Hematopoietic stem cell transplantation in multiple sclerosis.

It is widely accepted that the main common pathogenetic pathway in multiple sclerosis (MS) involves an immune-mediated cascade initiated in the peripheral immune system and targeting CNS myelin. Logically, therefore, therapeutic approaches to the disease include modalities aiming at downregulation of the various immune elements that are involved in this immunological cascade. Since the introduction of interferons in 1993, more specific immunoactive drugs have been introduced, but still most of them can, at best, effectively modulate only the early relapsing phases of MS. The more progressed phases of the disease are not efficiently amendable by the existing immunomodulatory drugs. Moreover, localized and compartmentized inflammation in the CNS, which seems to be mostly responsible for the chronic axonal damage and resulting progression of disability, is less affected by the current drugs. A more radical approach to suppress all the inflammation in MS, including that into the CNS, could theoretically be achieved with high-dose immunosuppression using strong cytotoxic medications and resetting of the immune system by hematopoietic stem cell transplantation (HSCT). HSCT, both allogeneic and autologous, has been tried as a novel therapeutic approach in various autoimmune diseases. During the last 15 years several (mostly open) clinical studies evaluated the effect of HSTC on MS patients; the published papers showed that a high proportion of the HSCT-treated MS patients were stabilized, or even improved after the transplantation and have generally indicated a beneficial effect on disease progression. In this review, the rationale of HSCT and the summary of the results of the existing clinical trials are presented. Despite the fact that it is difficult to collectively summarize the results of all the trials, due to lack of uniformity in the conditioning and treatment protocols and of completed controlled studies, these clinical studies have provided a strong ‘proof of concept’ for HSCT in MS and have significantly contributed to our understanding of the advantages and disadvantages of each approach and HSCT protocol.

Multiple sclerosis (MS) is a chronic, immunemediated inflammatory disease of the CNS that causes demyelination and loss of motor, sensory and cognitive function [1]. Clinically, most MS patients experience recurrent episodes (relapses) of neurological impairment (relapsing-remitting MS). Early neurological dysfunction may resolve spontaneously partially or completely, but usually the course of the disease becomes chronic and progressive (primary progressive MS or secondary progressive MS), leading to accumulating motor, sensory and cognitive deficits.
The hallmark pathological feature of MS at the early stages is the demyelinating plaque with perivenular inflammation and mononuclear cell infiltration [1]. Inflammation leads to damage or loss of oligodendrocytes and demyelination leads to disruption of the conduction of neuronal signals in the affected regions. In the initial stages of MS, compensatory pathways (such as the upregulation of ion-channels in the affected areas) may partially restore conduction and reverse the neurological dysfunction. As the disease progresses significant axonal loss and eventually neuronal damage, occurs and the lost function becomes permanent and nonreversible [2].
It is widely accepted that the inflammatory process in MS is caused or propagated by an It is widely accepted that the main common pathogenetic pathway in multiple sclerosis (MS) involves an immune-mediated cascade initiated in the peripheral immune system and targeting CNS myelin. Logically, therefore, therapeutic approaches to the disease include modalities aiming at downregulation of the various immune elements that are involved in this immunological cascade. Since the introduction of interferons in 1993, more specific immunoactive drugs have been introduced, but still most of them can, at best, effectively modulate only the early relapsing phases of MS. The more progressed phases of the disease are not efficiently amendable by the existing immunomodulatory drugs. Moreover, localized and compartmentized inflammation in the CNS, which seems to be mostly responsible for the chronic axonal damage and resulting progression of disability, is less affected by the current drugs. A more radical approach to suppress all the inflammation in MS, including that into the CNS, could theoretically be achieved with high-dose immunosuppression using strong cytotoxic medications and resetting of the immune system by hematopoietic stem cell transplantation (HSCT). HSCT, both allogeneic and autologous, has been tried as a novel therapeutic approach in various autoimmune diseases. During the last 15 years several (mostly open) clinical studies evaluated the effect of HSTC on MS patients; the published papers showed that a high proportion of the HSCT-treated MS patients were stabilized, or even improved after the transplantation and have generally indicated a beneficial effect on disease progression. In this review, the rationale of HSCT and the summary of the results of the existing clinical trials are presented. Despite the fact that it is difficult to collectively summarize the results of all the trials, due to lack of uniformity in the conditioning and treatment protocols and of completed controlled studies, these clinical studies have provided a strong 'proof of concept' for HSCT in MS and have significantly contributed to our understanding of the advantages and disadvantages of each approach and HSCT protocol.

Hematopoietic stem cell transplantation in multiple sclerosis
Review autoimmune cascade, involving mainly T cells that target myelin self antigens [3,4], possibly through mechanisms known as molecular mimicry (cross-reactive antigens expressed by viruses or other microorganisms and myelin components) [5]. An alternative hypothesis is that myelin-specific T cells, that are present 'naturally' may expand to critical pathogenic quantities due to malfunctioning immunoregulatory mechanisms (such as those involving the Th2, Th3 and CD8 + T cells and the T regulatory cells: Tr1 and Tregs) [6].
Although the autoimmune hypothesis is attractive and supported by concrete data (including the efficacy of immunomodulatory treatments in MS), the initial insult that initiates the whole immune-mediated cascade is still obscure. Environmental, genetic and infectious factors also seem to play an important role in MS pathogenesis. Specifically, putative infectious agents, if and when they are involved, may represent one of the triggers of the autoimmune process [6], rather than the primary target of the infiltrating cells. In any case, T cells of the Th1 and Th17 phenotype, specific for myelin antigenic epitopes seem to represent the common final pathogenetic effector pathway, regardless of the initial insult of the disease [7][8][9].
Logically, the therapeutic approaches to the disease include modalities aiming at downregulation of the various immune elements that are involved in this immunological cascade. Since the introduction of interferons in 1993 many more specific immunoactive drugs have been introduced, but still, most of them can at best effectively modulate only the early relapsing phases of MS. The more progressed phases of the disease are not efficiently amendable by the existing immunomodulatory drugs. Moreover, localized and compartmentized inflammation in the CNS, which seems to be mostly responsible for the chronic accumulating axonal damage and the resulting progression of disability is less affected by the current drugs [10][11][12].
A more radical approach to suppress all the inflammation in MS, including that in the CNS, could theoretically be achieved with high dose immunosuppression using strong cytotoxic medications and resetting of the immune system by hematopoietic stem cell transplantation (HSCT).

Hematopoietic stem cell transplantation
HSCT is unique among stem cell-related treatments as it does not primarily aim at neuroregeneration but rather at replacement/resetting of the whole immune system, that is, the rebuilding of all the series of immune cells from progenitor hematopoietic stem cells (HSCs), following the destruction of the 'old' immune system by radical immunosuppression (defined as 'conditioning'). An additional and more controversial mechanism of action of HSCs is their putative transdifferentiation into cells from the neuronal lineage (which may theoretically lead to neuroregeneration) and their neuroprotective/neurotrophic effects [13][14][15].
HSCT, both allogeneic and autologous, is one of the hottest areas of clinical immunology. In autologous HSCT (AHSCT), the (autoimmune-prone) immune system is partially or completely (depended on the intensity of the conditioning protocol) wiped out, followed by reinfusion of the patient's own HSC ( identified as CD34 + ). The idea behind this approach is that during the process of immune reconstitution, the status of tolerance to self-proteins may be re-established and the newly developing, 're-educated' lymphocytes will hopefully no longer carry the immunological memory of previous autoimmunity. In allogenic HSCT, the patient's 'faulty' immune system is 'replaced' by a healthy one (deriving from the donor of HSCs) and, hopefully, during this process the host autoimmune clones are eradicated. In the latter case, a 'graft-versus-autoimmunity' (GVA) effect [16], may theoretically further contribute to the putative therapeutic beneficial effect, but long-term immunosuppressive therapy of the transplanted patients is also needed to prevent rejection and graft-versus-host disease (GVHD). Whether this last intervention will be capable of achieving the 'holy grail' of self-tolerance is still not established given the complexity of the pathogenesis of MS and autoimmune diseases in general, including the persisting antigenicity of (possibly) altered 'self ' proteins and the involvement of genetic factors, 'embedded' in the stem cell genome, as well.

Different protocols for HSCT & their advantages & disadvantages
The protocols of AHSCT in general include, two stages: first, the harvesting of donor HSC, which are collected either from the bone marrow (by bone marrow aspiration) or from the peripheral blood (by cytapheresis). Since the number of circulating peripheral blood HSC (defined as CD34 + ) is very low, mobilization is induced using cyclophosphamide and granulocyte colony-stimulating factor. The target number of harvested CD34 + cells is 3-5 × 10 6 /kg. The prepared HSCs are subsequently cryopreserved.
The second phase involves the preparation of the host for the transplantation by suppression or ablation of the immune system using cytotoxic therapy of varying intensity. The conditioning regimens consist of a combination of chemotherapy/cytotoxic modalities that may include cyclophosphamide, busulfan, total body irradiation (TBI) or -most often the combination of BCNU (carmustine), etoposide, ARA-C (cytosine arabinoside) and melphalan (BEAM). Bone marrow or HSCs are thawed and infused to the recipient patient, either unmanipulated or following sorting for the expression of the CD34 + surface marker. In some of the protocols, for further elimination of the mature circulating lymphocytes, antithymocyte globulin is administered (in vivo T-cell depletion). However, such radical depletion of the mature lymphocytes is associated with a more significant and longer immuno-incompetence stage and more risk of infection, whereas on the other hand, there is no concrete evidence for a higher beneficial effect of T-cell depleted HSCT in ADs and particularly MS. Granulocyte colony-stimulating factor may be administered after transplantation especially if the neutrophil count remains low, and may be combined with steroidal treatment to avoid the risk of flare of autoimmunity or neurologic adverse effects. Antibiotics and acyclovir are usually administered for up to 3 months following HSCT for opportunistic infection prophylaxis.
In AHSCT, the goal of the conditioning regimen is radical immune suppression and its complete rebuilding from new stem cells. As mentioned above, the intensity of the immuno suppressive regimen may vary. In the only published study that compared the outcome of myeloablative regimens (TBI or busulfan) to less intense regimens, the mortality of TBI or busulfan-based regimens was significantly greater (fourfold higher) than less intense regimens without improvement in disease-free survival. While further registry studies analyzing the outcomes between myelo ablative and nonmyeloablative regimens are needed, currently there is no data to support the superiority of more intense myeloablative regimens containing either TBI or busulfan over less intense nonmyeloablative regimens with agents such as cyclophosphamide. Contrary to this, a systematic review of patients that have received high intensity radiation-based conditioning regimens showed that these patients actually responded less to the treatment [17].
The goal of allogeneic HSCT is to alter the genetic predisposition to disease by changing the stem cell compartment. It is not clear whether a full chimera (in which all HSCs are reconstituted from the donor) or mixed chimerism (with co existence of both donor and recipient immune cells) is more effective for the control of autoimmunity in MS. In animal models, either full donor chimerism or mixed chimerism was shown sufficient to induce remission and prevent relapse of autoimmune diseases [18,19]. Therefore, unlike malignancies where mixed chimerism is associated with a higher rate of disease relapse, such a chimeric state seems to be beneficial in autoimmune diseases.
On the other hand, full donor chimerism in malignancies has been complicated by a high rate of GVHD, an immune-mediated disease in which allogeneic donor lymphocytes are targeting host proteins. In mixed chimerism the incidence of GVHD is significantly lower. However, mixed chimerism is often unstable in humans and may eventually lead to graft rejection. In any case, to prevent or control GVHD following allogeneic HSCT, the use of immunosuppressive medications (often at significant doses) is necessitated. The latter, together with higher treatment mortality rates, make allogeneic HSCT a less attractive option for the majority of patients with MS, except possibly in the case of 'mini-transplants' (using lower intensity, nonmyeloablative chemotherapy) with less intensive conditioning. However, mini-transplants still have a significant mortality related to acute and chronic GVHD, which is greater than that of autologous stem cell transplantation.

Rationale for HSCT in MS & autoimmunity
Various experimental and clinical studies have provided the scientific rationale for the utilization of autologous and allogeneic HSCT in MS. Animal studies showed that when spleen and/or whole marrow cells were transferred to immunosuppressed New Zealand Black, BXSB and Murphy Roths Large mice they could reproduce murine lupus [20]. Conversely, transfer of bone marrow from healthy mice could cure the autoimmune manifestations of lupus in Murphy Roths Large and New Zealand Black mice [21]. In other animal studies, experimental autoimmune disease were efficiently ameliorated or even cured following syngeneic or allogeneic HSCT and replacement of the 'faulty' immune system that produced the autoreactivity by a 'healthier' one [22][23][24][25][26]. Allogenic HSCT was shown to be more effective in animal studies, possibly due to an additional GVA effect (specific or generalized anti-autoimmune effect) induced by the donor's immune cells [16]. It was shown that such active alloreactivity is associated with the greatest GVA effect [16,[27][28][29][30]. The reported clinical observation that patients affected by autoimmune diseases who were treated with allo-HSCT due to coexisiting hematological malignancy were ultimately cured of the autoimmune disease, further supports the rationale for using HSCT in autoimmune diseases [31][32][33]. Theoretically, the de novo immune system reconstitution, in the presence of the self-autoantigens in the thymus, may induce a resetting/restart of the immune system and reinduction of self-tolerance.
The scientific basis for this therapeutic approach in neuroimmunological diseases came initially from the practical studies by our group [25,26,34], which were later further supported by others [35,36]. These animal studies showed that it was possible to effectively suppress and even 'cure' experimental autoimmune encephalomyelitis (EAE), which serve as animal models for human MS, by means of autologous or 'pseudoautologous' HSCT [25,26,[34][35][36][37]. It has been shown in mouse EAE, both actively induced (by immunization with myelin homogenate, proteolipid protein-PLP or myelin oligodendrocyte glycoprotein) and passively induced (with transfer of myelin antigens specific T-cell lines), that autologous/syngeneic bone marrow transplantation (BMT) can prevent the appearance of motor weakness and induce long-term antigen-specific tolerance; in chronic EAE, HSCT could even reverse chronic neurological disability [23,25,26,34]. However, it has to be emphasized that although EAE may serve as a model of the inflammatory phase of MS, there are substantial histopathological differences between this model and human disease. Moreover, in EAE the disease is caused by immunization techniques and the animals are inbred strains, a fact that certainly does not correspond to the genetic variability among MS patients.

Worldwide clinical experience with HSCT in MS
In this review, the authors searched the internet (PubMed and NIH clinical trials sites) for all the published clinical studies with HSCT in MS since 1990. The search keywords used were: 'HSCT', 'bone marrow transplantation', 'multiple sclerosis' and 'clinical trials'. According to the European BMT registry, among the 300,000 transplantations during the last 15 years, there were 1089 cases of autoimmune disease that were treated with HSCT/BMT; half of them were patients with MS, from various countries including USA, Europe, Israel, China, Brazil, Russia and Canada.
There are at least 393 registered transplantations in MS patients, (157 male and 236 females, aging from 16 to 65 years, 48% with secondary progressive MS, 24% other progressive forms and 22% relapsing-remitting). An overall 95% 5-year survival rate and 62% progression-free survival at 3 years and 50% in longer follow-up were seen, with 100-day transplantation-related mortality ranging between 1.3 and 7.7%. At 10 years, the survival rate was 90% and the progression-free survival around 65% for secondary and 40% for primary progressive MS [38][39][40]. In addition, AHSCT favorably affected the quality of life of MS patients. When parameters like Review the MS quality of life questionnaire were examined, a significant improvement in both composite scores and in many of the individual scores of the questionnaires were observed [41]. The most used (75%) conditioning protocol is with cyclophosphamide and granulocyte colony stimulating factor (for mobilization) and BEAM for conditioning (in 50%). Younger patients (<40 years of age) and those with less progressed disease had better outcomes with progression-free survival of 67% (n = 45) in relapsing MS and 31% (n = 103) in progressive forms of MS.
However, indications for chronicity and/or recurrence of autoimmunity in MS, have also been reported following autologous or allogeneic BMT as evidenced by the persistence of oligoclonal antibodies in the cerebrospinal fluid and the continued demyelination and axonal damage [42]. Such relapses, post-HSCT, could be explained either by changes in the autoantigens profile (epitope spreading) or by a new de novo developing autoimmune process. Additional causes for disease recurrence might include (but are not limited to) insufficient immune cytoreduction, reintroduction of autoreactive lymphocytes in the stem cell graft or persistent immunogenetic risk factors.
In the first pioneer study by Fassas et al., 15 patients with progressive MS were transplanted with peripheral blood HSCs [43]. Six of the transplanted patients had mild transient neurotoxic side effects in the early stages following post transplanation. In a 6-month follow-up neurologic improvement was detected in a substantial proportion of the patients, as evidenced both by the expanded disability status scale (EDSS) follow-up (7 out of 15) and the Scripps neurologic rating scale (15 out of 15) clinical score changes. Two patients experienced a relapse and one deteriorated at 3 months post-HSCT. The same group reported the long-term results of their Phase I/II study, which included the clinical and MRI outcome of 35 patients with aggressive MS treated with HSCT. Disease progression-free survival was higher (44%) in patients with active disease, as compared with those with no MS activity prior to the treatment (10%). Two patients died at 2 months and 2.5 years from transplant-related complications. HSCT demonstrated an impressive effect in suppressing disease activity in MRI as evidenced by the absence of gadolinium-enhancing lesions. Active lesions were reduced after mobilization and most prominently post-HSCT [44].
In another study, no significant clinical benefits were documented in patients with progressive MS following TBI and HSCT [45]. In a more recent study by the same group [46], the safety and clinical efficacy of autologous nonmyeloablative HSCT was evaluated in 21 patients with relapsing-remitting MS. In this study, peripheral blood HSCs were mobilized with cyclophosphamide and filgrastim. After a mean of 37 months, 16 of the patients were free of relapses, and all were free of progression. Patients improved in EDSS, and in the paced auditory serial addition test, the timed 25-foot walk and the quality of life assessments. In another study Saccardi et al., treated nonprimary progressive MS patients with high disease activity despite conventional treatments with cytotoxic conditioning and infusion of unmanipulated peripheral blood stem cells [41]. All patients were stabilized or improved after HSCT; three patients subsequently deteriorated, one beyond the baseline.
MRI did not show any active lesions post-transplantation with the exception of one lesion in one patient 4.5 years post-HSCT. In general, in this study, HSCT induced a prolonged clinical stabilization in the patients with severe progressive MS.
In another study, 50 patients with various types of MS were treated with HSCT. No deaths were reported. A total of 28 patients improved (at least 0.5 point decrease in the EDSS as compared with the baseline, confirmed at 6 months). Furthermore, 17 patients were stabilized for at least 6 months post-HSCT and 72% of the patients were free of disease progression for 6 years post-HSCT. The MRI of the patients who were stabilized did not show any new lesions or activity [47].
Ni et al. reported the safety and efficacy of AHSCT in 21 progressive MS patients during a 3 year follow-up period [48]. Krasulová et al. also reported 10 years of experience with highdose immunoablation and AHSCT in aggressive MS [49]. A total of 70.8% of the patients were free from progression after 3 years and 29.2% were still stable after 6 years. Younger patients (<35 years) with relapsing course and those with disease duration of less that 5 years and relapsing course had a better outcome. No deaths were reported within 100 days post-AHSCT. In a recent publication, Mancardi et al. reported the outcome and follow-up of all MS patients treated with AHSCT using an intermediate intensity conditioning regimen, in Italy from 1996-2008 [50]. Clinical and MRI outcomes of 74 patients were collected after a median follow-up period of 48.3 months. Overall, this study showed that AHSCT, with the above conditioning regimen, had a sustained beneficial effect in aggressive MS cases unresponsive to conventional therapies, halting disease progression and even inducing a sustained clinical improvement, especially in relapsing forms of MS.
All the published trials with HSCT in MS are summarized in Table 1. In general, patients with severe CNS inflammation experienced a more substantial improvement in their disability. Patients with low disability treated with a nonmyeloablative conditioning protocol had the most impressive response [46]. Patients with primary progressive course and those with long disease duration are less likely to benefit from HSCT [45]. This has also been supported by histological examinations showing continuous demyelination and axonal damage despite the suppression of the inflammatory process [51][52][53].
In general, when trying to summarize the results of HSCT in MS patients, it must be taken into account that most of the published studies were open and an inherent selection bias for both the patient and the investigators could have influenced the reported outcomes. In most of the trials the treated patients had high EDSS scores and a progressive course, that is, they were patients with disease less amendable to treatment.
Despite this, the most dramatic beneficial effects were actually more prominent in the 'malignant' cases of MS [49,54]. HSCT is very effective in suppressing the inflammation in MS, as indicated by the elimination of the gadolinium-enhancing lesions in the MRI and the stop of clinical relapses.
However, the effect of HSCT on disease progression does not seem to be solely attributed to the immunosuppression induced by the conditioning treatment myelin-specific T-cells are initially eliminated after the treatment, but are eventually largely    Review restored within a year post-HSCT [55]. In general, following HSCT, memory T cells were shown to decrease, whereas, naive CD4-positive thymic cells are expanded along with a reconstitution of a broad clonal diversity and renewal of myelin antigens clonal specificities [55]. These findings may explain the long-term beneficial effects of HSCT through an induction of a new level of immunological tolerance and restoration of the regulatory immune mechanisms.

Reservations -recurrence of the activity of MS, brain atrophy & toxicity
Recurrency of MS, relapses of the disease and continuation of the accumulation of disability, demyelination and axonal loss after HSCT have been reported in some cases. The reported continuation of the presence of oligoclonal antibodies (one of the laboratory hallmarks of MS) in the cerebrospinal fluid, indicates an ongoing intrathecal production of antibodies, even after the transplantation [56]. The continuation of inflammation could be related either to a change in autoantigens (epitope spreading) and T-cell reactivity profile or to de novo autoimmune activity, but it may also be attributed to previous activity of MS, which possibly started before the transplantation. It appears apparently, that even when strong suppression of the inflammatory process in MS is achieved with HSCT, still demyelination, axonal damage and accumulation of disability may not be completely stopped. The latter may also indicate that the main two pathogenetic processes in MS ( inflammation and degeneration) are not always in good correlation.
Case report neuroimaging studies have indicated that brain atrophy seems to continue and even surprisingly accelerate following HSCT [57]. However, this finding may be related to a 'pseudoatrophy' caused by the resolution of edema in the areas of brain lesions or be a late result of the pre-HSCT disease activity and/or a result of the intensity of conditioning regiments (possible toxic effect). In any case, the rate of atrophy was shown to decrease after the second post-HSCT year [58,59].
One of the main limitations of HSCT is the toxicity caused by the conditioning regimens and the resulting side effects that include opportunistic infections, organ damage, transient progression of neurological signs and marked mortality rate.
Side effects, death or development of secondary autoimmune diseases were seen even during the late stages of post-HSCT.
Procedure-related mortality was significantly higher before 2000 (7.3%) but dropped to less that 1.3% in reports after 2000. The overall mortality rate reported by the European Group for Blood and Marrow Transplantation registry was 6% (n = 85) until 2006 and 5.3% in 2011 (n = 185).
In order to reduce the mortality risk 'lighter' protocols were suggested and used by some researchers, such as with cyclophosphamide 200 mg/kg.b.wt. plus alemtuzumab or antithymocyte globulin. These regimens are certainly less toxic but also possibly less efficient than the more radical ones (such as with BEAM -carmustine/etoposide/araC/melphalan) [39,40], and they induce immunomodulation by changing the balance between T regulatory and other T cells.

Autologous versus allogeneic BMT
Few anecdotal case reports indicate that allogeneic HSCT may have even higher (that autologous) efficacy in autoimmune diseases and MS [53,60,61]. This is a logical assumption based on the possible genetic predisposition to autoimmunity in the affected individuals and the fact that in AHSCT, the immune system is restored from the pool of progenitors, which carry the same genome, including the putative genetic defects that led to autoimmunity in the first place. On the other hand, immunoincompetence is proved to be longer in allogeneic HSCT than in AHSCT. The reasons for this delayed immune reconstitution include less successful donor-host interaction, which is required for competent immune response and the need for immunosuppression in order to suppress both host-versus-graft reaction and GVHD. Prolonged immunocompetense after allogeneic HSCT is associated with more toxicity, higher incidence of side effects (mainly infections) and higher mortality rates (especially following allogeneic HSCT). Moreover, it is often difficult to distinguish symptoms of relapse of MS/autoimmunity following allogeneic HSCT since the clinical features of chronic GVHD may sometimes be similar to those of other autoimmune diseases and even MS (although such an MS-like presentation is very rare) [62]. For instance, the first descriptions of chronic GVHD were based on the resemblance with signs of autoimmune diseases such as Sjogren's disease, systemic sclerosis and primary biliary cirrhosis. There are recent reports of altered immunity even years postallogeneic HSCT with clinical and laboratory characteristics of autoimmune activity [63].

Other protocols of HSCT
The introduction of reduced intensity conditioning (RIC) or nonmyeloablative conditioning transplants [64] has revolutionized clinical HSCT, reducing the period of immuno-incompetence, expanding the use of HSCT to patients with comorbidities and abolishing age limits. With this approach, the toxicity of HSCT is lower (no complete immunoablation; no use of TBI) and GVA effects may be induced as well in the case of allogeneic HSCT [65,66]. In our center, two patients, were treated with such allogeneic 'mini-transplant' protocols (nonmyeloablative HSCT, using intermediate, nonmyeloablative doses of fludarabine and cyclophosphamide) with an impressive response. The first patient was a 42-year old male patient suffering from secondary progressive MS of 8 years duration, with EDSS of 6.5 and severe progression (from EDSS of 5.5 to 6.5) during the year preceding the treatment, not related to a relapse; the second, was a 48-year old male patient with secondary progressive MS, disease history of 12 years and progression of one degree in the EDSS scale during the previous year (EDSS at inclusion: 6.5). They are both now, 10 years after the procedure, free of any disease activity (no relapses, no progression and no new lesions in the MRI) and with mild improvement in their neurological function, despite their initial high disability score (EDSS of 6.5). Both patients have chronic mild GVHD (only skin involvement) and one treated, since the transplantation, with low dose (1-2 mg/day) tarcolimus. However, the overall clinical experience during the last decade with reduced-intensity conditioning HSCT in hematological malignancies has not demonstrated significant superiority of this protocol. It seems therefore, that for the meantime 'conventional' AHSCT should be preferably used for patients with severe MS and other autoimmune diseases. If there are syn geneic twins, these may be the best candidates for donors. Allogeneic HSCT, which is associated with higher toxicity/mortality but also with possibly better clinical outcomes, should be considered in younger patients with high disease activity and low disability scores.

Expert commentary
Cumulative experience from the last 15 years has proven the efficacy of autologous HSCT in MS, especially in terms of suppression of inflammation and the relapses of the disease. The effect of HSCT was more pronounced in young patients and those with 'malignant' types of MS and high inflammatory activity. However, despite the strong suppression of the inflammatory process, disability may continue to accumulate in some of the treated patients and histopathological studies have shown new demyelination and new axonal loss and atrophy in HSCT-treated MS patients. One may argue that in most of the case studies in the literature that indicated such continuation and/or recurrency of disease progression, the histopathologies/autopsies were obtained within few weeks to few months after the procedure and therefore, it cannot be totally excluded that the detected activity of MS had already started before the transplantation. Such relapses, post-HSCT, could be explained either by changes in the autoantigen profile (epitope spreading) or by a de novo developing autoimmune process. Additional causes for disease recurrency might include (but are not limited to) insufficient immune cytoreduction, reintroduction of autoreactive lymphocytes in the stem cell graft or persistent immunogenetic risk factors.
Histopathological studies did not reveal any lymphocytic infiltrations (neither T, nor B cells) and the only evidence of inflammation was related to microglia and macrophages. Moreover, in the case of reported deterioration following allogeneic HSCT for hematological malignancy, the findings may be in part attributed to GVHD, especially since in this case as contrasted to previous reports with autologous transplantation, high numbers of lympho cytic infiltrations were seen in the CNS, that may indicate a novel, possibly GVHD-related autoimmune process.
The treatment-related mortality of autologous (and certainly more of allogeneic) HSCT is not negligible and it ranges between 1-5%. It is clear that there is a strong reverse association of mortality rates with the specialization level of the center in which the transplantation was performed. The procedure therefore, should be performed in specialized centers, in which mortality is significantly lower. Nonmyeloablative protocols, which are associated with less toxicity, are preferable. Allogeneic HSCT may be considered in selected cases.
In conclusion, it is important to emphasize that HSCT does not represent a cure for MS but may be considered as a second-or third-line therapeutic option, especially in young patients with severe disease, high inflammatory activity and low disability scores. Controlled studies over regular immunosuppressive protocols would definitely provide more concrete answers concerning the efficacy and the place of HSCT in the management of MS.

Five-year view
Over the next 5 years it is expected that further refinement of HSCT technology will improve the safety of the procedure and reduce the treatment-related risks. Novel therapeutic techniques involving allogeneic or semi-allogeneic HSCT using nonmyeloablative conditioning, are expected to emerge and be applied in several centers. Such modified protocols may increase the beneficial effects and reduce the procedure-related mortality and morbidity.
The differences in trial design that led to varied outcomes in HSCT underline the need for harmonization of end points using suitable clinical and surrogate markers (novel MRI and electrophysiological techniques) to evidence neuronal regeneration and restoration of neurological dysfunction. It is expected that future studies will utilize such end points. To our view, HSCT will still be considered during the next 5 years as a second-or third-line therapeutic option, especially for young patients with severe disease, high inflammatory activity and low disability scores who do not respond to the conventional immunomodulatory treatments.

Financial & competing interests disclosure
The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.
Paid writing assistance was provided by Alexandra Mahler.

Key issues
• More that 500 patients with multiple sclerosis (MS) and more than 1000 with other autoimmune diseases have been treated with hematopoietic stem cell transplantation (HSCT) worldwide. • The mostl used protocols (75%) include cyclophosphamide with granulocyte colony stimulating factor for mobilization and BEAM for conditioning (in 50%). • Post-HSCT mortality is reported to range between 1-5%, and has been lower during the last few years as compared with 15 years ago; toxicity is generally influenced by the patient's age, by existing comorbidities and by the experience of the treating group. • Cumulative data of the 10-year follow-up of the treated MS patients reveal an overall survival rate of 90%. All studies have uniformly reported a significant reduction or even disappearance of the inflammatory activity in MRI, which is retained by time. The reported disease progression-free survival rate at 3 years ranges between 60-80%; at 10 years, the rate is approximately 65% for patients with secondary progressive MS and 40% in primary progressive MS. All studies have consistently reported a strong reduction in the annual relapse rate, which was frequently accompanied by an improvement in neurological signs and disability. Patients with more active disease, those at younger age (<40 years), and those with low disability scores are more likely to benefit from HSCT (67% response rate in relapsing MS vs 31% in progressive MS). Patients with long standing disease and those with a primary progressive course (and a prominent neurodegenerative component), may not benefit from HSCT. • Recurrency of MS, relapses of the disease and continuation of the accumulation of disability, demyelination and axonal loss, have been reported in some cases, with a persistence of demyelination and of the presence of oligoclonal antibodies in the cerebrospinal fluid. This could be related to either a development of de novo autoimmunity or changes in autoantibodies/T-cell reactivity profile (epitope spreading) but it may as well be attributed to previous activity of MS, which possibly started before the transplantation. It apparently appears, that even when strong suppression of the inflammatory process in MS is achieved with HSCT, still demyelination, axonal damage and accumulation of disability may not be completely stopped. The latter may also indicate that the main two pathogenetic processes in MS (inflammation and degeneration) are not always in close correlation. • Case report neuroimaging studies have indicated that brain atrophy seems to continue and even accelerate following HSCT. However, this finding may be related to a 'pseudoatrophy' caused by the resolution of edema in the areas of brain lesions and it slows down after the second post-HSCT year. • Immuno-incompetence is considerably longer following allogeneic rather than autologous HSCT. The risk for relapse is lower after allogeneic HSCT but the mortality and toxicity are higher. In the meantime, autologous HSCT should be preferably used as a second-or third-line therapeutic option, especially in young patients with severe disease, high inflammatory activity and low disability scores. If there are syngeneic twins, these may be the best candidates for donors. Allogeneic HSCT may be considered in selected cases. The use of nonmyelobalative protocols may also be considered. • Controlled studies, which are still lacking should be performed to evaluate the possible superiority of HSCT over regular immunosuppressive protocols.