Besides the transfer of pro-inflammatory cells, which is heightened in DACDs, the relevance of immunosuppressive drugs on T_FH cells after kidney transplantation and transplant longevity is being increasingly appreciated, too. Although CD4⁺ T_FH cell subpopulations reciprocally orchestrate their preeminent transcriptional regulation, immunosuppressive drugs can have a profound effect in governing subset skewing (96). Thymoglobulin induction therapy (e.g., antithymoglobulin/ATG) was reported to deplete effector CD4⁺ and CD8⁺ T cells, whilst preserving allograft permissive FoxP3⁺ T_FR cells (97, 98). This warrants further studies investigating possible therapies that draw on T_FR cell-mediated effects. This could potentially enable long-term drug-free allograft permissiveness (96, 99). This appears to be of particular importance since kidney transplant patients who have suffered from prior AMR were found to display higher ratios of IL-21⁺ T_FH cells whilst their T_FR cell population was decreased both within the graft and in the circulation. Besides, sirolimus was also found to reduce the T_FR cell population even further (100). Permissiveness may also be facilitated actively by B cells since B cell depletion or IL-10 deficiency were shown to skew the tolerogenic environment towards an increased IL-21⁺ T_FH cell population decreasing T_FR cells in follicles ( Figure 1 ) (101). By extension, basic leucine zipper ATF-like transcription factor (BATF) inhibition, a transcription factor for T_H-17 and T_FH-17 cells alike, could be linked to enhanced FoxP3 levels coinciding with a reduction in retinoic-acid-receptor-related orphan nuclear receptor gamma (RORγt), IL-17A, and IL-4, thus, generating tolerance after transplantation (102). Calcineurin inhibitors like cyclosporine A also impact on T_FH cells. Although no effect on T_FH subtypes concerning T_FH-1, T_FH-2, and T_FH-17 cells was observed in healthy volunteers under transient cyclosporine A medication, a profound reduction in the pro-inflammatory markers IFN-γ, IL-17A, and IL-21, produced by T_FH-1 and T_FH-17 cells, respectively, was observed (103). This poses the question how the commonly employed permanent triple immunosuppressive therapy comprising not only of calcineurin inhibitors, but also corticosteroids and inosine-5’-monophosphate dehydrogenase (IMPDH) inhibitors, all known suppressors of T cell functionality (103–105), can lead to T_FH cell dysregulation culminating in alloantibody formation and potentially allograft loss.

Together, this argues for a better understanding of the effects of the lifelong post-transplant immunosuppressive regimen on the follicular and extrafollicular immune cell compartment to prevent sensitization to donor-specific HLA molecules or autologous non-HLA molecules with an eminent focus on the control of prevalent T_FH cell responses whilst maintaining a strong T_FR surveillance to induce allotolerance in kidney transplant recipients (KTxs).

Under physiological circumstances, the evolution of a humoral memory (LLPCs) allows high-affinity antibodies to be established within a matter of days upon re-exposure of an antigen to protect us from microbial threats (106). However, auto- and alloantigenic settings constantly expose T_FH cells to antigens and disturb modulating checkpoints provided by T_REG cells, T_FR cells, follicular cytotoxic T (T_FC) cells, or even regulatory B cells. Hence, the excessive T_FH cell stimulation may take a life of its own. Strategies to prevent this from happening exploit cytotoxic T-lymphocyte-associated protein 4 (CTLA-4)-specific immunoglobulin or IL-21 receptor (IL-21R) antagonists to prevent alloimmune responses (107, 108).

Hence, the surveillance of immune subtype compositions may turn out as a valuable tool in identifying KTxs at risk for the formation of auto- and alloantibodies and subsequent allograft rejection.

Considering the selective influence on T_FH cells, studies on autoimmune diseases have shaped our understanding of persistent T_FH cell stimulation resulting in dysregulated responses. For example, unrestricted clonal expansion of T_FH cells allows the generation of ectopic lymphoid structures (ELS), which are unphysiological tertiary lymphoid follicles featuring stimulation and clonal expansion of antigen-specific B cells (109). These ELS are being established because of an antigen response yet to be cleared (110), and are formed by a regulated and well-orchestrated expression of the lymphoid chemokines C-C motif ligand (CCL)19, CCL21, and CXCL13 (111–113). Persistence of these lymphokines (with T_FH cells being a major source of CXCL13) (9) allows PSGL-1^LOWCD40L⁺ T_FH-17 cells to maintain aggregates in an ICOS-dependent manner with B cells in extrafollicular ELS, inevitably preserving an immune response with a serious risk for the formation of autoantibodies of the IgG2a/IgG2b or IgG1/IgG3 type, tissue destruction, and ultimately the development of an autoantibody-mediated pathology (114–116). Autoimmune conditions and others have been shown to be associated with ELS, which are a major compartment for ample extrafollicular T_FH cell accumulation (117). To date, several studies have implied T_FH cells especially in the context of glomerulonephritis where IL-17 drives inflammation and autoantibody-induced kidney injury which can be considered to be key determinants for autoimmune disease activity (118–121), whilst a disrupted T_FH cell response has been shown to reduction disease activity, respectively (122–124). It appears intuitive that especially endothelial injury and incessant inflammation with enhanced HLA class II up-regulation can expose autoantigens for T_FH cell sensitization (125, 126). Despite the predominance of literature, particularly from human studies, on the importance of ELS for chronic inflammation caused by infections, autoimmune diseases, cancer, or transplantation (127), the contribution of derailed T_FH cell activity in the induction of autoimmunity within secondary lymphoid tissues (SLOs) including the spleen, lymph nodes, and mucosal-associated lymphoid tissue (MALT) (128), may be equally important. For example, a murine study investigating the importance of T_FH cells in the induction of rheumatoid arthritis found T_FH cell frequencies to be significantly increased in the spleen and joint-draining lymph nodes following disease induction. Here, CXCR5^-/- mice were protected from autoimmune arthritis by abrogating T_FH-B cell interactions within SLOs stressing its importance in the induction of autoimmune inflammation (129). In fact, not only murine models show that immune responses in SLOs may not only precede the formation of ELS and tertiary lymphoid organs (TLOs) (130), but also maintain an active role during autoimmune conditions as is observed in patients suffering from rheumatoid arthritis, who feature follicular hyperplasia and active GC responses in SLOs (131–133).

Together, a skewed T_FH cell response featuring lack of T_FH-17 cell control can lead to the establishment of ELS enhancing the likelihood of T_FH-17-B cell aggregates within these extrafollicular spaces for alloantibodies and alloreactive LLPCs to occur. Understanding how the disrupted T_FH cell subset equipoise can be restored may hold therapeutic potential to circumvent the need for immunosuppressive therapies with its inherent adverse effects and risk for infections (134).

Although more and more details are being understood regarding characteristics of HLA epitopes to perform appropriate epitope matching in clinical settings, there remains much to be learned. To date, HLA-matching in kidney transplantation is only performed for HLA-A, -B, -C, -Bw4, -Bw6, -DR, -DR51/52/53, -DQA1, - DQB1, and -DPB1 antigens (135). Nevertheless, mismatches are mostly inevitable and enhance the risk of dnDSA formation (136). However, it appears that not every epitope is equally immunogenic. Aubert et al. have found that low-level donor-specific HLA antibodies remain controversial in terms of predicting the risk of graft failure (136, 137). Indeed, studies have proposed dnDSA directed at HLA-DQ to be a risk factor for late allograft failure that feature histological alterations in accordance with chronic AMR (138–140).

Although dnDSA appear to form at relatively low frequencies (~15%) and often only several years after transplantation, they commonly facilitate detrimental consequences (141, 142). Independent risk factors were reported to be HLA-DR mismatches and non-adherence to immunosuppressive therapy. Upon emergence, dnDSA could progress and show antibody-mediated graft injury without impaired graft function (142).

Evidence suggesting AMR to be a consequence of the presence of dnDSA effectively challenged the dogma of calcineurin inhibitor toxicity and chronic allograft nephropathy (141). In fact, capillaritis was found to forecast allograft dysfunction culminating in AMR (143–145). Not much later, Lefaucheur et al. identified four distinct patterns of kidney rejection with capillaritis in association with AMR illustrating the poorest outcomes (146). Not all dnDSA could be paired with ensuing AMR (147, 148). In that regard, dnDSA properties would matter with respect to MFI, complement-binding capacity, and IgG subclass composition (147, 149). Conversely, the clinical phenotype of AMR cannot be linked with the presence of circulating dnDSA in all cases challenging the criteria for diagnosis. In that regard, evidence comprised histological acute tissue injury, current antibody interaction with endothelium as in C4d deposition, and serological evidence of dnDSA detection or non-HLA donor-specific antibodies (150, 151). The identification of intragraft dnDSA has been proposed to enable clinicians to correctly diagnose AMR, which would not meet these criteria (151, 152). However, if serologically detectable, dnDSA presence has been attributed to a poor graft outcome imposed by AMR (139, 149, 153–155).

A clear link between AMR and T_FH cells, particularly regarding IL-21 production, has been established (156). KTxs suffering from chronic allograft rejection were found to display distinctive increases in circulating T_FH cells with impaired controllability given a reduced PD-1 expression (157). Accordingly, PD-1 expression fine-tunes T_FH cell responses by suppressing follicular T cell recruitment, confining T_FH cell localization within GCs, and increasing the stringency of GC affinity selection via suppressed phosphoinositide 3-kinase activity upon PD-L1 ligation (158). Another study found stable circulating T_FH cell numbers with decreased IL-21 production. However, their ability to stimulate alloantigen-specific B cells to produce IgG was maintained (159). Stable numbers of T_FH cells have also been found by Chen et al., however, indicating a skewing toward IL-21-producing T_FH-17 and T_FH-2 cell subpopulations in AMR (100). To control this dysregulation of T_FH cells, Rodriguez-Barbosa et al. have found that whilst the CD40/L pathway could be used, the B and T lymphocyte attenuator (BTLA) pathway was dispensable (160). Exploiting the CD40/L axis revealed reduced clonal B cell expansion associated with curtailed GC-T_FH cell numbers and a blunted IL-21 secretion (161). It appears that challenging cellular receptor-ligand interactions between T_FH-B cells may alleviate the burden of AMR, which studies estimate to be responsible for 30-50% of allografts to fail (156). Likewise, protection from AMR could be mediated by strengthening the control of T_FH cells and Ag-specific B cells by T_FR cells (96, 100, 101, 155).

Therefore, T_FH cells are not only capable of inducing dnDSA formation but can also further augment preexisting DSA levels following alloantigen recall (162). Conversely, T_FR cells curtail T_FH cell-directed B cell help by preventing dnDSA formation. Whilst lack of T_FH cells control drives severe AMR, T_FR cells are less involved in this process (162). These observations appear to translate to the human setting. Here, KTxs with immunogenic tolerance towards their allograft have disrupted T_FH cell functionality characterized by lack of IL-21 production, although in humans T_FR cells control AMR (100, 163). In a similar notion, IL-21⁺ T_FH cells and activated B cell responses (ASCs, CD86⁺CD38⁺) together with serum IL-21 levels were proposed as biomarkers for AMR in KTxs (164, 165). A more profound understanding if certain T_FH cell-related pathways predisposed to the formation of complement-fixing or non-complement fixing dnDSA, would help to develop non-invasive biomarker-guided risk stratification and molecular refinement tools to prevent the emergence of dnDSA altogether. In fact, Louis et al. recently demonstrated that T_FR cells and transitional B cells were selectively reduced in KTxs with AMR (166). Characteristically, both populations comprising of CXCR5⁺ T_FR cells and CD21^- transitional B cells that had vanished expressed T-bet. Their loss coincided with enhanced inflammatory antibody responses, microvascular inflammation, and allograft failure (166). Previous studies have highlighted the importance of T-bet to act as a repressor of PD-1 and other inhibitory receptors such as LAG-3, CD160, and BTLA in adaptive immune cells (167), possibly identifying T-bet expression as a canonical immune checkpoint to drive alloreactivity in T cells ( Figure 2 ) (168). However, it must be recognized that noncanonical pathways in the absence of T_FH-B cell interactions need to be considered that may also contribute to the production of alloantibodies or certain subtypes of allospecific immunoglobulins such as IgG2c where T_FH cells may be dispensable (169, 170).

Subclinical rejection episodes, a sequelae of aberrant allosensitization possibly causing loss of T_FR cells, are considered prognostic indicators for chronic AMR (171, 172). Adjusting signaling pathways culminating in the absence of IL-21 production in T_FH cells prevents dnDSA formation due to lack of cognate B cell help and thus fosters allotolerance (163). Hence, defying allosensitization of T_FH cells, e.g., via engagement of the co-stimulatory blockade receptor CTLA-4 (100, 173), enhanced mTOR immunosuppression (174, 175), blockade of the CD40/L axis (67, 160, 161), or antagonizing antibodies against the IL-21R (108, 156, 159), may allow to further refine or counteract chronic AMR in kidney allografts. This strategy may promote longevity of kidney transplants and the quality of life in transplanted patients. Mechanisms underlying this deleterious allosensitization of naïve T cells to becoming alloreactive T_FH cells are conceivable. Comparing immunosuppression with tacrolimus against co-stimulation blockade of CD80 and CD86 using the CTLA-4-Ig belatacept conferred diminished seroconversion rates to influenza vaccination (176). Hence, understanding how to prevent the selection of alloreactive T_FH cell effector and memory clones will be an important area of future investigations. For example, a recent cohort study in patients with a positive dnDSA status showed an enhanced alloreactive T_FH cell pool in response to donor-specific HLA antigens. Despite continuous immunosuppression compared with healthy controls, an augmented IL-21 production and proliferative response in these T_FH cells upon stimulation in vitro was reported suggesting either incessantly activated or more easily recruited signaling domains even years after kidney transplantation (103). Further studies are needed to better define how to circumvent unwanted allograft-directed immune responses, whilst maintaining T cell-mediated antigen-specific B cell help in the context of the host’s immune protection and response to vaccination. GC responses emerge as prominent areas to be more meticulously studied. For example, KTxs were shown to lack GC responses to mRNA vaccination against severe acute respiratory syndrome coronavirus type 2 (SARS-CoV-2), a feature well characterized to be associated with neutralizing antibodies against SARS-CoV-2 (177). A prominent feature of this missing GC response in this study was a blunted T_FH cell signature, which is considered instrumental in orchestrating architecture and functionality of GCs (6–10). One may speculate that allograft-draining lymph nodes’ GCs may present a unique architecture in patients having developed dnDSA compared to other allograft-unassociated draining lymph nodes. Recent studies have shown that the trauma of the surgical intervention requires lymphatic vessels to heal, which in murine kidney transplantation showed lymphatic endothelial cells to abundantly release CCL21 by which stromal lymph node remodeling was fostered. Accordingly, dendritic cell enrichment was observed thus increasing the chance for successful alloreactive T cell priming. The causative role of the lymphatic system pertaining to the kidney allograft could be established in retrospective analyses in humans (178). Furthermore, the formation of TLOs within the allograft itself were observed in human kidneys undergoing chronic rejection (179, 180). Besides the formation of de novo lymphatic angiogenesis, the contribution of increased lymphatic flow may be another factor by which cellular trafficking, alloimmunity, and vasculopathy are being propagated, which in the context of heart transplantation found donor cell-trafficking to allograft-draining lymph nodes, increased lymphatic vessel are, and allograft infiltration of CD4⁺ and CD8⁺ T cells as well as CD68⁺ macrophages (181). In fact, draining lymph nodes and TLOs after small-bowel transplant rejection were enriched in CXCR3⁺ host T cells stimulated by donor-derived type 1 helper T cell-related chemokines (IP-10) suggesting their possible contribution also in other solid organ transplantation contexts such as the kidney (182). Since alloreactive T cells are susceptible toward inhibition by standard immunosuppressive drugs such as corticosteroids or calcineurin inhibitors, an advanced understanding of potentially derailed signaling pathways downstream of calcineurin in cases where patients are suspected of allograft rejection despite appropriate drug levels is needed (183).

Recent studies highlight the importance to better understand cytotoxic T cell responses in GCs. Some cytotoxic T cells can acquire CXCR5 expression (28), thus enabling them to enter GCs. Compared with STAT5⁺CXCR5⁺CD8⁺ T_FC cells, PD-1⁺CXCR5⁺CD8⁺ T_FC cells were found to be a biomarker for AMR. In this study, PD-1⁺CXCR5⁺CD8⁺ T cells were associated with chronic allograft dysfunction following kidney transplantation (184). Indeed, CXCR5⁺ CD8⁺ T cells with IFN-γ-producing abilities can be sampled from the peripheral blood in higher quantities in KTxs who remain DSA-free (185). Studies in a murine system of AMR using adoptive cell transfers in CCR5 KO mice equally show reduced frequencies of CXCR5⁺CD8⁺ T_FC cells following AMR of the kidney transplant (186), which confer cytotoxicity towards alloprimed IgG⁺ B cells. This could be a means by which incessant adaptive immune cell activation in allograft-draining lymph nodes may be held in check in transplanted patients who remain dnDSA-free.

Together, these studies implicate both T_FH cell and T_FR cell subsets as relevant entities to risk-stratify patients concerning potential dnDSA formation (89). Further studies are needed to better define lineage-commitment trajectories of T_FC cells to comprehend how to therapeutically intervene in cases of ongoing AMR. This may hold the potential to maintain allograft longevity even in cases of severe AMR.

Having discussed the relevance of T_FH cells and their cellular modulators for the formation of HLA-specific dnDSA and the importance of T_FH cell subsets for the disclosure of autoantigens and autoimmune conditions to emerge, the relevance of non-HLA antibodies remains to be outlined.

Just over a decade ago, Dragun et al. reported refractory vascular rejection in KTxs caused by non-HLA autoantibodies against the angiotensin II type 1 receptor (AT₁R) (187). Similarly, transplant glomerulopathy (TG) a consequence of autoantibody formation against perlecan and agrin, both compounds of the glomerular basement membrane, was described (188). It appeared that the risk for TG was even enhanced if both dnDSA and non-HLA antibodies were coincidentally present (189). To address these descriptions, a study conducted by Amico et al. found around 2.3% of patients of their cohort, who experienced early AMR, to be due to non-HLA antibodies (190). Progressive loss of self-tolerance to the autoantigens k-α-1 tubulin, collagen type V, and the collagen I was linked to increased risk of primary graft dysfunction in lung transplantation, which in turn would augment alloimmune responses inclining towards bronchiolitis obliterans syndrome (BOS) (191, 192). The importance of activating antibodies against AT₁R is by far the most thoroughly studied. AT₁R and the C-terminal fragment of perlecan (LG3) were studied in pregnant transplanted women and appeared to trigger allograft rejection. The mechanism of formation did not require allosensitization and a lack of correlation suggested different mechanisms of generation (193). It was suggested that renal ischemia, alterations to the intragraft microenvironment, and alloimmunity may be some of the various factors predisposing for AT₁R antibodies, which augment the ischemic condition by further contraction of the renal vasculature (187, 194, 195). The exposure of other cryptic antigens culminates in the emergence of non-HLA antibodies, which was described for the antigens LG3, vimentin, the endothelin type A receptor (ET_AR), and other extracellular proteins and intermediate filaments. But also, more recently, their ability to bind to antigens present on apoptotic cells activating complement was appreciated ( Figure 3 ) (195–198).

Evidence also suggests F_C-independent effects of non-HLA antibodies like augmented neointima formation with an accumulation of smooth muscle cells, mesenchymal stem cells advocating for their regulatory function via ERK1/2 signaling and interactions with α2β1 integrins in obliterative vascular remodeling during rejection (199, 200). But also other conditions have seen the discovery of non-HLA autoantibodies including systemic sclerosis with or without associated pulmonary arterial hypertension (201, 202).

It was shown that apoptotic cells release exosome-like vesicles, which, mediated by the 20S proteasome, incite non-HLA antibody formation. This coincided with enhanced T_FH and GC B cells (203). The conceptual framework of apoptosis taking place constantly may allow us to explain why some non-HLA autoantibodies may be present pre-transplant. In fact, a study by Nagele et al. reported ample naturally occurring immunoglobulin (Ig) G autoantibodies, which were influenced by age, gender, and disease (204).

Together, copious amounts of studies highlight the importance of auto- and allosensitization of T_FH cells in generating a permissive environment to stimulate cognate Ag-specific B cells toward the production of both non-HLA autoantibodies and dnDSA. Therefore, clinicians need to carefully monitor patients to identify derailed homeostatic interactions even in the absence of measurable auto- or allosensitization (205). It is noteworthy that compartments reflecting dysbalanced interactions (ELS, allograft draining lymph nodes) are not part of the regular clinical routine suggesting that further research is needed to define appropriate biomarkers that may reflect the current state of health of these compartments. Moreover, a more fundamental understanding about how these antibodies evolve from a T_FH cell perspective may identify interventions potent to either decrease the intensity of the antibodies detected or even reverse their presence after all.