CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

774 HIV INFECTION IMPAIRS MYCOBACTERIUM TUBERCULOSIS-SPECIFIC CD4 T-CELL RESPONSES Patrizia Amelio 1 , Jerry Hella 2 , Klaus Reither 3 , Lujeko Kamwela 2 , Omar Lweno 2 , Anneth Tumbo 2 , Linda Geoffrey 2 , Song Ding 4 , Giuseppe Pantaleo 1 , Claudia Daubenberger 3 , Matthieu Perreau 1 1 Lausanne University Hospital, Lausanne, Switzerland, 2 Ifakara Health Institute, Dar es Salaam, Tanzania, United Republic of, 3 Swiss Tropical and Public Health Institute, Basel, Switzerland, 4 EuroVacc Foundation, Amsterdam, Netherlands Background: HIV infection is the major risk factor predisposing for 1) Mycobacterium tuberculosis (Mtb) infection and 2) progression from latent tuberculosis infection (LTBI) to tuberculosis disease (TB). Since, pulmonary TB (PTB) usually occurs in HIV-infected individuals with higher CD4 T-cell counts as compared to other opportunistic pulmonary infections such as Pneumocystis pneumonia, we hypothesized that progression from LTBI to PTB might not only be due to CD4 T-cell depletion but also to Mtb-specific CD4 T-cell impairment. Methods: To test this hypothesis, Mtb-specific CD4 T-cell cytokine profiles and transcription factor expression profiles were investigated in untreated Tanzanian individuals suffering from LTBI or PTB and compared to those of untreated Mtb/HIV co-infected individuals suffering from LTBI or PTB. Results: We show that Mtb-specific CD4 T-cell cytokine profiles of HIV negative individuals with LTBI or PTB are primarily composed of polyfunctional Th1 (IFN-γ/TNF-α/IL-2) and Th2 cells (IL-4/IL-5/IL-13). In contrast, the cytokine profiles of Mtb-specific CD4 T cells of Mtb/HIV co-infected individuals with LTBI or PTB were dominated by single TNF-α, single IFN-γ and dual IFN-γ/TNF-α, and reduction of polyfunctional Th1 and Th2 cells (P<0.05). The skewing of Mtb-specific CD4 T-cell cytokine profiles in Mtb/HIV co-infected individuals was associated with a significant increase of T-bet expression (P<0.05) and a significant reduction of Gata-3 expression in memory CD4 T cells (P<0.05). Taken together these results indicate that HIV infection significantly influences Mtb-specific CD4 T-cell cytokine and transcription factor expression profiles. Interestingly, the proportion of IL-2-producing Mtb-specific CD4 T cells inversely correlated with the percentage of Mtb-specific CD4 T cells expressing PD-1 (r = -0.697; P<0.05),. Finally, we showed that the serum levels of IL-1α, IL-6, IFN-α2, IFN-β, IFN-ω, IL-23, MCP-1, IP-10 and CRP were significantly reduced in Mtb/ HIV co-infected individuals with PTB as compared to HIV negative individuals with PTB (P<0.05), suggesting that HIV infection significantly suppresses Mtb- induced systemic pro-inflammatory cytokine response. Conclusion: Taken together, this study suggests that HIV infection significantly impairs functionally favourable Mtb-specific CD4 T-cell responses in Tanzanian individuals suffering from LTBI or PTB. 775 VARIATION IN THE NLRP3 GENE IS ASSOCIATED WITH INFLAMMATION AND MORTALITY IN HIV/TB Shruthi Ravimohan 1 , Kebatshabile Ngoni 2 , Neo Tamuhla 2 , Caroline Tiemessen 3 , DrewWeissman 1 , Gregory Bisson 1 1 University of Pennsylvania, Philadelphia, PA, USA, 2 Botswana–UPenn Partnership, Gaborone, Botswana, 3 National Institute for Communicable Diseases, Johannesburg, South Africa Background: Nearly 10-20% of advanced HIV/tuberculosis (TB) co-infected patients die despite antiretroviral therapy (ART) initiation. While hyper- inflammation likely contributes to death, the underlying mechanisms are unclear. We hypothesized that common variation in innate immune genes may play a role by modulating inflammation. As single nucleotide polymorphisms (SNP) in inflammasome pathway genes have been linked to increased risk for several inflammatory diseases and variable levels of inflammation, we investigated the association between these SNPs and early mortality in HIV/TB patients initiating ART. Methods: We conducted a sub-analysis of a prospective cohort study of advanced HIV/TB patients initiating ART in Botswana. We determined variation at eight loci in the NLRP3, CARD8, IL1β, IL-18, and P2RX7 genes. We assumed a dominant model of inheritance for analyses, where patients homozygous for the major allele were compared to those heterozygous or homozygous for the minor allele. In unadjusted analyses, we determined the association between SNPs and death. We used a logistic regression model and adjusted for potential confounders of this association one at a time. For SNPs associated with death, we explored their relationship with pre- and post-ART levels of systemic inflammatory markers using Wilcoxon rank sum tests. Results: Ninety-four (55%) of 170 patients enrolled in the parent study had samples available for analysis. Of the 94 patients, 82 (87%) were survivors

468). Overall, 279 (66%) had ≥1 TB-related symptom, 197 (46%) had a CRP >5 mg/L, and 42 (10%) had a positive TB culture. Sensitivity for TB of both CRP and the WHO symptom screen was 90.5% (95% CI 77.4-97.3), however the specificity of CRP was 58.5% (95% CI 53.4-63.5) compared to 37.1% (95% CI 32.2-42.1) for the symptom screen. The negative likelihood ratio (LR-) for CRP was lower than for symptom screen (0.16 vs. 0.26), indicating better performance of CRP as a “rule-out” test. CRP accuracy was similar to symptom screen in persons with CD4>200 (LR- 0.3 vs. 0.31) but better in those with CD4≤200 at highest risk for TB (LR- 0.11 vs. 0.36). Using CRP to screen for TB resulted in 228 persons screening negative, compared to 146 using the symptom screen - a 56% increase in persons eligible for IPT and ART without requiring confirmatory TB testing, and not missing any additional active TB cases. Conclusion: CRP was as sensitive as the WHO symptom screen to diagnose TB in HIV-infected outpatients and was substantially more specific. Preserved sensitivity and higher specificity in higher CD4 strata are relevant for ART initiation in the growing proportion of persons eligible for ART with higher CD4 counts. Using CRP to exclude active TB in persons with HIV could reduce the need for time-consuming and costly diagnostic TB testing compared to current practice using the WHO symptom screen.

Poster Abstracts

773 PLASMA INDOLEAMINE 2, 3-DIOXYGENASE, A BIOMARKER FOR TUBERCULOSIS IN HIV INFECTION

Clement G. Adu-Gyamfi 1 , Tracy Snyman 1 , Chris Hoffmann 2 , Neil A. Martinson 1 , Richard E. Chaisson 2 , Jaya A. George 1 , Melinda S. Suchard 1 1 University of the Witwatersrand, Johannesburg, South Africa, 2 Johns Hopkins University, Baltimore, MD, USA Background: There exist no biomarker for diagnosing and/ or predicting active TB in HIV infected patients. Indoleamine 2, 3-dioxygenase (IDO) is an immunoregulatory enzyme which breaks down tryptophan to kynurenines. We evaluated whether IDO activity, as measured by Kynurenine-to-Tryptophan ratio, could diagnose or predict active TB disease in HIV infected adults. Methods: Using ultra-performance liquid chromatography mass spectrometry, we measured Kynurenine and Tryptophan concentrations in plasma of 32 HIV infected patients who developed active TB followed up prospectively. We compared with 70 HIV infected control subjects from the same cohort who did not develop TB, matched by age, sex and CD4 count, and 37 unmatched HIV infected patients diagnosed with pneumonia. Results: At time of TB diagnosis, IDO activity was significantly higher in TB patients than controls (P < 0·0001). Six months prior to TB diagnosis IDO activity was significantly higher than controls in all those who later developed TB (P < 0·0001). After 6 months of TB treatment, IDO activity in TB patients declined to similar levels as that of controls. IDO activity was 4 fold higher in TB patients than pneumonia patients, and could distinguish them. Using a receiver operating characteristic curve, IDO activity gave a sensitivity of 97%, specificity of 99% positive and negative predictive values of 89% and 100% for detecting active TB disease. Conclusion: Plasma IDO activity is suitable as a biomarker of active TB in HIV positive patients.

CROI 2018 290