Classical complement and inflammasome activation converge in CD14highCD16- monocytes in HIV associated TB-immune reconstitution inflammatory syndrome

Abstract

Inflammasome-derived cytokines, IL-1β and IL-18, and complement cascade have been independently implicated in the pathogenesis of tuberculosis (TB)-immune reconstitution inflammatory syndrome (TB-IRIS), a complication affecting HIV+ individuals starting antiretroviral therapy (ART). Although sublytic deposition of the membrane attack complex (MAC) has been shown to promote NLRP3 inflammasome activation, it is unknown whether these pathways may cooperatively contribute to TB-IRIS. To evaluate the activation of inflammasome, peripheral blood mononuclear cells (PBMCs) from HIV-TB co-infected patients prior to ART and at the IRIS or equivalent timepoint were incubated with a probe used to assess active caspase-1/4/5 followed by screening of ASC (apoptosis-associated speck-like protein containing a CARD domain) specks as a readout of inflammasome activation by imaging flow cytometry. We found higher numbers of monocytes showing spontaneous caspase-1/4/5+ASC-speck formation in TB-IRIS compared to TB non-IRIS patients. Moreover, numbers of caspase-1/4/5+ASC-speck+ monocytes positively correlated with IL-1β/IL-18 plasma levels. Besides increased systemic levels of C1q and C5a, TB-IRIS patients also showed elevated C1q and C3 deposition on monocyte cell surface, suggesting aberrant classical complement activation. A clustering tSNE analysis revealed TB-IRIS patients are enriched in a CD14highCD16- monocyte population that undergoes MAC deposition and caspase-1/4/5 activation compared to TB non-IRIS patients, suggesting complement-associated inflammasome activation during IRIS events. Accordingly, PBMCs from patients were more sensitive to ex-vivo complement-mediated IL-1β secretion than healthy control cells in a NLRP3-dependent manner. Therefore, our data suggest complement-associated inflammasome activation may fuel the dysregulated TB-IRIS systemic inflammatory cascade and targeting this pathway may represent a novel therapeutic approach for IRIS or related inflammatory syndromes.

Fig 1. Transcriptional profile of peripheral blood from TB-IRIS and TB non-IRIS patients after ART.

(A) Volcano plot of differentially expressed genes in TB-IRIS patients compared to TB-HIV co-infected patients who did not develop IRIS. The top differentially upregulated (log2 fold change 1.5, p = 0.05) genes related to complement and inflammasome pathways are highlighted. (B) Molecule activity prediction was performed using ingenuity pathway analysis (IPA) to predict downstream effects. Molecules in red were those found to be differentially abundant in the signatures of TB-IRIS patients when compared to TB non-IRIS participants. Molecules in light pink were predicted to be activated and arrows show predicted relationships that lead to activation. (C) Canonical pathways significantly differentially regulated are listed according to their p value, the regulation z-score algorithm to identify pathways that are upregulated (positive z-score).

Fig 2. Elevated numbers of monocytes containing FLICA⁺ASC-speck formation is found during TB-IRIS events.

(A) PBMCs from healthy donors (n = 22) and TB-IRIS (n = 14), TB-non IRIS (n = 9) and Non-TB, non-IRIS (n = 7) patients, before (Week 0) or after ART initiation (Week 2–8), were incubated with the fluorochrome inhibitor of caspase-1/4/5 (FLICA), stained for monocyte identification and intracellular ASC and acquired by using Imaging flow cytometry. Representative images showing co-localization of active caspase-1/4/5 with ASC Specks were selected from a TB-IRIS patient, after Bright Detail Similarity R3_MC_11-ASC_2-FLICA was applied inside the Speck⁺ gate. Images show respectively: BF (brightfield), FLICA and ASC fluorescences followed by a composite image containing BF and the fluorescence of ASC and FLICA merged. Note: The fluorescence intensity of the images may have been modified, without affecting results and statistics. (B and C) The number of monocytes showing spontaneous FLICA⁺ASC-Speck formation was quantified after application of Modulation_Morphology (M11,Ch11)_11-ASC feature, followed by Bright Detail Similarity R3_MC_11-ASC_2-FLICA, inside the “Total Monocytes” gate, by using IDEAS software. Data are presented as median with interquartile range. *P < 0.05, **P < 0.01 and ****P < 0.001, when Mann-Whitney test was applied. (D) Longitudinal analysis of pre-ART versus post-ART timepoints for the amount of FLICA⁺ASC-Speck⁺ monocytes/mL found in TB-IRIS (left, n = 8) and TB non-IRIS (right, n = 8) patients. (E) Absolute differences (Δ) between pre-ART versus post-ART timepoints were calculated for TB-IRIS and TB non-IRIS patient samples. Spearman correlations between plasma levels of IL-18 or IL-1β and caspase-1/4/5 activity (F and G, respectively) found on TB-IRIS and TB non-IRIS patient samples. CS = corticosteroids.

Fig 3. Quantification of canonical inflammasome activation within distinct monocyte subsets in different stages of pyroptosis.

FLICA⁺ASC-speck⁺ monocytes/ml were quantified in PBMCs from healthy volunteers and patients after modulation_morphology (M11,Ch11),11-ASC, followed by bright detail similarity R3_MC_11-ASC_2-FLICA features were applied in all data points, inside Live/Dead AQUA^low (A) or Live/Dead AQUA^high (B) gates. *P < 0.05, **P < 0.01 and ****P < 0.001, when Mann-Whitney test was applied. (C) Representative images of FLICA⁺ASC-speck⁺ monocytes inside the Live/Dead AQUA^high gate within CD14^highCD16^- subset (left panel) or CD14^lowCD16⁺ subset (right panel) were selected, showing respectively: BF1 (brightfield), CD16, CD14, Live/Dead (L/D) and a composite image containing the fluorescence of Live/Dead AQUA, FLICA and ASC.

Fig 4. TB-IRIS is accompanied by complement overactivation.

(A) Serum or (B)(C) plasma samples from TB-IRIS and TB non-IRIS patients, before (Week 0) or after ART initiation (Week 2–8), were collected and tested for classical complement activity (A), C1q (B) or C5a quantification (C). Membrane-bound C1q (mC1q) (D), C3 (E) and CD59 (F) expression levels were quantified on total monocytes from TB-IRIS, TB non-IRIS and Non-TB, non-IRIS patients, by flow cytometry. Expression levels (MFIs) were calculated as fold change over their experimental respective HCs. Data are presented as median with interquartile range. *P < 0.05 and **P < 0.01, when Mann-Whitney test was applied. Longitudinal analysis of pre-ART versus post-ART timepoints for mC1q (G) and C3 (H) expression levels for TB-IRIS (n = 8, left side) and TB non-IRIS (n = 7, right side) patients. *P < 0.05 was considered statistically significant when Wilcoxon signed-rank test was applied.

Fig 5. Association between complement MAC deposition and inflammasome overactivation on monocytes during IRIS event.

(A) Overlay of monocyte subsets clusters from HLADR⁺DUMP^-CD14⁺ and CD16⁺ population (“Total monocytes”) of TB-IRIS (n = 3) and TB non-IRIS (n = 3) patients on top of the t-SNE map based on CD14, CD16, C9 and FLICA expression. (A–right panel) Representative histograms of CD14, CD16, C9 and FLICA expression levels in each monocyte subset presented in the t-SNE plot. (B) Individual t-SNE maps for each group of samples (TB-IRIS and TB non-IRIS) based on expression levels of C9, FLICA, CD14 and CD16. Circles are drawn to highlight cell populations discussed in the text. (C) C9 and (D) FLICA expression levels (MFIs) were calculated as fold change over their respective experimental HCs for TB-IRIS (n = 10), TB non-IRIS (n = 9) and Non-TB, non-IRIS patients (n = 7), post-ART, within the distinct indicated monocytes subsets. Data are presented as median with interquartile range. *P < 0.05, **P < 0.01, when Mann-Whitney test was applied. Spearman correlations between FLICA fold change levels (E) or plasma levels of IL-18 (F) and C9 fold change levels found on monocytes from TB-IRIS and TB non-IRIS patient samples.

Fig 6. Effect of exogenous complement-mediated inflammasome activation on HCs and patient cells.

(A) Absolute numbers of healthy control human monocytes stimulated with designated concentrations of rabbit serum (RS) or heat-inactivated (HI)-RS (30 min at 56°C) for 2 hours, were enumerated by using counting beads by flow cytometry. HI-RS-stimulated cultures were considered 100% viable. (B) IL-1β secretion of 5% RS or 5% HI-RS-stimulated healthy control monocytes was determined at culture supernatants by ELISA. Cells were treated or not with the NLRP3 inhibitor, MCC950 (3 μM) or with KCl (10 mM) solution. Numbers represent the means ± SEM (n = 3). ****P < 0.001 when compared with control or untreated groups. Data are representative of three independent experiments. (C) IL-1β secretion by PBMCs from three different HCs, TB non-IRIS or TB-IRIS patients, at IRIS timepoint, was determined by ELISA after cells were incubated or not with MCC950 (3 μM) and exposed to 3% of RS or HI-RS for 2 hr. Data are presented as median with interquartile range. *P < 0.05 when compared with control or untreated groups after Mann-Whitney test was applied.