U.S. Government Defends Begal Research

Dear Ms. Miers,

I am an experimental psychologist who has several years of experience in research. Some of that research is in neuroscience. I have performed surgical procedures on several animal species and have served as principal investigator on autism research projects, worked in the cancer research center of the Salk Institute, and on a spinal cord research project. I have received a copy of an information sheet, sent by your office, to my Congressional Representative, John Shadegg. The information sheet refers to a research project funded by the National Institute of Neurological Disorders and Stroke. The title of this project is Allogenic Glioma in Immune Competent Dogs, Grant number: 1RO1NS38316, Principal Investigator: Michael Berens.

I have carefully reviewed, many, many times, the following documents and conversations:

1. Arizona State University Institutional Animal Care and Use Committee Minutes for this project.
2. ASU subject records for animals who served as subjects in this project.
3. Data submitted in 1999-2000 continuation forms to NINDS.
4. Several papers describing other canine glioma models and differences in the routes of invasion by cells in naturally occurring tumors as compared to cells in implanted tumors (as in Dr. Berens' animals).
5. Letters and memos from the office of extramural programs at NIH raising issues of concern regarding IACUC reviews done of animal research studies at Barrow from 1986 through 1998.
6. Correspondence from Geoffrey Pilkington, a scientist in the UK listed as a contributor to the canine glioma project.
7. I have spoken with the assistant Executive Director of AAALAC International and I am aware of the serious noncompliance problems faced by Barrow in 1998.
8. I have spoken with surgeons and veterinarians who are familiar with fetal surgeries and who have reviewed the ASU minutes and the 1999-2000 continuation data.

You state that Dr. Berens is attempting to develop a canine model of a brain tumor similar enough to a human tumor to be suitable to test reliably therapies for humans. I wonder if you are aware of the recent publications by Pilkington and others (1992) describing in detail the fundamental ways in which brain tumors artificially implanted in the brain of an animal differ from naturally occurring tumors in animals and humans with respect to patterns of invasion? Since the final step in the creation of the "allogenic model" is to remove a subcutaneous tumor and implant it directly on the brain of a four-month to six-month old puppy, this model appears to be subject to these problems. I have received a direct communication from Dr. Geoffrey Pilkington stating to me that he will not support implanting of cells into living animals. Since Dr. Pilkington is listed incorrectly in Dr. Berens' grant application to NINDS as supporting the development of an allogenic model, I feel that his comments, together with his expertise, are of some relevance.

I also wonder at the premise that this model will create a tumor similar to naturally occurring tumors in other respects. A fundamental manipulation in the creation of this allogenic canine model is the implanting of glioma cells in fetal puppies prior to thirty-seven days in gestation. The stated intent of this manipulation is to introduce the cells at a time when the immune system will not recognize them as non-self, i.e. to desensitize the immune system in such a way that the use of immunosuppressive drugs will not be required. This is far removed from the normal way in which tumors develop in children and in adults. No such manipulation occurs in them. In any case, the empirical data through 1999-2000 indicate that only three puppies have been documented to host a subcutaneous tumor in ten years.

I also am offended by your reference to children as one of the primary patient populations who are struck by these gliomas and your attempt to use an emotional response to this fact to justify relentlessly negative research over a ten year period at the cost of animals who have been wasted and died. I do not discount the suffering and death of these children but wonder why NINDS funds research with such abysmal results and uses children as one excuse. I have worked for years with children who have autistic disorder, Asperger's syndrome, PDD NOS, and Down's syndrome. The fact that these children have devastating disorders, and in the case of Down's Syndrome, often have their life span shortened, does not indicate that unsuccessful research is merited. The attempt to create an animal model of autism has all but been abandoned because it led to such abysmal results and took time and resources from more promising lines of research. I find your allusion to children as justification for abysmally failing research reprehensible. As a psychologist who worked with parents of children from very special populations, I know that to hold false hope of positive research results when none are indicated is both cruel and unethical. Sometimes hope lies in abandoning unpromising lines of research in favor of those that will bring substantial gains in the understanding and treatment of diseases and disorders that affect both children and adults.

It is my impression, based on the material that I have reviewed, that Dr. Berens was not funded on the first submission of this project and that several reviewers expressed concerns. The project was funded after the second submission. I am well aware that reviewers from both NINDS and the National Advisory Neurological Disorders and Stroke Council reviewed this grant. However, neither group was made privy to seven years of relentlessly negative pilot data generated during the years 1990 through 1997 while the project was housed at Arizona State University.

During that time, Dr. Berens was approved to use 479 animals. Twenty-one litters were aborted during fetal surgery. To date, one-hundred-and-twenty puppies have not developed a predicted subcutaneous tumor and have been used as subjects for stereotaxic surgery practice and then euthanized or used as donors for ancillary parts of this project. Fifty-six animals were stillborn or born with severe birth defects including hydrocephalus and missing limbs. One puppy may have had a subcutaneous tumor but was euthanized before it was discovered. Seven animals died of parvovirus. Forty-four females were used for breeding and are unaccounted for. Two puppies were moved to the final stage of a brain tumor.

One case of an animal hosting a brain tumor from this project is documented in the only published paper showing demonstrable data from this project in ten years (Berens et all, Neoplasia, 1999). The tumor in that animal had destroyed the entire left side of the brain. What were the clinical endpoints for this dog? The Arizona State University Veterinarian communicated to me that a subcutaneous tumor was never documented in any animal that he had seen. It is these concerns that, in part, led ASU to terminate this protocol in 1999. I quote from minutes of the ASU IACUC meeting, October10, 1997. "When pressed for a clear statement of the success rate over a seven year period of research, Dr. Berens estimated the success rate at approximately 5%. When asked if he would continue the research if improvements were not noted in the next three years, Dr. Berens did offer that he would be discouraged and perhaps would terminate the research." Data from the 1999-2000-continuation submission by Dr. Berens to NINDS indicate that the failure rate during that period was 98%.

You state that animal welfare concerns have been addressed and that the Barrow IACUC has approved this research. I find it odd that NINDS chooses to accept the approval of the Barrow IACUC, noncompliant with respect to de novo reviews and agreements to submit semiannual reports for over a decade, rather than the appropriate review by the ASU IACUC over a nine year period.

The ASU IACUC has not been guilty of ongoing noncompliance with PHS requirements for conducting animal research on funded projects. On the other hand, Barrow certainly was guilty of such oversight and noncompliance for more than a decade. In fact, Dr. James O'Donnnell, the head of the NIH performance site visit team to Barrow in 1998 wrote, in a letter to the Barrow Institutional Official, that if the institution continued in its present behavior (with respect to reviews and other noncompliance issues) it jeopardized itself and the entire scientific community.

The Chair of the Barrow IACUC at that time was Thomas Hamm who signed the 1996 animal welfare assurance indicating that review processes were being done correctly when in fact they were not. Dr. Hamm's own project, Recurrent Renshaw Circuits, was one of the projects most noncompliant with respect to the de novo review process. Dr. Hamm continues as the Chair of the Barrow IACUC. This is the IACUC that continues to review this project. ASU had such serious reservations about this research that, in 1997, the project was put on restricted approval and reviewed yearly. Whom do we believe? As a scientist, I believe the members of the ASU IACUC who were cognizant of the empirical data and animal welfare issues, of Dr. Berens' own admission of negative results, who were taking their review and reporting responsibilities seriously, and who chose to terminate the research. The policy of enforced self-regulation, with respect to an animal used in research, is only as good as the IACUC that is responsible for the review process.

I am including subject records based on ASU and NIH documents. Dr. Berens continues to euthanize every healthy animal not developing a subcutaneous tumor. Since almost every puppy does not develop a tumor, he continues to euthanize almost every puppy in this project. Several prominent members of this community have approached both Dr. Berens and Dr. Hamm about a more effective and aggressive adoption program for these puppies, neither Dr. Berens nor Dr. Hamm have been responsive to these inquiries. May I remind NINDS that it is the taxes of these community members that support this project. In addition, the NIH continuation forms indicate that the high incidence of aborted litters and puppies born dead continues. There is no protection for these animals. The protection for these animals ceased in 1999 with the termination of review by ASU. At that time sixty-eight healthy puppies had been euthanized. The number has risen to one-hundred-and-twenty. The failure rate of this project is abysmal. There is no dignity in NINDS continuing to support such carnage.

It is quite clear that the point of this research is to create an allogenic canine model of a glioma by injecting fetal puppies with glioma cells from sick dogs, removing predicted subcutaneous tumors at four to six months of age from the puppies, and then implanting the tumors, using stereotaxic surgery, into the puppy's brain. The theoretical point and premise of this research is quite clear, however, may I point out your statement beginning, "If this process is successful..." The relationship of ASU to this project, prior to NIH funding in 1997, provides much of the empirical data for the project's real probability of success. The concerns of ASU became so grave that, from 1997 through 1999, the project was approved with restrictions and reviewed yearly. When the project was reviewed in 1999, no substantial progress in surgical procedures or empirical data had been made and the IACUC voted to terminate the protocol. This lack of progress is reflected in the paucity of published papers resulting from the project. Dr. Berens has published extensively his in vitro work, Dr. Berens has published one paper on the canine model, presented two papers, and perhaps a third, presented anonymously. No matter how many times he presents the same data or ancillary data based on cell analyses, there is no more evidence of success in creating the canine model, the stated intent of the project. I am including below summary tables from ASU data and NINDS data.

Summary of data from Arizona State University Subject Records. These records do not account for all dogs used in the project. ASU subject records indicate Dr. Berens was approved to use four-hundred-and-seventy nine animals from 1990 through 1999. Including aborted litters, the records account for approximately three-hundred-and sixty dogs and puppies. 1990-1999 summary. Successful procedures based on the number of implant surgeries 1990 to 2000 = 2/38 = 95% . 0526 (Failure rate = 95%). Successful procedures based on the estimated number of puppies available for implant 1990 to 2000 = 2/322 = . 006 (Failure rate = 99%)

Summary of data from Arizona State University Subject Records

Accounting of Animals: Data reported by Dr. Berens in 1999-2000 NINDS Continuation forms. 1999-2000 summary: Data from 1999-2000 continuation forms for number of subcutaneous tumors in puppies implanted: Puppies born dead, preterm, or aborted = 38 Puppies born live and euthanised according to protocol = 11, Success rate: 1/47 = .02 (98% failure rate)

Data provided in the 1999-2000 continuation submission to NINDS indicate that this project continues to fail. Although, in the text of the continuation forms (p. 3), Dr. Berens states that seven fetal implant surgeries were performed. But one table, presented in the submission forms, illustrates the results of eight surgeries. During this period, only one puppy in forty-seven animals may have developed a subcutaneous tumor. Thirty-seven animals were born preterm, stillborn or died shortly after birth. Nine puppies were born with no tumor and euthanized. This again is a 98% failure rate.

The stated intent of this project is to create an allogenic canine glioma model. ASU did not feel, after nine years, that the research had shown any real progress toward that goal. Data submitted to NIH in the 1999-2000 continuation forms do not demonstrate any improved results.

I am struck by the fact that, in the 1999-2000 NINDS submission, Dr. Berens attributed poor results to needed refinements of surgical techniques and gestational staging. These are the same problems discussed over a nine-year period in earlier ASU IACUC reviews. As did members of the ASU IACUC, I believe that it is time to acknowledge this project has failed.

It may or may not be true that the growth and invasion patterns of glioma in dogs more closely approximate those in humans, but a model that is not empirically available, holds no use for the study of tumors in dogs or humans. In addition, some researchers suggest that tumors implanted on the brain of an animal have very limited use and applicability to the study of gliomas and treatment of gliomas in dogs or humans. Several large glioma research projects exist in this country and in the United Kingdom in which veterinarians carefully screen and recruit animals with naturally occurring tumors for participation in these projects to circumvent the problems inherent with artificially implanted tumors in animals. Perhaps the true intent of this funded canine glioma project is, as stated by Dr. Berens in a letter to me, to create animals for use in FDA testing and similar uses. There are enough canine glioma models available for this purpose already.

I am quite clear on the rationale for the fetal implant, however, there is little indication that the fetal implant has worked in ten years: 1) one reported subcutaneous tumor in 1999-2000, and 2) two reported subcutaneous tumors from 1990 to 1999. If more data are available, why is it not being submitted to NIH and published? Dr. Berens has secured a patent on the fetal implant. Members of the ASU IACUC suggested that Dr. Berens adopt the Salcman et al. procedure of implanting two-day old puppies with brain tumors by implanting the cells subcutaneously and debulking the tumors to prolong the life of the puppies. Dr. Berens declined this suggestion. This procedure is not patented by Dr. Berens.

Apparently, researchers in the UK and in this country who use clinical cases of spontaneously arising canine gliomas have either been able to recruit adequately and successfully to circumvent the problems you describe or acknowledge that several other canine models of gliomas already exist when dogs hosting spontaneous tumors are not available.

Following a relentlessly negative attempt to create this model, Dr. Berens research was terminated by the ASU IACUC. One of the primary concerns of this IACUC was Dr. Berens' continued destruction of healthy puppies not developing subcutaneous tumors. One-hundred-and-twenty puppies have been destroyed. The head of the ASU animal care facility acknowledged that their facility refused to participate in the euthanasia of any more healthy puppies and to house any more puppies for this project. Dr. Berens is using healthy puppies as subjects for stereotaxic surgery practice, Dr. Berens has aborted at the minimum twenty-one litters of puppies, created congenital birth defects and stillbirths in others, and continues along this same course. The Arizona State University IACUC also reviewed this project in full board reviews. These reviews were done during the period in which Barrow and St. Joseph's were not doing full board de novo reviews of continuing projects. In fact, only the ASU reviews were compliant with PHS policy. During that time, Dr. Berens attempted forty-three fetal implant surgeries. Twenty-one litters were aborted by the procedure, thirty-six puppies were born with severe birth defects, still born, born preterm or died shortly after birth. Seven puppies died of parvovirus at Barrow. Sixty-eight puppies survived the fetal implant procedure. Sixty-seven did not show any evidence of a subcutaneous tumor and were euthanized and used for practice of stereotaxic surgical procedures and as tissue donors for some aspect of this project. One puppy, possibly hosting a tumor, was euthanized by mistake. In a nine-year period, Dr. Berens reported that only two animals were moved to the stated final step of this project and harboring a brain tumor.

ASU voiced the following concerns:

1. Dr. Berens' surgical procedures. This concern was so serious that, in 1998, Dr. Berens was required to report in detail the results of each fetal surgery including the number of puppies in each litter, the number of aborted litters, and the number of congenital defects.
2. Dr. Berens was unable to identify fetal puppies who received glioma injections and those who did not. A member of the ASU IACUC acknowledged that this problem was never resolved.
3. The overall failure rate of the project was 95% in nine years; 3) other canine glioma models had been developed including one in immune competent puppies (Salcman, Scott, Schepp, Knipp, & Broadwell, 1982).
4. The waste of animals and prolonged confinement of healthy animals.
5. The lack of clarity in both the NIH protocol and ASU protocols regarding parameters of treatment for puppies exposed to experimental radiation and chemotherapy. When questioned about support for puppies exposed to experimental radiation and chemotherapy, Dr. Berens stated that he would monitor the white count of each puppy and "back off when it can't take anymore."

The ASU IACUC had such growing concerns over this research that it culminated in notification for Dr. Berens to withdraw the protocol or face termination of it. The policy of enforced self-regulation is only as strong as the IACUC that supervises the research.

Again, I point out that an IACUC at ASU that has no history of noncompliance issues refused to support this research. An IACUC at Barrow with a history of noncompliance issues documented by NIH, now asserts no animal welfare concerns exist.

You state that Dr. Berens' research has advanced less rapidly than NINDS had hoped. You also state that a fundamental problem has been a puppy's immune response. Minutes from the ASU IACUC indicate that a fundamental problem has been the ability to make the fetal surgery work at all. Over a ten-year period, Dr. Berens has had the help of many veterinarians and experts who have attempted to make the surgery work with continuing negative results. With respect to immunological tolerance in puppies who develop subcutaneous tumors, one-hundred-and-twenty puppies have survived this fetal implant procedure and have not developed a tumor.

Since Dr. Berens failed to do a critical control study at the beginning of this project to identify the number of years over which a subcutaneous tumor may develop, and since he continues to euthanize all the healthy animals who do not develop subcutaneous tumors, how does NINDS begin to understand when, in life, these tumors may develop, or attribute their lack of development to complicated hypotheses related to immunological and genetic characteristics?

Finally, you state that Dr. Berens has created several brain tumors using this allogenic fetal implant procedure. I have seen empirical evidence of one such tumor in one dog (Berens et al. (1999), Neoplasia), ASU saw evidence from that dog and one other, the NINDS continuation forms do not supply empirical evidence of any others. One again I wonder why Dr. Berens submitted to your agency a grant application markedly missing seven years of relentlessly negative data covering the period 1900 through1997, the current data continue to reflect the same pattern of results, and why your agency continues to support this project?