Allogeneic haematopoietic stem cell transplantation (HSCT) can cure a number of haematological malignancies such as leukaemias and lymphomas.1,2 The aim of using high-dose chemoradiotherapy is to kill as many malignant cells as possible and then rescue the patient using a haematopoietic stem cell graft. HSCT can also cure non-malignant disorders of the haematopoietic system, such as severe combined immunodeficiency and severe aplastic anaemia.3–5 HSCT may also be used as enzyme replacement to treat a variety of inborn errors of metabolism.6–11 Patients undergoing HSCT may have severe complications including toxicity of conditioning, infections by bacteria, fungi and viruses, and immunological complications such as graft failure and graft-versus-host disease (GVHD). Donor T-cells are responsible for triggering GVHD after activation by recipient antigens, which are expressed on host cells in the form of major histocompatibility complex, class I or class II antigens, viral antigens or minor antigenic peptides, including epithelial cell-associated antigens.12–16 Antigen-presenting cells, dendritic cells or macrophages present the antigens to T-cells. Cytokines then stimulate T-helper cells, which release interleukin 2 (IL-2), activating cytotoxic T-cells. Natural killer (NK) cells and macrophages also participate in this process. The main target organs for acute GVHD in humans include the skin, gut and liver. Acute GVHD is graded on a scale from 0 to IV, where grade IV is life-threatening.17 Risk factors for acute GVHD include human leucocyte antigen (HLA) disparity between recipient and donor, female donor to male recipient, the host environment and seropositivity for several herpesviruses.12,18,19
As prophylaxis, cyclosporine and methotrexate were equally effective in preventing GVHD.20 However, the combination of cyclosporine and a short course of methotrexate was superior to monotherapy and has been the gold standard for prevention of GVHD.21,22 An effective way to prevent GVHD is T-cell depletion of the graft. However, this may increase the risk of graft failure and leukaemic relapse.23,24
In unrelated donor transplants, antithymocyte globulin (ATG) decreased the risk of acute GVHD and improved survival.25–27 A dose-finding study of ATG given to patients receiving unrelated transplants showed that a low dose increased the risk of severe acute GVHD, while a high dose increased the risk of serious infectious complications.28 Intermediate doses of ATG were balanced to give an optimal outcome. First-line treatment of acute GVHD include high doses of steroids29 and it was seen that early treatment may decrease the risk of severe acute GVHD.30 Patients with severe acute GVHD who failed treatment of steroids have an extremely poor outcome.31 Other treatments that have been tried for severe acute GVHD include ATG, monoclonal antibodies against T-cells (for instance OKT-3), anti-IL-2 antibodies, anti-IL-2 receptor antibodies, psoralene and ultraviolet light, mycophenolate mofetil, pentostatin, sirolimus, thalidomide and almost all immunosuppressive treatments.32 Transplant-related mortality increases with increasing grade of acute GVHD.33
Chronic GVHD generally appears from 3 months and later after HSCT and resembles autoimmune disorders.34 Symptoms include generalized sicca syndrome, oral mucositis, oesophageal and vaginal strictures, malabsorption, wasting, liver disease, pulmonary insufficiency, bronchiolitis obliterans, myositis, neuropathy and immune deficiency.35 Chronic GVHD may be graded as limited or extensive, or mild, moderate or severe.36 Chronic GVHD may be treated by steroids, cyclosporine, tacrolimus, azathioprine, 1 Gy of total body irradiation, thalidomide, mycophenolate mofetil, sirolimus or anti-B-cell antibodies.12,37
The graft-versus-cancer effect
Weiden et al.38 reported that patients with GVHD, especially chronic GVHD, had an increased probability of being in remission. Twins who underwent HSCT and did not develop GVHD ran a higher risk of relapse than recipients of grafts from HLA-identical siblings.39,40 A European study showed that chronic GVHD had a stronger antileukaemic effect than acute GVHD in patients with acute leukaemia;41 the best leukaemia-free survival was seen in patients with mild acute and mild chronic GVHD.33,42 A graft-versus-leukaemia effect may also be seen in the absence of GVHD.43 This effect is not different using HLA-matched unrelated donors compared with HLA-identical sibling transplants.44 Bacigalupo et al.45 showed that in patients with acute myeloid leukaemia a high dose of cyclosporine was associated with an increased risk of leukaemic relapse, compared with a lower dose.45 We attempted to take advantage of the graft-versus-leukaemia effect by reducing the immunosuppressive therapy.46 By giving low doses of cyclosporine combined with methotrexate and discontinuing immunosuppression early, we were able to increase the risk of mild acute and chronic GVHD, and reduce the risk of leukaemic relapse. This strategy improved the long-term leukaemia-free survival.47 It is also possible to enhance the graft-versus-leukaemia effect by treatment with immunocompetent donor T-cells, known as donor lymphocyte infusion (DLI).48 Patients receiving DLI at an early stage, for example molecular or cytogenetic relapse, have a better response to this therapy than patients treated for more advanced haematological relapse.49 There also appears to be an antileukaemic cell-dose effect. Recipients of syngeneic grafts receiving a higher than median donor nucleated cell dose > 3 × 108/kg had a reduced risk of relapse compared with those receiving a lower cell dose.50 In another study, we found that recipients of peripheral blood stem cell transplants from HLA-identical sibling grafts receiving more than the median dose (≥ 6 × 106 CD34+ cells/kg) had a reduced risk of relapse and improved leukaemia-free survival compared with those receiving a lower stem cell dose < 6 × 106/kg.51
Haematopoietic stem cell transplantation has also been tried in patients with metastatic solid tumours and a poor prognosis. Childs et al.52 were the first to report an HSCT-associated antitumour effect in metastatic renal carcinoma. This group reported an overall response rate of 44% in 50 patients who underwent HSCT for metastatic renal carcinoma.53 Such an antitumour effect in renal carcinoma was confirmed by our group and in Europe.54,55 Survival was improved particularly in those treated with DLI and who developed chronic GVHD, being 70% at 3 years after HSCT.54
Other solid tumours in which HSCT has induced an antitumour response include metastatic colon carcinoma, ovarian carcinoma and pancreatic carcinoma.56–59 We combined liver transplantation as debulking with HSCT in patients with advanced liver cancer.60,61 Most patients with various advanced solid tumours died of progressive disease.55 However, 7 years ago we performed Whipple’s surgery in two patients with pancreatic carcinoma, followed by HSCT. Both patients are alive and free of tumour at the time of publishing; in contrast, 16 control patients who did not have an HLA-identical sibling donor and only underwent radical Whipple’s surgery died from tumour progression within 1 year (Omazic B and Ringdén O, Karolinska Institutet, 2014, unpublished data).
Home care during the neutropenic phase after allogeneic haematopoietic stem cell transplantation
Haematopoietic stem cell transplantation patients are, as a rule, treated in laminar airflow rooms or in isolation rooms in the hospital for protection against infectious complications.62–65 For 15 years we challenged this routine by letting patients living within 1 or 2 hours’ driving distance from our hospital be treated at home after HSCT.66 The patients received conditioning and the graft at our unit, but after this they went home. An experienced nurse from our unit visited the patient once or twice daily until engraftment and an absolute neutrophil count of > 0.5 × 109/l was achieved. When 36 patients had been treated at home, we compared them with control patients treated in the hospital.67 The patients spent a median of 16 days at home (range 0–26 days), but if they had a fever or other problems they were admitted to the hospital for antibiotics and could then return home. Before discharge to the outpatient clinic after HSCT, the home-care patients spent a median of 4 days (range 0–39 days) in the hospital. Among the advantages, the home-care patients were discharged earlier [relative risk (RR) 0.33, P = 0.03], had fewer days on total parenteral nutrition (RR 0.24, P < 0.01), fewer acute GVHD grades II–IV (RR 0.25, P = 0.01), lower transplant-related mortality (TRM) (RR 0.22, P = 0.04), and lower costs (RR 0.37, P < 0.05) than the control patients treated in hospital. Survival at 2 years was 70% in the home-care group compared with 57% in the control group (P < 0.03). Thereafter, home care became routine.
An unexpected finding was that home-care patients were less likely to develop grade II–IV acute GVHD, 17% compared with 44% in the control patients (P < 0.01). There may be several reasons for this. First, patients at home had better nutrition; second, there was a trend for the prevalence of bacteraemia and infections, which can trigger GVHD, to be lower in home-care patients; and, third, the bacterial environment at home may be safer than in hospital, where many strange and resistant bacterial strains may exist. As an example, gnotobiotic mice have a lower risk of developing life-threatening GVHD.68,69 A study in patients undergoing HSCT for severe aplastic anaemia also found that those treated in laminar airflow rooms were less likely to develop acute GVHD than those treated in regular hospital rooms.70
It is also cheaper to be treated at home.66 Home-care patients can be discharged earlier to the outpatient clinic and required less total parenteral nutrition, antibiotics and, in addition, TRM is reduced and more patients survive.
We feared that the home-care patients might have an increased risk of relapse compared with patients treated in the hospital, as there is an association between GVHD and a reduced risk of leukaemic relapse, as discussed above. There is also a close correlation between acute and chronic GVHD.35,71 Therefore, we completed a long-term follow-up of outcome in the home-care patients with emphasis on chronic GVHD, relapse and survival.72 The cumulative incidence of chronic GVHD was 52% in the home-care group compared with 57% in the hospital controls. TRM was 13% and 44% respectively (P = 0.002). The probability of relapse was similar in the two groups, 39% in those treated at home and 29% in the hospital controls. Four-year survival was 63% in the home-care patients as opposed to 44% in the controls (P = 0.04). In multivariate analysis, the only factor significant for poor survival was acute GVHD grades II–IV (relative hazard 2.41, P = 0.005). If GVHD was excluded, home care was associated with better survival (relative hazard 0.23, P = 0.019).
In contrast to many other modalities decreasing GVHD, home care seems to be optimal because it reduced acute GVHD and TRM, but had no effect on chronic GVHD and relapse. Thereby, survival was improved.
A study at our centre found a correlation between the number of days with no oral intake, before the diagnosis of acute GVHD, and the incidence of grade III–IV acute GVHD. In multivariate analysis, no oral intake was associated with acute GVHD grades III–IV (odds ratio 7.66; P = 0.016).73 Therefore, we suspected that the lower incidence of acute GVHD among home-care patients may be due to better oral intake compared with patients isolated in hospital.67 We therefore carried out a new matched-pair analysis with 76 home-care patients matched with a similar number of hospital controls – matched for age, sex, diagnosis, disease status, type of donor, conditioning, ATG, stem cell source and GVHD prophylaxis. Multivariate analysis showed that improved oral nutrition was strongly correlated with a decreased risk of acute GVHD grades II–IV [hazard ratio (HR) 0.95, P = 0.014] compared with home care (HR 0.49, P = 0.06).74 The conclusion from this study was that oral nutrition is a strong factor that contributes to the reduced risk of acute GVHD in patients treated at home compared with those treated in the hospital, where total parenteral nutrition is more common.
As a result, we improved oral nutrition in the hospital by having a special nutritional team of nurses. This team had the task of encouraging the patients in the hospital to increase their oral intake, with the aim of decreasing the incidence of acute GVHD. This nutritional team was introduced at our centre on 1 September 2006. We completed a new matched-pair analysis of 146 home-care patients and 146 hospital control patients.75 In this study we compared four groups, original home-care patients and original control patients (before 1 September 2006), and new home-care patients and new control patients (after 1 September 2006 when the nutritional team was introduced). We found that oral nutrition was indeed significantly improved in the hospital. The median oral intake in the hospital control patients before 1 September 2006 was 18.5 kcal/kg/day, compared with 25.6 kcal/kg/day in the new control patients (P = 0.002). However, this did not decrease the risk of acute GVHD in the new control patients. Among the four groups, the prevalence of acute GVHD grades II–IV was 15% in the original home-care patients (n = 76). Among the other three groups, acute GVHD grades II–IV occurred in 32–44% of the patients, but TRM did not differ between the four groups. There was a correlation between oral nutrition and low incidence of acute GVHD (P = 0.02). However, there was a stronger correlation in univariate analysis between number of days at home and incidence of acute GVHD (P = 0.005). In multivariate analysis, home care versus hospital was associated with a lower incidence of acute GVHD (HR 0.41, P = 0.02), but number of days at home was even more significant (HR 0.92, P = 0.004). Oral nutrition was not associated with a decreased risk of acute GVHD (HR 0.98, P = 0.13). Probability of relapse was the same in the home-care patients, 30%, as the hospital control patients, among whom the rate of relapse was 28%. There was a trend for better survival in the home-care patients than in the hospital control patients (P = 0.07). The conclusion from 15 years’ experience of home care was that home care is safe and that many days spent at home is correlated with a low risk of acute GVHD. Subsequent to this, we now encourage home-care patients to stay at home as many days as possible and to be sent back home even after readmission to the hospital and treatment with antibiotics – home care is encouraged.75 This practice has now started in the USA and Germany.
Stromal cells for treatment of graft-versus-host disease and toxicity
Definition and properties of mesenchymal stem cells
Mesenchymal stem cells (MSCs) are isolated from bone, fat and fetal tissues, cord blood and the placenta.76–78 Friedenstein et al.79 were the first researchers to describe MSCs.79 MSCs are rare and, in the bone marrow, they have been estimated to account for 1 out of 10 000 nucleated cells. MSCs may be useful in regenerative medicine because they can differentiate into several cells of mesenchymal cell lineage, such as bone, cartilage, tendon, cardiomyocytes, muscle or fat.80,81 MSCs may also be useful for haematopoietic support because they secrete several cytokines that influence the differentiation of haematopoietic stem cells82,83 and they stain positive for CD29, CD73, CD90, CD105 and CD166.81 However, there is no specific marker to identify MSCs as they are negative for haematopoietic markers such as CD34, CD45 and CD14. MSCs from bone marrow have the capacity to differentiate into bone, cartilage and fat after addition of exogenous factors.79,81,84 MSCs are not true stem cells because they cannot maintain a whole tissue compartment and cannot regenerate; however, they are multipotent in vivo and have the capacity to differentiate after in utero infusion in to newborn mice and chicken embryos.85,86
Immunogenicity and homing
Mesenchymal stem cells express HLA class I molecules on the cell surface, but not HLA class II84 – only after stimulation with interferon-γ can HLA class II be expressed on the cell surface. MSCs induce very low immunogenicity and low proliferation even after differentiation to bone, chondrocytes or adipocytes.87 MSCs are not lysed by cytotoxic T-cells to the same extent as target leucocytes from the same individual.88 NK cells capable of lysing K562 leukaemia target cells induced little lysis of MSCs. Fas-ligand and co-stimulatory molecules, such as B7-1, B7-2, CD40 and CD40L, are not expressed on MSCs.89 Human MSCs were rejected when they were injected into infarcted rat myocardium, demonstrating that xenografting rejection occurs.90 Even in the case of auto- or allo-transplantation, MSCs do not seem to be long-lived because they are extremely difficult to detect in vivo after infusion into humans.91 Infused MSCs first home to the lung and thereafter to the liver and spleen, and are subsequently detected in small numbers in almost all organs.92,93 We were able to demonstrate DNA from donor MSCs at autopsy in the gut, abdominal lymph node and urinary bladder, associated with GVHD of the gastrointestinal tract and haemorrhagic cystitis.94,95
I became interested in MSCs because they have the ability to inhibit T-cell alloreactivity. Thus, the mixed lymphocyte culture (MLC) reaction was inhibited by MSCs.84,87,89,96,97 Lymphocyte response to phytohaemagglutinin was inhibited in enriched CD3+, CD4+ and CD8+ T-cells by MSCs.98 MSCs constantly inhibited MLC at high concentrations, but variably inhibited and stimulated MLC when used at low concentrations (1:1000).84 We also found that MSCs induce suppression in MLC after differentiation to osteocytes, chondrocytes and adipocytes.87 Interferon-γ stimulation of MSCs enhanced suppression in MLC using undifferentiated or differentiated MSCs. When MSCs were added to the MLC, specific cytotoxic T-cell lysis was inhibited.88 However, no inhibition was seen when MSCs were added in the cytotoxic phase of the 51Cr release assay. Inhibition by human alloreactivity in vitro in MLC was caused by soluble factors, because MSCs inhibited response in MLC even if they were separated by a transwell membrane.88 MSCs affect T-cells, B-cells, dendritic cells and NK cells, and almost the whole immune system.99 However, B-cells are stimulated to IgG secretion when co-cultured with MSCs.100 Regulatory T-cells and activated T-cells are increased after MSC stimulation.101 A variety of factors are used by MSCs to suppress immune responses. Some of these factors are constitutively produced by MSCs, such as HLA-G,102 prostaglandin E2103 and galectines.104,105 MSCs can also be activated by stimulation of their toll-like receptors.105 An important mediator of suppression by MSCs is the T-cell inhibitory enzyme indoleamine-2,3-deoxygenase (IDO).106 MSC-produced IDO is involved in the induction of regulatory T-cells and inhibition of Th17 differentiation.107,108 MSC-derived IDO also promotes differentiation of macrophages towards an M2 phenotype.109 Activated MSCs can modulate adaptive immune cells through contact-dependent mechanisms that include activation of the PD-1 passway,110 fas-mediated T-cell apoptosis,111 engagement of VCAM-1 and ICAM-1,112 or through up-regulation of CD39 and increase in adenosine production.113 Nitric oxide synthesis is a main mediator of MSC-induced suppression in mice.114
Mesenchymal stem cells and animal models of graft-versus-host disease
Several mouse studies found no effect on GVHD using MSCs at any dose.115,116 Some studies showed improved survival following infusion of MSCs at day +2 or day +20, but not at earlier or later infusion.117 In that study, interferon-γ-activated MSCs were used. Min et al.118 used IL-10-transduced MSCs and reported improved survival. CXCR4-transduced MSCs also improved survival and lowered histopathological damage in mice with GVHD.119 MSCs failed to prevent acute GVHD in the canine model;120 thus, the animal data have not shown any convincing evidence for MSCs or other stromal cells as a useful treatment for GVHD.
Clinical experience using stromal cells for treatment of acute and chronic graft-versus-host disease
It was safe to infuse MSCs, as demonstrated in pilot studies in which these cells were given to promote engraftment following autologous or allogeneic haematopoietic stem cell transplantation.91,121 Bartholomew et al.96 showed in a baboon model that MSCs prolonged skin allograft survival.96 At our unit, a 9-year-old boy with acute lymphoblastic leukaemia developed therapy-resistant grade IV acute GVHD with voluminous haemorrhagic diarrhoea and highly elevated bilirubin. GVHD was non-responsive to treatment with cyclosporine, high-dose prednisolone, repeated pulses with methylprednisolone, extracorporeal psoralen combined with ultraviolet A and several infusions of infliximab (Remicade®, Merck Sharp & Dohme Ltd, White House Station, NJ, USA) and daclizumab (Zenapax®, Hoffman-LaRoche, Basel, Switzerland). Based on our knowledge that MSCs inhibited alloreactivity in vitro despite HLA incompatibility between MSCs and stimulator and responder cells, I aspirated bone marrow from his HLA-haploidentical mother and 2 × 106 MSC/kg was infused with subsequent normalization of stool and bilirubin within a week of MSC infusion.98 Subsequently, he had minimal residual disease of acute lymphoblastic leukaemia and therefore I decided to discontinue treatment with cyclosporine. Following this, GVHD reappeared with haemorrhagic voluminous diarrhoea and bilirubin increased to 350 mmol/l. In response to this, we infused 1 × 106 MSC/kg from his mother, which had been stored frozen in liquid nitrogen. The patient responded again and was sent home.
In the initial compassionate-use study, seven more patients were included.94 Although some patients did not respond, and several died from infection, survival was significantly better than that of similar patients not treated with MSCs. Among those patients, resolution of GVHD was seen in the gastrointestinal tract, liver and skin. A patient with chronic GVHD had a partial response. This study prompted me to initiate a larger European phase II study in patients with therapy-resistant acute GVHD treated in five centres.122 This study included 55 patients treated with MSCs at a median dose of 1.4 × 106 MSCs/kg (range 0.4–9× 106 MSCs/kg). The patients received from one to five infusions of MSCs from HLA-identical siblings, HLA-haploidentical donors and a majority from unrelated HLA-mismatched third-party donors. Complete response to MSC infusion was seen in 30/55 patients (55%) and partial response was seen in nine patients. Children tended to respond better to the treatment, with 68% responding, compared with 43% of adults (P = 0.07). Among the complete responders, 2-year survival was 52%, which was significantly higher than the 2-year survival of 16% among those with partial or no response (P = 0.018). Our findings that MSCs could cure life-threatening acute GVHD opened up a new field of cell therapy in regenerative medicine. MSCs are now used for a variety of autoimmune disorders such as multiple sclerosis and Crohn’s disease, and have also been used to treat myocardial infarctions.
Subsequent to our promising results, there has been a number of reports using various types of MSCs to treat acute GVHD; Fang et al.123 treated six patients with adipose tissue-derived MSCs for steroid-refractory acute GVHD and von Bonin et al.124 used MSCs expanded in platelet lysate medium for GVHD. Overall, there are reports from several publications on 190 patients who have been treated with a cell dose ranging from 0.4 to 9.0 × 106/kg from 1 to 21 doses, with a complete response of 52%, a partial response of 23% and no response in 25% of the patients (studies summarised in Ringdén125). Osiris Therapeutics (MD, USA) used expanded MSCs for grade II–IV acute GVHD with an initial response of 94% and a complete response of 77%.126 Subsequently, Osiris Therapeutics carried out a prospective double-blind placebo-controlled phase III study in which patients with grade II–IV acute GVHD were randomised to either Prochymal® (Osiris Therapeutics, Columbia, MD, USA) or placebo in a ratio of 2:1.127 Overall complete response at 28 days was 45% in the Prochymal group and 46% in the placebo group. However, among 61 patients with acute GVHD of the liver, complete response was 76% in the Prochymal group and 47% in the placebo group (P = 0.026). In 71 patients with gastrointestinal acute GVHD, complete response was 88% in the Prochymal group compared with 64% in the placebo group (P = 0.018). In this study, there was also a trend for a better outcome in children than in adults, and so Prochymal is now registered in Canada and New Zealand for the treatment of severe acute GVHD in children. There are few long-term reports of the use of MSCs for the treatment of acute GVHD; at our centre, 31 patients treated with MSCs for acute GVHD or haemorrhagic cystitis were followed for more than 4 years.128 Survival was better for those patients who had received MSCs from passages 1–2, 75% at 1 year, as opposed to 21% among those receiving MSCs from passages 3–4 (P < 0.01).
Although bone marrow-derived MSCs are the most employed, adipose-, umbilical cord-, or placenta tissue-derived stromal cells have also been utilized for treatment of acute GVHD.123,129,130 In a mixed-lymphocyte reaction, we compared three different sources of stromal cells from the placenta, the fetal membrane, the umbilical cord and from placental villi.131 Decidual stromal cells from the fetal membrane were the most consistent in inhibiting mixed lymphocyte culture, and they were therefore selected for a clinical pilot study. Subsequently, we used decidual stromal cells for the treatment of severe steroid-refractory grade III–IV acute GVHD.130 The median age of the patients was 57 years and there was one child in the study. Two patients had a complete response, four had a partial response and three patients became long-term survivors. This study showed that decidual stromal cells could cure life-threatening acute GVHD.
Stromal cells for treatment of chronic GVHD
Chronic GVHD resembles autoimmune diseases and, since MSCs had an effect on experimental autoimmune diseases, chronic GVHD would be an ideal disease for treatment with stromal cells.132 In the first patient we treated, we saw a partial response by resolution of liver GVHD after MSC treatment.94 Several reports included 61 patients treated for chronic GVHD with a complete response rate of 26%, a partial response rate of 48% and no response in 26%.125 The MSC dose ranged from 0.2 to 20 × 106/kg and 1–11 doses were given. Weng et al.133 reported on 19 patients with refractory chronic GVHD and saw response in 74% of the patients.133
Stromal cells for tissue toxicity and haemorrhages
Mesenchymal stem cells seem to home to sites of injury.134 When treating patients with acute GVHD, we found that haemorrhages stopped when the patients were treated with MSCs.94 Therefore, we used MSCs for haemorrhagic cystitis.95 The first two patients had life-threatening haemorrhagic cystitis grade V, and both died from multi-organ failure, but transfusion requirements were dramatically decreased after MSC infusions. Among 10 patients treated at an earlier stage of haemorrhagic cystitis, eight had a complete response and two were non-responders. Gross haematuria disappeared after a median of 3 days (range 1–14 days). We also treated major gastrointestinal haemorrhages in a 68-year-old man who had multispecific anti-HLA antibodies and was refractory to platelet transfusions. Surgery was impossible, but this patient responded to MSC infusion after an infusion of 2 × 106 MSCs/kg pooled from two donors.135 After this, we studied the effect of MSCs on the coagulation system and found that there was a profound effect on coagulation.136,137 We also saw that MSCs could cure colon perforation and peritonitis twice in a patient with HSCT;95 this finding was confirmed in a Japanese patient.138
Graft-versus-host disease causes morbidity and mortality after HSCT and severe steroid-refractory acute GVHD is a life-threatening disease. GVHD may also induce an anticancer effect. Care at home instead of isolation in hospital may decrease the risk of acute GVHD without harnessing the graft-versus-leukaemia effect and thereby improve survival. Therefore, home care should be encouraged. Stromal cells from various sources have immunomodulatory properties and can heal damaged tissue. We found that they could reverse life-threatening acute GVHD in some, although not all, patients. Stromal cells have also been used successfully for treating chronic GVHD and tissue damage including haemorrhages. Although promising, more research is required to make stromal cell therapy an established treatment for GVHD.