The current management of autoimmunity involves the administration of immunosuppressive drugs coupled to symptomatic and functional interventions such as anti-inflammatory therapies and hormone replacement. Given the chronic nature of autoimmunity, however, the ideal therapeutic strategy would be to reinduce self-tolerance before significant tissue damage has accrued. Defects in, or defective regulation of, key immune cells such as regulatory T cells have been documented in several types of human autoimmunity. Consequently, it has been suggested that the administration of ex vivo generated, tolerogenic immune cell populations could provide a tractable therapeutic strategy. Several potentially tolerogenic cellular therapies have been developed in recent years; concurrent advances in cell manufacturing technologies promise scalable, affordable interventions if safety and efficacy can be demonstrated. These therapies include mesenchymal stromal cells, tolerogenic dendritic cells and regulatory T cells. Each has advantages and disadvantages, particularly in terms of the requirement for a bespoke versus an ‘off-the-shelf’ treatment but also their suitability in particular clinical scenarios. In this review, we examine the current evidence for these three types of cellular therapy, in the context of a broader discussion around potential development pathway(s) and their likely future role. A brief overview of preclinical data is followed by a comprehensive discussion of human data.
- Cellular therapies
- tolerogenic dendritic cells
- regulatory T-cells
- mesenchymal stromal cells
- TR1 cells
- rheumatoid arthritis
- type 1 diabetes
- Crohn’s disease
- multiple sclerosis
- systemic lupus erythematosus
- graft versus host disease
- autoimmune thyroiditis
- myasthenia gravis
This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: http://creativecommons.org/licenses/by/4.0/
Statistics from Altmetric.com
- Cellular therapies
- tolerogenic dendritic cells
- regulatory T-cells
- mesenchymal stromal cells
- TR1 cells
- rheumatoid arthritis
- type 1 diabetes
- Crohn’s disease
- multiple sclerosis
- systemic lupus erythematosus
- graft versus host disease
- autoimmune thyroiditis
- myasthenia gravis
The complexity of immune tolerance mechanisms presents abundant opportunities for its breakdown, leading to the development of autoimmunity. In most cases, the precise pathogenesis of autoimmunity remains unknown but the genetic polymorphisms that underpin, for example, rheumatoid arthritis (RA), indicate that antigen presentation, cytokine dysregulation and the regulation of lymphocyte activation all play key roles. Furthermore, the clustering of different autoimmune diseases within families attests to common genetic predisposition and pathogenic mechanisms. However, for most autoimmune diseases, the provoking autoantigen(s) have not been defined and, critically, the predilection for the joint in RA versus the brain in multiple sclerosis (MS) versus the pancreas in diabetes mellitus remains enigmatic. Ultimately, the immune system can be viewed as a delicate balance of activation vs tolerance, with multiple mechanisms acting to maintain homeostasis.
Historically, management of autoimmune disorders involved managing end-organ manifestations such as insulin replacement in diabetes and control of pain and inflammation in conditions such as RA (table 1). During the second half of the 20th century the discovery of glucocorticoids and, subsequently, immunosuppressant medications enabled modification of the autoreactive process with reduced tissue damage and even improved life expectancy in diseases such as systemic lupus erythematosus (SLE). The 21st century has seen the biologics revolution with potent, targeted therapies that neutralise key proinflammatory cytokines or interfere with lymphocytes themselves. And, most recently, potent synthetic signalling pathway inhibitors are providing a further means to modulate immune reactivity.1 Nonetheless, current management options rarely lead to cure, or drug-free remission, and most patients require long-term maintenance therapy to control disease manifestations. For example, in RA, approximately 30% of patients achieve sustained remission, but 50% of these will flare if treatment is discontinued. The proportion that flare is usually higher once patients have moved on to more potent biological therapies.2 Because immunosuppressants downregulate the normal adaptive immune system, it is not surprising that several of the therapies in table 1 are associated with an enhanced infection risk, including opportunistic infections, and the development of malignancy. This is in addition to disease comorbidities and drug-specific side-effects, for example, with chronic glucocorticoids. In extreme cases, haematopoietic stem cell transplantation has been used to treat autoimmunity but, with rare exceptions, this intervention has not proved curative.3 4
The holy grail of treatment for autoimmunity would be the reinstatement of immune tolerance. So-called therapeutic tolerance induction offers the opportunity to ‘reset’ the diseased immune system to a state of immune tolerance, theoretically providing for long-term, drug-free remission.5 While multiple strategies have proven effective in animal models of autoimmunity and transplantation, translation to the clinic has been slow. Multiple explanations have been offered, relating to disease stage, therapeutics employed, and the need for better biomarkers of tolerance, among others. Nonetheless, because of the slow progress with therapeutics that target the immune system, such as biologic drugs and peptides, recent strategies have focused on the use of tolerogenic cells themselves.
Tolerogenic cell types
In recent years, investigators have turned their attention to the ex-vivo expansion or differentiation of ‘tolerogenic’ immune cells, followed by their adoptive transfer, as a potential route to therapeutic tolerance induction. To a large degree, these strategies have been catalysed by advances in bio-manufacturing in general, with robust and scalable processes leading to the efficient manufacture of advanced cellular therapies.6 To date, three main types of tolerogenic cell have been the focus of therapeutic strategies in humans.
Mesenchymal stromal cells
Mesenchymal stromal cells (MSCs) are spindle-shaped, plastic-adherent, progenitor cells of mesenchymal tissues with multipotent differentiation capacity.7 MSCs can modulate innate and adaptive immune cells including dendritic cells (DC), natural killer cells (NK) cells, macrophages, B-lymphocytes and T-lymphocytes. This occurs via both cell-cell contact and paracrine interactions through several soluble mediators including indoleamine-2,3-dioxygenase (IDO), prostaglandin E2 and transforming growth factor β.8–10 These and other mechanisms have been summarised in figure 1. By definition, MSCs can differentiate into bone, chondrocytes and adipose tissue in vitro; they are phenotypically positive for CD105, CD73 and CD90 and negative for haematopoietic markers CD45, CD34, CD14, CD11b, CD3 and CD19.7 11 They do not express Class II MHC molecules unless stimulated by interferons7 and lack costimulatory molecules such as CD40, CD80 and CD86.
Exposure to proinflammatory cytokines IFN-γ, TNF and IL-1β10 and activation by exogenous/endogenous danger signals such as bacterial products and heat shock proteins through Toll-like receptor 3 (TLR3) ‘licenses’ MSCs to become immunosuppressive12; in contrast, activation through TLR4 confers a proinflammatory signature and, under some conditions, TLR3 signals may do the same.12 13 The immunomodulatory functions of MSC include their ability to: inhibit T cell proliferation and promote their differentiation into regulatory T cells (Tregs);14 inhibit the CD4+ T cell induced differentiation of B-cells into plasma cells and directly inhibit B-cell proliferation, differentiation and chemotaxis.15 Although MSCs reside in most postnatal organs and tissues,16 they are readily harvested from bone marrow, adipose tissues, umbilical cord blood and Wharton’s jelly (figure 2).
Tolerogenic dendritic cells (tolDC)
DCs are best recognised for their antigen presenting functions in driving immune responses against pathogens and tumour cells. However, DC also play crucial roles in co-ordinating central and peripheral tolerance processes, such that absent or deficient DC associate with an increased tendency to develop autoimmunity.17 18 Furthermore, in autoimmunity, DC are skewed to a proinflammatory state, producing more proinflammatory cytokines and leading to activation and differentiation of autoreactive T cells.19
Immature DC are usually regarded as tolerogenic, whereas mature DC can exert either tolerogenic or immunogenic functions depending on signals received during maturation from the microenvironment and invading pathogens. For instance, bacterial lipopolysaccharides induce immunogenic maturation of DC by upregulating surface MHC complexes and T cell costimulatory molecules (CD80, CD86),20 21 while schistosomal lysophosphatidylserine, anti-inflammatory cytokines (eg, IL-10) and glucocorticoids induce a tolerogenic phenotype.18 Tolerogenic dendritic cells (tolDC) induce peripheral tolerance by induction of anergy and deletion of T cells,22 blockade of T cell expansion23 and induction of regulatory T cells (Tregs).24 25 Tregs in turn induce the regulatory properties of DC (figure 1). These mechanisms have already been reviewed.26 27
Several methods can be used to produce stable tolDC ex vivo, with limited or no capacity to transdifferentiate into immunogenic DC. Common methods include inhibiting the expression of immune-stimulatory molecules (CD80/CD86 and IL-2)28–30 or stimulating constitutive expression of immunosuppressive molecules such as IL-4, IL-10 and CTLA-4,31–35 through genetic engineering. Also, exposing differentiating DC ex-vivo to drugs such as dexamethasone and vitamin D336 37 or immunosuppressive cytokines such as IL-10 and TGF-β38–40 and lipopolysaccharides41 can be used to produce tolDC. These and other methods have been extensively reviewed elsewhere.42
Regulatory T cells (Tregs)
Tregs are a subset of T cells expressing CD4, CD25 and intracellular Forkhead Box P3 (FoxP3) protein that inhibit the functions of effector T cells as well as other immune effector cells and so are essential for immune tolerance.43 44 They mediate their effects by producing immunosuppressive cytokines and by cell-to-cell contact, following stimulation via their antigen-specific T cell receptors (TCR). These mechanisms also modulate other immune responses in an antigen-non-specific manner through ‘bystander suppression’ and ‘infectious tolerance’.45 46 Treg depletion and dysfunction have been implicated in a variety of autoimmune disorders including type 1 diabetes, RA, SLE and, classically, with an inherited deficiency of FoxP3, immune dysregulation polyendocrinopathy enteropathy X linked syndrome.47 48 These findings support the possibility that ex-vivo expansion and transfusion of autologous or allogeneic Tregs could provide an effective therapeutic strategy for unwanted immunopathology such as autoimmunity.
In the past, the lack of reliable Treg surface markers and the resultant possibility of simultaneously isolating and transfusing proinflammatory T cells slowed the development of protocols for Treg isolation and expansion.5 More recent studies have used CD4, CD25 and CD127 cell surface markers to isolate CD4+CD127lo/-CD25+ Tregs from blood.49 50 Other types of regulatory T cells exist, such as T-regulatory type 1 (Tr1) cells, which secrete IL-10.51 These are a distinct population of regulatory T cells that only transiently express FoxP3, on activation.52 They coexpress CD49b and LAG-3, and secrete high levels of IL-10 but low amounts of IL-4 and IL-17. Suppression is dependent on IL-10 and TGF-β and they kill myeloid antigen-presenting cells via granzyme B release.
Migration of tolerogenic cells
MSCs, Tregs and tolDC express a host of homing receptors that are important for their transmigration from the tissue of administration (eg, skin or vascular system) to activation sites (eg, regional lymph nodes) and, ultimately, to the target organs. For instance, FoxP3+ Tregs express CC receptor 7 (CCR7), CCR4, CCR6, CXC receptor 4 (CXCR4) and CXCR5. They also express CD103 (integrin αEβ7) (whose ligand is E-cadherin expressed by epithelial cells) and CD62L (L-selectin) (whose ligands are the lymph node and mucosal lymphoid tissue endothelial cell addressins CD34, GlyCAM-1 and MAdCAM-1).53 Activated tolDC express CCR7 and migrate to CC chemokine ligand 19 (CCL19),54 underpinning migration to regional lymph nodes. MSCs, on the other hand, express a restricted set of chemokine receptors (CXCR4, CX3CR1, CXCR6, CCR1, CCR7) and have shown appreciable chemotactic migration in response to the chemokines CXC ligand 12 (CXCL12), CX3CL1, CXCL16, CCL3 and CCL19.55 MSCs may also exert tolerogenic effects in distant tissues via extracellular vesicles.10 It is clearly important that migration potential is considered during the generation of cellular therapies.
Cellular therapies for therapeutic tolerance
What could cellular therapies achieve?
Numerous preclinical studies using animal models of autoimmune disorders have shown potent tolerogenic effects of these various immune modulatory cells, although some mechanisms of action remain unclear. Animal models do not faithfully replicate all mechanisms of human autoimmunity but positive results have provided the scientific basis to catalyse clinical trials.
Mesenchymal stromal cells (MSCs)
The first ever preclinical study of MSCs in an autoimmune setting was in experimental auto-immune encephalomyelitis (a model for MS).56 MSCs were effective in treating the disease and were shown to be strikingly effective if injected before or at the onset of disease. Further studies in experimental MS buttressed this finding57–60 and showed that MSCs control disease through inhibition of CD4+ Th17 T cells,58 generation of CD4+CD25+FoxP3+ Tregs60 and through hepatocyte growth factor production.59 Therapeutic efficacy was also observed in the MRL/Lpr61 and NZB/W F162 63 mouse models of SLE. MSCs were effective in collagen-induced arthritis,64 65 Freund’s adjuvant-induced arthritis and K/BxN mice with spontaneous erosive arthritis.66 These studies have been reviewed elsewhere.10
Results from early clinical trials in MS showed good tolerability and some potential efficacy67–70 (table 2A) associated with increased number of Tregs in the peripheral blood of patients.67 In the most recent controlled study,70 13 patients received MSCs while 10 patients received conventional MS treatment. The active treatment group showed a more stable disease course and a transient increase in immunomodulatory cytokines. A placebo-controlled dose-ranging study of mesenchymal-like cells derived from placenta in patients with MS71 used a distinct type of cell with immunomodulatory and regenerative properties, which do not fully meet ISCT criteria for MSCs (and therefore not included in table 2A). Their phenotype includes CD10+, CD105+ and CD200+; they are CD34- and, like MSCs, do not express class II HLA or costimulatory molecules CD80, CD86. The cells appeared safe and well tolerated in patients with relapsing remitting MS and secondary progressive MS.
In RA, MSCs were well-tolerated and showed preliminary efficacy with improvements in clinical outcomes when combined with disease-modifying anti-rheumatic drugs (DMARDS).72 73 In the first placebo-controlled randomised trial of MSCs in RA,73 40 patients who had failed at least two biological DMARDS received intravenous infusions of adipose-derived MSCs at varying dose, while 7 patients received placebo. Adverse events were few and included fever and respiratory tract infections; however, serious adverse events included a lacunar infarction. Clinical outcomes, especially DAS28-ESR, showed a dose-dependent improvement.
The first case series of MSC in patients with SLE was published in 2009.74 Four patients with cyclophosphamide/glucocorticoid-refractory SLE were treated with bone marrow-derived MSCs. After 12–18 months of follow-up, all showed improvement in disease activity, renal function and serological markers. Subsequent studies, mainly by the same group, have confirmed that MSCs are safe in SLE and reported promising results such as improvement in renal function, proteinuria, SLE disease activity indices, anti-dsDNA titre and circulating Tregs.75–80 In the most recent multicentre study, up to 60% of treated patients achieved either major or partial clinical response as determined by British Isles Lupus Activity Group scores.80 However, a relapse rate of 12.5% at 9 months may warrant repeated infusions of MSCs. An analysis, by the same group, of four patients with diffuse alveolar haemorrhage in SLE using high resolution CT scan showed resolution of lung pathology after treatment with MSCs.81
A serious complication of Crohn’s disease is perianal fistulae. MSCs have been extensively studied in Crohn’s disease for their immunomodulatory properties and for their ability to differentiate into mesodermal tissues with tissue repair capabilities (table 2B). Results in Crohn’s disease are encouraging with patients who received MSCs experiencing significant improvement in fistulae while reporting just minor side effects.82–90 The unprecedented success of MSCs in a recently concluded phase III multicentre clinical study in Crohn’s disease across seven European countries and Israel implies that MSCs could become a treatment of choice for Crohn’s fistulae refractory to conventional treatment. In this study,90 212 patients with Crohn’s disease-associated fistulae received intralesional injections of either MSCs or placebo. Fifty per cent of the treatment group achieved combined clinical and radiological remission at 24 weeks compared with 34% of the placebo group, with only minor adverse effects reported. MSC have also been successfully embedded in an absorbable biomaterial and surgically delivered for the treatment of fistulae associated with Crohn’s disease.91 In this study, 12 patients safely received MSC embedded in a Gore fistula plug with fistula healing rate of 88.3% at 6 months.
MSCs have also been used in several trials to prevent and treat graft versus host disease (GVHD). In a multicentre phase II study, 55 patients with steroid resistant severe acute GVHD received MSCs at a median dose of 1.4×106 cells, obtained either from HLA-identical sibling donors, haploidentical donors or third-party HLA-mismatched donors. Up to 30 patients achieved complete clinical response independent of cell source.92 In a recent phase II study, prophylactic MSCs were successfully used to prevent GVHD following HLA-haploidentical stem cell transplantation.93
A potential advantage of MSC therapy over some other tolerogenic therapies is that their lack of MHC class expression means that they can be derived from either an autologous or allogeneic source with little or no risk of immune rejection.10 Thus, cryopreserved allogeneic MSC could become an ‘off-the-shelf’ therapy rather than a bespoke therapy requiring preparation at the point of delivery. In tables 2A and 2B, the source of MSC is indicated for each trial listed.
Tolerogenic dendritic cells (tolDC)
In an early murine experiment, allogeneic DC transfer from diabetic non-obese diabetic (NOD) mice to prediabetic NOD mice prevented development of diabetes in the latter.94 The hypothesis was that the diabetic NOD mice DC contained pancreatic antigens that conferred immunoregulatory properties, possibly by targeting regulatory T cells specific to those antigens. Since then, many preclinical studies have demonstrated that ex vivo generated DC, with an anti-inflammatory or tolerogenic phenotype, can effectively suppress or ‘switch off’ auto-immune disorders such as diabetes,95 96 arthritis,97 MS,98 99 autoimmune thyroiditis100 and myasthenia gravis.39 In most studies, tolDC were pulsed with antigens to confer specificity: bovine serum albumin for bovine serum albumin-induced arthritis,97 pancreatic islet lysate for diabetes,95 encephalitogenic myelin basic protein peptide 68–86 (MBP 68–86) for MS99 and thyroglobulin for autoimmune thyroiditis.100 Interaction of autoreactive T cells with such partially mature or ‘deviated’ DC results in their loss of functionality (anergy), apoptosis or acquisition of regulatory function. The majority of the studies aimed at prevention of autoimmunity by administering tolDC in the predisease state (either prophylactically or immediately post-immunisation).39 95 96 100 However, tolDC also arrested established disease,39 41 97 with similar outcomes to prophylactic models.98 These studies have been summarised elsewhere.42
The first clinical trial of tolDC in a human autoimmune disorder was in type 1 diabetes101 (table 3). In this study, 10 million autologous DC were safely administered intradermally into patients two times a week for a total of 4 doses, without serious adverse effects. Two forms of DCs were used: immature ‘control DC’ cultured from monocyte precursors using IL-4 and GM-CSF and immunosuppressive DC (iDC) genetically manipulated ex-vivo to block the expression of costimulatory molecules CD80/CD86.101 TolDC were not loaded with autoantigens in this trial. Some therapeutic efficacy was suggested as some patients showed elevated c-peptide levels post-treatment, indicative of increased endogenous insulin production. In a phase I single centre study, tolDC were also safely infused intraperitoneally in patients with refractory Crohn’s disease and showed some potential efficacy.102 Other studies of TolDC in autoimmunity are in inflammatory arthritis: the AuToDeCRA study where autologous tolDC were loaded with autologous synovial fluid as a source of autoantigen103 and the Rheumavax study where autologous tolDC were exposed to citrullinated peptides to confer antigen specificity and administered intradermally to patients with RA.104 In the phase I AuToDeCRA study, DC were injected arthroscopically into an inflamed knee joint, as a robust test of their stability and safety in an inflamed environment. There was no evidence that the procedure provoked a flare of symptoms. In a study published only as an abstract, recombinant autoantigen-loaded tolDC were administered subcutaneously to patients with RA at doses of 0.5×107 and 1.5×107 cells. Dose-dependent efficacy was reported, especially in autoantigen positive patients and autoantibody titres also decreased.105 Other trials in Crohn’s disease, RA and MS are ongoing and results are yet to be published.27
A potential advantage of (autoantigen-loaded) tolDC compared with MSC is their capacity to specifically target autoreactive T cells, without non-specific immune suppression.103 104 Other similar antigen-specific cells are actively being investigated, especially in transplantation. These include regulatory macrophages (Mregs),106–108 myeloid derived suppressor cells109 and MSC-conditioned monocytes.110 While other applications remain preclinical, regulatory macrophages have been studied in humans in the context of renal transplantation. In a recent case report,108 two patients received donor-derived Mregs at doses of 7.1×106 and 8×106 cells/kg intravenously prior to receiving living donor renal transplants. Both patients were eventually weaned from steroids over 10 weeks leaving maintenance low dose tacrolimus. Transfused Mregs were shown to secrete IL-10 and suppress T cell proliferation by cell-cell contact and IFN-γ induced IDO activity.108 Both patients showed increased numbers of circulating Tregs post-transplant and a peripheral blood gene expression profile indicative of tolerance according to the Indices of Tolerance (IOT) research network.111
Regulatory T cells
‘Natural’ CD4+CD25+FoxP3+ regulatory T cells (Tregs) play a central role in immune tolerance in health. While the evidence is not always definitive, Treg defects or deficiencies have been implicated in several autoimmune diseases.47 112 As with MSCs and DCs, considerable effort has therefore been dedicated to developing methodologies to isolate and expand these cells, as a potential tolerogenic therapy for autoimmune disease. Isolation uses the cell surface markers CD4, CD25 and usually CD127low. Subsequent expansion generally uses anti-CD3, anti-CD28 and IL-2 (figure 2). The expanded cells can, in theory, be rendered disease-specific by expansion in the presence of relevant autoantigens or genetic manipulation of TCR expression.113 Expanded Tregs have been used preclinically to treat murine models of autoimmunity, especially type 1 diabetes114–118 and, in some studies, Tregs were expanded with DCs to confer antigen specificity. In humans, early trials took place in patients with GVHD following bone marrow transplantation. For example, transfusion of HLA partially matched allogeneic umbilical cord blood derived Tregs at a dose of 0.1–30×105 Treg/kg, following double umbilical cord blood transplantation, was associated with a reduced incidence of acute GVHD when compared with identically treated controls without Treg.119 Tregs have also been used in a phase I study to prevent GVHD by infusing donor-specific ex-vivo expanded Tregs prior to haploidentical haematopoietic stem cell transplantation without post-transplantation GVHD prophylaxis.120
The first description of expanded Treg administration in human autoimmunity was in children with type 1 diabetes.121 Ten children received intravenous injections of autologous Tregs in two dosing cohorts (10×106 and 20×106 cells/kg) and followed for 6 months (table 4). A matched control group was used to compare clinical improvement after infusion. The treatment group, on average, had lower insulin requirements at 6 months compared with their matched controls. In an extension of this study, a higher dose of up to 30×106 cells/kg was well tolerated and associated with some clinical improvement after 12 months (reduction in insulin requirement and higher C-peptide levels).122 In a recent study in adults with newly diagnosed type 1 diabetes,50 a dose escalation protocol was used to assess the maximum tolerated dose of Tregs. Patients received intravenous infusions of Tregs up to a target dose of 2.3×109 cells, experiencing no serious adverse effects. In vitro analysis showed that expansion of the Tregs increased the overall number of cells and their functional activity/potency. In this study, the DNA of expanded Tregs was labelled with deuterium, allowing in vivo tracking. Up to 25% of transfused Tregs survived in the peripheral blood after 1 year. Furthermore, deuterium did not appear in other lymphocyte populations suggesting expanded Tregs were stable after administration. Autologous Tr1 cells were also well tolerated when administered intravenously in 20 patients with Crohn’s disease with associated improvement in disease activity.123
Concerns have been raised about the potential plasticity of Tregs in relation to their reliability as a cellular therapy. Natural Tregs form a relatively small proportion of peripheral blood CD4+ T cells and express no unique surface marker to facilitate their isolation. Nonetheless, enrichment of CD127-/low cells generally suffices to minimise contamination with activated T cells. However, the propensity for expanded Tregs to express IL-17 was noted some years ago, with evidence suggesting that CD4+CD25+FoxP3+ Tregs can undergo transformation to pathogenic Th17 cells after repeated expansion.124–126 These studies demonstrated that epigenetic instability of the FoxP3 and retinoic acid receptor-related orphan receptor (RORC) loci accounted for the potential for Th17 (de-)differentiation. Further investigation demonstrated that both loci were stable in ‘naïve’ (CD45RA+) Tregs, when compared with memory (CD45RO+) Tregs.126 127 Therefore, use of CD45RA as an additional marker for Treg isolation should minimise expansion-induced epigenetic instability and produce a more homogenous tolerogenic Treg population, with low risk of Th17 transformation. In mice, evidence exists for cells that coexpress FoxP3 and RORγT, the murine equivalent of the Th17-lineage defining marker RORC.128 Despite a capacity to differentiate into either classical Tregs or Th17 cells, these cells demonstrated a regulatory function in murine diabetes.
The development of Tr1 cells as a therapy is at an earlier stage than regulatory T cell therapy. They can be expanded ex vivo from PBMC or CD4+ T cells. One method, using an IL-10 secreting DC (DC-10), can generate allospecific Tr1 cells for potential use in haematological or solid organ transplantation. An alternative technique generated ova-specific Tr1 cells for a phase 1b/2a clinical trial in Crohn’s disease.123
In vivo expansion of regulatory T cells
IL-2 is a key cytokine for T cell activation and proliferation. Furthermore, because natural Tregs express high levels of CD25, the IL-2 receptor alpha chain, they are highly sensitive to stimulation by IL-2. In patients with cancer treated with peptide vaccine129 and DC-based vaccine immunotherapy,130 131 administration of IL-2 (with a rationale to expand effector T cells) actually led to in-vivo expansion of Tregs. This led to the theory that IL-2, particularly at low doses, will preferentially expand Tregs, informing preclinical experiments and clinical trials in autoimmunity. In a cohort of patients with chronic refractory GVHD, low dose IL-2 administration (0.3–1×106 IU/m2) increased Treg:Teff ratio, with improvement in clinical symptoms and enabling tapering of steroid dose by a mean of 60%.132 Similarly, low dose IL-2 (1–2×105 IU/m2) post-allogeneic SCT in children prevented acute GVHD when compared with those who did not receive low dose IL-2.133
Treatment of patients with Hepatitis C virus-induced, cryoglobulin-associated vasculitis with IL-2 at a dose of 1.5×106 IU once a day for 5 days followed by 3×106 IU for 5 days on weeks 3, 6 and 9 was associated with clinical improvement in 80% of patients as well as a reduction in cryoglobulinaemia and normalisation of complement levels.134 In a phase I trial in type 1 diabetes, administration of 2–4 mg/day of rapamycin and 4.5×106 IU IL-2 thrice per week for 1 month led to a transient increase in Tregs but a paradoxical worsening of β-cell function, associated with an increase in circulating NK-cells and eosinophils.135 In SLE, a Treg defect associates with disease activity and appears secondary to defective endogenous IL-2 production.136 Exogenous low dose IL-2 appears to both reverse the biological defect and provide a potential therapeutic strategy.136–138
A common finding in trials of low dose IL-2 to treat autoimmunity is that effects are transient, declining once treatment is discontinued. Effects may not be limited to natural Tregs but also extend to FoxP3+CD8+ T cells, at least in type 1 diabetes.139 However, an optimum dosing regime is yet to be defined. Results from a recent adaptive dose-finding study in 40 patients with type 1 diabetes suggest that the optimal dose of a single injection of IL-2 that will induce 10% and 20% increases in Tregs over 7 days were approximately 0.10×106 IU/m2 and 0.5×106 IU/m2, respectively.140 This study also showed that the mean plasma concentrations of IL-2 at 90 min postinjection, even at the lowest doses, were higher than the hypothetical Treg-specific therapeutic window determined in vitro (0.015–0.24 IU/mL). This was associated with a dose-dependent transient desensitisation of Tregs (downmodulation of the beta subunit of IL-2 receptor (CD122)) and a decrease in the number of circulating Tregs and other lymphocytes, which improved 2 days after injection. These findings may explain the lack of response seen in some patients who have received daily injections of low-dose IL-2. A follow-on study by the same group investigated the optimum frequency of administration of IL-2 in type 1 diabetes.141 Results show that the optimum regimen to maintain a steady state increase in Treg of 30% and CD25 expression of 25% without Teff expansion was 0.26×10 IU/m2 every 3 days.142
It is unclear at this juncture whether in vivo expansion of Tregs might provide a superior therapeutic option in autoimmunity than ex vivo expansion and readministration. Conceivably the two modalities could be combined. Other attempts have been made to expand Tregs in vivo. One method is the administration of autoantigen in Freund’s incomplete adjuvant. In a phase I trial, a single dose of insulin-β-chain in IFA was administered intramuscularly to patients with type 1 diabetes.143 Treatment was well tolerated and appeared to stimulate robust antigen-specific regulatory T cell populations in the treatment arm up to 24 months, although there was no statistically significant difference in mixed meal stimulated c-peptide responses compared with the control group. Other methods are the probiotic use of whole helminths or their unfractionated products and administration of purified excretory/secretory helminths’ products. In preclinical studies using animal models of RA, MS, Crohn’s disease and type 1 diabetes, they induce Tregs (and other regulatory cells) in vivo and prevent autoimmunity.144–146 However, clinical trials are yet to show consistent encouraging results in humans.145
Where are we now?
Results to date from human clinical trials have shown that cellular therapies are, at minimum, safe and feasible, and therefore worth exploring further in our pursuit of therapeutic tolerance induction. The regenerative properties of MSCs could additionally provide an element of tissue replenishment, repairing some of the damage that inevitably accompanies autoimmunity. However, most of the studies outlined in this review are at the very earliest phases of clinical development. Phase II and, ultimately, phase III studies will be needed to confirm their efficacy. Furthermore, as with any tolerogenic therapy in autoimmunity, clear objectives are required for efficacy trials. In transplantation, ‘operational tolerance’ is present when immunosuppression can be removed without allograft rejection. The situation is less clear in autoimmunity. Re-establishment of self-tolerance should equate with life-time drug-free remission, which has been demonstrated in some animal models when tolerogenic cells are administered both prophylactically and therapeutically.42 95 However, tolerance takes time to develop and tolerogenic therapies may not reduce symptoms in the short-term, necessitating the temporary continuation of more conventional therapies. Furthermore, immunosuppressive drugs and glucocorticoids could potentially interfere with tolerance induction as previously suggested for calcineurin inhibitors.147 Careful clinical trial designs will therefore be fundamental in order to identify, robustly, tolerance induction. In the short term, this is likely to require immune monitoring, for example, using autoantibody arrays and MHC-peptide tetramers, in order to track and interrogate the quality and quantity of the autoantigen-specific response.148 149 To date, cellular therapy trials have only occasionally incorporated experimental medicine end-points, for example, to measure longevity of cells, their distribution in vivo or to determine appropriate dosage.123 140 It is important that future trials adopt a similar philosophy, both to advance therapeutic development and also for ethical reasons.
Other factors to consider during the development of tolerogenic cellular therapies include the route of delivery. For more standard therapeutics, the main decision is usually oral vs parenteral delivery. For cellular therapies, the route has to be parenteral but the decision is potentially more sophisticated. For example, where might TolDC regulate an aberrant autoimmune response? In the target tissue, the draining lymph nodes, the central lymphoid organs? Route of delivery is likely to influence the therapy's ultimate destination, and treatment development needs to encompass work that demonstrates the cells express appropriate homing receptors. And then, there are the more standard developmental questions such as dosage and frequency of administration—a true tolerogenic therapy should only require a single ‘course’ of treatment but, in a patient with a propensity to autoimmunity, regular re-treatments may be required to keep autoreactivity at bay. Choice of autoantigen is also critical for certain cellular therapies. And last, cost-effectiveness has to be demonstrated for any novel treatment. However, the health economics would be very different for a tolerogenic therapy if it could truly avoid the need for chronic immunosuppressive therapy and its complications, not to mention the ravages of autoimmunity-associated tissue damage and comorbidities, such as cardiovascular disease.
The costs of isolating and expanding cells for therapy are significant but collaborations across academic research centres and commercial partners will solve some logistical challenges of clinical grade manufacture. Such challenges include cell source, cell isolation and expansion techniques, culture media and reagents, potency markers and genetic manipulation techniques where required (figure 2). These need to be standardised to ensure reproducibility because different cell manufacturing techniques will lead to subtle or even unidentified phenotypic differences in the final product. For example, it is unclear whether different types of tolDC, manufactured using distinct techniques, will have significantly different clinical effects.150 Measurement of potency is therefore a critical step prior to the release and administration of any cellular therapy product.151
At one point, the costs of cell manufacturing were envisaged to be a potential barrier to the development of immunomodulatory cell therapies. However, with the success of cellular therapeutics such as chimeric antigen receptor T cells for cancer, significant investment has been made in relevant technologies. For example, closed bioreactors can enable manufacture of large quantities of GMP-grade cells within a shorter period of time than labour-intensive, open culturing in flasks and bags.152 Such technologies are inherently adaptable, and therefore transferrable to different types of cellular therapy,153 helping to achieve cost-effectiveness and reducing batch-to-batch variability.
Eventually, and assuming positive results, comparative effectiveness trials across cell types (MSCs, TolDC and Tregs) may be required to determine which products are best suited for different forms and stages of autoimmunity. For example, MSCs, because of their regenerative capacity, may be favoured in conditions such as Crohn’s disease and MS where tissue regeneration would be advantageous. On the contrary, Tregs may be preferred in diseases with documented evidence of Treg dysfunction such as type 1 diabetes and SLE, because ex-vivo expansion of Tregs can reverse Treg dysfunction.154 The effects of different cell types is being investigated in transplantation in The ONE Study.155 In this collaborative study, different immunosuppressive cell populations (tolerogenic macrophages, myeloid derived suppressor cells, tolDC, monocytes conditioned by MSCs, IL-10 induced DCs and rapamycin-conditioned DCs) are manufactured from the same leukapheresis product, removing one element of variability when comparing these very different therapies. Cells are then studied in different disease contexts to determine the best approach to treatment. It may also prove possible to combine different cells to produce synergistic effects.
As tolerance can break down many years before the onset of clinical disease, it is also important to consider the optimal timing of cellular therapies. Detection of preclinical autoimmunity may provide a window of opportunity to treat and cure these diseases with safe interventions before symptom onset and before tissue damage has accrued. Epitope spreading, with broadening of the autoimmune repertoire alongside the non-specific effects of tissue damage, might render therapeutic tolerance induction more difficult in established disease, despite phenomena such as infectious tolerance and linked suppression.156 Appropriate immune monitoring will be even more important in disease, as a means to establish benefit in the absence of symptoms or signs. In-depth studies of allograft recipients who have achieved operational tolerance have identified biomarkers that appear specific for the tolerant state. These may be useful for monitoring attempts at tolerance induction prospectively.157
It is an exciting time for tolerogenic cellular therapies. Rapid advances can be expected in the short to medium term catalysed by progress in manufacturing technologies, advances in the development of immune monitoring techniques and the identification of tolerance biomarkers, alongside an acceptance that earlier treatment may be ethically justified if the therapeutic target is tolerance induction. Whether any, or all, of the cells discussed in this review will ultimately demonstrate robust tolerogenic effects must await formal clinical trials of efficacy; and we should be as certain as we can be that the timing, route and dosing of therapy is optimal before conducting the ‘definitive’ studies. These are not easy challenges but they are tractable and, currently, there is a large amount of intellectual energy directed at solving them.
Handling editor Josef S Smolen
Contributors The outline of the article was developed by JDI and CHM. The first draft of the article was prepared by CHM. Subsequent drafts were prepared by CHM following suggestions and amendments by JDI. Figures and tables were prepared by CHM and approved by JDI. Both authors approved the final submitted article and its revision in response to Reviewers' comments.
Funding This work was supported by the National Institute for Health Research Newcastle Biomedical Research Centre based at Newcastle upon Tyne Hospitals NHS Foundation Trust and Newcastle University. The authors’ work is supported by the Arthritis Research UK (ARUK)-Newcastle Biomedicine Experimental Arthritis Treatment Centre, the ARUK Centre of Excellence in RA Pathogenesis and by the European Commission Innovative Medicines Initiative Rheuma-Tolerance for Cure (RT-CURE).
Disclaimer The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health.
Competing interests None declared.
Patient consent Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.