To the Editor:

In their recent Perspective, Bianco et al. indicated that uncertainties regarding the nature, identity, function, mode of isolation and handling of mesenchymal stem cells (MSCs) have a major impact on their envisioned therapeutic use1. Although there is merit in this work, we would like to highlight some limitations in the definition of therapeutic reagents and debate the basic principles that characterize experimental medicine.

The authors clearly summarize the evidence underlying the identification of specific skeletal self-renewing stem cells and progenitors and indicate that these are system defined and are therefore both structurally and functionally different from nonprogenitor connective tissue cells—sometimes referred to as MSCs—that can be found in almost every organ. We fully endorse the need to keep these two cellular entities separate to avoid confusion between stem and progenitor functions and the regulation of tissue homeostasis. This distinction is fundamental in understanding the rationale behind the current therapeutic applications of mesenchymal 'stromal' cells. The definition of stromal cells has undergone a paradigm shift in recent years. In addition to the traditional view of stromal cells supporting and organizing a parenchymal framework, numerous studies have unveiled their important role in modulating tissue inflammation and, as a consequence, promoting tissue repair. Such properties are in principle independent of any stem or progenitor activity, although the heterogeneity in their composition may imply some degree of overlap2. These immunomodulatory features—rather than the ability to differentiate into multiple lineages—have fuelled attempts to test the therapeutic potential of these cells in immune-mediated disorders. In this context, we believe that at least some of these attempts were justified.

A further issue that we feel deserves correction is the notion that stromal cell therapies are derivatives of alchemy. Such a position ignores some of the methodologies that have been used in experimental medicine to successfully develop new treatments. Initial laboratory studies have suggested that adherent bone marrow–derived and ex vivo–expanded stromal cells suppress lymphocyte proliferation in vitro3, leading to the hypothesis that these cells might exert similar effects in vivo. On the basis of these experimental findings, a single patient with steroid-resistant acute graft-versus-host disease was treated with stromal cells, and the impressive clinical response that was observed4 formed the basis for a range of clinical trials in this area5. At the time of clinical introduction, in vivo evidence of the efficacy of this type of treatment in experimental animals was lacking and only became available later. From the clinical perspective, attempts to design proper studies should include better standardization of the cellular production process and the development of potency assays. However, clinical studies remain valid even if the composition of the therapeutic product and the putative underlying mechanisms of efficacy remain to be defined.

Although animal experiments may provide a proof of principle for clinical studies, their predictive value with respect to safety and efficacy is limited6. Despite an evidence-based clinical practice, many innovations in medicine follow an empirical rather than scientific approach. A variety of effective treatments have been introduced in the absence of a solid scientific basis or mechanistic understanding. These clinical developments have been focused primarily on outcome and not on the science or underlying mechanisms. Once successful, these studies have provided unique opportunities for basic science to provide an understanding of the biology and mechanisms of treatment efficacy. The introduction of inoculation against smallpox by Chinese Taoist alchemists in the tenth century was entirely empirical but was successful without even a vague notion of the immune system that mediated protection against disease. The most informative example in recent times is perhaps hematopoietic stem cell transplantation. Its success did not involve any understanding of the composition of the bone marrow graft, but the consequent combined efforts of laboratory research and clinical studies were fundamental for the definition and isolation of hematopoietic stem and progenitor cells, dendritic cells, natural killer cells and T and B cells, as well as their subsets.

There is little doubt that the design of clinical studies in the field of stromal cell therapies can be improved. These studies should target a well-selected group of patients, and appropriate follow up of clinical and laboratory parameters is crucial. Studies should perhaps not focus exclusively on clinical outcome but should also examine the mechanisms and biomarkers behind treatment efficacy. In this process, basic research will be invaluable, but it would be foolish to negate a clinically efficacious reagent because its biological function is not fully understood. The further advancement of this field will provide new opportunities for productive interactions between scientists and clinicians.