Identification of immune biomarkers for use in early HIV detection and monitoring in sub-Saharan Africa = Identificación de biomarcadores de respuesta inmunitaria para la detección temprana y monitorización del VIH en África sub-Sahariana

Abstract

BACKGROUND AND RATIONALE

Acute HIV infection (AHI) is the period between the acquisition of the human immunodeficiency virus (HIV) and the development of HIV-specific antibodies that define seroconversion. AHI is characterised by high HIV viral replication and, in most cases, a transient non-specific febrile illness that typically occurs around 2 weeks after the HIV- transmission event. Primary HIV infection (PHI) is generally considered the period up to 6 months after infection and is a rapidly evolving phase characterized by the stepwise gain in positivity for the detection of HIV-RNA and HIV-specific antibodies. Different HIV antigen specificities appear in sequence after HIV transmission as do immunoglobulin G (IgG) subclass responses. As such, using different diagnostic tools, PHI has been categorised into ‘Fiebig stages’ that are useful in approximating infection date with relative accuracy.

As a result of high viremia in bodily fluids, individuals are considered hyper-infectious during AHI. In areas of high HIV incidence, this phenomenon could contribute greatly to fuelling the worldwide HIV pandemic. Despite the importance of early diagnosis and treatment to reduce onward transmissions and prevent substantial irreversible immunological damage, AHI represents a ‘window period’ during which persons infected with HIV are commonly undiagnosed. Routine second generation HIV-rapid test algorithms provide negative or indeterminate results for up to 6-8 weeks after infection. During this time, HIV can only be diagnosed by detecting the presence of the virus itself. The current gold-standard test for confirming viremia is HIV-RNA viral load (VL) testing. However, technical and financial constraints make this technique very limited in low-income areas such as Sub-Saharan Africa, where the prevalence of AHI among febrile patients may reach 3%.

VL testing is also used to monitor the efficacy of anti-retroviral treatment (ART). Achieving effective ART monitoring is a key determinant to ensure viral suppression and reach the UNAIDS 90-90-90 targets. Although considerable international efforts have resulted in a dramatic increase in ART coverage in the last years, relatively little progress has been achieved in the development of simple, accurate and affordable tools that allow proper surveillance of ART efficacy. VL monitoring is important for timely diagnosis of virological failure (VF) to allow early adherence interventions, prevent further transmissions and avoid delays in regimen switches that could lead to disease progression or emergence of drug resistances. Detecting virological failure depends on VL testing; whose availability is very limited in low and middle-income countries (LMIC) due to cost and operational constraints. The alternative to VL has often been clinical and/or immunological monitoring, which frequently results in patients remaining on failing ART as well as unnecessary regimen switches. Indeed, recent cross-sectional surveys reported that around 25%-35% of individuals on ART in Mozambique had detectable HIV viremia levels.

As viremia increases during PHI, there is a striking cascade response of inflammatory cytokines. Significant efforts have been made to characterise host and viral proteins present

during AHI aiming to identify biomarkers of progression or key pathological pathways that could be targeted to minimize HIV-induced immune damage over the course of infection. Subsets of T cells can be defined by their specificity, surface phenotype or degree of maturation, and any or all of these parameters can be affected by HIV infection. During PHI, many cells of the immune system show signs of extensive activation and a progressive loss of resting subsets. Generally, untreated HIV-infection is characterized by progressive CD4 T- cell depletion and CD8 T-cell expansion. The profound CD4 T-cell depletion is linked directly to the risk for opportunistic infections and mortality, while CD8 T-cell activation and exhaustion have been observed to be strong correlates of disease progression. Such alterations of CD4 and CD8 homeostatic mechanisms lead to progressive loss of the naïve and memory T-cell pool, resulting in an imbalance in T-cell phenotypes. Similarly, after HIV infection, accelerated aging of T cells or immunosenescence has been also associated with risk of adverse clinical events in HIV infected individuals.

The challenges of identifying individuals during the AHI phase have resulted in a lack of critical information that constrains the development of therapeutic interventions. In this thesis, we provide a longitudinal characterization of T-cell subsets and the expression of soluble inflammatory biomarkers over the first year after PHI in a cohort of Mozambican adults and compare these changes with Chronic HIV-infection (CHI). Additionally, we assess the predictive power of these soluble biomarkers as surrogates of viremia for detection of AHI in seronegative febrile individuals and for identification of virological failure in ART- treated subjects.

METHODS

This thesis is based on research conducted at the Barcelona Institute for Global Health (ISGlobal)/ Hospital Clinic-Universitat de Barcelona in Spain, the AIDS Research Insitute/ Germans Trias i Pujol Research Institute (IGTP) and at the Centro de Investigaçao em Saude de Manhica (CISM) in Mozambique.

To longitudinally analyze individuals during PHI, a screening based on reported-fever and pooled VL- testing was used to identify AHI in HIV-seronegative adults presenting at the Manhiça District Hospital (MDH), Mozambique. HIV-uninfected and chronically HIV infected individuals, both ART-naïve and on first-line ART, were also recruited at MDH in the context of the study. Plasma levels of inflammatory and immune biomarkers were subsequently determined by Luminex and ELISA, anti-HIV antibodies were analysed by flow-cytometry and Western Blot (WB) and T-cell phenotyping was performed through multi-panel flow- cytometry analysis.

To evaluate biomarkers in treated HIV infected individuals, samples from a resistance survey study performed in 2013 were retrospectively analysed.

The thesis is presented as a collection of four articles, of which three are accepted for publication in peer-reviewed international journals, and one is under review for publication.

KEY RESULTS

The findings from the studies that constitute this thesis provide a further characterization of the dynamics of immune response biomarkers over the different stages of HIV infection among adults in Mozambique.

Immune response biomarkers across Fiebig staging at primary HIV infection

A total of 85 AHI individuals were identified in our cohort as seronegative or indeterminate for rapid test and positive for plasma HIV-RNA among Mozambican adults seeking health care at the MDH. This represents an AHI prevalence of 3% among seronegative individuals reporting febrile illness. Soluble biomarkers, including inflammatory cytokines and general and HIV-specific antibody subtypes, were determined over different Fiebig stages at PHI, together with clinical and immunological characteristics. We compared cytokine levels between individuals at pre- (Fiebig I-IV) and post-seroconversion stages (Fiebig V-VI) at the screening visit. Thus, we identified a signature of four cytokines composed of BAFF, MCP-1, sCD163 and MIG that is highly associated with the PHI phase prior to development of the HIV-specific humoral response as determined by standard Western Blot serology.

Longitudinal analysis of soluble and cellular biomarkers along the HIV infection process

After longitudinal follow up of the PHI individuals, T-cell subsets and the expression of soluble inflammatory and immune biomarkers were characterized over the first year after of infection. Although plasma HIV viremia, CD4 and CD8 T cell counts undergo a rapid stabilization after HIV infection, several immunological parameters, including Th1Th17 CD4 T cells and activation or exhaustion of CD8 T cells continue to decrease even after 9 months post-infection. Importantly, no sign of immunosenescence was detected over the first year of HIV infection and no significant changes were observed in the Tregs population. Levels of IP-10, MCP-1, BAFF, sCD14, TNFR2 and TRAIL were significantly overexpressed at the first month of infection and underwent a prompt decrease in the following months. However, MIG and CD27 levels started to increase 1 month after infection and remained over- expressed for almost one year post-infection. Early levels of plasma TNFR2, sCD27, BAFF, IL- 10 and sCD14 cytokines were significantly associated with later levels of exhausted CD4 T- cells or with CD8 T-cell activation.

Biomarkers as an accurate tool to identify acute HIV infection in febrile individuals

In order to evaluate whether levels of a single or a combination of biomarkers had predictive accuracy to identify AHI among HIV-seronegative adults presenting with reported fever at the MDH, plasma levels of 49 inflammatory biomarkers from AHI (n=61) and non- HIV infected outpatients (n=65) were compared. The cytokine IP-10 demonstrated the best predictive accuracy for AHI detection (AUC=0.88 [95%CI 0.80-0.96]). A cut-off value of IP- 10≥161.6pg/mL provided a sensitivity of 95.5% (95%CI 85.5-99.5) and a specificity of 76.5% (95%CI 62.5-87.2) for AHI identification. Thus, an IP-10-based screening for subsequent AHI identification with VL could reduce the number of VL determinations necessary by 75%.

After a cost-effectiveness analysis of this IP-10-based approach, we concluded that the implementation of an IP-10 screening test could avert from 21 to 84 new infections and save from US$176,609 to US$533,467 to the health system per 1,000 tested patients.

IP-10 predictive power to detect virological failure in treated patients

Due to the strong association with VL, we hypothesized that plasma IP-10 levels could be a surrogate marker of detectable viremia in ART-treated individuals. Consequently, we found that IP-10 levels were significantly higher in ART-treated subjects with detectable VL (108.2 pg/mL) as compared to those with undetectable VL (38.0 pg/mL) (U-test p<0.0001) in a cohort of 316 HIV-infected individuals on ART for more than a year. An IP-10 univariate model demonstrated high accuracy for prediction of detectable viremia (AUC=0.85 [95% Confidence Interval (CI) 0.80-0.90]). Using a cut-off value of IP-10≥44.2 pg/mL, the IP-10 model identified viremic ART-treated subjects with 91.9% sensitivity (95% CI 83.9-96.7) and 59.9% specificity (95% CI 52.0-67.4). Accordingly, we found that such IP-10 screening for potential virological failure would reduce by 43% the number of VL determinations required to monitor the same number of patients and this approach could potentially save 38% of VL-derived costs to the health system.

CONCLUSIONS AND RECOMMENDATIONS

AHI prevalence in febrile seronegative adults presenting at the MDH was found to be 3% as reported in 2008, having thus remained unchanged in this population in the last 5 years. After quantification of the soluble biomarkers across the different Fiebig stages described in PHI, we identified a signature of four cytokines composed of BAFF, MCP-1, sCD163 and MIG that was highly associated with the pre-seronversion phase as determined by WB serology. These effectors could provide clues for the development of vaccine or immunomodulatory strategies aimed at reducing the irreversible immune damage inflicted during PHI.

Throughout characterization of cellular and soluble immune biomarkers along the different phases of HIV infection, we found that activated and effector memory CD8 T-cells together with Th1Th17 cells continued to decay several months after control of viremia. These findings indicate that balance in the T-cell compartments occurs months after viremia or CD4 count stabilize, suggesting persistent immune dysfunction in many different T cell subsets and raises the potential need for early initiation of therapy that could limit immunological damage.

From the 49 soluble biomarkers that were assessed in febrile seronegative individuals, IP-10 demonstrated the best predictive power for AHI detection, providing a sensitivity of 95.5% and a specificity of 76.5%. Implementation of an IP-10-based screening for subsequent AHI diagnosis with VL is a cost effective strategy that could avert up to 84 new infections and save up to US$500,000 to the health system per 1,000 tested patients.

IP-10 is also an accurate biomarker to screen individuals on ART for virological failure (VF), identifying 91.9% of patients with detectable viremia with a specificity of 59.9%. Employing

IP-10 as a screening tool to target the individuals on ART most likely to require VL testing would reduce by 43% the number of VL determinations required to control viral suppression and it could save 38% of VL-derived costs to the health system.

Thus, IP-10 quantification could be developed as a screening tool to identify both AHI in febrile seronegative individuals and VF in patients on ART with subsequent viral load testing. The implementation of these algorithms would facilitate AHI diagnosis and ART- monitoring in LMIC, as sub-Saharan Africa. Nevertheless, further research is necessary to explore the impact of other common HIV comorbidities on HIV-induced levels of IP-10, validate the IP-10 predictive power and optimize model cut-off values in other populations.

NTRODUCCIÓ: La fase aguda de la infecció pel virus de la immunodeficiència humana (VIH) és el període comprès entre l'adquisició del virus i el desenvolupament d'anticossos específics que defineixen la seroconversió i es coneix com AHI per les seves sigles en anglès (Acute HIV Infection). La AHI es caracteritza per un alt nivell de virus en els fluids corporals i, en la majoria dels casos, un quadre transitori de febre no específica. Com a resultat, els individus es consideren híper-infecciosos durant l'AHI i en zones o poblacions vulnerables d'alta incidència de VIH, aquest fenomen podria contribuir en gran manera a la pandèmia mundial del VIH. El diagnòstic precoç i l'inici primerenc del tractament són intervencions clau per reduir les potencials transmissions i prevenir un substancial dany immunitari irreversible. Malgrat la seva importància, el AHI representa un "període finestra" que pot durar fins a 2 mesos durant el qual les persones infectades només pot ser diagnosticades mitjançant la prova de càrrega viral (CV), que consisteix a detectar el ARN del virus en sang. La prova de CV també és necessària per al monitoratge de l'eficàcia del tractament antiretroviral. No obstant això, les restriccions logístiques, tècniques i financeres fan que aquesta prova de CV sigui d'accés molt limitat a les zones d'escassos recursos com l'Àfrica Subsahariana. OBJECTIUS: En aquesta tesi, es proporciona una caracterització longitudinal dels subconjunts de cèl·lula del sistema immune i l'expressió de biomarcadores inflamatoris solubles durant el primer any d'infecció en una cohort d'adults moçambiquesos i es comparen amb els nivells en la fase crònica de la infecció. A més, també s'avalua el poder predictiu d'aquests biomarcadores solubles per a la detecció de la AHI en individus seronegatius amb febre i per a la identificació de fallada terapèutica en individus tractats. RESULTATS: Es van testar un total de 4011 adults a l'Hospital Distrital de Manhiça (Moçambic), dels quals 3% dels individus seronegatius que van reportar símptomes febrils es trobaven en fase de AHI. Dels 49 biomarcadores solubles que es van avaluar, IP-10 va demostrar tenir el millor poder predictiu per a la detecció de AHI, proporcionant una sensibilitat del 95.5% i una especificitat del 76.5%. La implementació d'un cribratge de pacients amb simptomatologia febril basat en IP-10 per al posterior diagnòstic de AHI amb CV és una estratègia rendible que podria evitar fins a 84 noves infeccions en països d'alta incidència de VIH i estalviar més de 500,000US$ al sistema de salut per cada 1,000 pacients analitzats. L'IP-10 va demostrar també ser un biomarcador precís per detectar els casos de fallada terapèutica entre individus en tractament, identificant el 91.9% dels pacients amb viremia detectable amb una especificitat del 59.9%. CONCLUSIONS: La quantificació de la proteïna IP-10 podria desenvolupar-se com una eina per identificar tant AHI en individus seronegatius febrils com fallada terapèutica en pacients tractats mitjançant confirmació posterior per CV. La implementació d'aquests algorismes facilitaria el diagnòstic de AHI i el monitoratge del tractament en àrees d'escassos recursos, com l'Àfrica sub-Sahariana.

INTRODUCCIÓN:

La fase aguda de la infección por el virus de la inmunodeficiencia humana (VIH) es el periodo comprendido entre la adquisición del virus y el desarrollo de anticuerpos específicos que definen la seroconversión, y se conoce como AHI (en inglés Acute HIV Infection). La AHI se caracteriza por un alto nivel de virus en los fluidos corporales y, en la mayoría de los casos, un cuadro transitorio de fiebre no específica. Como resultado, los individuos se consideran híper-infecciosos durante la AHÍ y en zonas o poblaciones vulnerables de alta incidencia de VIH, este fenómeno podría contribuir en gran medida a la pandemia mundial del VIH. El diagnóstico precoz y el inicio temprano del tratamiento son intervenciones clave para reducir las potenciales transmisiones y prevenir un sustancial daño inmunitario irreversible. A pesar de su importancia, el AHI representa un "periodo ventana" que puede durar hasta 2 meses durante el cual las personas infectadas sólo puede ser diagnosticadas mediante la prueba de carga viral (CV), que consiste en detectar el ARN del virus en sangre. La prueba de CV también es necesaria para la monitorización de la eficacia del tratamiento anti-retroviral. Sin embargo, las restricciones logísticas, técnicas y financieras hacen que esta prueba sea de acceso muy limitado en las zonas de escasos recursos como el África Subsahariana.

OBJETIVOS:

En esta tesis, se proporciona una caracterización longitudinal de los subconjuntos de células del sistema inmune y la expresión de biomarcadores inflamatorios solubles durante el primer año de infección en una cohorte de adultos mozambiqueños y se compara estos cambios con la infección por el VIH en la fase crónica. Además, también se evalúa el poder predictivo de estos biomarcadores solubles para la detección de la AHI en individuos seronegativos con fiebre y para la identificación de fallo terapéutico en individuos tratados.

RESULTADOS:

Se testaron un total de 4011 adultos en el Hospital Distrital de Manhiça, de los cuales 3% de los individuos seronegativos que reportaron síntomas febriles se encontraban en fase de AHI. De los 49 biomarcadores solubles que se evaluaron, IP-10 demostró tener el mejor poder predictivo para la detección de AHI, proporcionando una sensibilidad del 95.5% y una especificidad del 76.5%. La implementación de un cribado de pacientes con sintomatología febril basado en IP-10 para el posterior diagnóstico de AHI con CV es una estrategia rentable que podría evitar hasta 84 nuevas infecciones en países de alta incidencia de VIH y ahorrar más de 500,00US$ al sistema de salud por cada 1,000 pacientes analizados. El IP-10 demostró también ser un biomarcador preciso para detectar los casos de fallo terapéutico entre individuos en tratamiento antirretroviral, identificando el 91.9% de los pacientes con viremia detectable con una especificidad del 59.9%.

CONCLUSIONES:

La cuantificación de IP-10 podría desarrollarse como una herramienta para identificar tanto AHI en individuos seronegativos febriles como fallo terapéutico en pacientes tratados mediante confirmación posterior por CV. La implementación de estos algoritmos en la práctica clínica facilitaría el diagnóstico de AHI y la monitorización del tratamiento en áreas de escasos recursos como el África sub-Sahariana, permitiendo así un mayor control de la pandemia del VIH.