM, several side effects were noted that might limit the doses that could be transfused into patients [20] and its development was terminated [21]. OxygentTM was implicated with an increased incidence of stroke in coronary bypass patients [22] and its clinical trials were terminated [4]. PerftoranTM has been approved for human use in Russia since 1999, however, no clinical studies with PerftoranTM has been conducted outside Russia [23]. Other PFC products in development include Pher-O2TM from Sanguine Corp. (Pasadena, CA, USA), and OxycyteTM from Synthetic Blood International, Inc. (Kettering, OH, USA) [21].
4
1.4 Hemoglobin-Based Oxygen Carriers Hemoglobin is a tetrameric protein contained inside erythrocytes capable of binding and releasing oxygen. Hb found in adult humans is referred to as hemoglobin A. Tetrameric Hb or Hb in its native form consists of two pairs of αβ dimers held together by non-covalent bonds with a total molecular weight of 64 kDa [16]. The concentration of Hb in human blood is ~ 150 mg/mL. The oxygen binding capacity of a Hb molecule (tetrameric Hb) depends on the presence of a non-polypeptide unit called the heme group. The iron atom at the center of a heme group can bind one molecule of oxygen. Therefore, one molecule of Hb, which possesses four heme groups, has the potential to bind four molecules of oxygen. Hb is an allosteric protein, and its oxygen binding properties are affected by interactions between neighboring oxygen binding sites (heme groups). The binding of oxygen to Hb is cooperative (the binding of an oxygen molecule to Hb enhances the binding of additional oxygen molecules). The cooperative binding of oxygen to Hb gives native Hb its characteristic sigmoidal-shaped oxygen dissociation curve [16, 17]. In contrast to the non-cooperative binding of oxygen to PFC molecules, the cooperative binding of oxygen to Hb allows oxygen to be bound and released within a narrow physiological range of pO2s ranging from 40-100 mmHg [3, 17]. The cooperativity of oxygen binding to Hb provides for an approximate 70% increase in oxygen delivery in an active capillary bed compared to non-cooperative oxygen binding [17]. The oxygen binding properties of human Hb contained within RBCs is also allosterically regulated by the presence of the inorganic molecule 2,3-diphosphoglycerate (2,3-DPG), which is also contained within RBCs at approximately the same 5
concentration as Hb. 2,3-DPG reduces the oxygen affinity of Hb by keeping the Hb molecule in the deoxy conformation. In the presence of 2,3-DPG, oxygen bound to Hb can be easily released in low pO2 regions of the body [17]. The oxygen binding properties of Hb is indicated by its P50 (the pO2 where 50% of the Hb oxygen binding sites are saturated with oxygen), and cooperativity coefficient (which indicates the cooperativity of oxygen binding to Hb [17]). The P50 and cooperativity coefficient of human Hb contained within RBCs is 26 mmHg and 2.3, respectively [24]. The valence of the iron atom in the heme group determines the oxygen binding capacity of Hb. Hb auto-oxidizes from the 2+ valence state to form non-oxygen binding methemoglobin (metHb), where the iron atom is now in the 3+ valence state [16]. Hb auto-oxidation generates several highly oxidative species, such as superoxide anions, and hydrogen peroxide, which can damage proteins, carbohydrates and nucleic acids, stimulate the peroxidation of unsaturated lipids, and promote the production of a range of cytotoxic molecules [4]. MetHb is structurally less stable than Hb, and readily releases the heme ring into the blood stream. Due to its hydrophobic nature, the heme ring can intercalate into the membranes of endothelial cells where it has the potential to catalyze the peroxidation of lipids, therefore resulting in cell damage and eventual cell death [25]. The rate of human Hb auto-oxidation in-vitro is 1% per hour at physiological temperature (37oC) [26]. Hb auto-oxidation also occurs within RBCs. However, the metHb level of RBCs is regulated at 1% [27] by enzymatic pathways involving metHb reductases, and by non-enzymatic pathways involving ascorbic acid and glutathione [16]. When cell-free Hb was infused into animals, vasoconstriction and kidney damage was observed [11, 19, 25]. Vasoconstriction is thought to occur via two mechanisms. In
6
the first mechanism, tetrameric Hb is able to extravasate through the capillary walls, where it accumulates in the smooth muscle cells, sequestering NO, which then leads to vasoconstriction [25]. Vasoconstriction is also believed to be the result of an autoregulatory response to supra-physiological amount of oxygen being delivered to tissues via the capillary beds. In this second mechanism, tetrameric Hb delivers excess oxygen to capillaries in the circulatory system. The body responds to these high oxygen levels by constricting blood vessels, thus reducing the available surface area for oxygen diffusion in order to maintain normal oxygen delivery [20]. In addition to vasoconstriction, kidney damage also occurred as a result of infusion of tetrameric Hb into the circulatory system. In fact, tetrameric Hb is rapidly filtered and removed by the renal tubules of the kidneys. In the renal tubules, the heme groups dissociate from the tetramer and induce renal toxicity [11, 19]. Given the various problems associated with the use of tetrameric Hb as an oxygen carrier, the development of Hb-based oxygen carriers (HBOCs) is now directed towards stabilizing the Hb molecule in order to eliminate the adverse physiological effects associated with cell-free Hb, while maintaining the physiological oxygen transporting ability of Hb [3]. HBOCs can be characterized generally as either acellular or cellular oxygen carriers.
1.4.1 Acellular Hemoglobin-Based Oxygen Carriers Acellular HBOCs are prepared by molecular modification of tetrameric Hb. Polymeric Hb is synthesized by intra- and inter-molecularly cross-linking tetrameric Hb, while solely intra-molecular cross-linking tetrameric Hb produces cross-linked Hb [3, 4].
7
Recombinant Hb is synthesized by fusing the two α subunits to prevent dissociation of the resulting tetramer into dimers [18]. Tetrameric Hbs have been conjugated with bioinert, water-soluble polymers to form conjugated Hbs [3, 4]. Several commercial acellular HBOCs are currently in development and clinical trials. Three polymeric Hb dispersions currently in advanced stages of development are [4, 5]: PolyHemeTM (Northfield Lab, Evanston, IL, USA), HemopureTM (Biopure Corp., Cambridge, MA, USA), and HemolinkTM (Hemosol, Inc., Mississauga, ON, Canada). The first two are glutaraldehyde polymerized Hbs, while the later is an o-raffinose polymerized Hb. PolyHemeTM was transfused into the victim of an auto accident on a compassionate use basis to treat severe hemorrhagic shock [19, 28]. This sustained the patient for several days until hemorrhage was controlled and erythropoiesis compensated for blood loss [19, 28]. This product was also successfully used in conjunction with erythropoietin therapy to treat a critically anemic woman who suffered from persistent colonic bleeding [29]. In April 2001, HemopureTM was approved for clinical use in South Africa [19, 28]. This product was successfully used to support a patient with autoimmune hemolytic anemia [19]. Recently, o-raffinose polymerized Hb was shown to be less stable with respect to auto-oxidation, oxidative modifications, and the integrity of the heme group compared to native Hb and glutaraldehyde polymerized Hb [4]. PolyHemeTM, HemopureTM and HemolinkTM are currently in phase III clinical trials [4]. Clinical studies showed that perioperative use of these products can reduce the number of allogeneic RBC units, and increase the avoidance rate of allogeneic transfusion in emergency bleeding, and vascular, cardiac and noncardiac surgeries [5]. Requests for approval have 8
been filed for these products in the U.S., U.K. and/or Canada, and are pending [2]. Development of a diaspirin-cross-linked Hb, HemAssistTM (Baxter International, Round Lake, IL, USA), was terminated in 1998 because of discouraging results in phase III clinical trials [20]. Research on a cross-linked recombinant Hb, OptroTM (Somatogen, Boulder, CO, USA) was temporarily halted, but its development was continued when Baxter purchased Somatogen in 1998 [21]. Baxter recently halted the clinical development of this product without further explanation [4]. Other commercial products currently in development are PHPTM (Apex Bioscience, Research Triangle Park, NC, USA), a poly(oxyethylene) conjugated Hb, and PEG-HemoglobinTM (Enzon, Piscataway, NJ, USA), a poly(ethylene glycol) conjugated Hb [4]. Unfortunately, acellular HBOCs were observed to induce vasoconstriction when transfused into animals, presumably due to NO sequestering and/or an over-oxygenation auto-regulatory response [20, 30, 31]. Moreover, acellular HBOCs were shown to exhibit short circulatory half-lives, and are therefore most suitable for short-term applications [28, 32].
1.4.2 Cellular Hemoglobin-Based Oxygen Carriers In the case of cellular-based HBOCs, Hb molecules are encapsulated inside the aqueous core of semi-permeable membrane shells. The structure of these oxygen carriers is similar to erythrocytes. The membranes protect surrounding tissues and blood components from direct contact with potentially toxic tetrameric Hb [4]. Encapsulating Hb inside vesicles permits manipulation of its physiochemical properties, and circulation lifetime [4, 25]. The vesicle membrane reduces recognition of encapsulated Hb by the reticuloendothelial system (RES), therefore, prolonging the intravascular lifetime of the
9
HBOC [25]. Surface modification of the vesicle membrane can further improve the intravascular persistence, and colloidal state of these oxygen carriers [33, 34]. Catalases, reductants and reductases can be readily coencapsulated with Hb to suppress metHb formation inside these oxygen delivery vehicles [35-37]. Currently, there are no cellular HBOCs in clinical trials. Two types of cellular HBOC membranes are being explored, namely: nanocapsules composed of biodegradable polylactide and polyglycolide, and liposomes composed of naturally occurring phospholipids [2, 25]. Both membranes have no blood group antigens on the surface, making them suitable as a universal oxygen carrier. Nanocapsules are stronger and more porous than liposomes, and thus require less membrane material. However, proteins and small hydrophilic molecules can pass through the membrane, a potential problem not encountered with liposomes. This can be especially problematic in the case of chloride ions, which are naturally abundant in blood plasma and are known to promote Hb oxidation [32, 38, 39]. In contrast, the permeability of liposomes to charged molecules is extremely low due to the hydrophobic nature of the membrane [40]. Hence, liposomes are essentially only permeable to water and non-polar molecules, such as, oxygen. The diameter of these vesicles can be easily engineered to be more than 100 nm, which prevents their filtration by the capillaries and the kidney tubules [25]. It was observed that liposome encapsulated Hb (LEH) dispersions with an average diameter of 250 nm could easily permeate through artificial capillaries with diameters of 3 µm and less, while erythrocytes could hardly traverse these artificial capillaries [41]. In humans, the capillary diameter is ~ 10 µm, and during circulatory failure such as in the case of hemorrhagic shock, the capillary typically constricts [41]. This study demonstrated quite convincingly that LEHs can easily 10
permeate through capillary blockages, and are potentially suitable oxygen carriers in cases of trauma and routine surgery. However, unmodified LEHs were observed to have short circulation half-lives (~12-18 hours in rabbits) [34, 42], and intervesicular aggregation and fusion occurred after several days of storage [33]. To alleviate these problems, the surface of LEHs was conjugated with poly(ethylene glycol) (PEG), a biocompatible and water-soluble polymer [4, 25, 43]. This advanced design of LEH is referred to as PEG-LEH. PEGylation improves not only the intravascular circulation and colloidal state of PEG-LEHs, but also their biocompatibility [33, 34, 43, 44]. However, the PEG surface density and length of PEG chains that can be conjugate onto liposome bilayers is severely limited [45, 46]. In general, increasing the PEG surface coverage and molecular weight of the PEG chains increases the circulation half-life, colloidal state, and biocompatibility of PEG-LEH dispersions. Thus, there is a physical limit to optimizing the physical properties of PEG-LEH dispersions [47-49]. For an optimum PEG surface density of 10 mole% with 5000 Da PEG chains, the circulation half-life of this PEG-LEH dispersion was reported to be 48 hours in rabbits or ~ 36 hours in rats [34, 50]. In contrast, the circulation lifetime of human RBCs is 120 days [16]. Moreover, conjugation of 5000 Da and lower molecular weight PEG chains may not effectively prevent complement activation in-vivo [51]. Complement activation is
associated with anaphylactic reactions and activation of other proteolytic plasma cascades in-vivo [51]. The results of these studies emphasize the need for the continuing development of a reliable HBOC.
11
1.5 Objective and Scope of Research The ultimate goal of our research is to create an effective and safe HBOC for clinical use. Due to the advantages of cellular HBOCs compared to their acellular counterparts [52], we concentrated our research on the improved development of cellular HBOCs. In this dissertation, liposomes and PEGylated liposomes were employed as invivo delivery systems to transport oxygen [39, 53, 54]. In addition, we selected a new type of membrane shell to transport oxygen via hemoglobin encapsulation: polymer vesicles, also called polymersomes [55]. The design of polymersomes offers 100% PEG surface density with longer PEG chains compared to PEG-LEH dispersions [56-58]. Hence, polymersomes can overcome the aforementioned limitations encountered with PEG-LEH dispersions. Moreover, the hydrophobic membranes of polymersomes can be engineered to be thicker than those of liposomes, which contribute to better mechanical strength [56, 58, 59]. Hb molecules were readily encapsulated inside the aqueous core of various vesicles to create LEH, PEG-LEH and polymersome encapsulated Hb (PEH) dispersions. Bovine Hb, which was shown to be a compatible replacement for human Hb [60], was used in our research due to its availability and cheaper cost compared to human Hb.
1.5.1 Design Criteria of Cellular Hemoglobin-Based Oxygen Carriers In order to serve as an effective oxygen carrier, vesicles loaded with Hb must satisfy the following design criteria of a cellular HBOC: 1. Cellular HBOCs should possess monodisperse size distributions, since the circulation half-life of vesicles is influenced by vesicle size [61]. Moreover, a
12
monodisperse vesicle suspension was shown to have simple exponential clearance kinetics compared to a heterogeneous dispersion, thus permitting easier control and manipulation of the circulation lifetime [61]. 2. Cellular HBOCs should possess diameters larger than 100 nm to prevent filtration by the capillaries and kidney microtubules [25], and less than 250 nm since larger vesicles are easily recognized and removed by the RES [62]. 3. Cellular HBOCs should possess high Hb encapsulation efficiencies. PEG-LEH with an encapsulated Hb concentration of 38 mg/mL was effective in resuscitating hamsters experiencing hemorrhagic shock [63]. However, cellular HBOCs should exhibit high Hb encapsulation efficiencies, in order to limit the amount of membrane materials required to produce HBOCs. This design criteria will limit saturation of the RES by the membrane materials, and therefore, avoid the potential adverse physiological effects of RES blockage. 4. For application in routine surgery, cellular HBOCs should possess normal physiological oxygen binding properties in order to maintain normal tissue oxygenation. HBOCs should possess a P50 and cooperativity coefficient in the range of 21-30 mmHg and 2-2.6, respectively [17]. These oxygen binding parameters indicate oxygen binding properties comparable to that of human RBCs. However, for resuscitation of patients experiencing hemorrhagic shock, a cellular HBOC with a P50 less than that of human RBCs (26 mmHg) is desired in order to target oxygen delivery to tissues with low pO2s [20]. 5. Cellular HBOCs should possess encapsulated metHb levels less than 1% [27], similar to that of RBCs, since metHb cannot bind and release oxygen. In
13
addition, metHb level >10% was observed to significantly impair tissue oxygenation when infused into rats [64]. 6. Cellular HBOCs should possess an apparent viscosity of 4 cP, similar to that of human blood, in order to preserve the shear forces in the microcirculation [20]. 7. Cellular HBOCs should possess an oncotic pressure of 50 mmHg, in order for the oxygen carrier to act as a plasma expander in the resuscitation of patients experiencing hemorrhagic shock, where the patients typically experience severe loss of blood volume [20]. However, for application in routine surgery, the oncotic pressure of cellular HBOCs should be similar to that of human blood (25 mmHg) in order to maintain normal vascular volume [41]. 8. Cellular HBOCs should persist intravascularly for as long as possible. The circulation lifetime of human RBCs is 120 days [16]. 9. Cellular HBOCs should possess oxygen binding parameters conducive to targeting oxygen delivery to 1) tissues experiencing normal physiological oxygenation (oxygen partial pressure entering the capillary is ~100 mmHg) [17]; and 2) tissues experiencing hypoxic oxygenation (oxygen partial pressure entering the capillary is < 40 mmHg) [65].
1.5.2 Outline of Dissertation In this work, the potential of LEH, PEG-LEH and PEH dispersions to function as cellular HBOCs was evaluated by investigating the following physical properties: vesicle size distribution, Hb encapsulation efficiency, oxygen affinity (as indicated by P50 and cooperativity coefficient), and encapsulated metHb level. The apparent viscosity and
14
oncotic pressure of the dispersions were not measured, since the viscosity and oncotic pressure of vesicle suspensions can be easily engineered post-production by simply adjusting the total concentration of vesicles and/or via the addition of human serum albumin to the HBOC suspension [63]. In order to characterize the ability of these HBOCs to transport oxygen to tissues, oxygen transport in a capillary and its surrounding tissue was simulated using the Krogh tissue cylinder model [66]. The ability of these HBOCs to transport oxygen was evaluated under several scenarios: 1) various oxygen partial pressures entering the capillary (100, 30 and 15 mmHg) to simulate oxygen transport under normal physiological conditions (oxygen partial pressure entering the capillary is 100 mmHg), severe hypoxia (oxygen partial pressure entering the capillary is 30 mmHg), and an extremely low tissue oxygenation (oxygen partial pressure entering the capillary is 15 mmHg); 2) various P50s and cooperativity coefficients exhibited by LEH, PEG-LEH, PEH dispersions, and human RBCs; and 3) various oxygen consumption rates that are present in tissues such as the pancreas, brain, and muscle. HBOC dispersions that exhibit the best potential as cellular HBOCs (based on the aforementioned design criteria) will be selected for animal studies in order to measure their circulation half-life and biocompatibility. The outline of our research to evaluate the potential of LEH, PEG-LEH and PEH dispersions as cellular HBOCs is presented below. In Chapter 2, the experimental procedures to measure the aforementioned physical properties of LEH, PEG-LEH and PEH dispersions are presented. A new experimental technique consisting of an asymmetric flow field-flow fractionator (AFFFF) coupled to a multiangle static light scattering (MASLS) photometer and a differential interferometric refractometer (DIR) was used to measure the absolute size distribution of
15
various vesicle dispersions [39, 53, 55]. In previous studies [42, 67-69], only the average size of LEH and PEG-LEH dispersions was measured. In contrast, the true size distribution of a vesicle suspension can be measured using the AFFFF-MASLS-DIR setup [39, 53, 55]. Moreover, we developed a novel technique to measure the Hb encapsulation efficiency of a vesicle dispersion using the AFFFF-MASLS-DIR set-up [39]. Unlike other methods for determining Hb encapsulation efficiency [42, 67, 69], this technique is convenient, accurate and can be performed in the sample’s aqueous environment [39]. Chapter 3 is devoted to elucidating the physical properties of LEH dispersions. To avoid confusion with PEG-LEHs, LEHs presented in this chapter will be referred to as unmodified LEHs (ULEHs). ULEHs of various sizes and encapsulated Hb concentrations were prepared. The aqueous core of ULEHs was engineered to exhibit near-physiological ionic strength, and an ionic strength less than physiological ionic strength. The aim of this study was to investigate the effect of in-vivo osmotic pressure on the integrity of ULEH bilayers [39]. The results of this study will contribute to the development of mechanically strengthened ULEHs, which can withstand the lyzing effect of the osmotic pressure exerted by blood. In addition, the influence of the size, ionic strength of the aqueous core and Hb concentration used to prepare ULEHs on the oxygen carrying and binding ability of ULEHs was analyzed [39]. Chapter 4 focuses on elucidating the physical properties of PEG-conjugated LEHs (PEG-LEHs). PEG-LEHs of various sizes, encapsulated Hb concentrations, PEG molecular weights, and ionic strengths of the liposome aqueous core were prepared. Moreover, a reductant was coencapsulated inside PEG-LEHs to reduce metHb formation.
16
The aim of this study was to investigate the influence of these variations in PEG-LEH preparation on the oxygen carrying and binding ability of PEG-LEHs. In addition, the simulated effect of in-vivo osmotic pressure on the integrity of PEG-LEH bilayers was evaluated [53]. In Chapter 5, we introduce a novel cellular HBOC: polymersome encapsulated Hb (PEH) [55]. To our knowledge, this is the first study that explores the potential of polymersomes as an oxygen carrier. PEHs of various sizes, PEG lengths, membrane thicknesses and encapsulated Hb concentrations were prepared, and their oxygen carrying and binding ability were investigated. To suppress Hb oxidation, a reductant was coencapsulated in the PEH aqueous core. In this study, we aim to prove that PEH dispersions exhibit good potential as a cellular HBOC. Chapter 6 is devoted to simulating in-vivo oxygen transport of various oxygen carriers in a capillary and its surrounding tissue under various conditions using a Krogh tissue cylinder model [66]. By analyzing the in-vitro physical properties and in-vivo oxygen transport simulations of LEH, PEG-LEH and PEH dispersions, we aim to select the dispersions, which exhibit the best potential to function as cellular HBOCs. Finally, our recommendations and future plans for the continued development of a reliable cellular HBOC are presented in Chapter 7.
17
CHAPTER 2
MATERIALS AND METHODS
The procedures to prepare unmodified liposome encapsulated Hb (ULEH), PEGconjugated liposome encapsulated Hb (PEG-LEH), and polymersome encapsulated Hb (PEH) dispersions and the methods to measure the physical properties of these vesicle dispersions are presented in this chapter. The following physical properties were investigated: vesicle size distribution, Hb encapsulation efficiency, oxygen affinity (as indicated by the P50 and cooperativity coefficient), and encapsulated metHb level.
2.1 Preparation of Hemoglobin Stock Solution
2.1.1 Purification of Bovine Hemoglobin Bovine erythrocytes were extracted, centrifuged and collected in 3.8% sodium citrate by Animal Technologies, Inc. (Tyler, Texas, USA). Packed bovine erythrocytes were first centrifuged at 6000 rpm for 20 minutes (2-3oC), and the supernatant removed. The cell pellet was washed with three volumes of ice-cold isotonic saline solution (0.95 g NaCl per 100 mL) to one volume of sedimented erythrocytes, this was followed by gentle mixing and swirling for 2-3 minutes. The mixture was centrifuged again at 6000 rpm for
18
20 minutes (2-3oC), and the supernatant discarded. The washing step was repeated twice. Erythrocytes were lyzed by adding two volumes of hypotonic, 15 mOsm phosphate buffer (PB) at pH 7.4 to one volume of washed erythrocytes, followed by gentle mixing and swirling for 2-3 minutes. The mixture was left to stand for one hour in an ice bath. Lyzed erythrocytes were then filtered through a 0.05 µm Minikros Sampler hollow fiber cartridge (Spectrum Laboratories, Inc., Rancho Dominguez, CA, USA) at 4oC. The retained fluid was recycled, and the filtrate (Hb solution) collected. The extracted Hb solution was then filtered through a 50 kDa hollow fiber cartridge (Spectrum Laboratories, Inc.) at 4oC to concentrate the Hb solution. The concentrated Hb solution was then diluted with phosphate buffer (PB) or phosphate buffered saline (PBS) at physiological pH of 7.3 to the desired Hb concentration. PB and PBS were made using salts obtained from Sigma-Aldrich (St. Louis, MO, USA).
2.1.2 Assay of Hemoglobin and Methemoglobin Concentrations To quantify the Hb concentration [39, 70], the Hb solution was diluted to 10 mg/mL (solution A) with deionized water, and the absorbance was measured at 630 nm (L1) against a blank reference (deionized water) using a UV visible spectrophometer (Varian, Inc., Palo Alto, CA, USA). One drop of KCN solution (1 part 10% KCN and 1 part 0.05 M phosphate, pH 7.6) was added to one mL of the Hb solution and mixed. The ensuing reaction converted methemoglobin to cyanomethemoglobin, which does not absorb at 630 nm. After two minutes, the absorbance was read at 630 nm against a blank reference (L2). One mL of solution A was diluted with 9 mL of deionized water, and mixed with one drop of 20% potassium ferricyanide. After two minutes, one drop of 10% 19
KCN was added. The mixture was measured at 540 nm against a reference consisting of 10 mL of deionized water and one drop each of 20% potassium ferricyanide and 10% KCN (L3). The concentration of metHb and Hb were calculated using these formulas:
Concentration of metHb (mM) = (L1-L2)/3.7 x dilution factor of solution A
(2.1)
Total concentration of Hb (mM) = L3/11 x dilution factor of solution A x 10
(2.2)
The purity of the extracted Hb solution was established by size exclusion chromatography under dissociating condition (0.5 M MgCl2 mobile phase), which gave rise to one peak corresponding to αβ dimers resulting from the dissociation of tetrameric Hb. N-Acetyl-L-Cysteine (NAC) was added to the Hb solution to limit metHb formation: 1.2 mM for PEG-LEH dispersions and 5 mM for PEH dispersions. The resulting Hb solution was used when the metHb level was less than 3% for all subsequent experiments.
2.2
Preparation
of
Unmodified
Liposome,
PEG-Liposome
and
Polymersome
Encapsulated Hemoglobin Dispersions
2.2.1 Materials Dimyristoyl-phosphatidylcholine (DMPC), dimyristoyl-phosphatidylglycerol
(DMPG), and cholesterol in a mole ratio of 10:9:1 was used to prepare ULEHs [39]. This combination of lipids was selected based 20 on several reasons. Unsaturated
phosphatidylcholine (PC) was shown to oxidize encapsulated Hb within a few days, therefore saturated PC, DMPC in this study, was found to be most suitable in limiting Hb oxidation [42]. Cholesterol was added to the liposome bilayer in order to overcome several known problems. Intravascularly, cholesterol in the membrane of RBCs transferred to the bilayer of liposomes composed of pure phospholipids due to the cholesterol concentration gradient, resulting in osmotically fragile erythrocytes. Hence, the addition of cholesterol to the bilayer of liposomes reduced the cholesterol concentration gradient, and prevented cholesterol transfer. Moreover, cholesterol enhanced the resistance of liposomes to fusion and lysis [71], and the impermeability of liposomes to small ions [72], such as chloride ions, which are naturally abundant in the blood plasma [16] and have been shown to induce Hb oxidation [38]. Negatively-charged DMPG was incorporated in the bilayer to impart a negative charge to vesicles, since neutral liposomes displayed a strong tendency to aggregate. Moreover, it was observed that infusion of DMPG into mice was safe, and that addition of 5 mole% of DMPG resulted in the most optimum Hb encapsulation efficiency [42]. PEG-LEH dispersions were prepared using a combination of DMPC, cholesterol, dimyristoyl-phosphoethanolamine-methoxy[poly(ethylene glycol)] (DMPE-PEG),
DMPG and α-tocopherol in a mole ratio of 43:40:10:5:2 [53]. Two molecular weights of PEG were used in this study: 550 and 2000 Da. Since lipid bilayers are susceptible to oxidation, α-tocopherol was incorporated in the bilayers to suppress bilayer oxidation [42]. To further suppress Hb oxidation, 1.2 mM of the reductant NAC was coencapsulated inside PEG-LEHs [35]. Phospholipids and cholesterol were obtained
21
from Avanti Polar Lipid, Inc. (Birmingham, AL, USA), while NAC and α-tocopherol were purchased from Sigma-Aldrich. PEH dispersions were prepared using biocompatible poly(butadiene)-
poly(ethylene glycol) (PBD-PEO) diblock copolymers. Hydrophilic PEO is the structural equivalent of PEG [73], and PBD is also known to be bioinert as it is commonly used as the hydrophobic block in biocompatible triblock copolymers [74]. In addition, infusion of empty polymersomes composed of PBD-PEO copolymers into rats showed no toxic effects over an eight-week period [47]. Poly(butadiene(1,4 addition)-b-ethylene oxide) with PBD-PEO molecular weights of 22-12.6 and 5-2.3 kDa, and poly(butadiene(1,2 addition)-b-ethylene oxide) with PBD-PEO molecular weights of 2.5-1.3 and 1.8-0.9 kDa were used in this study [55]. The PBD blocks of the first two copolymers are linear, while the PBD blocks of the last two are branched. Diblock copolymers were purchased from Polymer Source (Dorval, QC, Canada). Five mM of NAC [33, 35] was coencapsulated inside PEHs to suppress metHb formation.
2.2.2 Preparation Procedure ULEH, PEG-LEH and PEH dispersions were prepared using a film rehydration technique, followed by extrusion to regulate vesicle size [39, 53, 55]. Various methods of vesicle preparation, such as reverse-phase evaporation [75], dehydration-rehydration [69, 76], detergent dialysis [77], extrusion [68, 78] and microfluidization [79], have been employed in the past. The last two techniques are gentle on the physical integrity of encapsulated Hb, and therefore do not adversely affect its oxygen binding properties.
22
However, extrusion gives more control over the particle size distribution compared to microfluidization [68, 79], and can be readily scaled up for industrial production. To prepare ULEH, PEG-LEH and PEH dispersions, first 20 mg/mL of phospholipid combination or one mg/mL of copolymer was dissolved in chloroform. One mL of phospholipid or copolymer chloroform solution was then uniformly coated onto the inside wall of a round-bottom glass vial, followed by evaporation of chloroform for 4 hours using a Buchi R-205 rotary evaporator (Buchi Analytical, Inc., New Castle, DA, USA). In our study, we observed that a very thin film of diblock copolymer had to be formed in order to yield polymersomes, otherwise the copolymer molecules selfassembled into a mixture of polymersomes, micelles and solid copolymer precipitates in aqueous solution [55]. The creation of a thin film was achieved on the laboratory-scale by introducing one mg/mL of copolymer in chloroform into a 50 mL round-bottom glass vial prior to chloroform evaporation. The resulting phospholipid or copolymer film was rehydrated with one mL of Hb solution in PB or PBS with or without NAC. Using either PB or PBS as the extrusion buffer created liposome aqueous cores with ionic strengths less than physiological ionic strength (PB) or near-physiological ionic strength (PBS). Polydisperse, giant ULEHs, PEG-LEHs or PEHs were spontaneously formed during the rehydration step. The size of polydisperse vesicle dispersion was regulated via extrusion of the resultant suspensions through polycarbonate membranes starting from a larger pore diameter successively down to the final pore diameter. Ten passes were executed for each successive step, and 25 passes were done at the final step to ensure formation of monodisperse suspensions [39]. Empty unmodified liposomes, PEGylated liposomes and polymersomes were also prepared in an identical way (except that the Hb solution was
23
replaced with PB or PBS with or without NAC) as the controls. Vesicle dispersions were used immediately after preparation. Polycarbonate membranes and extruders were purchased from Avanti Polar Lipids, Inc.
2.3 Vesicle Size Distribution
2.3.1 Introduction It is well known that the size of the oxygen carrier influences its in-vivo circulation half-life [61]. In addition, monodisperse dispersions were shown to have simple exponential clearance kinetics compared to heterogeneous dispersions, thus permitting easier control and manipulation of the vesicle’s circulation half-life [61]. In our study, we employed an asymmetric flow field-flow fractionator (AFFFF) coupled to a multi-angle static light scattering (MASLS) photometer and a differential interferometric refractometer (DIR) to measure the absolute size distribution of liposome and
polymersome dispersions [39]. AFFFF-MASLS-DIR possesses several advantages compared to other techniques which had been used to measure the size of particles in a particle suspension such as: transmission electron microscopy (TEM) [80], sedimentation field-flow fractionation (SFFF) [81], gel filtration chromatography [82], nuclear
magnetic resonance (NMR) spectroscopy [83] and dynamic light scattering (DLS) [84]. In TEM, the dispersion of particles has to undergo several preparation steps, which may destroy or alter the particles and introduce artifacts. Moreover, the measurement is usually conducted in air or vacuum, not in the particle’s original aqueous environment [80]. SFFF is a powerful technique, yet a relatively complex centrifugation apparatus is
24
required [81]. Gel filtration chromatography possesses several unsolved limitations, such as limited adsorption range, long elution times, and lipid adsorption to the column [82]. DLS [84] and NMR spectroscopy [83] are gentle towards particles, and can yield accurate information about average particle size. However, prior knowledge of the particle’s geometry is required in order to determine particle size. Thus, the size distribution and particle shape can not be measured simultaneously and unambiguously. Previous attempts at using DLS to measure particle size were performed in batch mode. Hence, DLS is most suitably applied to relatively monodisperse samples. The idea of combining cross-flow FFF with DLS was explored as a mean of measuring the size distribution of particles in a dispersion [85]. The physical separation of the particle dispersion using FFF is followed by collection of each of the eluting fractions and their subsequent analysis in batch mode by DLS. In addition to the tedious measurement of each collected fraction, there are no direct means of measuring the actual particle size of each eluting fraction. Hence, the result has been qualitative at best, and depends critically on the reproducibility of the operating conditions [85]. In contrast, classical light scattering or MASLS measurements [86] can provide more information about the size and shape of particles in suspension versus DLS without prior information of the particle’s geometry. The angular distribution of scattered light makes it possible for MASLS to determine the shape (geometry) of particles in solution, the root mean square radius of gyration, and the weight averaged molecular weight of the dispersion. Combined with AFFFF and DIR, the AFFFF-MASLS-DIR experimental setup can measure the differential and cumulative number fraction size distributions of vesicle dispersions. Various moments of the size distribution, such as weight-, number-,
25
and z-averaged, can be calculated readily from the number fraction distribution [85, 8789].
2.3.2 Experimental Procedure An Eclipse AFFFF coupled in series to an 18-angles Dawn EOS MASLS photometer equipped with a linearly polarized 30 mW gallium-arsenide laser operating at 690 nm, and an Optilab DSP DIR (Wyatt Technology Corp., Santa Barbara, CA, USA) were used to measure the size distribution of unmodified liposome, PEG-liposome and polymersome dispersions. During a typical experiment, AFFFF would separate the particle dispersion based on particle size. The resultant eluent was then sent through the flow cell of the MASLS photometer, which continuously measured the particles’ Raleigh ratio at 16 angles. The eluent subsequently went through the flow cell of the DIR, which continuously measured the molecules’ refractive index [39]. Typical chromatograms of ULEH, PEG-LEH and PEH dispersions monitored using the 90o light scattering and differential refractive index detectors are shown in Figures 2.1 and 2.2. The chromatograms of ULEH, PEG-LEH and PEH dispersions show a definite separation between unencapsulated Hb and vesicles. The first peak corresponds to unencapsulated Hb fractionated from the suspension, while the second peak corresponds to ULEHs, PEG-LEHs or PEHs in the suspension. Light scattering spectra were analyzed using the ASTRA™ software (Wyatt Technology Corp.). The mobile phase for all experiments was PBS at physiological pH = 7.3 and T = 37oC, filtered through 0.2 µm filters. This set-up is capable of measuring particles with molecular weights ranging from ~500 Da to 108 Da, and mean square radii ranging from 10 nm to 500 nm [39]. 26
Figure 2.1. Staggered chromatograms of the DIR (dotted line) and 90o light scattering (solid line) detector signals for the following dispersions: ULEHs extruded through 100 nm pore diameter membranes in PB (panel A) and PBS (panel B), at an initial Hb concentration of 200 mg/mL; PEG-LEHs grafted with 2000 Da PEG, extruded through 200 nm pore diameter membranes in PB (panel C) and PBS (panel D), at an initial Hb concentration of 300 mg/mL; PEG-LEHs grafted with 550 Da PEG, extruded through 400 nm pore diameter membranes in PB (panel E) and PBS (panel F), at an initial Hb concentration of 300 mg/mL.
27
Figure 2.2. Staggered chromatograms of the DIR (dotted line) and 90o light scattering (solid line) detector signals for the following dispersions: PEHs composed of 22-12.6 kDa PBD-PEO copolymer (panel A), 5-2.3 kDa PBD-PEO copolymer (panel B), 2.5-1.3 kDa PBD-PEO copolymer (panel C) and 1.8-0.9 kDa PBD-PEO copolymer (panel D), extruded through 400 nm pore diameter membranes in PBS, at an initial Hb concentration of 300 mg/mL.
2.3.3 Theoretical Background
2.3.3.1 Flow field-flow fractionation theory In the AFFF fractionator system, the channel flow was run parallel to the polymer membrane, while the cross flow was perpendicular to the membrane. The trapezoidal AFFFF channel was bounded on the top by a glass plate and on the bottom by a membrane with a 10 kDa-MWCO [90-93]. The AFFFF channel was pre-equilibrated by initiating a cross flow of 3 mL/min, and a channel flow of 0 mL/min for a total of 3 28
minutes. During the focus-inject mode of operation, the sample was injected into the channel, where it was subjected to a cross flow of 3 mL/min and a channel flow of 0 mL/min. Particles exposed to the cross flow accumulated towards the membrane surface, where they rapidly equilibrated via molecular diffusion into a thin ellipsoidal shaped sample front. Instead of diluting our sample, we manipulated injection time (with the injection flow rate set at 0.1 mL/min) to achieve the most optimum light scattering and refractometer signals. The injected football-shaped sample volume was then reduced by applying a cross flow of 3 mL/min and a channel flow of 0 mL/min for 3 minutes. After this last focusing step, elution of the sample begun, and the ellipsoidal shaped sample region was exposed to a channel flow of 1 mL/min and a cross flow, which was linearly decreased from 1 mL/min to 0 ml/min over a 30 minute interval. The elution was continued for another 10 minutes without cross flow. The membrane does not permit the passage of the particles, therefore the particles remain permanently confined in the channel. The cross flow imposes a Stokes drag force that impinges particles against the membrane, thereby building up a higher particle concentration at the membrane than elsewhere. The resulting particle concentration gradient at the membrane gives rise to a diffusive flux in the upward direction. The effect of cross flow on the particle elution time is dependent on the particle diameter. The drag on smaller particles is less and their Brownian motion is greater, thus smaller particles will experience a higher elevation from the membrane (accumulation wall) and be transported faster through the channel due to the parabolic velocity profile of the channel flow [88, 90-94]. This behavior is characterized by an exponential distribution of particles near the accumulation wall. The particle’s elution time, tr, depends on νc 29
(channel flow velocity), νx (cross flow velocity), the channel thickness, w, and the particle diffusion coefficient, Dv via the following equation [85, 87-89]:
tr ≈
w 2υ x πηw 2υ x = d 6 Dvυ c 2kTυ c
(2.3)
where η is the solvent viscosity, k is Boltzmann’s constant, T is the absolute temperature and d is the particle’s diameter. Since the particle’s diffusion coefficient is inversely proportional to the particle’s diameter, the retention time is directly proportional to the particle’s diameter.
2.3.3.2 Light scattering theory In the MASLS flow mode of operation, light scattered from the particle dispersion was detected with 16 detectors. The light scattering intensity profile was recorded as a function of time at a rate of one full spectra every second. Here, the time it takes to collect one complete spectra will be called a slice. The amplitude of the scattering wave vector, q, is defined as [84]:
q = (4π n/ λ 0 )sin(θ / 2)
(2.4)
where n = 1.3316 is the refractive index of the buffer solution, θ is the scattering angle, and λ 0 = 690 nm is the wavelength of the incident light beam in vacuum.
30
The excess Rayleigh ratio, Rθ , describes the light scattered from a particle dispersion at a particular angle. The following expression represents the scattering due to particles, which is equal to the difference in scattering between the particles in solution and buffer solution.
⎛ 2 Iθ − Iθ ,solvent ⎞ ⎟ Rθ = r ⎜ ⎝ Io V ⎠
(2.5)
Here, r is the distance between the scattering volume and the detector, Iθ is the scattered intensity of the solution, Iθ ,solvent is the scattered intensity of the solvent, Io is the intensity of the incident beam, and V is the volume of the scattering medium. This definition of
Rθ is very robust, since it corrects for any stray light present in the scattering volume
along with any fluctuations in laser power. Determination of the size of a monodisperse system of particles can be derived from the following expression proposed by Zimm [95, 96] for a vertically polarized monochromatic light source:
R = MP(θ ) − 2A2 cM 2 P 2 (θ ) Kc
(2.6)
where K = 4π 2 n 2 (dn / dc ) /( λ o 4 NA )
2
(2.7)
K is the optical constant of the hemoglobin solution, N A is Avogadro’s number which is equal to 6.023 x 1023, c is the concentration of hemoglobin in solution, M is the molecular weight of the solute, P(θ ) is the theoretically derived form factor which is a 31
function of the size, shape, and structure of the particle, and A2 is the second virial coefficient. Since Zimm’s expression only takes into account single contacts between particles, it is typically used to model particle scattering in dilute solutions. To measure the size and shape of particles, values of K and c are not needed, instead equation 2.6 can be normalized with the light scattering intensity of the 90o
detector to yield:
R(θ ) P(θ ) ≈ o R(90 ) P(90o )
(2.8)
This equation is used to regress characteristic dimensions of scatterers from the experimental light scattering spectra. For our controls, which were empty liposomes and polymersomes, the shape factor of an infinitely thin spherical shell was used:
⎡ sin qR ⎤ P(θ ) = ⎢ ⎥ ⎣ qR ⎦
2
(2.9)
where R is the shell radius. For LEHs and PEHs, the shape factor of a solid sphere was used:
⎤ ⎡ 3 P(θ ) = ⎢ 3 3 (sin qR − qR cos qR )⎥ ⎦ ⎣q R
2
(2.10)
where R is the radius of the sphere.
32
If the molar mass and concentration of each slice comprising the elution peak is known, it becomes a trivial process to calculate the following moments of the root mean square (rms) radius of each eluting peak: Number-averaged mean square radius:
∑(
Rn = r 2
n
=
ci r2 ) Mi i c ∑ i Mi
(2.11)
Weight-averaged mean square radius:
Rw = r
2 w
=
2 ∑ (c i r i )
∑ ci
(2.12)
Z-averaged mean square radius:
Rz = r
2 z
=
2 ∑ (c i M i r i )
∑ (c i M i )
(2.13)
Here Mi and r 2
i
represent the molar mass and mean square radius of the ith
slice. In our case, we replaced the mean square radius with the radius derived from either the sphere or thin shell form factor model. The degree of polydispersity of the eluting fraction can be easily calculated via either one of the following definitions: Rw/Rn or
Rz/Rn. The last expression weights larger particles more than the former.
33
Knowledge of the hydrodynamic radius of each eluting slice also allows us to calculate both the cumulative and differential size distribution. The weight fraction of particle dispersion having a mean square radius less than r 2 , W ( r 2 ) , is defined as:
∑
W ( r2 ) =
r
2 '
c
2
< r
r2
'
∑
all r 2
'
c r2
(2.14)
where c
r
2 '
is the mass concentration of the slice having a mean square radius of r 2 .
'
The differential size distribution is defined as:
x( r ) =
2
dW ( r 2 ) d (log r 2 )
(2.15)
2.3.3.3 Differential interferometric refractometer DIR is used to measure the change in refractive index of the Hb solution with respect to the solvent of the ith slice ( ∆ni ). This can be related to the Hb concentration of the ith slice (ci) via the differential refractive increment of the solution with respect to the change in solute concentration, dn / dc . Therefore ∆ci = ∆ni /(dn / dc ) , where ∆ci = ci , since the differential refractive index baseline represents pure solvent. At T = 25oC and
λ0 = 690 nm, we measured a dn / dc of 0.185 mL/g for Hb. Since the refractive index is a
34
strong function of wavelength and temperature, all measurements of dn / dc and static light scattering measurements were performed at the same wavelength and temperature.
2.4 Hemoglobin Encapsulation Efficiency
2.4.1 Introduction In addition to economical considerations, cellular HBOCs with high Hb encapsulation efficiencies are required to prevent overloading the patient’s RES system with phospholipids or copolymers [97]. Hence, an accurate method of quantifying Hb encapsulation is critical. The most common method for measuring the Hb encapsulation efficiency of a vesicle suspension involved disrupting the vesicle membrane with a detergent solution before quantifying the Hb content photometrically [76, 98, 99]. This process exposed Hb to detergent, which denatured Hb, and thus led to incomplete removal of turbidity [69]. Moreover, since this method assumed that the total volume of membrane lipids was negligible compared to the total volume of particles, this approach overestimated the Hb encapsulation efficiency [99]. A method to determine Hb encapsulation efficiency without permeabilizing the lipid bilayer of LEH particles was introduced by Brandl et al. [69]. Unentrapped Hb was first removed by size exclusion chromatography on Ultragel columns, followed by photometric assay of the unentrapped Hb concentration, where the entrapped Hb concentration could be calculated. To quantify the lipid content, the lipid was radiolabelled with 14C and quantified using a scintillation counter. Although this technique was used to overcome the weaknesses of the first method, Brandl et al. [69] stated that the photometric assay correction for solution
35
turbidity was not applicable to LEH dispersions. Farmer et al. [42] proposed another technique, which required centrifuging the vesicle suspension, and approximating the entrapped Hb concentration by multiplying the Hb concentration in the aliquot with the percent hemocrit. However, this method assumed that the total volume of lipids and the intervesicular spaces were negligible, and led to overestimation of the encapsulated Hb concentration. In a previous publication [39], we developed a novel technique for measuring the Hb encapsulation efficiency of ULEHs, PEG-LEHs and PEHs using AFFFF-MASLSDIR. In contrast to the previous methods, we demonstrated that our new technique is convenient, gentle on the integrity of encapsulated Hb and vesicle membrane, and provides an accurate measurement of Hb encapsulation efficiency. In addition, the oxygen carrying particles can be measured in their original aqueous environment. Note that this technique is also applicable for measuring the loading capacity of any in-vivo drug delivery systems.
2.4.2 Experimental Procedure To calculate Hb encapsulation efficiency, we measured the DIR signal of eluting unencapsulated Hb, which was separated from ULEHs, PEG-LEHs or PEHs by the AFFFF, and the DIR signal of the reference (pure Hb solution of Hb concentration identical to the initial Hb concentration of ULEHs, PEG-LEHs or PEHs). Since the area under the DIR peak corresponding to unencapsulated Hb is proportional to the unencapsulated Hb concentration, the Hb encapsulation efficiency can be calculated as follows [39]:
36
⎡ unencapsulated Hb area ⎤ Encapsulation Efficiency = ⎢1 − ⎥ × 100% reference area ⎣ ⎦
(2.16)
The area under the peak was calculated using the Scientist Software (Micromath Scientific Software, Inc., St. Louis, MO, USA) with an error controlled Runge-Kutta integrator. Staggered chromatograms of the DIR peaks of the reference Hb solution and unencapsulated Hb separated from ULEH, PEG-LEH or PEH dispersions are shown in Figures 2.3 and 2.4. When ULEH, PEG-LEH or PEH dispersions were fractionated, a definite separation between unencapsulated Hb and ULEHs, PEG-LEHs or PEHs was observed (see Figures 2.1 and 2.2). The purity of the Hb solution was established earlier; hence, we are convinced that no traces of vesicles or contaminants contribute to the unencapsulated Hb signal. However, Hb-loaded vesicles and particles in the literature [69, 78, 100] were made using different phospholipid or polymer concentrations than our formulations. Therefore, we calculated the weight ratio of entrapped Hb to either phospholipids or copolymers (R) comprising the membrane, such that the Hb loading capacities of ULEHs, PEG-LEHs and PEHs measured in this work can be compared to the Hb loading capacities reported in the literature [69, 78, 100]. R was calculated by dividing the entrapped Hb concentration in mg/mL with the phospholipid or copolymer concentration in mg/mL. The entrapped Hb concentration was calculated by multiplying the encapsulation efficiency with the initial Hb concentration used to prepare ULEH, PEGLEH, and PEH dispersions.
37
Figure 2.3. Staggered chromatograms of DIR signals of the Hb reference (solid line) and unencapsulated Hb (dotted line) of: ULEHs extruded through 400 nm pore diameter membranes in PB (panel A) and PBS (panel B), at an initial Hb concentration of 150 mg/mL; PEG-LEHs grafted with 2000 Da PEG, extruded through 200 nm pore diameter membranes in PB (panel C) and PBS (panel D), at an initial Hb concentration of 300 mg/mL; PEG-LEHs grafted with 550 Da PEG, extruded through 400 nm pore diameter membranes in PB (panel E) and PBS (panel F), at an initial Hb concentration of 300 mg/mL.
38
Figure 2.4. Staggered chromatograms of DIR signals of the Hb reference (solid line) and unencapsulated Hb (dotted line) of: PEHs composed of 22-12.6 kDa PBD-PEO copolymer (panel A), 5-2.3 kDa PBD-PEO copolymer (panel B), 2.5-1.3 kDa PBD-PEO copolymer (panel C) and 1.8-0.9 kDa PBD-PEO copolymer (panel D), extruded through 400 nm pore diameter membranes in PBS, at an initial Hb concentration of 300 mg/mL.
2.5 Oxygen Binding Properties Freshly prepared ULEH, PEG-LEH or PEH dispersions were dialyzed overnight in PBS (pH 7.3) at 2-3oC using 100 kDa molecular weight cutoff dialysis bags (Spectrum Labs, Rancho Dominguez, CA, USA) at a sample:PBS v/v ratio of 1:1000 to separate unencapsulated Hb from the vesicle suspension. In preparation for oxygen affinity measurements, 0.5-4 mL of dialyzed sample was lightly mixed with 20 µL of Additive-A and 10 µL of Anti-Foaming Agent, and diluted with Hemox solution to a total volume of 39
5 mL [39, 53, 55]. The oxygen dissociation curve of ULEH, PEG-LEH and PEH dispersions was measured using a HemoxTM-Analyzer from TCS Scientific Corp (New Hope, PA, USA) at physiological temperature (T=37oC). The P50 is defined as the partial pressure of oxygen when 50% oxygen saturation is achieved (Y = 0.5) and was measured from the oxygen dissociation curve [16]. The cooperativity coefficient (n) was regressed using the Hill equation [24]. It was shown that the sigmoidal shape of the oxygen dissociation curve can be described by the following equilibrium reaction:
Hb(O2 ) n ↔ Hb + nO2
(2.17)
Here, n represents the number of oxygen molecules that bind to Hb to form oxyhemoglobin (Hb(O2)n). The rate expression for the appearance of Hb in an equilibrium state is:
d [ Hb] = k1 [ Hb(O2 ) n ] − k −1[ Hb][O2 ]n = 0 dt
(2.18)
This equation can be rearranged to:
[ Hb(O2 ) n ] = K [ Hb][O2 ] n
(2.19)
where K is k-1/k1. The fraction of saturated Hb (Y) is defined as the following:
40
Y=
[ Hb(O2 ) n ] [ Hb(O2 ) n ] + [ Hb]
(2.20)
This equation can be simplified using equation 2.19 and expressed in terms of partial pressure of oxygen (pO2) using the Henry’s law. Henry’s law is presented below:
pO2 = H oxygen [O2 ]
(2.21)
where Hoxygen is the Henry’s law constant for oxygen in blood and [O2] is the concentration of dissolved oxygen in blood. Thus, equation 2.20 becomes:
Y=
K [O2 ] n ( pO2 ) n = n 1 + K [O2 ]n H oxygen + ( pO2 ) n K
(2.22)
Equation 2.22 is called the Hill equation and can be rearranged to:
H oxygen Y ln = n ln( pO2 ) − n ln 1−Y K
n
(2.23)
The plot of ln(Y/1-Y) versus ln(pO2) is linear with a slope equals to the cooperativity coefficient (n).
41
2.6 Encapsulated Methemoglobin Level To measure the metHb level inside ULEH, PEG-LEH and PEH particles, freshly extruded ULEHs, PEG-LEHs or PEHs were dialyzed overnight in PBS at 2-3oC (100 kDa MWCO) to remove unencapsulated Hb from the dispersions. The dialyzed sample was then vigorously mixed with a detergent solution at a concentration slightly above the detergent’s critical micellar concentration for a few seconds at room temperature to lyze the vesicle membranes. Octyl-β-D-glucopyranoside solution (103 mM) in PBS was used to lyze ULEHs and PEG-LEHs [42], while Triton-X solution (15 mM) in PBS was used to lyze PEHs [101]. The encapsulated metHb concentration was immediately measured with UV visible spectrophotometry (Varian, Inc.) according to the method described in section 2.1.2. In this case, deionized water was replaced with PBS buffer. The total concentration of entrapped Hb in the dispersion volume was calculated by multiplying the Hb encapsulation efficiency with the initial Hb concentration. The metHb level was calculated using the following formula:
MetHblevel (%) =
MetHb concentration (mM ) × 100 Total entrapped Hb concentration (mM )
(2.24)
42
CHAPTER 3
UNMODIFIED LIPOSOME ENCAPSULATED HEMOGLOBIN DISPERSIONS
3.1 Introduction Hb was encapsulated in the aqueous core of unmodified liposomes to create unmodified LEH (ULEH) dispersions. ULEHs were prepared using five initial Hb concentrations (300, 200, 150, 100 and 50 mg Hb per mL of buffer) and five extrusion pore diameter membranes (400, 200, 100, 80, and 50 nm) [39, 54]. The ultimate goal of this work is to infuse ULEHs into animals or humans as a cellular HBOC. In this setting, the infused ULEH dispersions will be subjected to physiological osmotic pressures. To mimic this osmotic environment in-vitro, we examined the size distributions of ULEHs in phosphate buffered saline (PBS), which mimics the salinity of blood plasma, at physiological pH and temperature of 7.3 and 37oC, respectively. The osmotic pressure inside ULEH particles extruded in PBS is expected to be slightly higher than the exterior environment (PBS) due to the presence of encapsulated Hb. In this case, water will diffuse into ULEH particles in order to equilibrate the osmotic pressure gradient, which will cause ULEH particles to swell. Swelling will render ULEH particles fragile and susceptible to rupture by intravascular shear forces. If ULEHs are lyzed, cytotoxic Hb molecules will be released into the blood circulation. Because of this concern, we also
43
investigated phosphate buffer (PB), which has an ionic strength less than blood plasma, as an alternative extrusion buffer. The lower ionic strength of PB is expected to decrease the osmotic pressure inside ULEHs extruded in PB compared to the osmotic pressure of the exterior environment (PBS), causing these ULEHs to shrink, and therefore preventing lysis of ULEHs in-vivo when ULEHs are infused into animals or humans. The results of this study will contribute to the development of mechanically strengthened ULEHs, which can withstand osmotically induced lysis in the blood stream. Furthermore, we investigated the influence of the initial Hb concentration, extrusion membrane pore diameter, and ionic strength of the extrusion buffer on the physical properties of ULEH dispersions. The physical properties of interest were: vesicle size distribution, Hb encapsulation efficiency, oxygen affinity (as indicated by the P50 and cooperativity coefficient), and encapsulated metHb level. For convenience, ULEHs extruded in PB will be referred to as PB-LEHs and those extruded in PBS will be referred to as PBS-LEHs.
3.2 Results and Discussion
3.2.1 Average Radius Number- (Rn), weight- (Rw) and z-averaged (Rz) radii and polydispersity indices (Rw/Rn and Rz/Rn) of ULEHs and their controls extruded in PB and PBS are presented in Table 3.1. Regardless of the initial Hb concentration, extrusion buffer and membrane pore diameter, extrusion through 400 nm pore diameter membranes produced unmodified liposomes smaller than the pore size, extrusion through 200 nm pore diameter membranes produced liposomes with diameters similar to the pore size, and extrusion
44
through 100 nm, 80 nm, and 50 nm pore diameter membranes produced liposomes larger than the pore size. This phenomenon can be explained by “equilibrium vesicle” theory. This theory states, there exists an “equilibrium vesicle” size which is the result of mixing entropy and molar bending energy of the vesicle suspension [102]. The formation of vesicles relies on the fact that the tectons (in this case, the phospholipids) energetically prefer a parallel molecular arrangement or the formation of bilayers. However, when the bilayers are so thick and the energy loss due to surface tension at the ends of the bilayers (due to exposing hydrophobic parts of the tectons to an aqueous environment) are so great, the bilayers will bend and form a closed vesicle. Bending of bilayers to form a closed vesicle requires bending energy. Note that although the bending energy per vesicle does not depend on the vesicle size, the molar bending energy does. Therefore, the molar bending energy drives toward formation of a small number of vesicles. However to maximize the disorder of the system, the mixing entropy of the system drives toward formation of a large number of vesicles, which consequently, favors formation of smallsized vesicles. Hence, the energy constraint defines a minimum vesicle size, while the entropy constraint defines a maximum vesicle size [102]. In this case, formation of ULEHs and empty liposomes with diameters larger than 100 nm and less than 300 nm were favored. It appeared that when liposomes smaller than 100 nm in diameter were created, the liposomes fused together to form larger liposomes. Indeed, this is an advantage, since liposomes with diameters less than 100 nm were shown to be rapidly filtered out of the systemic circulation by the capillaries and kidney microtubules [25], while liposomes with diameters in the range of 160-220 nm were observed to exhibit long intravascular circulation lifetimes [62].
45
TABLE 3.1
NUMBER- (Rn), WEIGHT- (Rw), Z-AVERAGED (Rz) RADII AND POLYDISPERSITY INDICES OF ULEH AND CONTROL DISPERSIONS EXTRUDED IN PB AND PBS
Extrusion buffer: PB Initial [Hb] (mg/mL) 300 Pore Radius (nm) 200 100 50 40 25 200 100 50 40 25 200 100 50 40 25 Rn (nm) 118.3 101.1 85.6 63 47 112.3 77.8 65.5 52.1 49.7 128.3 78.3 85.7 82.2 88.9 Rw Rz Rw/Rn Rz/Rn 1.01 1.06 1.08 0.98 1.06 1.02 1.02 1.01 1.04 1.04 1.06 1.02 1.07 1.08 1.05 1 1.08 1.13 0.98 1.11 1.03 1.04 1.04 1.08 1.08 1.09 1.04 1.11 1.12 1.09 Rn (nm) 130.1 117.8 79.3 56.3 50 139.4 160.5 88.4 66.3 90.2 154.8 105.4 71.8 64.2 53.2
Extrusion buffer: PBS Rw (nm) 129.9 118.6 80.3 57.9 52.1 141.6 160.8 103.1 67.1 103.2 159.2 106 73.7 71.7 55 Rz (nm) 129.7 119 81.1 59.9 53.8 141.8 159.8 109.2 67.7 107.3 160.8 106.3 75 74.7 56.6 Rw/Rn Rz/Rn 0.59 1.01 1.71 1.58 1.88 0.56 0.78 1.31 1.3 1.99 0.64 0.78 1.71 2.06 3.56 0.6 1.07 1.85 1.55 2 0.57 0.79 1.32 1.36 2.08 0.68 0.8 1.83 2.21 3.74
200
150
(nm) (nm) 119.1 118.7 107.4 109 92.4 96.5 61.9 61.9 49.9 52.2 114.5 115.9 79 80.6 66.1 68 54.2 56.1 51.9 53.5 136.4 140.3 80.1 81.5 91.5 94.8 88.5 92 93.5 96.5
46
TABLE 3.1
(Continued)
PB as extrusion buffer Initial Pore Rn [Hb] Radius (mg/mL) (nm) (nm) 100 200 121.5 100 90.4 50 59.5 40 56 25 47.1 50 200 134.2 100 88.9 50 67.3 40 50.6 25 52.1 0 200 126.3 100 88.2 50 67.3 40 54.6 25 51.3 Rw (nm) 121.7 91.5 62.1 58.1 50.6 136.9 90.4 69.3 50.8 53.9 128.9 88.6 67.1 54.8 52.2 Rz (nm) 121.8 92.2 64 59.7 53.2 138.6 91.4 71.2 52.3 55.6 129.8 88.5 67 55.1 52.9 Rw/Rn 1 1.01 1.04 1.04 1.07 1.02 1.02 1.03 1 1.03 1.02 1 1 1 1.02 Rz/Rn 1 1.02 1.08 1.07 1.13 1.03 1.03 1.06 1.03 1.07 1.03 1 1 1.01 1.03 Rn (nm) 137.3 104.2 81.7 66.5 53.2 146.1 109.6 77 62.8 57.1 125.4 83.2 75.7 70.3 50.1
PBS as extrusion buffer Rw (nm) 139.6 105.2 82.7 67.7 53.4 147.9 110.5 77.9 63.8 57.3 125.8 83.6 75.6 70.9 53 Rz (nm) 140.5 105.7 83.5 68.6 53.7 148.9 110.8 78.6 64.6 57.7 124.2 83.6 75.4 71.2 54.5 Rw/Rn Rz/Rn 0.61 0.9 1.19 1.4 1.88 0.67 0.89 1.35 1.27 2.08 1 1 1 1.01 1.06 0.61 0.92 1.24 1.45 2.02 0.68 0.9 1.39 1.27 2.16 0.99 1 1 1.01 1.09
47
3.2.2 The Effect of Osmotic Pressure on the Size of Unmodified Liposomes Encapsulated Hemoglobin The osmotic pressure gradient was calculated by subtracting the osmotic pressure inside the ULEH particle from the osmotic pressure outside the ULEH particle. A positive osmotic pressure gradient states that the osmotic pressure outside the ULEH particle is higher, and vice versa. Table 3.2 displays the calculated osmotic pressure gradients of ULEHs extruded in PB and PBS. We compared the measured mean radii of ULEHs extruded in PB versus PBS, and concentrated our analysis on vesicles extruded through the three largest membrane pore radii (200, 100 and 50 nm), since ULEHs extruded through smaller membrane pore radii fused together to form larger liposomes. We observed that PB-LEHs were generally smaller than PBS-LEHs. Since the osmotic pressure outside PB-LEH particles was higher, water diffused out of these liposomes, and resulted in shrinkage of PB-LEH particles. In contrast, a negative osmotic pressure gradient caused PBS-LEH particles to slightly swell. The osmotic pressure gradients of PB-controls and PBS-controls were 5,880.19 mmHg and 0 mmHg, respectively. As expected, PB-LEHs generally possessed larger mean radii than PB-controls, since the osmotic pressure gradient of PB-LEH particles were less than the osmotic pressure gradient of PB-controls. The mean radii of PBS-LEHs were larger than PBS-controls, since the osmotic pressure gradients of PBS-LEHs were negative. The differences in the osmotic pressure gradients between PB-LEH dispersions and PB-control dispersions ranged from 30 to 900 mmHg, while the differences in the osmotic pressure gradients between PBS-LEH dispersions and PBS-control dispersions ranged from 5 to 815 mmHg.
48
TABLE 3.2
OSMOTIC PRESSURE GRADIENTS (∆P) BETWEEN THE AQUEOUS CORE OF ULEHS AND THE EXTERIOR ENVIRONMENT (PBS)
Initial [Hb] Pore Radius (mg/mL) (nm) 300 200 100 50 200 100 50 200 100 50 200 100 50 200 100 50
PB-LEHs ∆P (mmHg) 4,965.40 5,651.55 5,699.40 5,624.67 5,535.61 5,727.33 5,561.68 5,663.12 5,652.73 5,525.89 5,857.88 5,668.37 5,850.73 5,835.87 5,862.56
PBS-LEHs ∆P (mmHg) -815.44 -635.7 -486.15 -186.58 -186.23 -192.94 -69.09 -97.06 -68.17 -141.96 -140 -95.95 -14.18 -12.96 -5.08
200
150
100
50
The osmotic pressure gradient of PBS-LEHs became more positive as the initial Hb concentration decreased. However, the measured mean radii of PBS-LEHs were not affected by the initial Hb concentration. Similarly, the osmotic pressure gradient of PBLEHs increased with decreasing initial Hb concentration, and the measured mean radii of PB-LEHs were not influenced by the initial Hb concentration. In section 3.2.4, the Hb encapsulation efficiencies of PB-LEHs and PBS-LEHs increased with increasing initial 49
Hb concentration. However, PB-controls and PBS-controls possessed similar mean radii, although the difference in the osmotic pressure gradients of both controls was 5,880.19 mmHg. Hence, it appeared that encapsulation of Hb inside liposomes made the liposomes more susceptible to the influence of osmotic pressure gradients. It was previously observed that encapsulated Hb molecules formed complexes with lipid bilayers, which was subsequently followed by intercalation of Hb molecules into the lipid bilayers [103]. Hb complex formation and intercalation were primarily due to hydrophobic interaction between the hydrophobic portions of the Hb protein and the hydrophobic lipid membrane. Here, we speculated that the interaction between encapsulated Hb molecules and the lipid bilayer caused ULEHs to become more susceptible to the influence of osmotic pressure gradients. Hence, increased Hb encapsulation is expected to render ULEHs fragile in the blood circulation.
3.2.3 Size Distribution The differential and cumulative size distributions of ULEH dispersions extruded in PB and PBS are shown in Figures 3.1 and 3.2, respectively. The size distributions of their corresponding controls are shown in Figure 3.3. PB-LEH dispersions prepared with an initial Hb concentration of 50 mg Hb/mL exhibited monodisperse size distributions with distribution widths of ~20 nm, regardless of the membrane pore diameter utilized (Figure 3.1.A). Monodisperse size distributions were observed in Figure 3.2.A (PBSLEHs, 50 mg/mL initial Hb concentration). PBS-LEHs extruded through 80 and 50 nm pore diameter membranes exhibited size distribution widths of 10 nm, while PBS-LEHs extruded through 400, 200 and 100 nm pore diameter membranes exhibited size
50
distribution widths of ~20 nm. The size distributions of PB-LEHs and PBS-LEHs (Figures 3.1.A and 3.2.A) appeared to be Gaussian.
Figure 3.1. Size distributions of ULEHs extruded in PB. Differential and cumulative distributions of PB-LEHs prepared with an initial Hb concentration of 50 mg Hb/mL are shown in panels A and B, respectively, 100 mg Hb/mL in panels C and D, 150 mg Hb/mL in panels E and F, 200 mg Hb/mL in panels G and H, and 300 mg Hb/mL in panels I and J. The following symbols represent the various membrane pore diameters used: 400 nm (– – –), 200 nm (– · · –), 100 nm (——), 80 nm (········), and 50 nm (– · –). 51
Figure 3.2. Size distributions of ULEHs extruded in PBS. Differential and cumulative distributions of PBS-LEHs prepared with an initial Hb concentration of 50 mg Hb/mL are shown in panels A and B, respectively, 100 mg Hb/mL in panels C and D, 150 mg Hb/mL in panels E and F, 200 mg Hb/mL in panels G and H, and 300 mg Hb/mL in panels I and J. The following symbols represent the various membrane pore diameters used: 400 nm (– – –), 200 nm (– · · –), 100 nm (——), 80 nm (········), and 50 nm (– · –).
52
Figure 3.3. Size distributions of unmodified controls, which are empty liposomes. Differential and cumulative distributions of controls extruded in PB are shown in panels A and B, respectively, controls extruded in PBS in panels C and D. The following symbols represent the various membrane pore diameters used: 400 nm (– – –), 200 nm (– · · –), 100 nm (——), 80 nm (········), and 50 nm (– · –).
With an initial Hb concentration of 100 mg/mL, PB-LEHs extruded through 400 nm pore diameter membranes exhibited a very monodisperse size distribution with a distribution width of only a few nm. However, PB-LEHs extruded through 200, 100, 80 and 50 nm pore diameter membranes exhibited wider size distributions widths, with distribution widths of ~20 nm (Figure 3.1.C). If ULEHs were extruded in PBS at an initial Hb concentration of 100 mg/mL (Figure 3.2.C), we noticed narrower size distributions regardless of the membrane pore diameter utilized, with distribution widths of 10 - 20 nm. For both PB-LEHs and PBS-LEHs, the size distributions were Gaussian in shape. When the initial Hb concentration was increased to 150 mg/mL, the size distributions of PB-LEHs widened with distribution widths of ~40 nm, except for PBLEHs extruded through 200 nm pore diameter membranes (Figure 3.1.E). PB-LEHs 53
extruded through 200 nm pore diameter membranes exhibited the narrowest distribution with a distribution width of 20 nm. All of these size distributions were Gaussian in shape, except for the size distribution of PB-LEHs extruded through 80 nm pore diameter membranes. Even the size distribution of PB-LEHs extruded through 50 nm pore diameter membranes appeared to be slightly skewed. Presumably, this was caused by intervesicular fusion, which was observed for liposome dispersions extruded through membranes smaller than 100 nm in diameter. The size distributions of PBS-LEHs prepared using an initial Hb concentration of 150 mg/mL generally exhibited narrower distribution widths of 20 nm, except for PBS-LEHs extruded through 400 nm pore diameter membranes, which exhibited a distribution width of ~40 nm (Figure 3.2.E). Moreover, all PBS-LEH size distributions were Gaussian in shape, unlike PB-LEHs. The size distributions of PB-LEHs prepared with an initial Hb concentration of 200 mg/mL were Gaussian with size distribution widths of ~20 nm, except for PB-LEHs extruded through 400 nm pore diameter membranes (Figure 3.1.G). PB-LEHs extruded through 400 nm pore diameter membranes exhibited the broadest size distribution width of ~30 nm. The presence of salt in the extrusion buffer (Figure 3.2.G) widened the size distributions. PBS-LEHs prepared with an initial Hb concentration of 200 mg/mL and extruded through 200 nm pore diameter membranes exhibited a distribution width of ~20 nm, while PBS-LEHs extruded through 400, 100, 80 and 50 nm pore diameter membranes generally possessed wider size distributions with distribution widths ranging from 40 to 55 nm. Although all size distributions were generally Gaussian in shape, PBSLEHs extruded through 100 nm pore diameter membranes displayed a slightly skewed size distribution.
54
PB-LEHs prepared using the highest initial Hb concentration (300 mg/mL) exhibited the widest size distribution (Figure 3.1.I). PB-LEHs extruded through 50 nm pore diameter membranes exhibited a distribution width of ~20 nm, while PB-LEHs extruded through 400, 200, 100 and 80 nm pore diameter membranes exhibited size distribution widths ranging from 40 to 60 nm. All of these size distributions were Gaussian in shape, except the size distribution of PB-LEHs extruded through 100 nm pore diameter membranes, which was slightly skewed. The size distributions became significantly more monodisperse with the presence of salt in the extrusion buffer (Figure 3.2.I). The size distributions of PBS-LEHs prepared with an initial Hb concentration of 300 mg/mL were Gaussian in shape, and exhibited distribution widths ranging from 10 to 20 nm. Without Hb encapsulation (Figure 3.3.A), PB-control dispersions exhibited Gaussian shaped size distributions. PB-controls extruded through 100 and 200 nm pore diameter membranes exhibited the most monodisperse size distributions with distribution widths of ~10 nm, while PB-controls extruded through 50 and 80 nm pore diameter membranes exhibited size distribution widths of 20 nm, and PB-controls extruded through 400 nm pore diameter membranes exhibited the widest size distribution with a distribution width of 40 nm. PBS-controls also displayed Gaussian shaped size distributions with distribution widths ranging from 10 to 20 nm (Figure 3.3.C). The size distributions of PB-LEHs slightly widened with increasing initial Hb concentration. However, the trend was less obvious in PBS-LEH dispersions. We observed that using PBS as the extrusion buffer generally caused the formation of more Gaussian-like and more monodisperse liposome size distributions, which have been
55
shown to exhibit simple exponential clearance kinetics compared to heterogeneous particles [61], thus permitting easier control and manipulation of liposome circulation lifetime. Sakai et al. [78], and Takeoka et al. [68] used different phospholipid and cholesterol formulations, and employed a Coulter particle analyzer (N4-SD) with a size distribution processor (SDP) to measure their liposome size distributions. Sakai et al. [67] verified their results using transmission electron microscopy (TEM). Their method of ULEH preparation was similar to ours, namely extrusion with a stepwise reduction of the membrane pore size down to the final pore diameter using PBS as the extrusion buffer. Sakai et al. [67] used a final extrusion pore diameter membrane of 220 nm, while Takeoka et al. [68] used a final pore diameter membrane of 200 nm. The N4-SD-SDP particle analyzer yielded a mean particle diameter of 251±87 nm for Sakai et al. [67], and 200±40 nm for Takeoka et al. [68]. TEM measurements yielded a mean diameter of 198±44 nm for Sakai et al. (number of LEHs measured = 55) [67]. Unfortunately, we can not directly compare our results with Sakai et al. [67] due to differences in the final pore size of the extruder membrane. Takeoka et al. [68] reported a similar mean radius with PBS-LEHs extruded through 200 nm pore diameter membranes although a different lipid combination was used. Both Sakai et al. [67] and Takeoka et al. [68] reported wider standard deviations compared to our result (we obtained a standard deviation of 10 nm for PBS-LEHs extruded through 200 nm pore diameter membranes), which indicated that the reproducibility of their methods to measure liposome size distribution was less accurate than our method. Mobed and Chang [104] measured the size distribution of liposomes composed of several lipid combinations (with distearoylphosphatidylcholine
56
and cholesterol as the principal lipids), extruded through 220 nm pore diameter membranes with PBS as the extrusion buffer, using a DAWN-F multi-angle light scattering photometer in a batch mode, and reported liposome mean diameters ranging from 154 to 220 nm. Farmer et al. [42] measured the size of ULEHs extruded through pore diameter membranes ranging from 3 µm to 200 nm using dynamic light scattering at a 90o angle in batch mode. Farmer et al. [42] reported that the polydispersity of their ULEH dispersions increased with increasing membrane pore diameter, but the polydispersity became constant as the membrane pore diameter approached 200 nm. In contrast, we investigated smaller ULEHs (extruded through 50-400 nm pore diameter membranes). In our case, the polydispersity indices did not appear to be influenced by the membrane pore diameter (refer to Table 3.1), which was in agreement with Farmer et al.’s observation [42].
3.2.4 Hemoglobin Encapsulation Efficiency Tables 3.3 and 3.4 display Hb encapsulation efficiencies (E%) and weight ratios of entrapped Hb to phospholipids (R) of ULEHs extruded in PB and PBS, respectively. Since ULEHs with diameters less than 100 nm could not be prepared, we only measured the Hb loading capacities and the oxygen binding properties of ULEHs extruded through 400, 200 and 100 nm pore diameter membranes. Theoretically, the sodium counter ions should mask the charge on the phospholipids, resulting in a decrease in the interlamellar spacing of multilamellar vesicles formed in the initial hydration step, thereby decreasing the aqueous trapped volume and Hb encapsulation efficiency [105]. As predicted, indeed extrusion in PB generally resulted in higher Hb encapsulation efficiencies, which is a 57
desired property of a cellular HBOC. The encapsulation efficiencies of PB-LEHs and PBS-LEHs generally increased with increasing initial Hb concentration.
TABLE 3.3
ENCAPSULATION EFFICIENCY (E%), WEIGHT RATIO OF ENTRAPPED HB TO PHOSPHOLIPIDS (R), P50, COOPERATIVITY COEFFICIENT (n), AND ENCAPSULATED METHB LEVEL OF PB-LEHS
MetHb Pore n E% R P50 level Radius (nm) (%) (mg Hb/ mg lipids) (mmHg) (%) 200 70.0 10.5 27 2.4 24 100 17.5 2.6 26 2.5 3 50 13.8 2.1 * * ** 200 200 29.3 2.9 22 2.2 0 100 39.5 4.0 23 2.3 3 50 17.5 1.8 22 2.3 0 150 200 48.7 3.7 21 2.2 0 100 33.2 2.5 23 2.3 25 50 34.8 2.6 23 2.4 0 100 200 81.3 4.1 19 2.1 0 100 5.1 0.3 24 2.4 0 50 48.6 2.4 23 2.4 55 50 200 13.5 0.3 23 2.3 0 100 20.3 0.5 22 2.1 0 50 8.1 0.2 23 2.0 4 ** The encapsulated metHb level > 50%. * The values of P50 and n could not be accurately measured due to high metHb level. Initial [Hb] (mg/mL) 300
58
TABLE 3.4
ENCAPSULATION EFFICIENCY (E%), WEIGHT RATIO OF ENTRAPPED HB TO PHOSPHOLIPIDS (R), P50, COOPERATIVITY COEFFICIENT (n), AND ENCAPSULATED METHB LEVEL OF PBS-LEHS
MetHb Pore n E% R P50 level Radius (nm) (%) (mg Hb/ mg lipids) (mmHg) (%) 200 62.4 9.4 * * ** 100 48.6 7.3 31 2.4 2 50 37.2 5.6 36 2.9 7 200 200 21.4 2.1 27 2.4 2 100 21.4 2.1 26 2.2 6 50 22.1 2.2 28 2.4 9 150 200 10.6 0.8 27 2.3 8 100 14.9 1.1 29 2.9 14 50 10.4 0.8 26 2.4 20 100 200 32.6 1.6 21 2.0 4 100 32.1 1.6 25 2.4 0 50 22.0 1.1 23 2.5 12 50 200 6.5 0.2 * * ** 100 6.0 0.2 19 2 20 50 2.3 0.06 * * ** ** The encapsulated metHb level > 50%. * The values of P50 and n could not be accurately measured due to high metHb level. Initial [Hb] (mg/mL) 300
The highest encapsulation efficiencies and R of both PB-LEH and PBS-LEHs were achieved at the highest initial Hb concentration (300 mg/mL), and the lowest values were achieved at the lowest initial Hb concentration (50 mg/mL), as expected. The weight ratio of encapsulated Hb to total cell membrane components of human RBCs is about 6.7 [106]. This R value can be mimicked using ULEHs prepared with an initial Hb 59
concentration of 300 mg/mL and extruded through 400 nm pore diameter membranes. In this chapter, we will not discuss the total entrapped Hb concentration of ULEH dispersions, since the total entrapped Hb concentration in the dispersion can be easily engineered by adjusting the concentration of the liposomes and/or the initial Hb concentration. Brandl et al. [69] reported phospholipid/entrapped Hb molar ratios of 9797±1218, 1521±164 and 656±241 (corresponding to initial Hb concentrations of 10, 20 and 60 mg/mL, respectively, with mean diameters of 2628, 1750 and 788 nm, respectively). For comparison sake, we converted Hb encapsulation efficiencies of PBSLEHs extruded through 400 nm pore diameter membranes into the appropriate molar ratios. The estimated phospholipid/entrapped Hb molar ratio of PBS-LEHs extruded through 400 nm pore diameter membranes using initial Hb concentrations of 300, 200, 150, 100 and 50 mg/mL are 13.5, 59.0, 159.5, 77.6 and 776.7, respectively. According to our results, the phospholipid/entrapped Hb molar ratio increased (the equivalent of decreasing Hb encapsulation efficiency) with decreasing initial Hb concentration. Although Brandl et al. [69] used a different lipid composition and produced larger LEHs using a dehydration/rehydration method, Brandl et al.’s results [69] exhibited the same trend as our results. Using Brandl et al.’s technique [69] to measure Hb encapsulation efficiency [69], Sakai et al. [78] reported an R value of 1.61 and Takeoka et al. [68] reported a value of ~1.7. Sakai et al. [78] used an initial Hb concentration of 100 mg/mL. Although Sakai et al. [67] prepared slightly larger LEHs (the extrusion membrane pore diameter was 220 nm), it is also interesting to compare their data with ours. The R value reported by Sakai
60
et al. [67] is identical with the R value of PBS-LEHs extruded through 200 nm pore diameter membranes at an initial Hb concentration of 100 mg/mL. Since Takeoka et al. [68] employed extrusion to a final pore diameter of 200 nm, direct comparison with their result is possible. Takeoka et al. [68] used an initial Hb concentration of 450 mg/mL, and obtained a lower R value compared to PBS-LEHs extruded through 200 nm pore diameter membranes at an initial Hb concentration of 300 mg/mL (R =7.29). This is in agreement with an observation reported by Zheng et al. [76], in which the upper limit of Hb encapsulation was reached at 25 wt% of the initial Hb concentration (263.16 mg/mL).
3.2.5 Oxygen Binding Properties The P50 and n of human RBCs are 26 mmHg and 2.3, respectively [24], while the measured P50 and n of cell-free bovine Hb are 26 mmHg and 2.5, respectively. ULEHs extruded both in PB and PBS exhibited P50s and cooperativity coefficients comparable to those of human RBCs and cell-free bovine Hbs, regardless of the extrusion membrane pore sizes and initial Hb concentrations utilized (refer to Tables 3.3 and 3.4). The oxygen dissociation curves of bovine Hb and ULEHs are presented in Figure 3.4. Here, Y is the fraction of Hb saturated with oxygen, and pO2 is the partial pressure of oxygen in mmHg. Although we only show the oxygen dissociation curves of ULEHs extruded through 400 nm pore diameter membranes at an initial Hb concentration of 200 mg/mL, Figure 3.4 represents the typical shape of oxygen dissociation curves for ULEH dispersions. Unlike human hemoglobin, bovine hemoglobin responds to chloride ions as its allosteric effector [107]. The presence of NaCl in the PBS extrusion buffer shifted the oxygen dissociation curves to the right, due to chloride ion binding to Hb, thus affecting oxygen binding
61
through allosteric interactions. The oxygen dissociation curves of PB-LEHs and PBSLEHs exhibited a sigmoidal shape, similar to that of human RBCs [16]. Taken together, these results indicate that encapsulating Hb inside liposomes does not compromise the oxygen binding properties of encapsulated Hb.
Figure 3.4. Oxygen dissociation curves of bovine Hb (panel A), and ULEHs prepared using an initial Hb concentration of 200 mg Hb/mL and extruded through 400 nm pore diameter membranes in PB (panel B) and PBS (panel C).
62
3.2.6 Encapsulated Methemoglobin Level Tables 3.3 and 3.4 display the encapsulated metHb levels of ULEHs extruded in PB and PBS, respectively. Various groups [78, 108, 109] observed that encapsulating Hb inside liposomes increased the encapsulated metHb level due to peroxidation of the lipid bilayers, which enhanced Hb oxidation. This phenomenon was observed in PBS-LEHs, but not in PB-LEHs. The metHb levels of PB-LEHs were generally very low (mostly 0%). Quantitative studies conducted by Wallace, et al. [38] demonstrated that anions, in particular chloride ions, promoted auto-oxidation of Hb and hence, resulted in the higher metHb levels observed for PBS-LEHs. We also observed that the initial Hb concentration did not influence the metHb level. LaBrake and Fung [108] found that small unilamellar vesicles enhanced Hb oxidation more than large unilamellar vesicles. However, our results do not support this previous observation.
3.3 Concluding Remarks The oxygen binding properties of ULEHs, regardless of the extrusion pore diameter, initial Hb concentration, and ionic strength of the extrusion buffers, were comparable to those of human RBCs. This indicated that ULEHs exhibited good potential as cellular HBOCs. Regardless of the extrusion pore diameter membranes utilized, only unmodified liposomes with diameters ranging from 100 to 300 nm could be prepared. When PB was used as the extrusion buffer, ULEHs with higher Hb encapsulation efficiencies and lower metHb levels were prepared. However, ULEHs extruded using PBS formed more monodisperse dispersions. Hence, these results presented a dilemma in selecting the proper extrusion buffer for facilitating the design of ULEHs as a cellular
63
HBOC. Moreover, our results suggest that the interaction between encapsulated Hb and lipid bilayers make ULEHs more susceptible to the influence of osmotic pressure gradients, and will therefore render ULEHs fragile in the blood stream. This potential problem can be solved by conjugation of a biocompatible polymer, poly(ethylene glycol) (PEG) to the surface of lipid bilayers. This study of PEG-conjugated LEHs is presented in Chapter 4.
64
CHAPTER 4
POLY(ETHYLENE GLYCOL)-CONJUGATED LIPOSOME ENCAPSULATED HEMOGLOBIN DISPERSIONS
4.1 Introduction
4.1.1 Background In order to improve the intravascular circulation, biocompatibility and colloidal state of unmodified liposome encapsulated Hb (ULEH) dispersions, the surface of ULEHs has been conjugated with poly(ethylene glycol) (PEG) [26, 33, 34, 44, 62]. PEG is a hydrophilic and bioinert polymer, which has been extensively used in drug-delivery systems and is declared to be safe by the U.S. Food and Drug Administration (FDA) [110]. It was proposed that PEGylation creates a steric hydrophilic barrier surrounding each liposome, preventing the deposition and adsorption of opsonizing plasma proteins onto the surface of liposomes, and hence prolonging the vesicles’ intravascular persistence [34, 62, 111]. While the circulation half-lives of ULEHs were reported to range between ~12-18 hours in rabbits [34, 42], the circulation half-life of LEHs grafted with 10 mole% of 5000 Da PEG was shown to be ~ 48 hours in rabbits [34]. Infusion of ULEHs into laboratory animals was observed to trigger complement activation [112, 65
113], which is associated with anaphylactic reactions and activation of proteolytic plasma cascades in-vivo [51]. However, PEGylation of LEHs successfully attenuated these adverse reactions [44]. Unlike acellular HBOCs, PEG-LEH dispersions were observed to be non-vasoactive when infused into animals [114, 115]. The steric barrier created by PEG conjugation stabilizes the colloidal state of PEG-LEHs, and thus, prolongs the storage life of PEG-LEH dispersions [33]. PEG-LEH dispersions stored in a deoxygenated state at 4 and 23oC were shown to be stable for one year [33]. In contrast, intervesicular aggregation and fusion occurred in ULEH dispersions after a few days of storage [33]. Laboratory animals subjected to hemodilution with PEG-LEH dispersions (80%-exchange transfusion in hamsters and 90%-exchange transfusion in rats) were observed to maintain normal tissue oxygenation and hemodynamics (due to the suppression of intervesicular aggregation in-vivo) during and after transfusion [41, 106]. Moreover, a bolus infusion of PEG-LEH dispersions into rats did not appear to induce any irreversible signs of deteriorative damage to the liver, spleen, lungs, kidney, heart, and pancreas [116]. Daily repeated infusions of PEG-LEH dispersions into rats for a period of 14 days showed no deteriorative signs in organ function (spleen, liver, kidney, heart, skeletal muscle, lungs, intestine, brain, testis, skin, thymus, stomach). However, the pancreatic function of rats appeared to change [97]. Meanwhile, rabbits subjected to hemorrhagic shock (85% blood withdrawal) were successfully resuscitated with transfusion of PEG-LEH dispersions, and lived normally for 6 months at which time they were sacrificed. It was observed that transfused PEGLEHs supported the animals until hemopoiesis compensated for blood loss [26]. Likewise, PEG-LEH dispersions were successfully used to resuscitate hamsters
66
experiencing hemorrhagic shock (50% blood withdrawal) [63]. These studies demonstrate the potential of PEG-LEHs as a cellular HBOC for applications ranging from routine surgery to trauma.
4.1.2 Research Overview The research presented in this chapter was directed toward improving the design of PEG-LEHs as a cellular HBOC [53]. In this work, PEG-LEH dispersions were prepared via extrusion through 100, 200 and 400 nm pore diameter membranes, since our previous study of ULEHs (Chapter 3) demonstrated that only liposomes with diameters from 100 to 300 nm could be prepared via extrusion. Unmodifed liposomes extruded through membrane pores smaller than 100 nm were observed to be unstable, and these liposomes fused together to form liposomes with diameters larger than 100 nm [39]. An initial Hb concentration of 300 mg/mL was used to prepare PEG-LEH dispersions in order to achieve high Hb encapsulation efficiencies. Both phosphate buffer (PB) and phosphate buffered saline (PBS) at physiological pH (7.3) were used as extrusion buffers. Liposomes were conjugated with two different PEG molecular weights: 550 and 2000 Da. Since it is desired to obtain PEG-LEH suspensions with low encapsulated metHb levels, especially for PEG-LEHs extruded in PBS, 1.2 mM of N-Acetyl-L-Cysteine (NAC) was coencapsulated to suppress metHb formation [35]. The size distribution, Hb encapsulation efficiency, oxygen affinity (as indicated by the P50 and cooperativity coefficient), and encapsulated metHb level of PEG-LEH dispersions prepared using various extruder membrane pore diameters, ionic strengths of the extrusion buffers, and PEG lengths were measured [53]. The influence of these variations on the physical
67
properties of PEG-LEH dispersions was analyzed in order to understand and identify experimental parameters, which affect the function of PEG-LEHs as an oxygen carrier. In addition, the simulated effect of in-vivo osmotic pressure on the physical integrity of PEG-LEH bilayers was evaluated. For convenience, LEHs grafted with 2000 Da PEG and extruded in PB will be referred to as PB-PL2000, and LEHs grafted with 550 Da PEG and extruded in PB will be referred to as PB-PL550. Likewise, LEHs grafted with 2000 and 550 Da PEG and extruded in PBS will be referred to as PBS-PL2000 and PBSPL550, respectively, and the corresponding PEG-conjugated controls will be referred to as PB-C2000, PB-C550, PBS-C2000, and PBS-C550.
4.2 Results and Discussion
4.2.1 Average Radius The number- (Rn), weight- (Rw), z-averaged radii (Rz), and polydispersity indices (Rw/Rn and Rz/Rn) of PEG-LEH and PEG-control dispersions extruded in PB and PBS are presented in Tables 4.1 and 4.2, respectively. PEG-LEHs and PEG-controls extruded through 100 nm pore diameter membranes exhibited mean diameters that were larger than the membrane pore diameter, while those extruded through 200 nm pore diameter membranes exhibited mean diameters that were similar to the membrane pore diameter, and those extruded through 400 nm pore diameter membranes exhibited mean diameters that were smaller than the membrane pore diameter. We previously observed this trend with ULEHs and their controls [39]. This phenomenon was caused by the opposing energetic and entropic forces, which selected an “equilibrium vesicle” size distribution
68
(Chapter 3). Here, it appeared that the equilibrium size range of PEG-LEH and PEGcontrol dispersions extruded in PB and PBS was ~140-350 nm in diameter. In contrast, the equilibrium size range of ULEHs and their controls was ~100-300 nm in diameter. The incorporation of “wedge-shaped” PEG-lipid molecules into liposome bilayers composed of tubular-shaped phospholipid molecules induced lateral expansion of the bilayers surface [117], which consequently promoted the formation of larger PEG-LEHs and PEG-controls [53].
TABLE 4.1
NUMBER- (Rn), WEIGHT- (Rw), Z-AVERAGED RADII (Rz) AND POLYDISPERSITY INDICES OF PEG-LEH AND PEG-CONTROL DISPERSIONS EXTRUDED IN PB
Initial [Hb] (mg/mL) 2000 Da PEG 300
Pore Radius (nm)
Rn (nm)
Rw (nm)
Rz (nm)
Rw/Rn
Rz/Rn
0
200 100 50 200 100 50
118.9 97.8 113.5 100.2 93.4 78.1
120 98.2 113.7 100 92.7 79.4
120.6 98.1 112.4 99.3 92.2 79.6
1.01 1 1 1 0.99 1.02
1.01 1 0.99 0.99 0.99 1.02
550 Da PEG 300
0
200 100 50 200 100 50
89.6 75.1 72 100.2 90.7 62.9
88.3 78.3 77.6 100 91 65 69
89.3 80.4 80.5 99.3 91.1 66.4
0.99 1.04 1.08 1 1 1.03
1 1.07 1.12 0.99 1 1.06
TABLE 4.2
NUMBER- (Rn), WEIGHT- (Rw), Z-AVERAGED RADII (Rz) AND POLYDISPERSITY INDICES OF PEG-LEH AND PEG-CONTROL DISPERSIONS EXTRUDED IN PBS
Initial [Hb] (mg/mL) 2000 Da PEG 300
Pore Radius (nm)
Rn (nm)
Rw (nm)
Rz (nm)
Rw/Rn
Rz/Rn
0
200 100 50 200 100 50
122.8 121 83.9 139.7 96 75.2
131.7 125.1 86.7 139.7 98.1 75.3
132.3 124.8 85.6 139 97.3 75
1.07 1.03 1.03 1 1.02 1
1.08 1.03 1.02 0.99 1.01 1
550 Da PEG 300
0
200 100 50 200 100 50
120.1 95.3 63.9 174.7 96.8 70.1
117.2 107.3 67.2 171.2 106.8 70
114.9 109.5 69 165 109.7 71.1
0.98 1.13 1.05 0.98 1.1 1
0.96 1.15 1.08 0.94 1.13 1.01
The osmotic pressure inside the aqueous core of PEG-LEHs extruded in PBS was expected to be higher than the osmotic pressure of the exterior environment (PBS), due to the presence of encapsulated Hb molecules. Whereas the osmotic pressure inside the aqueous core of PEG-LEHs extruded in PB was expected to be lower than the osmotic pressure of the exterior environment (PBS), due to the lower ionic strength of PB. However, unlike unmodified liposomes, the ionic strength of the extrusion buffer, the initial Hb concentration, and the Hb encapsulation efficiency (the encapsulation 70
efficiency data will be presented in section 4.2.3) did not affect the mean radii of PEGLEHs and PEG-controls. Lateral expansion of the liposome bilayer surface, induced by incorporation of PEG-lipid molecules into the lipid bilayer, rigidified the bilayer [117]. Hence, expansion and contraction of PEG-liposome bilayers due to the osmotic pressure gradient between the aqueous core of the liposome and the exterior environment (PBS) was suppressed by PEG conjugation [53]. Moreover, we previously observed that the interaction of encapsulated Hb molecules with the lipid bilayer of ULEHs caused ULEHs to be more susceptible to the influence of osmotic pressure gradient, therefore rendering ULEHs vulnerable to lysis in-vivo (Chapter 3). In contrast, the deteriorating effect of encapsulated Hb was not observed for PEG-LEH dispersions. Hence, PEGylation not only improves the intravascular persistence, biocompatibility and colloidal state of PEGLEHs, but also strengthens the lipid bilayer against the adverse effect of osmotic pressure gradient, and the deteriorating interaction with encapsulated Hb molecules.
4.2.2 Size Distribution The size distributions of PEG-LEH dispersions extruded with an initial Hb concentration of 300 mg/mL and their corresponding controls (empty PEG-liposomes) are displayed in Figures 4.1 and 4.2, respectively. Although the polydispersity indices of PEG-liposome dispersions indicated that PEG-LEH and PEG-control dispersions were monodisperse (indices were close to one), closer inspection revealed that the size distributions of PEG-liposomes extruded in PB were narrower compared to those extruded in PBS, regardless of the PEG molecular weight. The size distribution widths of PB-PL2000 and PB-PL550 dispersions extruded through 100, 200 and 400 pore diameter
71
membranes were 10-20 nm and 40 nm, respectively. In contrast, the size distribution widths of PBS-PL2000 and PBS-PL550 extruded through 100, 200 and 400 pore diameter membranes were 20-60 nm and 30-60 nm, respectively. Electrostatic repulsion between the chloride ions in the PBS extrusion buffer, and the negatively charged DMPG and DMPE-PEG phospholipids in the bilayers may disrupt the liposomes’ dispersion state. Conjugation with longer PEG chains (2000 Da) resulted in narrower size distributions for PEG-LEH dispersions extruded in the same buffer. Longer PEG chains were more effective in suppressing aggregation and fusion between PEGylated liposomes, hence creating more monodisperse suspensions. Monodisperse liposome suspensions are desired, since liposomes with different sizes exhibited different circulation half-lives [62] and heterogeneous vesicle suspensions exhibited complex clearance kinetics [61]. A similar trend was also observed with PEG-control dispersions. While the size distribution widths of both PB-C2000 and PBS-C2000 extruded through 100, 200 and 400 pore diameter membranes were almost similar, 15-40 nm, the size distribution widths of PB-C550 and PBS-C550 were 35-50 nm and 40-150 nm, respectively. Thus, we observed that PEG-conjugated liposome dispersions extruded in PB were more monodisperse than those extruded in PBS, and conjugation with longer PEG chains promoted the formation of narrower size distributions.
72
Figure 4.1. Size distributions of PEG-LEH dispersions extruded with an initial Hb concentration of 300 mg/mL. Differential and cumulative size distributions of 2000 Da PEG-LEH dispersions extruded in PB are displayed in panels A and B respectively; 2000 Da PEG-LEH dispersions extruded in PBS in panels C and D respectively; 550 Da PEGLEH dispersions extruded in PB in panels E and F respectively; and 550 Da PEG-LEH dispersions extruded in PBS in panels G and H respectively. Dashed lines represent: 400 nm pore diameter membranes, solid lines represent: 200 nm pore diameter membranes, and dotted lines represent: 100 nm pore diameter membranes.
73
Figure 4.2. Size distributions of PEG-control dispersions. Differential and cumulative size distributions of 2000 Da PEG-control dispersions extruded in PB are displayed in panels A and B respectively; 2000 Da PEG-control dispersions extruded in PBS are displayed in panels C and D respectively; 550 Da PEG-control dispersions extruded in PB are displayed in panels E and F respectively; and 550 Da PEG-control dispersions extruded in PBS are displayed in panels G and H respectively. Dashed lines represent: 400 nm pore diameter membranes, solid lines represent: 200 nm pore diameter membranes, and dotted lines represent: 100 nm pore diameter membranes.
74
Since this is the first study that measured the absolute size distribution of PEGLEHs, we can only compare the average sizes of PEG-LEH and PEG-control dispersions to the average sizes of PEG-liposomes measured by other groups. Extrusion though 220 nm pore diameter membranes produced PEG-LEH dispersions with mean diameters of 222±62 nm [33], 230±81 nm [118], 259±82 nm [119], and 250±80 nm [41], while microfluidization technique produced PEG-LEH dispersions with a mean diameter of 193 nm [34]. The best comparison is with Awasthi et al. [62], who prepared empty PEGliposome dispersions with mean diameters of 136.2, 165.5, 209.2, 275 and 318 nm via extrusion. Awasthi [50] experienced difficulties in preparing PEG-liposome dispersions with diameters less than 100 nm via extrusion, while microfluidization yielded PEGliposomes with diameters ranging from 80 to 120 nm. If anionic lipids were used (in our case, DMPG is negatively charged), it would not have been possible for liposomes with diameters larger than 320 nm to be prepared via extrusion [50]. This observation confirms our finding that only PEG-liposomes with diameters ranging from 140 to 350 nm could be prepared by extrusion.
4.2.3 Hemoglobin Encapsulation Efficiency Hb encapsulation efficiencies (E%) and weight ratios of entrapped Hb to phospholipids (R) of PEG-LEHs extruded in PB and PBS are displayed in Tables 4.3 and 4.4, respectively. Here, we did not discuss the entrapped Hb concentration of PEG-LEH dispersions, since the entrapped Hb concentration can be easily engineered by adjusting the concentration of the liposomes and/or the initial Hb concentration. We found that the encapsulation efficiencies of PB-PL2000, PB-PL550 and PBS-PL550 dispersions were
75
very low (less than 11%) and the R values were significantly less than the weight ratio of encapsulated Hb to total cell membrane components of human RBCs (~6.7) [106], regardless of the extruder membrane pore diameters utilized. The R values of PEG-LEHs prepared in this study were much lower compared to ULEHs. The R values of ULEHs extruded in PB and PBS were 2.1-10.5 and 5.6-9.4, respectively, using 400, 200 and 100 nm pore diameter membranes at an initial Hb concentration of 300 mg/mL [39]. This behavior was expected, since the PEG chains were also conjugated onto the inner leaflets of the liposome bilayers, hence decreasing the volume of the liposome aqueous core available for Hb encapsulation. Interestingly, the Hb encapsulation efficiencies of PBSPL2000 dispersions were high (27-36%), regardless of the extruder membrane pore diameters utilized. The highest Hb encapsulation efficiency (36.3%) was obtained by extruding PBS-PL2000 through the largest pore diameter membrane (400 nm) at an initial Hb concentration of 300 mg/mL. The R values of PBS-PL2000 were closer to the R value of human RBCs [106]. This was not expected, since conjugation of longer PEG chains (2000 Da) onto lipid bilayers should decrease the liposome core volume available for Hb encapsulation. We postulated that extrusion in PBS created a highly charged environment in the aqueous core of PBS-PL2000. This caused the PEG tails to collapse into a globular, mushroom-like conformation in order to shield the hydrophobic portion of the lipid membrane from the highly charged aqueous core [120], therefore liberating more core volume for Hb encapsulation. Due to the decrease in the ionic strength of the liposome aqueous core, PEG chains in the aqueous core of PEG-liposomes extruded in PB assumed an extended, brush-like configuration [53]. However, this does not explain the low Hb encapsulation efficiencies of PBS-PL550 dispersions. We found that PBS-
76
PL550 dispersions were stabilized by thermal undulations mechanism (liposomes stabilized by thermal undulations exhibit fluctuating bilayers). Here, the fluctuating bilayers of PBD-PL550 caused extensive Hb leakage and hence, PBS-PL550 dispersions possessed low R values [53].
TABLE 4.3
ENCAPSULATION EFFICIENCY (E%), WEIGHT RATIO OF ENTRAPPED HB TO PHOSPHOLIPIDS (R), P50, COOPERATIVITY COEFFICIENT (n), AND ENCAPSULATED METHB LEVEL OF PEG-LEHS EXTRUDED IN PB
Pore Radius (nm) 2000 Da PEG 200 100 50 550 Da PEG 200 100 50
E% (%)
R (mg Hb/mg lipids)
P50 (mmHg)
n
MetHb level (%)
4.3 10.5 7.5 5.2 3.9 6.0
0.6 1.6 1.1 0.8 0.6 0.9
24.7 22.5 21 31 28 38
2.3 2.7 2.7 2.2 2.2 2
<5 <5 <5 <5 <5 <5
77
TABLE 4.4
ENCAPSULATION EFFICIENCY (E%), WEIGHT RATIO OF ENTRAPPED HB TO PHOSPHOLIPIDS (R), P50, COOPERATIVITY COEFFICIENT (n), AND ENCAPSULATED METHB LEVEL OF PEG-LEHS EXTRUDED IN PBS
Pore Radius (nm) 2000 Da PEG 200 100 50 550 Da PEG 200 100 50
E% (%)
R (mg Hb/mg lipids)
P50 (mmHg)
n
MetHb level (%)
36.3 30.9 27.4 1.4 10.7 1.0
5.5 4.6 4.1 0.2 1.6 0.2
25 38 38 33 25 17
2.5 2.1 2.1 2.2 2.6 2.2
<5 <5 <5 <5 <5 <5
Phillips et al. [34] reported an R value of ~0.14 for LEHs grafted with 10 mol% of 5000 Da PEG with a mean diameter of 193 nm and prepared by microfluidization. We expected to obtain higher Hb encapsulation efficiencies than Phillips et al. [34], since our LEHs were grafted with shorter PEG chains. Sakai et al. [41] managed to increase the Hb encapsulation efficiency of PEG-LEHs by developing a preparation procedure to conjugate PEG chains exclusively onto the outer surface of PEG-LEHs. In this procedure [41], vesicle size regulation was achieved via extrusion. The R values of PEG-LEHs prepared using this procedure were reported to be 1.90 (259 nm mean diameter, 0.13 mol% 5000 Da PEG) [106], 1.79 (~250 nm mean diameter, 0.13-0.30 mol% 5000 Da PEG) [41, 97], and 1.75 (252 nm mean diameter, 0.30 mol% 5000 Da PEG) [116]. 78
Recently, a new preparation procedure involving a freeze-thawing cycle followed by extrusion to improve the Hb encapsulation efficiency of PEG-LEHs and to decrease the amount of time required to extrude PEG-LEHs was developed [121]. Sou et al. [121] reported that extrusion of PEG-LEH suspensions using this new method was 30 times faster than Sakai et al.’s procedure [41], and the encapsulated metHb levels could be maintained at <2%. However, the highest lipid concentration that can be used in order to shorten the extrusion time was 20 mg/mL, while Sakai et al. [41] used a lipid concentration of ~57 mg/mL. In this new procedure, PEG chains were also conjugated onto the inner leaflets of the lipid bilayers, therefore producing PEG-LEHs with R values of 1.7-1.8 (250±20 nm diameter) [121], not higher than the values yielded using Sakai et al.’s procedure [41].
4.2.4 Oxygen Binding Properties and Encapsulated Methemoglobin Level The measured P50, cooperativity coefficient (n) and encapsulated metHb level of PEG-LEH dispersions extruded in PB and PBS are presented in Tables 4.3 and 4.4, respectively. Regardless of the type of extrusion buffer, PEG molecular weight, and membrane pore size utilized, all PEG-LEH dispersions exhibited P50 and n values comparable to the measured P50 and n of bovine Hb (26 mmHg and 2.5, respectively), and human RBCs (26 mmHg and 2.3, respectively) [24]. The oxygen dissociation curves of bovine Hb and PEG-LEH dispersions are displayed in Figure 4.3. Here, Y is the fraction of Hb saturated with oxygen, and pO2 is the partial pressure of oxygen in mmHg. Although only oxygen dissociation curves of PEG-LEHs extruded though 400 nm pore diameter membranes are shown, Figure 4.3 represents typical oxygen dissociation curves
79
of PEG-LEH dispersions. Extrusion in PBS shifted the oxygen dissociation curves to the right, since chloride ions allosterically regulate oxygen binding to bovine Hb [107]. All curves displayed a sigmoidal shape, similar to that of human RBCs [16]. Surface modification with PEG clearly did not compromise the oxygen binding properties of PEG-LEH dispersions. Although chloride ions were reported to promote Hb oxidation [38, 39], we did not observe any differences in the metHb levels of PEG-LEHs extruded in PB versus PBS. This study demonstrated that coencapsulated NAC effectively suppressed metHb formation. However, the encapsulated metHb levels of PEG-LEHs prepared by Sakai et al. were reported to be <3% [41, 116], since 5 mM of the reductant DL-homocysteine was coencapsulated. The encapsulated metHb levels of PEG-LEHs could be maintained at <2% by utilizing carbonylhemoglobin (carbonylHb) to prepare PEG-LEH dispersions [121]. In this procedure, Hb will be converted into carbonylHb by bubbling CO gas through the Hb solution. However, a reductant, reductase or catalase should be coencapsulated, in order to maintain low metHb levels during storage. Hence, we will coencapsulate 5 mM reductant in the future to maintain low metHb levels.
80
Figure 4.3. Oxygen dissociation curves of bovine Hb (panel A), PEG-LEHs grafted with 2000 Da PEG and extruded in PB (panel B) and PBS (panel C), and PEG-LEHs grafted with 550 Da PEG and extruded in PB (panel D) and PBS (panel E). All PEG-LEH dispersions were extruded through 400 nm pore diameter membranes.
81
4.3 Concluding Remarks PEG-LEH dispersions exhibited oxygen binding properties comparable to that of human RBCs. Coencapsulation of 1.2 mM NAC effectively maintained the encapsulated metHb level < 5%, even though PBS was used as the extrusion buffer. These properties demonstrate that PEG-LEHs possess good potential as a cellular HBOC [53]. Similar to unmodified liposomes, only PEGylated liposomes with diameters ranging from 140 to 350 nm could be prepared via extrusion. It was observed that PEG conjugation rigidified the liposome bilayer and hence, attenuated, or even eradicated, the influence of osmotic pressure gradient on PEG-LEHs, which can render the LEHs susceptible to lysis in-vivo. In contrast, the interaction of encapsulated Hb molecules with the lipid bilayers of ULEHs caused ULEHs to be more susceptible to the influence of osmotic pressure gradient. This deteriorating effect of encapsulated Hb was not observed with PEG-LEH dispersions. Hence, PEGylation of liposomes strengthens lipid bilayers against the adverse effects of osmotic pressure gradients, and the deteriorating interaction with encapsulated Hb molecules. PEG-LEHs extruded in PB exhibited more monodisperse dispersions compared to PEG-LEHs extruded in PBS. This was opposite the trend observed with ULEH dispersions. As expected, Hb encapsulation efficiencies of PEG-LEHs were low, since the aqueous core volume available for Hb encapsulation was reduced by PEG chains conjugated onto the inner leaflet of the liposome bilayer. Interestingly, the Hb encapsulation efficiencies of LEHs grafted with 2000 Da PEG and extruded in PBS were high. We postulated that this was caused by the collapse of PEG chains in the aqueous core into a mushroom-like configuration, which increased the core volume available for
82
Hb encapsulation. Sakai et al. [41] developed a preparation procedure to conjugate PEG chains exclusively onto the outer surface of LEHs. Therefore PEG-LEH dispersions with higher Hb encapsulation efficiencies can be prepared. However, the concentration of PEG chains that can be successfully conjugated onto the surface of liposomes is limited (a maximum PEG surface coverage of 10 mole% of 5000 Da was previously reported [34]). In the next chapter, we will introduce a novel polymer vesicle, which exhibits good potential as an in-vivo delivery system for oxygen transport and whose superior physical properties can overcome the disadvantages of PEGylated liposomes.
83
CHAPTER 5
POLYMERSOME ENCAPSULATED HEMOGLOBIN DISPERSIONS
5.1 Introduction
5.1.1 Background Although PEGylated liposome encapsulated Hb (PEG-LEH) dispersions exhibited good potential as a cellular HBOC, the surface coverage and molecular weight of PEG chains that can be stably conjugated onto the liposome bilayer are severely limited [45, 46]. Increasing the PEG molecular weight increases the tendency of PEG-lipid molecules to form into a separate micellar phase. However, increasing the PEG surface density and the length of PEG chains conjugated onto the surface of liposomes improves the intravascular circulation, biocompatibility and colloidal state of PEG-LEHs [47-49]. Hence, there is a limit to optimizing the circulatory half-life, biocompatibility and colloidal state of PEG-LEHs. More specifically, the maximum circulation half-life of liposomes grafted with 10 mol% of 5000 Da PEG was observed to be 48 hours in rabbits [34] or ~36 hours in rats [50]. Vesicles conjugated with PEG chains shorter than 20 kDa were found to be less effective in suppressing complement activation [51, 122].
84
Moreover, the mechanical strength of liposomes is limited by their universal membrane thickness (3-4 nm) [40, 123]. The physical limitations encountered with PEG-LEHs can be alleviated by employing polymersomes as an in-vivo delivery system to transport oxygen throughout the systemic circulation. Polymersomes, which are composed solely of amphiphilic diblock copolymers, overcome the disproportionality of PEG-lipids by increasing both the hydrophobic and hydrophilic masses of the copolymer diblocks [59, 124]. Consequently, polymersomes are 100% surface PEGylated with longer PEG chains compared to PEG-liposomes [56-58]. These unique physical properties impart various advantages to the design of polymersomes as a cellular HBOC. The PEG chains on polymersomes automatically adopt a mushroom-brush intermediate structural
conformation or a brush conformation due to steric repulsion between neighboring PEG chains. It was observed that PEG brushes adopting these conformations were more effective in preventing phagocytosis and suppressing activation of the human complement system [51]. Longer PEG chains (20 kDa or more), which were found to be more effective in suppressing complement activation, can be easily incorporated into polymersomes [51, 122]. Polymersomes with diameters ranging from 50 nm to 50 µm were shown to be readily prepared in aqueous solution [47, 58, 59, 73]. The hydrophobic membrane thickness of polymersomes can be engineered from 3 to 40 nm, by simply varying the molecular weight of the copolymer hydrophobic block [59]. Similarly, the length of PEG chains on polymersomes can be engineered from 1.2 to 30 kDa by varying the molecular weight of the hydrophilic block [55, 59]. Increasing the membrane thickness of polymersomes will increase the vesicle’s mechanical strength [59, 123, 125], 85
while increasing the length of PEG chains on polymersomes will impart longer intravascular persistence, and improved biocompatibility and colloidal state compared to PEG-liposomes [47, 73]. More specifically, giant polymersomes with membrane thickness ~ 8 nm were shown to be almost an order of magnitude tougher and at least ten times less permeable to water than liposomes [56, 73]. This unique physical feature reduces the ability of polymersomes to osmotically lyze. In contrast, liposomes are prone to osmotic lysis because of their thin membranes (3-4 nm), which compromises their mechanical stability [39, 54]. Increasing the membrane thickness of polymersomes imparted better resistance toward membrane solubilization by various detergents [123]. The bending rigidities (mechanical strength) of polymersomes with membrane thicknesses of 8 and 14.8 nm were found to be 33.3 and 465.5 kT, respectively (k is Boltzmann’s constant and T is absolute temperature) [125]. In contrast, the bending rigidity of liposomes was found to range from 13.3 to 21.4 kT [125]. It was observed that the membrane rigidity of polymersomes increased with increasing membrane thickness, while the surface elasticity was scale-independent [59, 126]. Polymersomes were stable for several months in saline solution, and for at least 5 days in blood plasma under well-mixed quasi-physiological condition [47, 57, 73]. Polymersomes exhibited no in-surface thermal transition (lamellar surface transition and membrane area expansion with increasing temperature) up to 60oC. In contrast, liposomes experienced thermal transition at ~ 23-25oC [42, 73]. If the surface transition is at, or near, the vesicle preparation or storage temperature, leakage of the vesicle content will occur.
86
Incubation of cell cultures with polymersomes showed no adverse effects or vesicle-cell adhesion after 5 days under turbid condition [47], and after 48 hours under stagnant condition [73]. A mixture of polymersomes with whole blood also showed no signs of adhesion to RBCs for several hours without sustained mixing. In addition, infusion of five mg of polymersomes into rats (300-400 g body weight) did not induce toxic effects over an eight week period [47]. The circulation half-life of 100 nm-diameter empty polymersomes with PEG brushes ranging from 1200 to 3680 Da and hydrophobic core thicknesses = 9.6–14.8 nm were shown to be 10-20 hours longer than those of PEGliposomes with similar sizes. The circulatory half-life of polymersomes appears to depend mainly on the molecular weight of the PEG brush, and not on the hydrophobic core thickness, due to the obfuscation of the hydrophobic core by the dense PEG brush [47]. Therefore, there is flexibility in the choice of hydrophobic core chemistry. Considering the aforementioned advantages of polymersomes, we are confident that polymersome encapsulated Hb (PEH) dispersions will function as a superior HBOC compared to PEG-LEHs.
5.1.2 Research Overview In this chapter, we introduce PEH dispersions as a novel type of cellular HBOC. PEH dispersions can be readily prepared by encapsulating Hb inside the aqueous core of polymersomes [55]. To our knowledge, this is the first study that explores the potential of polymersomes as an oxygen carrier [55]. Here, we aim to prove that PEH dispersions possess physical properties that satisfy the design criteria of a cellular HBOC (Chapter 1). To accomplish this, PEH dispersions of various sizes, hydrophobic membrane
87
thicknesses and PEG chain lengths were prepared and their physical properties were investigated [55]. The physical properties measured in this study included: vesicle size distribution, Hb encapsulation efficiency, oxygen affinity (as indicated by the P50 and cooperativity coefficient) and encapsulated metHb level. Bioinert polyethylene oxide (PEO) and polybutadiene (PBD) were chosen as the hydrophilic and hydrophobic block of the diblock copolymers, respectively [47, 73, 74]. PEO is the structural equivalent of PEG [73]. The structural properties of the PBD-PEO copolymers used in this study are summarized in Table 5.1. The hydrophobic membrane thickness (d) of the polymersomes were taken or extrapolated from the literature [59]. For convenience, poly(butadiene(1,4 addition)-b-ethylene oxide) with molecular weights of 22-12.6 and 5-2.3 kDa will be referred to as L1 and L2, respectively. Poly(butadiene(1,2 addition)-b-ethylene oxide) with molecular weights of 2.5-1.3 and 1.8-0.9 kDa will be referred to as B1 and B2, respectively. The PBD blocks of L1 and L2 copolymers are linear, whereas the PBD blocks of B1 and B2 copolymers are branched. These copolymers are nearly monodisperse based on the measured ratios of the weight-averaged to number-averaged molecular weight (Mw/Mn < 1.1). PEHs were prepared using an initial Hb concentration of 300 mg/mL and extruded in PBS to achieve high Hb encapsulation efficiencies. Based on the literature [33], 5 mM of NAC was coencapsulated to maintain encapsulated metHb level less than 3%. Our previous studies of ULEHs and PEG-LEHs [39, 53] showed that liposomes with diameters smaller than 100 nm were unstable and fused together to form larger liposomes (refer to Chapters 3-4). Hence, in this study, PEHs were prepared via extrusion through 200 and 400 nm pore diameter membranes.
88
TABLE 5.1
NUMBER-AVERAGED MOLECULAR WEIGHT (Mn), PEG LENGTH, HYDROPHILIC MASS FRACTION (fhydrophilic), AND HYDROPHOBIC MEMBRANE THICKNESS (d) OF PBD-PEO DIBLOCK COPOLYMERS USED IN THIS STUDY
Diblock copolymer 22-12.6 kDa (L1) 5-2.3 kDa (L2) 2.5-1.3 kDa (B1) 1.8-0.9 kDa (B2)
Mn (kDa) 34.6 7.3 3.8 2.7
PEG length (kDa) 12.6 2.3 1.3 0.9
fhydrophilic 0.36 0.32 0.34 0.33
d (nm) ~35 12-13 9-10 8-9
5.2 Results and Discussion
5.2.1 Size Distribution and Average Radius The size distributions of PEH and control dispersions are shown in Figures 5.1 and 5.2, respectively. The number- (Rn), weight- (Rw), z-averaged radii (Rz) and distribution widths extracted from the differential size distributions of PEH and control dispersions are shown in Tables 5.2 and 5.3, respectively. L2 (5-2.3 kDa) and B2 (1.80.9 kDa) diblock copolymers were observed to have poor solubility in aqueous solution, while L1 (22-12.6 kDa) and B1 (2.5-1.3 kDa) diblock copolymers completely dissolved in PBS or Hb solutions. During the rehydration step in either Hb or PBS solutions, L2 and B2 diblock copolymers formed solid copolymer precipitates, micelles and polymersomes [55]. Even with vigorous sonication, the copolymer precipitates did not 89
completely dissolve in aqueous solution. In agreement with the literature[127, 128], polymersome formation was favored upon an increase in the total molecular weight of the copolymers. The morphology of copolymer aggregates in solution is governed by a balance of the contributions to the free energy of aggregation: 1) core-chain stretching, 2) interfacial energy and 3) intercoronal chain interactions [127]. Thus, the morphologies can be controlled by many factors which influence one or more of the three free-energy contributions [127]. The Gibbs free energy becomes more negative with increasing total copolymer molecular weight [128]. Thus, the morphology shifts from micelles/open bilayers structures to vesicles with increasing total copolymer molecular weight in order to decrease the interfacial tension between the hydrophobic portion of copolymer and the aqueous solution [128]. Since the bending modulus of polymers increases with increasing molecular weight, the formation of vesicles should be easier in long-chained copolymers compared to short-chained copolymers [127]. However, other factors also determine the morphology of copolymer aggregates, such the length of the hydrophilic block [127] and the hydrophilic/hydrophobic block ratio [129]. The aggregate structure changes from spherical to cylindrical micelles to vesicles to micro-size compound micelles with decreasing length of the hydrophilic block [127] and the hydrophilic/hydrophobic block ratio [129]. Therefore, in this study, we can not yet determine the molecular weight required for the copolymers to completely self-organize into polymersomes. Since L1 and B1 copolymers completely dissolved in PBS or Hb solutions, we will discuss these copolymers first. It should be noted that L1 and B1 copolymers could also self-assemble in aqueous solution to form micelles in addition to polymersomes.
90
However, the presence of micelles was not detected in the AFFFF-MASLS-DIR chromatograms. Therefore, we deduced that most or all of the L1 and B1 diblock copolymers self-assembled into polymersomes.
Figure 5.1. Differential and cumulative size distributions of PEH dispersions composed of 22-12.6 kDa PBD-PEO copolymers extruded through 400 (solid lines) and 200 nm (dotted lines) pore diameter membranes are shown in panels A and B, respectively; PEH dispersions composed of 5-2.3 kDa PBD-PEO are shown in panels C and D; PEH dispersions composed of 2.5-1.3 kDa PBD-PEO are shown in panels E and F; and PEH dispersions composed of 1.8-0.9 kDa PBD-PEO are shown in panels G and H. 91
Figure 5.2. Differential and cumulative size distributions of control (empty polymersome) dispersions composed of 22-12.6 kDa PBD-PEO copolymers extruded through 400 (solid lines) and 200 nm (dotted lines) pore diameter membranes are shown in panels A and B, respectively; control dispersions composed of 5-2.3 kDa PBD-PEO are shown in panels C and D; control dispersions composed 2.5-1.3 kDa PBD-PEO are shown in panels E and F; and control dispersions composed 1.8-0.9 kDa PBD-PEO are shown in panels G and H.
92
TABLE 5.2
NUMBER- (Rn), WEIGHT- (Rw), AND Z-AVERAGED RADII (Rz) AND SIZE DISTRIBUTION WIDTHS OF PEH DISPERSIONS
Pore radius (nm) 22-12.6 kDa (L1) 200 100 5-2.3 kDa (L2) 200 100 2.5-1.3 kDa (B1) 200 100 1.8-0.9 kDa (B2) 200 100
Rn (nm) 80.5 77.4 119 85.7 99.6 109.8 262.3 102.1
Rw (nm) 83.7 81.6 120.5 93.9 111.7 108.9 210.5 119.2
Rz (nm) 80.8 80.4 121.5 97.3 117.3 109.6 192.1 125.2
Distribution width (nm) 60 65 50 80 35 15 125 120
93
TABLE 5.3
NUMBER- (Rn), WEIGHT- (Rw), AND Z-AVERAGED RADII (Rz) AND SIZE DISTRIBUTION WIDTHS OF CONTROL (EMPTY POLYERSOME) DISPERSIONS
Pore radius (nm) 22-12.6 kDa (L1) 200 100 5-2.3 kDa (L2) 200 100 2.5-1.3 kDa (B1) 200 100 1.8-0.9 kDa (B2) 200 100
Rn (nm) 49.4 54.4 145.5 70.8 220.1 195.3 224.3 77.2
Rw (nm) 83.5 60.6 149.9 77.9 175.6 154.4 206.8 99
Rz (nm) 65.8 53.7 138.3 73.7 158.1 137.4 182.5 100.9
Distribution width (nm) 75 50 50 50 170 110 190 80
We observed that B1 PEH dispersions were more monodisperse than L1 PEH dispersions, regardless of the extruder membrane pore sizes utilized. However, L1 control dispersions were more monodisperse than B1 control dispersions. The size distribution widths of L1 PEH and control dispersions were almost similar (differed by ~10 nm), while those of B1 control dispersions were significantly larger compared to B1 PEH dispersions. The behavior observed for the control dispersions was expected, since the molecular weight of the PEO block of L1 control polymersomes (12.6 kDa) was significantly greater than that of B1 control polymersomes (1.3 kDa). Hence, it was expected that the colloidal state of L1 control dispersions should be better than that of B1 94
control dispersions due to the superior steric shielding afforded by the longer PEO block of the L1 copolymer. However, the trend observed for PEH dispersions was exactly opposite to what we expected. In the case of LEH dispersions, it was observed that encapsulated Hb molecules formed complexes with lipid bilayers, which was subsequently followed by intercalation of Hb molecules into the bilayers [103]. Complex formation and Hb intercalation were primarily attributed to the hydrophobic interaction between Hb and the lipid membrane. To date, little is known regarding protein adsorption into pure
polymeric bilayers. However, Pata and Dan predicted that transmembrane proteins incorporation into pure polymeric bilayers is possible [130]. In other words, it is possible for Hb molecules to intercalate into the hydrophobic core of the polymersomes. It appeared that the intercalation of Hb molecules into the hydrophobic membranes of polymersomes composed of branched hydrophobic blocks (B1) promoted monodispersity. Past studies [54, 111, 131] reported that vesicle dispersions stabilized by spontaneous curvature mechanism (indicated by vesicle bending constant >> kT where k is the Boltzmann’s constant and T is 37oC) exhibited narrower size distributions compared to those stabilized by thermal undulations mechanism (bending constant ≈ kT or the thermal energy). In liposomes, it was observed that incorporation of wedge-shaped PEG-lipids into the tubular-shaped phospholipids bilayers induced lateral expansion of the bilayers, which rigidified the bilayers and shifted the vesicle stabilization mechanism to spontaneous curvature [53, 132]. We speculated that the Hb intercalation into the hydrophobic core of B1 polymersomes also induced membrane lateral expansion, which increased the rigidity of the membranes and consequently, shifted the vesicle stabilization 95
mechanism to spontaneous curvature. Therefore, monodisperse B1 PEH dispersions were formed. However, it is not clear why this phenomena occurred in B1 PEH dispersions, but not in L1 PEH dispersions. Nevertheless, in designing PEH dispersions for use as a cellular HBOC, it is very important to prepare a monodisperse dispersion, since the circulation half-life and kinetics of clearance of these vesicles are determined by their sizes [61]. Hence, encapsulation of Hb into polymersomes composed of copolymer with branched PBD blocks (B1 copolymer), which appeared to improve their monodispersity, contributed favorably to the design of PEHs as a cellular HBOC. The average radius of B1 dispersions was larger than that of L1 dispersions by 2030 nm for PEH dispersions and by 90-150 nm for control dispersions, regardless of the extruder membrane pore sizes utilized. Since the bending modulus of polymer increases with increasing molecular weight [127, 133], the formation of large vesicles which have less curvature compared to small vesicles is favored by copolymer with lower molecular weight (in this case, B1 copolymer). Vesicles with radii larger than 50 nm and less than 125 nm are desired in order to obtain long circulation lifetimes [25, 62]. Hence, we demonstrated that PEHs with radii of 50-125 nm could be easily prepared. The average radii of L1 PEHs were larger than those of L1 controls, whereas the average radii of B1 PEHs were smaller than those of B1 controls. Our first guess was that encapsulated Hb in the aqueous core of polymersomes caused the osmotic pressure inside the polymersomes to be larger than the osmotic pressure of the exterior environment (PBS buffer). Therefore, the osmotic pressure gradient drove water osmosis into the
96
polymersomes and caused L1 PEHs to swell. Water osmosis into and out of the aqueous core was encountered with LEHs [39]. Empty polymersomes with (membrane thickness) d = 8 nm were found to be at least ten times less permeable to water than common liposomes [56]. However, d of L1 polymersomes was extrapolated to be ~35 nm (refer to Table 5.1), meaning that the permeability of L1 polymersomes to water is even lower. Although the encapsulation efficiencies of B1 PEH dispersions were higher than L1 PEH dispersions (the encapsulation efficiency data will be presented in section 5.2.2) and B1 PEH membranes were thinner than L1 PEHs (refer to Table 5.1), B1 PEH particles did not swell, but shrank. Hence, the swelling and shrinking of polymersomes when Hb was encapsulated could not be caused by the osmotic pressure gradient between the aqueous core of polymersomes and the exterior environment. This behavior can be attributed to Hb molecules intercalation into the membranes of B1 polymersomes. This observation is in agreement with our speculation that Hb intercalation into B1 PEH bilayers shifted the vesicle stabilization mechanism to spontaneous curvature, since vesicle dispersions stabilized by spontaneous curvature typically possessed smaller radii compared to those stabilized by thermal undulations [131, 132]. We have not yet found the explanation for the increase of L1 PEHs’ radii due to Hb encapsulation. We also observed that there was no significant difference in the average radii of L1 and B1 PEH dispersions extruded through different membrane pore sizes. The average radii of control particles extruded through 200 nm pore diameter membranes were smaller than those extruded through 400 nm pore diameter membranes, which was
97
expected. However, the average radius of B1 control particles extruded through 200 nm pore diameter membranes was larger than the extruder membrane pore size. We previously encountered the same trend with respect to both unmodified and PEGylated liposomes [39, 53]. Here, it appeared that polymersomes favored an “equilibrium vesicle” radii of ~80 nm for L1 PEH dispersions, ~100 nm for B1 PEH dispersions, 50-60 nm for L1 control dispersions, and 150-200 nm for B1 control dispersions. L2 and B2 diblock copolymers formed solid copolymer precipitates, micelles and polymersomes in aqueous solution. After the solid copolymer precipitates were removed from the dispersions, and the micelle and polymersome mixtures were injected into the AFFFF-MASLS-DIR system, the presence of micelles was not observed in the light scattering spectra as a separate chromatographic peak. Hence, the size distributions and average radii presented here are those of L2 and B2 polymersomes. Note that only polymersomes are capable of encapsulating Hb. It was observed that the size distributions of B2 control and PEH dispersions were broader than L2 control and PEH dispersions, regardless of the extrusion membrane pore sizes utilized. This was expected since the longer PEG chains of L2 copolymers (2.3 kDa) improved the colloidal state of L2 polymersome dispersions compared to B2 polymersomes, which possessed shorter PEG chains (0.9 kDa). Interestingly, unlike B1 PEH dispersions, B2 PEH dispersions did not exhibit narrow size distributions. Hence, we observed that Hb intercalation into the polymersome bilayers, which affected the vesicle size distributions, only occurred in B1 polymersome dispersions. The unique
property of B1 copolymer which contributes to this phenomena is not yet known.
98
The average radii of L1 controls < those of L2 controls < those of B1 controls < those of B2 controls, because of the decreasing molecular weight of the copolymers, i. e. decreasing bending moduli [127, 133]. Similar trend was also observed with the average radii of L1, L2, B1 and B2 PEHs. We observed no consistent trend between the average radii of L2 PEHs and those of L2 controls, and between the average radii of B2 PEHs and those of B2 controls. Unlike the trend observed with the average radii of the L1 and B1 polymersomes, L2 and B2 control and PEH dispersions extruded through 100 nm pore radius membranes were smaller than those extruded through 200 nm pore radius membranes. The average radii of L2 control particles and L2 PEH particles extruded through 200 and 100 nm pore radius membranes were less than the membrane pore sizes, however, this was not the case with respect to B2 control particles and B2 PEH particles. According to the “equilibrium size” theory discussed in Chapter 3, the average radii of L2 control and PEH particles extruded though 200 nm pore radius membranes must be the “maximum equilibrium radii,” while the average radii of L2 control and PEH particles extruded though 100 nm pore radius membranes must be the “minimum equilibrium radii.” However, the “equilibrium size” of B2 control and PEH particles must be larger than the 200 and 100 nm pore radius membranes utilized in this study, which caused the average radii of the B2 control and PEH particles to be larger than the extruder membrane pore radii utilized.
99
5.2.2 Hemoglobin Encapsulation Efficiency Hb encapsulation efficiencies (E%), and weight ratios of estimated entrapped Hb to diblock copolymer (R) of PEH dispersions are shown in Table 5.4. L1 PEH dispersions extruded through 400 and 200 nm pore diameter membranes exhibited lower encapsulation efficiencies and R. This was expected, since the longer PEO chains of L1 copolymer (12.6 kDa), which were also present on the inner leaflets of the polymersomes, reduced the core volume available for Hb encapsulation. In addition, the hydrophobic membrane of L1 polymersomes was much thicker than that of L2, B1 and B2 polymersomes, which further reduced the core volume available for Hb encapsulation. However, the encapsulation efficiency and R of L1 and B1 PEH dispersions extruded through 400 nm pore diameter membranes were the most reproducible, while those of L1 and B1 PEH dispersions extruded through 200 nm pore diameter membranes and those of L2 and B2 PEH dispersions had poor reproducibility. The “equilibrium” radii of L1 and B1 PEHs were close to the size of the 100 nm pore radius membranes. When a vesicle suspension is extruded through a pore membrane of a certain size, the resultant vesicle suspension will be smaller than the pore size. Here, L1 and B1 PEHs extruded through 100 nm pore radius membranes formed PEHs with radii less than the “equilibrium vesicle” radii dictated by the energy and entropy constraints. Therefore, these PEHs reassembled after extrusion to form PEHs possessing “equilibrium” radii, and allowed Hb leakage during the process. Since this process was not within our control, the reproducibility of E% and R of L1 and B1 PEH dispersions extruded through 100 nm pore radius membranes was poor. Due to the competition between the formation of
100
diblock copolymer precipitates, micelles and polymersomes during the rehydration step of L2 and B2 copolymer films, the concentrations of L2 and B2 PEH dispersions produced in each experiment were colloidally unstable, resulting in poor reproducibility.
TABLE 5.4
ENCAPSULATION EFFCIENCY (E%), WEIGHT RATIO OF ENTRAPPED HB TO COPOLYMER (R), P50, COOPERATIVITY COEFFICIENT (n) AND ENCAPSULATED METHB LEVEL OF PEH DISPERSIONS
Pore radius (nm) 22-12.6 kDa (L1) 200 100 5-2.3 kDa (L2) 200 100 2.5-1.3 kDa (B1) 200 100 1.8-0.9 kDa (B2) 200 100
E% (%) 5.3±0.4 2.7±1.2 11.1±7.5 3.9±0.8 10.1±0.9 12.2±2.7 7.5±2.4 11.6±7.7
R P50 (mg Hb / mg copolymer) (mmHg) 16.0±1.2 8.1±5.9 33.4±22.6 11.6±2.5 30.3±2.5 36.5±8.2 22.4±7.1 34.7±23.0 22 30 28 27 25 27 30 30
n
MetHb level (%) <1 <1 <1 <1 <1 <1 <1 <1
2.4 2.5 2.1 2.1 2.1 2.5 2.4 2.4
Alternative design strategies, which have been used to create cellular HBOC include: PEGylated actin-containing liposomes, lipogel particles and nanoscale hydrogel 101
particles (NHPs) encapsulating Hb. The weight ratios of entrapped Hb to phospholipids in PEG-LEH dispersions reported in the literature were 5.5 (~250 nm diameter, 10 mol% of 2 kDa PEG) [53], 0.2 (~230 nm diameter, 10 mol% of 0.55 kDa PEG) [53], 1.90 (259 nm mean diameter, 0.13 mol% of 5 kDa PEG) [106], 1.79 (~250 nm mean diameter, 0.13-0.30 mol% of 5 kDa PEG) [41, 97], 1.75 (252 nm mean diameter, 0.30 mol% of 5 kDa PEG) [116], and ~0.14 (193 nm diameter, 10 mol% of 5 kDa PEG) [34]. Current designs of cellular HBOCs include PEGylated liposome encapsulated actin-hemoglobin dispersions, and lipogel particles and nanoscale hydrogel particles (NHPs) encapsulating Hb. Liposome encapsulated actin-hemoglobin (LEAcHb) dispersions grafted with 2 kDa PEG exhibited R values of 4.02 and 6.90 for LEAcHbs with average diameters of 272 and 281 nm, respectively [134]. Lipogel particles and NHPs encapsulating Hb possessed R values ranging from ~2.57 to ~7.31 for particles with an average diameter of 200-360 nm [100]. Meanwhile, ULEHs exhibited R values of 7.3 and 9.4 for ULEHs with average diameters of ~238 and ~260 nm, respectively [39]. The weight ratio of Hb to the total cell membrane components of human RBCs is ~ 6.7 [106]. Although smaller in size compared to other cellular HBOCs, polymersomes were able to encapsulate higher concentrations of Hb compared to PEG-liposomes, PEGylated actin-containing liposomes, unmodified liposomes and even human RBCs, while possessing thicker hydrophobic membranes and, for L1 PEHs, longer PEG coronas. PEHs also possessed higher Hb encapsulation efficiencies compared to lipogel particles and NHPs [55].
102
5.2.3 Oxygen Binding Properties The P50 and cooperativity coefficients (n) of PEH dispersions are shown in Table 5.4. The oxygen dissociation curves of bovine Hb and PEH dispersions extruded through 200 nm pore diameter membranes are shown in Figure 5.3. Here, Y is the fraction of Hb saturated with oxygen, and pO2 is the partial pressure of oxygen in mmHg. The oxygen dissociation curves of PEH dispersions extruded through 400 nm pore diameter membranes are identical to those extruded through 200 nm pore diameter membranes. Since the hydrophobic membranes of polymersomes can be thicker than the membranes of liposomes and human RBCs, there was a concern that the permeability of the polymersomes to oxygen was significantly decreased and consequently, the oxygen binding capacity of PEH dispersions would be adversely affected. However, the P50 and n of PEH dispersions were comparable to both the measured P50 and n of bovine Hb (26 mmHg and 2.5, respectively) and the P50 and n of human RBCs (26 mmHg and 2.3, respectively) [24]. The oxygen dissociation curves of PEH dispersions exhibited a sigmoidal shape, similar to that of human erythrocytes [16]. These results indicated that the oxygen binding properties of the Hb was not compromised by encapsulation inside polymersomes [55]. This was true even for L1 PEH dispersions, which were expected to have ~35 nm hydrophobic membrane thickness. The P50 and n of PEH dispersions were also comparable to the values exhibited by PEG-LEHs [53], LEAcHbs [134] and ULEHs [39]. However, the P50 of lipogel particles and NHPs encapsulating Hb were reported to be lower than human RBCs (ranging from 9.9±1.9 to 14.4±0.1 mmHg [100]), which indicated an increase in the oxygen affinity of those oxygen carriers.
103
Figure 5.3. Oxygen dissociation curves of bovine Hb (panel A) and PEH dispersions extruded through 200 nm pore diameter membranes. Panel B: PEHs composed of 2212.6 kDa PBD-PEO; panel C: 5-2.3 kDa PBD-PEO; panel D: 2.5-1.3 kDa PBD-PEO; and panel E: 1.8-0.9 kDa PBD-PEO.
104
5.2.4 Encapsulated Methemoglobin Level The encapsulated metHb levels of PEH dispersions are presented in Table 5.4. PEH dispersions, regardless of the types of diblock copolymers and extrusion membrane pore sizes utilized, possessed encapsulated metHb levels less than 1%. The metHb level of the stock Hb solution used to create PEHs was 0.5-1%. This demonstrated that coencapsulated NAC effectively suppressed Hb oxidation. However, the metHb levels of PEG-LEHs were reported to be <3% [41, 116], although 5 mM of DL-homocysteine was coencapsulated. This was due to the peroxidation of the phospholipid membranes, which enhanced Hb oxidation [42], and the rigorous preparation procedure that the LEHs were subjected to in order to incorporate PEG chains onto the outer surface of the LEHs [41]. This indicates that, unlike liposomes [39, 42], the bilayers of polymersomes do not enhance Hb oxidation. Moreover, although the presence of chloride ions in the extrusion buffer was found to enhance Hb oxidation [38, 39], the presence of chloride ions in this study did not affect the metHb level of PEH dispersions. Although the metHb level of 272 nm diameter-LEAcHbs was <1%, the metHb level of 281 nm diameter-LEAcHbs was 7.4% [134]. Meanwhile, the metHb level of ~238 and ~260 nm diameter-ULEHs were 1.6% and 99%, respectively [39]. Hb oxidation in LEAcHbs and ULEHs was caused by Clmediated phospholipid peroxidation. The metHb levels of lipogel particles and NHPs containing Hb were even worse with values ranging from 9.3±3.7% to 26.0±5.0% [100].
105
5.3 Concluding Remarks Polymersomes offer 100% PEG surface coverage, longer PEG chains and thicker hydrophobic membranes, which contribute to improved intravascular circulation, biocompatibility, mechanical strength and colloidal state of PEH dispersions compared to PEG-LEH dispersions [47, 53, 56, 59, 73, 123]. Either possessing branched (2.5-1.3 kDa) or linear (22-12.6 kDa) PBD blocks, PBD-PEO diblock copolymers with higher molecular weights self-assembled into polymersomes in Hb solution, while copolymers with lower molecular weights (1.8-0.9 kDa and 5-2.3 kDa) formed a mixture of solid copolymer precipitates, polymersomes and micelles. Hence, we recommend utilization of diblock copolymers with higher molecular weights in future studies. The minimum molecular weight of copolymer required for complete formation of polymersomes is not known, and will have to be investigated in the future. PEH dispersions exhibited oxygen binding properties similar to human RBCs, even for PEHs possessing very thick hydrophobic membranes. This indicates that PEHs possess good potential as a cellular HBOC. PEH dispersions composed of 22-12.6 kDa and 2.5-1.3 kDa PBD-PEO copolymers exhibited higher Hb encapsulation efficiencies and lower encapsulated metHb levels compared to other potential cellular HBOCs, namely, PEG-LEHs, LEAcHbs, ULEHs, lipogel particles and NHPs. In addition, encapsulation of Hb into polymersomes did not induce metHb formation, hence low concentrations of coencapsulated metHb-reducing systems will be sufficient to maintain low encapsulated metHb levels, even for the duration of storage lifetime. These results [55] demonstrate the superior physical properties of PEHs compared to PEG-LEHs, which will facilitate their further development as a cellular HBOC.
106
CHAPTER 6
SIMULATION OF IN-VIVO OXYGEN TRANSPORT IN A CAPILLARY AND TO THE SURROUNDING TISSUE UTILIZING AN OXYGEN CARRIER
6.1 Introduction In this chapter, we simulated in-vivo oxygen transport in a capillary and to the surrounding tissue utilizing oxygen carriers with various oxygen binding properties. Simulations were based on the Krogh tissue cylinder model [66]. By analyzing the results of these simulations, we aimed to evaluate the efficacy of ULEH, PEG-LEH and PEH dispersions as a cellular HBOC for possible transfusion applications ranging from routine surgery to hemorrhagic shock. In-vivo oxygen transport was simulated under the following scenarios: • Simulations were conducted using the highest P50 and cooperativity coefficient (n), the average P50 and n, and the lowest P50 and n of ULEH, PEG-LEH and PEH dispersions [39, 53, 55], which were 38 mmHg and 2.9, 28 mmHg and 2.4, and 17 mmHg and 2, respectively. Cellular oxygen carriers with higher oxygen affinity could be developed by simply encapsulating polymeric Hb inside vesicles. For this case, simulations were conducted with a P50 of 7 mmHg and n of 2 [135].
107
Simulations utilizing the P50 and n of human RBCs (26 mmHg and 2.3, respectively) [24] were performed as a control. • Simulations were performed: under normal physiological conditions, where the oxygen partial pressure entering the capillary (pO2,inlet) is 100 mmHg [17]; during severe hypoxia (induced by more than 40% blood volume loss) where the pO2,inlet is below 40 mmHg (in this study, we used 30 mmHg); and under a significantly low pO2,inlet of 15 mmHg (such a low pO2,inlet is seldom encountered in actual practice, since patients suffering from severe hypoxia are already in critical condition (51% mortality rate) [136]) [65, 136, 137]. • Simulations were performed using the oxygen consumption rate of muscle tissue and the oxygen consumption rates of brain and pancreas tissues, which are the same [24, 138]. • For comparison, simulations were performed for the case where no oxygen carrier was present in the capillary, i.e. oxygen is present in the blood solely as dissolved oxygen.
6.2 Theoretical Background A bed of capillaries in tissue can be represented as a repetitive arrangement of capillaries surrounded by a cylindrical layer of tissue. The Krogh tissue model considers a single capillary in tissue as a cylinder with radius rc and length L, surrounded by a cylindrical layer of tissue with radius rT. Figure 6.1 shows a schematic of the Krogh tissue cylinder. We can use the Krogh tissue cylinder model to simulate oxygen transport in a capillary and its surrounding tissue. The following derivation of the oxygen 108
concentration profiles in a capillary and the surrounding tissue is taken from the literature [24, 139].
rT r rc
Capillary
r+∆r
V z z+∆z
Krogh tissue cylinder
Figure 6.1. Geometry of the Krogh tissue cylinder model.
The following assumptions are applied to our system in order to simplify the analysis: 1. The particulate nature of blood and the mass transfer resistance of RBCs are neglected. 2. The tissue region is treated as a continuous phase, ignoring the fact that the tissue is comprised of discrete cells. 3. Blood is assumed to be in plug flow, flowing through the capillary with an average velocity of V. Let us first consider the concentration of oxygen in blood flowing through the capillary. Note that oxygen is present in the blood as both dissolved oxygen and oxygen bound to Hb molecules, which are encapsulated in RBCs/ULEHs/PEG-LEHs/PEHs. 109
Only dissolved oxygen can diffuse through the capillary wall into the tissue region, where it is subsequently consumed by the tissue. Oxygenated Hb thus acts as a sink for dissolved oxygen. The conservation equation for species i in cylindrical coordinates is given by [139]:
⎡ 1 ∂ ⎛ ∂C i ∂C i ∂C i ν θ ∂C i ∂C i +ν r + +ν z = Di ⎢ ⎜r ∂t ∂r r ∂θ r ∂r ⎝ ∂r ∂z ⎣
2 2 ⎞ 1 ∂ Ci ∂ Ci ⎤ + 2 + ⎟ ⎥ + Ri 2 ∂z 2 ⎦ ⎠ r ∂θ
(6.1)
where t is time, z is the axial coordinate in the capillary, r is the radial coordinate in the capillary, θ is the angular coordinate in the capillary, and νr, νθ, νz are the velocities in the radial, angular and axial directions, respectively. Ci the concentration of species i, Di is the diffusivity of species i in the medium, while Ri is the production or consumption rate of species i. The first term on the left side of equation 6.1 is the accumulation rate of the species i, the last three terms on the left side of equation 6.1 describe the transport of species i by convection, the Di term on the right side of equation 6.1 describes the transport of species i via diffusion, while the last term on the right side of equation 6.1 describes the generation or consumption of species i. If species i is considered oxygenated Hb in the blood flowing through the capillary, Ci is C’ or the concentration of oxygen bound to Hb, while Ri is RHbO or the production rate of oxygenated Hb. There are no diffusive terms for oxygen bound to Hb, since oxygenated Hb is encapsulated in the RBC/ULEH/PEG-LEH/PEH dispersions and cannot pass through the membranes of these vesicles. Oxygenated Hb is carried inside these vesicles by blood, flowing through the capillary with an average velocity of V. After applying these assumptions, equation 6.1 is
110
simplified to obtain the mass balance for the concentration of oxygen bound to Hb in blood flowing through the capillary:
∂C' ∂C' +V = RHbO ∂t ∂z
(6.2)
For the case of dissolved oxygen in blood, Ci is C or the concentration of dissolved oxygen in blood, Di is D or the diffusivity of oxygen in blood, and Roxygen is the production rate of dissolved oxygen in blood. Here, dissolved oxygen is assumed to diffuse only in the z and r directions. Blood is assumed to be in plug flow with an average velocity of V. After applying these assumptions to equation 6.1, we obtain the mass balance for the dissolved oxygen concentration in blood flowing through the capillary:
⎡ 1 ∂ ⎛ ∂C ⎞ ∂ 2 C ⎤ ∂C ∂C +V = D⎢ ⎜r ⎟ + 2 ⎥ + Roxygen ∂t ∂z ⎣ r ∂r ⎝ ∂r ⎠ ∂z ⎦
(6.3)
Since the rate of generation of dissolved oxygen must be equal to the rate of disappearance of oxygen bound to Hb (remember that oxygenated Hb acts as a sink for dissolved oxygen in blood), RHbO = -Roxygen. We can relate the concentration of dissolved oxygen to the concentration of oxygen bound to Hb using the following equations:
111
∂C ∂C' ∂C ⎛ ∂C' ⎞ = ⎜ ⎟=m ∂t ⎝ ∂C ⎠ ∂t ∂t ∂C ∂C' ∂C ⎛ ∂C' ⎞ = ⎜ ⎟=m ∂z ⎝ ∂C ⎠ ∂z ∂z
(6.4)
where m is a dimensionless parameter, which is related to the slope of the oxygen dissociation curve. Equations 6.2 and 6.3 can be combined using equation 6.4 and Henry’s law to express the dissolved oxygen concentration in terms of the oxygen partial pressure of blood (pO2). Henry’s law is presented below:
pO2 = HC
(6.5)
where C is the dissolved oxygen concentration in blood and H is the Henry’s law constant of oxygen in the blood. Thus, the mass balance for dissolved oxygen in blood in terms of pO2 is:
⎡ 1 ∂ ⎛ ∂pO2 ⎞ ∂ 2 pO2 ⎤ ∂pO2 ∂pO2 (1 + m ) + V (1 + m ) = D⎢ ⎜r ⎟+ ⎥ ∂t ∂z ∂z 2 ⎦ ⎣ r ∂r ⎝ ∂r ⎠
(6.6)
The parameter m can be related to pO2 using Henry’s law and recognizing that the fraction of saturated Hb or Y = C’/C’sat, where C’sat is the saturated concentration of oxygen bound to Hb. Hence, m can be calculated using the following equation:
m=
dY dC' = HC' sat dpO2 dC
(6.7) 112
The term dY/d pO2 can be calculated using Hill’s equation:
( pO2 )n Y= n n P50 + ( pO2 )
(6.8)
where the P50 is the pO2 when 50% of the Hb’s oxygen-binding sites are filled with oxygen, and n is the cooperativity coefficient, which describes the cooperativity of oxygen binding to Hb. Therefore, equation 6.7 can be expressed as:
n m = nP50 HC' sat
(P
n pO2 −1
n 50
n + pO2 )
2
(6.9)
In the tissue region, Ci is CT or the oxygen concentration in the tissue, Di is DT or the diffusivity of oxygen in the tissue and Ri is Γmetabolic or the metabolic oxygen consumption rate in the tissue. Here, there is no oxygen transport by convection in the tissue region, and oxygen is assumed to diffuse only in the z and r directions. Applying these assumptions to equation 6.1, we obtain the mass balance for the oxygen concentration in the tissue region. Using Henry’s law, the mass balance can be expressed in terms of the oxygen partial pressure in tissue ( pO T ) as the following: 2
T T ⎡ 1 ∂ ⎛ ∂pO2 ∂pO2 ⎜r = DT ⎢ ⎜ ∂t ⎢ r ∂r ⎝ ∂r ⎣
T ⎞ ∂ 2 pO2 ⎟+ ⎟ ∂z 2 ⎠
⎤ T ⎥ − Γ metabolic H ⎥ ⎦
(6.10)
where HT is the Henry’s law constant of oxygen for the tissue region. 113
The boundary conditions for equations 6.6 and 6.10 are listed below: Blood region: BC 1: BC 2: 0 ≤ z ≤ L, 0 ≤ r ≤ rc z = 0, pO2 = pO2,inlet
r = 0, ∂pO 2 =0 ∂r
BC 3:
r = rc , pO 2 = pO T 2
and D
∂pO 2 ∂pO T 2 = DT ∂r ∂r
BC 1 states that blood enters the capillary with a known pO2. BC 2 assumes that the oxygen tension profile in the capillary is symmetric with respect to the center of the capillary. BC 3 assumes that the capillary wall has negligible mass transfer resistance, and states that the oxygen flux must be continuous at the interface between the capillary and the tissue region. Tissue region: BC 4: 0 ≤ z ≤ L, rc ≤ r ≤ rT z = 0, ∂pO T 2 =0 ∂z
BC 5:
∂pO T 2 =0 z = L, ∂z r = rT , ∂pO T 2 =0 ∂r
BC 6:
BC 4 and BC 5 state that oxygen cannot leave the tissue at both axial ends by axial diffusion. BC 6 assumes that the oxygen profile in the tissue region between capillaries a distance of 2rT apart is symmetrical (remember that a capillary is part of a bed of capillaries in the tissue). Simulations were conducted with the following physiological parameters [24]: • Length of capillary (L) = 0.1 cm. 114
• • • • •
Radius of capillary (rc) = 5 µm (the inner radius of a capillary). Average blood velocity flowing through the capillary (V) = 0.05 cm/s. The diffusivity of oxygen in tissue (DT) = 6.3 X 10-6 cm2/s. The diffusivity of oxygen in blood (D) = 0.91 X 10-5 cm2/s. Blood perfusion rate (q) = 2 mL blood/cm3 tissue/min. The Krogh tissue cylinder radius (rT) can be calculated using the following
formula:
rT = rc
V qL
(6.11)
Here, we calculated rT to be 20 µm. To obtain reasonable analytical solutions for equations 6.6 and 6.10, further simplifications need to be made to our system. We examined the steady-state solutions, hence eliminating the time derivatives in equations 6.6 and 6.10. The axial diffusion of oxygen in blood is assumed to be negligible compared to the axial convection of oxygen in blood. The validity of this assumption is verified through the following argument. The amount of oxygen transported by axial convection in blood is on the order of:
Oxygen transport by axial convection = πrc2VC o
(6.12)
115
Whereas the amount of oxygen transported by axial diffusion in blood is on the order of:
Oxygen transport by axial diffusion = πrc2
DC o L
(6.13)
where C0 is the oxygen concentration entering the capillary. The ratio of equation 6.12 to equation 6.13 is called the Peclet number (Pe), whose magnitude represents the importance of axial convection in comparison to axial diffusion. The assumption of ignoring axial diffusion of oxygen is valid if PeL = VL/D >> 1. In this study, we calculated PeL to be 550 >> 1. Thus, we can ignore axial diffusion of oxygen within the capillary in comparison to axial convection. Axial diffusion of oxygen in the tissue region is assumed to be negligible compared to radial diffusion of oxygen in the tissue. The amount of oxygen transported by radial diffusion in tissue is on the order of:
⎛ D T C0 Oxygen transport by radial diffusion = ⎜ ⎜r −r ⎝ T c
⎞ ⎟2π rc L ⎟ ⎠
(6.14)
While the amount of oxygen transported by axial diffusion in tissue is on the order of:
D T C0 π (rT2 − rc2 ) Oxygen transport by axial diffusion = L
(6.15)
116
The ratio of equation 6.14 to equation 6.15 describes the importance of radial diffusion in comparison to axial diffusion in the tissue region. The assumption is valid if the ratio >> 1.
2rc L2 Radial diffusion = Axial diffusion (rT − rc )(rT2 − rc2 )
(6.16)
In this study, we calculated the ratio to be 1,778 >> 1. Hence, we can ignore the axial diffusion of oxygen in comparison to the radial diffusion of oxygen in the tissue region. The metabolic oxygen consumption rate in the tissue region is assumed to follow Michaelis-Menten kinetics [24]:
Γ metabolic =
Vmax C T Km + CT
(6.17)
where CT is the concentration of oxygen in the tissue, Vmax is the maximum reaction rate and Km (Michaelis constant) is the oxygen concentration in the tissue when the reaction rate is one-half the maximum reaction rate. When CT >> Km, the maximum reaction rate occurs and the reaction rate is essentially independent of CT. For many biological reactions, this simplification typically applies, and we can reasonably assume that
Γmetabolic ≅ Vmax = a constant.
By ignoring the axial diffusion of oxygen in the capillary, and assuming that the system is at steady-state, equation 6.6 can be simplified to:
117
(1 + m )V ∂pO2
∂z
⎛ 1 ∂ ⎛ ∂pO2 = D⎜ ⎜ r ∂r ⎜ r ∂r ⎝ ⎝
⎞⎞ ⎟⎟ ⎟ ⎠⎠
(6.18)
By ignoring axial diffusion of oxygen in the tissue region, and assuming that the system is at steady-state, we can simplify equation 6.10 to:
T d ⎛ dpO2 ⎜r dr ⎜ dr ⎝
⎞ rΓ metabolic H T ⎟= ⎟ DT ⎠
(6.19)
In practical applications, the measured pO2 in the blood vessels is the radiallyaveraged pO2 values. This is obtained by integrating equation 6.18 from 0 to rc. The result is the following equation:
∂pO2 d c 2π (1 + m )V ∫ pO2 rdr = 2π Drc ∂r dz 0
r
(6.20)
rc
The radially-averaged pO2 in the blood at a given axial position in the capillary ( pO 2 )
is defined as:
2π pO2 =
rc
∫ pO rdr
2 0
π rc2
(6.21)
118
Applying BC 3 and writing equation 6.20 in terms of pO 2 , we obtain:
(1 + m )
d pO2 dz
=
2 D dpO2 Vrc dr
=
rc
T 2 D T dpO2 Vrc dr
(6.22)
rc
Note that pO 2 is a function of the axial position within the capillary. Applying BC 3 and BC 6 to equation 6.19, we obtain an equation which describes the partial pressure of oxygen within the tissue region at a given axial and radial position in the tissue:
2 rc2 Γ metabolic H T ⎡ ⎛ r ⎞ ⎤ rT2 Γ metabolicH T ⎛ r ⎞ ⎢1 − ⎜ ⎟ ⎥ − pO (r , z ) = pO2 ( z ) − ln⎜ ⎟ ⎜r ⎟ ⎜ ⎟ 4DT 2DT ⎢ ⎝ rc ⎠ ⎥ ⎝ c⎠ ⎦ ⎣ T 2
(6.23)
Now, we can solve equation 6.22 by taking the derivative of equation 6.23 with respect to
r and evaluate the derivative at rc. Substituting the derivative at rc into equation 6.22, and
integrating the resulting equation from 0 to L, we obtain an equation which describes the radially-averaged oxygen partial pressure in the blood at a given axial position in the capillary:
pO2 ( z ) = pO2,inlet
HT Γ − metabolic (1 + m)V
⎡⎛ r ⎢⎜ T ⎜ ⎢⎝ rc ⎣
2 ⎤ ⎞ ⎟ − 1⎥ z ⎟ ⎥ ⎠ ⎦
(6.25)
119
Under certain conditions, anoxic regions (regions where no oxygen is present) may develop in the tissue. In this case, the pO T is 0 mmHg. 2 We can also calculate the total amount of oxygen released (OR) into the tissue using the following equation:
⎡ pO ⎛ ( pO2 ,inlet )n ( pO2 ,out )n − pO2 ,out OR = ⎢ 2 ,inlet + C' sat ⎜ n − n n n ⎜ P + ( pO H P50 + ( pO2 ,out ) ⎢ 2 ,inlet ) ⎝ 50 ⎣
⎞⎤ ⎟⎥Vπ rc2 ⎟⎥ ⎠⎦
(6.26)
This equation takes into account both the concentration of dissolved oxygen and the concentration of oxygen bound to Hb. Here, pO2,out is the pO 2 at z = L. In this study, the oxygen partial pressure profiles within a capillary and its surrounding tissue exposed to various: oxygen carriers with various oxygen binding parameters (P50 and cooperativity coefficients), pO2,inlets, and oxygen consumption rates, and for the case where no oxygen carrier is present in blood (C’sat = 0) were simulated using MatLab (The MathWorks, Inc., Natick, MA, USA). We also calculated the total amount of oxygen released into the tissue for each oxygenation scenario. The MatLab code for the simulation is presented in Appendix I. The Henry’s law constant for normal blood at 37oC is 0.74 mmHg/µM (we assume that H = HT) and C’sat = 8800 µM. The oxygen consumption rate (Γmetabolic) in muscle tissue is 100 µM/s, while the oxygen consumption rate in brain and pancreas tissue is 25.9 µM/s [24, 138].
120
6.3 Results The oxygen dissociation curves of ULEH/PEG-LEH/PEH dispersions are shown in Figure 6.2. Figures 6.3 to 6.8 display the oxygen partial pressure profiles within a capillary and its surrounding tissue under various oxygenation scenarios. Tables 6.1 and 6.2 display the total amount of oxygen released in pancreas/brain and muscle tissue, respectively.
Figure 6.2. Oxygen dissociation curves of cellular oxygen carriers with P50 and cooperativity coefficients of 38 mmHg and 2.9 (− · −), 28 mmHg and 2.4 (⎯), and 17 mmHg and 2 (·····), respectively.
121
A. pO2,inlet = 100 mmHg
B. pO2,inlet = 30 mmHg
C. pO2,inlet = 15 mmHg
D. pO2,inlet = 100 mmHg
E. pO2,inlet = 30 mmHg
F. pO2,inlet = 15 mmHg
Figure 6.3. Oxygen partial pressure profiles (pO2) in a capillary and the surrounding Krogh tissue region for the case where no oxygen carrier is present in the blood. Z and r are the axial and radial coordinates in the capillary, respectively. Panels A, B and C represent pancreas and brain tissue, while panels D, E and F represent muscle tissue. 122
A. pO2,inlet = 100 mmHg
B. pO2,inlet = 30 mmHg
C. pO2,inlet = 15 mmHg
D. pO2,inlet = 100 mmHg
E. pO2,inlet = 30 mmHg
F. pO2,inlet = 15 mmHg
Figure 6.4. Oxygen partial pressure profiles (pO2) in a capillary and the surrounding Krogh tissue region simulated using the P50 and cooperativity coefficient of human RBCs (26 mmHg and 2.3, respectively). Z and r are the axial and radial coordinates in the capillary, respectively. Panels A, B and C represent pancreas and brain tissue, while panels D, E and F represent muscle tissue. 123
A. pO2,inlet = 100 mmHg
B. pO2,inlet = 30 mmHg
C. pO2,inlet = 15 mmHg
D. pO2,inlet = 100 mmHg
E. pO2,inlet = 30 mmHg
F. pO2,inlet = 15 mmHg
Figure 6.5. Oxygen partial pressure profiles (pO2) in a capillary and the surrounding Krogh tissue region simulated using the highest P50 and cooperativity coefficient observed for ULEH/PEG-LEH/PEH oxygen carriers (38 mmHg and 2.9, respectively). Z and r are the axial and radial coordinates in the capillary, respectively. Panels A, B and C represent pancreas and brain tissue, while panels D, E and F represent muscle tissue. 124
A. pO2,inlet = 100 mmHg
B. pO2,inlet = 30 mmHg
C. pO2,inlet = 15 mmHg
D. pO2,inlet = 100 mmHg
E. pO2,inlet = 30 mmHg
F. pO2,inlet = 15 mmHg
Figure 6.6. Oxygen partial pressure profiles (pO2) in a capillary and the surrounding Krogh tissue region simulated using the average P50 and cooperativity coefficient observed for ULEH/PEG-LEH/PEH oxygen carriers (28 mmHg and 2.4, respectively). Z and r are the axial and radial coordinates in the capillary, respectively. Panels A, B and C represent pancreas and brain tissue, while panels D, E and F represent muscle tissue. 125
A. pO2,inlet = 100 mmHg
B. pO2,inlet = 30 mmHg
C. pO2,inlet = 15 mmHg
D. pO2,inlet = 100 mmHg
E. pO2,inlet = 30 mmHg
F. pO2,inlet = 15 mmHg
Figure 6.7. Oxygen partial pressure profiles (pO2) in a capillary and the surrounding Krogh tissue region simulated using the lowest P50 and cooperativity coefficient observed for ULEH/PEG-LEH/PEH oxygen carriers (17 mmHg and 2, respectively). Z and r are the axial and radial coordinates in the capillary, respectively. Panels A, B and C: represent pancreas and brain tissue, while panels D, E and F represent muscle tissue. 126
A. pO2,inlet = 100 mmHg
B. pO2,inlet = 30 mmHg
C. pO2,inlet = 15 mmHg
D. pO2,inlet = 100 mmHg
E. pO2,inlet = 30 mmHg
F. pO2,inlet = 15 mmHg
Figure 6.8. Oxygen partial pressure profiles (pO2) in a capillary and the surrounding Krogh tissue region simulated using a high oxygen affinity-oxygen carrier (P50 and cooperativity coefficient of 7 mmHg and 2, respectively). Z and r are the axial and radial coordinates in the capillary, respectively. Panels A, B and C represent pancreas and brain tissue, while panels D, E and F represent muscle tissue. 127
TABLE 6.1
TOTAL AMOUNT OF OXYGEN RELEASED IN PANCREAS AND BRAIN TISSUES (µmol/s) EXPOSED TO VARIOUS OXYGEN CARRIERS WITH DIFFERENT OXYGEN BINDING PROPERTIES (P50 and COOPERATIVITY COEFFICIENT (n)), AND OXYGENATION SCENARIOS (pO2,inlet)
Oxygen binding properties of oxygen carriers No RBCs/oxygen carriers P50 = 26 mmHg, n = 2.3 P50 = 38mmHg, n = 2.9 P50 = 28 mmHg, n = 2.4 P50= 17 mmHg, n = 2 P50 = 7 mmHg, n = 2
pO2,inlet = 100 mmHg 5.3 x 10-9 1.9 x 10-8 2.0 x 10-8 1.9 x 10-8 1.7 x 10-8 1.1 x 10-8
pO2,inlet = 30 mmHg 1.6 x 10-9 2.8 x 10-8 3.1 x 10-8 2.9 x 10-8 2.6 x 10-8 1.9 x 10-8
pO2,inlet = 15 mmHg 8.0 x 10-10 3.4 x 10-8 2.3 x 10-8 3.7 x 10-8 2.9 x 10-8 2.5 x 10-8
TABLE 6.2
TOTAL AMOUNT OF OXYGEN RELEASED IN MUSCLE TISSUE (µmol/s) EXPOSED TO VARIOUS OXYGEN CARRIERS WITH DIFFERENT OXYGEN BINDING PROPERTIES (P50 and COOPERATIVITY COEFFICIENT (n)), AND OXYGENATION SCENARIOS (pO2,inlet)
Oxygen binding properties of oxygen carriers No RBCs/oxygen carriers P50 = 26 mmHg, n = 2.3 P50 = 38mmHg, n = 2.9 P50 = 28 mmHg, n = 2.4 P50 = 17 mmHg, n = 2 P50 = 7 mmHg, n = 2
pO2,inlet = 100 mmHg 5.3 x 10-9 5.0 x 10-8 5.7 x 10-8 5.1 x 10-8 4.2 x 10-8 1.0 x 10-8
pO2,inlet = 30 mmHg 1.6 x 10-9 1.0 x 10-7 1.2 x 10-7 1.1 x 10-7 8.1 x 10-8 5.0 x 10-8
pO2,inlet = 15 mmHg 8.0 x 10-10 7.7 x 10-8 2.3 x 10-8 6.4 x 10-8 1.5 x 10-7 7.3 x 10-8
128
6.4. Discussion Anoxic regions immediately developed in the capillary and tissue of the pancreas, brain and muscle, when the blood was devoid of human RBCs or artificial oxygen carriers (Figure 6.3). This was expected, since oxygen has a very low solubility in blood plasma or any aqueous media [17]. Because of this, the oxygen carrying capacity of blood plasma is very low, and as a result large anoxic regions formed in these tissues. When human RBCs were present in blood, anoxic regions did not develop in pancreas and brain tissues under normal physiological conditions (pO2,inlet = 100 mmHg), during severe hypoxia (pO2,inlet = 30 mmHg), and when a significantly low pO2,inlet of 15 mmHg was used (Figures 6.4.A, B and C, respectively). Under normal physiological conditions, muscle tissue was adequately oxygenated (Figure 6.4.D). However, an anoxic corner developed in muscle tissue exposed to conditions of severe hypoxia (Figure 6.4.E), and the muscle almost immediately became anoxic when the pO2,inlet was set at 15 mmHg (Figure 6.4.F). These results were expected, since the oxygen consumption rate in muscle tissue was four times the rate in pancreas and brain tissues. When the capillaries and tissues were oxygenated with an HBOC possessing an oxygen affinity lower than that of human RBCs, the pancreas and brain tissues were properly oxygenated under normal physiological conditions, and during severe hypoxia (Figures 6.5.A and B, respectively). However, a significant anoxic area developed when the pO2,inlet was set at 15 mmHg (Figure 6.5.C). The muscle tissue was also properly oxygenated under normal physiological conditions (Figure 6.5.D). However, the end corner of the muscle tissue became anoxic during severe hypoxia (Figure 6.5.E), and a large area of the muscle tissue became anoxic when the pO2,inlet was set at 15 mmHg 129
(Figure 6.5.F). Comparing these oxygen tension profiles with those of human RBCs (Figure 6.4), we observed that human RBCs were more effective in oxygenating pancreas, brain and muscle tissues experiencing severe hypoxia compared to HBOCs with low oxygen affinity. The oxygen tension profiles in pancreas, brain and muscle tissues oxygenated with oxygen carriers possessing an oxygen affinity comparable to human RBCs (Figure 6.6) were similar to the profiles of human RBCs (Figure 6.4), as expected. However, the muscle tissue was slightly better oxygenated when the pO2,inlet was set at 15 mmHg (Figure 6.6.F), compared to the case where tissue oxygenation was carried out by human RBCs (Figure 6.4.F), i.e. the anoxic region developed in muscle tissue was slightly smaller in size. The pancreas and brain tissues were properly oxygenated under normal physiological conditions, during severe hypoxia, and at the pO2,inlet of 15 mmHg when oxygen was transported by HBOCs with an oxygen affinity slightly higher than human RBCs (Figures 6.7.A, B and C, respectively). In agreement with the trend observed before, muscle tissue was properly oxygenated under normal physiological conditions, while an anoxic corner developed downstream during severe hypoxia (Figures 6.7.D and E, respectively). A significant area of the muscle tissue became anoxic when the pO2,inlet was set at 15 mmHg (Figure 6.7.F). However, the muscle tissue was better oxygenated compared to the case where tissue oxygenation was carried out by human RBCs (Figure 6.4.F) and by oxygen carriers with oxygen affinity comparable to human RBCs (Figure 6.6.F).
130
When the oxygen affinity of the oxygen carrier was further increased, the pancreas and brain tissues were well oxygenated under normal physiological conditions, during severe hypoxia and when the pO2,inlet was set at 15 mmHg (Figures 6.8.A, B and C, respectively). Interestingly, the muscle tissue rapidly became anoxic under normal physiological conditions if the oxygen affinity of the oxygen carrier was significantly higher than that of human RBCs (Figure 6.8.D). This oxygen partial pressure profile was almost identical to the profile where no oxygen carrier was present in the blood (Figure 6.3.D). Since the oxygen affinity of the HBOCs was very high, it was difficult for the HBOCs to release oxygen at high pO2s (see the oxygen dissociation curves in Figure 6.2). During severe hypoxia (Figure 6.8.E), the oxygenation of muscle tissue was worse compared to the previous case (Figure 6.7.E), i.e. a larger anoxic corner formed in the muscle tissue. When pO2,inlet of 15 mmHg was used, a significant area of the muscle tissue was anoxic (Figure 6.8.F). However, the muscle tissue was better oxygenated compared to the case where tissue oxygenation was carried out by human RBCs (Figure 6.4.F) and by oxygen carriers with oxygen affinity comparable to human RBCs (Figure 6.6.F). As expected, the amount of oxygen released into the tissues when the blood was devoid of RBCs or artificial oxygen carriers was less than the oxygen unloading when the blood carries RBCs or oxygen carriers (Tables 6.1 and 6.2). The amount of oxygen released in the pancreas and brain decreased with increasing oxygen affinity under normal physiological conditions, and during severe hypoxia (Table 6.1). Human RBCs and oxygen carriers with oxygen affinity comparable to human RBCs released the highest concentration of oxygen in the pancreas and brain when the pO2,inlet was set at 15 mmHg. 131
Likewise, the amount of oxygen released in the muscle decreased with increasing oxygen affinity under normal physiological conditions, and during severe hypoxia (Table 6.2). When the pO2,inlet was set at 15 mmHg, oxygen carriers with oxygen affinity slightly higher than human RBCs (P50 = 17 mmHg and n = 2) released the highest amount of oxygen in the muscle . Based on the oxygen partial pressure profiles and the amount of oxygen released in the tissues, we observed that HBOCs with oxygen affinities similar to human RBCs are most suitable for oxygenating tissues under normal physiological conditions. Whereas, HBOCs with oxygen affinities higher than human RBCs are most suitable for oxygenating tissues under hypoxic conditions, in order to target oxygen delivery to low pO2 tissues. Our simulation results are in agreement with results reported in the literature [65, 140-142]. Theoretical models of oxygen transport and animal studies [65, 140, 141] suggested that HBOCs with high oxygen affinity (P50 less than 20 mmHg [65]) were suitable for resuscitating patients experiencing severe hypoxia. The optimum P50 for treating severe hypoxia was found to range between 15-16 mmHg [65, 140]. In contrast, resuscitation of hamsters with PEG-LEH dispersions (P50 = 9 mmHg) resulted in decreased microvascular perfusion [140]. In order to reduce the occurrence of anoxic regions induced by the presence of an oxygen carrier possessing a P50 of 9 mmHg, blood flow is shunted away from the microcirculation. This behavior was predicted by our simulations, since muscle tissue oxygenated with oxygen carriers possessing a P50 of 17 mmHg (Figure 6.7.E) exhibited less anoxic regions compared to muscle tissue oxygenated with oxygen carriers possessing a P50 of 7 mmHg (Figure 6.8.E). 132
It was generally believed by the blood substitute community that lower oxygen affinity HBOCs (compared to the oxygen affinity of human RBCs) were appropriate for oxygenating tissues under normal physiological conditions [65]. However, recent research showed that low oxygen affinity HBOCs over-oxygenated tissues, which lead to autoregulatory induced vasoconstriction and decreased functional capillary density in the peripheral circulation [142]. Hence, based on our results and literature [65, 140-142], we suggest that HBOCs with oxygen affinity comparable to that of human RBCs are appropriate for transfusion applications in routine surgery, in order to maintain proper oxygenation of tissues under normal physiological conditions. Whereas, oxygen carriers with high oxygen affinity are suitable for transfusion applications in hemorrhagic shock (where the pO2,inlet drops to be below 40 mmHg).
6.5 Concluding Remarks Taken together, our results show that a cellular HBOC possessing an oxygen affinity comparable to that of human RBCs (P50 = 26 mmHg and cooperativity coefficient = 2.3 [24]) is suitable for transfusion applications in routine surgery. However, for transfusion applications in trauma, a cellular HBOC with an oxygen affinity higher than that of human RBCs is more effective at delivering oxygen to oxygen-depleted regions of the body. Based on our research and literature reports [65, 140], the optimum P50 of oxygen carriers employed to resuscitate trauma patients should be ~ 15-17 mmHg. Cellular HBOCs with oxygen affinity comparable to human RBCs were readily prepared through encapsulation of unmodified Hb inside liposomes or polymersomes [39, 53-55].
133
While cellular HBOCs possessing high oxygen affinity can be created by encapsulation of polymeric Hb inside liposomes or polymersomes.
134
CHAPTER 7
CONCLUSIONS AND PROPOSED FUTURE STUDIES
7.1 Conclusions The oxygen binding properties of unmodified liposome encapsulated Hb (ULEH) dispersions were comparable to that of human RBCs, which indicated that ULEH dispersions displayed good potential as a cellular HBOC [39]. We observed that only unmodified liposomes with diameters ranging from 100 to 300 nm could be prepared via extrusion [39]. ULEH dispersions extruded in phosphate buffer (PB) possessed higher Hb encapsulation efficiencies, while the size distributions of ULEH dispersions extruded in phosphate buffered saline (PBS) were more monodisperse. However, the physical
integrity of ULEHs in PBS at pH 7.3 and 37oC (which mimics the salinity, pH and temperature of blood plasma) was unstable post-production, due to osmosis of water across the lipid bilayers into and out of the liposome core. This observation implies that ULEH dispersions will be osmotically fragile in the blood stream. In addition, our results suggested that encapsulated Hb molecules interacted directly with the lipid bilayers, thus rendering ULEH dispersions susceptible to the influence of osmotic pressure gradient [39].
135
This potential problem can be addressed by grafting poly(ethylene glycol) (PEG) molecules onto the lipid bilayers, which suppresses the expansion and contraction of lipid bilayers [53]. Moreover, we observed that PEGylation of LEH dispersions strengthened the bilayers against the deteriorating interaction with encapsulated Hb molecules previously observed with ULEH dispersions [39]. Similar to unmodified liposomes, only PEGylated liposomes with diameters ranging from 140 to 350 nm could be prepared via extrusion [53]. The oxygen binding properties of PEG-LEH dispersions were comparable to that of human RBCs, which indicated that PEG-LEH dispersions displayed good potential as a cellular HBOC [53]. However, the Hb encapsulation efficiencies of these vesicles were low, except for LEH dispersions grafted with 2 kDa PEG and extruded in PBS. We speculated that this was caused by the collapse of PEG chains conjugated onto the inner leaflets of liposomes into a mushroom-like conformation, which liberated more liposome aqueous core volume for Hb encapsulation. Opposite to the trend observed for ULEH dispersions, the size distributions of PEG-LEH dispersions extruded in PB were more homogeneous compared to those extruded in PBS. Moreover, the PEG surface coverage and length of PEG brushes that can be stably conjugated onto the liposome bilayer are severely limited [45, 46]. Thus, there is a limit to optimizing the intravascular circulation, biocompatibility and colloidal state of PEG-LEH dispersions. In order to alleviate these limitations, we employed a novel type of polymer vesicle, namely polymersomes, as an oxygen carrier [55]. Various studies [56-58] have demonstrated the superior PEG shielding ability of polymersomes compared to PEG-liposomes. The shielding ability of polymersomes is primarily attributed to the 100% PEG surface coverage of these vesicles, and the fact that
136
the molecular weight of these PEG chains can be synthesized to be quite large. These unique properties contribute to the long intravascular lifetime, high level of biocompatibility and improved colloidal state (which translates into longer storage lifetime) of polymersomes [47, 56, 73]. Since polymersomes can be synthesized with thicker hydrophobic membranes compared to liposomes, polymersomes are mechanically stronger than liposomes [56, 123, 125, 143]. In our studies, we observed that PEH dispersions possessed higher Hb encapsulation efficiencies compared to ULEH, PEGLEH, PEGylated actin-containing LEH, lipogel particle and nanoscale hydrogel particle dispersions loaded with Hb [55]. Unlike phospholipid bilayers, the membranes of polymersomes did not enhance oxidation of Hb encapsulated in the aqueous core of polymersomes [55]. Although the membranes of polymersomes are thicker than phospholipid bilayers, the membranes are permeable to oxygen and allow oxygen transport across the membranes. PEH dispersions were shown to have oxygen binding properties comparable to that of human RBCs [55]. Simulation of in-vivo oxygen transport in a capillary and to surrounding tissues (Chapter 6) demonstrated that PEH dispersions can be engineered for transfusion specific applications ranging from blood replacement during routine surgery to treatment of hemorrhagic shock. In conclusion, we observed that PEH dispersions were readily prepared, and exhibited good potential as a cellular HBOC, while offering superior physical properties, which may alleviate the limitations and problems encountered with current cellular-based oxygen carriers, such as, ULEH, PEG-LEH, PEGylated actin-containing LEH, lipogel particle and nanoscale hydrogel particle dispersions loaded with Hb [55]. However, the
137
current state of research on the potential of PEH dispersions to function as an effective oxygen carrier is at an early stage, and further work will be required in order to develop PEH dispersions into a reliable cellular HBOC.
7.2 Proposed Future Studies Our research plan for the next five years in the continuing development of an effective and reliable cellular HBOC is summarized below. In our present work, we investigated the physical properties of PEH dispersions composed of PBD-PEO diblock copolymers with molecular weights of 22-12.6, 5-2.3, 2.5-1.3 and 1.8-0.9 kDa. However, 5-2.3 and 1.8-0.9 kDa PBD-PEO copolymers formed a mixture of PEH dispersions, solid copolymer precipitates and micelles in Hb solution, while 22-12.6 and 2.5-1.3 kDa PBDPEO copolymers completely dissolved in Hb solution to form PEH suspensions [55]. In an unpublished study, we observed that 11.8-5.3 kDa PBD-PEO copolymer (with a linear PBD block) did not dissolve in aqueous solution. Hence, we propose to expand our investigation by preparing PEH dispersions composed of PBD-PEO copolymers with structural properties displayed in Table 7.1. The hydrophobic membrane thickness (d) of polymersomes in Table 7.1 were extrapolated from the literature [59].
138
TABLE 7.1
TYPE OF PBD BLOCK, NUMBER-AVERAGED MOLECULAR WEIGHT (Mn), PEG LENGTH, HYDROPHILIC MASS FRACTION (fhydrophilic), AND HYDROPHOBIC MEMBRANE THICKNESS (d) OF PBD-PEO DIBLOCK COPOLYMERS
PBD-PEO (kDa) 70.0-30.0 70.0-30.0 58.3-25.0 58.3-25.0 46.7-20.0 46.7-20.0 23.3-10.0 23.3-10.0 11.7-5.0
PBD block linear branched linear branched linear branched linear branched branched
Mn (kDa) 100.0 100.0 83.3 83.3 66.7 66.7 33.3 33.3 16.7
PEG length (kDa) 30.0 30.0 25.0 25.0 20.0 20.0 10.0 10.0 5.0
fhydrophilic 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3
d (nm) ~74 ~74 ~63 ~63 ~52 ~52 ~29 ~29 ~17
PEH dispersions prepared from these copolymers will possess a variety of membrane thicknesses and PEG lengths. PEH dispersions will be prepared using a film rehydration technique in 300 mg/mL Hb solution, followed by extrusion through 200 and 400 nm pore diameter membranes to regulate the size of the dispersions [55]. PEH dispersions will be engineered for specific transfusion applications ranging from blood replacement during routine surgery to treatment of hemorrhagic shock. For transfusion applications in routine surgery, PEH dispersions must possess an oxygen affinity comparable to that of human RBCs, which can be achieved by encapsulating bovine Hb inside polymersomes. However, for transfusion applications in hemorrhagic shock, PEH dispersions with high oxygen affinity will be required (see Chapter 6). This type of PEH 139
dispersion will be created by encapsulating polymeric Hb, which exhibit high oxygen affinity [135], inside polymersomes. Both types of PEH dispersions will be prepared in future studies. Based on the circulation half-lives of empty polymersomes in rats measured by Photos et al. [47], we extrapolated the circulation half-lives of PEH dispersions (derived from the entries in Table 7.1) in rats (Figure 7.1).
Figure 7.1. Circulation half-life of PEH dispersions in rats (t1/2) versus length of PEG block. The data represented by the solid circles were collected by Photos et al. [47], while the solid squares represent PEH entries in Table 7.1 that were extrapolated from Photos et al.’s data [47].
We estimate that the circulation half-lives of these PEH dispersions (~31-128 hours in rats) will surpass those of PEG-LEH dispersions (maximum of 48 hours in rabbits [34] or ~36 hours in rats [50]). In addition, these PEH dispersions will possess thicker hydrophobic membranes (~17-74 nm) compared to liposomes (3-4 nm) [17]. We will select copolymers from the entries of Table 7.1, which completely dissolve in Hb solution to self-assemble into PEH particles. 140
The in-vitro physical properties of the selected PEH dispersions will be evaluated in order to determine their potential as an oxygen carrier. The physical properties of interest are: vesicle size distribution, Hb encapsulation efficiency, oxygen affinity (as indicated by P50 and cooperativity coefficient), encapsulated metHb level, and hydrophobic membrane thickness. The first four physical properties will be measured according to the procedures described in Chapter 2. Empty polymersomes will be imaged by cryogenic transmission electron microscopy, and the membrane thickness of polymersomes will be measured [144]. A recent study suggested that transmembrane proteins could incorporate into the bilayers of polymersomes [130]. This presents a concern that potentially toxic Hb may be trapped on the surface of PEHs. In this case, infusion of PEHs transporting toxic Hb ligands on the surface into mammals may cause an adverse immune response. In order to avoid this potential problem, we will improve the purification process of PEH dispersions by employing a Haptoglobin (Hp)-affinity column in conjunction with the dialysis procedure used in Chapters 3-5 [39]. Hp is a human serum protein, which forms a stable complex with Hb [145]. The dialysis procedure will remove the bulk of unencapsulated Hb molecules from the dispersion, while further purification of impure PEH dispersions with a Hp–affinity column will capture trace amount of unencapsulated Hb molecules in the dispersion. Based on the analysis of the in-vitro physical properties of PEH dispersions, we will select PEH dispersions which satisfy the design criteria of a cellular HBOC (the design criteria was presented in Chapter 1). These selected PEH dispersions will be subjected to a series of in-vitro biocompatibility tests to evaluate their potential to activate the complement system and the vesicles’ interaction with human cells. In
141
addition, the level of endotoxin contaminant in each PEH dispersion will be determined. Endotoxin, also called LPS, is the component of the outer membrane of gram-negative bacteria, which is released into free solution upon disruption of intact bacteria [146]. Endotoxin is known to activate the complement system, and may compromise the results of the in-vitro complement activation assay [146]. The activation of the complement system in-vivo is associated with anaphylactic reactions, and activation of other proteolytic plasma cascades [51]. The capacity of each PEH dispersion to activate the complement system will be evaluated using the sheep erythrocyte hemolysis assay [147, 148]. Furthermore, PEH dispersions will be allowed to interact with cultured human aortic endothelial cells and vascular smooth muscle cells. These cell lines were selected, since they comprise the cells lining blood vessels, which will directly interact with PEH dispersions in-vivo. The level of PEH biocompatibility will be determined by testing the viability of these cells after incubation with PEH dispersions [47, 149]. PEH dispersions, which exhibit low complement activating potential, and high level of human cellbiocompatibility, will be selected for subsequent animal studies. These selected PEH dispersions will be fluorescently labeled with a biocompatible dye and infused into Sprague-Dawley rats, in order to determine their circulation half-life [47]. Fluorescence microscopy will be used to quantify the concentration of polymersomes in the systemic circulation [47]. The time required for removal of 50% of the initially injected PEH dispersion from the systemic circulation will be considered the circulatory half-life of the PEH dispersion. During these studies, the health of the rats will be observed to determine the presence of any toxic effects related to PEH administration. Furthermore, the viscosities and oncotic pressures of PEH
142
dispersions will be engineered to be 4 cP and 25 mmHg, respectively (for transfusion applications in blood replacement during routine surgery), or 4 cP and 50 mmHg (for transfusion applications in treatment of hemorrhagic shock), respectively, through adjustment of the PEH concentration and addition of human serum albumin (HSA) [20, 41, 63, 150]. For applications in routine surgery, the oncotic pressure of the PEH dispersion will be prepared to be identical to that of human blood in order to maintain normal vascular blood volume [41, 106]. However, for applications in hemorrhagic shock, the oncotic pressure of the PEH dispersion should be higher than that of human blood in order for the dispersion to act as a plasma expander [20, 63]. PEH/HSA dispersions will be used to resuscitate Sprague-Dawley rats, in order to evaluate the dispersions’ efficacy as an oxygen carrier in routine surgery and hemorrhagic shock [41, 63, 106]. To simulate the physiological conditions present in routine surgery, rats will be subjected to 90% exchange blood transfusion with a PEH/HSA dispersion [41]. Whereas to simulate the physiological conditions present in hemorrhagic shock, 50% of the rat’s blood will be withdrawn and after some specified amount of time has elapsed the removed blood will be subsequently replaced with an equal volume of PEH/HSA dispersion [63]. The following systemic and microvascular responses of rats will be continuously monitored in both cases: mean arterial pressure, heart rate, blood flow rate, blood pH, oxygen partial pressure and CO2 partial pressure [41, 63, 106]. PEH/HSA dispersions that exhibit the best potential as a cellular HBOC, will be those that exhibit long circulatory half-life, induce no significant elevation in mean arterial pressure after resuscitation, and are able to maintain normal heart rate, blood flow rate, blood pH, oxygen and CO2 blood tensions after resuscitation. 143
In addition to these studies, we will also design alternative PEH dispersions, which may improve the in-vitro and in-vivo properties of PEH dispersions as an oxygen carrier. Although coencapsulation of 5 mM of NAC was proven to be effective in maintaining encapsulated metHb levels of PEH dispersions at less than 1% [55], the metHb reduction system should also be capable of maintaining low metHb levels in-vivo until the vesicles are cleared from the blood circulation and in-vitro during the PEH dispersion storage lifetime. Therefore, we propose an alternate method to suppress Hb oxidation using an enzymatic-reduction system [26, 151]. This system imitates the Embden-Meyerhoff reduction pathway in human RBCs by coencapsulation of coenzyme nicotinamide adenine dinucleotide (NAD) and the substrates glucose, adenine and inosine in the aqueous core of polymersomes [26, 151]. Ogata et al. [26, 151] reported that this system was capable of reducing the rate of metHb formation from 1% per hour to 0.5% per hour in-vitro, and from 1.5% per hour to 0.7% per hour in-vivo. Transfused PEH dispersions must be able to withstand the shear forces in the circulatory system, which can induce premature lysis of PEHs in-vivo. The hydrophobic membrane of polymersomes can be cross-linked via free-radical polymerization to further strengthen the membrane [152, 153]. Discher et al. [124] successfully created polymersomes with cross-linked membranes using PBD-PEO diblock copolymers. Our study (Chapter 5) showed that the Hb encapsulation efficiencies of PEHs decreased with increasing PEO blocks, since the PEO chains are also present in the inner leaflets of the polymersomes. This potential problem can be solved by engineering PEHs with asymmetric membranes [154, 155]. In this design, the outer leaflets of the polymersomes will be assembled from an amphiphilic copolymer with long PEO chains,
144
while the inner leaflets will be assembled from an amphiphilic copolymer with short PEO chains. Therefore, PEHs possessing a high Hb encapsulation efficiency and long PEO corona can be produced. Moreover, the choice for the hydrophobic core chemistry of polymersomes is flexible, since the intravascular half-life of polymersomes appeared to be mainly dependent on the molecular weight of the PEG brush, and not on the hydrophobic core thickness, due to the obfuscation of the hydrophobic core by the dense PEG brush [47]. Therefore, if the PBD block of polymersomes is observed to induce cytotoxic effects, the PBD hydrophobic block of the PEO-PBD diblock copolymers can be replaced with other biocompatible hydrophobic polymers, such as poly(lactide) [156].
145
REFERENCES
1. World Health Organization. www.who.int, (2005). 2. T. M. S. Chang, A New Red Blood Cell Substitute. Crit. Care Med., 32: 612-613 (2004). 3. Y. O. Schumacher and M. Ashenden, Doping with Artificial Oxygen Carriers. Sports Med., 34: 141-150 (2004). 4. A. I. Alayash, Oxygen Therapeutics: Can We Tame Haemoglobin? Nature, 3: 152-159 (2004). 5. T. Standl, Artificial Oxygen Carriers: Hemoglobin-Based Oxygen Carriers - Current Status 2004. Transfus. Med. Hemother., 21: 262-268 (2004). 6. National Blood Data Resource Center. www.nbdrc.org, (2005). 7. North Central Blood Services. www.yourbloodcenter.org, (2005). 8. R. M. Winslow, Blood Substitutes. Adv. Drug Delivery Reviews, 40: 131-142 (2000). 9. H. G. Klein, Blood Substitutes: How Close to a Solution? Oncology, 16: 147-151 (2002). 10. W. J. Williams, E. Beutler, A. J. Erslev and R. W. Rundles, Preservation and Clinical Use of Blood and Blood Components. Hematology. McGraw-Hill Book Company, New York (1977). 11. R. Rabinovici, The Status of Hemoglobin-Based Red Cell Substitutes. Reviews, 3: 691-697 (2001). 12. B. T. Kjellstrom, Blood Subtitutes: Where Do We stand Today? J. Intern. Med., 253: 495-497 (2003). 13. American Association of Blood Banks. www.aabb.org, (2005). 14. Minnesota Department of Health. www.health.state.mn.us, (2005). 146
15. S. E. Hill, Oxygen Therapeutics-Current Concepts. Can. J. Anesth., 48: S32-S40 (2001). 16. C. Smith, M. Lieberman and A. Marks, Marks' Basic Medical Biochemistry: A Clinical Approach. Lippincott Williams & Wilkins, Philadelphia (2004). 17. D. Voet and J. G. Voet, Biochemistry. John Wiley & Sons, Inc., New York (2003). 18. K. C. Lowe and E. Ferguson, Benefit and Risks Perceptions in Transfusion Medicine: Blood and Blood Substitutes. J. Intern. Med., 253: 498-507 (2003). 19. C. P. Stowell, Hemoglobin-Based Oxygen Carriers. Curr. Opin. Hematol., 9: 537-543 (2002). 20. R. M. Winslow, Current Status of Blood Substitute Research: Towards a New Paradigm. J. Intern. Med., 253: 508-517 (2003). 21. P. E. Keipert, Oxygen Therapeutics ("Blood Substitutes"). Where Are They, and What Can We Expect? Adv. Exp. Med. Biol., 540: 207-213 (2003). 22. E. Niiler, Setbacks for Blood Substitutes Companies. Nature Biotechnol., 20: 962-963 (2002). 23. E. Maevsky, G. Ivanitsky, L. Bogdanova, O. Axenova, N. Karmen, E. Zhiburt, R. Senina, S. Pushkin, I. Maslennikov, A. Orlov and I. Marinicheva, Clinical Results of Perftoran Application: Present and Future. Art. Cells, Blood Subs., and Biotech., 33: 37-46 (2005). 24. R. L. Fournier, Basic Transport Phenomena in Biomedical Engineering. Taylor & Francis, Philadelphia (1999). 25. J. G. Riess, Oxygen Carriers ("Blood Substitutes")-Raison d'Etre, Chemistry and Some Physiology. Chem. Rev., 101: 2797-2919 (2001). 26. Y. Ogata, Evaluation of Human Hemoglobin Vesicle as an Oxygen Carrier: Recovery from Hemorrhagic Shock in Rabbits. Polym. Adv. Technol., 11: 301-306 (2000). 27. B. Faivre, P. Menu, P. Labrude and C. Vigneron, Hemoglobin Autooxidation/Oxidation Mechanisms and Methemoglobin Prevention or Reduction Processes in the Bloodstream. Art. Cells, Blood Subs., and Immob. Biotech., 26: 17-26 (1998). 28. E. Moore, Blood Subtitutes: The Future is Now. J. Am. Coll. Surgeons, 196: 1-17 (2003).
147
29. G. Allison and C. Feeney, Successful Use of a Polymerized Hemoglobin Blood Substitute in a Critically Anemic Jehovah's Witness. Southern Med. J., 97: 12571258 (2004). 30. A. S. Rudolph, Biomaterial Biotechnology Using Self-Assembled Lipid Microstructures. J. Cell. Biochem., 56: 183-187 (1994). 31. A. S. Rudolph, A. Sulpizio, P. Hieble, V. MacDonald, M. Chavez and G. Feuerstein, Liposome Encapsulation Attenuates Hemoglobin-Induced Vasoconstriction in Rabbit Arterial Segments. J. Applied Phys., 82: 1826-1835 (1997). 32. T. M. S. Chang, Future Prospects for Artificial Blood. Trends Biotechnol., 17: 61-67 (1999). 33. H. Sakai, K. Tomiyama, K. Sou, S. Takeoka and E. Tsuchida, Poly(ethylene glycol)Conjugation and Deoxygenation Enable Long-Term Preservation of HemoglobinVesicles as Oxygen Carriers in a Liquid State. Bioconjugate Chem., 11: 425-432 (2000). 34. W. T. Phillips, R. W. Klipper, V. D. Awasthi, A. S. Rudolph, R. Cliff, V. Kwasiborski and B. A. Goins, Polyethylene Glycol-Modified LiposomeEncapsulated Hemoglobin: A Long Circulating Red Cell Substitute. J. Pharmacol. Exp. Ther., 288: 665-670 (1999). 35. R. O. Wright, B. Magnani, M. W. Shannon and A. D. Woolf, N-Acetylcysteine Reduces Methemoglobin in Vitro. Ann. of Emergency Med., 28: 499-503 (1996). 36. Y. Teramura, H. Kanazawa, H. Sakai, S. Takeoka and E. Tsuchida, Prolonged Oxygen-Carrying Ability of Hemoglobin Vesicles by Coencapsulation of Catalase in Vivo. Bioconjugate Chem., 14: 1171-1176 (2003). 37. S. Takeoka, H. Sakai, T. Kose, Y. Mano, Y. Seino, H. Nishide and E. Tsuchida, Methemoglobin Formation in Hemoglobin Vesicles and Reduction by Encapsulated Thiols. Bioconjugate Chem., 8: 539-544 (1997). 38. W. J. Wallace, J. C. Maxwell and W. S. Caughey, A Role for Chloride in the Autoxidation of Hemoglobin under Conditions Similar to Those in Erythrocytes. FEBS Lett., 43: 33-36 (1974). 39. D. R. Arifin and A. F. Palmer, Determination of Size Distribution and Encapsulation Efficiencies of Liposomes Encapsulated Hemoglobin by Asymmetric Flow FieldFlow Fractionation and Multi-Angle Light Scattering. Biotech. Prog., 19: 17981811 (2003). 40. J. Israelachvili, Intermolecular and Surface Forces. Academic Press, Inc., San Diego (1995). 148
41. H. Sakai, S. Takeoka, S. I. Park, T. Kose, H. Nishide, Y. Izumi, A. Yoshizu, K. Kobayashi and E. Tsuchida, Surface Modification of Hemoglobin Vesicles with Poly(ethylene glycol) and Effects on Aggregation, Viscosity and Blood Flow during 90% Exchange Transfusion in Anesthesized Rats. Bioconjugate Chem., 8: 23-30 (1997). 42. M. C. Farmer and B. P. Gaber, Liposome-Encapsulated Hemoglobin as an Artificial Oxygen-Carrying System. Methods Enzymol., 149: 184-200 (1987). 43. D. Bhadra, S. Bhadra, P. Jain and N. K. Jain, Pegnology: a Review of PEG-ylated Systems. Pharmazie, 57: 5-29 (2002). 44. T. Ishida, R. Maeda, M. Ichihara, K. Irimura and H. Kiwada, Accelerated Clearance of PEGylated Liposomes in Rats after Repeated Injections. J. Contr. Release, 88: 35-42 (2003). 45. F. K. Bedu-Addo, P. Tang, Y. Xu and L. Huang, Effect of Polyethyleneglycol Chain Length and Phospholipid Acyl Chain Composition in the Interaction of Polyethyleneglycol-Phospholipid Conjugates with Phospholipid: Implications in Liposomal Drug Delivery. Pharm. Res., 13: 710-717 (1996). 46. K. Shimada, S. Matsuo, Y. Sadzuka, A. Miyagishima, Y. Nozawa, S. Hirota and T. Sonobe, Determination of Incorporated Amounts of Poly(ethylene glycol)Derivatized Lipids in Liposomes for the Physicochemical Characterization of Stealth Liposomes. Int. J. Pharm., 203: 255-263 (2000). 47. P. J. Photos, L. Bacakova, B. Discher, F. S. Bates and D. E. Discher, Polymer Vesicles In Vivo: Correlations with PEG Molecular Weight. J. Contr. Release, 90: 323-334 (2003). 48. A. Chonn, S. C. Semple and P. R. Cullis, Association of Blood Proteins with Large Unilamellar Liposomes In Vivo - Relation to Circulation Lifetime. J. Biol. Chem., 267: 18759-18765 (1992). 49. J. Lee, J. Kopecek and J. D. Andrade, Protein-Resistant Surfaces Prepared by PEGContaining Block Copolymer Surfactants. J. Biomed. Mater. Res., 23: 351-368 (1989). 50. V. D. Awasthi, personal communication, (2003). 51. S. M. Moghimi and J. Szebeni, Stealth Liposomes and Long Circulating Nanoparticles: Critical Issues in Pharmacokinetics, Opsonization and ProteinBinding Properties. Prog. Lipid Res., 42: 463-478 (2003). 52. H. Sakai, M. Yuasa, H. Onuma, S. Takeoka and E. Tsuchida, Synthesis and Physicochemical Characterization of a Series of Hemoglobin-Based Oxygen 149
Carriers: Objective Comparison between Cellular and Acellular Types. Bioconjugate Chem., 11: 56-64 (2000). 53. D. R. Arifin and A. F. Palmer, Physical Properties and Stability Mechanisms of Poly(ethylene glycol) Conjugated Liposome Encapsulated Hemoglobin Dispersions. Art. Cells, Blood Subs., and Biotech., 33: 137-162 (2005). 54. D. R. Arifin and A. F. Palmer, Stability of Liposome Encapsulated Hemoglobin Dispersions. Art. Cells, Blood Subs., and Biotech., 33: 113-136 (2005). 55. D. R. Arifin and A. F. Palmer, Polymersomes Encapsulated Hemoglobin: A Novel Type of Cellular-Based Artificial Blood Substitutes. Biomacromolecules, (in press) (2005). 56. B. Discher, Y.-Y. Won, D. Ege, J. Lee, F. S. Bates, D. E. Discher and D. A. Hammer, Polymersomes: Tough Vesicles Made from Diblock Copolymers. Science, 284: 1143-1146 (1999). 57. D. E. Discher and A. Eisenberg, Polymer Vesicles. Science, 297: 967-973 (2002). 58. B. Discher, D. A. Hammer, F. S. Bates and D. E. Discher, Polymer Vesicles in Various Media. Curr. Opin. Coll. Inter. Sci., 5: 125-131 (2000). 59. H. Bermudez, A. K. Brannan, D. A. Hammer, F. S. Bates and D. E. Discher, Molecular Weight Dependence of Polymersome Membrane Structure, Elasticity and Stability. Macromolecules, 35: 8203-8208 (2002). 60. H. Sakai, Y. Masada, S. Takeoka and E. Tsuchida, Characteristics of Bovine Hemoglobin as a Potential Source of Hemoglobin-Vesicles for an Artificial Oxygen Carrier. J. Biochem., 131: 611-617 (2002). 61. R. L. Juliano and D. Stamp, The Effect of Particle Size and Charge on the Clearance Rates of Liposomes and Liposome Encapsulated Drugs. Biochem. Biophys. Res. Commun., 63: 651-658 (1975). 62. V. D. Awasthi, D. Garcia, B. A. Goins and W. T. Phillips, Circulation and Biodistribution Profiles of Long-Circulating PEG-Liposomes of Various Sizes in Rabbits. Int. J. Pharm., 253: 121-132 (2003). 63. H. Sakai, S. Takeoka, R. Wettstein, A. G. Tsai, M. Intaglietta and E. Tsuchida, Systemic and Microvascular Responses to Hemorrhagic Shock and Resuscitation with Hb Vesicles. Am. J. Physiol. Heart Circ. Physiol., 283: H1191-H1199 (2002).
150
64. R. Linberg, C. D. Conover, K. L. Shum and R. G. L. Shorr, Hemoglobin Based Oxygen Carriers: How Much Methemoglobin is too Much? Art. Cells, Blood Subs., and Biotech., 26: 133-148 (1998). 65. M. Kavdia, R. N. Pittman and A. S. Popel, Theoretical Analysis of Effects of Blood Substitute Affinity and Cooperativity on Organ Oxygen Transport. J. Appl. Physiol., 96: 2122-2128 (2002). 66. A. Krogh, The Number and Distribution of Capillaries in Muscles with Calculations of the Oxygen Pressure Head Necessary for Supplying the Tissue. J. Physiol., 52: 409-415 (1919). 67. H. Sakai, K. Hamada, S. Takeoka, H. Nishide and E. Tsuchida, Physical Propeties of Hemoglobin Vesicles as Red Cell Substitutes. Biotechnol. Prog., 12: 119-125 (1996). 68. S. Takeoka, T. Ohgushi, K. Terase, T. Ohmori and E. Tsuchida, Layer-Controlled Hemoglobin Vesicles by Interaction of Hemoglobin with a Phospholipid Assembly. Langmuir, 12: 1755-1759 (1996). 69. M. Brandl and G. Gregoriadis, Entrapment of Haemoglobin into Liposomes by the Dehydration-Rehydration Method: Vesicle Characterization and In Vivo Behavior. Biochim. Biohys. Acta, 1196: 65-75 (1994). 70. B. L. Oser, Hawk's Physiological Chemistry. McGraw Hill Book Company, New York (1965). 71. D. Papahadjopoulos, G. Poste, B. E. Schaeffer and W. J. Vail, Membrane Fusion and Molecular Segregation in Phospholipid Vesicles. Biochim. Biohys. Acta, 352: 1028 (1974). 72. D. Papahadjopoulos and H. K. Kimelberg, Phospholipid Vesicles (Liposomes) as Models for Biological Membranes: Their Properties and Interactions with Cholesterol and Proteins. Prog. Surf. Sci., 4: 141-144 (1973). 73. J. Lee, H. Bermudez, B. Discher, M. Sheehan, Y.-Y. Won, F. S. Bates and D. E. Discher, Preparation, Stability, and In Vitro Performance of Vesicles Made with Diblock Copolymers. Biotech. Bioeng., 73: 134-145 (2001). 74. A. Kidane, T. McPherson, H. S. Shim and K. Park, Surface Modification of Polyethylene Terephthalate Using PEG-Polybutadiene-PEG Triblock Copolymers. Colloids Surf., B: Biointerfaces, 3-4: 347-353 (2000). 75. F. Szoka and D. Papahadjopoulos, Procedure for Preparation of Liposomes with Large Internal Aqueous Space and High Capture by Reverse-Phase Evaporation. Proc. Natl. Acad. Sci. USA, 75: 4194-4198 (1978). 151
76. S. Zheng, Y. Zheng, R. L. Beissinger and R. Fresco, Microencapsulation of Hemoglobin in Liposomes Using a Double Emulsion, Film Dehydration/Rehydration Approach. Biochim. Biophys. Acta, 1196: 123-130 (1994). 77. C. G. Choquet, G. B. Patel, T. J. Beveridge and G. D. Sprott, Formation of Unilamellar Liposomes from Total Polar Lipid Extracts of Methanogens. Appl. Environ. Microbiol., 58: 2894-2900 (1992). 78. H. Sakai, K. Hamada, S. Takeoka, H. Nishide and E. Tsuchida, Physical Properties of Hemoglobin Vesicles as Red Cell Substitutes. Biotech. Prog., 12: 119-125 (1996). 79. A. Vivier, J. C. Vuillemard, H. W. Ackermann and D. Poncelet, Large-Scale Blood Substitute Production Using a Microfluidizer. Biomat., Art., Cells and Immob. Biotech., 20: 377-397 (1992). 80. S. U. Egelhaaf, E. Wehrli, M. Muller, M. Adrian and P. Schurtenberger, Determination of the Size Distribution of Lecithin Liposomes: A Comparative Study Using Freeze Fracture, Cryoelectron Microscopy and Dynamic Light Scattering. J. Microscopy, 184: 214-228 (1996). 81. R. Dreyer, E. Hawrot, A. C. Sartorelli and P. P. Constantinides, Sedimentation Field Flow Fractionation of Fused Unilamellar Vesicles: Comparison with Electron Microscopy and Gel Filtration. Anal. Biochem., 175: 433-441 (1988). 82. S. Lesieur, C. G. Madelmont, M. Paternostre and M. Ollivon, Study of Size Distribution and Stability of Liposomes by High Performance Gel Exclusion Chromatography. Chem. Phys. Lipids, 64: 57-82 (1993). 83. D. P. Hinton and C. S. Johnson, Diffusion Ordered 2D NMR Spectroscopy of Phospholipid Vesicles: Determination of Vesicle Size Distribution. J. Phys. Chem., 97: 9064-9072 (1993). 84. J. H. VanZanten, Characterization of Vesicles and Vesicular Dipersions via Scattering Techniques. Marcel Dekker, New York (1996). 85. P. J. Wyatt, Submicrometer Particle Sizing by Multiangle Light Scattering following Fractionation. J. Colloid Interface Sci., 197: 9-20 (1997). 86. J. H. VanZanten and H. G. Monbouquette, Characterization of Vesicles by Classical light Scattering. J. Colloid Interface Sci., 146: 330-336 (1991). 87. B. A. Korgel, J. H. VanZanten and H. G. Monbouquette, Vesicle Size Distributions Measured by Flow Field-Flow Fractionation Coupled with Multiangle Light Scattering. Biophys. J., 74: 3264-3272 (1998).
152
88. W. Wittgren and K. G. Wahlund, Fast Molecular Mass and Size Characterization of Polysaccharides Using Asymmetrical Flow Field-Flow Fractionation-Multiangle Light Scattering. J. Chromatogr. A., 760: 205-218 (1997). 89. D. Roessner and W.-M. Kulicke, On-Line Coupling of Flow Field-Flow Fractionation and Multi-Angle Laser Light Scattering. J. Chromatogr. A., 687: 249-258 (1994). 90. A. Litzen, Separation Speed, Retention, and Dispersion in Asymmetrical Flow FieldFlow Fractionation as Functions of Channel Dimensions and Flow Rates. Anal. Chem., 65: 461-470 (1993). 91. A. Litzen and K. G. Wahlund, Zone Broadening and Dilution in Rectangular and Trapezoidal Asymmetrical Flow Field-Flow Fractionation Channels. Anal. Chem., 63: 1001-1007 (1991). 92. A. Litzen, J. K. Walter, H. Krischollek and K. G. Wahlund, Separation and Quantification of Monoclonal Antibody Aggregates by Asymmetrical Flow FieldFlow Fractionation and Comparison to Gel Permeation Chromatography. Anal. Chem., 212: 469-480 (1993). 93. K. G. Wahlund and J. C. Giddings, Properties of an Asymmetrical Flow Field-Flow Fractionation Channel Having One Permeable Wall. Anal. Chem., 59: 1332-1339 (1987). 94. H. Brenner and D. A. Edwards, Macrotransport Processes. Butterworth-Heinemann, Stoneham (1993). 95. B. H. Zimm, The Scattering of Light and the Radial Distribution Function of High Polymer Solutions. J. Chem. Phys., 16: 1093-1099 (1948). 96. B. H. Zimm, Apparatus and Methods for Measurements and Interpretation of the Angular Variation of Light Scattering; Preliminary Results on Polystyrene Solutions. J. Chem. Phys., 16: 1099-1116 (1948). 97. H. Sakai, Y. Masada, H. Horinouchi, E. Ikeda, K. Sou, S. Takeoka, M. Suematsu, M. Takaori, K. Kobayashi and E. Tsuchida, Physiological Capacity of the Reticuloendothelial System for the Degradation of Hemoglobin Vesicles (Artificial Oxygen Carriers) after Massive Intravenous Doses by Daily Repeated Infusions for 14 Days. J. Pharmacol. Exp. Ther., 311: 874-884 (2004). 98. M. C. Farmer, A. S. Rudolph, K. D. Vandegriff, M. D. Hayre, S. A. Bayne and S. A. Johnson, Liposome-Encapsulated Hemoglobin: Oxygen Binding Properties and Respiratory Function. Biomat. Art. Cells, Art. Org., 16: 289-299 (1988).
153
99. S. V. Deshpande and R. L. Beissinger, Liposome-Encapsulated Hemoglobin Using Film Hydration Processing to Form Artificial Red Blood Cells. Biomat. Art. Cells and Immob. Biotech., 21: 135-151 (1993). 100. J. N. Patton and A. F. Palmer, Photopolymerization of Bovine Hemoglobin Entrapped Nanoscale Hydrogel Particles within Liposomal Reactors for Use as an Artificial Blood Substitute. Biomacromolecules, 6: 414-424 (2005). 101. R. R. C. New, Liposomes: A Practical Approach. The Practical Approach Series. Oxford University Press, New York (1997). 102. M. Antonietti and S. Forster, Vesicles and Liposomes: A Self-Assembly Principle Beyond Lipids. Adv. Mater., 15: 1323-1333 (2003). 103. J. Szebeni, H. Hauser, C. D. Eskelson, R. R. Watson and K. H. Winterhalter, Interaction of Hemoglobin Derivatives with Liposomes. Membrane Cholesterol Protects against the Changes of Hemoglobin. Biochemistry, 27: 6425-6434 (1988). 104. M. Mobed and T. M. S. Chang, Comparison of Polymerically Stabilized PEGgrafted Liposomes and Physically Adsorbed Carboxymethylchitin and Carboxymethyl/glycolchitin Liposomes for Biological Application. Biomaterials, 19: 1167-1177 (1998). 105. H. Hauser, Some Aspects of the Phase Behavior of Charged Lipids. Biochim. Biophys. Acta, 772: 37-50 (1984). 106. H. Sakai, A. G. Tsai, K. H., S. I. Park, S. Takeoka, H. Nishide, E. Tsuchida and M. Intaglietta, Subcutaneous Microvascular Responses to Hemodilution with a Red Cell Substitute Consisting of Polyethyleneglycol-modified Vesicles Encapsulating Hemoglobin. J. Biomed. Mater. Res., 40: 66-78 (1998). 107. C. Fronticelli, E. Bucci and C. Orth, Solvent Regulation of Oxygen Affinity in Hemoglobin. J. Biol. Chem., 259: 10841-10844 (1984). 108. C. C. LaBrake and L. W. M. Fung, Phospholipid Vesicles promote Human Hemoglobin Oxidation. J. Biol. Chem., 267: 16703-16711 (1992). 109. Y. Yoshida, K. Kashiba and E. Niki, Free Radical-Mediated Oxidation of Lipids Induced by Hemoglobin in Aqueous Dispersion. Biochim. Biophys. Acta, 1201: 165-172 (1994). 110. W. Martindale, Martindale Extra Pharmacopoeia. Royal Pharmaceutical Society, London (1996). 111. D. D. Lasic and F. Martin, Stealth Liposomes. CRC Press, Boca Raton (1995). 154
112. J. Szebeni and C. R. Alving, Complement-Mediated Acute Effects of LiposomeEncapsulated Hemoglobin. Art. Cells, Blood Subs., and Immob. Biotech., 27: 2341 (1999). 113. J. Szebeni, N. M. Wassef, H. Spielberg, A. S. Rudolph and C. R. Alving, Complement Activation in Rats by Liposomes and Liposome-Encapsulated Hemoglobin: Evidence for Anti-Lipid Antibodies and Alternative Pathway Activation. Biochem. Biophys. Res. Comm., 205: 255-263 (1994). 114. K. Nakai, A. Usuba, T. Ohta, M. Kuwabara, Y. Nakazato, R. Motoki and T. A. Takahashi, Coronary Vascular Bed Perfusion with a Polyethylene GlycolModified Hemoglobin-Encapsulated Liposome, Neo Red Cell, in Rats. Art. Organs, 22: 320-325 (1998). 115. H. Sakai, H. Hara, M. Yuasa, A. G. Tsai, S. Takeoka, E. Tsuchida and M. Intaglietta, Molecular dimensions of Hb-based O2 carriers determine constriction of resistance arteries and hypertension. Am. J. Physiol. Heart Circ. Physiol., 279: H908-915 (2000). 116. H. Sakai, H. Horinouchi, Y. Masada, S. Takeoka, E. Ikeda, M. Takaori, K. Kobayashi and E. Tsuchida, Metabolism of Hemoglobin-Vesicles (Artificial Oxygen Carriers) and their Influence on Organ Functions in a Rat Model. Biomaterials, 25: 4317-4325 (2004). 117. D. Marsh, Elastic Constants of Polymer-Grafted Lipid Membranes. Biophys. J., 81: 2154-2162 (2001). 118. S. Wakamoto, M. Fujihara, H. Abe, H. Sakai, S. Takeoka, E. Tsuchida, H. Ikeda and K. Ikebuchi, Effects of Poly(ethylenelglycol)-Modified Hemoglobin Vesicles on Agonist-Induced Platelet Aggregation and RANTES Release In Vitro. Art. Cells, Blood Subs., and Immob. Biotech., 29: 191-201 (2001). 119. H. Sakai, A. G. Tsai, K. H., S. I. Park, S. Takeoka, H. Nishide, E. Tsuchida and M. Intaglietta, Subcutaneous Microvascular Responses to Hemodilution with a Red Cell Substitute Consisting of Polyethyleneglycol-modified Vesicles Encapsulating Hemoglobin. J. Biomed. Mater. Res., 40: 66-78 (1997). 120. D. D. Lasic, Novel Applications of Liposomes. Reviews, 16: 307-321 (1998). 121. K. Sou, Y. Naito, T. Endo, S. Takeoka and E. Tsuchida, Effective Encapsulation of Proteins into Size-Controlled Phospholipid Vesicles using Freeze-Thawing and Extrusion. Biotech. Prog., 19: 1547-1552 (2003). 122. V. C. F. Mosqueira, P. Legrand, A. Gulik, O. Bourdon, R. Gref, D. Labarre and G. Barrat, Relationship between Complement Activation, Cellular Uptake and
155
Surface Physicochemical Aspects of Novel PEG-Modified Nanocapsules. Biomaterials, 22: 2967-2979 (2001). 123. V. Pata, F. Ahmed, D. E. Discher and N. Dan, Membrane Solubilization by Detergent: Resistance Conferred by Thickness. Langmuir, 20: 3888-3893 (2004). 124. B. M. Discher, H. Bermudez, D. A. Hammer, D. E. Discher, Y.-Y. Won and F. S. Bates, Cross-linked Polymersome Membranes: Vesicles with Broadly Adjustable Properties. J. Phys. Chem. B, 106: 2848-2854 (2002). 125. H. Bermudez, D. A. Hammer and D. E. Discher, Effect of Bilayer Thickness on Membrane Bending Rigidity. Langmuir, 20: 540-543 (2004). 126. H. Aranda-Espinoza, H. Bermudez, F. S. Bates and D. E. Discher, Electromechanical Limits of Polymersomes. Physical Rev. Lett., 87: 208301-1 208301-4 (2001). 127. J. A. Opsteen, J. J. L. M. Cornelissen and J. C. M. van Hest, Block Copolymer Vesicles. Pure Appl. Chem., 76: 1309-1319 (2004). 128. H. Shen and A. Eisenberg, Block Length Dependence of Morphological Phase Diagrams of the Ternary System of PS-b-PAA/dioxane/H2O. Macromolecules, 33: 2561-2572 (2000). 129. S. Forster and M. Konrad, From Self-Organizing Polymers to Nano- and Biomaterials. J. Mater. Chem., 13: 2671-2688 (2003). 130. V. Pata and N. Dan, The Effect of Chain Length on Protein Solubilization in Polymer-Based Vesicles (Polymersomes). Langmuir, 20: 5493-5500 (2004). 131. H. T. Jung, B. Coldren, J. A. Zasadzinski, D. J. Iampietro and E. W. Kaler, The Origins of Stability of Spontaneous Vesicles. Proc. Natl. Acad. Sci. USA, 98: 1353-1357 (2001). 132. S.-Y. Kang, B.-S. Seong, Y. S. Han and H.-T. Jung, Self-Organization of Amphiphilic Polymer in Vesicle Bilayers Composed of Surfactant Mixtures. Biomacromolecules, 4: 360-365 (2003). 133. S. A. Safran, P. Pincus and A. Andelman, Theory of Spontaneous Vesicle Formation in Surfactant Mixtures. Science, 248: 354-356 (1990). 134. S. L. Li, J. Nickels and A. F. Palmer, Liposome-Encapsulated Actin-Hemoglobin (LEAcHb) Artificial Blood Substitutes. Biomaterials, 26: 3759-3769 (2005).
156
135. J. H. Eike and A. F. Palmer, Oxidized Mono-, Di-, Tri-, and Polysaccharides as Potential Hemoglobin Cross-Linking Reagents for the Synthesis of High Oxygen Affinity Artificial Blood Substitutes. Biotech. Prog., 20: 953-962 (2004). 136. S. M. Hameed, W. C. Aird and S. M. Cohn, Oxygen Delivery. Crit. Care Med., 31: S658-S667 (2003). 137. S. Shafi and D. R. Kauder, Fluid Resuscitation and Blood Replacement in Patients with Polytrauma. Clin. Orthop., 422: 37-42 (2004). 138. G. A. Truskey, F. Yuan and D. F. Katz, Transport Phenomena in Biological System. Pearson Prentice Hall, Upper Saddle River (2004). 139. W. M. Deen, Analysis of Transport Phenomena. Oxford University Press, New York (1998). 140. H. Sakai, A. G. Tsai, R. J. Rohlfs, H. Hara, S. Takeoka, E. Tsuchida and M. Intaglietta, Microvascular Responses to Hemodilution with Hb Vesicles as Red Blood Cell Substitutes: Influence of O2 Affinity. Am. J. Physiol., 276: H553-562 (1999). 141. A. G. Tsai, K. D. Vandegriff, M. Intaglietta and R. M. Winslow, Targeted O2 Delivery by Low-P50 Hemoglobin: A New Basis for O2 Therapeutics. Am. J. Physiol. Heart Circ. Physiol., 285: H1411-H1419 (2003). 142. A. G. Tsai and M. Intaglietta, The Unusual Properties of Effective Blood Substitutes. Keio. J. Med., 51: 17-20 (2002). 143. J. Lee, M. Santore, F. S. Bates and D. E. Discher, From Membranes to Melts, Rouse to Reptation: Diffusion in Polyermersome Versus Lipid Bilayers. Macromolecules, 35: 323-326 (2002). 144. A. Walter, P. K. Vinson, A. Kaplun and Y. Talmon, Intermediate Structures in the Cholate-Phosphatidylcholine Vesicle Micelle Transition. Biophys. J., 60: 13151325 (1991). 145. J. C. Wejman, D. Hovsepian, J. S. Wall, J. F. Hainfeld and J. Greer, Structure of Haptoglobin and the Haptoglobin-Hemoglobin Complex by Electron Microscopy. J. Mol. Biol., 174: 319-314 (1984). 146. M. B. Gorbet and M. V. Sefton, Biomaterial-Associated Thrombosis: Roles of Coagulation Factors, Complement, Platelets and Leukocytes. Biomaterials, 25: 5681-5703 (2004). 147. M. M. Mayer, Kabat and Mayer's Experimental Immunochemistry. Charles C. Thomas, Springfield (1961). 157
148. D. Labarre, B. Montdargent, M.-P. Carreno and F. Maillet, Strategy for In Vitro Evaluation of the Interactions Between Biomaterials and Complement System. J. Applied Biomat., 4: 231-240 (1993). 149. S. I. Ertel, B. D. Ratner, A. Kaul, M. B. Schway and T. A. Horbett, In Vitro Study of the Intrinsic Toxicity of Synthetic Surfaces to Cells. J. Biomed. Mater. Res., 28: 667-675 (1994). 150. A. G. Tsai and M. Intaglietta, High Viscosity Plasma Expanders: Volume Restitution Fluids for Lowering the Transfusion Trigger. Biorheology, 38: 229237 (2001). 151. Y. Ogata, H. Goto, T. Kimura and H. Fukui, Development of Neo Red Cells (NRC) with the Enzymatic Reduction System of Methemoglobin. Art. Cells, Blood Subs., and Immob. Biotech., 25: 417-427 (1997). 152. H.-P. Hentze and E. W. Kaler, Polymerization of and within Self-Organized Media. Curr. Opin. Coll. Inter. Sci., 8: 164-178 (2003). 153. Y. Y. Won, H. T. Davis and F. S. Bates, Giant Wormlike Rubber Micelles. Science, 283: 960-963 (1999). 154. R. Stoenescu and W. Meier, Vesicles with Asymmetric Membranes from Amphiphilic ABC Triblock Copolymers. Chem. Commun., 3016-3017 (2002). 155. S. Schrage, R. Sigel and H. Schlaad, Formation of Amphiphilic Polyion Complex Vesicles from Mixtures of Oppositely Charged Block Ionomers. Macromolecules, 36: 1417-1420 (2003). 156. F. Meng, C. Hiemstra, G. H. M. Engbers and J. Feijen, Biodegradable Polymersomes. Macromolecules, 36: 3004-3006 (2003).
158
APPENDIX A
MATLAB CODE FOR SIMULATION OF OXYGEN TRANSPORT
clear clc %Enter the parameters Pin=100; %Inlet oxygen tension, mmHg H=0.74; %Henry's constant, mmHg/microM Csat=8800; %Concentration of saturated oxyHb, microM P50=7; %mmHg n=2; %Cooperativity coefficient V=0.05; %Average blood velocity in a capillary, cm/s L=0.1; %Length of a capillary, cm rc=0.0005; % Capillary inner radius, cm DT=6.3*(10^-6); %Oxygen diffusivity in tissues, cm2/s Rm=100; %Tissue metabolic oxygen consumption rate, 25.9 microM/s in pancreas and brain, and 100 microM/s in muscle rT=0.0020; %Radius of Krogh tissue cylinder, cm count=0; P(1,1)=Pin; m(1,1)=0; zcount=0; iteration=0.005; for z=0:iteration:L zcount=zcount+1; if zcount>1 if zcount==2 Pave=Pin; elseif zcount==3 Pave=(Pin+P(2,1))/2; elseif zcount>3 Pave=(P(zcount-1,1)+P(zcount-2,1)+P(zcount-3,1))/3 159
end %Calculate the value of m using the average of the last three oxygen tensions m(zcount,1)=n*(P50^n)*H*Csat*((Pave^(n-1))/(((P50^n)+(Pave^n))^2)); %Calculate radially-averaged oxygen tension in the blood (P) with respect to the axial position in capillary P(zcount,1)=Pin-(((Rm*H)/((1+m(zcount,1))*V))*(((rT/rc)^2)-1)*z); %In anoxic regions, the oxygen tension is 0 mmHg if P(zcount,1)<=0 P(zcount,1)=0; elseif P(zcount-1,1)<=0 P(zcount-1,1)=0; end end dummy=P(zcount,1); rvar=0.5; rcount=0; step=0.00001; for r=rc+step:step:rT rcount=rcount+1; rvar=rvar+(step*1000); count=count+1; zaxis(count,1)=dummy; %Calculate the oxygen tension in tissue (PT) PT(zcount,rcount)=P(zcount,1)-((rc^2)*Rm*H*(1-((r/rc)^2))/(4*DT))((rT^2)*Rm*H*(log(r/rc))/(2*DT)); %In anoxic regions, the oxygen tension is 0 mmHg if PT(zcount,rcount)<=0 PT(zcount,rcount)=0; end dummy=PT(zcount,rcount); yaxis(count,1)=z; 160
xaxis(count,1)=rvar; end end %Plot the oxygen tension in a capillary and tissue with respect to the radial and axial positions plot3(xaxis,yaxis,zaxis,'.') grid on set(gca,'YDir','reverse') zlabel('pO_2 (mmHg)','FontSize',20) ylabel('z (cm)','FontSize',20) xlabel('r X 10^-^3 (cm)','FontSize',20) xlim ([0.0005 0.003]) axis tight %Calculate the total concentration of oxygen consumed in the tissue region (OC) OC=Csat*(((Pin^n)/((P50^n)+(Pin^n)))-((P(zcount,1)^n)/((P50^n)+(P(zcount,1)^n))));
161