Reaction from StemCells, Inc. to Two Papers in Stem Cell Reports on the Efficacy of Human NSCs in Mouse Models of Alzheimer’s Disease and Spinal Cord Injury

The two articles in this issue of Stem Cell Reports authored by UCI researchers are based on sponsor-supported collaborations that were conducted in the effort to develop therapies for serious and debilitating neurological disorders that have no effective therapy: specifically, cervical spinal cord injury (SCI) and Alzheimer’s disease (AD) (Anderson et al., 2017, Marsh et al., 2017).

As industry sponsor of four separate INDs involving human neural stem cell (HuCNS-SC) transplantation in human subjects (over a course of more than a decade), StemCells, Inc. held as an absolute that all emerging data from preclinical studies were to be carefully analyzed for any safety concerns and potential impact on clinical development. The data from each of the reports published in this issue were carefully reviewed by the company and discussed at length with the respective senior authors. The outcome of the animal study by Anderson et al. (2017) was disclosed to the clinical investigators for the company’s Pathway Study and to the U.S. FDA in regulatory filings. The observations of focal cellular clusters from Marsh et al. (2017) had been previously observed and extensively analyzed in other in vivo studies and was also disclosed to the FDA as part of required regulatory filings. After careful analysis, the company reached the conclusion that neither study indicated a safety concern.

We respectfully disagree with several interpretations made by each senior author regarding the analysis of the animal data and their relevance for the clinical studies. Brevity prohibits a point-by-point response to the various results in the two papers, but none of the conclusions reached in either report required alteration to then ongoing clinical testing or informed consent. The paper by Anderson et al. (2017) (1) assumes the animal model of chronic cervical spinal cord injury is predictive of efficacy in the human setting, (2) does not acknowledge that the mild cervical hemi-contusion injury is not validated as a translational model, and (3) fails to emphasize that neither the research nor clinical cell line (when directly compared) achieved evidence of efficacy. The paper by Marsh et al. (2017) confirms that HuCNS-SC was not effective in the specific model of AD, but also raises the possibility that the observations of focal clusters of cells represents a safety concern, despite (1) any adverse clinical, behavioral, or histological observations in the animals and (2) the assessment by an independent veterinary pathologist that the histology was consistent with normal in vivo behavior of neural stem cells and did not represent a concern for neoplasia.

To date, after 10 years of clinical trial experience involving HuCNS-SC transplantation in more than 50 subjects (with clinical follow-up extending to 5 years post transplant), absolutely no safety concerns regarding the HuCNS-SC cells have been identified.

We share the senior authors’ frustration regarding the failed cervical SCI and negative AD outcomes of the respective preclinical studies and acknowledge the complexity of successfully translating cellular therapy in human patients for neurological disorders. However, we note that the ultimate test of efficacy resides in the human setting, and the limitations of animal models for predicting efficacy in human neurological disorders are well recognized. Thus, it is important to interpret data derived from animal studies in the larger context of the weak correlation between success in animals and success in humans. Clinical development for a treatment with a strong therapeutic hypothesis is therefore first and foremost based on identifying safety issues in preclinical studies and in the subsequent accrual of human safety profiles. Nonetheless, given the limitations of animal models for predicting efficacy, we have been very reassured by the signals of efficacy, albeit modest, observed in the early human studies conducted to date with HuCNS-SC cells (across multiple cell banks), and that it is these emerging signals that has supported further clinical investigation.

In this spirit, we sincerely hope investigators focused on preclinical studies in cellular therapy continue to push the field forward, but we also acknowledge that no animal model can fully recapitulate the experience of testing human subjects, nor become, in the absence of safety concerns, the exclusive factor on which clinical testing is based. In closing, we wish to express our gratitude to the physicians and patients who understand the nature of clinical translation and that human outcomes ultimately determine whether therapeutic testing continues to advance.