Document Sample

       The primary motivation behind this research is improving the success of the

treatment of physical and chemical injuries to the cornea. The limbus is the barrier

that divides the conjunctival epithelium from the corneal epithelium and contains the

progenitor population for the corneal epithelium. In severe injuries that destroy

significant amounts of the corneal epithelium, the progenitor population can also be

vastly reduced leading to limbal stem cell deficiency and causing the surface of the

cornea to be invaded by conjunctival epithelium. Limbal stem cell deficiency can

arise due to both acquired and inherited disorders (Thoft et al., 1989; Tseng, 1989;

Tseng, 1996; Dua and Azuara-Blanco, 2000) although acquired disorders make up the

majority of clinical cases (Ang and Tan, 2004). Such acquired conditions include

chemical injuries, Stevens-Johnson syndrome, multiple surgeries to the eye, ocular

cicatricial phemphigoid, neurotrophic keratopathy, peripheral ulcerative keratitis and

contact lens-induced keratopathy. There are several approaches to the treatment of

limbal stem cell deficiency but they are separated by distinct differences.

       The most commonly used treatment is the replacement of the damaged limbal

tissue with a donor limbal graft. Unfortunately this approach results in almost half of

the grafts performed failing within a year of the operation. While the central cornea is

essentially non-vascularised, the limbus is highly vascularised and may allow

interaction with the patient’s immune system, thereby leading to the high incidence of

graft failure. In order to circumvent this problem, limbal autografts have been used in

cases where only one eye has been injured and the other has been left untouched

(Kenyon and Tseng, 1989; Tsai and Tseng, 1994; Dua and Azuara-Blanco, 2000). The

drawback of this approach is that it requires healthy tissue to be removed from the

unaffected eye, thereby risking the long-term maintenance of the healthy eye in the


         Another approach, developed to reduce this risk, involves the biopsy of a

smaller sample of the healthy limbus followed by the cultivation of the epithelial cells

that it contains. This can be achieved either through explant culture of the biopsy

directly (Tseng, 1994; Akpek and Foster, 1999) or through enzymatic digestion of the

biopsy and the cultivation of the subsequent uniform cell suspension (Kenyon and

Tseng, 1989; Tsai and Tseng, 1994; Dua and Azuara-Blanco, 2000). The latter

method employs a similar technique to that which is used to produce cultured skin

grafts whereby keratinocytes are cultured in the presence of mitotically inhibited

mouse 3T3 fibroblasts in a specialised media (Rheinwald and Green, 1975). As the

keratinocytes proliferate and spread, the mouse cells are dislodged into the culture


         Thus, while the cultivation of the autologous keratinocytes has been well

established, there is still a mechanical challenge in applying the cultivated

keratinocytes to the injured eye. Human amnionic membrane has been used routinely

in ophthalmic surgical procedures for a significant period of time and shares

properties of the basement membrane of the limbus (Kim and Tseng, 1995). In fact,

human amnionic membrane has been used to varying degrees of success to treat

ocular surface injuries as a sole treatment (Lee and Tseng, 1997; Tseng et al., 1997;

Azuara-Blanco et al., 1999). These characteristics make it a promising choice for use

as a grafting substrate for the autologous keratinocyte culture. In fact, similar

approaches employing cultivated autologous limbal epithelial cultures grown on

amnionic membrane have been used to treat limbal stem cell deficiency to varying

degrees of success (Schwab, 1999; Schwab et al., 2000; Tsai et al., 2000; Cohen,

2001; Grueterich et al., 2002; Grueterich and Tseng, 2002; Xi et al., 2003; Nakamura

et al., 2004). The major differences between these previous studies and this work lie

in the assessment of phenotype of the cultivated autografts and the criteria of

assessing successful treatment.

       This chapter describes the treatment and clinical outcomes of five patients

treated from late 2001 until mid 2003 using a hybrid of various approaches across

keratinocyte culture and ophthalmic surgery. The patients were treated for limbal

epithelium deficiency resulting from a variety of clinical causes ranging from thermal

and chemical injury to suspected viral infection and artefact of previous surgeries. The

level of success of these treatments was noticeably variable. This variation in itself

may be due to differences in the cause and severity of each injury. In any case, this

study is the first to assess the phenotype of autologous grafts used in clinical practice

through western blotting, RT-PCR and immunohistochemistry performed upon the

same samples. In particular, prior to this work a detailed analysis of p63 isoforms in

patient cultures had not been performed.


       The procedure used to produce the autologous grafts was mostly consistent

over the course of the study, with only minor modifications made as the study

progressed. Any modifications made are noted where applicable.

       The technique commenced with the removal of two 2mm2 biopsies of limbal

epithelium from the superior and inferior limbus of the patient’s healthier eye. These

biopsies were removed by an ophthalmic surgeon in an operating theatre and were

immediately placed into sterile PBS before being transported back to the laboratory at

room temperature. Upon arrival in the laboratory the biopsies were rinsed in an

antibiotic antimycotic solution for 30 minutes before being digested for 10 minutes at

37°C in PBS containing 0.25% trypsin and 1 mM EDTA to produce a uniform cell

suspension. This cell suspension was centrifuged at 300g for 5 minutes and

subsequently resuspended in culture media to inactivate the trypsin. The primary

expansion of the keratinocytes was achieved using a slightly modified version of the

Rheinwald and Green method (Rheinwald and Green, 1975). The culture media

consisted of DMEM/F12 supplemented with 1 µg/ml insulin, 0.4 µg/ml

hydrocortisone, 10 ng/ml epidermal growth factor, 180 µM adenine, 10 ng/ml cholera

toxin with 10% gamma irradiated fetal bovine serum (FBS) also added.

       The patient cells were cultured in the presence of gamma irradiated mouse

3T3 fibroblasts at a cell density of 5x104/cm2 for the first two passages. This typically

took 2 weeks and produced approximately 107 autologous cells. Initially, the cultures

were grown in 35mm-diameter culture dishes (P0) and, after being passaged the first

time, were reseeded into 75cm2 flasks to give passage 1 (P1). When passaged,

aliquots of 106 cells in 90% FBS/10% dimethylsulfoxide (DMSO) were transferred to

liquid nitrogen storage for later analysis and further propagation. Once the P1 cultures

reached confluence they were passaged for the last time and the autologous grafts

were constructed.

       Pieces of human amniotic membrane (AM) cut with a diameter of 35mm were

used as a transport substrate to seed the cultured patient cells. Before seeding the AM,

the amniotic epithelium was removed by incubating the AM in 0.25% trypsin and 1

mM EDTA for 10 minutes at 37°C and then gently scraping the cells away with a

plastic cell scraper. Once the epithelium was removed, the AM was washed in PBS

once again. Two pieces of stripped AM were placed into the centre wells of a 6 well

plate and secured with silicon rings before the cultured patient cells were added at a

cell density of 2x105/cm2. No 3T3 fibroblasts were added to these cultures. The two

grafts were cultured for another 2 weeks before one was fixed in 4%

paraformaldehyde and examined immunohistochemically, while the other was grafted

onto the patient’s injured eye. The grafting procedure is shown in Figure 5.1.

       Before attaching the graft to the injured eye any invading conjunctiva was

removed by scraping the surface back to 3 or 4 mm past the limbal interface (see

Figure 5.1A and B). During the first treatment of the first patient, the graft was

applied face down with the cultivated epithelium in contact with the surface of the

eye. After observing an underwhelming response and consultation with Ivan Schwab,

all subsequent grafts were applied in a flipped orientation (i.e. the cultivated

epithelium facing away from the surface of the eye). Figure 5.1 C, D and E,

demonstrate the improved approach, with the graft placed face up in position covering

the affected area and sutured into place through the underlying stroma and adjacent

conjunctiva. A therapeutic contact lens was worn for 2 months following the

operation to aid in protecting the graft (Figure 5.1F). Further details of each individual

grafting operation are described in the case studies that follow.

       This chapter also employed the techniques of immunocytochemistry (2.1.3),

RNA analysis (2.1.4), western blot analysis (2.1.5) and immunohistochemistry (2.1.6)

described in the General Methods chapter.

       Dr Andrew Apel provided access to clinical case notes which allowed for the

detail of each patient injury and treatment history. In addition, Dr Apel provided both

clinical and surgical photos included in the results section (5.3) of this chapter.

       Ethics approval for this research was obtained through the Royal Brisbane and

Women’s Hospital Human Research Ethics Committee (Research Protocol: 2001/040)

and the Queensland University of Technology (Ethics #: 2412H).

Figure 5.1: Application of the Autologous Graft. (A) The invading conjunctival

epithelium was scraped away; (B) Excess fluid and loose material was cleaned; (C)

The graft was peeled from its backing paper support; (D) The graft was positioned to

cover the injured surface with the cultivated epithelium facing away from the surface

of the eye; (E) The graft was sutured to the surface of the eye; (F) A therapeutic

contact lens was placed over the graft to protect from dislodgment.

5.3.   RESULTS

       5.3.1. Patient 1: Mr AA

Date of Birth: 25/01/52

First Appointment: 24/08/99

       A thermal injury to the patient’s left eye, which occurred in August of 1999,

had induced limbal deficiency. A pannus of vascular tissue had also formed over the

surface of the cornea causing some visual opacity. In October of 2001 the visual

acuity of the patient’s left eye was restricted to hand movements.

       On the 23rd of October 2001 a peritomy was performed to remove scar tissue.

At the same time a biopsy of healthy limbus was taken from the right eye. The

autologous graft was applied on the 19th of November 2001 with the cultivated patient

cells being placed facedown on the corneal surface. Figure 5.2A shows the condition

of the surface of the injured eye before application of the graft. The graft split and

broke down over the next few months as seen in Figure 5.2B, where a split is apparent

in the 1-2 clock hour position and Figure 5.2C, where 11 months later conjunctival

tissue has regrown. A second graft was constructed early in 2003 using patient cells

that had been stored in liquid nitrogen since being cultivated from the original biopsy

taken in 2001. This graft was applied, this time with the patient cells facing outwards,

on the 2nd of April 2003. In addition to the change in the orientation of the graft, a

therapeutic contact lens was applied to protect the surface from the mechanical

disruption of the eyelids. The appearance of the eye surface immediately prior to the

second grafting is shown in Figure 5.2D. One month after the grafting, blood vessels

were visibly spreading inwards from the limbus as seen in Figure 5.2E. This pattern

continued over the next several months. On the 20th of November 2003 the surface

was still relatively stable, however the invasive blood vessels were continuing to

spread. In July of 2004 the visual acuity of the patient’s left eye was restricted to hand


             Immunological Analysis

       During the first grafting operation the excess material of the graft was trimmed

away as it was sutured to the surface of the eye. As previously noted, the graft had

been applied facedown so that the cultivated cells were sandwiched between the

surface of the eye and the supporting AM basement membrane. This excess material

was collected and processed, to be later examined for K3, K14 and K19. As shown in

Figure 5.3, the graft expressed significant amounts of all three cytokeratins. The graft

itself had an epithelium that was only 3 or 4 cells thick and the staining for each

antibody was spread evenly throughout, at similar intensity. This graft subsequently

degraded and another attempt was made.

       In the case of Mr AA it was possible to remove the remnants of the first

constructed graft for immunological analysis while the second graft was being

applied. This allowed for further assessment of cytokeratin expression in the graft

after 17 months in vivo. The retrieved remains of the first graft were assessed for K3

and p63 using the respective antibodies against these markers. Figure 5.4D

demonstrates strong K3 expression in the remnants while A and B identify restricted

expression of p63 in a limited number of cells. Due to the nature of the tissue samples

and the way in which the first graft was applied, it is difficult to orientate the remnants

of the earlier graft in respect to the suprabasal epithelium and the supporting basement


       At each subsequent treatment, two identical grafts were produced to allow for

one to be applied to the patient’s injury while the second, if not required due to

problems with the first, could be analysed for molecular markers. The second attempt

to treat Mr AA produced such a supplementary graft and it was employed for

immunological analysis. Figure 5.5 illustrates the variety in the thickness of

epithelium produced in the graft and also the presence of K3. Once more the

suprabasal cells express K3 much more strongly than those cells basally located.

Figure 5.2: Condition of AA Injured Eye during Treatment. (A) Surface of the

injured eye immediately prior to the first application of the autologous graft; (B) The

condition of the eye 2.5 months following the first grafting; (C) Appearance of

surface 14 months after first graft; (D) Immediately prior to application of second

graft, 16.5 months after first; (E) 5 weeks after second grafting.

Figure 5.3: Immunohistochemistry of First AA Graft. (A, C, E) Sections of the

constructed autologous graft applied 19/11/01. Scale bar included in A represents

100µM and is valid for all images; (B) The same section as A, stained for keratin 3;

(D) The same section as C, stained for keratin 14; (F) The same section as E, stained

for keratin 19.

Figure 5.4: Immunohistochemistry of First AA Graft Remnants Removed before

Second Grafting. (A) Hoechst staining of sectioned remnants. Scale bar included in

A represents 100µM and is valid for all images; (B) The same section as A, stained

for p63; (C) Section of first graft remnant; (D) The same section as C, Hoechst-

stained nuclei (blue) and keratin 3 staining (red).

Figure 5.5: Immunohistochemistry of Second AA Graft. (A, C) Sections of the

constructed autologous graft applied 02/04/03. Scale bar included in A represents

100µM and is valid for all images; (B) The same section as A, stained for Keratin 3;

(D) The same section as C, stained for keratin 3.

       5.3.2. Patient 2:Mr MP

Date of Birth: 23/07/26

First Appointment: 09/11/98

       The patient originally injured the surface of his right eye in November 1992

when some caustic soda exploded and burnt one side of his face. This accident also

damaged the function of his lower eyelid, causing some laxity of movement and

splitting of the tissue. The burn also caused some trichiasis, where the eyelashes grow

inward and scratch the eye’s surface. The left eye suffered preretinal fibrosis and

some cataract, which was removed in 1995, though this was not caused by the same

injury. The patient had borderline hystolic hypertension and hypercholesterolaemia.

The visual acuity of the patient’s injured eye prior to the autologous graft treatment

was reduced to counting fingers.

       A biopsy of limbal epithelium was taken from the left eye on the 19th of March

2002 but was lost due to infection during culture. A second biopsy was taken on the

18th of April 2002 and was successfully cultured. The transplant of the autologous

graft took place on the 4th of May after a peritomy of the conjunctiva was performed

to remove invading cells on the surface of the cornea. Figure 5.6A shows the state of

the corneal surface before the transplant. The graft was attached to about half the

surface of the cornea of the right eye, covering the nasal region from 12 to 6:30 in

clock hour segments. One month after the operation the graft was intact but some

peripheral blood vessels were noted in the superior and inferior limbal position. The

patient commented that the eye ached at times and occasionally “gummed up”.

        At the end of August of 2002 (see Figure 5.6B) the graft remained smooth but

there still seemed to be invading blood vessels. In effort to halt this, a second graft

was applied on the 7th of October 2002. The surface of the cornea at this point is

shown in Figure 5.6C. One fortnight after the second graft the surface had settled

noticeably (Figure 5.6D) and maintained a stable surface 4 months later (Figure 5.6E).

On the 26th of May 2003, seven and a half months after the second graft, a

conventional corneal graft was performed in an attempt to repair stromal damage

related to the injury. Figure 5.6F shows the patient’s cornea one week following this

procedure. At the same time cataract extraction was performed and an intraocular lens

was implanted. One month after the graft the eye was comfortable and vision had

improved. A subsequent appointment in October confirmed these observations

(Figure 5.6G). The visual acuity of the patient’s right eye was 6/18 in December of

2003. In April 2004 the patient’s right eye was still cloudy in the area peripheral to the

donor graft but was much more comfortable. The surface of the cornea was also much

more smooth and stable. The visual acuity of the right eye was 6/9 in November of


             Immunological Analysis

        Two grafts were applied to Mr MP’s eye within 6 months using cultivated

cells from the same biopsy. The duplicate graft produced during the second treatment

was sectioned and stained for p63 as well as keratin 14 and keratin 19. The results are

shown in Figure 5.7. The graft had a continuous epithelium of between 1 and 2 layers

and, as displayed in Figure 5.7B, many of these cells expressed p63. Keratin 14

(Figure 5.7D) was present in greater concentration than keratin 19 (Figure 5.7F) but

both were seen in the majority of the epithelial cells.

       This case also gave the unique opportunity to assess the phenotype of the

grafted cells almost seven months after they had been applied to the cornea. Figure

5.8 demonstrates the keratin 3 expression of the central corneal epithelium recovered

after a conventional corneal graft was performed and a button of the patient’s own

cornea was removed. The expression is identical to that seen in donor tissue (see

Chapter 3).

Figure 5.6: Condition of MP Injured Eye during Treatment. (A) Appearance of

surface 10 years after injury but before autologous graft treatment; (B) Corneal

surface almost 4 months after first graft; (C) Immediately prior to second graft, 5

months after first; (D) 2 week following second grafting; (E) 4.5 months after the

second graft was applied; (F) 1 week following conventional corneal graft, 8 months

after the second autologous graft; (G) 4 months after conventional graft, 1 year after

autologous graft.

Figure 5.7: Immunohistochemistry of MP Graft. (A, C, E) Sections of unused

duplicate graft prepared concurrently with the graft applied 07/10/02, E is still

attached to the backing paper support. Scale bar included in A represents 100µM and

is valid for all images; (B) The same section as A, stained for p63; (D) The same

section as C, stained for keratin 14; (F) The same section as E, stained for keratin 19.

Figure 5.8: Immunohistochemistry of MP Corneal Button. (A) Section of corneal

button removed 26/05/03 after 2 autologous grafting operations. Scale bar included in

A represents 100µM and is valid for all images. (B) The same section as A, stained

for keratin 3.

       5.3.3. Patient 3: Mr JK

Date of Birth: 27/09/29

First Appointment: 11/06/02

       This patient had developed limbal deficiency in his left eye as a consequence

of repetitive surgery conducted on his eye. In February 1996 an intraocular lens

exchange and conventional corneal transplant were performed on the left eye. The

patient also suffered from glaucoma and the lens of his right eye had been removed,

making him surgically aphakic. His right eye also had minor scarring from cataract

removal. Prior to beginning the autologous graft treatment the visual acuity of his left

eye was 6/36+1.

       On the 12th of June 2002 biopsies of limbal epithelium were taken from both

eyes to assess which surface produced the healthiest culture when grown in vitro. The

biopsy from the right eye produced cells with morphology more similar to normal

keratinocyte phenotype than the culture derived from the left eye biopsy. An

autologous graft was therefore constructed using cells from the right eye biopsy

culture and was subsequently transplanted onto the surface of the left eye’s cornea

(seen in Figure 5.9A before the operation) on the 21st of October 2002. Figure 5.9B

shows the condition of the eye one month after the graft while Figure 5.9C

demonstrates the reduced surface inflammation four months following the operation.

A follow up assessment was done seven months following the transplant to determine

its success (Figure 5.9D). The patient commented that he felt greater comfort and it

was also observed that no neovascularisation had appeared in the grafted region.

Visual acuity had decreased possibly due to a haze caused by the amniotic membrane

on which the graft was grown. The surface of the cornea was smooth and there was

punctate staining seen in the peripheral host corneal rim. The next appointment was

six months later in November of 2003. The surface was still smooth and healthy and

the patient’s comfort was improving. The vague haze of the amniotic membrane was

still present but had faded since the previous appointment. Upon examination in May

of 2004 and the patient was very pleased with the comfort of the graft. The surface

was quiet but still distorted with the amniotic membrane causing some haze. The

visual acuity of his left eye at this appointment was 6/24. In December of 2004 the

visual acuity was 6/36.

            Immunological Analysis

       The duplicate graft prepared was fixed and stained for p63 and keratin 3 as

shown in Figure 5.10. The epithelium varied between 1 or 2 cells at the thinnest areas

up to 4 or 5 cells in the thickest. As shown in Figure 5.10B and C, there was positive

staining for p63 in the graft but only in a limited number of cells. The presence of

keratin 3 was abundant and spread throughout the layers of the epithelium to differing

intensities (Figure 5.10E).

Figure 5.9: Condition of JK Injured Eye during Treatment. (A) Appearance of the

injured eye immediately prior grafting; (B) Surface of eye 1 month after graft; (C) 4

months after grafting; (D) 7 months following grafting.

Figure 5.10: Immunohistochemistry of JK Graft. (A, D) Sections of unused

duplicate graft prepared concurrently with graft applied 21/10/02. Scale bar included

in A represents 100µM and is valid for all images; (B) The same section as A with

Hoechst-stained nuclei; (C) The same section as A and B, stained for p63; (E) The

same section as D, Hoechst-stained nuclei (blue) and keratin 3 staining (red).

       5.3.4. Patient 4: Mr WG

Date of Birth: 05/11/38

First Appointment: 10/09/02

       This patient had limbal deficiency in his right eye with a relatively normal

epithelium on the upper part of his left eye. A firecracker exploding in his face when

he was 9 years old caused the injury. His right eye also had developed some cataract.

Prior to beginning the autologous grafting treatment the visual acuity of his right eye

was 6/60.

       A biopsy of healthy epithelium was taken from the second hour position of his

left eye for cultivation on the 23rd of September 2002. The completed autologous graft

was applied to the full surface of his right eye on the 21st of October 2002. The graft

was protected by a therapeutic contact lens that prevented its dislodgment. Due to

technical difficulty no image of the injured eye was taken prior to treatment. In the

month following the surgery (Figure 5.11A) the patient experienced ache in his eye as

well as some photophobia. In January of 2003 (Figure 5.11B), three months following

the surgery, the patient still reported occasional ache and some blood vessels were

observed at the edge of the graft. At this time the therapeutic lens was removed. In

April of 2003 (Figure 5.11C), six months after the operation, superficial vessels were

observed in both the superior and inferior limbal regions of the right eye. The patient

continued to suffer some photophobia but felt that it had decreased. The visual acuity

of the patient’s right eye was 6/24 at this point. One year following the application of

the autologous graft the surface of the cornea was much more stable. Visual acuity

had not improved greatly, however the patient noted a greater level of comfort than

prior to the operation. When the patient was examined at the end of March 2004 there

had been no changes in the condition of the cornea. On the 12th of October 2004 the

visual acuity of the treated eye was 6/60.

            Immunological Analysis

       The sections produced from the supplementary WG graft were particularly

fragile. Figure 5.12 presents the immunohistochemistry performed to examine the

presence of p63 and keratin 3 in the graft. Figures 6.12A, B and D demonstrate the

thin epithelium of the graft, rarely thicker than a monolayer. A small number of cells

did express p63 but the intensity of this expression was quite low. Keratin 3 was

expressed throughout the epithelium (Figure 5.12E) but only in low amount, possibly

due to the basal position of the cells present. A stronger signal was seen in a number

of cells no longer in contact with the basement membrane.

Figure 5.11: Condition of WG Injured Eye during Treatment. (A) Appearance of

the surface of the injured eye 1 month after the application of the graft; (B) 3 months

following grafting; (C) 6 months after grafting.

Figure 5.12: Immunohistochemistry of WG Graft. (A, D) Sections of unused

duplicate graft prepared concurrently with graft applied 21/10/02. Scale bar included

in A represents 100µM and is valid for all images; (B) The same section as A with

Hoechst-stained nuclei; (C) The same section as A and B, stained for p63; (E) The

same section as D, Hoechst-stained nuclei (blue) and keratin 3 staining (red).

       5.3.5. Patient 5: Mrs BB

Date of Birth: 05/06/21

First Appointment: 18/02/03

       This patient had developed total limbal deficiency in her right eye due to

infection by herpes zoster ophthalmicus. No dysplasia was apparent in the eye and her

left eye maintained a healthy limbus. She was not suffering any discomfort but the

vision in her right eye was blurred while the vision in her left was normal. Before

starting the autologous graft treatment the visual acuity of her right eye was 6/36.

       The condition of the injured eye prior to this treatment is shown in Figure

5.13A. The biopsy was taken from her left eye on the 2nd of April 2003 and grown

until the 30th of April. At this time the complete autologous graft was applied to the

full surface of her right eye after a peritomy was performed to clear the surface of

conjunctival tissue. Figure 5.13B shows the surface immediately before the procedure

and 6.13C shows the cornea after the graft was applied. One month following the

operation she was not feeling any significant irritation and her visual acuity had

improved slightly. Two months post-graft (Figure 5.13D) there was still no obvious

change or irritation although there was evidence of a small epithelial defect in the

centre of the graft. At the end of August 2003, four months after the graft was applied,

the surface had developed some rough spots and opacity. Some discharge from the

right eye was also noticed in the mornings. The patient was examined again on the

16th of December 2003. At this time the patient’s comfort level had improved from

prior to the graft but the surface of the eye was still rough and lacking sufficient

lubrication at times. The visual acuity of her right eye in June of 2004 was 6/60.

            Immunological Analysis

       An epithelium of between 2 to 4 cells thickness was present in the

supplementary graft produced for Mrs BB as seen in Figure 5.14A, B and D. Staining

for p63 (Figure 5.14C) was observed in less than 10% of the epithelial cells, similar to

the levels seen in other patient grafts. The expression of keratin 3 (Figure 5.14E) was

also similar to previous grafts, being located throughout the epithelium to varying


Figure 5.13: Condition of BB Injured Eye during Treatment. (A) Appearance of

surface at first appointment; (B) Surface condition immediately prior to graft

operation; (C) Surface following peritomy and graft transplant; (D) 2 months

following grafting.

Figure 5.14: Immunohistochemistry of BB Graft. (A, D) Sections of unused

duplicate graft prepared concurrently with graft applied 30/04/03. Scale bar included

in A represents 100µM and is valid for all images; (B) The same section as A with

Hoechst-stained nuclei; (C) The same section as A and B, stained for p63; (E) The

same section as D, Hoechst-stained nuclei (blue) and keratin 3 staining (red).


       To further examine the phenotype of the cultivated autologous cells produced

for each patient, the RNA and protein was extracted from cell samples which had

been stored in liquid nitrogen following the initial propagation. RNA and protein was

also extracted from ME180 cells to act as a positive control for the p63 analysis.

       A summary of RT-PCR results is presented in Figure 5.15A. Reactions using

the Pex 1 primers were run first to confirm the presence of mRNA for use as template.

All samples produced strong bands confirming appropriate RNA concentrations.

Next, all samples were used in reactions employing the four p63 primer combinations

to assess the presence of p63 mRNA. One set of primers was designed to amplify all

p63TA isoforms (‘TA all’), another all p63∆N isoforms (‘∆N all’). The final two sets

amplify the α isoforms, p63TAα and p63∆Nα specifically. All patient cells produced

bright bands in both the ‘∆N all’ reactions as well as the ∆Nα specific reactions. AA,

MP, WG and BB produced visible bands in the TAα specific reactions. JK also

produced the correct band for TAα but it was noticeably fainter. Only BB produced

any visible band in the ‘TA all’ reactions. For all reactions correct bands were seen in

the ME180 samples.

       A summary of western blot results is presented in Figure 5.15B. A polyclonal

antibody recognising GAPDH was used to confirm equivalent protein concentration

between the samples used in the western blot analysis. All the samples produced

bands of similar intensity. The samples were next assessed using a monoclonal keratin

3 antibody. Bands were seen in all samples although they were quite faint, with the

MP sample providing the darkest band. When screened with the monoclonal p63

antibody, which is reactive to all p63 isoforms, each of the samples produced bands at

66 kDa and 62 kDa, although transfer complications distorted the 66 kDa band

somewhat in the JK sample. The 62 kDa bands in each sample were noticeably fainter

than the 66 kDa bands in each sample. This matched the result seen in the ME180

sample although the 62 kDa bands in the patient samples were relatively fainter than

those seen in the ME180 sample.

Figure 5.15: RT-PCR and Western Blotting of Cultivated Patient Limbal

Epithelium. (A) Display of RT-PCR products run on 1% agarose gels from reactions

performed using mRNA extracted from the cultivated epithelial cells produced from

the patient biopsies along with negative controls using DEPC treated water instead of

template mRNA and positive controls using mRNA extracted from the ME180 cell

line. Pex 1 primers were used as a positive control to confirm presence of mRNA in

all extracted samples prior to examination using the p63 primers. Two forward

primers and 2 reverse primers were used in combination to amplify product for all the

p63TA and p63∆N isoforms (‘TA all’ and ‘∆N all’ respectively) as well as

specifically the p63TAα and p63∆Nα isoforms. (B) Protein was extracted from the

same samples as the mRNA and then examined with Western blotting. Samples were

run on 4%-20% gradient polyacrylamide gels and then transferred to either

nitrocellulose or PVDF membranes. A polyclonal antibody against glyceraldehyde-3-

phosphate dehydrogenase (GAPDH) was used to confirm equal protein concentration

between samples. Monoclonal antibodies against both keratin 3 and p63 were used to

examine the expression of these proteins. The p63 antibody was reactive against all

p63 isoforms so the isoforms were identified by molecular weight.


       The aim of this chapter was to examine the phenotype of limbal epithelial cell

cultures used in the treatment of limbal stem cell deficiency in 5 patients and to assess

this phenotype in relation to the success of the autologous grafting approach that was

performed upon each. The markers examined in all patient samples were keratin 3 and

p63 while the first 2 patients’ constructed grafts were also examined for keratin 14

and 19. Each patient experienced improvement after the treatment was given,

although the degree of improvement was varied. The marker expression observed in

all patients was very consistent, despite the variability of their treatment outcomes.

       Of all the patients treated, Mr AA had the least improvement after treatment.

Over the course of three years and two grafting operations the surface epithelium was

repaired so that conjunctival invasion was slowed. The first grafting attempt failed

possibly due to the orientation of the graft. Initially it was decided to place the graft

face down, with the cultivated cells in direct contact with the surface, upon the

assumption that this would improve the integration. Unfortunately the graft degraded

within a few months rather than integrating into the corneal structure. The second

graft was applied in a flipped position, with the AM support base in contact with the

stripped-back corneal basement membrane. In the second attempt the graft was also

protected by a therapeutic contact lens for 2 months, to lessen the mechanical

disruption of blinking upon the graft. This approach seemed to allow the graft to

integrate much more successfully, however blood vessels still invaded. It is possible

that this vascularisation occurred through the underlying stroma, which was also

damaged by the thermal injury, but was not treated by the autologous grafting. Since

the second graft applied did not degrade as quickly as the first this approach was

employed in subsequent grafting operations for all other patients.

       Mr MP was the most successful of the patients treated. The initial graft was

applied to just over half of the injured eye and the graft was successful in retarding the

conjunctival ingrowth. This inspired a second attempt applied to the entire surface in

order to give a complete covering. This graft was also absorbed well, leaving the

corneal epithelium in a state almost identical to a healthy eye. However, this success

once again highlighted the limitations of this technique. The graft only treats the

surface epithelium so any damage inflicted deeper is not addressed. The conventional

cornea graft which was performed allowed the distorted stroma of the patient’s cornea

to be replaced with a healthy donor stroma and the repair of the epithelium through

the autologous grafting means that the patient’s own cells can maintain the surface.

       Mr JK experienced greater comfort following the grafting. The visual acuity

was initially seen to decrease, this symptom was likely due to the slight opacity of the

supporting AM of the graft but later appointments indicate that this is a temporary

concern as the graft integrates. Once again the condition of the corneal surface was

improved by the treatment, indicating that the procedure was successful in restoring

much of the patient’s ability to maintain a renewable corneal epithelium.

       Similar observations were seen in Mr WG. The major improvement observed

was the restoration of a smooth corneal surface, indicating healthy cell regeneration,

and a greater feeling of comfort. However, in Mr WG case, superficial blood vessels

were again observed invading from the limbus. This could be due to stromal damage,

as in other cases, and also it should be noted that the injury itself was inflicted 55

years earlier meaning the scarring had been well established.

       Mrs BB’s injury differed to that of the other patients in that it was inflicted via

microbial infection rather than chemical or physical injury. Nevertheless, the grafting

in her case still appeared to provide improvement in her injury. The major difference

in this case was the slight opacity and a roughness to the surface of the corneal

epithelium. This may be related to an observation that the eye lacked sufficient

lubrication at times, which could produce greater mechanical abrasion on the graft.

Overall however, her comfort had improved following the surgery and the graft was

integrating well.

       While the clinical outcomes of each graft varied, the immunological and

molecular biological observations were strikingly similar. All grafts were stained for

p63 and keratin 3 with grafts from the first 2 patients, Mr AA and Mr MP, also being

screened for keratin 14 and 19. All the autologous grafts exhibited limited expression

of p63 restricted to less than 10% of observed cells, which is comparable to that seen

in situ in the limbus (Pellegrini et al., 2001). The keratin 3 staining profile was also

similar to that seen in situ (Schermer et al., 1986) with stronger expression observed

in suprabasal cells compared to the basal population although the basal cells did

express keratin 3. Keratin 14 was observed in the constructed grafts, confirming its

role in the attachment of limbal epithelial cells to AM (Meller et al., 2002) and

indicating a highly proliferating population of cells. The presence of keratin 19 in the

grafts of the first 2 patients is encouraging due to its expression in the proposed stem

cell niche (Lauweryns et al., 1993). Taken as a whole, this immunochemical data

would indicate an immature, yet specified population of epithelial cells.

       The western blotting also demonstrated similar patterns of keratin 3 and p63

expression respectively across the patient samples. The keratin 3 levels were

somewhat faint, perhaps indicating a slightly less mature phenotype since keratin 3

expression increases as cells mature (Schermer et al., 1986). The expression of p63

was definitely skewed more prominently towards the isoform present in the 66 kDa

band although there was also a band at 62 kDa in all samples. Based upon the

calculated molecular weights of the p63 isoforms listed in Chapter 4 (Figure 4.1), this

would indicate the p63∆Nα and p63TAβ isoforms respectively. This profile supports

previous work indicating predominant expression of p63∆Nα in corneal epithelium

(Yang et al., 1998; Liefer et al., 2000). The presence of the p63∆Nα and p63TAβ

isoforms would indicate an immature phenotype, as the TA class of p63 is expressed

during embryogenesis to induce formation of a stratifying epithelium, while the

truncated ∆N class appears just prior to the cells becoming fully differentiated (Koster

et al., 2004). However, recent work has indicated that p63 expression in vitro is more

widespread, identifying both the progenitor population as well as their immediate

transient amplifying progeny (Salehi-Had et al., 2005). While this might indicate that

the cultivated cell phenotype is further removed from the stem cell population, even

so it does confirm that the cells have great proliferative potential.

       The results of the RT-PCR analysis also present interesting data. The ∆N

isoforms were clearly present in the mRNA of the samples. In particular, the ‘∆Nα’

bands were very strongly amplified. Surprisingly the TA bands were not as easily

amplified. The primer set used was designed to amplify all the TA isoforms (‘TA all’)

as well as the ‘TAα’ isoform specifically. There were faint bands present in the

‘TAα’ reactions, however in the ‘TA all’ reaction the bands were either very faint or

not visible at all. Obviously, due to the positive result in the ‘TAα’ reaction, there

should be some band present in the ‘TA all’ reaction as this result would otherwise

indicate a failure in the reaction against all the TA isoforms including TAp63α. From

the reliable data produced both the ‘TAα’ isoform and ‘∆Nα’ isoform are being

transcribed in the patient cultivated cells. Since the RT-PCR data was not quantitative

it cannot be said that one isoform was expressed at a higher level than the other but

when viewed in conjunction with the western blot data these results support the

assertion the ∆Np63α isoform is the prominent form of p63 expressed in

keratinocytes (Yang et al., 1998; Liefer et al., 2000).


       The results indicate that cultivated grafts of limbal epithelium from different

patients generally display similar characteristics. However, based on the clinical

observations, it is obvious that the application of similar grafts to significantly

different injuries leads to quite variable outcomes.

       All patient biopsies taken were successfully cultured which led to the

construction of autologous grafts. The phenotypic analyses that were performed

demonstrate that both the cultivated epithelium and the grafts created from the

different donor biopsies exhibit similar properties in the areas examined. Across all

the patients the surface of the corneas were more stable for extended periods after the

grafting was performed. Another improvement observed in all patients was an

increased level of comfort with their eye. From these observations it is clear that the

treatment is valuable in improving the condition of the corneal surface, however there

is still much room for further improvement and possibly customising of the approach

to suit different injuries more effectively. The nature of the original injury and the

condition of the corneal surface due to its occurrence will intimately affect the long-

term success and survival of the graft.