Kidney Failure Associates With T Cell Exhaustion and Imbalanced Follicular Helper T Cells

Individuals with kidney failure are at increased risk of cardiovascular events, as well as infections and malignancies, but the associated immunological abnormalities are unclear. We hypothesized that the uremic milieu triggers a chronic inflammatory state that, while accelerating atherosclerosis, promotes T cell exhaustion, impairing effective clearance of pathogens and tumor cells. Clinical and demographic data were collected from 78 patients with chronic kidney disease (CKD) (n = 42) or end-stage kidney disease (ESKD) (n = 36) and from 18 healthy controls (HC). Serum cytokines were analyzed by Luminex. Immunophenotype of T cells was performed by flow cytometry on peripheral blood mononuclear cells. ESKD patients had significantly higher serum levels of IFN-γ, TNF-α, sCD40L, GM-CSF, IL-4, IL-8, MCP-1, and MIP-1β than CKD and HC. After mitogen stimulation, both CD4+ and CD8+ T cells in ESKD group demonstrated a pro-inflammatory phenotype with increased IFN-γ and TNF-α, whereas both CKD and ESKD patients had higher IL-2 levels. CKD and ESKD were associated with increased frequency of exhausted CD4+ T cells (CD4+KLRG1+PD1+CD57−) and CD8+ T cells (CD8+KLRG1+PD1+CD57−), as well as anergic CD4+ T cells (CD4+KLRG1−PD1+CD57−) and CD8+ T cells (CD8+KLRG1−PD1+CD57−). Although total percentage of follicular helper T cell (TFH) was similar amongst groups, ESKD had reduced frequency of TFH1 (CCR6−CXCR3+CXCR5+PD1+CD4+CD8−), but increased TFH2 (CCR6−CXCR3−CXCR5+PD1+CD4+CD8−), and plasmablasts (CD3−CD56−CD19+CD27highCD38highCD138−). In conclusion, kidney failure is associated with pro-inflammatory markers, exhausted T cell phenotype, and upregulated TFH2, especially in ESKD. These immunological changes may account, at least in part, for the increased cardiovascular risk in these patients and their susceptibility to infections and malignancies.


INTRODUCTION
Despite significant improved survival in patients with chronic kidney disease (CKD) and end-stage kidney disease (ESKD) in the past decades, the all-cause mortality among these patients has remained exceedingly higher when compared to general population (1,2). In particular, cardiovascular diseases, infections, and malignancies represent the leading causes of the mortality independent of concomitant comorbidities, such as diabetes or hypertension, suggesting a unique and systemic pathobiology in kidney failure (3)(4)(5).
In response to pathogen exposure or tissue damage, the body generates physiological and necessary inflammation to restore homeostasis. However, prolonged inflammatory signals alter T cell functions, leading to a state of exhaustion (6). Exhausted T cells, both CD4 + and CD8 + , are characterized by the loss of their effector functions and the expression of inhibitory receptors such as PD-1 and KLRG1. Co-expression of these inhibitory receptors can be of additive effects (6). Functionally, exhausted T cells have impaired cytokine production upon stimulation, but retain some cytotoxicity and degranulation capability (7). Exhausted T cells might be reversed back to fully functional T cells; however, these functional changes become irreversible if stimulus such as inflammation sustains, and eventually lead to cell death (6,8). Clinically, T cell exhaustion has been associated with unresolved infection and drug-resistant tumors, and emerging immunotherapy such as anti-PD-1 was designed to reverse functional deficits of exhausted T cells (7,9). In addition, a single-cell analysis also suggested the existence of T cell exhaustion in the development of atherosclerosis (10). Therefore, it is possible that significantly increased risks of cardiovascular death, infection, and malignancy in kidney failure are mediated by exhausted T cells, as a markedly increased T cell exhaustion was seen in a pediatric CKD cohort (11). Follicular helper T cells (T FH ), a subset of CD4 + T cells expressing CXC-chemokine receptor 5 (CXCR5), was initially discovered in the germinal centers (GC) of human tonsillar tissues (12).
To produce high affinity antibodies against pathogens, B cell requires the assistance from T FH to complete its maturation and class switch recombination inside the GC. Work related to T FH in the kidneys has largely focused on kidney transplantations, and upregulated T FH has been implicated in less favorable outcomes especially due to increased humoral rejections (13).
On the other hand, T FH are also linked to atherosclerosis, autoimmune diseases and some lymphoid tumors, suggesting their potential pathogenic role in promoting chronic inflammation (12,14).
To date, data on T cell exhaustion or T FH in adult CKD and ESKD remains scarce and incomplete. We hypothesized that an association between impaired renal function and exhausted T cell as well as follicular helper T cells, might explain the increased risk of infection, cancer, and cardiovascular disease among CKD and ESKD patients (15)(16)(17)(18)(19)(20), and performed a comprehensive immunophenotyping on T cells.

Study Population
We included 42 CKD patients (stage 1-5 or with proteinuria) seen at Complejo Hospitalario de Navarra, Pamplona, Spain (Biobank Navarrabiomed; n = 32), or S. Orsola-Malpighi Hospital, University of Bologna, Bologna, Italy (n = 10), 36 ESKD patients who received hemodialysis for > 3 months at the Mount Sinai Kidney Center (n = 21), Northwestern Hospital in Chicago (n = 13), or Complejo Hospitalario de Navarra, Pamplona, Spain (Biobank Navarrabiomed; n = 2), and 18 agematched healthy controls (HC). All individuals were enrolled from November 2017 to December 2019 ( Table 1). Exclusion criteria were pregnancy, recent (< 3 months) infectious episode requiring hospitalization, history of kidney transplant, use of immunosuppressive medications at the time of enrollment, inability to give consent, active malignancy. Samples and data from patients included in this study were processed following standard operating procedures with the appropriate approval of the Ethics and Scientific Committees of the participating centers.

Flow Cytometry Analysis
PBMC were isolated from peripheral blood by Ficoll gradient and frozen for batched analysis. We designed six multicolor flow cytometry panels to quantify 60 T cell subsets along with two B cell subsets, and calculated the CD4 + /CD8 + T cell ratio (Supplementary Table 1

Intracellular Cytokine Staining
We performed intracellular staining for IL-17, IL-2, IFN-γ, and TNF-α with extracellular markers for CD4 + and CD8 + after 5 h stimulation by mixture of PMA (Fisher, Waltham, MA), ionomycin (Fisher) and Golgi Plug (Millipore Sigma, Burlington, MA) at 37 • C. We fixed and permeabilized the cells using Intracellular Fixation and Permeabilization Buffer Set (eBioscience), according to the manufacturer's instructions. Data were acquired (> 1 × 10 6 events) on a 3-laser FACSLyric flow cytometer (BD Biosciences) and analyzed with FlowJo R software. Intracellular cytokines were recorded as percentages of CD4 + and CD8 + T cells.

Statistical Analyses
Results were presented as mean and standard deviation or standard error unless stated otherwise. Comparison of continuous variables between groups was performed by Kruskal-Wallis test, unpaired t-test, and categorical variables by two-sided chi-square, where applicable. P < 0.05 was considered as statistically significant. No correction was made for multiple testing. Statistical analysis was performed using GraphpadPrism R version 8.4.2 software package (Graphpad Software Inc., San Diego, CA).

Naïve, Effector, and Memory CD4 + T Cells
Consistent with prior reports (17)(18)(19)(20), we observed fewer CD3 + T cells in CKD (39.5 ± 13.8%) and ESKD (30.3 ± 12.8%) when compared to HC (48.4 ± 10.1%) (Figure 2A). In total T cells, there was no difference in the percentage of total CD4 + T cells among the three groups. Between CKD and ESKD, ESKD group had a significantly lower percentage of total CD8 + T cells with a higher CD4 + /CD8 + ratio (Figures 2B-D). Absolute numbers of CD4 + and CD8 + T cells of CKD and ESKD groups demonstrated similar findings to the percentage difference (Supplementary Figure 2), although we did not have absolute numbers in healthy controls during the collection process.

Dysfunctional and Regulatory CD4 + T Cells
Chronic inflammation has been associated with the emergence of T cell dysfunction (6). To test the hypothesis that a similar phenomenon occurs in patients with kidney failure, we measured phenotypically exhausted and anergic T cells. We found a significant increase of exhausted CD4 + (KLRG1 + PD1 + CD57 − ) and anergic T cells (KLRG1 − PD1 + CD57 − ) in both CKD and ESKD patients. (Figures 3H,I). Importantly, exhausted CD4+ T cells showed no cytokine production (Supplementary Figure 4).
Previous reports showed reduced levels of regulatory T cells (Treg) (21), a cell subset of CD4 + T cells essential for immune tolerance and eliminating chronic inflammation (22). Herein, we found a markedly decreased frequency of active Treg cells (CCR4 + CD45RA − CD25 + CD127 low− CD4 + CD8 − CD3 + ) in both CKD and ESKD groups, whereas the percentage of total Treg (CD25 + CD127 low CD4 + CD8 − CD3 + ) in ESKD group was statistically higher than CKD but of no difference compared to HC (Figures 3J,K).

Naïve, Memory, Effector, and Dysfunctional CD8 + T Cells
Similar to CD4 + T cells, we observed a decreased frequency of naïve CD8 + T cells in CKD patients, while that of ESKD patients was close to HC (HC: 19.2 ± 11.1% vs. CKD: 10.8 ± 9.1% vs. ESKD: 14.83 ± 11.0%) ( Figure 4A). ESKD patients had the lowest effector and memory CD8 + T cells among the groups (Figures 4B,C). In intracellular cytokine analyses, IFN-γ and TNF-α were significantly elevated after mitogen stimulation in ESKD. We also noticed a non-significant trend toward higher IL-2 levels in CKD and ESKD groups, but no difference in IL-17 across the groups (Figures 4D-G). Similar to CD4 + T cells, we found significantly more exhausted and anergic CD8 + T cells in ESKD patients (Figures 4H,I).

DISCUSSION
Our study shows that kidney failure is associated with a proinflammatory state, characterized by increased T cell production of pro-inflammatory cytokines and high circulating levels of T FH2 and plasmablasts. These alterations may lead to a progressive dysfunction of immune cells, as supported by increased exhausted CD4 + and CD8 + T cells. Altogether, these findings were suggestive of persistent and active pro-inflammatory T cells in both CKD and ESKD patients. Chronic inflammation can be triggered by reciprocal activation of innate and adaptive immunity (26). In our serum cytokine analysis, we identified a significant pro-inflammatory phenotype characterized by upregulated sCD40L, GM-CSF, IL-4, IL-8, MCP-1, and MIP-1β in the ESKD cohort, apart from the prototypical cytokines such as IFN-γ and TNF-α. CD40L is a costimulatory molecule predominantly expressed on activated CD4 + T cells. Activation of CD40-CD40L axis leads to activation of B cells and cytokine secretion, which in turn activate T cells as positive feedback (27). Increased sCD40L has been associated with chronic HIV infection (28), cancers (29), and pulmonary arterial hypertension (30), and could potentially serve as a biomarker of disease activities. Intriguingly, upregulated circulating sCD40L also enhanced expression and release of sCD40L in platelets (31), supporting the clinical observation of hypercoagulability in ESKD patients. IL-4 is produced by follicular helper T cells and has been shown to regulate B cell maturation in GC (32), which is consistent with our findings in upregulated T FH2 and increased plasmablasts. In addition, IL-4 (33) and MIP-1β (34) are known to polarize macrophage into M1 macrophage, a subtype of macrophage robustly producing cytokines and chemokines such as IL-1β and TNF-α. Although we did not include macrophage into our analysis, a prior review (35) has delineated the pro-inflammatory role of macrophages in kidney failure patients. Of note, the conventional intermittent hemodialysis membranes do not remove middle molecules (500 Da-15 kDa), and therefore the higher levels of pro-inflammatory cytokines were not due to poor clearance compared with others of no statistical significance. Thus, our results support a chronically activated inflammatory state in ESKD.
Regarding T cells, first we noted a similar finding of worsening lymphopenia in patients with CKD and ESKD. However, we did not observe a decreased or inverted ratio of CD4 + /CD8 + T cells as previously reported by others (11,19). This discrepancy might be explained by the unknown status of CMV infection of our patients, as Crépin et al. (18) reported that inverted ratio was only found in CMV-seropositive patients in their cohort. Lymphopenia in CKD/ESKD patients, especially naïve CD4 + and CD8 + T cells and T central memory, was previously attributed to decreased thymic output, increased apoptosis or shortening of telomere length (36). We also observed reduced naïve CD4 + and CD8 + T cells in CKD but not in ESKD cohort. Of note, we strictly defined naïve T cells as CD45RA + CD45RO − CD27 + CD28 + in our study rather than CD45RO − CCR7 + , which might represent a different population when compared to others (37). It is unclear whether the mobilization of naïve T cells in our ESKD group was mediated by intermittent removal of uremic toxins by hemodialysis or other factors.
Our data demonstrate an association between ESKD and reduced effector CD4 + and CD8 + T cells. Whether this reduction is due to recruitment of these cells to the peripheral tissues or reflects an absolute decline goes beyond the scope of the current study. However, after mitogen stimulation, T cells from ESKD patients demonstrated a pro-inflammatory phenotype with an increase in IFN-γ, IL-2 and TNF-α, whereas CKD patients only had a higher level of IL-2. Altogether, these findings were suggestive of persistent and active pro-inflammatory T cells especially in ESKD patients.
Intriguingly, we found that both CKD and ESKD patients had higher frequencies of exhausted and anergic CD4 + and CD8 + T cells. T cell receptor (TCR) signaling is essential for T cell differentiation into effector cells as well as appropriate clonal expansion (38) and defective TCR signaling have been described in elderly individuals (39) and chronic inflammatory diseases (40). In our study, phenotypically exhausted T cells were not capable of effectively producing cytokines. Therefore, it is tempting to speculate that, upon chronic stimulation (possibly driven by uremic toxins, amongst other stimuli), T cells become dysfunctional, leading to increased risk of infections and malignancies (1).
Despite the inconsistent results of Treg frequency in ESKD patients when compared to HC (17,21,41,42), impaired ex vivo Treg proliferation and suppressive capacity has been shown in patients with ESKD (21). Treg counteracts effector T cells to promote self-tolerance and reduce alloimmunogenicity. Forkhead box P3 (FOXP3), a pivotal transcription factor in Treg homeostasis, is degraded in chronic inflammation by pro-inflammatory cytokines such as TNF-α (43). Thus, downregulated Treg can result in excessive inflammatory responses and aberrant autoimmune diseases. In our cohort, we observed a slight increase in the total frequency of Treg in ESKD patients but a significant decrease of active Treg (CCR4 + CD45RA − CD25 + CD127 low− CD4 + CD8 − CD3 + ) in both CKD and ESKD patients. This supports the hypothesis that the chronic inflammatory milieu in ESKD impairs Treg function, instead of affecting their number. A similar finding of upregulation of total Treg was considered to be a reactive response in a cohort ANCA-associated vasculitis (AAV) patients (44).
Upregulation of T FH cells has been associated with chronic inflammation in cardiovascular (45,46) and autoimmune diseases (47). In our cohort, we observed T FH1 , a possibly protective subtype against viral infection, clearly diminished in ESKD. On the contrary, the presence of pro-inflammatory T FH2 and T FH17 was associated with systemic lupus erythematous (SLE) (48) and rheumatoid arthritis (49), respectively, and T FH2 was remarkably increased in our ESKD cohort. To date, there was only one small cohort of ESKD patients, where Assing et al. (50) reported a decreased frequency of peripheral T FH from apoptosis. The authors also suggested the reduced T FH could explain antibody deficiency in ESKD from a disturbed interaction between T FH and B cells in GC. However, the authors did not further classify the subtypes of T FH in their analysis. Indeed, B cells lymphopenia has been shown in patients with CKD or ESKD due to B cell apoptosis and poor B cell differentiation (51,52), but the mechanism remains poorly understood especially in the context of T FH and B cell interactions. In line with T FH findings, we also observed a higher percentage of circulating plasmablasts, suggesting an increase of B cell maturation within GC. Surprisingly, the upregulated plasmablasts did not result in a similar increase of plasma cells in ESKD patients. After entering circulation from secondary lymphoid tissue, plasmablasts further differentiate into plasma cells, the predominant antibody-secreting cells, and later transition into long-lived plasma cells. This process requires necessary survival factors and environment (bone marrow or spleen) to maintain circulating plasma cells in the periphery (53). For instance, deletion of B cell maturation antigen (BCMA), a B-cell activating factor (BAFF) receptor, was shown to affect the survival of long-lived plasma cells in a murine model (54). Downregulated BAFF receptor on transitional B cells has been implicated in deficient B cell differentiation in ESKD patients (51) but whether it also occurs in plasma cells requires further studies.
Our study has certain limitations. The numbers of our patients were relatively small and the cross-sectional nature of the study prevents testing the association between immunological abnormalities and cardiovascular outcomes, incidence of malignancies, or infectious complications. We did not have information of CMV infection status of our cohort, which could partly contribute to T cell exhaustion.
In conclusion, we showed that kidney failure is associated with a pro-inflammatory state that may be responsible for the exhausted phenotype of T cells and the skewed upregulation of T FH2 , especially in ESKD. These immunological changes may be, at least in part, responsible for their increased cardiovascular risk and the augmented susceptibility to infections and malignancies.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by IRB-20-03454, Mount Sinai Hospital. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
PC designed the study. SH, SB, CC, and MH performed all the experiments. JM, AA, BM, WZ, and LG provided patient samples. SH, JL, SM-WY, GLM, and PC helped with data interpretation and wrote the initial draft of the paper. All authors approved the paper.

FUNDING
This study has been supported by an investigator-initiated grant from the Renal Research Institute. We want to particularly acknowledge the patients and the Biobank Navarrabiomed (PT17/0015/0007) integrated in the Spanish National Biobanks Network for their collaboration.