Increased Plasma CXCL10 May be a Marker of Increased Risk of Immune Restoration Disease Associated With Nonviral Pathogens

Abstract

To the Editors: Commencement of antiretroviral therapy (ART) in HIV patients with a recently treated or unrecognised opportunistic infection may result in immune restoration disease (IRD) in those patients who experience the restoration of a pathogen-specific immune response that causes immunopathology.1 The immunopathology may be different for different pathogens.2 Plasma biomarkers which may predict some forms of IRD have been identified,3–7 but knowledge remains incomplete. We have undertaken further studies using plasma samples from patients enrolled into the INITIO study. The INITIO study was a large international randomized trial of ART treatment strategies in ART-naive HIV patients who were allocated 2 nucleoside analogue reverse transcriptase inhibitors (didanosine and stavudine) with either a nonnucleoside reverse transcriptase inhibitor (efavirenz) or a protease inhibitor (nelfinavir) or both.8 Here, we examined levels of proinflammatory cytokines/chemokines and other markers of immune activation (IL-6, IL-18, soluble(s) CD8, sCD26, sCD30, CXCL10, and vascular endothelial growth factor) in plasma samples collected before ART from patients recruited to the INITIO study who experienced IRD and matched controls. We considered if these markers might reveal more about the prediction of IRD. Forty-three patients who had an infectious disease event during the first 24 weeks of ART were identified from the INITIO database. Twenty-four of these patients were classified as having probable IRD because the plasma HIV RNA level had declined by >1 log10, and the infectious disease event had occurred during the first 12 weeks of ART.9 As the number of patients experiencing IRD with individual pathogens was small, levels of plasma biomarkers were compared in patients classified as having “viral IRD” only (Kaposi sarcoma: n = 4, cytomegalovirus disease: n = 2, herpes simplex virus disease: n = 1, progressive multifocal leucoencephalopathy [PML]: n = 1, lymphoma: n = 1) or “nonviral IRD” only (Mycobacterium avium complex disease: n = 3, Toxoplasma gondii disease: n = 2, Pneumocystis carinii pneumonitis: n = 2, Cryptococcal meningitis: n = 1, Mycobacterium tuberculosis disease: n = 1, Mycobacterium avium complex disease and Candidiasis: n = 1) and matched controls. Each IRD patient was matched with 2 non-IRD controls by sex, pre-ART CD4+ T-cell count and country of origin where possible. For 6 “nonviral IRD” cases, samples were available from only 1 control. The median time of IRD onset was 6.5 weeks after the start of ART (range: 1–24 weeks). All patients provided written and informed consent. Plasma levels of IL-6, CXCL10, and vascular endothelial growth factor were measured using BD Cytometric Bead Array Flex Sets (BD Biosciences, San Jose, CA). Three hundred events were collected per analyte using a BD FACSArray machine and BD FACSArray System Software v1.0.3 (BD Immunocytometry Systems, San Jose, CA). Analysis was performed using FCAP Array Software v1.0.1 (BD Biosciences). All samples were diluted 1/5. The lowest limit of detection was 5 pg/mL. Plasma levels of IL-18 (MBL, International, Woburn, MA) and sCD30 (Bender Medsystems; Vienna, Austria) were measured by enzyme-linked immunosorbent assay. Levels of sCD26 were measured by an enzyme capture assay.10 Soluble CD8, a marker of CD8+ T-cell activation, was assayed in the plasma of patients who developed IRD associated with cytomegalovirus and JC virus (PML) infection and matched controls using an enzyme-linked immunosorbent assay (Bender Medsystems). Pre-ART levels of immune markers in IRD and non-IRD patients were compared using the Mann–Whitney U test. Statistical significance was defined as P < 0.05. Age, sex, pre-ART CD4+ T cell count, HIV RNA load, smoking status, and treatment arm were similar in IRD cases and controls (data not shown). Plasma levels of IL-6 and sCD8 were below the lower limit of detection of the assay (<5 pg/mL and <15.6 U/mL, respectively). Plasma levels of IL-18, sCD26, sCD30, and CXCL10 were similar in IRD cases and controls pre-ART when all cases of IRD were considered together (P = 0.18–0.76; data not shown). However, pre-ART levels of CXCL10 were significantly higher in “non-viral IRD” cases when compared with matched controls (P = 0.008) (Fig. 1). Pre-ART levels of CXCL10 did not differ between “viral IRD” cases and controls (P = 0.74) or between “nonviral IRD” cases and “viral IRD” cases (P = 0.21) (Fig. 1). Pre-ART plasma levels of IL-18, sCD26, and sCD30 did not differ between “nonviral IRD” cases and controls (P = 0.56–0.90; data not shown), “viral IRD” cases and controls or between “nonviral IRD” and “viral IRD” cases (P = 0.14–0.82; data not shown).FIGURE 1: Pre-ART Levels of CXCL10 in “nonviral IRD” cases (n = 10) (denoted by ) and matched controls (n = 14) (denoted by ), and ‘viral IRD’ cases (n = 9) (denoted by ) and matched controls (n = 18) (denoted by ). Levels of CXCL10 were compared between groups using a Mann–Whitney U test.We have previously shown that increased plasma levels of CXCL10 are associated with IRD in HIV patients with Hepatitis B virus or M. tuberculosis infection and might therefore be a marker of disease immunopathogenesis.4,11 Pre-ART levels of CXCL10 in plasma from unstimulated whole blood cultures were predictive of “paradoxical” TB-IRIS4 and pre-ART CXCL10 responses to region of difference 1 (RD1) antigens of M. tuberculosis were predictive of TB-IRIS and ART-associated tuberculosis (ART-TB; most cases of which are equivalent to “unmasking” TB-IRIS) (Oliver B, BSc, unpublished data, December 2011). Here, pre-ART plasma levels of CXCL10 were higher in patients who subsequently developed a “nonviral IRD” in contrast to matched controls and patients with “viral IRD”. Increased CXCL10 production may therefore be a marker of active immune responses against not just M. tuberculosis but also other nonviral pathogens. However, as CXCL10 is produced during both innate and adaptive immune responses,12 it is not possible to determine what type of immune response is being generated. Data from the FIRST Study13 and a study by Porter et al14 indicated that high pre-ART plasma levels of C-reactive protein and D-dimer identified patients with a high risk of developing IRD and/or an AIDS-defining event. Haddow et al7 also demonstrated higher serum C-reactive protein levels in patients who developed “unmasking TB-IRIS” but did not show a difference in CXCL10 levels from controls. Here, we measured specific markers of immune function and activation. We show that these markers are not predictive of a single type of IRD but pre-ART levels of CXCL10 were elevated in patients who developed a “non-viral IRD,” compared with matched controls. These findings provide support for our previous suggestions that the immunopathogenesis of IRD may be different for different types of pathogen2 and that markers of immune responses against some pathogens are present before ART is commenced.3,4 We suggest that plasma CXCL10 levels should be assessed as a marker of IRD risk in future studies. ACKNOWLEDGMENTS The authors would like to thank the personnel at the Genomic Unit, Centro Nacional de Microbiología, Instituto de Salud Carlos III, for technical assistance in sequencing; Ana Sánchez for processing patients' data; and José Antonio Taboada for his support of this study.