1
INTRODUCTION: RESPONSE TO REVIEWERS
The applicant appreciates the reviewers’ thoughtful review and critique. As a result of
responding to the reviewers’ comments, the application has been significantly revised. The
Aims and Specific Aims are regrouped to monitor the effect of tumor growth and chemotherapy
on total tissue and intracellular Na+, water diffusion, and perfusion in subcutaneous (SC) (Aim 1)
and orthotropic (Aim 2) hepatocellular carcinoma (HCC) rat models. Specific Aims (a, b, and c)
in each Aim are dedicated to monitoring changes in untreated tumors (a), after chemotherapies
(b), and to studying the mechanisms of these changes (c). In the Preliminary Data section, new
experimental data concerning monitoring of perfusion and diffusion components of water
apparent diffusion coefficient (ADC) in diethylnitrosamine (DEN) -treated liver have been added.
Specific Aim 4 from the previous proposal has been deleted due to the first reviewer’s
recommendation.
1. Innovation of the proposal: The first and second reviewers raised the concern that
studies on proton and sodium magnetic resonance imaging (MRI) assessment of
tumor chemotherapy have been previously done by multiple investigators.
As a tool for monitoring chemotherapeutic response of the body tumors, diffusion weighted
(DW) 1H MRI was used by many investigators and 23Na NMR mostly by us. However, those
studies were performed using predominantly SC tumor models. The measurement of water
ADC in abdominal tumors is very challenging due to its sensitivity to respiratory, cardiac, and
other physiological motions. Furthermore, neither single quantum (SQ) nor triple quantum-
filtered (TQF) 23Na MRI were used in studies of animal models of HCC. The central hypothesis
of our proposal is that 23Na MRI techniques can provide effective methods independent from
these motions for monitoring and imaging post-treatment changes in HCC. The MR experiments
outlined in the proposal will also provide information about the physiological role of the
transmembrane Na+ gradient and its relationship with cellular bioenergetics and pHi during
untreated growth and post-treatment tumor transformation. TQF 23Na MRI in monitoring early
physiological and metabolic chemotherapy effects, such as changes in intracellular Na+ will be a
more reliable tool for monitoring chemotherapeutic response in orthotopic HCC.
Unlike in our previous 2002-2005 grant when only SC tumors, such as radiation induced
fibrosarcoma (RIF-1) or 9L glioma, were the subjects of the studies, one of the goals of the
current project is to investigate the compartmental and metabolic differences and similarities in
tumors growing orthotopicaly and SC.
2. Human studies: The first reviewer stated that without a clear understanding of what
the obtained animal data mean, studies in humans are considered premature.
In our new application, we decided to delete the Specific Aim 4 and test the feasibility of SQ and
TQF 23Na MRI for monitoring the efficiency of chemotherapy in liver cancer patients in our future
applications.
3. Use of 23Na MRI for monitoring chemotherapy response: The fifth reviewer raised the
concern that some MR characteristics can affect the corresponding interpretation of
the results making them ambiguous.
A relatively long echo-time (TE) is used to achieve short sweep-width (3,700 Hz) and optimize
the signal-to-noise-ratio (SNR) in the 23Na images. The noise in an image is directly proportional
to the square-root of the sweep-width; thus decreasing the sweep-width should increase the
image SNR. However, decreasing the sweep-width increases the acquisition time and TE,
resulting in signal loss due to T2* relaxation. The optimum TE and sweep-width that give the
maximum SNR were calculated as described by Vinitski et al. (1987) from the relaxation
characteristics of the tumor 23Na signal and other imaging parameters used in the study. In
addition, the weighted signal summation technique was employed in the two phase-encoding
directions to further improve SNR (Bansal, Seshan, 1995). Our previous data with SC 9L glioma
2
and RIF-1 have shown that SQ and TQF 23Na signal intensity (SI) are very sensitive to
compartmental, physiological, and metabolic post-chemotherapy changes (Winter et al., 2001;
Babsky et al., 2007).
4. Statistical data analysis: The third reviewer raised some questions about statistical
analysis of data and recommended to use a statistical collaboration in the project.
Despite the fact that human studies has been deleted in new proposal the dedicated
biostatistician (Dr. Susan Perkins) will be involved in the project for proper analysis of animal
MRI data. Further details regarding statistical analysis have been added, as well as justification
of sample sizes in the animal studies.
5. Sensitivity of DW 1H and 23Na MRI methods: The forth reviewer raised the doubt about
the sensitivity of the proposed NMR methods and proposed that careful dose
response studies would be one approach which might allow development of a model
of disease response.
Non-invasive DW 1H and 23Na MRI techniques have the potential to identify intra-tumor
physiologic and metabolic changes prior changes in tumor volume. The increase in water ADC
and total tissue Na+ was demonstrated for SC RIF-1 tumors (Zhao et al., 1996) and intracranial
9L glioma (Schepkin et al., 2005) even before tumor shrinkage. Furthermore, even small early
changes in water diffusion and 23Na SI can be identified by MRI methods. In our previous
studies, DW 1H and 23Na MRI showed a ~20% increase in water ADC and SQ Na+t after
treatment of RIF-1 tumor with cyclophosphamide or 5-fluorouracil (see Preliminary Data
section). These relatively small changes have been already identified at 48 hours post-
treatment. In addition, conventional methods usually are painful and invasive (e.g. biopsy) or
radiation exposure (radiography, CT, PET). The dose response studies for Specific Aims 1b and
2b will test four different dose combinations of doxorubicine (DOXO), 5-fluorouracil (5FU), and
benzamide riboside (BR).
6. Animal models of HCC: The first reviewer raised the concern that the rationale for
choosing three animal models of HCC is not clear.
Monitoring of chemotherapeutic response of liver and other body tumors using DW 1H MRI has
been performed predominantly using SC tumor models. This model is not sensitive to motion
effect and allows easy measurements of tumor volume by caliper. However, chemotherapeutic
effects in SC HCC can differ from intrahepatic (IH) HCC due to differences in development of
circulatory system, tumor temperature, oxygenation, and pH etc., which can all affect drug
delivery and action. For example, it is well known that many human HCCs are better perfused
compared to the normal liver [McLennan et al., 2005]. In contrast SC tumors are generally
hypoxic. Therefore, effect of different chemotherapeutic drugs will be estimated in both SC (Aim
1) and IH HCC (Aim 2). N1S1 cell inoculation produces multiple well-encapsulated distinct
masses of transplanted (TP)-IH HCC without visible metastatic growth in other organs. This
model allows monitoring tumor volume changes using 1H MRI and comparing them to SC HCC
data. From other side, spontaneous (SP) DEN-induced HCC is a comprehensive and realistic
model for study of the mechanisms of carcinogenesis and post-therapy changes that closely
imitates the different types of benign and malignant hepatomas in humans (Li et al., 2005;
Hemmings et al., 2002). DEN also stimulates a variety of liver lesions such as steatosis,
hepatitis, fibrosis, and cirrhosis that are seen in human disease. The comparison of changes in
23
Na and DW 1H MRI SI in the three tumor models of HCC will also allow us to make a
conclusion about the general or specific nature of these changes.
7. Tumor volume: The first reviewer raised the concern that it is not clear why the
investigator chose to start MRI measurements when tumors reach rather large sizes.
3
In the first 7-10 days after cell inoculation SC HCC is not visible and thus cannot be measured
by caliper. In both control and treated groups, the initial tumor volumes will match each other,
and thus even the small changes after the chemotherapy can be identified.
8. Sodium MRI as a marker of chemotherapy efficiency: The fifth reviewer raised the
concern that the central hypothesis that sodium MRI may replace water ADC MRI,
because sodium imaging is less sensitive to motion artifacts, is more an indication of
avoiding the problems then solving them.
Sometimes avoiding problems is more fruitful than solving them. Measurement of water ADC in
abdominal tumors is very challenging due to its sensitivity to respiratory motion. In trying to
solve this problem, we used a number of approaches, such as 1) respiratory gating of MR
signal, 2) threshold of ADC map to eliminate the artifact in ADC values higher than free water
diffusion (>3 mm2/s), 3) threshold of the noise level in ADC map, 4) comparison of water ADC in
IH HCC and in surrounding liver tissue, and 5) use of a SC HCC model. Each of these
approaches decreases motion effect but none solve the problem completely. However, similar
to DW 1H MRI, total tissue Na+ measurement by SQ 23Na MRI reflects primarily the changes in
ECS and are not sensitive to liver motion. In addition, intracellular Na+ measurement by TQF
23
Na MRI also is not affected by motion and mostly reflects physiological and metabolic changes
in the tumor. Furthermore, shift reagent aided 23Na MRS allows one to measure ADC in extra-
and intracellular space simultaneously as is shown in the Preliminary Data section for ischemic
skeletal muscle. Thus, the monitoring of Na+ level in HCC by 23Na MR can allow non-invasive
evaluation of the pre- and post-treatment structural and metabolic changes in HCC, because
23
Na MRI is not affected by motion effect and Na+ transmembrane sodium gradient reflects the
activity of many physiological and metabolic processes in tumor tissue.
9. Resistant vs. non-resistant tumors: The third reviewer raised concern that there is no
indication of how “resistant” and “non-resistant” tumor will be defined.
Resistant and non-resistant tumors will be identified by the post-treatment changes in the tumor
volume (as it was done for SC-implanted 9L glioma (see Preliminary Data section)), histology,
and clonogenic assay data. To our knowledge, approximately 50% of HCCs will be responsive
to the most effective drug therapies used in the proposal.
10. PI responsibility: The second reviewer raised concern that it is a bit unclear which of
the two PIs will be in charge.
In the new proposal, Dr. Babsky will be the sole PI and will be responsible for the overall
direction and planning of the studies. Dr. Bansal will serve as a Co-investigator, responsible for
developing and planning NMR experiments.
11. Surrounding tumor normal tissue: The third reviewer raised concern that many of
the animal experiments will involve comparisons between paired tumor and
surrounding normal tissue, which are not specifically addressed.
In case of TP- and spontaneous (SP)-HCCs, the nearby liver tissue will be used as surrounding
normal tissue. In case of SC-HCC, the nearby skeletal muscle will used for this purpose.
12. The latest advances: The fifth reviewer raised concern that the application did not
demonstrate knowledge of current achievements and recent advances using sodium
or diffusion MRI to evaluate tumor therapy.
The latest publications of the recent advances using sodium or diffusion MRI to evaluate tumor
therapy have been added and discussed in the Background and Significance section.