Preservation of lymphocyte functional fitness in perinatally-infected and treated HIV+ pediatric patients displaying sub-optimal viral control.

Background: Host-pathogen dynamics associated with HIV infection are quite distinct in children versus adults. We interrogated the functional fitness of the lymphocyte responses in two cohorts of perinatally infected HIV+ pediatric subjects with early anti-retroviral therapy (ART) initiation but divergent patterns of virologic control. We hypothesized that sub-optimal viral control would compromise immune functional fitness.
Methods: The immune responses in the two HIV+ cohorts (n = 6 in each cohort) were benchmarked against the responses measured in age-range matched, uninfected healthy control subjects (n = 11) by utilizing tests for normality, and comparison [the Kruskal-Wallis test, and the two-tailed Mann-Whitney U test (where appropriate)]. Lymphocyte responses were examined by intra-cellular cytokine secretion, degranulation assays as well as phosflow. A subset of these data were further queried by an automated clustering algorithm. Finally, we evaluated the humoral immune responses to four childhood vaccines in all three cohorts.
Results: We demonstrate that contrary to expectations pediatric HIV+ patients with sub-optimal viral control display no significant deficits in immune functional fitness. In fact, the patients that display better virologic control lack functional Gag-specific T cell responses and compared to healthy controls they display signaling deficits and an enrichment of mitogen-stimulated CD3 negative and positive lymphocyte clusters with suppressed cytokine production. Conclusions: These results highlight the immune resilience in HIV+ children on ART with sub-optimal viral control. With respect to HIV+ children on ART with better viral control, our data suggest that this cohort might potentially benefit from targeted interventions that might mitigate cell-mediated immune functional quiescence.

Introduction

It is estimated that at the end of 2019, about 2 million children are currently living with the Human Immunodeficiency Virus (HIV)[1]. One of the major success stories in the fight against the Acquired Immunodeficiency Syndrome (AIDS) has been the development of combination anti-retroviral therapy (ART)[2]. Timely access to ART has contributed to a dramatic reduction in the mother to child transmission rates of HIV (perinatal HIV transmission accounted for only 65 out of the ~38,000 new HIV diagnoses in the US in 2018), and 50% of HIV-infected children die by the second year of life if they do not receive ART[3-5]. In contrast to ART, efforts to develop a successful vaccine against HIV have faced multiple setbacks[6]. This likely reflects the fact that we do not yet fully understand the complex host pathogen dynamics associated with HIV infection. Hence, in order to attempt a cure for this disease we need to further enhance our understanding of the virus–host interactions, be aware of its nuances, and perhaps look beyond a one size fits all approach. Related to this point, there are some intriguing differences between adults and children infected with HIV. Between 5% and 10% of ART naive, HIV+ children are slow progressors with clearly detectable viremia but they still maintain normal, age-range associated CD4 T cell counts, and they do not harbor HLA class I alleles that have been described to mediate protective CD8 T cell responses[7-9]. Prevailing evidence indicates that there does not appear to be a good correlation between viral load, immune activation, and CD4 counts in children[7,8,10]. Furthermore, HIV-associated disease can progress quickly in ART naïve children even if they harbor HLA class I alleles deemed protective and despite mounting an early and active virus-specific CD8 T cell response (presumably by CD8 T cell-mediated cytotoxicity of infected CD4 T cells)[7-9]. ART-treated children display a more rapid rebound in CD4 T cell numbers compared to adults, and a greater proportion of naïve T cells populate the peripheral T cell repertoire following ART-mediated viral control in the pediatric age group[11-14]. Moreover, the extent to which CD4 T cell function recovers following viral suppression is greater in children than that observed in treated, aviremic adults[11,13,15,16]. Age-related changes in thymic output quite likely contribute to these differences between ART-treated adults and children[14]. Interestingly, the viral reservoir in children is not populated with escape variants to the extent observed in adults in spite of the fact that immune selection pressure induces escape variants early after infection in children[7-9,17,18]. Collectively, this has led to the speculation that manipulating the immune response in children might present better odds for a pathway to a cure against HIV. This concept is also gaining traction due to the fact that the expectation of faithful drug compliance (especially during adolescence) and lifelong adherence to ART that is initiated in childhood is perhaps unrealistic[19-21]. This realization coupled with the cost and the attendant side effects of prolonged therapy is prompting a serious re-assessment of what might constitute the most effective long-term management approach for this disease[22,23].

In this study, we explored the hypothesis that sub-optimal viral control will compromise immune functional fitness in perinatally infected HIV+ patients with early ART initiation[24,25]. We evaluated the immune functional fitness of two US based cohorts of perinatally infected HIV+ study subjects that initiated ART within the first year of life. Cohort 1 consists of individuals that displayed better virologic control compared to cohort 2 study subjects reflecting more consistent adherence, as well as a lower degree of viral resistance to ART. The results from both cohorts were benchmarked against responses measured in age-range matched, uninfected control subjects. We measured polyfunctional responses (IFNγ, TNFα, IL-2, and IL-21) and degranulation potential (CD107a) by intracellular cytokine staining assays in defined lymphocyte subsets following PMA/ionomycin stimulation, as well as following Gag-potential T cell epitope (PTE)-peptide pool treatment. Additionally, we queried our mitogen-stimulated intracellular cytokine data using unsupervised clustering analyses to uncover novel populations that clearly segregated healthy control donors from the HIV-exposed subjects. We also utilized phosflow to track the phosphoprotein signature in the lymphocyte subsets following treatment with IFNγ, IFNα, IL-2, IL-4, IL-6, IL-7, IL-10, IL-15, IL-21 and after anti-CD3 stimulation. Finally, we examined serum Ab responses to childhood vaccines against Varicella, tetanus, measles and Haemophilus influenzae b (Hib).

Our data reveal that sub-optimal viral control did not subvert immune functional fitness. On the other hand, optimal viral suppression was characterized by undetectable Gag-specific T cells, subdued lymphocyte signaling responses, and an enrichment in unique, predominantly CD3 negative lymphocyte clusters, that displayed suppressed cytokine responses.

Methods

Study subjects

Perinatally infected HIV+ study subjects with early ART initiation (n = 6 each in cohorts 1 and 2), and age-range matched, HIV-negative control subjects (n = 11), were recruited between June 2018–November 2019, from the Special Infectious Diseases (SID) clinic, and the Academic General Pediatrics (AGP) clinic at the Ann and Robert H. Lurie Children’s Hospital of Chicago (Supplementary Table 1). Fourteen subjects were African–American, four identified as white, and race was unknown/not reported for five subjects. In terms of ethnicity, seven study subjects identified as Hispanic/Latino. This study was carried out in accordance with the recommendations of the Human Research Protection Program Guidelines of our Institutional Review Board (IRB). This protocol was approved by our IRB (No. 2018-1541; Study Title: Evaluation of immune functional fitness in perinatally infected HIV+ study subjects with early ART initiation). All study subjects ≥18 years gave written informed consent in accordance with the Declaration of Helsinki. Study subjects between the ages of 12–17 years provided written assent and in addition we obtained written informed consent from their legal guardians. Written informed consent was obtained from the legal guardians of all study subjects below the age of 12 years. For the purpose of the measles serology assays that were performed on de-identified samples at the Centers for Disease Control and Prevention (CDC), the CDC Human Research Protections Office determined that this study was not human subjects research and exempt from the CDC Institutional Review Board (IRB) review.

Measurement of peripheral T cell counts and viral load measurements

We are a Clinical Laboratory Improvement Amendments (CLIA)-certified and College of American Pathologists (CAP)-accredited laboratory. The peripheral CD4 and CD8 T cell counts for the study subjects were performed using our clinically validated flow-cytometry based immunophenotyping assay that utilizes the BD MultitestTM 6 color TBNK Reagent with TrucountTM tubes (BD Biosciences, San Jose, CA) (Supplementary Data). From 1999 to November 2008, viral load quantification was performed using the Roche Amplicor HIV-1 Monitor test (Roche Molecular Systems, Inc., Branchburg, NJ). After November 2008, the Abbott Real Time HIV-1 assay (Abbott Molecular Inc., Des Plaines, IL) was deployed to measure the viral load (Supplementary Data). For the purpose of this study a cut-off value of >200 HIV-RNA copies/ml of plasma was used to define a detectable viral load.

Staining for intracellular cytokines and CD107a by flow cytometry using whole blood samples

We used intracellular cytokine staining per published guidelines[26-28]. Whole blood sample aliquots were treated with medium alone or stimulated with phorbol 12-myristate 13-acetate (PMA) (10 ng/mL) and Ionomycin (Ion) (1 µg/mL), or the HIV-1 potential T cell epitopes (PTE) Gag peptide pool (2 μg/ml) (NIH AIDS Reagent Program, Germantown, MD; Catalog No. 12437) for 5 hours at 37 °C in the presence of Brefeldin A (Golgi-Plug) (BD Biosciences, San Jose, CA) [or Monensin (Golgi-Stop) (BD Biosciences, San Jose, CA) for CD107a assays]. The following MAb clones directed against the human antigens listed below were used at the manufacturer recommended doses to detect the intracellular cytokines: IFNγ (clone 4S.B3); TNFα (clone MAb11); IL-21 (clone 3A3-N2), and IL-2 (clone MQ1-17H12) (all from Biolegend, San Diego, CA). Relevant isotype control Abs (Biolegend, San Diego, CA) were used at matching doses to determine background level staining (Supplementary Table 2). For CD107a detection, the relevant isotype-control Ab (mouse IgG1, κ-PE; clone X40; BD-Biosciences, San Jose, CA) or the CD107a-PE Ab (clone H4A3; BD-Biosciences, San Jose, CA) was added at the initiation of stimulation[26,28]. MAb directed against human CD3 and CD8 were used for surface staining (BD Biosciences, San Jose, CA). Lymphocytes were gated based on light scatter characteristics followed by gating on CD3+ and CD3-negative populations. CD8+ and CD8− T cells were identified within the CD3+ population. Additional subsets were similarly delineated based on surface or cytosolic marker expression within the indicated parent population. The samples were acquired using FACS Canto-II instruments (Becton-Dickinson, Franklin Lakes, NJ) (a million total events/tube were acquired for the Gag-PTE-treated tubes) and data were analyzed using FlowJo software (version ≥ 10.6) (BD Biosciences, San Jose, CA). Each study subject was analyzed once (Supplementary Data).

Unsupervised (automated) clustering analyses

Cluster Identification, Characterization, and Regression (CITRUS) (Beckman Coulter, Brea, CA) is an automated clustering algorithm designed with the purpose of identifying statistically significant differences in the biological properties of cellular populations identified in multiple experimental groups[29]. CITRUS was used to compare the relative abundance of cells in each subset compared to the whole population, as well as the median expression levels of the four cytokines between the healthy control samples and cohorts 1 and 2 (Supplementary Data). For evaluating the abundance feature, we selected either CD3+ or CD3− lymphocytes as the starting population, and subsequently selected CD8, IFNγ, TNFα, IL-2 and IL-21 as the clustering channels. To evaluate the median feature, lymphocytes were selected as the starting population, and CD3 and CD8 were placed in the clustering channels, and the four cytokines (IFNγ, TNFα, IL-2 and IL-21) were listed as the statistics channels. For both the median and abundance features, we utilized the Predictive-Nearest Shrunken Centroid [Partitioning Around Medoids-R (PAMR)] association model, and an equal number of events were evaluated from each fcs file. The CITRUS run was repeated a minimum of four times for every group pairing (Control versus Cohort 1, Control versus Cohort 2, and Cohort 1 versus Cohort 2) for assessing the abundance and median cluster characterizations.

Phosflow analyses for evaluating the phosphorylation status of signaling nodes

We performed phosflow staining using established protocols[26,30,31]. Whole blood sample aliquots were treated for 15 min at 37 oC with phosphate buffered saline (PBS) or the following recombinant human cytokines: IFNγ, IL-4, IL-10, IL-15 (BD Biosciences, San Jose, CA), IFNα (R&D Systems, Minneapolis, MN), IL-2 (PeproTech, Rocky Hill, NJ), IL-7 (Thermo Fisher Scientific, San Diego, CA), IL-6 and IL-21 (Biolegend, San Diego, CA). The cells were then fixed, permeabilized and stained with the MAb targeting the following molecules: anti-human human CD3, CD8 (Biolegend, San Diego, CA), phospho-STAT1 (pY701; clone 4a), phospho-STAT3 (pY705; clone 4), phospho-STAT5 (pY694; clone 47), and phospho-STAT6 (pY641; clone 18) (all from BD Biosciences, San Jose, CA). For measuring phospho-ZAP70, whole blood sample aliquots were incubated with either mouse anti-human CD3 MAb (clone UCHT1) (Biolegend, San Diego, CA) or mouse IgG1, κ (clone MOPC-21; Biolegend, San Diego, CA) for 30 min on ice, washed and then treated with goat anti-mouse IgG (clone Poly4053; Biolegend, CA) on ice for an additional 30 min before transferring the tubes to a 37 °C water bath for 5 min. The cells were then fixed, permeabilized and stained per manufacturer’s instructions with an antibody targeting the phospho-ZAP70 (PY319) (clone 17A/P-ZAP70; BD Biosciences, San Jose, CA) and with a non-competing Ab clone targeting CD3 (clone SK7) and a MAb to CD8 (Biolegend, San Diego, CA). 50,000–100,000 events/tube were acquired on FACS Canto-II instruments (Becton-Dickinson, Franklin Lakes, NJ) and the data were analyzed using Cytobank software (Cytobank, Santa Clara, CA). Each study subject was analyzed once (Supplementary Data).

Microbial serology assays

IgG responses to tetanus toxoid (TT) and Haemophilus influenzae type b (Hib) were evaluated in duplicate with clinically validated enzyme immunoassays (EIA) that utilize the VaccZymeTM TT-IgG and the VaccZymeTM Hib-IgG kit from Binding Site (Edgbaston, Birmingham, UK) (Supplementary Data). The IgG responses to Varicella-zoster virus (VZV) were also evaluated in duplicate with a clinically validated assay that utilizes the CAPTIATM EIA kits manufactured by Trinity Biotech (Jamestown, NY) (Supplementary Data). All three EIA assays were run on the fully automated, open DS2 platform (Dynex Technologies, Chantilly, VA). The measles virus serology assays were performed at the Viral Vaccine Preventable Diseases Branch at the Centers for Disease Control and Prevention (Atlanta, GA) (Supplementary Data). Measles immunity was defined as a serum IgG titer >120 measured by plaque reduction neutralization (PRN). A PRN titer of >120 was considered protective for measles and is equivalent to 120 mIU/ml based on use of the WHO second international standard[32]. The IgG response to the measles virus was measured by EIA using the measles IgG test system (Zeus Scientific, Branchburg, NJ), and the avidity of measles-specific IgG antibodies was tested by modification of a commercial measles IgG EIA (Captia Measles IgG, Trinity Biotech, Jamestown, NY), as described by Latner et al. and Mercader et al.[32,33].

Statistics and reproducibility

For the mitogen-induced intracellular cytokine secretion, CD107a expression assays, phosflow analyses and vaccine antigen-specific Ab responses, the D’Agostino–Pearson omnibus normality test (α = 0.05) for normal distribution was performed followed by the Kruskal–Wallis and Dunn’s multiple comparison’s tests (α = 0.05) to identify any potential statistically significant differences in the patterns measured between the three cohorts (Prism version ≥8.0, GraphPad, San Diego, CA). The statistical comparison between the control and cohort-1 subjects depicted in the form of scatter-plot column graphs for the subsets identified through the CITRUS analyses utilized the two-tailed Mann–Whitney U test (Prism version ≥8.0, GraphPad, San Diego, CA). Statistical assessments for the Gag PTE-pool induced intracellular cytokine secretion and CD107a surface mobilization assays also utilized the D’Agostino–Pearson omnibus normality test (α = 0.05) followed by the two-tailed Mann–Whitney U test (Prism version ≥8.0, GraphPad, San Diego, CA). For all statistical analyses, measurements were taken from distinct samples. The figure legends indicate the number times each sample was evaluated.