Review article: stem cell therapies for inflammatory bowel disease – efficacy and safety


  • This commissioned review article was subject to full peer-review.

Dr J. Panés, Department of Gastroenterology, Hospital Clínic of Barcelona, Villarroel 170, Barcelona 08036, Spain.


Aliment Pharmacol Ther 2010; 32: 939–952


Background  Drugs available for the treatment of inflammatory bowel disease fail to induce and maintain remission in a significant number of patients.

Aim  To assess the value of stem cell therapies for treatment of inflammatory bowel disease based on published studies.

Methods  Publications were identified through a MEDLINE search using the Medical Subject Heading terms: inflammatory bowel diseases, or Crohn’s disease, or ulcerative colitis, and stem cell, or stromal cell or transplant.

Results  Haematopoietic stem cell therapy as a primary treatment for inflammatory bowel disease was originally supported by animal experiments, and by remissions in patients undergoing transplant for haematological disorders. Later, transplantation specifically performed for patients with refractory Crohn’s disease showed long-lasting clinical remission and healing of inflammatory intestinal lesions. Use of autologous nonmyeloablative regimens and concentration of the procedures in centres with large experience are key in reducing treatment-related mortality. Initial trials of mesenchymal stem cell therapy with local injection in Crohn’s perianal fistulas had positive results.

Conclusions  Autologous haematopoietic stem cell transplant changes the natural course of Crohn’s disease, and may be a therapeutic option in patients with refractory disease if surgery is not feasible due to disease location or extension.


Standard medical therapy for inflammatory bowel diseases (IBD) is directed against the inflammatory and immune processes that are known to play an important role in the disease process. Medical therapy is of variable success in ameliorating cardinal symptoms of the disease (diarrhoea, abdominal pain), in treating extraintestinal manifestations (fatigue, anorexia, fever, weight loss, arthralgias, and skin, eye, liver, and kidney manifestations), and in preventing complications (stricture, fistula, abscess). Currently, therapy is most often implemented in a stepwise fashion, progressing through aminosalicylates [sulfasalazine, mesalazine (mesalamine)], antibiotics (metronidazole, ciprofloxacin), corticosteroids, immunosuppressive medications including tioguanine (thioguanine) compounds (mercaptopurine, azathioprine), methotrexate, and ciclosporin, and finally anti-TNF drugs. This common approach is predicated on the addition of more potent medications to agents that are believed to be safer but that may also be less effective. Despite an optimized use of this treatment paradigm, the need for intestinal resection in Crohn’s disease has remained stable.1 Primary and secondary failure to respond to approved therapies and, in some cases, inability to provide a surgical solution to a particular patient due to extension and/or location of lesions represent unmet needs in the treatment of IBD.

Two streams of research, experimental and clinical, are the origin of the increasing utilization of stem cell therapies for severe immune-mediated diseases (IMIDs) including IBD. The considerable excitement surrounding the stem cell field was initially based on the unique biological properties of these cells and their capacity to self-renew and regenerate tissue and organ systems; later, the immunomodulatory ability of stem cell therapy has become also apparent. The definition of stem cell conventionally indicates the haematopoietic ones, with reference to the myeloid and lymphoid lineages. However, a distinct lineage is now known to consist of mesenchymal stromal cells. Similar to haematopoietic stem cells (HSC), mesenchymal stem (stromal) cells (MSC) are rare in the bone marrow, representing 1 in 10 000 nucleated cells.2 The ontogenic relationship between HSCs and MSCs remains unclear.

Haematopoietic stem cells

Stem cells are undifferentiated cells that have the capability of both self-renewal and differentiation into mature specialized cells through replication. Two types of stem cells have been recognized, embryonic stem cells and adult stem cells. Human embryonic stem cells are isolated from a 50- to 150-cell, 4- to 5-day-old blastocyst. Embryonic stem cells can generate every specialized cell in the human body and, while capable of indefinite ex vivo proliferation, in vivo exist only during embryogenesis. Adult stem cells exist in all types of tissues throughout the body and function as a reservoir to replace damaged or ageing cells. Under physiological conditions, adult stem cells are mostly restricted in their differentiation to cell lineages of the organ system in which they are located.

Embryonic stem cells have a great versatility and were envisioned as a promising option for cell therapy, but compared with adult stem cells are currently difficult to control due to their tendency to form tumours containing all types of tissue (teratomas). In contrast, adult stem cells do not have this tendency, and follow traditional lineage-specific differentiation patterns, fulfilling their physiological homologous function of replacing normal turnover, ageing, or damaged tissues. Due to the inability to harvest or expand stem cells from most adult organs efficiently and safely (e.g. liver, gastrointestinal tract, heart, brain), a majority of human stem cell trials have focused on clinical applications for HSCs, MSCs, or both, which can be easily obtained in clinically sufficient numbers from peripheral blood, bone marrow, or umbilical cord blood and placenta. Human haematopoietic progenitor cells are identified by glycoproteins CD34+, CD133+, or both. Most human marrow or blood CD34+ or CD133+ cells are committed progenitors, and only a minority are lifelong repopulating stem cells. A CD34- or CD133-enriched HSC product will reconstitute lifelong haematopoiesis and may be easily purified from the marrow or peripheral blood.3

Mesenchymal stem cells

When cells from a bone marrow aspirate are cultured in plastic flasks, haematopoietic cells and HSCs do not adhere to the plastic and are removed with change of media. The remaining plastic-adherent cells were originally termed colony-forming unit fibroblasts because they formed fibroblast-like colonies ex vivo. Subsequently, these adherent cells have been termed MSCs, an abbreviation for both mesenchymal stromal cells and mesenchymal stem cells. The former refers to the ability of MSCs to contribute to the structural matrix of bone marrow and to support haematopoiesis; the latter describes the ability of MSCs to differentiate under various ex vivo culture conditions into different mesenchymal-derived cells.4 MSCs may be isolated from bone marrow, skeletal muscle, adipose tissue, synovial membranes and other connective tissues of human adults as well as cord blood and placenta.

Mesenchymal stem cells have no unique phenotypic marker. The minimal criteria by the International Society of Cellular Therapy to define MSCs are (i) plastic-adherent in culture; (ii) expression of CD105, CD73, and CD90; (iii) lack of expression of haematopoietic markers such as CD45, CD34, CD14, CD11b, CD19, CD79a and HLA-DR; and (iv) able to differentiate into osteoblasts, adipocytes and chondrocytes.3 The ratio of MSCs to marrow mononuclear cells is estimated to be only 10–100 MSCs per million marrow cells.5 Despite relatively low numbers, a 2-mL aspirate of bone marrow can be expanded 500-fold ex vivo to 12 billion to 35 billion MSCs within 3 weeks.5

Depending upon the environment, MSCs give rise to many cell lineages including epithelial cells, astrocytes, osteoblasts, chondrocytes, adipocytes and muscle, promoting regeneration of damaged tissue in vivo.4In vitro, MSCs have vast proliferative potential, can clonally regenerate and can give rise to differentiated progeny. They also exhibit antiproliferative and anti-inflammatory properties in vitro and in vivo, making them candidates for treatment of IMIDs.6 Regardless of whether or not MSCs are true stem cells, clinical benefit from MSCs may not require sustained engraftment of large numbers of cells or differentiation into specific tissues. It is possible that a therapeutic benefit can be obtained by local paracrine production of growth factors and by the provision of temporary antiproliferative and immunomodulatory properties.7

Until recently, it was assumed that MSCs enjoyed ‘immune privilege’ in allogeneic settings, neither exerting nor being subject to immunological reaction. However, recent data suggest that in an immunocompetent host, allogeneic MSCs will be eliminated. In vitro studies indicate that MSCs possess immunosuppressive properties. Rodent, baboon and human MSCs suppress T and B-cell lymphocyte proliferation in mixed lymphocyte cultures (MLCs) or induced by mitogens and antibodies in a dose-dependent manner. The suppression is MHC independent, and in human cell cultures, the magnitude of suppression is not reduced when the MSCs are separated from the lymphocytes in transwells, indicating that cell–cell contact is not required.

The mechanism(s) underlying the immunosuppressive effect of MSCs remain(s) to be fully clarified with sometimes conflicting data, probably reflecting variable definitions and experimental conditions. MSCs exert important immunomodulatory functions both in vitro and in experimental in vivo models. Bone marrow-derived MSCs can suppress cytotoxic and IFN-γ T-cell proliferation in a cell contact-dependent mechanism through engagement of the inhibitory molecule PD-1. Also, MSCs have been shown to trigger generation of Treg, as well as to induce antigen-dependent Treg proliferation directly.4 On the other hand, MSCs can modulate the innate immune response which is known to play a key role in the pathogenesis of CD. In particular, MSCs inhibit dendritic cell differentiation and maturation, act on resting NK cells and decrease the respiratory burst and apoptosis of neutrophils. Interestingly, MSCs obtained from adipose tissue (AMSCs) have been shown to exert immunosuppressive effects in vitro and inhibit peripheral blood mononuclear cell proliferation and IFN-γ production while increasing IL-10 secretion.8

Data demonstrating the fate of transplanted MSCs in vivo are scarce. In rats, radiolabelling experiments following intra-arterial and intravenous infusion of MSCs show localization mainly in the lungs and secondarily in the liver and other organs. Estimated levels of engraftment in these tissues ranged from 0.1% to 2.7% and were similar in the autologous and allogeneic settings.9 MSCs homing to tumours is of concern. Human MSCs have been shown to localize to a murine xenogenic breast cancer via monocyte chemotactic protein-1 (MCP-1)10 in a severe combined immunodeficiency (SCID) mouse model. Although this suggests that MSCs may provide a potential therapeutic delivery system for cancer therapy, it may pose long-term safety issues in autoimmune disease treatment.

Lessons from haematopoietic stem cell studies as a treatment for IMIDs

The concept and practice of HSC transplant (HSCT) as a primary treatment for IMIDs began in the late 1990s, and was originally supported by decades of animal experiments and by unanticipated remissions of inflammatory diseases observed in patients undergoing HSCT for haematological disorders or cancer. In 1995, an international committee was established that developed guidelines on entry criteria and transplant protocols for severe autoimmune diseases.11 Furthermore, a database was created to collect clinical data that would enable monitoring of feasibility, toxicity and efficacy of the different treatment protocols. Analysis of the pooled data has yielded relevant information on trends with regard to mortality, type of protocols used and diseases targeted.

Haematopoietic stem cell transplant includes conditioning (high dose chemotherapy, antilymphocyte antibodies and/or total body irradiation) followed by infusion of HSC. The HSC are usually mobilized from bone marrow to peripheral blood and harvested by apheresis. In some protocols of autologous transplantation for autoimmune diseases, the graft is depleted of T cells (e.g. selection of CD34 cells) to reduce the likelihood of reinfusing presumed disease-causing T cells.12 Allogeneic HSCT should theoretically be associated with a higher probability of cure than autologous HSCT, as the genetic background of immune cells would be changed, and there is a graft-vs.-host reaction that removes the presumed disease-causing T cells that could survive conditioning.13 However, due to graft-vs.-host disease, allogeneic transplantation is associated with higher transplant-related morbidity and mortality. For this reason, more autologous transplants have been performed for IMIDs. The rationale of autologous HSCT for autoimmune diseases is to immune reset, i.e. to generate new self-tolerant lymphocytes after chemotherapy-induced elimination of self- or auto-reactive lymphocytes (i.e. lymphoablation). Allogeneic HSCT is based on the rationale of both immune reset (similar to autologous HSCT) and of correcting the genetic predisposition to disease by reinfusing nondisease-prone HSCs from a normal donor.

Over the last two decades, HSCT has become increasingly safe due to reduced conditioning procedures, use of peripheral blood stem cells, and improvement in supportive measures and patient selection. This has lead to the application of HSCT to treat IMIDs refractory to available therapies. Treatment-related mortality for autologous HSCT of autoimmune diseases in the European Group for Blood and Marrow Transplantation registry is approximately 7%,11 and some trials have reported rates of up to 23%.14 Although mortality is reduced with greater experience and more careful patient selection, it has justifiably dampened enthusiasm for the field. Autologous HSCT for IMIDs may be performed with either myeloablative or nonmyeloablative regimens.15 Myeloablative regimens use cancer-specific treatments that destroy the entire marrow compartment, including marrow stem cells, resulting in irreversible and lethal marrow failure if HSCs are not reinfused. Nonmyeloablative regimens are designed specifically for autoimmune diseases, i.e. for lymphoablation without irreversible destruction of marrow stem cells. Following a nonmyeloablative regimen, haematopoietic recovery will occur without infusion of HSCs; however, autologous HSCs provide support and shorten the duration of chemotherapy-induced marrow suppression.15

The essential argument in favour of nonmyeloablative regimens is that treatment-related mortality needs to be very low for nonmalignant diseases, and nonmyeloablative regimens appear safer than myeloablative regimens.14 A second consideration in favour of nonmyeloablative regimens is that independent of using a myeloablative or nonmyeloablative regimen, disease relapse may occur; autologous HSCT for autoimmune diseases should not be viewed as a cure, but rather as changing the natural history of disease. A third point in favour of nonmyeloablative regimens is that IMIDs may undergo long-lasting spontaneous remissions despite an initial aggressive course, and no clinical reliable predictor has been identified for any of the IMIDs. It is debatable whether a subset of patients who may remit without HSCT should be exposed to myeloablative regimens, especially those including total body irradiation, which cause a relatively high incidence of a more lethal disease, i.e. myelodysplastic syndrome and leukaemia.15

Both early and late toxicity are a consequence of the regimen used for transplantation and of the increase in transplant-related mortality that occurs with increased intensity of the transplant regimen. Treatment-related mortality is less than 1% for nonmyeloablative, less than 2% for low-intensity myeloablative, and 13% for high-intensity myeloablative regimens.16 A number of reports combined data from patients treated with different conditioning regimens or from those with different diseases complicating interpretation, because toxicity is both regimen-specific and disease-specific. Although transplant regimen intensity may correlate with remission duration, it remains unclear if, at some point in dose intensity, a response plateau occurs independent of any further increase in regimen intensity. In this regard, it has been observed that in patients with systemic sclerosis, autologous HSCT results in remarkable reversal of skin tightness, improved joint flexibility and quality of life, and reversal of pulmonary alveolitis.17, 18 Two studies of nonmyeloablative regimens demonstrated 0% (0/10) and 4% (1/26) rates of treatment-related mortality,17, 18 respectively, while a study using a myeloablative approach including total body irradiation reported a rate of 23% (8/34).14 Importantly, both myeloablative14 and nonmyeloablative18 approaches reported an identical 5-year event-free survival of 64%.

Independent of whether myeloablative or nonmyeloablative regimens are used, another complication, late secondary autoimmune disorders, may arise from some agents used in the conditioning regimen. The initial standard nonmyeloablative regimen of cyclophosphamide and rabbit antithymocyte globulin (rATG) was well tolerated. A second-generation nonmyeloablative regimen used cyclophosphamide and a broader- and longer-acting agent, alemtuzumab, instead of rATG. Potential life-threatening secondary autoimmune cytopaenias, including idiopathic thrombocytopaenic purpura, autoimmune neutropaenia, and autoimmune haemolytic anaemia, occurred late (2–18 months) after transplantation in patients receiving regimens containing alemtuzumab.19 A third-generation nonmyeloablative regimen, termed ‘rituximab sandwich’, entails one dose of rituximab given before and after cyclophosphamide and rATG. To date, this regimen has been well tolerated.

Another factor with a notable influence on mortality is the experience of the centre in the use of HSCT for treatment of a particular IMID, as the transplant procedure may be associated with disease-specific complications that benefit from specific prevention and/or treatment. In a single experienced centre, nonmyeloablative autologous HSCT for patients with systemic lupus erythematosus, when performed as salvage therapy for treatment-refractory disease, resulted in marked serological, clinical and organ improvement, with 2% (1/50) treatment-related mortality and 50% probability of maintaining remission for 5 years.20 In comparison, a multicenter analysis of HSCT for systemic lupus erythematosus that included both myeloablative and nonmyeloablative regimens from 23 different centres reported a similar 55% 5-year disease-free survival, but treatment-related mortality was 13% (7/53).21

Timing of HSCT relative to disease duration and/or complications may be another key factor affecting treatment outcome. For multiple sclerosis, original HSCT studies were performed predominantly in patients with secondary progressive and, to a lesser extent, primary progressive disease. In this subset of patients, intense myeloablative regimens generally failed to improve neurological disability or to convincingly halt or change the rate of progressive neurological disability.16 Despite lack of clinical benefit in patients with progressive multiple sclerosis, magnetic resonance imaging evidence of inflammation was abrogated, while loss of brain volume continued for 2 years before subsiding.22 In retrospect, the predominant pathophysiology in primary and secondary progressive multiple sclerosis is axonal degeneration, which would not be expected to improve after autologous HSCT, a method that allows delivery of intense immune suppression. Learning from these studies, a trial of autologous HSCT for relapsing-remitting multiple sclerosis, which is an immune-mediated inflammatory disease, was performed with a safer nonmyeloablative regimen. There was no treatment-related mortality, no disease progression, and two-thirds of patients had significant improvement in neurological function.23 The relevance of timing of cell therapy may not only depend on the appearance of irreversible damage but also on the differential characteristics of immune activation in early and late disease, which has been shown also in CD.24

The lessons learned from multiple sclerosis – i.e. treat early while the disease is inflammatory and use nonmyeloablative regimens with low risk of treatment-related mortality – were applied to patients with type 1 diabetes by using a nonmyeloablative regimen and selecting patients within 6 weeks of diagnosis before complete loss of insulin-producing islet cells. Autologous nonmyeloablative HSCT resulted in insulin-free remission of type 1 diabetes in 20 of 23 patients, and 12 patients maintained this status for a mean of 31 months (range 14–52 months). There was no mortality.25

Recently, allogeneic HSCT using a sibling’s HSCs has also been reported for treatment of several IMIDs,26, 27 including CD (see below). As it changes genetic predisposition to disease, allogeneic HSCT is considered more likely to cure autoimmune diseases compared with autologous HSCT. Graft-vs.-host disease (GvHD), an often more lethal immune-mediated disease, is not an acceptable risk following allogeneic HSCT for autoimmune disorders rather than for malignancies and should be minimized by depletion of lymphocytes from the donor graft. Although yet unproven, some animal and limited human data26–28 suggest that an allogeneic graft-vs.-autoimmunity effect may occur without GvHD via use of a lymphocyte-depleted graft.

Haematopoietic stem cell transplant for inflammatory bowel disease

The possibility that HSCT might be an effective treatment for IBD arose initially from reports of improvements in IBD patients who had HSCT for other reasons, followed by case reports and case series of transplantation specifically performed for Crohn’s disease (CD) patients. Tables 1 and 2 summarize the publications on direct and indirect evidence of efficacy.

Table 1.   Myeloablative HSCT for treatment of cancer in patients with IBD
YearStudyPrimary conditionType HSCTIBDIBD complicationsIBD duration (years)AgeGender
Cell selectionConditioningMortalitySAEsOutcome
  1. CD, Crohn’s disease; F/M, female/male; GvHD, graft-vs.-host disease; HL, Hodgkin’s lymphoma; HSCT, haematopoietic stem cell transplantation; IBD, inflammatory bowel disease; MDS, myelodysplastic syndrome; NA, information not available; NHL, non-Hodgkin’s lymphoma; SAE, severe adverse event; UC, ulcerative colitis.

1993Drakos et al.29NHLAutologous1 CDFistula
22411/0NoBCNU 300 mg/m2,
Etoposide 600 mg/m2,
CTX 60 mg/m2,
ara-C 600 mg/m2,
melphalan 120 mg/m2
00Remission 6 months
1996Castro et al.33Breast cancerAutologous1 CDNA11N/A1/0N/ACTX 5.525 g/m2,
cisplatin 165 mg/m2,
BCNU 0.6 g/m2
0NARemission >2 years
1998Kashyap and Forman 34NHLAutologous1 CDPerianal disease
Obstruction Surgery
8210/1NoTBI 200cGy × 6,
VP-16 60 mg/kg,
CTX 100 mg/kg,
00Remission 7 years
1998Talbot et al.32LeukaemiaAllogeneic1 CDSurgery7350/1NoCTX 200 mg/kg
Busulfan 16 mg/kg
0GvHDRemission 8 years
1998Lopez-Cubero et al.30LeukaemiaAllogeneic6 CDFistulae 2
Obstruction 1
Surgery 2
(median 11)
27–460/6NoCTX 120 mg/kg
TBI 12–13.5 Gy
1 sepsis 97 days
1 MI 4.5 years
1 suicide 5.8 years
Acute GvHD 6
Chronic GvHD 4
Remission 4–15 years in 4, relapse 1.5 years in 1
2000Musso et al.35HLAutologous1 CDSurgery10300/1NoIdarubicin melphalan00Remission 3 years
2001Marti et al.38Breast cancerAutologous1 UCNone7571/0NoCTX
carboplatin thiotepa
00Relapse at 20 months
2002Soderholm et al.36LeukaemiaAutologous1 CDObstruction, Fistula, Surgery3571/0NoCTX 60 mg/kg ×2
TBI 6x 2Gy
0Sepsis KlebsiellaClinical & endoscopic Remission 5 years
2003Ditschkowski et al.31MDS 1/leukaemia 10Allogeneic4 UC 7 CDFistulae 3
Obstruction 2
Surgery 4
Median 10 years27–55
Median 41
6/5CD34 enriched in 4 casesMyeloablative various;
TBI in 7
1 pulmonary fungal infectionAcute GvHD 8
Pulmonary fungal infection 1
Remission in 10, (median follow-up 34 months)
2007Anuma-konda et al.37NHLAutologous1 CDNA16321/0NACHOP0NARemission, relapse after 8 years
Table 2.   Haematopoietic stem cell transplant as primary treatment for Crohn’s disease
YearStudyType HSCTnIBD complicationsCD duration (years)AgeGender
MobilizationCell selectionConditioningPrevious TmtsMaintenanceSAEsOutcome
  1. CD, Crohn’s disease; F/M, female/male; HSCT, haematopoietic stem cell transplantation; MTX, methotrexate; NA, information not available; SAE, severe adverse event.

2003Kreisel et al.44Autologous1Fistula
14360/1CTX 2 g/m2/day × 2 G-CSF 480 μg/dayYesCTX 200 mgAntibiotics steroids azathioprine anti-TNFNo treatment 9 months, Prednisone + MTX at month 90Clinical remission,
Mild endoscopic lesions at 9 months
2004Scimèet al.43Autologous1EIM2.5550/1CTX 2 g/m2 G-CSF 10 μg/kg/dayYesCTX 200 mg
equine ATG 90 mg
Azathioprine anti-TNF
No tmt0Clinical remission, mild endoscopy lesions 5 months
2005Oyama et al.41Autologous12Fistula 6 Obstruction 3 Perianal 4 Surgery 910 (1.5–20)27 (15–38)6/6CTX 2 g/m2 G-CSF 10 μg/kg/dayYesCTX 200 mg
equine ATG 90 mg
Thiopurine 12
Anti-TNF 12
1 Pred + MTX
11 no
1 Surgery stenosis month 511 Remission month 12
1 recurrence month 18
2008Cassinotti et al.42Autologous4Obstruction 1
Surgery 3
Perianal 3
GI bleeding 1
NA31 (26–45)1/3CTX 1.5 g/m2 G-CSF 10 μg/kg/dayNoCTX 200 mg
rabbit ATG 7.5 mg/kg
Thiopurine 4
Anti-TNF 4
Antibiotic 20Remission at 3 months
1 relapse mo4
Endoscopic remission month 12 2/2
2009Glocker et al.45Allogeneic1Fistula
990/1Alemtuzumab 1 mg/kg fludarabine 180 mg/m2 treosulfan 42 mg/m2 thiotepa 10 mg/kgMTX thalidomide anti–TNFNA0Remission 2 years

Effect of HSCT performed for cancer on IBD

In 1993, the first report was published on a 41-year-old female with CD who underwent autologous HSCT for non-Hodgkin’s lymphoma.29 The patient remained symptom-free during the 6-month reported period of follow-up. Since then, a total of 25 patients have been reported who underwent HSCT for cancer (Table 1).

A series from Seattle30 was reported in 1996 of six patients with CD who underwent allogeneic HSCT for leukaemia. Five of these patients had active CD at the time of transplant. All patients received cyclophosphamide and total body irradiation as conditioning therapy. Two patients had sclerosing cholangitis. One patient died of sepsis at 3 months post-transplant. The remaining five patients were in clinical remission more than 1 year post-transplant despite discontinuation of immunosuppression. Four of these five patients had sustained remission of CD at 54–183 months post-transplant. One patient had a relapse of CD 18 months after transplantation. Interestingly, the relapsed patient had mixed donor-host haematopoietic chimerism, whereas the four patients who achieved sustained remission were assumed to be complete chimeras as they had GvHD in the first 2 years after HSCT.

Another series from Germany31 reported in 2003 of seven patients with CD and four with UC who underwent allogeneic HSCT for haematological malignancy. Six of the 11 patients had active disease at the time of transplantation. All patients received cyclophosphamide as part of different conditioning regimens, and all patients except two received also total body irradiation. Ten patients went into remission following HSCT with a median follow-up time of 34 months. In the one patient with recurrent IBD symptoms after transplant, endoscopy and radiological examinations did not demonstrate any IBD related lesion. There was one transplant-related mortality reported at 10 months from an opportunistic infection. The usefulness of this study for the determination of whether allogeneic HCT may cure CD is limited by the fact that at the last follow-up, only two patients were off of immunosuppressive drugs.

Between 1996 and 2007, seven additional single case reports of patients suffering IBD who underwent HSCT for treatment of cancer were published. One patient with CD underwent allogeneic HSCT and had a sustained remission 8 years after transplant.32 Five patients with CD received an autologous HSCT; in four, the disease remained in remission during follow-up lasting 2–8 years, and one patient had a clinical relapse at 8 years.33–37 Finally, one patient with UC underwent autologous HSCT and had a relapse of the IBD 20 months after transplant.38

Overall, 22 of 25 patients have achieved clinical remission over a median follow-up of 20 months. Only two of these patients are on any ongoing medication for IBD suggesting that HSCT may result in long-lasting remission without any need for medications. There have been two transplant-related deaths among the 25 cases, both related to infectious complications, and other two cases died of unrelated causes: one myocardial infarction and one suicide.

Although these studies were not designed to investigate the effect of HSCT on the course of IBD, they support the notion that elimination of self- or auto-reactive lymphocytes (i.e. lymphoablation) and generation of new self-tolerant lymphocytes after chemotherapy (immune reset) affords long-lasting remissions in patients with IBD. The central role of the immune system in the pathogenesis of CD is further supported by another observation coming also from the HSCT field. A healthy patient undergoing allogeneic HSCT developed severe CD soon after transplant.39 Investigations showed that the transplanted stem cell had a pathogenetic mutation of the NOD2-gene, compatible with the notion that the CD developed when the normal immune response was disrupted by an anomalous handling of bacterial products by the transplanted immune cells. This body of evidence suggested that long-term remissions may be achieved by HSCT in CD and form the basis for prospective studies.

HSCT as a primary treatment for CD

In IBD, HSCT is currently being tested for the treatment of CD, given that ulcerative colitis has a relatively simple therapeutic option for refractory cases with surgery.

The first documentation of the efficacy of autologous HSCT for treatment of CD came from the Chicago group in 2003.40 The response to this therapy in the first four patients was excellent; all patients achieved clinical remission with no significant untoward event from the transplantation. These initial cases are also integrated in the final report of the phase I study41 including 12 patients with active moderate-severe CD (CDAI 250–358) refractory to conventional therapies including anti-TNF treatment. A majority of these patients had a history of surgery. Mobilization was with cyclophosphamide 2 g/m2 and G-CSF 10 μg/kg/day, and conditioning with cyclophosphamide 200 mg/kg and equine ATG 90 mg/kg. The peripheral blood stem cells were CD34+ enriched. Follow-up was at 6 and 12 months post-HSCT and then yearly and included small bowel radiography and colonoscopy at each visit. There was an early and sustained clinical remission (CDAI ≤ 150) in 11 of 12 patients after a median follow-up of 18.5 months. Radiographic and colonoscopy findings improved more gradually over months to years. One patient experienced CD relapse at 15 months post-transplant. All patients tolerated the HSCT well and there were no transplant-related deaths. One patient died of accidental cause at 37 months post-HSCT; there was no evidence of active CD at this time.

A second phase I-II study from Milan was published in 2008.42 The study included four patients of similar characteristics, all had failed immunosuppressant treatment and anti-TNF therapy and two of them had undergone multiple surgical resections. All patients had active disease (CDAI 258–404) at the time of transplantation. Mobilization was with cyclophosphamide 1.5 g/m2 and G-CSF 10 μg/kg/day, and conditioning with cyclophosphamide 200 mg/kg and rabbit ATG 7.5 mg/kg. No cell immunoselection was carried out in this study. Three months after transplant, all patients had achieved clinical remission, and complete endoscopic remission was achieved in 2/3 patients. Three of the four patients had sustained remission after a median follow-up of 16.5 months. No mortality was observed in the two series of patients.

In the first study coming from the group in Chicago,41 the authors observed improvement of the CD in most patients following the HSC mobilization that was attributed to the immunomodulatory effects of the drugs used in this phase. In contrast, in the study from Milan,42 the authors reported a deterioration of the clinical condition of the patients during mobilization. Observations of two additional case reports published so far using HSCT as a primary treatment for CD also suggest that sustained clinical remission with HCT is likely not due just to cyclophosphamide and G-CSF. Scime et al.43 reported a case of a 55-year-old man who showed no improvement in colonic lesions 1 month after mobilization therapy, but achieved significant endoscopic and histological improvement after autologous HSCT. Five months after HCT, the patient remains in clinical remission. Kreisel et al.44 describe a 36-year-old man who continued to have persistent histological changes and recurrent CD 9 months after mobilization therapy with cyclophosphamide and G-CSF (before HCT). Ten months after HSCT, the CD was improved, in clinical remission with minimal histological changes. The mechanisms underlying the beneficial effect of HSCT in Crohn’s disease remain unclear. Initially, a beneficial effect is likely to result from eradication of autoreactive T cells and memory cells because of the direct lymphoablative effects of drugs used in the conditioning regimen; this process is non-specific and recruitment of inflammatory cells will be restricted for at least 2–3 months until their ablated bone marrow source reconstitutes with cells derived from the graft. Later, there may be an effect of altered immune reconstitution. To gain an insight into this question, the European Bone Marrow Transplantation (EBMT) Group, along with the European Crohn’s and Colitis Organization (ECCO), is currently conducting a randomized trial on autologous HSCT (ASTIC trial; to evaluate the potential clinical benefit of haematopoietic stem cell mobilization followed by high-dose immune ablation and HSCT compared with haematopoietic stem cell mobilization only followed by best clinical practice in patients with refractory CD.

In 1995, the international committee established for development of guidelines on entry criteria and transplant protocols for IMIDs recommended that autologous HSCT should be preferred to allogeneic due to lower risk of severe toxicity.11 Although at that time a genetic predisposition to CD had been established based on familial aggregation studies, not a single susceptibility gene had been identified. Now more than 50 disease susceptibility genotypes are known highlighting the complex heritability of this polygenic disease in the majority of cases. However, a recent study, based on genetic-linkage analysis and candidate-gene sequencing on samples from two unrelated consanguineous families with children who were affected by early-onset inflammatory bowel disease, identified three distinct homozygous mutations in genes IL10RA and IL10RB that perfectly segregated with disease phenotype.45 A member of the family with the IL10RB mutation suffered a very severe CD since the 3rd month of life with proctitis and perianal abscesses, which required multiple surgical interventions and finally colostomy; multiple enterocutaneous fistulas originating from the small intestine required bowel resections and ileostomy; and recurrent episodes of intestinal obstruction. At 9 years of age, he underwent allogeneic HSCT using as a donor an HLA-matched sibling unaffected not carrying the mutation. Inflammatory anal fistulas resolved shortly after transplantation. The patient has remained in continuous remission from ileocolitis; he gained weight and had no further episodes of intestinal pseudo-obstruction during the follow-up period of 2 years after HSCT. This is a stepping-stone study showing a curative approach to severe CD by means of allogeneic HSCT, which was justified based on the monogenic cause of the disease in this case.

Mesenchimal stem cells for treatment of IMIDs

The past 5 years have seen major interest in the use of MSCs in the treatment of autoimmune and inflammatory disorders. Human trials of MSCs for numerous immune-mediated diseases are being discussed and have been initiated in patients with GvHD. As MSCs can be easily obtained and culture-expanded, bone marrow- or adipose tissue-derived MSCs from third parties or from the original marrow donor have been infused to modulate refractory GvHD. In nonrandomized trials, 94% of patients with acute GvHD responded to intravenous infusion of MSCs.46

Animal models of tissue protection

It may be impossible, to separate the anti-inflammatory, immunomodulatory and tissue-protective ‘trophic’ effects of MSCs. An immunosuppressive effect of MSCs in vivo was first suggested in a baboon model, in which infusion of ex vivo expanded donor or third-party MSCs delayed the time to rejection of histo-incompatible skin grafts.47 MSCs also downregulate bleomycin-induced lung inflammation and fibrosis in murine models48 if given early (but not late) after the induction. A similar effects were seen in a murine hepatic fibrosis model using an MSC line,49 and in a rat kidney model of ischaemia–reperfusion injury in which syngeneic MSCs, but not fibroblasts, were used.50The notion that early treatment is necessary to provide significant clinical and pathological amelioration is also supported by observations in an experimental autoimmune encephalomyelitis (EAE) murine model.51 The responses are dependent on time of MSCs treatment (the earlier the better), and are reversed with IL-2 treatment, indicating that anergy rather than apoptosis had occurred.

Other experimental observations suggest that stem cell combinations may afford higher benefit that purified single cells lines. In a murine model of streptozocin-induced diabetes mellitus, clinical improvement was documented following transplantation with a combination of bone marrow-derived cells and syngeneic and allogeneic MSCs following sublethal irradiation; however, cell product alone was not effective.52 The proposed mechanism was regeneration of recipient-derived islet cells and immunosuppression of autoreactive T cells.

In experimental models of IBD, adipose-derived MSCs ameliorate clinical and microscopic signs of colitis, reduce systemic and mucosal pro-inflammatory cytokine production, increase IL-10 secretion and induce Tregs in mesenteric lymph nodes of colitic mice.53 MSCs emerge, therefore, as potent direct modulators of the innate immune system that preferentially target inflamed tissues and modulate T cell responses. MSCs from adipose tissue, in contrast to bone marrow MSCs, present the additional benefit of being more easily accessible and expanded in vitro. The potential anti-inflammatory actions of MSCs are summarized in Figure 1.

Figure 1.

 Mechanisms of action of mesenchymal stem cells (MSC) in inflammatory bowel disease. In active Crohn’s disease (right part of the figure), destruction of the epithelium or decreased antimicrobial defence by epithelial cells leads to an increased exposure to commensal bacteria that activate innate immune cells of the lamina propria. Activated mature antigen presenting cells (APC) acquire the ability to drive T-cell differentiation and activation into effector cells (IFN-γ producing Th1 or IL-17-producing Th17, depending on the nature of the response). An inflammatory infiltrate with a Th1 and Th17 phenotype accumulates in the lamina propria. Increased proinflammatory cytokine production by APC and macrophages induces up-regulation of adhesion molecules in both lymphatic and blood vessels that irrigate the intestinal mucosa and produce chemotactic factors, favouring recruitment of leucocytes and therefore further amplifying the inflammatory response. Transplanted MSCs migrate to inflamed tissues where they promote epithelial regeneration and tissue healing, and exert immuno-regulatory functions (left part of the figure). MSC inhibit APC maturation, T-cell proliferation and IFN-γ production. Transplanted adipose-derived MSC also decrease production of pro-inflammatory cytokines such as IL-6, TNF-α, and IL-12 by lamina propria mononuclear cells. Furthermore, MSCs can produce both IL-10 and TGF-b, which inhibit Th1 differentiation while promoting Treg differentiation, and restoring tolerance. Modified from Panes and Salas.7

MCS for treatment of human IMIDs

Successful pre-clinical studies using MSCs in models of autoimmunity, inflammation and tissue damage have paved the way for clinical trials. In humans, ex vivo expanded allogeneic MSCs have been infused in several phase I studies.54–56 No adverse events during or after MSCs infusion have been observed and no ectopic tissue formation has been noted. After infusion, MSCs remain in the circulation for no more than an hour. Although durable stromal cell chimerism has been difficult to establish, low levels of engrafted MSCs have been detected in several tissues. It is possible that sufficient therapeutic benefit is obtained by local paracrine production of growth factors and the provision of temporary immunosuppression by MSCs infusion. Infusion of haploidentical MSCs to a patient with steroid-resistant severe acute GvHD of the gut and liver promptly improved liver abnormalities and intestinal function.57 Upon discontinuation of ciclosporin, the patient’s acute GvHD recurred, but was still responsive to a second MSC infusion. Lymphocytes from the patient, when investigated on multiple occasions after MSCs infusion, continued to proliferate against lymphocytes derived from the haploidentical MSCs donor in coculture experiments. This suggests an immunosuppressive effect of MSCs in vivo, rather than a development of tolerance. The EBMT is currently sponsoring clinical trials of MSCs for the prevention and treatment of acute GvHD through the Developmental Committee and interim results of the treatment trial have recently been published. Thirty out of 55 steroid-resistant acute GvHD patients had a complete response with no immediate toxicity. Median dose of allogeneic MSC was 1.4 × 106/kg body weight up to five times. Four patients had a recurrence or de-novo malignancy. Patients who achieved a complete response had a reduced transplant-related mortality at 12 months (37% vs. 72%) and higher overall 2-year survival (52% vs. 16%).56

Autologous bone marrow-derived MSCs have been shown to be potently antiproliferative to stimulated T cells from normal individuals and from patients with IMIDs (rheumatoid arthritis, systemic sclerosis, Sjögrens syndrome and systemic lupus erythematosus). In patients with systemic sclerosis, the MSCs were normal with respect to proliferation, clonogenicity and differentiation to bone and fat.58 Encouraging results from an initial phase I clinical trial using locally administered ASCs to treat complex perianal fistula59 have been confirmed in a phase II multicentre randomized trial including patients with complex perianal fistulas (cryptoglandular origin, n = 35; associated with Crohn’s disease, n = 14), observing fistula closure in 17 (71%) of 24 patients who received ASCs in addition to fibrin glue compared with 4 (16%) of 25 patients who received fibrin glue alone, with similar fistula healing in Crohn’s and non-Crohn’s subgroups.60 The results of two trials of intravenous MSCs have been reported in abstract form. An open label phase II trial using allogeneic bone marrow-derived adult MSCs for the treatment of refractory luminal CD showed a clinical response defined as a reduction in CDAI of at least 100 points in three of nine treated patients.61 Results of another completed phase I trial of intravenous autologous bone marrow derived MSCs that included nine patients confirmed that the treatment is feasible and safe, but without apparent benefit for patients with severe refractory luminal disease.62

There are several ongoing phase I/II clinical trials of MSCs treatment in autoimmune disease, including Crohn’s disease, multiple sclerosis, systemic lupus erythematosus and type 1 diabetes mellitus, with other proposals under consideration for the treatment of other IMIDs such as vasculitis. For these studies, it is important to set clear therapeutic targets and attempt to identify uniform cell products. It is especially important to define the source and type of MSC (autologous or allogeneic), to standardize cell expansion conditions and to adopt uniform trial protocols. In addition, an international interdisciplinary registry of MSC-treated patients has been launched to allow uniform collection of long-term safety data.


The use of adult HSCs is rapidly expanding beyond the traditional applications for malignancy. In IMIDs, chemotherapy to ablate the disease-causing immune system is followed by infusion of unmanipulated or purified CD34+ HSCs to reconstitute immunopoiesis or haematopoiesis. A large experience exists on the use of HSCT as treatment for MS, systemic sclerosis and lupus. Important lessons from these studies indicate that nonmyeloablative regimens should be preferred to myeloablative ones, that patients should receive this cell therapy before irreversible damage occurs, while there is still a considerable component of reversible inflammatory lesions, and that morbidity and mortality are minimized in experienced centres.

A growing body of evidence shows that CD is also a highly responsive disease to this form of therapy. Time has now come for the scientific evaluation of the transplant-based procedures against acceptably best non-transplant treatments, and also for the comparison of various forms of stem cell therapies in randomized controlled trials. A robust rationale coming from experimental studies also sets the base for testing MSCs, in human IBD, but initial phase I and II studies have produced mixed results.

Notwithstanding the fascinating perspective of immune re-education and regulation, it is unrealistic to believe that cell-based therapies can eradicate the immune disease, but data available so far indicate that at least HSCT is able to change the natural course of Crohn’s disease.


Declaration of personal interests: None of the authors has competing interests related to the information provided in the manuscript. Declaration of funding interests: This work was supported in part by grant SAF2009/07272 from Ministerio de Ciencia e Innovación, Spain, to JP, and grant TRA-097 from Ministerio de Sanidad y Política Social, Spain to JP.