SYNOPSIS: The goal of this proposal is to use human embryonic stem (ES) cells to develop a humanized model to identify mechanisms by which cancer cells escape the normal constraints on self-renewal as well as mechanisms controlling their spread. Aim 1 will use human ES cells to create teratomas in "humanized" chimeric immunodeficient mice that will facilitate establishment of primary human tumors. The first year of the grant will be used to fully characterize the relationship of the cancer stem cells with the normal stromal cells in patients' tumors and in several xenograft tumors both by histology and by gene expression. During this time, teratomas will be established in the mice and cancer stem cells will be injected in to them. Aim 2 will use the humanized immunodeficient mouse model to study metastasis of solid tumor cancer stem cells. Aim 3 is to understand whether the human immune system recognizes cancer stem cells using Weisman’s SCID mouse model with engrafted human immune cells.
IMPACT AND SIGNIFICANCE:
The overall goal of this proposal is to use hESCS to develop a better model system to study the mechanisms by which cancer cells escape the normal constraints on self-renewal and the mechanisms of metastatic spread. The proposal’s hypothesis is based on the emerging theory that cancer growth is dependent on a population of continuously proliferating cancer stem cells within a heterogeneous tumor that interact with the surrounding tumor stroma or niche, and in this way the niche may help maintain or support the cancer stem cell population. The goal of the project overall is to develop an animal model of human tumor growth that better reflects cancer in humans. This would be a useful tool.
This is a potentially high impact application, that proposes to sort out the role of the microenvironment, and immunological response, in human cancer xenografts, by adding to the mix teratoma cells from hESCs. The applicant is a pioneer of the stem cell model of cancer, who here wishes to examine the role of stromal elements in tumor morphology and metastasis. The system of hES teratomas for an environment of human tumors in mice has already been published as a useful platform, so the real innovation here is the use of cancer stem cells in the xenografts.
QUALITY OF THE RESEARCH PLAN: The proposed research is innovative in its study and characterization of cancer stem cells. This grant will use 1 NIH approved hESC line and 2 hESC lines developed by collaborator Dr Baker (Stanford University) that are not approved by the federal government. These hESC will be used to develop teratomas in mice to provide a human tumor microenvironment in immunodeficient mice in order to study cancer stem cells in a more natural environment. This system has been reported by the Srorecki group as advantageous. The justification for using the non-NIH approved hESCs is that the one NIH approved line may not produce adequate human stroma, even though there is some mention of engineering stroma to add separately. A large body of parallel experiments on characterization of human tumors will add value to the studies.
The first aim is to create a “humanized” chimeric immunodeficient mouse to facilitate establishment of primary human tumors in the mouse. Using approaches at which the investigator is expert, he will inject tumors into NOD/SCID mice, along with ES cell-derived embryoid bodies to provide a humanized stroma. The methods are well established, largely because of the investigator’s own body of work. Endpoints include gene expression studies by Pat Brown and “genomics” analysis by Steve Quake, although neither is listed in the collaborator section of the grant. While the idea is compelling, the aim, and the grant as a whole, are limited by a lack of specificity. How many tumors? What sort of endpoints can be quantified? How many cell lines? How many ESC-derived complementing lines? The investigator is already doing this using NIH-approved lines, and new lines might not even be necessary. Another potential problem is potential incompatibility between the stroma of one individual and the tumor of another, whether because of genetic variation or the marked age difference and source of the stromal cells. An additional weakness is that it is not explained how the “stemness” of the teratoma derived cells will be more relevant or provide new specific information, compared to stroma from the patient himself.
The second aim is to use the humanized immunodeficient mouse model to study metastasis of solid tumor cancer stem cells. Here they will determine whether the human teratomas serve as a metastatic site for human colon tumors. One endpoint is whether the tumors “most closely resemble the original patients’ tumor”, but that is quite vague. Another endpoint is whether cancer stem cells invade into the periphery of the tumor, but again it is not clear how this will be measured, and how statistically meaningful conclusions will be drawn. The third aim is in collaboration with the Weissman lab, to use infiltrating lymphocytes from the patient’s own tumor, or by reconstituting the mouse immune system, to determine whether they recognize the xenografts, using a peptide-MHC tetramer assay developed by Mark Davis. It is an intriguing idea, but the follow-up to these studies is not laid out.
Overall it is an intriguing proposal with potentially high significance, but it is lacking in detail and clarity. The projects propose to test the hypothesis that ESCs can be used to generate a “human” microenvironment that will allow cancer stem cells to form tumors that will more closely resemble the tumors in humans. This is essentially a model system developed by Tzukerman and Skorecki, published in PNAS in 2003 and subsequently in Cancer Research in 2006.
The research plan is comprised of three independent but moderately high risk aims. In the first aim, they propose to make a humanized chimeric immunodeficient mouse. It is important to note that they are not dealing with somatic cell chimerism or with immunological chimerism. They are simply implanting hESCs or embryoid bodies to create a teratoma into which colon cancer cells would then be transplanted. This is straightforward and should be readily accomplished as it was established already by Tzukerman et al. Again, the goal here is to study primary human tumor growth by microarray profiling in this model human environment. Gene expression profiling will be performed through the Stanford University microarray facility and they will compare results of colon tumor cells grown in this environment to standard mouse xenografts to 100 colon tumor specimens. Aim 2 will use a similar humanized immunodeficient mouse model to study metastatic potential of solid tumor cancer stem cells, again from human colon cancers. The question to be addressed is why human cancer xenografts in immunodeficient mice do not generally demonstrate the metastatic potential that they do in their human host. The hypothesis is that the presence of a hESC-derived teratoma may provide a permissive environment for growth of cancer stem cells or alternatively chemotactic environment that promotes homing. Generally, simple straightforward experiments are described. However, the research plan is descriptive and limited in that emphasis is not placed on how outcomes will be assessed and in this case, the possibility of lymphatic spread of tumor cells is not considered. They propose to use GFP-marked ESC lines to distinguish the cancer cells from the hESC-derived human teratoma cells yet the source of this cell line is not identified. In the third aim, the PI proposes to generate a different type of “humanized” immunodeficient mouse model that is made with BM cells derived from the same hosts as the cancer cells to test the activity of human T cells against colon cancer stem cells in a human teratoma environment. This series of experiments is incompletely laid out and there are many potential pitfalls including that this model system is an older humanized model with incomplete immune reconstitution and the possibility that human cells inoculated into these animals may not only recognize tumor antigens but also human alloantigens present on the hESC-derived teratoma which may confound interpretation of the results. In general, the proposal is descriptively written with little emphasis on technical details that are relevant
STRENGTHS: The primary focus of the proposal is on the emerging concept of the importance of the stem cell niche in cancer stem cell biology, an original and important idea for cancer research. Dr Clark and his collaborators are outstanding and leaders in the field of cancer stem cells. The model system thus generated will be useful for cancer biology. Previous research work has focused on head and neck cancer stem cells and breast cancer stem cells focusing on fractionation of CD44+CD24- lineage- cancer stem cells. Significant preliminary data in this area is presented which is relevant to the proposal if similarly fractionated CSCs from colon cancers are proposed to be studied in comparison to other cell fractions or unfractionated colon cancer cells. However, this is not clearly stated. There is significant cross-disciplinary cooperation with Dr. Weisman, a pioneer in hematopoetic stem cell biology and Dr. Baker, an assistant professor of genetics who directs the ESC efforts at Stanford University whose lab has established several new hESC lines and is expert in maintaining these lines and forming teratomas. Essential expertise to conduct the proposed studies is in place. All the resources and facilities necessary to perform the described studies, including Dr. Brown’s microarray facility, are also available.
WEAKNESSES: The core aspect of both Aims 1 and 2 is to compare the expression patterns/behavior of colon cancers as they exist in human hosts (from surgery specimens) to colon cancer cells grown as mouse xenografts vs. colon cancer cells grown in the environment of hESC- or embryoid body-derived teratomas. The endpoints for these studies are not adequately described and do not seem to be quantitative. Firstly they propose to compare histology which is clearly nonquantitative and will be rather descriptive making conclusions difficult. Secondly, they will be comparing pieces of heterogeneous tumors by microarray analysis and it is not clear that they will obtain meaningful information because of the variety of human tissue types within the teratomas which will confound the data. That is, genes expressed may be from the colon cancer cells or the tumor cells and the source will not be able to be resolved with this methodology.
Many details and specifics of the methodology as well as the outcome measures are not carefully considered. For example, in one case they propose specialized flow cytometry to isolate various components of the tumor and study them individually through genomics analysis at the microarray facility. However, they do not state by what antigens or antibodies they will separate these cell populations.
In general, consideration of NK cell mediated immunological activity against the tumors or lack thereof is not considered. The rationale for studying to non-approved ESC lines that were recently developed at Stanford University is weak. As discussed above, this work is already going on with an NIH-approved line. Finally, the source of GFP-tagged cells is not identified.
In Aim 3, many details of the proposed studies and experimental design are not included: for example, how many cells are needed for the tumor infiltrating lymphocyte studies and how many cells will one obtain from the human tumor specimens? Simple considerations such as how much tumor tissue will be sent for research vs. clinical use are also not mentioned; if limiting then this may preclude these studies from being performed. Also in Aim 3, an older humanized mouse system is proposed, and there are newer, more robust humanized models recently published. The studies in this aim do not consider that the human BM-derived immune system that develops and/or the human colon cancer TIL that will be infused into the animals may not only recognize the colon cancer tumor antigens but also hESC\teratoma-derived alloantigens potentially confounding the interpretation of results.
DISCUSSION: There was some discussion about whether the fact that this was primarily a cancer biology grant that would yield little new information about hESCs made it non-responsive to the RFA. The conclusion was that it was within the purview of the RFA. There was general agreement that although the applicant is regarded as "brilliant" the proposal was very poorly written, illogical in places with many experimental details missing, and with important collaborators referred to, but with no documentation or formal indication of their participation. There was some discussion about the recent productivity of the investigator, but his recent move was regarded as a mitigating circumstance. Finally, although sophisticated about cancer biology, the proposal seemed to lack requisite knowledge and expertise in stem cell biology.