Since the early 1970s, hematologists have been using hematopoietic stem cell transplantation (HSCT) to cure patients with some forms of leukemia, even though the cause of the disease is still not fully understood. Currently many different autologous (from self) and allogeneic (from another person, usually a sibling) transplant protocols are available for patients with malignant and nonmalignant disorders. There is also an international list of donors for patients without an HLA-matched family member, but such so called matched unrelated donor transplants have a higher risk of reacting against the recipient's tissues, called graft-versus-host disease (GVHD). Autologous protocols are less toxic because GVHD does not occur, but have the disadvantage that the curative graft-versus-tumor effect is also absent. Obviously for inherited disorders such as sickle cell disease, only an allogeneic graft would be effective.

Ten years ago it was proposed that a similar HSCT approach could be applied to severe refractory autoimmune disease (AD) (1, 2) based on supportive animal model data and coincidental observations in humans undergoing HSCT for malignancy. The concept was that although drugs such as cyclophosphamide were effective in severe AD, the dose was limited by the inherent bone marrow toxicity. By using hematopoietic stem cells already collected and stored to rescue the marrow, one could exceed this threshold.

In addition, there is evidence from animal studies that when the ablated immune system returns (immune reconstitution) the animal may be tolerant to antigens that previously had caused the autoimmunity. This gave hope that a “once only” treatment may suffice for long periods of remission, because it is likely that antigens inciting autoimmunity in humans remain ubiquitous. However, not all animal models show tolerance induction, and as the clinical program proceeded, only approximately one-third of the treated patients achieved such a remission. The reasons are not apparent, and mechanistic studies are part of either current or planned prospective controlled trials (see below).

Currently ∼1,000 patients have received an HSCT (predominately autologous) as treatment for autoimmune disease, and the data from phase I–II studies have been sufficiently encouraging such that phase III randomized studies are proceeding in Europe and the US for systemic lupus erythematosus, systemic sclerosis, and multiple sclerosis. These data have recently been fully reviewed and published and are partly summarized in Table 1 (3).

Table 1. Number of cases in the autoimmune diseases hematopoietic stem cell transplantation database of the European Group for Blood and Marrow Transplantation/European League Against Rheumatism as of November, 2005
Neurologic disorders
 Multiple sclerosis204
Rheumatologic disorders
 Systemic sclerosis120
 Rheumatoid arthritis85
 Juvenile idiopathic arthritis64
 Systemic lupus erythematosus77
 Mixed connective tissue disease4
 Behçet's disease8
 Psoriatic arthritis2
 Ankylosing spondylitis2
 Sjögren's syndrome1
 Wegener's granulomatosis6
Hematologic immuncytopenias
 Immune thrombopenia16
 Pure red cell aplasia7
 Autoimmune hemolytic anemia11
 Evans syndrome (immune thrombocytopenia9
  and hemolytic anemia) 
Gastrointestinal disorders
 Inflammatory bowel disease5

Initial transplant-related mortality was 9%, and was higher in some diseases such as systemic sclerosis probably due to a generally poor condition of patients at the time of transplant. Although GVHD is not a risk with autologous HSCT, the procedure is still aggressive and associated with significant morbidity and mortality. Early toxicity is related to direct organ damage either from the agents used (e.g., cyclophosphamide cardiotoxicity) or infection and bleeding during the 10–12 days of aplasia following the immunosuppressive conditioning period. Late toxicity relates to developing malignancy such as leukemia due to the chemotherapy and/or radiation exposure. So far, the toxicity seen has been consistent with the known adverse events associated with conventional indications of HSCT.

Since adopting stricter selection criteria based on the phase I–II experience, transplant-related mortality is now considerably lower; in the ongoing Autologous Stem Cell Transplantation International Scleroderma trial (data available at URL: there has been no transplant-related mortality among the 65 patients randomized to either arm of the trial.

Recent data also suggest that clinical remission is maintained despite full immune reconstitution following autologous HSCT for multiple sclerosis, reinforcing the concept that a resetting of immune tolerance rather than full ablation of autoimmunity is possible (4). In other words, although not all signs of autoimmunity are eradicated (e.g., autoantibodies) a drug-free durable clinical remission is obtained. The hematopoietic stem cells in this context are probably not directly therapeutic, but act as a rescue for the ablated bone marrow during the period of reconstitution or recovery following the conditioning chemotherapy.

During this period of close collaboration between rheumatologists and hematologists, the emerging potential of mesenchymal stem cells (MSCs) in the pathophysiology and treatment of autoimmune diseases has been the focus of interdisciplinary collaboration.

Mesenchymal stem cells

  1. Top of page
  2. Mesenchymal stem cells
  3. Future application of MSCs in human autoimmune disease
  4. Conclusion

MSCs are multipotent stromal cells of the mesodermal lineage, which appear in a wave in the fetal blood between gestation weeks 7–12 and populate the stroma of hematopoietic and other tissues (Figure 1). When discussing stem cells, it is important to distinguish between embryonal stem cells that may be totipotent (capable of producing a complete individual) and pluripotent (capable of producing any tissue or organ). In adults, most stem cells are multipotent (capable of producing many lineages of a parent tissue; see glossary, Appendix A). The hematopoietic stem cell is the best-studied example.

thumbnail image

Figure 1. Stem cell development (Courtesy of Dr. Felipe Prosper Cardoso). ESC = embryonic stem cells.

Download figure to PowerPoint

Although there is speculation and some data to suggest the presence of pluripotent stem cell reservoirs in adults, it is likely that most adult stem cells have a multipotent capacity in vivo. In addition, despite demonstrated transdifferentiation of adult stem cells, i.e., shifting from one lineage to another, this effect is likely to be of little functional significance in vivo.

In adults, cells similar to the fetal MSCs are found in bone marrow, fat, muscle, and synovium. Due to the cell's scarcity in vivo and the current inability to prospectively isolate them, most of our knowledge comes from MSCs generated in vitro. In the absence of an agreed standardized marker, MSCs are typically defined as cells with a fibroblast morphology that are adherent to plastic. MSCs do not express CD14 (a monocyte marker), CD34 (a hematopoietic stem cell and early progenitor marker), or the common leukocyte antigen CD45. By flow cytometry, the cells stain positive for a number of molecules, including CD73, CD105, and CD90. Some of the surface epitopes on MSCs are related to cell trafficking (integrin and adhesion molecules) and may account in part for the ability of the MSCs to home to distressed tissues. The hallmark of MSCs is their trilineage potential in vitro, i.e., the ability to differentiate into bone, cartilage, and fat.

Whether or not MSCs are true stem cells remains to be established. The capacity to establish and maintain a complete tissue compartment has not yet been demonstrated for MSCs. For this reason the International Society of Cellular Therapy recently proposed the name multipotent mesenchymal stromal cell. However, MSCs are clearly multipotent in vitro and in vivo (5, 6).

The role of MSCs in adults is not fully clear, and most of our knowledge comes from in vitro studies, echoing the sentiments of Robey, “we have learned to recognize stem cells, not necessarily from by what they do in their dependent organism, but rather by what we can make them do in the laboratory” (7).

In contrast to the relative lack of information concerning the in vivo role of MSCs in health and disease, there has been intense activity concerning the application of ex vivo expanded MSCs in various clinical settings. The most explored is tissue engineering, with programs aimed at expanding and differentiating MSCs into bone, cartilage, and muscle, especially myocardium after infarct. In addition, MSCs are excellent vehicles for gene therapy, because they maintain expression of tranfected genes for up to 40 divisions. These aspects of MSCs are beyond the scope of this article, and will not be reviewed.

Recently, the ability of MSCs to exert an antiproliferative, immunomodulatory, and antiinflammatory effect has focused attention on them as potential therapeutic agents in various clinical settings including autoimmune diseases. An added attraction is their capacity for ex vivo expansion up to a billion-fold without loss of their multipotent properties, and their apparent immune privilege, i.e., they neither initiate nor are subject to immune reaction across allogeneic barriers.

The first evidence of the immunomodulatory properties of MSCs emerged in the late 1990s. Addition of MSCs to allogeneic mixed lymphocyte cultures (MLCs) profoundly reduced lymphocyte proliferation. The suppression was independent of HLA, and occurred when the added MSCs were autologous to the stimulatory or the responder lymphocytes or derived from a third party (8–11). The suppressive effect increases if the MSCs are exposed to interferon-γ (IFNγ) prior to the MLCs. This finding may be surprising since IFNγ induces HLA class II expression. It is not simply the result of apoptosis or T cell anergy, since lymphocyte proliferation may be restored if the MSCs are removed and the lymphocytes restimulated in a secondary MLC (12). The factor(s) mediating suppression by human, but not murine MSC appears to be soluble as proliferation is still reduced if MSC and lymphocytes are separated in a transwell system. Several factors have been proposed to mediate the suppressive effect, including transforming growth factor beta, hepatocyte growth factor, prostaglandin E2, and indoleamine 2,3-dioxygenase–mediated tryptophan depletion. Similar suppression of a range of proliferative T cell and B cell responses to various agents has been demonstrated, including phytohemagglutinin and monoclonal stimulating antibody CD3 for T cells (9), and interleukin-4 combined with CD40 ligand for B cells (13).

Animal models of autoimmune disease were tested with some contradictory results. Models of multiple sclerosis in mice responded favorably (14), whereas a mouse arthritis model did not (15). The latter was attributed to excessive tumor factor necrosis α in the environment, and clearly more work is needed.

Other nonimmunologic models of tissue injury and distress have been studied with positive outcomes. A rat model of renal ischemic/reperfusion injury was protected with syngeneic (same strain of inbred animal) MSC intra-arterial infusion, an effect not obtained with similar numbers of syngeneic fibroblast cells (16). Hepatologists have protected the liver from necrosis induced by the chemical carbontetrachloride (17), and in the lung, injury caused by the cytotoxic agent bleomycin was attenuated by MSCs (18). This effect is particularly relevant for application to human autoimmune diseases in which inflammation-induced fibrosis plays a role. In all cases, the timing of the MSC protection was critical (i.e., early and not late) suggesting certain “tipping points” after which processes become chronic.

In human studies, more than 100 patients have received MSCs (19). In most cases, the MSCs were derived from an allogeneic HLA-matched donor and the cells given in the context of HSCT for malignancy or inborn errors of metabolism. Infusion of MSCs was well tolerated and no side effects were noted. Although acute toxicity appears low, the identification of long-term unwanted effects will require a prospective registry of treated patients. Although one murine model of melanoma suggested an increased rate of metastasis (20), an increased risk of tumor formation has not been reported in humans to date. In fact, the fate of MSCs once infused remains elusive to some extent. After infusion in humans, MSCs can only be detected in the circulation during the first few hours. MSCs can be detected in the lung, the gastrointestinal tract, and other organs such as the kidney, skin, and thymus in experimental animals. However, only 0.5–2.7% of the infused MSCs are traceable, and the fate of the remainder is unknown.

Protection from an immunologic reaction by MSC infusion was reported in a 9-year-old boy who received a matched unrelated donor HSCT for leukemia (21). Severe acute steroid-resistant GVHD of the gut and liver was reversed by infusion of haploidentical MSCs derived from the patient's mother. Studies further evaluating the potential of MSCs both to prevent and to treat steroid-resistant GVHD and to reduce tissue damage post myocardial infarction are ongoing. Other studies aim to evaluate whether MSC infusion enhances hematopoietic recovery when the stem cell dose is limiting.

Future application of MSCs in human autoimmune disease

  1. Top of page
  2. Mesenchymal stem cells
  3. Future application of MSCs in human autoimmune disease
  4. Conclusion

Currently, further in vitro studies are confirming the immunosuppressive potential of autologous and allogeneic MSCs from patients with autoimmune disease, and further animal models of autoimmune disease are being examined. In the meantime, an internationally coordinated effort involving the International Society for Cellular Therapy and the Stem Cell Subcommittee of the European Group for Blood and Marrow Transplantation is underway to standardize MSC definition and expansion according to good manufacturing practice conditions.

If the toxicity and safety data in animals and humans confirm the current favorable impression, it is hoped that clinical trials performed under good clinical practice and good manufacturing practice conditions in severe active autoimmune disease may become a reality within the next few years.


  1. Top of page
  2. Mesenchymal stem cells
  3. Future application of MSCs in human autoimmune disease
  4. Conclusion

Autologous HSCT for severe autoimmune disease has been shown in phase I–II studies to be effective in many individuals, and is currently being evaluated in several international prospective, randomized trials. Cellular therapy also offers the unique potential of utilizing the complex paracrine potential of adult multipotent stem cells capable of homing to and modulating a local environmental event. This is already a reality with hematopoietic stem cell transplantation that acts as support following immunoablative therapy, and may also eventuate with MSCs. The advantage of the latter is their vast ex vivo expansion capacity, apparent low toxicity, immune privileged status, and their capacity to home to distressed tissue and exert an antiinflammatory and antiproliferative effect.

Because of the many disparate groups involved in MSC biology and application, there is a need for internationally standardized definitions, good manufacturing practice expansion protocols, and clinical trial design. In addition, a database for assessing long-term toxicity, including tumor surveillance, should be initiated across all fields.


  1. Top of page
  2. Mesenchymal stem cells
  3. Future application of MSCs in human autoimmune disease
  4. Conclusion


  1. Top of page
  2. Mesenchymal stem cells
  3. Future application of MSCs in human autoimmune disease
  4. Conclusion

Stem cell. On division, one daughter cell has the potential to form a complete tissue compartment, and the other retains its stem property.

Embryonic stem cells. Isolated from the inner cell mass of the blastocyst (present at the implantation stage of embryonic development several days after fertilization).

Totipotent cells. In mammals, totipotent cells have the potential to become either any type in the adult body or any cell of the extraembryonic membranes (e.g., placenta). The only totipotent cells are the fertilized egg and the first 4 or so cells produced by its division. This is evidenced in mammals by the occurrence of identical twins, triplets, etc. The nomenclature totipotent stem cells is a misnomer: these cells fail to meet the second criterion (i.e., on division, one daughter cell does not retain its stem properties).

Pluripotent stem cells. These are true stem cells, with the potential to make any differentiated cell in the body. However they are unable to contribute to making the extraembryonic membranes derived from the trophoblast, and therefore cannot make a complete individual.

Multipotent stem cells. Also true stem cells, but they can only differentiate into a limited number of types (i.e., the bone marrow contains multipotent stem cells that give rise to all the cells of the blood, but not to other types of cells). Multipotent stem cells are found in adult animals; perhaps most organs in the body (e.g., brain, liver) contain them where they can replace dead or damaged cells or, under some conditions, malignancies.