Is belief larger than fact: expectations, optimism and reality for translational stem cell research

We built a dataset of newspaper articles using Factiva.com (Dow Jones & Co.), a comprehensive database of major newspapers from the USA and UK, which also includes all major newspapers in Canada. We searched without date restrictions for: ('stem cell' or 'stem cells') AND (therapy or therapies or therapeutic or therapeutics or treatment or treatments). Factiva identified and removed duplicates. We then analyzed both our CT dataset described below (title, brief summary, and detailed description for each CT) and the full-text newspaper dataset, using a custom word frequency analysis program that enabled a set of documents to be searched using multiple words or phrases, separated by and/or. The result of each search query was not the number of times that a search term occurred; rather, it was the number of documents containing the term(s). We compared word frequencies using χ² analyses in two time blocks: 1990 to 2000 and 2001 to 2010. While the broad search strategy may have captured newspaper articles that were only tangentially related to translational SC research, our keyword analysis extracted newspaper articles from this set that specifically referred to SC types and disease categories.

We searched the term 'stem cell*' in the world's largest and most comprehensive database for CTs: ClinicalTrials.gov and a second portal that captures international trials not registered with the US National Institutes of Health (NIH): the World Health Organization's International Clinical Trials Registry Search Portal [17]. The latter included trials from Australia and New Zealand, UK, Brazil, China, India, South Korea, Cuba, Germany, Iran, Japan, Africa, Sri Lanka, and The Netherlands. We removed non-SC-related trials that were nevertheless captured by the search and duplicates (trials registered in more than one database). We constructed a database of CT information using all fields, limiting our analysis to CTs registered before 1 January 2011.

We coded all CTs in the dataset for the principle disease/condition addressed, and CTs could fall within more than one category. We then developed inclusion and exclusion criteria to identify groundbreaking or novel SC therapies and coded the trials based on their description supplemented by related publications and websites. The latter was necessary, for example, to characterize a therapeutic agent identified only by its proprietary name in the trial description.

Our definition of novel SC therapies was CTs that involved: (1) the use of stem or precursor cells to stimulate non-hematopoietic organ regeneration (for example, limbal SCs for corneal regeneration or hematopoietic SCs for cardiac repair); (2) the use of agents to stimulate stem or precursor cell action for regenerative or therapeutic purposes; (3) the use of established hematopoietic SC transplantation procedures for novel indications (for example, autoimmune or congenital diseases); (4) the use of novel agents or processes for stem or precursor cell collection, purification or expansion; and (5) the use of genetically modified stem or precursor cells. These categories fall predominantly within Mason et al.'s definition of regenerative medicine [18].

We excluded CTs from our 'novel' category that were observational in nature, for example, some CTs measured circulating endothelial precursor cells without a therapeutic goal; involved an established SC therapy for an established indication, most commonly hematopoietic SC transplantation for hematological malignancies; or investigated supportive measures surrounding an SC therapy, for example, antibiotics to prevent infection in hematopoietic SC transplant recipients.

Newspaper articles focused mainly on human embryonic SCs and neurological conditions with significant coverage of cardiovascular disease and diabetes as well. Newspaper coverage peaked in 2001 and again in 2005, coinciding with US President George Bush's 2001 Directive limiting the use of federal funds for derivation of new hESC lines and similar discourse following his re-election at the end of 2004 (Figure 1A) [4]. Notably, a similar spike in coverage did not accompany US President Barack Obama's Executive Order, signed 9 March 2009, reversing President Bush's Directive. We discussed the framing of the issues in newspaper coverage around hESC research in a previous paper [4]. The key result here is that hESCs dominated newspaper coverage of SC research, with slight coverage of the discovery of iPSCs in November 2007, mainly focused on their being an 'ethical' alternative to hESCs (Figure 1A) [6]. This leaves hESCs front and center in the public mind with respect to the type of stem cell in clinical development. In contrast, while newspapers referred to adult SCs and occasionally their tissue of origin, they rarely referenced terms such as hematopoietic and mesenchymal, or synonyms for these such as blood SCs. In addition, newspaper articles focused disproportionately on neurological conditions, primarily, multiple sclerosis, stroke, Parkinson's disease, spinal cord injury and Alzheimer's disease with significant coverage also of cardiovascular disease and diabetes (Figure 2A). Neurological diseases were significantly linked to coverage of hESC research and therapies. Promising research in therapies for ocular conditions such as corneal repair and macular degeneration received a small but steady amount of coverage from 2000 onwards.

In contrast, all CTs (n = 3,404) were dominated by use of adult SCs, primarily hematopoietic SCs, with some trials using umbilical cord blood derived SCs. There were an increasing number of trials using mesenchymal stem cells (MSCs) from 2007 (Figure 1B). The latter dominated our novel classification for CTs, most of which were in phases I and II (Figure 3). Figure 4 shows the global landscape for innovative CTs, with higher proportions using MSCs in Asia and the Middle East.

As of 2011, iPSCs had not entered CTs, and the best-known trial using hESCs, started in October 2010 by Geron for spinal cord injury, was halted in November 2011 after use in only four trial subjects. Three additional trials by Advanced Cell Technology (Santa Monica, CA, USA) began in July 2011: two using retinal pigment epithelial cells derived from hESCs for Stargardt's macular dystrophy and one for dry age-related macular degeneration [20]. CTs for cancers and graft-versus-host disease (GVHD), which composed most of the 'other diseases' category, dominated the SC CT landscape (Figure 2B). Nevertheless, there was a steady increase in CTs for cardiovascular disease beginning in 2001 and a small number of trials for neurological, ocular conditions and diabetes (note that the key word analysis for diabetes also captured CTs for associated morbidities such as diabetic foot). Our detailed examination of novel CTs was dominated by trials using MSCs and for conditions other than hematological cancers.

Our criteria for novel SC CTs resulted in the selection of 639 (19%) CTs that, if successful, would become the next generation of regenerative treatments. A total of 512 of these CTs clearly indicated study phase, and, of these, the majority were safety trials in phases I or II (Figure 4). Only 54 novel CTs were in phases II/III or III and 17 were in phases III/IV or IV (Table 1). The majority of clinical trials in phase III/IV (n = 11) used drug treatments to activate endothelial progenitor cells for therapeutic purposes in cardiovascular diseases (which may be associated with other diseases such as diabetes and ankylosing spondylitis). A further three advanced stage trials used stents to facilitate vascular repair by capturing endothelial progenitor cells in the stent and three trials addressed metabolic syndrome, critical limb ischemia, and breast deformities following a lumpectomy.

While the public and policymakers expect therapies for neurological conditions, there was limited activity in what may be considered novel application of SC therapies. Among the novel CTs, only 69 trials addressed neurological conditions (Table 2) and only 1 in our dataset, the Geron (Menlo Park, CA, USA) trial, used hESCs. The majority of trials used hematopoietic stem cells or MSCs. For example, clinical trials for multiple sclerosis used autologous transplantation of MSCs or hematopoietic SCs following chemotherapy to achieve immunomodulation and facilitate regeneration of the central nervous system. Two trials, being conducted in India, for Parkinson's disease used MSCs.

The majority of novel trials were publicly funded; however, some industry involvement was evident. Industry trials targeted both orphan and common diseases. Three phase I trials by StemCells Inc. (Newark, CA, USA) employed cultured allogeneic neural SCs (HuCNS-SC) to treat Pelizaeus-Merzbacher disease and neuronal ceroid lipfuscinosis. Ischemic stroke was the focus of significant industry involvement with 6 of 18 CTs sponsored by Stem Cell Therapeutics (Toronto, Canada), ReNeuron (Guildford, UK), and Stempeutics Research (Bangalore, India). The phase I trial by ReNeuron used a manufactured neural SC line (CTX cells) while the four Stem Cell Therapeutics trials attempted to use a proprietary drug (NTx-265) to promote neural regeneration. However, news reports indicated the lack of demonstrable effect in the latter CTs [21].

Of all the other diseases highlighted in the media, 208 CTs were for cardiovascular disease: 75% targeted the heart and 25% the peripheral vascular system. Almost all of these trials involved MSCs, mobilization of hematopoietic SCs, other purified bone marrow-derived SCs, or endothelial progenitor cells for regenerative and supportive purposes. There were 27 novel trials for diabetes mellitus, and these CTs attempted to treat both Type I and Type II diabetes using hematopoietic and MSCs from adipose tissue, umbilical cord blood, and bone marrow. Private sector companies were involved as collaborators or sponsors in eight of the trials, including Osiris Therapeutics (Columbia, MD, USA), Adistem (Hong Kong, China), Transition Therapeutics (Toronto, Canada), Cellonis Biotechnology (Beijing, China), and Genzyme (Cambridge, MA, USA).

Stem cell biology is highly complex and variable; using cells as therapeutic agents bears few parallels to the use of small molecule drugs. This creates significant challenges for regulators developing the testing necessary to assure the quality of novel technologies and therapies, for businesses that require reproducible large scale production capacity, and for the health systems that need to find the funds for what will likely be personalized and expensive therapies. As an example, there are valid concerns about manufacturing consistency of cell cultures and the genetic stability of cell lines. Both hESCs and iPSCs have been shown to experience genetic and phenotypic drift in long-term in vitro culture. There are further concerns about iPSCs due to the degree to which differentiated adult cells must be manipulated to become pluripotent [26, 27]. Given that there is a potential for late tumor formation resulting from inadvertent mutagenesis in the creation of SC products, prolonged monitoring of animals used in preclinical studies and long-term monitoring of CT participants will be required. SCs that are manipulated or manufactured or used in a non-homologous manner [28] require more stringent oversight to ensure the safety of the recipient.

Regulators are still developing systems specifically for complex cell therapies. The conventional preclinical regulatory studies for absorption, distribution, metabolism, excretion, and toxicity of drugs cannot be applied to cell-based therapies, in part because of an inability to track differentiation and migration of transplanted cells [15]. In contrast to drugs, SC based therapies may result in long-term engraftment, with amplification of the administered dose of SC and further growth and differentiation of the SC progeny. Regulators, in consultation with the SC research community, are attempting to develop a proportionate response which will balance patient safety yet will not stifle innovation in this field.

Developing or modifying regulatory pathways takes time, trust, and communication between regulators and the research community, all of which are undermined by hype [15, 28] or public representations of SC research that do not reflect its current state and the realistic developmental timelines for specific diseases or conditions, as documented by this study. Even with cooperation between SC researchers and regulators, oversight may prove difficult in a translational environment that has gone global, as demonstrated in Figure 3. Significantly, our results show that CTs testing novel SC approaches are occurring beyond the borders of North America and Europe. Indeed, while there are SC trials for novel application increasingly being registered from sites in India, China, and Iran, and to a lesser extent in Brazil, Mexico and Malaysia, these may use SC with only rudimentary characterization or for situations with as yet poorly understood biological rationale; the very types of trials concerning to the ISSCR. Ethical issues and conflicts of interests are more likely to arise where inadequately developed regulatory and consent frameworks exist. This will be of special concern in an arena such as the complex regulatory environment for cell therapies that necessitate long-term monitoring and follow-up for participants.

Media reporting of therapeutic benefits, especially for neurological conditions, may exacerbate the 'therapeutic misconception' in all countries [29]. While our analysis demonstrates that most SC CTs are in phases I or II, participants may conflate such early stage safety trials with therapeutic benefit. This mistaken therapeutic belief, which undermines informed consent, may arise from poorly structured consent forms or from the descriptions of CTs in trial registries [30]. In reality, most therapies fail in phase I and II, and those that are proven both safe and effective take between US$200 million and US$1 billion in investment over a span of 10 to 15 years before they are approved and adopted by the medical community [31].

A further challenge highlighted in our results of public versus private support of trials is that commercial enterprise has been slow to move into the cell therapy arena, primarily because of the lack of easily defined business models [32]. Cell therapies are not drugs and require the development of new manufacturing and distribution systems. While a few allogeneic products with central manufacturing have been developed, most notably Carticel from Genzyme and Apligraf from Organogenesis, the closest model to many of the current CTs is hematopoietic SC transplantation (HSCT). HSCT is a high-cost, highly specialized treatment, requiring significant infrastructure and an interdisciplinary network of healthcare professionals [33, 34]. Its global use is highly correlated with gross national income, healthcare expenditure and the availability of transplant teams. Given these system costs, the bar for therapeutic value, which compares the benefits of SC therapies with existing treatments, will be high. The model also belies the anticipated health system savings promised by proponents of SC research. These financial pressures, such as the scientific and regulatory issues, are absent from media reports and are hidden from public discourse, contributing to the disconnected state of current expectations.