Cell-based therapeutic strategies for multiple sclerosis.

The availability of multiple disease-modifying medications with regulatory approval to treat multiple sclerosis illustrates the substantial progress made in therapy of the disease. However, all are only partially effective in preventing inflammatory tissue damage in the central nervous system and none directly promotes repair. Cell-based therapies, including immunoablation followed by autologous haematopoietic stem cell transplantation, mesenchymal and related stem cell transplantation, pharmacologic manipulation of endogenous stem cells to enhance their reparative capabilities, and transplantation of oligodendrocyte progenitor cells, have generated substantial interest as novel therapeutic strategies for immune modulation, neuroprotection, or repair of the damaged central nervous system in multiple sclerosis. Each approach has potential advantages but also safety concerns and unresolved questions. Moreover, clinical trials of cell-based therapies present several unique methodological and ethical issues. We summarize here the status of cell-based therapies to treat multiple sclerosis and make consensus recommendations for future research and clinical trials.

Introduction

With multiple approved disease-modifying therapies (DMTs), there is a broad range of options to treat relapsing-remitting multiple sclerosis (Ingwerson ). However, less progress has been made in the treatment of progressive forms of the disease (Shirani ). While the positive impact of treatment on reducing the frequency of relapses and accrual of relapse-related disability has been demonstrated, none of the currently available agents halt disease progression or directly promote repair of pre-existing CNS damage. Moreover, all of the approved therapies have potential adverse events that may compromise safety or adherence. All are expected to be ongoing life-long therapies as long as they remain safe and effective. Consequently, there is an imperative for new therapies that (i) are more effective in relapsing-remitting multiple sclerosis, particularly for patients with highly active disease who are at substantial risk for future disability; (ii) are effective in slowing or preventing progression; (iii) have the potential to reverse disability; and (iv) can be used safely with fewer delivery and adherence concerns.

Cell-based therapies have generated substantial interest as potential approaches to address these gaps by working through various mechanisms: regenerating the defective immune system that underlies multiple sclerosis by immunoablation followed by autologous haematopoietic stem cell transplantation (I/AHSCT); modifying both immune reactions and endogenous repair mechanisms using mesenchymal stem cells (MSCs) and other stem cells from bone marrow, adipose tissue, placenta, or other tissues; pharmacologic manipulation of endogenous stem cells to enhance their reparative capabilities; or replacing damaged or lost myelin-making oligodendrocytes by transplantation of oligodendrocyte progenitor cells (OPCs) or OPC-like inducible pluripotent stem cells (iPSCs) (Sarkar and Scolding, 2016).

In this review, we discuss the biology and potential utility of these cell-based therapeutic approaches in multiple sclerosis; summarize the progress made to date on testing in multiple sclerosis; discuss practical, scientific, clinical, regulatory, and ethical concerns; and make recommendations for future studies to move this therapeutic area forward. The review is based on an extensive literature search related to cell-based therapies for multiple sclerosis and on discussions at a consensus workshop, the International Conference on Cell-Based Therapies for Multiple Sclerosis, held 19–21 November 2015 in Lisbon Portugal under the auspices of the International Advisory Committee for Clinical Trials in Multiple Sclerosis (see Appendix 1 and Supplementary material for a list of conference participants).

Immunoablation followed by haematopoietic stem cell transplantation

Biological background and rationale

The rationale for I/AHSCT to treat multiple sclerosis is depletion of autoreactive effector cells with immunoablative agents (the conditioning regimen) followed by infusion of autologous haematopoietic stem cells to support immune system reconstitution with more normal immune function (Muraro ; Muraro and Abrahamsson, 2010). Analysis of circulating lymphocytes after I/AHSCT demonstrates reduction of circulating autoreactive effector T cells, predominantly Th17 rather than Th1, and emergence of recent thymic emigrants post-transplant, restoring a more regulatory milieu (Muraro , 2014; Darlington ; Arruda ). The degree of reconstituted T cell repertoire variability is related to the intensity of the conditioning regimen (Muraro ). Muraro reported the presence, pre-transplant, of circulating mucosal-associated invariant T cells (MAITs) characterized by a CD8+, CD161high phenotype (Abrahamsson ). These MAITs exert a pro-inflammatory effect by promoting production of several cytokines thought to be associated with the pathogenesis of multiple sclerosis, including interferon-gamma and interleukin-17 (Lovett-Racke ). After transplantation, there was a significant reduction of this population in the peripheral blood in parallel with an increase of regulatory CD4+, CD25high, CD127-, FoxP3+ T cells. The initial benefit of I/AHSCT probably results from this and other comparable alterations in immune function. However, some studies have detected re-emergence of autoreactive effector cells despite a high intensity conditioning regimen and persistence of efficacy (Darlington ). Thus, the mechanisms responsible for sustained benefit of I/AHSCT are less well understood.

Practical/procedural background

Appropriate patients for I/AHSCT

Recognition of patients with multiple sclerosis most likely to benefit from I/AHSCT has evolved. Initial studies mainly enrolled patients with longstanding severe progressive multiple sclerosis, when inflammatory features are less prominent and neurodegeneration is the main underlying mechanism (Trapp and Nave, 2008). Benefit generally was modest, although some patients exhibited sustained slowing or stabilization of disability, but improvement in neurologic function was rarely seen (Burt ; Mancardi ). Also, patients with more severe neurologic disability had increased risk of adverse events (Mancardi and Saccardi, 2008). More recent studies (Table 1) focused on relapsing-remitting multiple sclerosis and demonstrated that patients with active inflammatory features appear to derive the most benefit from this approach (Burt ; Saccardi ; Muraro ). As a result, the current recommendation is for studies of I/AHSCT to enrol patients with highly active relapsing-remitting multiple sclerosis reflected by clinical relapses and MRI lesion activity, time from diagnosis within 5 years, and suboptimal response to available regulatory-approved DMTs (Burt ; Saccardi ). These criteria apply to only a limited subset of patients with multiple sclerosis but help define those at high risk for future disability despite available therapy. These recommendations have been somewhat controversial, as they suggest a relatively aggressive therapeutic approach for patients who may have little established disability (Soelberg Sorensen, 2016).

Table 1

Contemporary clinical trials of I/AHSCT for multiple sclerosis

Study and study period	Design	n	Study population	EDSS median (range)	Mobilization	Conditioning regimen	Graft manipulation	Results
Italian (Mancardi et al., 2012) 1996–2008	Multicentre case series	74	RRMS n = 33, SPMS n = 41	6.3 (3.5–9.0)	Cy/GF	BEAM/ATG	None	PFSc 66% at 5 years
CIBMTR/EBMT (Muraro et al., 2013) 1995–2006	Multicentre case series	281	RRMS n = 45, progressive MS n = 236	6.5 (1.5–9.0)	chemo/GF 93%	Intensity: high 19%, intermediate 64%, low 27%	None 56%	PFSh 48% at 5 years
Swedish (Burman et al., 2014) 2004–2013	Multicentre case series	48	RRMS n = 34, progressive MS n = 7	RRMS: 2.5 (0–6.5), progressive MS: 6.5 (5.0–7.5)	Cy/GF	BEAM/ATG n = 41, Cy/ATG n = 7	None	DFSf 68% at 5 years
Northwestern (Burt et al., 2015) 2003–2014	Single centre case series	145	RRMS n = 118, SPMS n = 27	2.0 (3.0–5.5)b	Cy/GF	Cy/ATG, Cy/Alemtuzumab	None	64% improvement of EDSS at 4 years
ASTIMS (Mancardi et al., 2015) 2004–2009	Multicentre, randomized, open-label, active comparator, phase 2	24	RRMS n = 9, SPMS n = 12	6.0 (5.5–6.5)	Cy/GF	BEAM/ATG	None	79% reduction in new T2 MRI lesions
HALT-MS (Nash et al., 2015) 2006–2009	Multicentre, single-arm, phase 2	21	RRMS n = 21	4.5 (3.0–5.5)	Steroids/GF	BEAM/ATG	CD34+selection	EFSa 78% at 3 years
Russian (Shevchenko et al., 2015) 2005–2011	Single centre case series	99	RRMS n = 43, progressive MS n = 56	3.5 (1.5–8.5)	Steroids/GF	BEAM-like/ATGd	None	EFSe 88% at 3 years
Canadian (Atkins et al., 2016) 2001–2009	Multicentre, single-arm, phase 2	24	RRMS n = 12, SPMS n = 12	4.3 (3.0–6.0)	Cy/GF	Busulfan, Cy, ATG	CD34+ selection	EFSi 69.6% at 3 years

aEFS included: progression (increase of 0.5 in EDSS score compared to baseline, starting 6 months post-transplant and confirmed 3 months later), relapse (worsening or development of new neurologic sign and corresponding symptom lasting more than 48 h), MRI evidence of disease progression (two or more independent multiple sclerosis-related lesions gadolinium-enhancing or T2-hyperintense lesions on brain MRI performed 1 year or more after transplant) or death.

bRepresents interquartile range, the manuscript reported that 15 patients had the EDSS > 6.0.

cEFS: clinical and development of new brain lesions by MRI.

dBEAM-like: BEAM with dose reductions of etoposide, cytarabine and melphalan or the combination of carmustine and melphalan.

eEFS from the Russian study was freedom from progression (worsening of at least 0.5 points in EDSS for two consecutive assessments 3 months apart) or relapse (acute deterioration of neurologic function more than 24 h without other causes).

fDFS in the Swedish study classified events as relapse, new MRI manifestations, EDSS progression or death.

gDifferent progressive forms, the majority of participants had secondary progressive multiple sclerosis.

hPFS events included worsening of EDSS or initiation of new disease-modifying agent.

iEFS included clinical relapse, new or enhancing MRI lesion, or sustained EDSS worsening.

ASTIMS = autologous stem cell transplantation international multiple sclerosis; chemo = chemotherapy-based mobilization; CIBMTR = Center for International Blood and Marrow Transplant Research; DFS = disease-free survival; EBMT = European Blood and Marrow Transplant Group; EFS = event-free survival; GF = growth factors; HALT-MS = Hematopoietic cell autologous transplant for multiple sclerosis; MS = multiple sclerosis; PFS = progression-free survival; progressive MS = combined progressive forms of multiple sclerosis; RRMS = relapsing-remitting multiple sclerosis; SPMS = secondary progressive multiple sclerosis.

An important determinant of transplant success is the ability of patients to tolerate the conditioning regimen. Disease-related factors not only affect efficacy but also tolerability. In cancer patients, those with more advanced disease, either with active cancer at time of transplant or refractory to prior therapy, have a higher failure rate. This is not only due to inability to control the disease with higher doses of chemotherapy, but also the increase in transplant-related morbidity or mortality from the cumulative effect of prior treatments. Similarly, multiple sclerosis patients with more severe disability or progressive disease also tend to have higher rates of transplant-related morbidity and mortality (Mancardi and Saccardi, 2008; Muraro ). The effects of prior multiple sclerosis DMTs on efficacy or safety of I/AHSCT are unknown.

In allogeneic haematopoietic stem cell transplantation for haematologic malignancies, the overall impact of the cancer on general health (the estimated the ability to work, perform activities of daily living and the need for hospitalization) is correlated with transplant outcome; lower performance scores are associated with higher post-transplant mortality. Similarly, the presence of key comorbid conditions also impacts transplant outcome. A high score on the Haematopoietic Cell Transplantation Comorbidity Index, which includes 17 items comprising past medical history (stroke, myocardial infarction, arrhythmia, autoimmune disease, prior solid tumours), end organ function (pulmonary, hepatic, renal and cardiac), and weight (obesity), is associated with increased post-transplant mortality (Sorror , 2015; Elsawy and Sorror, 2016). Although comorbidities are less common in younger patients with multiple sclerosis, they have an important impact on multiple sclerosis disease outcomes (Marrie ). Their effects on the efficacy or safety of I/AHSCT to treat multiple sclerosis have not been explored.

Transplant procedure

I/AHSCT should be viewed as a multi-step process that leads to a combined therapeutic effect in multiple sclerosis. Adverse effects also can occur at each step. The typical sequence includes mobilization of peripheral blood haematopoietic stem cells, immunoablation via administration of a conditioning regimen, then infusion of haematopoietic stem cells to promote haematologic reconstitution.

Granulocyte colony stimulating factor (G-CSF) or granulocyte-macrophage colony stimulating factor (GM-CSF) are often used alone to mobilize haematopoietic stem cells from the bone marrow to the peripheral blood in healthy volunteer donors in an allogeneic donor transplant setting, but may cause worsening of neurologic manifestations in multiple sclerosis, either accentuation of pre-existing symptoms due to fever or bona fide relapses (Openshaw ). Therefore, in multiple sclerosis, the most common approach is administration of cyclophosphamide (Cy) plus G-CSF as a mobilizing agent, which helps deplete lymphocytes that will eventually be collected in the graft and lessens the chance of reinfusion of autoreactive T cells. Administration of Cy also may contribute to the therapeutic effect and adverse effects of I/AHSCT.

There are clear distinctions between the conditioning regimens used to treat malignancies and for non-malignant diseases such as multiple sclerosis. In the setting of AHSCT for malignancies, the conditioning regimen consists of high doses of chemotherapy to maximize disease control; the primary intent of subsequent haematopoietic stem cell infusion is to ‘rescue’ haematopoiesis. In the setting of allogeneic haematopoietic stem cell transplantation, the conditioning regimen can be classified within an intensity spectrum according to the type of chemotherapeutic agents and/or radiation selected, and their respective doses. High intensity regimens, also termed myeloablative, require haematopoietic stem cell infusion to prevent irreversible bone marrow damage. At the opposite end of the spectrum are lower intensity regimens, also termed non-myeloablative, which minimally affect haematopoiesis. Intermediate intensity regimens, also termed reduced intensity, fall in the middle of this spectrum. The indication, type of transplant, and population being treated determine the choice of regimen intensity. For example, in some instances for the same indication and transplant type, a high intensity regimen will be selected for patients younger than 65 years and a reduced intensity in patients older than 65 years. In general, high intensity regimens are selected to better control the cancer and minimize the risk of disease relapse, but are potentially associated with greater morbidity and mortality compared to lower intensity regimens.

In multiple sclerosis, like other autoimmune diseases, the optimal intensity of the conditioning regimen remains uncertain and is actively debated. The main objective is to balance lymphocyte depletion to eliminate pathologic autoimmunity with acceptable morbidity and mortality. The most commonly used conditioning regimen in multiple sclerosis has been BEAM (carmustine, etoposide, cytarabine and melphalan), which is considered an intermediate intensity regimen, combined with anti-thymocyte globulin (ATG) (Mancardi , 2015; Muraro ; Burman ; Nash ; Shevchenko ). High intensity regimens, such as total body irradiation and busulfan, were initially used in multiple sclerosis but were either abandoned or modified because of toxicity. Atkins reported the use of high dose busulfan and Cy (Bu/Cy) with infusion of a T cell depleted (CD34+ cell selection) autologous graft (Atkins ). During the study, several modifications to the busulfan regimen were made to improve tolerability. The route of administration was switched from oral to intravenous, the dose was reduced (though still considered in the high dose range), and the target dose was adjusted based on the first busulfan dose pharmacokinetics. Conversely, Burt and colleagues (2009, 2015) have advocated a low intensity or non-myeloablative regimen—Cy or alemtuzumab followed by ATG—reporting good efficacy in relapsing-remitting multiple sclerosis with reduced toxicity and no mortality.

One potential concern with use of a less intense conditioning regimen is suboptimal multiple sclerosis disease control. One study demonstrated that recipients of non-myeloablative regimens had early reappearance of MRI lesion activity post-transplant (Mancardi ). It is possible that the intensity of the conditioning regimen may need to be tailored to the clinical situation, although consensus on how to identify patients early with aggressive multiple sclerosis and poor prognosis is lacking. Another consideration is whether the specific drug combination within regimens considered the same intensity is associated with differential outcomes. Muraro analysed data on 281 transplant recipients with multiple sclerosis worldwide in a retrospect registry-based study. Conditioning regimens varied greatly and when they were grouped according to intensity, there was no correlation with outcome. This observation may have been due to most patients having progressive multiple sclerosis and the proportion of patients with relapsing-remitting multiple sclerosis was not sufficient to demonstrate differential efficacy. Thus, the optimal regimen, Cy/ATG, BEAM-ATG, or Bu/Cy remains uncertain and, based on the available data, all remain acceptable options.

It remains uncertain whether haematopoietic stem cell transplantation should be considered merely bone marrow rescue or if it contributes to the therapeutic benefit of I/AHSCT. While less intensive conditioning regimens may not necessitate haematopoietic stem cell transplantation, the infusion of haematopoietic stem cells serves two purposes: (i) to reduce morbidity by shortening the duration of pancytopenia; and (ii) to increase benefit by promoting immune reconstitution with broader clonal diversity without auto-reactivity. Characteristics of the graft have received relatively little attention in multiple sclerosis. Some studies have administered a largely unmanipulated graft (Burman ; Burt ; Mancardi ). Other studies have positively selected CD34+ cells ex vivo to remove any residual lymphocytes in the graft (Nash ; Atkins ). This step adds to the technical complexity of the transplant procedure but reduces the risk of reinfusion of potentially autoreactive lymphocytes. If haematopoietic stem cell transplantation is not merely rescue after immune depletion but, in fact, contributes to the efficacy of the procedure, further work is needed to optimize mobilization and graft processing to maximize potency.

Multiple sclerosis-related outcomes

Evaluation of therapeutic outcomes in multiple sclerosis is complex. Assessment of the success of a therapeutic intervention is more difficult than in I/AHSCT for malignant diseases, where transplant-related mortality, all-cause mortality, and/or malignant disease recurrence often are used as outcomes. Because multiple sclerosis is associated with only modest shortening of life-span (Goodin ), transplant-related mortality or all-cause mortality alone are not likely to be an informative efficacy outcome in studies of I/AHSCT in multiple sclerosis.

The outcomes most often used in multiple sclerosis clinical trials are relapses, confirmed worsening of disability measured by the Expanded Disability Status Scale (EDSS) or Multiple Sclerosis Functional Composite, MRI lesion activity and burden, and normalized whole brain volume (Cohen ). As summarized in Table 1, these endpoints have been used in trials of I/AHSCT in multiple sclerosis (Mancardi , 2015; Muraro ; Burman ; Burt ; Nash ; Shevchenko ; Atkins ). Because of the potential risk associated with I/AHSCT and to distinguish its efficacy from that of available highly effective multiple sclerosis therapies, some workshop participants favoured a stringent outcome be utilized in future trials of I/AHSCT in multiple sclerosis, specifically event-free survival with a composite outcome comprising clinical relapses, MRI lesion activity (new/enlarged T2-hyperintense lesions or gadolinium-enhancing lesions), confirmed disability worsening, and normalized whole brain volume. Whether the specific definitions of the outcome components should be those used for so-called ‘no evidence of disease activity’ (NEDA) in previous studies (Havrdova ; Giovannoni ; Kappos ; Rotstein ), or modified, e.g. to account for delayed efficacy in a highly active study population, was not decided at the workshop. Other workshop participants felt that early inhibition of inflammatory activity (clinical relapses and MRI lesion activity) had uncertain relation to long-term disease outcome (University of California San Francisco MS-EPIC Team ) and was not likely to distinguish I/AHSCT from available highly effective DMTs. They advocated focusing primarily on long-term disability accrual. Thus, an important issue, especially in relapsing-remitting multiple sclerosis, is how effectively I/AHSCT alters the long-term disease course, that is, delays or prevents development of progressive disease and disability accrual, compared to available therapies. This determination will require a randomized trial comparing I/AHSCT to DMTs with long-term follow-up. A related question is whether I/AHSCT affects the subsequent response to, or safety of, DMTs administered.

An additional efficacy outcome potentially relevant for clinical trials of I/AHSCT is confirmed improvement in disability, which has been demonstrated in trials of several multiple sclerosis DMTs (Jones ; Phillips ; Hauser ). Similarly, reversal of pre-existing disability also has been reported with I/AHSCT (Burt ; Atkins ). Although it is possible these interventions directly stimulate repair to some degree, it is more likely they unmask intrinsic repair mechanisms by effectively suppressing ongoing inflammatory damage (Chang , 2008, 2012).

Financial cost

The annual costs of multiple sclerosis DMTs range from ∼$50 000 to $70 000 (Hartung ) in the USA. The cost of I/AHSCT is ∼$120 000 in the USA, which is incurred mainly in the first year with minimal direct costs subsequently. Thus, in contrast to DMTs for which cost accrues indefinitely, the financial cost of I/AHSCT is largely front-loaded and may be less expensive overall. However, the procedure is not universally covered by health insurance in the USA, although there is variation across different centres and payors. Approaches in other transplant indications, e.g. myelodysplastic syndromes where Center for Medicare Services established Coverage with Evidence Development, allow coverage of transplant costs with the requirement for systematic and prospective data collection.

Clinical trials to date

I/AHSCT has been the most investigated cell-based therapeutic strategy for multiple sclerosis. Recent clinical studies are summarized in Table 1. Most studies were small or single centre case series with different patient populations, therapeutic protocols, and outcome measures. The published experience mostly comprises uncontrolled studies. The one randomized trial (Mancardi ) used mitoxantrone, an agent that is now largely less relevant as a comparator.

A recent retrospective analysis indicated that treatment with I/AHSCT achieved NEDA based on relapses, MRI lesion activity, and disability worsening in a higher proportion of multiple sclerosis patients (78–83% at 2 years) than reported in trials of the available DMTs (13–46%) (Sormani ). It should be noted that these studies had different patient populations and visit schedules, particularly the frequency of MRI scans, which can have a marked effect on NEDA rate. In addition to potent benefit on clinical measures and MRI lesion activity, I/AHSCT had potent efficacy on normalized whole brain volume loss. Following I/AHSCT there was initial acceleration of, followed by marked slowing after 2 years to levels approximating normal ageing (Roccatagliata ; Lee ).

Any evaluation of the utility of I/AHSCT needs to assess risk of mortality. In multiple sclerosis, mortality related to the disease may occur, but usually many decades after diagnosis. Thus, any therapy with significant risk of mortality will not readily be accepted. Mortality associated with I/AHSCT has decreased over the past two decades (Sormani ). With recent protocols, I/AHSCT is a safer procedure with mortality rates <5% (Muraro ), with some trials reporting no mortality (Burt ). Risks remain associated with the conditioning intensity, which necessitate careful patient selection (excluding participants with significant recent or chronic infection, liver disease, heart disease, etc.) and adequate supportive care during the 2–3-week aplastic phase. Optimal selection of patients, transplant procedure, timing of transplant, and post-transplant care help minimize the risk of transplant-related mortality. Some delayed adverse events occur late after I/AHSCT, but they are uncommon. The principal late adverse event is a 9% risk of a secondary autoimmune disorder within 5 years of I/AHSCT to treat autoimmune disease (Daikeler ). Thus, the front-loading of safety issues with I/AHSCT contrasts with multiple sclerosis DMTs, for which the risk of ongoing therapy accumulates over time related to chronic immune modulation or suppression.

Key questions/issues and recommendations

Workshop participants generally agreed on several consensus recommendations.

In aggregate, the available evidence suggests I/AHSCT has substantial and sustained efficacy in suppressing inflammatory disease activity in multiple sclerosis. However, at present, it remains uncertain where the benefit-risk-cost profile of I/AHSCT places it in the treatment for relapsing-remitting multiple sclerosis relative to other available highly effective DMTs.

Patients most likely to benefit from I/AHSCT are relatively young e.g. 50 years of age or less, with relatively short disease duration e.g. 5 years or less, have active relapsing-remitting multiple sclerosis and accumulating disability but still are ambulatory, and have ongoing disease activity despite DMT. I/AHSCT is unlikely to benefit patients with longstanding progressive multiple sclerosis without recent inflammatory features (clinical relapses or MRI lesion activity).

We recommend a formal, multicentre, randomized phase 3 trial, comparing I/AHSCT head-to-head versus currently available highly effective therapy(ies) in a defined patient population. Issues concerning the trial design were discussed extensively, but further details still need to be determined (Box 1). Nevertheless, there was a substantial interest in the development of and participation in such a trial.

If I/AHSCT is performed to treat individual patients in clinical practice, comprehensive safety and efficacy data should be collected, the outcomes submitted to existing registries such as the Autoimmune Disease Working Party of the European Society for Blood and Marrow Transplant (EBMT) (Autoimmune Disease Working Party 2016) and the Autoimmune Diseases and Cellular Therapies Working Committee of the Center for International Blood and Marrow Transplant Research (CIBMTR) (Center for International Blood and Marrow Research 2016), and the results published. However, it is strongly encouraged that efforts be made to enrol such patients into formal clinical trials of I/AHSCT when available.

Enhanced endogenous cell therapy including mesenchymal stem cells

Many stem cell types have potentially beneficial properties unrelated to trans-differentiation and cell replacement. These ‘non-canonical’ properties, some paracrine, may in some disorders play a greater therapeutic role than conventional cell replacement (Korbling and Estrov, 2003). In neurological disease, neural stem cells, MSCs from bone marrow or other sources including adipose tissue, and haematopoietic stem cells have all been shown to have therapeutic potential that depends on such non-canonical properties (Pluchino ; Uccelli ; Rice ). MSCs have attracted the most attention in this regard.

MSCs are present in most (possibly all) tissues (Da Silva Meirelles ; Phinney, 2012). Bone marrow contains various non-haematopoietic stem cells, including MSCs, and MSCs are themselves a heterogeneous population (Phinney, 2012). Within the bone marrow, they function to help maintain the haematopoietic stem cell developmental niche (Mendez-Ferrer ), but it is increasingly clear they also play a significant systemic role in repair in many tissues. In some diseased or damaged tissues, MSC differentiation into cells of the mesodermal lineage contributes to their putative benefit, for example, in liver and cardiac disease. Despite early reports of trans-differentiation into both neurons and oligodendrocytes (Woodbury ), this phenomenon probably does not play a significant role in potential repair-promoting effects of MSCs in the CNS. Rather, their multiple paracrine and other mechanisms of action are more relevant (Box 2), offering the prospect of ameliorating a number of the differing pathological processes contributing to tissue damage in multiple sclerosis through what might be termed ‘enhanced endogenous cell therapy’ (Korbling and Estrov, 2003; Rice and Scolding, 2004).

Some bone marrow cell subpopulations can reside for decades in the human brain after transplantation (Cogle ), though it is not yet clear which. There is no such evidence for MSCs specifically and, in fact, some evidence indicates that MSCs do not persist in tissues (von Bahr ). Thus, potential therapeutic benefit of MSC transplantation is likely to be self-limited, suggesting repeated administration will be necessary.

Source and cell production

The question of whether MSCs from different tissue sources are identical in all properties is not wholly resolved (Strioga ; Li ). This uncertainty is partly a consequence of the continuing absence of any unique identifying marker of MSCs. They are consequently defined by a range of properties (Dominici ). So, the argument becomes almost circular: do cells with this same defining range of properties have identical additional, non-defining properties? There may be potentially important differences between MSCs derived from different sources, e.g. adipose tissue and bone marrow, and these differences may, in the future, influence the choice of tissue source.

The optimal dose of MSCs in any therapeutic use remains unknown, but a common target is 1–2 × 106 cells/kg body weight—a number that makes using primary MSCs near impossible for clinical use. Therefore, protocols for expanding cells are widely used (Mosna ), though it has become clear that cycles of expansion significantly attenuate many reparative and neuroprotective properties. In addition, the typical yield limits repeat dosing. New approaches to expansion therefore continue to be explored (Hoch and Leach, 2014).

Some studies used mixed/unseparated cells (Rice ); others administered purified and culture-expanded MSCs (Karussis ; Yamout ; Bonab ; Connick ; Cohen ). Some authors have studied modified MSCs adapted to express and secrete particular neurotrophins, though less so in multiple sclerosis models. Others have pre-differentiated MSCs—for example, using the classic neural stem cell mitogenic combination of epidermal growth factor and basic fibroblast growth factor, combined with ‘neurosphere’ culture techniques, to produce cells with MSC-derived neural stem cell properties (Harris ). Lack of comparative studies of different cell products and of in vitro markers that relate to therapeutic efficacy preclude recommendations on the optimal cell production protocol.

As in all clinical cell therapy endeavours, there is need for rigorous and stringent quality and safety control in cell production with, in the case of artificially expanded cells, assessment of phenotype and karyotype, mutagenesis testing, and microbiological analysis (Dominici ; Mosna ). This safety aspect represents one significant reason for caution in considering patient requests to purchase treatments from commercial clinics, particularly in countries where medical facilities are arguably less well regulated.

Route of delivery

Directly injecting cells into specific lesions would provide little benefit in the diffuse grey and white matter involvement that characterizes multiple sclerosis. Cell therapy delivered systemically (as with any conventional drug) may have more rationale—and is safer. Most studies of MSCs have adopted intravenous injection (Rice ; Odinak ; Connick ; Li ; Llufriu ; Lublin ; Cohen ). Following intravenous injection, many cells are trapped in the lungs, but significant numbers still enter the CNS, become widely distributed, and can remain for decades, as shown in experimental models and in human subjects (Cogle ). Emerging evidence also suggests potential immune-modulating effects result from the interaction of MSCs and immune cells in the lung (Lee ; Odoardi ). Intra-arterial (carotid) delivery of bone marrow-derived MSCs has been explored in multiple system atrophy (Lee ), but not in multiple sclerosis to date. Concerns about micro-embolization have limited enthusiasm for this approach. Intrathecal delivery also has been tested in multiple sclerosis (Liang ; Riordan ; Karussis ; Yamout ). In the absence of a head-to-head comparison study, the optimal route of delivery remains uncertain.

Clinical results to date

Various groups have published small studies exploring feasibility and safety of MSC transplantation in multiple sclerosis (summarized in Table 2). These studies involved differing study populations, cell products, and routes of administration. The results generally supported the feasibility and safety of MSC transplantation in multiple sclerosis, as was expected based on studies in other conditions (Lalu ), including no evidence of ectopic tissue formation (von Bahr ). Transient aseptic meningitis was common with intrathecal delivery (Karussis ). Also, there have been case reports of acute disseminated encephalomyelitis after intrathecal MSC injection (Kishk ); a glioproliferative spinal cord tumour after intrathecal injection of a combination of mesenchymal, embryonic, and foetal neural stem cells (Berkowitz ); and severe visual loss in three patients with age-related macular degeneration after intravitreous injection of adipose tissue-derived stem cells (Kuriyan ). Some uncontrolled studies reported preliminary evidence of benefit on clinical, neurophysiological, or imaging outcomes (Rice ; Connick ; Cohen ).

Table 2

Published studies of MSC transplantation in multiple sclerosis

Reference	Study populationa (n)	Post-treatment follow-up (months)	Cell product	Route of administrationb	Efficacy outcomes	Adverse events (number of patients)
Bonab et al. (2007)	Progressive MS (10) unresponsive to treatment, EDSS 3.5–6.0	13–26 (mean 19)	Autologous, culture-expanded BM MSCs	IT	EDSS change (n): improved (1) stable (4), worse (5); MRI lesions (n): decrease (1), no change (7), increase (2)	No serious adverse events
Bonab et al. (2012)	SPMS (23), PRMS (2) unresponsive to treatment, mean EDSS 6.1	12	Autologous, culture-expanded BM MSCs	IT	Discontinued study (n = 3); EDSS change (n): improved (4) stable (12), worse (6); MRI lesion activity (n): none (15), present (6), no data (1)	Low-grade fever (all), nausea/vomiting (2), lower limb weakness (2), headache (3) No serious adverse events
Cohen et al. (2017)	RRMS (10), SPMS (14); clinical or MRI activity or worsening in the prior year; afferent visual pathway involvement; EDSS 3.5–6.5	6	Autologous, culture-expanded BM MSCs	IV	No benefit on MRI lesion activity; possible benefit on EDSS and whole brain MTR	No serious or severe adverse events
Connick et al. (2012)	SPMS (10), afferent visual pathway involvement, EDSS 5.5–6.5	5.8–10.2 (mean 7.0)	Autologous, culture-expanded BM MSCs	IV	Improvement in visual acuity and visual evoked response latency, increase in optic nerve area on MRI	Transient post-infusion rash (1), bacterial infection (2) No serious adverse events
Karussis et al. (2010)	RR or progressive MS (15 total) unresponsive to treatment, EDSS 4.0–8.0	6	Autologous, culture-expanded BM MSCs	IT (+IV in five)	Improved mean EDSS 6.7 to 5.9, no new enhancing MRI lesions at 6 months	Transient fever (10), headache (10), meningeal irritation and aseptic meningitis (1)
						No major adverse events
Li et al. (2014)	RR or SPMS (23), 13 treated versus 10 controls	12	Allogeneic human umbilical cord- derived MSCs	IV three times over 6 weeks, with corticosteroids	EDSS and relapses significantly improved compared to control group	None reported
Liang et al. (2009)	PPMS (1) unresponsive to treatment, EDSS 8.5	5	Allogeneic human umbilical cord- derived MSCs	IT + IV, with cyclophosphamide	Improved EDSS 8.5 to 5.5, decreased MRI lesion load	None reported
Llufriu et al. (2014)	RRMS (9) five treated versus four placebo, unresponsive to treatment, EDSS 3.0–6.0	12	Autologous, culture-expanded BM MSCs	IV	Non-significant decrease in cumulative number of enhancing MRI lesions	No serious adverse events
Lublin et al. (2014)	Treated (12): RRMS (7); SPMS (5) versus placebo (4): RRMS (3), SPMS (1); clinical or MRI activity or worsening in the prior year; EDSS 1.5–6.5	12	Allogeneic human placenta-derived MSCs	IV, one infusion (low dose) or two infusions (high dose)	No significant change in EDSS or enhancing MRI lesions	MS relapse (1), anaphylactoid reaction (1), superficial thrombophlebitis (1), headache (7), infusion site reaction (6)
Odinak et al. (2011)	MS, course not specified (8)	12	Autologous, culture-expanded BM MSCs	IV, three infusions	EDSS change (n): improved (6), stable (1), worse (1)	No significant adverse events reported
Rice et al. (2010)	Progressive MS with recent relapse (6); EDSS 4.5–6.5	12	Autologous Filtered, non-expanded whole BM aspirate	IV	Multi-modal evoked potential improvement	Transient increase in lower limb spasticity (2), urinary retention (1)
						No serious adverse events
Riordan et al. (2009)	RRMS (3), unresponsive to treatment	3–7	Autologous, non-expanded adipose SVF, allogeneic CD34+ cells, allogeneic MSCs	Two IV SVF infusions, multiple IV + IT infusions of CD34+ cells and MSCs	Clinical improvement	No side effects reported
Yamout et al. (2010)	SPMS (7); EDSS 4.5–7.5	12	Autologous, culture-expanded BM MSCs	IT (lumbar + cisternal)	EDSS change at 12 months (n): improved (5), stable (1), worse (1); MRI at 3 months (n): new/enlarging T2 lesions (5/7), enhancing lesions (3/7)	Transient encephalopathy with seizures (1), cervical and back pain (1) No serious adverse events

aNo control group unless otherwise noted.

bSingle administration and without concomitant multiple sclerosis disease therapy unless otherwise noted.

BM = bone marrow; IT = intrathecal; IV = intravenous; MS = multiple sclerosis; MTR = magnetization transfer ratio; PPMS = primary progressive multiple sclerosis; PRMS = progressive relapsing multiple sclerosis; RRMS = relapsing-remitting multiple sclerosis; SPMS = secondary progressive multiple sclerosis; SVF = stromal vascular fraction.

The consensus among workshop participants was that further clinical trials were warranted. Larger, controlled phase 2 studies of both unseparated, non-expanded bone marrow-derived cells (Rice , b) and purified, culture-expanded MSCs [MEsenchymal StEm Cells for Multiple Sclerosis (MESEMS)] (Freedman ) are underway.

Workshop participants identified several methodological issues concerning MSC transplantation in multiple sclerosis (Box 3).

Cell numbers and types

Cell dose currently is entirely empirical; there is little or no evidence indicating how many cells might be optimal. Similarly, there is no clear evidence whether multiple infusions would be needed, though this appears intuitively likely, and exploratory trials are underway (Rice ). A further issue is the type of cell preparation to be used. Most studies used mixed mononuclear cell preparations or purified culture-expanded MSCs; some investigators studied MSC-derived neural progenitors (Harris ). MSCs are much studied experimentally, and there is an attractive rationale in using a purified homogenous cell population (Freedman ). However, bone marrow mononuclear preparations include many cell types, and there is good experimental evidence of benefit of such mixed preparations in repairing demyelination (Akiyama ) and suppressing inflammation, as well as clinical evidence in patients with stroke and other diseases (Savitz ). We do not know which cells among the bone marrow population are the most valuable therapeutically; there is no known benefit from excluding cell populations. Indeed, some evidence points to the superiority in certain experimental situations of using mixed mononuclear preparations over purified MSCs. The former simpler approach requires fewer technical resources and avoids the potential risks of genetic instability (Miura ), infection (Uhlin ), and altered phenotype that may accompany multiple cell cycling in culture for expansion. At present, it appears reasonable for both approaches to be pursued.

It is unclear whether the underlying biology of multiple sclerosis might affect MSC function. Some studies have demonstrated similar growth in culture, differentiation potential, surface antigen expression, and immunomodulatory properties of MSCs isolated from multiple sclerosis subjects versus non-multiple sclerosis controls (Papadaki ; Mazzanti ; Mallam ; Kassis ). Other studies reported notable functional differences (de Oliveira ; Redondo ; Sarkar ), suggesting autologous cell transplantation might not be appropriate. Commercial studies using pooled culture-expanded (heterologous) MSCs in non-multiple sclerosis conditions have, thus far, had disappointing efficacy results. Issues of donor variance, immunogenicity, culture expansion, epigenetic reprogramming, senescence, and (perhaps particularly) cryopreservation and thawing (Francois ; Chinnadurai , 2016) may have contributed to these negative results.

Consensus recommendations

MSC and related cell therapy is an active area of research. Several phase 2 trials are already underway (Freedman ; Rice ), which should clarify whether this approach is a potentially efficacious treatment for multiple sclerosis and in what phase of the disease. Workshop participants agreed it remains important to monitor carefully for long-term adverse effects, perhaps through international registries.

If the concept that underlies MSC therapy proves to be of benefit in diseases like multiple sclerosis, it may be that enhanced endogenous cell therapy will ultimately give way to molecular treatments. If the main beneficial effects of MSC therapy are paracrine, these might be more conveniently reproduced by directly using the principal effectors elaborated by infiltrating cells. The problem, however, is that not all the multiple therapeutic capacities of MSC-related populations are uniformly activated in all disease situations. Rather, infiltrating cells probably ‘sense and react’, with specific pathways triggered in response to the tissue and form of tissue damage (Murphy ). This process is likely dynamic, with the profile of administered cells and of those infiltrating tissues evolving with the progress of each individual disease or injury. It may be challenging to reproduce this by administering molecules rather than cells.

Pharmacological manipulation of endogenous repair mechanisms

The traditional approach to develop remyelination-promoting pharmacologic therapies begins with basic studies of myelination and remyelination followed by development of agents that augment these processes. Such studies have identified a sizable number of candidate therapeutics (Kremer ). When the agents are novel, e.g. the anti-LINGO-1 monoclonal antibody opicinumab (Mi ), they then must go through the standard regulatory approval process of preclinical studies then phase 1–3 clinical trials (Mi ). However, a sizable number of already existing medications also may have the ability to promote remyelination (Kremer ); ‘repurposing’ these agents could expedite testing and regulatory approval. A complementary molecular approach involves using cultured OPCs or OPC-like iPSCs as the basis for high-throughput screening of libraries of already available drugs for their ability to stimulate remyelination (Deshmukh ; Mei ; Najm ). Molecules identified in the initial screens were further evaluated by increasingly stringent in vitro and in vivo testing, identifying the muscarinic antagonist benztropine, the antihistamine clemastine, the imidazole antifungal miconazole, and the topical steroid clobetasol as potential candidates for further testing (Deshmukh ; Mei ; Najm ).

Agents identified through the approaches described above are anticipated to have a wide range of mechanisms of action and pharmacological properties. Therefore, the design of proof-of-principle clinical trials in multiple sclerosis will vary according to the agent under study.

A pilot study of clemastine fumarate to promote inherent remyelination showed improvement on visual evoked potentials in participants with multiple sclerosis-related chronic optic neuropathy (Green et al., presented at the 2016 Annual Meeting of the American Academy of Neurology). The RENEW trial of opicinumab in acute optic neuritis demonstrated benefit on visual evoked potential latency recovery in the per-protocol analysis but not in the intention-to-treat analysis or on visual function or optical coherence tomography measures (Cadavid ). The SYNERGY trial of opicinumab in relapsing multiple sclerosis did not demonstrate benefit on the primary endpoint, percentage of participants with confirmed improvement on a composite outcome measure comprising EDSS, timed 25-foot walk, 9-hole peg test, and paced auditory serial addition test (Cadavid ).

A number of studies of such agents are underway or planned. The key question is whether medications identified through basic studies of myelination/remyelination or high throughput screening, in fact, promote remyelination in patients with multiple sclerosis and how to demonstrate it. Some of the theoretical problems applying to the potential use of remyelinating cells, as outlined below, also apply—for example, the question of whether degenerated axons can support remyelination.

Oligodendrocyte progenitor cells and induced pluripotent cells

Cell-based remyelinating strategies have long been of interest as a potential therapeutic approach in progressive multiple sclerosis. Glial progenitor cells expressing A2B5 but not polysialylated neural cell adhesion molecule can be isolated from foetal human brain (Windrem ), and when injected intracerebrally into hypomyelinating shiverer mice mediate widespread myelination and reversal of the clinical phenotype (Windrem ). These cells can be further purified by selecting for expression CD140a and platelet-derived growth factor alpha receptor, which yields the more potent myelinogenic fraction (Sim ). Similarly, iPSCs might be used as a source of oligodendrocytes, with the potential for autologous cells to be used (Thiruvalluvan ). There are a number of potential therapeutic targets for such cells (Goldman ), including genetic dysmyelinating disorders, traumatic brain and spinal cord injury, and acquired inflammatory demyelinating disorders such as multiple sclerosis.

Because OPCs are not expected to be capable of trafficking from blood or CSF into the CNS parenchyma, it is assumed that direct injection will be necessary for the cells to gain access to demyelinated lesions in multiple sclerosis, which introduces an additional level of technical and safety concerns. It appears that the cells have the capacity to migrate substantial distances within the CNS (Goldman ), so it may be possible to inject the cells into selected locations and still obtain widespread repair.

Another issue is that the current principal source of OPCs is foetal tissue, which provides a limited number of cells, given the finite proliferative capacity of OPCs in culture. Because such cells would be allogeneic, immunosuppression of the recipient is required to prevent rejection. It is reassuring that a previous study demonstrated that corticosteroids, interferon-beta, and azathioprine did not affect OPC proliferation, survival, or migratory capacity (Halfpenny and Scolding, 2003).

For the reasons outlined, generation of OPCs from iPSCs generated from the recipient is an attractive alternative approach. However, some studies of human iPSCs have detected frequent genetic modifications, including aberrant DNA methylation and mutations in genes implicated in cancer (Gore ; Hussein ; Lister ), raising the possibility of aberrant tissue formation or malignant transformation after transplantation. To design cell lines safe for human use, more research clearly is required concerning the mechanisms leading to genetic alterations in iPSCs. As a result, there will be substantial regulatory hurdles prior to human testing of such cells. These issues provide much of the impetus for identifying agents that stimulate remyelination by acting through intrinsic OPCs rather than relying on administration of exogenous cells.

A phase 1 trial currently is planned by the New York State Consortium for Cell Therapy to evaluate the feasibility and safety of intracerebral injections of escalating doses of OPCs into multiple locations at a single time point in patients with secondary progressive multiple sclerosis (Goodman, 2016). To prevent transplant rejection, participants will receive tacrolimus and mycophenolate mofetil for 6 months then mycophenolate mofetil alone. Safety studies will include clinical assessment, laboratory studies, and brain MRI.

The planned phase 1 study of OPC transplantation will focus on feasibility and safety as a prelude to proof-of-principle studies evaluating whether administration of exogenous OPCs augments remyelination in multiple sclerosis. Workshop participants supported further exploration of OPC transplantation in multiple sclerosis, but with some reservations (Box 4). Some studies have demonstrated sizable numbers of OPC in some chronically demyelinated lesions (Chang ), suggesting that lack of such cells is not the cause of inadequate remyelination in multiple sclerosis. Rather, the lack of factor(s) necessary to support and sustain remyelination, the presence of inhibitory factor(s) in the multiple sclerosis lesion environment, and inability of degenerated axons to support remyelination may be the main obstacles. Thus, administration of exogenous cells may not address the need. Even if the proof-of-principle with exogenous OPCs can be demonstrated with reasonable safety, there are several practical issues that need to be resolved, e.g. the appropriate cell dose and patient population most likely to benefit.

Ethical considerations

Cell-based therapies for multiple sclerosis are experimental, and strict adherence to ethical guidelines for human subject research (World Medical Association Declaration of Helsinki) helps preserve patient welfare and the integrity of the research process (Box 5) (Hyun ). Specific guidelines for human embryonic stem cell research broadly applicable to cell-based therapies for multiple sclerosis, including the most recent revision of international guidelines in 2016 (International Society for Stem Cell Research 2016; National Research Council and Institute of Medicine of the National Academies; Daley ), stress the ethics of procurement, derivation, banking, distribution, and use of cells and tissues, including somatic tissues and human totipotent or pluripotent stem cell lines.

Media attention has resulted in some cases of misrepresentation and exaggeration of therapeutic claims for cell-based therapies for multiple sclerosis and other diseases. In the consent process, patients need to be clearly informed that the cell-based therapy procedure is not ‘standard of care’. Background information must include what is known about the procedures they are considering and what the goals are of the study in which they will participate. Close attention must be paid to known safety concerns and the potential for unanticipated adverse events.

The existence of many stem cell clinics around the world has resulted in ‘medical tourism’ by patients who believe they have exhausted other routes of treatment and are willing sometimes to travel great distances to unregulated clinics for cell-based therapies (Lindvall and Hyun, 2009; Gunter ; New York Times, 2016). Patients are usually asked to pay for the care directly, without insurance reimbursement. Often such clinics are—for obvious reasons—based in jurisdictions with less stringent medical regulatory structures, and so there can often be little if any assurance of the expertise, quality of care (or even hygiene), or ethical standards of the provider centre, which is often unwilling or unable to seek more traditional financial support for their ‘research’. Freestanding stem cell clinics, which are as yet largely unregulated, have also opened in Western Europe and the USA, so the issue is becoming more widespread (Turner and Knoepfler, 2016). In fact, there is a proposed change in the law in the USA, the so-called REGROW Act, which would remove the requirement for formal clinical trials for regulatory approval of cell-based therapies. Caution against this change has been urged (Nature Editorial, 2016).

Workshop participants agreed that, at present, all cell-based therapies lack definitive evidence for efficacy in multiple sclerosis and, thus, should be considered experimental and only pursued in rigorous clinical trials and observational studies with the expectation that the results will be published. Clinics offering such therapies should, at minimum, confirm that individuals with appropriate qualifications, training, and experience administer the treatment. There should be a written treatment plan, including how complications will be monitored and managed, that can be reviewed and approved by the treating physician. In the case of I/AHSCT, workshop participants acknowledged that there are rare patients with highly aggressive multiple sclerosis not adequately controlled by available DMTs for whom this approach can appropriately be considered as part of clinical practice. In this case, it should be performed at centres with experience both in the procedure and with managing multiple sclerosis, and the outcomes submitted to existing registries and ultimately published.

Workshop participants considered the complex issues of patient-funded research, in which study participants provide financial support for clinical trials. We acknowledged that this approach might be a potential option to allow progress in the field given the limited availability of grant funding. However, a number of concerns were emphasized. Having participants fund research, particularly pay-to-participate, can accentuate therapeutic misconception; may compromise the scientific merit and integrity of the trial; and should only be undertaken if the trial is reviewed, approved, and monitored by an independent review body, such as an institutional review board or data safety monitoring committee for the protection of human subjects in research (Wenner ).

Future directions

Cell-based therapies in multiple sclerosis have been pursued experimentally for at least four decades (Blakemore, 1977), and the past few years have witnessed considerable progress. Each of the cell-based approaches discussed above has begun to enter clinical trials. Much work remains, however. There is no clear consensus about their relative roles, especially in comparison with available DMTs, the specifics of the procedures, or the most appropriate patient population and study design to demonstrate short- and long-term safety and efficacy.

In spite of these uncertainties, there was agreement among workshop participants that I/AHSCT appears to have potent efficacy in relapsing-remitting multiple sclerosis though with significant safety concerns. The principal question is where I/AHSCT should be placed in the overall treatment for relapsing-remitting multiple sclerosis. Other cell-based therapeutic strategies—MSC or OPC/iPSC transplantation, and manipulation of endogenous stem cells—may be more helpful in patients with progressive forms of multiple sclerosis where degenerative mechanisms predominate, but this hypothesis remains to be confirmed. Thus, all forms of cell-based therapy for multiple sclerosis should be considered experimental at this time. There may be rare patients with highly aggressive relapsing-remitting multiple sclerosis who have failed available therapies for whom I/AHSCT may be justified. Other than these patients, cell-based therapy of multiple sclerosis should be pursued only in rigorous clinical trials. In all cases, comprehensive safety and efficacy data should be collected and submitted to existing registries, with the expectation that the results will be published.

It seems clear that the most efficient approach to cell-based therapeutic trials for a relatively uncommon disease with approved, available DMTs like multiple sclerosis, is not through the independent effort of many disconnected clinical centres, but through the development of stable, inclusive networks of investigators involved in a spectrum of cell-based therapies. Such networks can function beyond the organization of a single multicentre clinical trial and can establish protocols; undertake studies; and importantly set up registries to record transplantation protocols and outcomes. The Autoimmune Disease Working Party of the EBMT and the Autoimmune Diseases and Cellular Therapies Working Committee of the CIBMTR (Center for International Blood and Marrow Research, 2016) maintain longstanding registries of patients with autoimmune disorders, including multiple sclerosis, undergoing I/AHSCT. More recently, networks have been established for MSC transplantation trials (MESEMS) (Freedman ). Given the experimental status of cell-based therapy, we recommend all patients undergoing these procedures either in trials or in clinical practice are recorded in registries. In addition, important biological questions remain for all forms of cell-based therapy. Therefore, well designed mechanistic studies using validated methods for sample procurement and handling also should be included.

Undertaking clinical trials is a costly enterprise, even without the creation and maintenance of networks and collaborative working groups. It is unlikely that any single source can meet all funding needs. Collaborative funding networks for such efforts have been created, for example, the International Progressive Multiple Sclerosis Alliance; the New York State Consortium for Stem Cell Therapy for Progressive Multiple Sclerosis; government agencies such as Immune Tolerance Network of the National Institute of Immunology Allergy and Infectious Disease and National Institute of Neurological Disorders and Stroke, the National Institute of Heart Lung and Blood, the National Cancer Institute, all of the United States National Institutes of Health; and others. Public-private funding consortia will likely be needed to raise sufficient funds to undertake studies and to create and maintain networks and registries. While there is documented interest in such support from private foundations, care needs to be taken to avoid confusion between the concepts of ‘care’ and ‘research’ in the absence of public support and independent oversight. The issues of private stem cell clinics, which ‘sell’ research that is supported by patients undergoing treatment, are fraught with ethical and practical concerns.

Given the interest in all forms of cell-based therapies for multiple sclerosis, and the increasing number of observational and randomized studies that are being done and proposed, there will likely be more opportunities than funds available. Prioritization among opportunities will be important for funders as well as investigators, and encouraging creation of research networks will encourage efficiencies for both scientific progress and expenditure of limited funds.

Concluding remarks

Immunoablation followed by autologous haematopoietic stem cell transplantation appears to have potent and durable efficacy in relapsing-remitting multiple sclerosis though with significant safety concerns. The principal question is where this therapeutic approach should be placed in the overall treatment for relapsing-remitting multiple sclerosis.

There may be rare patients with highly aggressive relapsing-remitting multiple sclerosis who have failed available therapies for whom I/AHSCT may be justified as part of clinical practice. Other than these patients, cell-based therapy of multiple sclerosis should be pursued in clinical trials.

Cell-based therapy—transplantation, mobilization, and pharmacologic manipulation—may be helpful in patients with progressive forms of multiple sclerosis where degenerative mechanisms predominate, but this hypothesis remains to be confirmed.

All forms of cell-based therapy for multiple sclerosis should be considered experimental at this time. When it is pursued, comprehensive safety and efficacy data should be collected and submitted to existing registries, with the expectation that the results will be published. Because important biological questions remain for all forms of cell-based therapy, mechanistic studies should be included.

Click here for additional data file.