Cryptococcosis-Associated Immune Reconstitution Inflammatory Syndrome Is Associated With Dysregulation of IL-7/IL-7 Receptor Signaling Pathway in T Cells and Monocyte Activation

Abstract

Background: 

Systemic levels of interleukin (IL)-7 at antiretroviral therapy (ART) initiation have previously been shown to be predictive of HIV-linked paradoxical cryptococcosis-associated immune reconstitution inflammatory syndrome (C-IRIS). We therefore explored IL-7/IL-7 receptor (IL-7/IL-7R) signaling pathway dysfunction, with related alterations in immune function, as a mechanism underlying C-IRIS.

Method: 

HIV-infected patients with cryptococcal meningitis who experienced C-IRIS (n = 27) were compared with CD4⁺ T-cell count-matched counterparts without C-IRIS (n = 27), after antifungal therapy and pre-ART initiation. Flow cytometry was used to assess T-cell and monocyte phenotypes and functions.

Results: 

Proportions of IL-7R+ CD4⁺ or CD8⁺ T cells correlated positively with CD4⁺ T-cell counts and proportions of central memory and naive CD4⁺ and CD8⁺ T-cell pre-ART (all r > 0.50 and P < 0.05); however, the former negatively correlated with CD4⁺ T-cell counts fold-increase on ART in non–C-IRIS but not C-IRIS patients. Higher frequencies of activated monocytes (CD14⁺CD86⁺ or CD14+HLA-DR+; P ≤ 0.038) were also observed in C-IRIS compared with non–C-IRIS patients, and those who failed to clear cryptococci from cerebrospinal fluid before ART had higher levels of activated monocytes (CD14+HLA-DR+, P = 0.017) compared with those who cleared. In multivariate regression, CD14+HLA-DR+ monocytes were independently associated with C-IRIS [hazard ratio = 1.055 (1.013–1.098); P = 0.009].

Conclusion: 

In contrast to non–C-IRIS patients, C-IRIS patients displayed a lack of association between proportions of IL-7R+ T cells and several markers of T-cell homeostasis. They also exhibited higher monocyte activation linked to cerebrospinal fluid cryptococcal culture positivity before ART. These data suggest a role for IL-7/IL-7R signaling pathway dysregulation in the pathogenesis of C-IRIS, possibly linked to monocyte activation and residual pathogen burden before ART.

INTRODUCTION

Cryptococcosis-associated immune reconstitution inflammatory syndrome (C-IRIS) is a common clinical complication in some patients with cryptococcal meningitis (CM), usually within 6 months of initiating combination antiretroviral therapy (cART).^1–4 C-IRIS presents clinically as a neurological deterioration during immune reconstitution after cART initiation. Mortality rates vary widely but may exceed 50%.^1,5–8 C-IRIS is believed to be caused by a paucity of “protective” immune responses, resulting in cryptococcal antigen accumulation before cART; followed by a storm of proinflammatory responses as the immune system recovers.^1,6,9,10 Severely immunocompromised patients who initiate cART with low CD4⁺ T-cell counts are at high risk of C-IRIS.^1,2 Paradoxical C-IRIS is the commonest form of C-IRIS and usually presents as a neurological deterioration in individuals with known CM before initiating cART.^1,4,6,10,11 Understanding the immunopathogenesis of C-IRIS may facilitate targeted interventions.

We recently demonstrated that plasma but not cerebrospinal fluid (CSF) levels of interleukin (IL)-5 and IL-7 preinitiation of cART were predictive of C-IRIS.¹² High IL-5 levels may reflect a generalized type 2 T-helper cell (Th2) environment that impairs the clearance of cryptococci, which is primarily dependent on Th1 responses in the absence of HIV infection.^6,9,10 Conversely, IL-7 is a hematopoietic cytokine that is involved in T-cell survival, proliferation, and differentiation. IL-7 may also play a role in T-cell activation and responses, as immune cell subsets repopulate on cART initiation and encounter residual cryptococcal antigens.^12–15 Levels of circulating IL-7 are believed to be primarily regulated by the rate at which the cytokine is consumed by T cells,^13,16 which in turn may depend on the expression and function of IL-7R on T cells. Defects in the IL-7R signaling pathway are associated with lymphopenia, with other cell lineages largely unaffected.^17,18 We therefore hypothesized that high plasma IL-7 levels in HIV patients with CM who developed C-IRIS reflect an impairment of the IL-7/IL-7 receptor (IL-7R/CD127) signaling pathway in T cells in general, or in Cryptococcus-specific T cells, after cART initiation. Defects in the IL-7/IL-7R signaling pathway may lead to aberrant T-cell responses to cryptococcal antigens as an underlying factor in the immunopathogenesis of C-IRIS. Alternatively, high IL-7 in C-IRIS patients may be just an indicator of generalized immune activation. It has also been shown that polymorphisms in the IL-7Rα gene are associated with faster CD4⁺ T-cell recovery after initiation of cART²⁸; and gene variations in IL-7Rα have been shown to affect IL-7R expression on CD4⁺ T cells in HIV-infected individuals.²⁹ Moreover, reduced IL-7R expression on T cells and increased plasma sCD127 have been demonstrated in late-presenting HIV-infected individuals.³⁰ In addition, evidence from a large body of literature indicates that activation of innate immune responses could play a major role in the pathogenesis of C-IRIS.^6,9,12,19

We hypothesized that the IL-7/IL-7R pathway may be differentially regulated in C-IRIS versus non–C-IRIS patients. We therefore sought to determine whether IL-7R expression on T cells negatively correlates with IL-7 levels, and whether IL-7R expression was independently associated with C-IRIS or T-cell numbers before cART and their recovery after cART initiation in either C-IRIS or non–C-IRIS patients. We then assessed whether T-cell and monocyte activation profiles before cART could distinguish between subsequent non–C-IRIS and C-IRIS conditions. Finally, previous work has demonstrated that cryptococcal mannoprotein (CMP)-specific CD4⁺ T-cell responses before cART may differ between non–C-IRIS and C-IRIS patients.^19,20 There is also evidence that cryptococcal strain variation can impact immune responses and therefore potentially influence C-IRIS risk.²¹ However, the magnitude of immune response to autologous cryptococcal strains and how this may impact on clinical outcomes after cART initiation is largely unknown. We therefore investigated whether cryptococcal-specific T-cell responses differed between C-IRIS and non–C-IRIS patients using autologous cryptococcal antigens (ACAs) isolated from cryptococci infecting each study participant versus mannoproteins derived from cap67 strain (CMP) as the recall antigens.

MATERIALS AND METHODS

Study Participants

This retrospective CD4⁺ T-cell count-matched case–control study was performed in participants enrolled into a larger study of 130 HIV-infected, cART-naive patients (mean age = 18 years, between 33 and 40.5 years) who experienced their first episode of CM in Durban, South Africa. Participants were consecutively enrolled into a longitudinal cohort study between August 2009 and March 2011.¹ Fifty-four patients with CM were included in this substudy. CM was diagnosed by CSF India ink and/or by cryptococcal antigen (CrAg) positivity (Cryptococcal Latex Agglutination System; Meridian Bioscience, Inc., Cincinnati, OH). Patients who experienced probable or possible C-IRIS while on cART (n = 27) were compared with CD4⁺ T-cell count-matched counterparts without C-IRIS in a 1:1 ratio. All study subjects were diagnosed with paradoxical C-IRIS, with all other forms of IRIS excluded from the study. All parameters of interest including IL-7 and cellular markers by flow cytometry were measured after antifungal treatment but before ART commencement, and at time of C-IRIS event in the C-IRIS group. Demographic and clinical characteristics of participants are provided in Table S2, Supplemental Digital Content, https://links.lww.com/QAI/B269, and detailed clinical findings have been presented elsewhere.¹

Cytokine Measurements

Plasma samples were collected after antifungal treatment and pre-cART initiation, and at the C-IRIS event and frozen at −80°C until use. IL-7 levels were quantified in a 17-plex-high-sensitivity Luminex kit on a Bio-Plex 200 system (Bio-Rad, Hercules, CA) as described in detail elsewhere.¹²

Preparation of Cryptococcal Antigens

Two sets of cryptococcal antigen preparations were used in the studies. CMP derived from a capsular strain cap67 (ATCC #52817) were purified from culture supernatants using Con A sepharose affinity chromatography as described.²² CMP has been shown to stimulate lymphoproliferative responses from cryptococcosis patients, including those with AIDS and immune reconstitution.¹⁹ CMP was stored in lyophilized aliquots at −80°C until use. ACAs were prepared by growing each patient-specific Cryptococcus strain and then performing an alkaline extraction, as described.²³ ACAs include a broad array of antigens, including mannoproteins. The pellet was suspended in 300 μL of 20 mM Tris-buffered saline containing a protease inhibitor cocktail of serine, cysteine, and metalloproteases inhibitors (Roche Diagnostic, Boston, MA). Protein concentration of the preparations was determined using the bicinchoninic acid. Cells were stimulated using a final concentration of 10 μg protein/mL.

Flow Cytometry

Cryopreserved peripheral blood mononuclear cells were thawed in RPMI 1640 (Sigma-Aldrich, Johannesburg, South Africa) supplemented with 10% heat-inactivated fetal calf serum, 100 U/mL penicillin, 100 g/mL streptomycin sulfate, and 1.7 mM sodium glutamate. After 2 hours of resting at 37°C in a 5% CO₂ incubator, half a million viable cells were aliquoted in a volume of 200 μL per well in a 96-well plate. Cells were washed in phosphate-buffered saline and then incubated with fixable near infrared (NIR) staining dye on APC-Cy7 (BioLegend, Inc., San Diego, CA) for dead cell exclusion for 30 minutes. Cells were then stained with the following antibodies, all from BD Biosciences, San Jose, CA, unless otherwise indicated: anti-CXCR3 (clone FUN-1) -BV421, anti-CD27 (clone M-T271) (BioLegend) -BV510, anti-CD45RA (clone 5H9) -quantum dot (Qdot)605 (Invitrogen, Carlsbad, CA), anti-CD4 (clone SK3) -BV711, anti-CD127 (clone M-A251) -BV786, anti-CD8 (clone MAb11) -Alexa F488, anti-PD-1 (clone RPA-T8) -PerCP-Cy5.5, anti-CD25 (clone SK7) -PE, anti-CD3⁻ (clone 5344.111) -PE-CF594, anti-CCR6 (clone G46-6) PE-Cy7, and anti-CCR7 (clone 150,503) -Alexa F 700. Cells were then washed and fixed (Perm/fix Medium A; Invitrogen).

Separately, one million viable cells/well were stimulated with staphylococcus enterotoxin B (SEB) and lipopolysaccharide (both from Sigma-Aldrich) as positive control at a concentration of 1 μg/mL each for 4.5 hours (to avoid downregulation of the CD14 molecule by lipopolysaccharide), CMP/ACA at a concentration of 10 μg/mL each for 18 hours in 5% CO₂ at 37°C. Unstimulated negative control and fluorescence minus one control wells were also added. Costimulatory antibodies, CD28 and CD49d (1 μg/mL each; BD Biosciences), were added to each well. Brefeldin A (BioLegend) was also added to each well after 1 hour of incubation. Cells were surface stained with: anti-CD86 (clone FUN-1) -Brilliant violet (BV)421, anti-CD38 (clone HIT2) -BV510, anti-CD14 (clone M5E2) -BV605, anti-CD134 (clone ACT35) -BV650, anti-CD4 (clone SK3) -BV711, anti-CD8 (clone M-A251) -BV786, anti-PD-1-PerCP-Cy5.5 (clone RPA-T8), anti-CD25 (clone SK7) -phycoerythrin (PE), anti-CD16 (clone 3G8) -PE-Cy5, anti-HLA-DR (clone G46-6) PE-Cy7, and anti-CD3 (clone SK7) -Alexa F 700. Subsequently, peripheral blood mononuclear cells were washed, fixed (Perm/fix medium A; Invitrogen), permeabilized (Perm/fix Medium B; Invitrogen), and intracellularly stained with anti-TNF-α (clone MAb11) -Alexa F488, anti-IL-2 (clone 5344.111) -PE-CF594, and anti-IFN-γ (clone 4S.B3) -Alexa F647. Cells were acquired on a BD LSRFortessa.

Data Analysis

Data were analyzed using the FlowJo version 10.01 (TreeStar, Inc., OR). Gating schemes are illustrated in the gating strategy figures. All parameters measured were based on fluorescence minus one controls. For intracellular cytokine staining, the background was considered as the frequency of cells producing cytokines in the absence of antigenic stimulation (unstimulated negative control) and was subtracted for each sample.²⁴

Statistical Analysis

The Mann–Whitney test was used to compare proportions of cells expressing a particular marker and intracellular cytokine responses before cART in C-IRIS individuals versus controls for both T cells and monocytes. The Wilcoxon signed-ranks test was used to compare baseline (pre-cART initiation) percentage of cells expressing a marker and intracellular cytokine responses versus the same parameters at the C-IRIS event (post-cART). To determine the predictors of C-IRIS in multivariable analysis, Cox proportional hazards regression analysis was performed using all parameters with a P value ≤0.1 in the univariate analysis. All statistical analyses were performed using GraphPad Prism v7.1 (La Jolla, CA) and Stata v13.0 (StataCorp, College Station, TX). Statistical significance was defined at P < 0.05.

Ethics Statement

Written informed consent was obtained from participants or next-of-kin (if the participant was not competent). Review boards at the University of KwaZulu-Natal, Monash University, and the University of Western Australia granted ethics approval.

RESULTS

Pre-cART CD4⁺ T-Cell Counts Correlated With Proportions of IL-7R+ CD4⁺ T Cells in Non–C-IRIS Patients Only but Not With Plasma IL-7 Levels in Either Group

Demographic and clinical characteristics of the cohort are presented in Table S2, Supplemental Digital Content, https://links.lww.com/QAI/B269, and detailed clinical findings have been reported elsewhere.^1,12 T-cell subsets were defined by CD45RA and CD27 expression such that naive T cells (Nv) were defined as CD27⁺CD45RA+; central memory (Cm) as CD27⁺CD45RA-; effector memory (Em) as CD27⁻CD45RA-; and terminal effector (Ef) as CD27⁻CD45RA+, as described.^25,26 The proportions of these cell subsets expressing IL-7R were measured ex vivo (see Fig. S1, Supplemental Digital Content, https://links.lww.com/QAI/B269). There was no correlation between proportions of T cells expressing IL-7R and plasma IL-7 levels before cART in both non–C-IRIS and C-IRIS patients (see Fig. S2, Supplemental Digital Content, https://links.lww.com/QAI/B269) as would have been expected from previous studies.^27,28 We next explored whether there was an association between percentage of IL-7R+ CD4⁺ T cells or plasma IL-7 levels pre-cART and pre-cART CD4⁺ T-cell count or fold-increase in CD4⁺ T-cell count after cART initiation according to C-IRIS outcome as the latter two have been reported to be predictors of C-IRIS.¹ Interestingly, in non–C-IRIS but not C-IRIS patients, pre-cART CD4⁺ T-cell counts positively correlated with the proportion of IL-7R+ total CD4⁺ T cells (r = 0.5423, P = 0.0035) (Fig. 1A) and central memory CD4⁺ T cells (r = 0.4891, P = 0.0096) (Fig. 1B). By contrast, fold-increase in CD4⁺ T-cell counts from pre-cART to 24 weeks of ART negatively correlated with pre-cART IL-7R+ total CD4⁺ T cells (r = −0.5747, P = 0.0041) (Fig. 1A) and central memory CD4⁺ T cells (r = −0.4818, P = 0.0199) (Fig. 1B). Before cART, CD4⁺ T-cell counts did not correlate with plasma IL-7 levels in either the non–C-IRIS or C-IRIS group (Fig. 1C). Collectively, these data suggest differential regulation of the IL-7/IL-7R signaling pathway in C-IRIS versus non–C-IRIS individuals.

FIGURE 1.: Correlations between pre-cART (baseline) proportions of IL-7R+ CD4⁺ T cells or plasma IL-7 levels and pre-cART CD4⁺ T-cell counts or fold-increase in CD4⁺ T-cell counts after ART in non–C-IRIS and C-IRIS patients. Proportions of (A) IL-7R+ total CD4⁺ T cells (CD4⁺CD127+) and (B) central memory CD4⁺ T cells (CD4⁺CD27⁺CD127+) positively correlated with pre-cART CD4⁺ T-cell counts but negatively correlated with fold-increase in CD4⁺ T-cell counts after ART in non–C-IRIS patients (blue), but not C-IRIS patients (red). C, There were no correlations between pre-cART plasma IL-7 levels and pre-cART CD4⁺ T-cell counts or fold-increase in CD4⁺ T-cell counts after ART in either study group. An analysis using IL-7R expression median fluorescence intensity (MFI) showed a similar trend.

Proportions of IL-7R+ Total and Central Memory CD8⁺ T Cells Were Lower in C-IRIS Patients Compared With Non–C-IRIS Controls Before cART, but Increased at the Time of C-IRIS

We further compared pre-cART proportions of IL-7R+ T cells in C-IRIS versus non–C-IRIS patients to determine whether proportions of IL-7R+ cells could distinguish patients who subsequently experienced C-IRIS from those who did not (Fig. 2). There were no significant differences in percentage of either total or central memory IL-7R+ CD4⁺ T cells (P = 0.199 and P = 0.394, respectively) or any other CD4⁺ T-cell subsets (data not shown). However, a trend was noted of lower percentages of IL-7R+ total CD8⁺ T cells and central memory CD8⁺ T cells in C-IRIS compared with non–C-IRIS individuals (P = 0.065 and P = 0.053, respectively) (Fig. 2A). In C-IRIS patients, IL-7R+ central memory CD4⁺ and CD8⁺ T cells increased between baseline and time of C-IRIS (P = 0.034 and 0.003, respectively; Fig. 2B).

FIGURE 2.: Comparison of proportions of IL-7R+ CD4⁺ and CD8⁺ T cells before cART in non–C-IRIS and C-IRIS patients and change in proportions of each cell type from pre-cART to C-IRIS in C-IRIS patients. A, Before cART, proportions of IL-7R+ total (CD4⁺CD127+) and central memory (CD4⁺CD27⁺CD127+) CD4⁺ T cells did not differ between non–C-IRIS and C-IRIS patients, but there was a trend toward higher IL-7R+ total (CD8⁺CD127+) and central memory (CD8⁺CD27⁺CD127+) CD8⁺ T cells in non–C-IRIS patients (P = 0.065 and 0.053, respectively). B, Proportions of total and central memory CD8⁺ T cells, but not CD4⁺ T cells, were higher at C-IRIS event compared with baseline. Analysis using IL-7R expression median fluorescence intensity (MFI) showed a similar trend.

Proportions of IL-7+ CD4⁺ T Cells Pre-cART Correlated With Pre-cART Proportions of Central Memory and Naive CD4⁺ and CD8⁺ T Cells in Non–C-IRIS Patients and Fold-Increase in Central Memory and Naive CD4⁺ T Cells at C-IRIS Event

Because IL-7 and IL-7R are involved in the production and homeostatic maintenance of T cells, we correlated proportions of T cells expressing IL-7R or plasma IL-7 levels pre-cART with pre-cART percentages of central memory (CD27⁺CD45RA−) and naive (CD27⁺CD45RA+) T cells in both non–C-IRIS and C-IRIS patients and with fold-increase in these cell populations from pre-cART to time of C-IRIS in C-IRIS patients. Similar to the observation for CD4⁺ T-cell counts (Fig. 1), in non–C-IRIS but not C-IRIS patients, the pre-cART proportions of central memory (CD4⁺CD27⁺CD45RA−) and naive (CD4⁺CD27⁺CD45RA+) CD4⁺ T cells showed highly significant positive correlations with proportions of IL-7R+ CD4⁺ T cells (r = 0.638, P = 0.0003 and r = 0.5433, P = 0.0034, respectively) (Fig. 3A). The percentage of IL-7R+ CD8⁺ T cells also significantly correlated with the percentages of central memory (CD8⁺CD27⁺CD45RA−) and naive (CD8⁺CD27⁺CD45RA+) CD8⁺ T cells (r = 0.6569, P = 0.0002 and r = 0.6801, P < 0.0001, respectively) but only in non–C-IRIS patients (Fig. 3A). By contrast, pre-cART proportions of none of these T-cell subsets correlated with plasma IL-7 levels (Fig. 3B). These data further suggest functional differences in the IL-7/IL-7R receptor signaling pathway of T cells in C-IRIS versus non–C-IRIS patients and suggest that the high IL-7 plasma levels observed in C-IRIS patients may be a marker of decreased IL-7 utilization by a dysfunctional IL-7/IL-7R pathway.

FIGURE 3.: Correlations of IL-7R+ CD4⁺ and CD8⁺ T cells or plasma IL-7 levels pre-cART with naive and central memory CD4⁺ and CD8⁺ T cells pre-cART in non–C-IRIS (blue) and C-IRIS (red) patients. A, Before cART, proportions of naive and central memory CD4⁺ and CD8⁺ T cells correlated with IL-7R+ CD4⁺ and CD8⁺ T cells in non–C-IRIS, but not C-IRIS, patients. B, Plasma IL-7 levels did not correlate with proportions of naive or central memory CD4⁺ or CD8⁺ T cells in either patient group. Analysis using IL-7R expression median fluorescence intensity (MFI) showed a similar trend.

We further assessed the association between proportion of IL-7R+ T cells pre-cART and fold-increase in central memory and naive T-cell subsets from pre-cART to the time of C-IRIS event in C-IRIS patients. The fold-increase in central memory and naive CD4⁺ T cells positively correlated with the proportion of IL-7R+ CD4⁺ T cells pre-cART (r = 0.4771, P = 0.0184 and r = 0.6504, P = 0.008, respectively) (Fig. 4A). However, similar correlation was not observed for the CD8⁺ T-cell compartment. By contrast, no correlations between fold-increase of these T-cell subsets and plasma IL-7 levels were observed (Fig. 4B), suggesting that in C-IRIS patients, increases in naive and central memory CD4⁺ T-cell subsets are dependent on the proportion of CD4⁺ T cells expressing IL-7R (CD4⁺CD127+) but not the amount of plasma IL-7.

FIGURE 4.: Correlations of IL-7R+ CD4⁺ and CD8⁺ T cells or plasma IL-7 levels pre-cART with fold-increase in proportions of central memory and naive CD4⁺ and CD8⁺ T cells at C-IRIS event in C-IRIS patients. A, Fold-increase in proportions of central memory and naive CD4⁺, but not CD8⁺, T cells at C-IRIS event correlated with proportions of IL-7R+ CD4⁺ T cells pre-cART. B, Plasma IL-7 levels pre-cART did not correlate with fold-increase in central memory or naive CD4⁺ or CD8⁺ T cells at C-IRIS event. Analyses using IL-7R expression median fluorescence intensity (MFI) showed a similar trend.

Development of C-IRIS Was Associated With Higher Monocyte Activation and CSF Cryptococcal Culture Positivity Before cART

We further sought to investigate the immunopathogenesis of C-IRIS by assessing T-cell and monocyte (CD14⁺ cells) activation. CD4⁺ and CD8⁺ T-cell activation was defined by expression of both CD38 and HLA-DR, whereas exhaustion was assessed by PD-1 expression. We observed comparable percentages of activated T cells between the C-IRIS and non–C-IRIS patients at pre-cART (see Fig. S3A, Supplemental Digital Content, https://links.lww.com/QAI/B269). These percentages did not change significantly at the time of C-IRIS event for those who developed C-IRIS (see Fig. S3B, Supplemental Digital Content, https://links.lww.com/QAI/B269). Interestingly, when monocyte activation was assessed by CD86 or HLA-DR expression on CD14⁺ cells as illustrated by the gating strategy (Fig. 5A) and representative flow plot (Fig. 5B), C-IRIS individuals had significantly higher activation than non–C-IRIS controls pre-cART (P = 0.036 and P = 0.038, respectively) (Fig. 5C). Furthermore, at the time of C-IRIS event, CD14⁺CD86⁺ cells were more frequent than pre-cART in C-IRIS patients (P = 0.017) (Fig. 5C).

FIGURE 5.: Activated monocytes were higher in C-IRIS than non–C-IRIS patients and in patients who had positive CSF cultures for cryptococci before cART. A, Gating strategy. B, Representative flow cytometry gating strategy for defining activated (CD86⁺ or HLA-DR+) monocytes (CD14⁺). CM108 did not experience C-IRIS, whereas CM034 did. C, Proportions of activated monocytes were higher in C-IRIS patients compared with non–C-IRIS patients before cART, and CD86⁺ monocytes increased further at the time of C-IRIS event. D, Proportions of activated monocytes (CD14+HLA-DR+) were higher in patients who had positive CSF cultures for cryptococci before cART. Comparison between non–C-IRIS versus C-IRIS was undertaken using the Mann–Whitney test.

Because studies have suggested that activation of innate immune responses in IRIS conditions may be largely caused by opportunistic infection antigen load,^11,29 we further investigated whether monocyte activation correlated with CSF cryptococcal antigen burden. We observed that monocyte activation (CD14⁺CD86⁺) was associated with CSF cryptococcal positivity pre-cART (P = 0.017), and CD14+HLA-DR+ monocytes were also higher in those individuals who failed to clear cryptococci from CSF pre-cART, although in the latter case the difference was not statistically significant (P = 0.108) (Fig. 5D).

We also explored whether the presumptive dysregulation of the IL-7/IL-7R signaling pathway in T cells of patients who developed C-IRIS might affect Cryptococcus-specific T-cell responses by measuring responses to ACA, derived from the Cryptococcus cultured from each patient's CSF, and to CMP from the cap67 strain of Cryptococcus (see Fig. S4, Supplemental Digital Content, https://links.lww.com/QAI/B269). Before cART, SEB-activated and ACA- and CMP-specific IFN-γ, IL-2, and TNF-α T-cell responses of C-IRIS individuals tended to be higher, albeit not statistically significant, than in non–C-IRIS patients (see Fig. S4A, Supplemental Digital Content, https://links.lww.com/QAI/B269 and see Fig. S4C, Supplemental Digital Content, https://links.lww.com/QAI/B269). At the time of C-IRIS, cases showed a decrease in the frequency of SEB-activated (P = 0.056), ACA-specific (P = 0.046), and CMP-specific (P = 0.005) CD4⁺ T-cell IFN-γ responses compared with pre-cART (see Fig. S4B, Supplemental Digital Content, https://links.lww.com/QAI/B269). Similarly, significant decreases of ACA-specific CD8⁺ T-cell IL-2 responses (P = 0.008) and CMP-specific IFN-γ responses (P = 0.022) were observed at C-IRIS event (see Fig. S4D, Supplemental Digital Content, https://links.lww.com/QAI/B269). We therefore did not find that ACA- or CMP-specific T-cell responses were lower before cART or rose after cART in C-IRIS patients.

Both Plasma IL-7 Levels and Proportion of Activated Monocytes Were Predictive of C-IRIS in a Multivariable Regression Analysis

Finally, we assessed the risk of C-IRIS occurrence in a ten-parameter multivariable Cox proportional hazards regression analysis where all pre-cART correlates from univariate analysis with a P value <0.1 were considered relevant (see Table S1, Supplemental Digital Content, https://links.lww.com/QAI/B269). High plasma IL-7 levels remained predictive of C-IRIS {hazard ratio (HR) = 5.7 [95% confidence interval (CI): 1.70 to 19.3]; P = 0.005}. In addition, the proportion of activated monocytes (CD14+HLA-DR+) [HR = 1.06 (95% CI: 1.01 to 1.10); P = 0.009] was also predictive of C-IRIS. There was also a nonsignificant trend for low proportions of IL-7R+ CD8⁺ T cells as a predictor of C-IRIS [HR = 0.836 (95% CI: 0.694 to 1.007), P = 0.059].

DISCUSSION

In this study of C-IRIS cases and non–C-IRIS controls matched for CD4⁺ T-cell counts, several findings provided evidence that dysregulation of the IL-7/IL-7R signaling pathway in T cells might contribute to the immunopathogenesis of C-IRIS and explain our previous observation that a high plasma IL-7 level is predictive of C-IRIS.^30,31 In non–C-IRIS patients, there was a positive correlation between proportion of total and central memory CD4⁺ T cells expressing IL-7R and CD4⁺ T-cell count before cART; however, no such correlation was observed in C-IRIS patients. Similarly, there was also a positive correlation between proportion of IL-7R+ T cells and the proportion of naive and central memory CD4⁺ and CD8⁺ T cells but only in non–C-IRIS patients. By contrast, the proportion of IL-7R+ total and central memory CD4⁺ T cells before cART negatively correlated with fold-increase in CD4⁺ T-cell counts in non–C-IRIS but not C-IRIS patients. These findings suggest that IL-7R may play a role in the homeostasis of T cells before cART initiation in non–C-IRIS patients but not C-IRIS patients. However, in C-IRIS patients, we demonstrated a positive correlation between the proportion of IL-7R+ CD4⁺ T cells pre-cART and the fold-increase in central memory and naive CD4⁺ T cells at the time of C-IRIS, which was much earlier than the time at which fold-increase in CD4⁺ T-cell counts was examined. Interestingly, plasma IL-7 levels did not correlate with pre-cART CD4⁺ T cell counts or proportions of IL-7+ central memory and naive CD4⁺ T cells, nor with fold-increase in any of these cell populations in any patient group at any time point. On assessing immune activation, we observed that pre-cART monocyte activation, linked to CSF cryptococcal culture positivity, was associated with C-IRIS. By contrast, we did not observe a relationship with T-cell activation, and baseline Cryptococcus-specific T-cell responses were similar in non–C-IRIS and C-IRIS individuals. In a multivariable regression analysis, IL-7 levels and the proportion of activated monocytes remained predictive of C-IRIS.

While acknowledging that we have not assessed the IL7/IL-7R signaling pathway directly, to the best of our knowledge, our study is the first to implicate dysfunction of the IL7/IL-7R signaling pathway in the immunopathogenesis of C-IRIS, although plasma IL-7 levels have previously been implicated.^6,12

Our finding that baseline CD4⁺ T-cell counts strongly correlated with proportions of T cells expressing IL-7R (but not baseline plasma IL-7 levels) in non–C-IRIS individuals but not in C-IRIS individuals strongly suggests a functional difference in the IL-7/IL-7R signaling pathway of T cells in non–C-IRIS and C-IRIS patients. It was previously demonstrated that cART initiation rescued IL-7R expression on T cells in HIV infection,³² and deficiencies in the IL-7R expression have been shown to result in severe lymphopenia.^17,18 The significant correlations of CD4⁺ T-cell counts and proportions of central memory and naive T cells with IL-7R expression in non–C-IRIS patients may indicate that the cellular machinery required for T-cell homeostasis before cART initiation remains intact, whereas this may not be the case in C-IRIS patients. Overall, our data are consistent with the current theory that a low CD4⁺ T-cell count is an underlying factor in C-IRIS immunopathogenesis. The data further suggest defects in IL-7R–dependent T-cell proliferation and maturation in C-IRIS but not non–C-IRIS HIV-1–infected patients with CM. Previously observed differences in plasma IL-7 levels between C-IRIS and non–C-IRIS patients may therefore reflect a dysfunction in IL-7R expression or in the signaling pathway.

Consistent with a recent study of TB-IRIS, there was no significant difference in T-cell activation or exhaustion between non–C-IRIS and C-IRIS patients before cART.³³ By contrast, others have suggested that CD8⁺ T-cell activation might be relevant in some instances of IRIS.^34,35 Interestingly, we observed that monocyte activation was associated with C-IRIS, in line with previous data implicating aberrant innate immune responses by neutrophils and monocytes in C-IRIS.^36,37 High monocyte activation was associated with CSF cryptococcal culture positivity before cART, suggesting that persisting antigens after antifungal therapy may induce a proinflammatory antigen presenting cell bias, consistent with high-plasma IL-6 levels observed in C-IRIS after cART commencement.^9,12,33 Our data suggest that the monocyte activation observed in C-IRIS may be due to high fungal load, which has already been shown to be a major risk factor for C-IRIS.^1,6,38 Considering that there may be interactions between components of the immune system, we performed a multivariable regression analysis of relevant components from the univariate analysis. Plasma IL-7 levels and monocyte activation remained strongly associated with C-IRIS.

In conclusion, these data provide evidence of a functional defect in the IL-7/IL-7R signaling pathway in HIV/CM coinfection that may contribute to the immunopathogenesis of C-IRIS. Specifically, the proportion of CD4⁺ or CD8⁺ T cells expressing IL-7R but not plasma IL-7 levels correlated with CD4⁺ T-cell counts and proportions of central memory and naive T cells in non–C-IRIS but not C-IRIS individuals, suggesting IL-7/IL-7R signaling pathway dysregulation in the C-IRIS cases. These data also demonstrate that monocyte activation, linked to CSF cryptococcal culture positivity before cART, is a risk factor for developing C-IRIS. The cause of the presumed IL-7/IL-7R signaling pathway dysregulation in C-IRIS individuals deserves further investigation.