Targeted Therapy of Multiple Sclerosis: a Case for Antigen-specific Tregs

1. Introduction

Multiple sclerosis is a chronic neuroinflammatory autoimmune disease in which autoreactive T cells and B cells infiltrate the blood-brain barrier and wrongly attack the myelin sheath, resulting in demyelination and axonal damage in the central nervous system (CNS). The condition typically manifests as a Relapsing-Remitting MS (RRMS), although the majority of patients advance to Secondary Progressive MS (SPMS) which is characterized by disease worsening without signs of remission (Lorscheider et al., 2016). About 10-15% of MS patients are initially diagnosed with Primary Progressive MS (PPMS), where the disease steadily worsens without any intervals of remission (Faissner et al., 2019, Oh & Bar-Or, 2022).

Currently, the therapeutic landscape of RRMS revolves around disease-modifying treatments (DMTs) that aim to reduce pathology related to relapses to achieve a favorable clinical course (Oh & Bar-Or, 2022). However, despite a long list of approved DMTs, patients still face the progression of disease and worsening of disabilities. Unfortunately, the DMT’s benefit in RRMS patients are typically not seen in patients with SPMS and PRMS (Faissner et al., 2019; Oh & Bar-Or, 2022). There is increasing evidence to suggest that the accumulation of axonal loss and neurodegeneration, in association with the local inflammation, leads to irreversible neurological disabilities associated with the clinical advancement in MS. Currently, drugs that can promote neural repair are absent (Absinta, Lassmann & Trapp, 2020; Klotz, Antel & Kuhlmann, 2023). Thus, there is a need to develop novel therapies that could provide solutions to the existing unmet needs of MS patients.

2. CNS is not Immune Privileged

Although the frequency of adaptive immune cells in healthy brain tissue is low, there is a steady state of lymphocyte migration across the blood-brain border, which is often incorrectly described as blood-brain barrier (Engelhardt & Ransohoff, 2012). In addition to low level migration into brain tissue, there is also a population of CNS-resident T_RM cells that, in experimental models, can cause local tissue damage even in the absence of further T cell infiltration from the blood (Vincenti et al., 2022). However, the infiltration of pathogenic T cells from the blood into the CNS plays an important role in the progression of MS. Hence, reducing the migration of T cells across the blood-brain border has been a successful strategy in the management of patients with RRMS. The antibody natalizumab blocks the a4 integrin subunit of VCAM-1 and MDdCAM-1, and thus disrupts T cell adhesion to endothelial cells, resulting in the inhibition of T cell infiltration into brain and spinal cord (Engelhardt & Ransohoff, 2012). Although Natalizumab, and other disease modifying biologics that prevent T cell infiltration, have been effective in RRMS, these interventions are ineffective in SPMS and PPMS. Mechanistically, prevention of T cell infiltration can reduce inflammation and cell damage caused by autoreactive immune cells during the RRMS phase of disease, but it cannot reverse extensive demyelination and tissue damage caused by resident T_RM cells. Despite the success of treatment with biologics, the current approaches do not enhance tissue repair or stimulate remyelination in patients with MS.

3. T_reg Biology and its Role in MS

T_reg cells are a subset of CD4+ T cells that are vital to prevent autoimmunity. FoxP3 is the master transcription factor for T_reg cells, and inherited defects in the FoxP3 gene lead to the failure of T_reg development which results in fatal autoimmunity in mice (Clark et al., 1999) and in the Immune dysregulation, Polyendocrinopathy, Enteropathy, X-linked (IPEX) syndrome in humans (Bennett et al., 2001). While the majority of T_reg cells develop naturally in the thymus, some can be reprogrammed from conventional CD4+ T cells in peripheral tissues. The T cell receptor (TCR) of thymus-derived T_reg cells have been selected for high affinity to self-peptides (Jordan et al., 2001), and these cells play an important role in maintaining systemic tolerance. By contrast, peripherally-induced pT_reg are formed in response low affinity self-peptides and in response to commensal microbiota and food antigens, and these cells play a role in tissue specific hemostasis and tolerance (Raffin, Vo & Bluestone, 2019). In vitro T_reg induction from conventional CD4+ T cells has been repeatedly demonstrated by TGF-β and IL-2 to mirror the pT_reg lineage commitment pathway (Chen et al., 2003), yet these iT_reg cells are prone to lose FoxP3 expression and thus their T_reg functionality. This so-called T_reg plasticity is to a large extent under the epigenetic control of the T_reg -Specific Demethylated Region (TSDR). While the TSDR is completely demethylated in tT_reg cells (Toker et al., 2013), it is partially methylated in pT_reg and iT_reg (Dhamne et al., 2013), which may favor loss of FoxP3 expression in these cells after in vitro restimulation (Floess et al., 2007; Toker et al., 2013). Similarly, activated effector T cells that can temporarily upregulate FoxP3 have a TSDR that is largely methylated (Baron et al., 2007).

T_reg cells are equipped with multifaceted mechanisms to achieve their two major roles - immune suppression and tissue homeostasis (Figure 1). Upon activation by cognate self-antigen and sensing a pro-inflammatory environment, T_reg cells are attracted to the site of inflammation, release anti-inflammatory cytokines and restrict further T cell activation by hindering antigen presentation. Additionally, local immunosuppression can be achieved by activated T_reg through depriving T-cell pro-survival nutrients and signals, such as IL-2, from the local environment (Duffy, Keating & Moalem-Taylor, 2019). The mechanisms by which T_reg can repair tissue damage are less well established. Tissue resident T_reg can have a direct regenerative ability, as demonstrated in skin and skeletal muscle (Li et al., 2018). Amphiregulin, an EGFR ligand crucial to tissue repair, is dispensable for the suppressive function of T_reg cells (Arpaia et al., 2015), while its production by T_reg was found to play a role during the chronic phase after stroke by accelerating neurological recovery (Ito et al., 2019). Moreover, the observation that T_reg can promote oligodendrocyte proliferation and differentiation in vitro, may explain how T_reg can enhance oligodendrocytes mediated repair of damaged CNS tissue in vivo, as described by Crawford and colleagues (Crawford et al., 2016). Mechanistic studies further revealed that CCN3, IL-10 and AIM2 expressed by T_reg play an important role in stimulating neuronal tissue repair (Dombrowski et al., 2017; Naughton et al., 2020; de la Vega Gallardo et al., 2020; Chou, 2021).

Recent studies have uncovered some abnormalities of T_reg in MS patients, yet it has been difficult to establish whether this is a cause or consequence of MS pathogenesis. Direct comparisons of the frequency of peripheral T_reg and infiltrated T_reg from cerebrospinal fluid between MS patients and healthy individuals had generated unconclusive results (Duffy et al., 2018). However, ex vivo phenotyping of peripheral T_reg of MS patients revealed reduced FoxP3 expression and impaired suppressive activity (Viglietta et al., 2004; Haas et al., 2007). A study of untreated RRMS patients showed an increased number of Th1-like, IFN-γ-producing FoxP3+ T cells compared to healthy controls (Dominguez-Villar, Baecher-Allan, & Hafler, 2011). This functional plasticity could be due to the highly pro-inflammatory environment of CNS lesions, characterized by IFN-γ, TNF-α, IL-1β, IL-6 and IL-17 (Duffy et al., 2018). The use of animal models has provided more mechanistic understanding of the role of T_reg cells in MS disease. Using the most widely used MS mouse model, Experimental Autoimmune Encephalomyelitis (EAE), it has been shown that T_reg depletion or dysfunction can promote disease progression which can then be rescued by the adoptive transferr of functional T_reg (Duffy et al., 2017; Stephens et al., 2009; Kim et al., 2018; Duffy, Keating & Moalem-Taylor, 2019). Considering the dual actions of T_reg cells, it is likely that the beneficial effect of adoptive T_reg therapy is the concerted result of immune suppressive activities and their ability to stimulate tissue repair.

4. Novel T_reg Based Therapies for MS

To ultimately achieve an increased number of functional T_reg in MS patient, a variety of non-cell-based and cell-based therapies are under development. Whilst the former aims at rescuing the patient’s own T_reg repertoire, cell-based therapies replenish the patient with ex vivo expanded and potentially engineered T_reg. The toolbox of non-cell based therapies comprises primarily three synergistic approaches, which are peptides for antigen-specific expansion, variants of IL-2 for preferential T_reg survival, and small molecules for reinforcing functionality (Eggenhuizen, Ng & Ooi, 2020). Although these strategies have not yet entered the clinic, novel biologic designs aiming at combining all three keys of the toolbox are emerging. Particularly, a recent proof-of-concept study demonstrated the feasibility and efficacy of microparticles decorated with MHC-II molecules containing peptides of myelin oligodendrocyte glycoprotein (MOG) used as a T_reg vaccine, which led to targeted expansion of MOG-specific T_reg cells via the incorporation of a modified IL-2 molecule stimulating the IL-2 receptor of T_reg, as well as simultaneous inhibition of effector T cell proliferation via the release of rapamycin (Rhodes et al., 2023). However, these types of strategies rely on the presence T_reg with the vaccine-targeted specificity in MS patients, and they carry the risk of inadvertently expanding pathogenic effector T cells that have the same specificity. Thus, adoptive transfer of antigen-specific T_reg represents an alternative therapeutic approach.

5. Antigen-Specific T_reg Cell Therapy for MS and its Benefits

Therapeutic T_reg can be generated in three ways (Figure 2), which include (i) polyclonal expansion or (ii) antigen-specific expansion of autologous T_reg obtained from patients, and (iii) genetic engineering of T_reg specificity (Raffin, Vo & Bluestone, 2019).

The observation that multiple autoantigens have been identified as the cause of MS, and more than one clone of autoreactive T cells can be found in patients (Bronge et al., 2022), raised the question whether adoptive transfer of polyclonal T_reg is preferable over T_reg cells with a single specificity. However, T_reg cells with specificity for one specific antigen, for example myelin basic protein (MBP), can achieve local suppression of effector T cells present in the same microenvironment, even if these effector cells are specific for several other proteins of the neuronal myelin sheet. For example, experiments in the murine EAE model showed that T_reg specific for MBP were able to suppress pathology caused by effector T cells specific for MOG (McGovern et al., 2022). The murine EAE model was also used to demonstrated that T_reg with defined specificity are more potent than polyclonal T_reg in controlling MS-like symptoms (Stephens et al., 2009; Kim et al., 2018), which is consistent with better disease control by antigen-specific T_reg cells observed in other autoimmune animal models (Tang et al., 2004; Tarbell et al., 2004; Zhou et al., 2004; Fujio et al., 2006; Tsang et al., 2008). The improved efficacy of antigen-specific over polyclonal T_reg is related to the enhanced antigen-driven migration of T_reg into target tissues, and the continued TCR-stimulation at the site of pathology which is required for T_reg to exert their optimal suppressive function (Eggenhuizen, Ng & Ooi, 2020).

6. TCR-Engineered T_reg Versus CAR-Engineered T_reg

A population of antigen-specific T_reg can be produced by ex vivo stimulation with antigen, or by genetic T_reg engineering. The frequency of naive T_reg that are specific for a desired antigen is very low, and the expansion of these cells by in vitro stimulation is technically difficult and unreliable, compared to the robust process of genetic engineering of T_reg specificity (Serra & Santamaria, 2019). Specificity engineering can be achieved by TCR and CAR (Chimeric Antigen Receptor) gene transfer. Both methods involve designing and transducing a transgenic receptor, yet each has its strengths and limitations. Whilst antigen recognition by transgenic TCR is restricted by the HLA genotype of patients, the antigen-binding domain of CAR is a synthetic single-chain variable antibody construct that allows antigen recognition irrespective of HLA. However, in contrast to TCRs, CARs can only recognize proteins present on the cell surface but not proteins in the cytosol or nucleolus, whereas HLA-presented peptides can be derived from any cellular protein for recognition by TCRs. Hence, the number of cellular proteins that can be targeted by TCRs is vastly greater than the proteins targetable by CARs. Currently, the preclinical development of CAR-T_reg for MS is lagging behind the use of TCR-engineered T_reg (Depil et al., 2020).

7. Source of T_reg

As shown in Figure 2, human T_reg cells can be isolated from Peripheral Blood (PB), Umbilical Cord Blood (UCB) or in vitro differentiated from Induced Pluripotent Stem Cells (iPSC). Autologous T_reg from PB remain the most common source of T_reg used in patients (Eggenhuizen, Ng & Ooi, 2020). The results of a clinical trial in patients with type 1 diabetes produced safety data of autologous ex vivo expanded polyclonal T_reg (Bluestone et al., 2015), prompting the development of T_reg therapies for other autoimmune diseases, including MS. In addition, several studies with UCB-derived T_reg have demonstrated that they displayed superior phenotypic stability and greater TCR repertoire diversity compared to PB-derived T_reg (Seay et al, 2017; Motwani et al., 2020). As UCB biobanks expand, the use of HLA-matched allogeneic T_reg population for the treatment of GvHD, and possibility certain autoimmune conditions, becomes a realistic possibility (Motwani et al., 2020).

iPSC is an attractive source for producing large numbers of ‘off the shelf’ T_reg with defined specificity and phenotypic and functional features. However, T_reg differentiation from iPSC is still at an early stage of development, and TCR+FoxP3 transduction combined with Notch-1 ligand stimulation was required to achieve T_reg differentiation (Haque et al., 2012; Haque et al., 2016). The use of iPSC-derived T_reg as ‘off the shelf’ medicines would require additional genetic engineering to avoid host-mediated recognition and rejection of the adoptively transferred cells (Raffin, Vo & Bluestone, 2019).

8. The Risks and Challenges of Antigen-Specific T_reg Cell Therapy

The challenges of T_reg-based adoptive cell therapy include common limitations of currently available techniques, and challenges specific to the treatment of MS. The first challenge is the yet inconclusive markers for the isolation of T_reg cells with high purity, phenotypic stability, and continued functionality. At present the markers CD4+ CD25+ CD127-/low are frequently used to purify T_reg. Unfortunately, staining for FoxP3, an intracellular protein, requires cell permeabilization and is therefore not suitable for the purification of life cells for adoptive therapy, although the vast majority of CD4+ CD25+ CD127-/low cells are FoxP3-positive and thus largely overcomes the inability of using FoxP3 for cell purification. However, T_reg isolation is further complicated by the observation that FoxP3-positive T_reg consist of distinct subsets with distinct functionality. For example, it was shown that tT_reg cells are more resistant to phenotypic plasticity than pTreg cells (Miyao et al., 2012), yet any differentiating phenotypic markers remain to be elucidated in humans (Raffin, Vo & Bluestone, 2019). At present, we do not know which Treg subset is most suitable for dampening the neuroinflammation in the CNS in the context of MS.

9. Off-Target Toxicities

Retro and lentiviral gene transfer, which so far has been used in most human applications of CAR or TCR-T cell therapy, can cause genome toxicity by insertion mutagenesis, which has resulted in fatal side effects with gene engineered stem cells (Hacein-Bey-Abina et al., 2003), while similar side effects have not been seen with gene engineered T cells. Owing to the heterodimer nature of TCR alpha and beta chains, mis-paring between transgenic TCR and endogenous TCR is a risk associated with TCR but not CAR engineering. Experiments in murine T cell transfer models have shown that TCR mis-pairing can result in severe side effects in lymphodepleted recipient mice (Bendle et al., 2010). In patients, toxicities caused by TCR mis-pairing have not yet been described, although a large number of patients have been treated over the past years. Nevertheless, strategies for modifying transgenic TCR to avoid pairing with endogenous TCR, such cysteine-modification (Cohen et al., 2007), domain-swapping (Bethune et al., 2016), constant region modification (Sommermeyer & Uckert, 2010) and variable region modification (Thomas et al., 2019), all reduce the risk of mis-pairing. The CRISPR/Cas9 precision gene-editing tool changed the landscape, as it allows specific disruption of TRAC and TRBC loci, which encode the TCRα and TCRβ chain, respectively. As recently demonstrated, the disruption of endogenous TCR expression and insertion of therapeutic CAR or TCR constructs into the TRAC locus has improved the in vivo functionality of engineered T cell products in preclinical models (Eyquem et al., 2017; Roth et al., 2018).

10. Systemic Immunosuppression

Inevitably, any drugs with immunosuppressive activity afford a risk of systemic immunosuppression, whereupon the likelihood of opportunistic infection or malignancy will be elevated. Being a cell therapy, adoptively transferred Treg cells are expected to persist long-term and therefore require careful assessment of possible side effects.

Safety features developed for TCR/CAR engineering for cancer cell therapy can be translated into controlling unwanted systemic immunosuppression of Treg-based therapy. One strategy is to use a suicide gene to control the in vivo survival and/or functionality of the adoptively transferred Treg cells. Numerous ‘suicide switched’ have been designed and assessed for their efficacy and safety in preclinical studies and in patients, exemplified by truncated EGFR (Paszkiewicz et al., 2016) and inducible Caspase 9 (Mestermann et al., 2019). Another approach to regulating Treg function in vivo is the introduction of synthetic receptors. The requirement of IL-2 for Treg proliferation, survival and function makes IL-2 a promising manipulative pathway (Chinen et al., 2016). Recently, an elegant study has demonstrated that a genetically modified IL-2 receptor binds selectively to a synthetic IL-2 molecule that contains a ‘matched’ modification, while wild type IL-2 does not bind to the modified receptor. Hence, engineered T cells expressing the modified IL-2 receptor responded in vivo only to administered of synthetic IL-2 containing the ‘matched’ modification but not to endogenous IL-2 (Sockolosky et al., 2018). This technology could be used to engineer Treg whose survival and function in vivo is controlled by the time and dose of synthetic IL-2 administration.

11. Plasticity of Functions

The functional plasticity of Treg cells has been demonstrated repeatedly both in vitro and in vivo following adoptive transfer. Upon cues from inflammatory cytokines and IL-2 deficiency, antigen-specific Treg cells can adopt an alternative transcriptional program and become effector T cells that may contribute to tissue pathology (Sawant et al., 2014). For example, in the murine EAE model researchers identified a population of exFoxP3-Treg that had lost expression of FoxP3 which was associated with the production of the effector cytokine IFNg (Bailey-Bucktrout et al., 2013). However, genetic engineering can be used to reinforce the Treg identity by forcing FoxP3 expression. This can be achieved by delivering a genetic construct that drives FoxP3 expression from a constitutively active promotor (McGovern et al., 2022), or by inserting a constitutively active enhancer/promotor into genome proximal to the FoxP3 gene (Honaker et al., 2022). Both approaches lead to sustained level of FoxP3 expression which ‘locks’ Treg cells into a stable phenotype. Both approaches can also be utilized to force FoxP3 expression in conventional CD4+ T cells and convert them into Treg-like cells that display suppressive activity in vitro and in vivo (Wright et al., 2009).

12. Conclusions

With its multifaceted immunosuppressive functions and remyelinating capacity, T_reg cells are undoubtedly an ideal cellular candidate for MS immunotherapy. Adoptive therapy with engineered T_reg is now at the exciting stage where the first human clinical trials are ongoing, and will soon provide eagerly awaited feasibility, safety and efficacy data. Success of T_reg therapy will depend on the selection of suitable target antigens that stimulate the suppressive activity and the tissue repair function of T_reg at the local site of disease, without any systemic impairment of the immune system.