Targeting T Follicular Helper Cells To Control Humoral Allogeneic Immunity

Abstract

Humoral allogeneic immunity driven by anti-HLA donor-specific antibodies (DSAs) and antibody-mediated mediated rejection (ABMR) significantly impede prolonged survival of organ allografts after transplantation. Although the importance of T follicular helper (T_FH) cells in controlling antibody responses has been long established, their role in directing DSA generation leading to ABMR was only recently appreciated in the clinical setting of organ transplantation. In this review, we provide a comprehensive summary of the current knowledge on the biology of human T_FH cells as well as their circulating (cT_FH) counterparts and describe their pivotal role in driving humoral alloimmunity. In addition, we discuss the intrinsic effects of current induction therapies and maintenance immunosuppressive drugs as well as of biotherapies on T_FH cells; and provide future directions and novel opportunities of biotherapeutic targeting of T_FH cells that have the potential of bringing the prophylactic and curative treatments of ABMR towards personalized and precision medicine.

Introduction

The immune conflict triggered by the development of anti-HLA donor-specific antibodies (DSAs) in genetically incompatible organ recipients with their donors has been recognized as a major limitation to prolonged survival of allografts, across all solid organ transplants including kidney, heart, lung and liver transplants.^1,2 Antibody-mediated rejection (ABMR) is the clinical manifestation of DSA pathogenicity and results from binding of DSAs to the allogeneic endothelium, which triggers microvascular inflammation and intimal arteritis through antibody-dependent cell-mediated cytotoxicity mediated by effector NK cells and monocytes, as well as complement activation through complement-dependent cytotoxicity.^3,4 ABMR represents a main immunological cause of allograft failure. Yet, adequate predictive biomarkers of humoral alloimmunity are still lacking, mostly because of the poor understanding of the complex cellular and molecular mechanisms underlying DSA and ABMR development. As a consequence, patients developing ABMR are still in need of adequate treatments.

Given that DSAs are isotype switched IgG antibodies directed against protein antigens, it is postulated that they are generated through T cell-dependent B cell responses. Although it could have been largely anticipated that T follicular helper (T_FH) cells, the CD4 T cell subset specialized in the control of B cell responses,⁵ would play a major role in the induction of DSA generation and their pathogenicity leading to ABMR, it was only in the past few years that their role was elucidated. A growing number of studies have demonstrated a pivotal role for T_FH cells in transplantation in animal models but also, more recently, in humans.^6,7

This review aims to summarize the most recent and relevant findings regarding the contribution of human T_FH cells to health and disease, and to humoral alloimmunity against organ transplants; and how these cells represent amenable therapeutic targets for treatment of ABMR.

Human T_FH cells in health and disease

a). Germinal center T_FH cells

Germinal center (GC) T_FH cells are pivotal in the optimal development of humoral immunity by controlling cognate antigen-specific B cell responses and the formation of GCs. While B cells may undergo Ig class-switch recombination prior to their differentiation into GC B cells, they proliferate, undergo somatic hypermutations and generate memory B cells as well as high-affinity plasma cells in GCs with the crucial help of T_FH cells.⁸ GC T_FH cells exhibit salient cellular, molecular and tissue-dynamic features that are essential to their development and function. These unique features markedly demarcate GC T_FH cells from the classical paradigm of effector CD4 T cell activation differentiation.

Differentiation of T_FH cells is a complex and highly controlled process that involves a succession of coordinated spatio-temporal interactions with different cell populations of antigen presenting cells (APCs), and unlike other populations of CD4 T cell effectors, no single event allows their complete differentiation.^9,10 Full differentiation of T_FH cells requires a multi-step process, with an initial activation mediated by dendritic cells (DCs) in the T cell zones of secondary lymphoid organs (SLO), a subsequent B cell interaction at the T-B border and then within GCs.¹¹ In the T cell zone of SLO, the indirect presentation of antigen by DCs to naive CD4 T cells, concomitant with the engagement of the costimulatory molecule ICOS, induces the expression of the transcription factor Bcl-6 and consecutively that of CXCR5.¹² A complex combination of cytokines including IL-6, IL-12, IL-21, IL-23 and TGF-β upstream STAT3/STAT4 pathways promotes human T_FH cell fate and counteracts the non-T_FH ones.^9,13 In contrast, IL-2 is a potent inhibitor of commitment to T_FH cell fate and the IL-2/STAT5 axis cooperates with the transcription factor Blimp-1 (a Bcl-6 antagonist) to inhibit T_FH cell differentiation.¹⁴ The initial dose of antigen and potent TCR signals in conjunction with the finely tuned costimulatory and cytokine signals provided by DCs are also strong determinants of T_FH cell generation and polarization.^15,16 In addition, sustained antigenic stimulation by B cells is also necessary for T_FH cell differentiation¹⁷ and once committed, T_FH cells continue to require continuous cognate GC B cell exposure for their maintenance (Figure 1).

Figure 1.

T_FH cell crosstalk with dendritic cell and B cell.

The phenotype and transcriptional profiles of T_FH cells are now well established. In addition to Bcl-6, several other transcription factors contribute to T_FH cell differentiation and maintenance including c-Maf, IRF4, Baft, LEF1 and TCF1.^10,18 STAT3 and STAT4 are of particular importance as they promote the expression and production of IL-21, the hallmark cytokine of T_FH cells.¹⁹ In addition, epigenetic changes, including histone modification and DNA methylation that control the expression of the abovementioned transcriptional programs, are increasingly recognized.²⁰ Regarding surface molecules, T_FH cells express a range of costimulatory receptors, that reinforce their cognate interaction with B cells and promote B cell activation (e.g. CD28, CD40L, ICOS, PD-1, OX40), and multiples cytokine receptors (IL-6R, IL-7Rα, IL-12Rβ, IL-21R) allowing self-response to cytokines, cell renewal, as well as activation and differentiation into memory T_FH cells (Figure 1).

The human deficiency of T_FH cells in numbers or function have highlighted their crucial role in defense against infection and cancer, as individuals displaying genetic deficiency for hallmark molecules lack T_FH cells and GCs, and exhibit increased susceptibility to infections and cancers. Conversely, increased T_FH cells in numbers or function due to genetic mutations or polymorphism of IL-21 can lead to increase susceptibility to autoimmunity.^21,22 Also, exponential number of studies have documented that the increase frequencies/numbers and function of T_FH cells with activation phenotypes (expressing increased ICOS, PD-1 or IL-21) promote protective antibody response against pathogens after vaccination or infection and can also promote deleterious autoantibody generation and their pathogenicity leading to autoimmunity and tissue damage.

b). Circulating T_FH cells

Because of difficulties in accessing lymphoid tissues and GCs in humans, the analysis of circulating T_FH (cT_FH) cells has proved valuable in understanding alterations of T_FH cell response that contribute to human diseases. cT_FH cells are blood memory CD4 T cells that express CXCR5 that are in a resting state in the absence of antigenic stimulation. Although the developmental origins of cT_FH cells remain debated, the phenotypic, transcriptional and epigenetic relationships between cT_FH and GC T_FH cells have been outlined by studies performed in genetically mutant mice and specimens from patients with primary immunodeficiency.^23,24 These cT_FH memory cells are mainly generated from cells either committed to the T_FH lineage, or from recirculated T_FH of the GC. Some of these cT_FH cells may develop before GC formation and therefore represent precursors of GC T_FH,²⁵ while others may emerge from memory GC T_FH cells at the resolution of GC responses.^26-28 However, more definitive arguments for a pre- or post-GC emergence and differentiation of cT_FH cells are still needed. Unlike GC T_FH, cT_FH cells display markedly distinct features: (i) absence of Bcl-6 expression; (ii) lower expression of CXCR5 and of (iii) CD40L, ICOS and PD-1 at resting state.^29,30 However, cT_FH similar to GC T_FH cells display common functional attributes such as the capacity to help B cells and promote their activation and differentiation into plasma cells.³¹ This helper capacity appears to be more evident for PD-1⁺ compared to PD-1⁻ cT_FH subsets and to other CD4⁺CXCR5⁻ T cell subsets. There is also substantial overlap in the transcriptional programs of cT_FH and GC T_FH cells, including the common expression of c-MAF, IRF4 and POU2AF.^25,29-33

Activated (PD-1⁺ICOS⁺) cT_FH cells with effector (CD45RO⁺CD62L⁻) memory phenotype are robust surrogate biomarkers of ongoing GC and antibody responses in multiple clinical settings including vaccination, acute and chronic infections, allergies and cancers. They emerge before antibodies in blood, as do plasmablasts, and therefore predict the magnitude and the pathogenicity of the antibodies generated. cT_FH cells have been shown to correlate with systemic CXCL13 levels, antibody titers and disease manifestations.³⁴ cT_FH cells are composed of heterogeneous subsets, and depending on the antigenic trigger, co-stimulation and inflammatory environment, can be composed of predominant cT_FH-Th1 (CXCR3⁺CCR6⁻), cT_FH-Th2 (CXCR3⁻CCR6⁻), cT_FH-Th1/17 (CXCR3⁺CCR6⁺) or cT_FH-Th17 (CXCR3⁻CCR6⁺) subsets. In addition to producing IL-21, these subsets have distinct capacities to secrete IFNg, IL-4 or IL-17 helper cytokines as well as to instruct naive or memory B cells to differentiate into plasma cells and produce Ig of different subclasses.³⁵

c). T peripheral helper (T_PH) cells.

More recently the spectrum of T_FH-like cells with capacity to provide B cell help has substantially expanded. Very recent studies have identified a novel subset of PD-1^hiCXCR5⁻ CD4⁺ T cells with B helper capacities, localized in inflamed join tissues of patients, and were therefore named T peripheral helper (T_PH) cells.³⁶ As for T_FH cells, T_PH cells were described as expressing CXCL13, CD40L, ICOS and the transcription factor c-Maf, but not Bcl-6,^37,38 and exerting B helper activities in a manner largely dependent on IL-21 and SLAMF5.³⁶ Because of their hallmark capacity to produce CXCL13 and their frequent co-localization with B cells, it is likely that T_PH cells play a pivotal role in the formation of tertiary lymphoid structures in target tissues. As T_FH cells, T_PH cells can recirculate and can be detected as circulating T_PH (cT_PH) cells in peripheral blood of healthy individuals and are found expanded in patients with lupus, Sjogren’s syndrome, IgG4-related disease, systemic sclerosis, IgA nephropathy and type I diabetes.^39,40 Intriguingly, unlike cT_FH cells that provide strong B cell help to naive B cells to produce IgG, IgE and IgA, cT_PH cells can only induce naive B cells to produce IgM.³⁷ cT_PH cells appear to strongly correlate with the accumulation of blood extrafollicular/atypical CD11c⁺ CD21⁻ memory B cells that also lack the expression of CXCR5.^41,42 Both CD11c⁺ CD21⁻ CXCR5⁻ B cells and T_PH cells are found in inflamed tissues such as rheumatoid arthritis synovial or lupus nephritis tissues.^36,43 As CD11c⁺ CD21⁻ CXCR5⁻ B cells, cT_PH cells display marked Th1-biased features such as the expression of T-bet and the chemokine receptors CCR2, CCR5 and CX3CR1, and thus likely respond to chemokines secreted by targets tissues, that attract both cT_PH cells and atypical B cells away from secondary lymphoid organs.^36,42 Interestingly, cT_PH cells may have cell cytotoxicity potential, as indicated by the expression of GZMA, GZMB and PRF1 in specific disease settings.^42,44

However, it still remains to understand: (i) the molecular control of T_PH versus T_FH cell differentiation, (ii) the transcriptional differences between T_PH and T_FH cells, (iii) whether there can be plasticity between T_PH and T_FH cells, and (iv) whether T_PH cells are only poised to help atypical memory B cells. Also, data regarding the existence of T_PH and cT_PH cells in patients mounting DSAs and developing ABMR are missing and warranted.

d). T follicular regulatory (T_FR) cells

Another T_FH-like cell subset with capacity to suppress GC and antibody responses is represented by T follicular regulatory (T_FR) cells. T_FR cells share both T_FH cell features (Bcl-6, CXCR5 and PD-1) and T regulatory (T_REG) cell markers (CD25, CTLA4 and FoxP3) and function (IL-10 and TGFβ).^45,46 Pre-T_FR cells are derived from thymic T_REGs and are CXCR5⁺ FoxP3⁺ cells localized at the T-B border.⁴⁷ While the initial stage of T_FR cell differentiation appears B cell independent, B cells are necessary for stabilization of the T_FR cell program, including the expression of TCF1 and Bcl-6.⁴⁸ CD25 signaling helps to inhibit T_FH cell program and favors a T_FR cell fate by promoting Blimp-1 and STAT5.⁴⁹ However, a distinct subset of CD25⁻ T_FR cells can be found in GCs and differs from the CD25⁺ T_FR cells by the downregulation of IL-2-dependent canonical T_REG features, but nonetheless retains suppressive capacities.^50,51

Circulating T_FR (cT_FR) cells, similar to cT_FH cells, appear to be less activated than those found in the SLOs, generally lacking the expression of PD-1, ICOS and Bcl-6 but retaining CXCR5, and expressing CD45RA and CCR7.⁵² Due to their relative resting state compared to the activated T_FR cells found in SLOs, cT_FR cells have weaker suppressive activity unless they are pre-activated.^31,52 cT_FR cells are shown to be generated at an early stage of the GC response prior to the involvement of B cells, as they are still generated in individuals genetically lacking B cells.⁵² cT_FR cells were found to be expanded 7 days after flu vaccination and to correlate with auto-antibody responses during Sjogren’s syndrome, suggesting that their expansion is indeed indicative of ongoing GC responses.⁵²

Some mechanisms of the suppression of humoral immunity by T_FR and cT_FR cells are similar to those used by T_REGS and are dependent on: (i) CTLA4, by regulation of CD28 signaling to T_FH cells via direct control of CD80 and CD86 expression on B cells,^53,54 (ii) upregulation of CD25 and FoxP3.⁵²

However, other mechanisms of suppression such as granzyme B-mediated cytotoxicity⁵⁵ and through the action of the ectoenzymes CD39 and CD73⁵⁶ were shown for T_REGS but not for T_FR cells. In addition, recent studies have demonstrated the potent T_FR cell capacity to regulate the early, but not late, GC responses and to control Ig responses to vaccines, allergens and autoantigens.^57,58

Role of T_FH cells in transplantation

The pivotal role of T_FH cells during humoral alloimmunity against organ transplants has been increasingly documented during the past 5 years. The first studies were performed using mouse^59,60 and nonhuman primate^61,62 models of allosentization and ABMR. These studies have demonstrated that ABMR can be used as an experimental model to study T_FH cell response to alloantigen and have allowed to identify hypertrophic GCs and increase in T_FH cell numbers and activation during ABMR. Also, these models are of particular interest to test whether existing or novel therapeutic agents have intrinsic effects on GC and T_FH cell responses. In transplant recipients, using blood cT_FH cells as a proxy for estimating T_FH cell functions in GCs proved promising for diagnostic and immune monitoring purposes. However, this approach has limitations and warrant further clinical investigations to measure to what extent these cells reflect what is occurring within GCs and their utility to monitor the intrinsic effects of therapies and immune response to treatments in ABMR.

a). Role of cT_FH cells in DSA generation

More recently, increasing number of studies have exemplified the implication of T_FH cells during humoral alloimmunity in humans. De Graav et al were the first to document the functional capacity of cT_FH cells from kidney recipients to promote B cell activation and Ig production, and the increased number of cT_FH cells post transplant in patients with preformed DSA⁶ It was next shown by other groups that cT_FH cells with activation features (PD-1⁺ICOS⁺ or PD-1^hiCCR7^lo) correlated with DSA generation and their MFI levels.⁶³ Of note, frequencies of PD-1⁺ICOS⁺ cT_FH in the first year was predictive of de novo DSA development after 1 year post transplant.⁶⁴ These findings nicely corroborate those in preclinical kidney transplantation models showing that T_FH cells are required for de novo DSA responses, as well as augmentation of secondary (memory) DSA responses following presensitization,⁶⁵ and that PD-1⁺ICOS⁺ cT_FH cells are biomarkers of DSA formation.⁶⁶ In contrast, PD-1⁺ICOS⁺ cT_FH cells were substantially decreased in frequency and functionally impaired in patients displaying allograft operational tolerance, who very rarely mount DSAs post transplant.⁶⁷

Like in animal models, human cT_FH cells display potent capacity to respond to donor antigen by producing IL-21. Increased IL-21⁺ cT_FH cells can be readily identified pretransplant in patients who further develop DSAs, unlike in those who do not develop DSAs posttransplant,^68,69 indicating that preformed T cell memory to donor is significantly enriched within cT_FH cell compartment.

b). Role of cT_FH cells in ABMR

Although the association between cT_FH cell activation, their reactivity to donor and DSA generation was established, their relationship to DSA pathogenicity and ABMR was unclear. Our group has recently shown that there was an expansion of overall cT_FH cell compartment during ABMR but also an enrichment for activated/proliferating Ki67⁺ICOS⁺ and surprisingly of early memory precursors CD45RO⁺CCR7⁺CD127⁺ cells that may replenish and sustain effector T_FH response during chronic alloantigen trigger, which was not previously documented.⁷⁰ Frequencies of Ki67⁺ICOS⁺ cT_FH cells coincided with concomitant expansion of blood Ki67⁺ activated (CD20⁺CD38^lo) B cells and antibody-secreting (CD20⁻CD38^hi) cells, indicative of coordinated cT_FH-B cell responses. Ki67⁺ICOS⁺ cT_FH cells also correlated with plasma CXCL13 and DSA MFI levels, suggesting a robust GC reaction that contribute to pathogenic DSA generation during ABMR. This finding is consistent with recent data in preclinical models showing the crucial role of T_FH cells in promoting DSA responses and ABMR, as deletion of T_FH cells at the time of transplant resulted in significantly less severe allograft ABMR in mice.⁶⁵

Our transcriptional analyses of cT_FH cells further supported an enrichment for both T_FH precursors (LEF1 and TCF1) and effector T_FH (CD28, CD40L, IL-21R) gene signatures, that were selective to patients with DSA+ABMR+, unlike those who did not develop DSA (DSA-) or developed DSA but did not progress to ABMR (DSA+ABMR-) patients.⁷⁰ This distinct transcriptional programing of cT_FH cells in ABMR resulted, not surprisingly, in their increased functional capacity to provide help to B cells, that differentiated into plasma cells in response to this vigorous cT_FH cell help and generated DSAs enriched in IgG3 isotype. Although previous studies have found that both CXCR3⁺ (Th1) and CXCR3⁻ (Th2 and Th17) cT_FH subsets were found increased in patients mounting DSAs,^64,68 it was unclear how these distinctly polarized cT_FH cells correlated with differential pathogenicity of DSAs and what polarized subsets were observed during ABMR. Our recent study showed that while heterogeneous expansion of Th1, Th2 and Th17 cT_FH subsets were present during ABMR, only Th2 and Th17 subsets were predominantly increased within Ki67⁺ICOS⁺ cT_FH, and these were associated with increased IgG3 DSA generation. This is consistent with the known roles of IL-17 and IL-21 in isotype switching towards IgG3.^71,72 While previous studies have shown an association between increased frequencies of activated cT_FH cells and allograft rejection,^6,63 it was unclear whether this increase in cT_FH cells was associated with specific patterns of allograft injury and with clinical outcome. Analyzing frequencies of cT_FH cells with histological and clinical profiles, we found that expansion of Ki67⁺ICOS⁺ as well as of CD45RO⁺CCR7⁺CD127⁺ cT_FH were associated with a more severe form of ABMR with poorer kidney allograft survival. These severe ABMR were characterized by increased arteritis and by more interstitial inflammation and tubulitis, that are features of T-cell mediated rejection (TCMR).⁷⁰ Thus, cT_FH cell phenotypic and functional profiles represent valuable surrogate biomarkers to predict and distinguish less robust phenotype (DSA+ABMR−) from pathogenic (DSA+ABMR+) humoral responses after transplantation.

c). Role of T_FH cells within allograft tissues

It is tempting to speculate that inflammatory cellular infiltrates found in allografts undergoing ABMR could be composed of increased cT_FH cells that would have migrated to target tissues. Indeed, it was previously found that T_FH cells could be detected within allografts from patients with acute TCMR⁶ and in those with mixed ABMR with TCMR lesions.^73,74 These T_FH cells can also be part of ectopic/tertiary lymphoid structures formed by dense cell clusters of activated B cells.⁷⁵ Our group also found that T_FH cells can be detected within allografts with chronic ABMR (Figure 2). Interestingly, a substantial number of CD4 T cells that were CXCR5⁻ could be also observed, suggesting a T_PH or CD4⁺ T conventional (non-T_FH) phenotype of these cells, concomitant with the presence of CXCR5⁺ (T_FH) cells during chronic ABMR (Figure 2C). As T_FH cells can be observed in cases showing TCMR, it remains to be investigated whether cT_FH cell transcriptional profile and function in patients with TCMR would significantly differ from those with ABMR.

Figure 2. T_FH cells within kidney allograft showing chronic ABMR.

Representative hematoxylin and eosin, and multiplex immunofluorescence staining performed on a kidney allograft biopsy (transplant explant biopsy) from a patient with incurable chronic ABMR, refractory to immunosuppressive treatments. Scale bars indicate 50μm.

d). Role of cT_FR cells in transplantation

In transplantation, cT_FR cells are increasingly described^76,77 and their frequency were found to be diminished in the context of acute and chronic rejection.⁷⁸ Moreover, using inducible T_FR cell deletion strategies, a recent study revealed that T_FR cells significantly inhibited de novo DSA formation, but had a minor role in controlling kidney allograft ABMR in mice.⁶⁵ However, in the context of DSA and ABMR responses in humans, their frequency, function and whether these cells can migrate to allograft tissues remains to be resolved. Thus, further attention on T_FR cells is needed in future studies to understand how to promote their development and suppressive function, as they could be instrumental to optimally control T_FH cell-dependent humoral responses.

Effects of current therapies on T_FH cells

Although the current induction therapies and maintenance immunosuppressive (IS) drugs used nowadays in the clinic were primarily designed to target and alter T cell responses, their intrinsic effects on T_FH cell differentiation and functions were not initially known but are now increasingly documented (Table 1). The effects of such therapies were measured by assessing cT_FH cells in patients but these effects on T_FH cells in GCs are also supported by multiple experimental models.

Table 1.

Effects of current therapies on human T_FH cells

Type of drug	Drug	Intrinsic effects on TFH cells	References
Induction therapy	Thymoglobulin	reduces cTFH numbers	63
		reduces cTFH numbers, induces proliferation of cTFH and promotes an effector memory (CD62L−) phenotype and their overexpression of CXCR3 and PD-1	68
		reduces cTFH numbers and their activation-induced upregulation of CD40L and ICOS	79
		reduces cTFH numbers and induces increased frequencies of PD-1+, ICOS+PD-1+ and PD1+CXCR3− cTFH at 12-months post transplant	64
	Basiliximab	no effect on cTFH numbers	63
		no effect on cTFH numbers and induces a central memory (CD62L+) cTFH phenotype	68
		no effect on cTFH numbers	79
Maintenance immunosuppression	Tacrolimus	reduces cTFH, GC-TFH precursor and TFH memory numbers. No effects on GC-TFH numbers. Inhibits PD-1 expression on cTFH and plasma cell formation in cTFH-B cell co-cultures	86
		inhibits naive CD4 T cell differentiation into TFH, reduces TFH expression of ICOS and PD-1, and their production of IL-21. Inhibits TFH-dependent proliferation and differentiation of B cells into plasma cells in TFH-B cell co-cultures	90
	Mycophenolate mofetil	reduces T cell expression of CD40L. inhibits T cell-dependent proliferation and differentiation of B cells into plasma cells in T-B cell co-cultures and in IL-21 stimulated B cells	96
		reduces T cell expression of CD40L and ICOS. inhibits T cell-dependent B cell production of IgG in T-B cell co-cultures	97
	Sirolimus	inhibits naive CD4 T cell differentiation into TFH. reduces TFH expression of ICOS and PD-1, and their production of IL-21. Inhibits TFH-dependent proliferation and differentiation of B cells into plasma cells in TFH-B cell co-cultures	90
		reduces total cTFH, PD-1+ cTFH and STAT3+ cTFH numbers	178
	Corticosteroids	reduces cTFH and Th1, Th2 and Th17-cTFH subset numbers	102

a). Effects of induction therapies on T_FH cells

Thymoglobulin, basiliximab and alemtuzumab, as compared to absence of induction therapy, have drastically reduced the incidence of acute rejection during the first year post transplant.

Thymoglobulin is a mixture of polyclonal IgGs recovered from the sera of rabbits or horses immunized with human thymocytes and therefore mostly aims to deplete the T cell compartment. Its intrinsic effects on cT_FH cells include a prolonged depletion in cell numbers, but also a shift from central (CD62L⁺) to effector (CD62L⁻) memory phenotype as well as the overexpression of PD-1 and CXCR3 consistent with pre-activated and Th1 phenotypes.^63,64,68 These effects are maintained throughout the first year post transplant. Conversely, the lymphopenia-induced proliferation of cT_FH cells was transient and maximum at 1-month post transplant and was observed selectively in patients further developing DSA in the first year post transplant, indicating that in these patients, thymoglobulin may promote a pre-activation state that may favor DSA development.⁶⁸ However, regardless of DSA development, thymoglobulin was able to diminish activation-induced upregulation of CD40L and ICOS on cT_FH cells, indicative of a partial control of cT_FH cell helper function.⁷⁹

Basiliximab is a monoclonal chimeric mouse-human antibody and nondepleting agent targeting the α-chain (CD25) of IL-2R. It was designed to reduce IL-2 binding to its receptor and therefore reduce IL-2 induced capacity for T cells to proliferate following activation. Unlike thymoglobulin, basiliximab does not induce cT_FH cell depletion nor proliferation, and maintain cT_FH cells in a central memory differentiation.^68,79 Given that IL-2 prevent T_FH cell differentiation and function, it could have been anticipated that basiliximab would promote increase cT_FH cell numbers and activation state via blockade of IL-2R, however this was not observed in the recent studies. IL-2/IL-2R is crucial for T_REG and T_FR optimal cell development and functions.⁸⁰ Is it therefore likely that basiliximab would also block the functions of T_REG and T_FR cells, thus preventing the suppression of T_FH cell response by these cells.

Alemtuzumab is a humanized monoclonal antibody directed against CD52, an immune receptor widely present on the surface of T and B cells. Alemtuzumab is known to induce rapid and profound T and B cell depletion,⁸¹ and to be associated with increased incidence of ABMR in patients, when compared to thymoglobulin.^82,83 This mechanism includes increased GC T_FH differentiation and promotion of IL-21 production by GC T_FH.⁸⁴

b). Effects of maintenance IS drugs on T_FH cells

In addition to induction therapy, the IS drug combination including calcineurin inhibitor (or mTOR inhibitor) with purine synthesis inhibitors (mycophenolate mofetil or azathioprine) and with corticosteroids, is currently considered as the standard-of-care maintenance IS regimen for organ transplant recipients. This IS combination regimen has dramatically reduced the incidence of acute rejection in the first year post transplant and transformed the clinical picture and spectrum of allograft rejection with increased incidence of subclinical and chronically evolving forms of ABMR. Although the T cell response is the central target of maintenance IS drugs, their effects on T_FH cell differentiation and functions were only recently described.

Calcineurin inhibitors (tacrolimus and cyclosporine) were designed to suppress calcineurin activity and activation of NFAT in T cells, which are induced after the engagement of TCR. Compared to cyclosporine and mTOR inhibitors, tacrolimus is more effective at preventing de novo DSA development in transplant patients.⁸⁵ This may be explained by the reduced numbers of CXCR5⁺PD-1⁺ cT_FH cells and CXCR5⁺PD-1⁺ GC T_FH precursors in patients treated with tacrolimus, while no effects were observed on B cells, T_FR and T_REGS. In vitro, tacrolimus is associated with significant reduction of PD-1 expression on cT_FH and of plasma cell formation in cT_FH-B cell co-cultures.⁸⁶ Tacrolimus also significantly reduces T_FH cell function and expression of Bcl-6 and of BATF/JUN/IRF4 complex, and reduces IL-6 and IL-21 production.⁸⁷

mTOR inhibitors (rapamycin and everolimus) engage FK binding proteins, forming a complex that inhibits mTOR in T cells. This inhibition prevents translation of mRNA-encoding proteins needed to enter the cell cycle and therefore impairing T cell proliferation and cytokine production.⁸⁸ In addition to this nonspecific mechanism, rapamycin is able to inhibit naive CD4 T cell differentiation into T_FH cells,^89,90 to reduce T_FH cell expression of ICOS and PD-1, and their production of IL-21.⁹⁰ Rapamycin also inhibits T_FH cell-dependent proliferation and diminished differentiation of B cells into plasma cells and antibody production in T_FH-B cell co-cultures to a degree similar to that achieved with T_FR cells.^90,91 This effect was also observed experimentally in vivo by increased heart allograft survival and reduced T_FH and B cell numbers post transplant in mice recipients treated with rapamycin compared to vehicle.⁹² Moreover, genetic studies have shown that while mTOR kinase complexes 1 and 2 may act distinctly, they are both essential for T_FH cell differentiation and GC responses under steady state and after antigen immunization.^93,94 However, in heart-transplanted mice treated with alemtuzumab, rapamycin promoted increased in PD-1⁺ICOS⁺ T_FH cells in lymph nodes, suggesting paradoxical effects on T_FH cells in post depletion settings.⁹⁵

Mycophenolate mofetil blocks cell division by inhibiting DNA synthesis and acts as a selective inhibitor of inosine-59-monophosphate dehydrogenase, a key enzyme in de novo synthesis of purine. While mycophenolate mofetil is known to be a significant inhibitor of B cell activation and proliferation, no studies have specifically investigated the effects of mycophenolate mofetil on T_FH cells, some have reported its effect on CD4 T-cell help to B cells. Two studies have shown that mycophenolate mofetil significantly reduced CD40L and ICOS expression on polyclonally activated T cells, and inhibited T cell-dependent proliferation and differentiation of B cells into plasma cells in T-B cell co-cultures and in IL-21 stimulated B cells.^96,97 This is consistent with clinical studies suggesting that mycophenolate mofetil may be efficient in preventing de novo DSA formation, when compared to azathioprine, another inhibitor of DNA synthesis.^98,99

Corticosteroids are among the oldest and most commonly used immune-modulatory drugs. Their mechanisms of action on T cells are numerous and include interference with cell receptors and their downstream signaling pathways and transcription factors, including TCR and its downstream kinases.^100,101 Specifically, corticosteroids were shown to markedly reduce cT_FH, Th1-, Th2- and Th17-cT_FH subset numbers in individuals with IgG4-related disease.¹⁰² Investigating the effects of corticosteroids on cT_FH cells in transplant patients are warranted, as a recent study showed an association between lack of corticosteroids and increased risk of subclinical ABMR.¹⁰³

Biotherapeutic targeting of T_FH cells

The success of biotherapy in organ transplantation has been recently illustrated by the use of belatacept, the first targeted biotherapy that received FDA approval for prophylaxis of organ rejection in kidney transplantation, which significantly diminish the incidence of de novo DSA and improve long-term allograft survival, when compared to cyclosporine.¹⁰⁴ It was then demonstrated that this decreased risk of de novo DSA development was explained by a direct effect of belatacept on T_FH cells and their crosstalk with cognate B cells.¹⁰⁵ In addition, the IL-6R antagonist tocilizumab, a biotherapy developed and FDA approved for treatment of rheumatoid arthritis, was also used in recent clinical trials in ABMR.¹⁰⁶ Its administration in patients with chronic ABMR resulted in stabilization of allograft function and DSA reduction. While biotherapeutic targeting of T_FH cells may result in complex immune modulation beyond T_FH cell responses, these successes further encourage and open the way for understanding the intrinsic effects of biotherapies on T_FH cells, and for developing novel biotherapies targeting these cells in organ transplantation (Table 2).

Table 2.

Biotherapeutic targeting of human T_FH cells

Type of drug	Drug	INN	Clinical setting	Trial registration	References
Costimulatory molecules agonists and antagonists	CTLA4-Ig	belatacept	transplantation	NCT00256750	104
	Anti-CD28	lulizumab	transplantation, Sjogren’s syndrome, lupus	NCT04066114, NCT02265744, NCT02843659	–
	Anti-CD40	bleselumab	transplantation	NCT01780844, NCT01279538	121, 122
	Anti-CD40L	dapirolizumab	lupus	NCT01764594	179
	Anti-ICOS	MEDI-570	lupus, lymphoma	NCT01127321, NCT02520791	–
	Anti-OX40	GBR 830	atopic dermatitis	NCT02683928	141
	Anti-CD70	cusatuzumab	acute myeloid leukemia	NCT03030612	147
Cytokines and cytokine receptors antagonist	Anti-IL-6	clazakizumab	transplantation, rheumatoid arthritis, COVID-19	NCT03444103, NCT01373151, NCT04363502	–
	Anti-IL-6R	tocilizumab	transplantation, rheumatoid arthritis, giant-cell arteritis, COVID-19	NCT01594424, NCT01119859, NCT01791153, NCT04356937	180, 181, 182, 183
	Anti-IL-21	BOS161721	lupus	NCT03371251	–
	Anti-IL-21	NNC114–0005	healthy subjects, rheumatoid arthritis	NCT01208506	163
	Anti-IL-21R	ATR-107	healthy subjects	NCT01162889	164
	Anti-IL-7Rα	OSE-127	healthy subjects	NCT03980080	–
	Anti-IL-7Rα	RN168	type 1 diabetes	NCT02038764	168
	Anti-IL-12/IL-23	ustekinumab	psoriasis, Crohn's disease, ulcerative colitis	NCT00454584, NCT00771667, NCT02407236	170, 171, 172

INN, international nonproprietary name

a). Costimulatory molecules agonists and antagonists

a1. CD28/CTLA4/B7 axis.

Therapeutic targeting of CD28/CTLA4/B7 has always been a primary goal in transplantation.¹⁰⁷ The use of a recombinant CD28-Ig that binds CD80/CD86 and blocks CD28 signaling was ineffective due to insufficient affinity for its B7 ligands, leading to the development of another recombinant protein: the CTLA4-Ig (belatacept).¹⁰⁸ Belatacept is a human fusion protein combining the extracellular portion of CTLA4, which has been mutated to confer greater binding avidity to CD80 (B7-1) and CD86 (B7-2), and the constant region fragment of human IgG1. CTLA4 binds CD80 and CD86 on APCs, preventing their interaction with CD28 and therefore inhibiting T cell activation. Although belatacept has been primarily developed to target DC-mediated priming of T cells, preclinical and recently human models have shown that belatacept could also unexpectedly inhibited antibody responses.⁶² Nonhuman primates treated with belatacept showed decreased in follicle size, GC formation and IL-21 production, indicating a role for belatacept in disruption of T_FH-B cell interaction.⁶² In mice models, delayed belatacept could inhibit ongoing antibody response, even though priming of allogenic T cells had already occurred. The administration of belatacept 14 days after sensitization successfully inhibited DSA production and collapsed GC responses.^109,110 These results suggested that CTLA4-Ig action on T_FH-B cell responses were prominent and independent of initial T cell priming by DCs. Leibler at al. confirmed that this mechanism also existed in humans.^105,111 Belatacept inhibited T_FH-B cell crosstalk in vitro by decreasing T_FH cell activation and decreasing B cell differentiation into plasma cells. Interestingly, belatacept-treated kidney recipients had lower frequencies of PD-1⁺ICOS⁺ cT_FH and effectors B cells than recipients on calcineurin inhibitors, indicating that the inhibition mechanism proposed also existed in vivo.¹⁰⁵ Consistently, interventional trials further confirmed that belatacept was associated with decreased incidence of de novo DSA, decreased DSA levels and also prevented the occurrence of ABMR.^112,113

More recently, agents more specifically inhibiting CD28 were developed, while leaving CTLA4 intact, to better control alloreactive T cell responses and prevent the undesirable T-cell mediated rejections observed in patients treated with belatacept. The use of such selective CD28 blockers in murine and primate models of transplantation has indeed revealed superior allograft survival results as compared to belatacept, an effect potentially dependent on the preservation of negative signaling through CTLA4.¹⁰⁷ Indeed, preservation of CTLA4-mediated signals with anti-CD28 is associated with enhanced T_REG functionality¹¹⁴ and better control of the proliferative response of preexisting T_FH cells than belatacept.¹¹⁵ In addition, anti-CD28 significantly reduces cT_FH, ICOS⁺PD-1⁺ cT_FH and GC T_FH cell numbers and inhibits GC T_FH cell proliferation, and GC T_FH cell-dependent B cell differentiation into plasma cells in GC T_FH-B cell co-cultures.^66,116 However, as loss of CD28 signals would also lead to alteration of T_REG and T_FR cell functions, the effects of anti-CD28 on these cells should be considered. The efficacy of the anti-CD28 lulizumab in combination with other adjunct therapies is currently tested in clinical trials in living donor kidney transplant recipients, or in patients with lupus or Sjogren’s syndrome (Table 2).

a2. CD40/CD40L axis.

CD40/CD40L molecules remain promising therapeutic targets in transplantation as blockade of this pathway has demonstrated strong effects on humoral responses. In a myriad of experimental models, inhibition of the CD40/CD40L axis resulted in markedly increased allograft survival and represents a potential way of inducing a tolerogenic environment.¹¹⁷ In mice, anti-CD40L therapy drastically reduces DSA production via terminating ongoing GC reactions.¹⁰⁹ In humans, interventional trials evaluating the safety and efficiency of anti-CD40L reagents were prematurely halted because of severe thromboembolic complications (related to CD40L expression on human platelets).¹¹⁸ Anti-human CD40, are now developed as safer and potential alternatives. Importantly, anti-CD40 also effectively disrupts GCs by reducing GC T_FH cell numbers and their IL-21 production, and attenuated DSA production in kidney transplanted monkeys.^62,119 Moreover, bleselumab (ASKP1240) abolished DSAs production in a monkey model of pancreatic islet transplantation.¹²⁰ Several clinical trials using bleselumab in kidney transplantation and lupus are ongoing (Table 2).^121,122

a3. LFA-1/ICAM axis.

LFA-1 (CD11a) is an integrin critical for trafficking of T_FH cells to GCs but also for delivering potent T cell costimulatory signals that augment Bcl-6 and their effector function.¹²³ In animals, anti-LFA-1 inhibits priming of naive alloreactive T cell responses, prolongs allogeneic cardiac and islet allograft survival^124,125 and is effective in mitigating donor-specific memory T cells that are independent of CD28 and CD40L.^126,127 Moreover, anti-LFA-1 reduces GC T_FH cell formation, serum IL-21 and DSA generation in a cardiac model of chronic ABMR following alemtuzumab induction.⁸⁴ In 2 clinical trials in allogeneic islet transplantation and 1 in kidney transplantation, patients were treated with the anti-LFA-1 efalizumab.^128-130 However, further trials using efalizumab were halted because of excessive incidence of progressive multifocal leukoencephalopathy and posttransplant lymphoproliferative disease.¹³¹

a4. ICOS/ICOSL axis.

ICOS/ICOSL axis can be blocked using recombinant ICOS-Ig or anti-ICOS antibody. So far, preclinical studies in nonhuman primates evaluating the efficacy of ICOS-Ig have shown unsuccessful results in preventing allograft rejection.^132,133 Anti-ICOS, however, appears to more effectively inhibit humoral responses; its administration resulted in decreased DSA generation in islet xenografts recipients,¹³⁴ and its combination with anti-CD40, also led to decrease in DSAs and improved histology as compared to anti-CD40 alone, in a cardiac model of chronic ABMR.¹³⁵ Anti-ICOS treatment appears promising, particularly when combined with other costimulatory blockade therapies, as compensatory pathways in B cells may overrule its requirement.¹³⁶ Several clinical trials in lupus and lymphoma using anti-ICOS are ongoing (Table 2).

a5. OX40/OX40L axis.

OX40 is a member of TNF superfamily of high interest in transplantation, particularly because of its prominent role in regulating alloreactive memory T cell and T_FH cell responses.¹³⁷ Although specific effects on T_FH cells were not evaluated, promising results showed that anti-OX40 significantly reduces alloreactive CD4 T cell survival and prevented skin allograft rejection in mice.^138,139 Moreover, alloreactive memory CD4 T cells that were relatively insensitive to either CD28 or CD40L blockade could be attenuated via anti-OX40L after skin allograft transplantation.¹⁴⁰ While the anti-OX40 GBR 830 is currently evaluated in atopic dermatitis, its role in clinical transplantation requires further investigation.¹⁴¹

a6. CD70/CD27 axis.

CD70/CD27 are other members of the TNF superfamily and mutually expressed on T_FH and B cells. CD70 activates PI3K/AKT pathway and stimulates cytokine production in T_FH cells,¹⁴² while CD70 synergize with the CD40 pathway¹⁴³ and play a role in B cell activation and antibody production in B cells.¹⁴⁴ CD27, the ligand of CD70, is expressed on activated/memory T_FH and B cells as well as plasma cells. In rats, anti-CD70 administration diminished memory T cells in recipient lymph nodes and prolonged corneal allograft survival.¹⁴⁵ Similar findings were observed in heart-transplanted mice treated with anti-CD70.¹⁴⁶ While there are ongoing clinical trials with the anti-CD70 cusatuzumab in acute myeloid leukemia, studies in transplant recipients are lacking.¹⁴⁷

a7. PD-1/PD-L1 axis.

PD-1 is the ligand for PD-L1, and is known to dampen TCR and CD28 signaling, to diminish the accumulation of T_FH cells and to control T_FH cell localization in GC.^148,149 In contrast to CTLA4-Ig that blocks T_FH cells from receiving costimulatory signals, or to anti-PD1 antagonists used in cancer immunotherapy, PD-L1-Ig induces an inhibitory signal through PD-1. PD-L1-Ig with or without anti-CD40L, was shown to prevent allograft rejection and prolong allograft survival in 2 experimental models; islet¹⁵⁰ and cardiac transplantation.¹⁵¹ However, as loss of PD-1 signals is also expected to enhance T_REG and T_FR cell functions,¹⁵² the effects of PD-L1-Ig on these cells should be considered. Clinical trials using PD-L1-Ig are lacking and warranted to validate these findings in humans.

b). Cytokines and cytokine receptor antagonists

b1. IL-6/IL-6R axis.

IL-6 plays a major role in T_FH cell differentiation and function¹⁵³ and in B cells, it promotes antibody production of all IgG subclasses¹⁵⁴ and plasmablast/plasma cell survival.¹⁵⁵ Plasmablasts produce large amounts of IL-6, which in turn, stimulate T_FH cells via IL-6R.¹⁵⁶ Anti–IL-6R (tocilizumab) administration resulted in reduction in T_FH and Th17 cells, and augmentation in T_REGS¹⁵⁷ in experimental models. In patients with chronic ABMR, tocilizumab resulted in stabilization of kidney allograft function and significant decreased in DSA MFI levels.¹⁰⁶ Anti-IL-6 (clazakizumab) recently showed promising results in late ABMR patients, including reduction of DSA levels and modulation of ABMR gene and histological activity (Table 2).

b2. IL-21/IL-21R axis.

While IL-21 is a pivotal cytokine for optimal T_FH-B cell crosstalk,⁴³ it may have double-edge sword effects in clinical settings.¹⁵⁸ IL-21R is highly expressed on T_FH cells but also on naive and activated memory B cells.¹⁵⁹ Donor-reactive cT_FH cells were shown to produce IL-21⁶⁸ and IL-21 is detected at high levels in serum in experimental models of ABMR.⁸⁴ In mice, anti-IL-21R administration could significantly delay skin allograft rejection¹⁶⁰ and anti-IL-21 significantly prolonged islet graft survival.¹⁶¹ In humans, anti-IL-21R led to reduction in STAT3 phosphorylation in cT_FH cells, increased plasma cells differentiation and IgG production in a coculture model of cT_FH-B cells stimulated with donor antigen.¹⁶² Safety and efficacy data are accumulating in clinical trials in lupus and rheumatoid arthritis for anti-IL-21,¹⁶³ and in healthy subjects for anti-IL-21R (Table 2).¹⁶⁴ Thus, these 2 blocking agents are not yet ready for prime time in transplantation.

b3. IL-7/IL-7R axis.

IL-7 is crucial for memory CD4 T cell development, homeostatic proliferation and long-live maintenance. It also plays a role in T_FH cell generation in vivo, and anti-IL-7 antibody markedly impairs the development of IgG responses.¹⁶⁵ Importantly, IL-7Rα (CD127) was shown to be significantly upregulated on cT_FH cell subset with a memory precursor phenotype during ABMR.⁷⁰ In mice, anti-IL7Rα has been shown to induce long-term skin allograft survival or islet allograft tolerance when combined with prior T cell depletion, by delaying T cell reconstitution, decreasing allospecific memory T cells and decreasing DSAs.¹⁶⁶ Conversely, in baboons, anti-IL-7Rα combined with thymoglobulin and tacrolimus did not promote increased allograft survival nor delayed T cell reconstitution.¹⁶⁷ Currently, clinical trials using anti-IL-7Rα are ongoing in healthy subjects and in type 1 diabetes.¹⁶⁸

b4. IL-12/IL-23/IL-12Rβ axis.

Unlike in mice, IL-12/IL-23 play an important role in the differentiation and polarization of human T_FH cells, therefore these cytokines represent potential targets to control humoral responses.¹⁶⁹ IL-12/IL-23 can be simultaneously targeted using anti-IL-12/IL-23 (ustekinumab), which has shown clinical efficacy in psoriasis, Crohn’s disease and ulcerative colitis.^170-172 Ustekinumab could be repurposed in transplantation, given that T_FH cell differentiation and IL-21 production strongly rely on both IL-12 and IL-23 signals.

c). Combination of targeted biotherapies

Targeting multiple cellular components and stages of the humoral alloresponse appears crucial for optimal control of DSA generation and ABMR. Of note, proteasome inhibitor (bortezomib), along with plasma cell depletion, paradoxically triggers homeostatic proliferation of GC T_FH and B cells, indicative of humoral compensation, a plausible explanation for the failure of the treatment to sustainably decrease DSA levels and reverse ABMR.¹⁷³ Therefore, most recent studies have investigated the combined effects of proteasome inhibitor (carfilzomib) combined with anti-CD28 (lulizumab) and showed reduction in lymph node T_FH cells, DSA levels and improved kidney allograft survival in allosensitized nonhuman primates.¹⁷⁴ Similar beneficial effects on T_FH cells and DSAs were observed using carfilzomib combined with belatacept¹⁷⁵ and using bortezomib combined with belatacept and anti-CD40¹⁷⁶ in the same primate transplant model. In humans, bortezomib combined with belatacept also resulted in decreased PD-1⁺ICOS⁺CD38⁺ cT_FH cells and DSAs in highly sensitized heart transplant candidates.¹⁷⁷ Thus, mitigating both upstream T_FH cell responses and terminal effector plasma cell responses appears promising for increase efficacy in depletion of DSAs and for prevention of rebound effects.

In conclusion, the recent recognition of the pivotal role of T_FH cells in directing DSA generation leading to ABMR has significantly improved our understanding of the cellular and molecular cues underlying the pathophysiology of ABMR. Although current maintenance IS drugs and induction therapies appear to have beneficial effects in tempering T_FH cell responses, further efforts are needed to develop efficient combinatorial biotherapies targeting T_FH cells, as well as other cellular participants to the humoral immune response for the prophylaxis and curative treatment of ABMR. This will represent an important step towards personalized and precision medicine in the field of organ transplantation.

Abbreviations

ABMR

antibody-mediated rejection

Bcl-6

B cell lymphoma-6

Blimp-1

B lymphocyte-induced maturation protein-1

cT_FH

circulating T follicular helper cell

cT_FR

circulating T follicular regulatory cell

CTLA4

cytotoxic T-lymphocyte-associated antigen 4

DC

dendritic cell

DSA

donor-specific antibody

GC

germinal center

HLA

human leucocyte antigen

ICOS

inducible T cell co-stimulator

IFN-g

interferon-gamma

Ig

immunoglobulin

IL

interleukin

IRF4

interferon regulatory factor 4

LFA-1

lymphocyte function-associated antigen 1

mTOR

mammalian target of rapamycin

PD-1

programmed cell death-1

STAT

signal transducer and activator of transcription

T_FH

T follicular helper cell

T_FR

T follicular regulatory cell

T_PH

T peripheral helper cell

T_REG

T regulatory cell

TNF

tumor necrosis factor