HIV-infected patients of all ages frequently underperform in response to seasonal influenza vaccination, despite virologic control of HIV. The molecular mechanisms governing this impairment, as well as predictive biomarkers for responsiveness, remain unknown. This study was performed in samples obtained prevaccination (T0) from HIV-infected children who received the 2012–2013 seasonal influenza vaccine. Response status was determined based on established criterion for hemagglutination inhibition titer; participants with a hemagglutination titer ≥1:40 plus a ≥4-fold increase over T0 at 3 wk postvaccination were designated as responders. All children had a history of prior influenza vaccinations. At T0, the frequencies of CD4 T cell subsets, including peripheral T follicular helper (pTfh) cells, which provide help to B cells for developing into Ab-secreting cells, were similar between responders and nonresponders. However, in response to in vitro stimulation with influenza A/California/7/2009 (H1N1) Ag, differential gene expression related to pTfh cell function was observed by Fluidigm high-density RT-PCR between responders and nonresponders. In responders, H1N1 stimulation at T0 also resulted in CXCR5 induction (mRNA and protein) in CD4 T cells and IL21 gene induction in pTfh cells that were strongly associated with H1N1-specific B cell responses postvaccination. In contrast, CD4 T cells of nonresponders exhibited increased expression of IL2 and STAT5 genes, which are known to antagonize peripheral Tfh cell function. These results suggest that the quality of pTfh cells at the time of immunization is important for influenza vaccine responses and provide a rationale for targeted, ex vivo Ag-driven molecular profiling of purified immune cells to detect predictive biomarkers of the vaccine response.

Understanding how the immune systems of immunocompromised individuals respond to current vaccines is important to develop tailored vaccines. Influenza viruses cause illnesses with higher morbidity and mortality in patients with acquired immunodeficiencies and can lead to deadly pandemics due to antigenic drift. Seasonal vaccination is recommended in young children, the elderly, and in immunocompromised individuals to prevent influenza infection and its complications. Protection from influenza is mediated through neutralizing Abs against the viral surface proteins hemagglutinin (HA) and neuraminidase. The dilution of serum capable of blocking the hemagglutination reaction between HA and RBCs defines the titer of influenza Abs. A titer ≥1:40 is considered protective in healthy adults and, together with a ≥4-fold increase in titer postvaccination, represents a positive response to vaccination. Ab responses to the influenza vaccine are generated following the germinal center (GC) reaction that occurs between B cells and T follicular helper (Tfh) cells, a CD4 T cell subset that is critical for affinity maturation and somatic hypermutation in Ag-primed B cells (1). Binding of the surface molecule CXCR type 5 on Tfh cells to its ligand CXCL13 is required for the homing of Tfh cells to lymphoid follicles. The cytokine IL-21 is an important secretory product of Tfh cells and plays a dominant role in the GC reaction (2, 3). In the peripheral circulation, a subset of circulating memory CD4 T cells that express CXCR5 are referred to as peripheral Tfh (pTfh) cells. They manifest functional properties of GC Tfh cells, including the capacity for IL-21 secretion, which represents the strongest correlate of Tfh cell function in the peripheral blood (4). Previous studies demonstrated a relationship between pTfh cell and B cell function in vaccine responders (Rs) at 3–4 wk postinfluenza vaccination, including the expansion of pTfh cells concurrently with influenza H1N1 Ag–induced production of IL-21 in pTfh cells in vitro and help by purified pTfh cells to autologous B cells in coculture experiments for influenza A/California/7/2009 (H1N1)-specific IgG production (57).

In patients with HIV infection, seasonal influenza vaccination has emerged as a useful model for probing immune competency by evaluation of the serologic response to the vaccine (5, 6, 810). In patients on combination antiretroviral therapy (ART), we observed pTfh cell functional deficiencies in response to H1N1 flu Ag that worsen with aging (8). However, repeated immunizations with Ags, such as H1N1, which has been retained in the vaccine since its introduction in 2009 to address the H1N1 pandemic, has altered the serologic landscape, and baseline titers can be high (1113). Although informative biomarkers of influenza vaccine response and efficacy, such as early transcriptional changes in blood (1416) and serum Abs to neuraminidase postvaccination, are under investigation (17), immunocompromised populations, such as those infected with HIV, require independent testing and validation to design appropriate vaccination regimens (1820).

In this study, we used a targeted multiplex RT-PCR approach to evaluate T0 gene-expression patterns by relevant cell types (i.e., CD4 T and pTfh cells) in HIV-infected children and adolescents. Our study focused on T0 samples that were stimulated with influenza vaccine Ag (H1N1) in vitro to explore predictive measures of vaccine response. We hypothesized that the pattern of expression of a selected set of genes associated with immune function in CD4 and pTfh cells would provide unique insight into the quality of pre-existing memory pTfh cells in the clinical setting, especially in the context of vaccine responses, regardless of pre-existing titers of hemagglutination inhibition (HAI) due to previous vaccinations. Our results suggest that ex vivo Ag stimulation is a useful approach to dissect molecular mechanisms of H1N1 memory responses and to identify novel correlates of response.

Participants were recruited from Children’s Hospital Bambino Gesù. ART-treated HIV-1 vertically infected individuals and HIV age-matched controls were enrolled between September and November of 2012. Written informed consent was obtained from all subjects or parents/legal guardians before enrollment, and the ethical committee of the Bambino Gesù Children’s Hospital approved the study. Participants were immunized with a single dose of inactivated influenza vaccine trivalent types A and B (Split Virion) VAXIGRIP (Sanofi Pasteur). The strains for the 2012–2013 season include A/California/7/2009 H1N1, A/Victoria/361/2011 (H3N2), and B/Wisconsin/1/2010 (B). PBMCs, serum, and plasma were collected prevaccination (T0) and at 21 d postvaccination (T1). Clinical surveys were performed over the course of the vaccination trial to exclude natural infection.

Ab titers to H1N1, H3N2, and B influenza strains were evaluated separately using an HAI assay, which was performed as previously described (21). HAI titers are expressed as the reciprocal of the highest serum dilution at which hemagglutination was prevented.

PBMCs from T0 and T1 were thawed and polyclonally activated in vitro for 5 d with CpG (2.5 mg/ml), as described (13). Briefly, ELISPOT 96-well filtration plates (Millipore) were coated with purified H1N1, H3N2, and B influenza inactivated virus particles and subsequently loaded with 2 × 105 cells per well. Processed membranes were scanned with an Eli.Scan and counted with ELISPOT Analysis Software V5.1 (both from A.EL.VIS).

Plasma IL-21 titers were measured using the human IL-21 platinum ELISA kit (eBioscience, San Diego, CA), following the manufacturer’s instructions.

A single vial of PBMCs (T0) was thawed, counted with a Countess Automated Cell Counter (Life Technologies), and reconstituted in complete RPMI 1640 medium at a concentration of 5 × 106 PBMCs per milliliter. Cells were cultured in the presence or absence of H1N1 peptides (5 μg/ml) overnight (16 h). Following the culture period, cells were labeled with mAbs against CD3 (AmCyan), CD4 (PerCP Cy5.5), CD45RO (ECD), CCR7 (Alexa Fluor 700), and CXCR5 (Alexa Fluor 647) and a live/dead marker (ViViD; Molecular Probes). All Abs were titrated previously. A four-way sorting mode was used on a FACSAria II (BD Biosciences) to sort cell populations of interest. A total of 500 cells per subset was sorted directly into 1.1-ml tubes containing CellsDirect One-Step PCR buffer and pooled TaqMan Gene Expression Assays (5 μl of 2× CellsDirect Reaction mix, 0.5 μl of Superscript III + Taq Polymerase, 2.5 μl of 0.2× TaqMan primer pool, and 1 μl of resuspension buffer). After sorting, tubes containing cells were immediately centrifuged (3000 rpm for 3 min) and kept on ice. Samples were transferred to PCR tubes, and reverse transcription and target-specific preamplification were performed on a C1000 Thermal Cycler (Bio-Rad) with the following cycle conditions (50°C for 20 min, 95°C for 2 min, 95°C for 15 s, and 60°C for 4 min; the last two steps were repeated for 18 cycles). Resulting cDNA was diluted 1:1 with TE Buffer (10 mM Tris, 1 mM EDTA [pH 8]) and stored at −20°C until further analysis.

Previously amplified samples were loaded onto a 96.96 Dynamic Array IFC (Fluidigm), according to the manufacturer’s instructions. Briefly, the assay premix was prepared with 3 μl of 20× TaqMan Gene Expression Assay (Applied Biosystems) and 3 μl of 2× Assay Loading Reagent (Fluidigm). The sample premix was prepared with 3 μl of TaqMan Universal PCR Master Mix (2×; Applied Biosystems), 0.3 μl of 20× GE Sample Loading Reagent (Fluidigm), and 2.7 μl of diluted cDNA (67.5 cell equivalent). A total volume of 5 μl was loaded into appropriate sample and assay wells. Integrated fluidic circuits were primed in the Fluidigm Controller HX unit prior to loading onto the Biomark HD instrument for standard RT-PCR. The panel of TaqMan gene expression assays was qualified on human PBMCs and T lymphocytes as previously described (22). Cycle threshold (Ct) values derived from Biomark experiments were normalized to 500 cells in cases in which the sorted population did not reach 500 using the following equations: determine y using 67.5/500 = y/x, where x is the number of cells sorted and y is the cDNA cell equivalent loaded onto the integrated fluidic circuit, and normalized Ct = reported Ct − log2 (67.5/y). Expression threshold (Et) values or log2 expression was used for all analyses and were determined using Et = 40 − Ct.

Et values were analyzed using the SINGuLAR Analysis Toolset 3.0 (Fluidigm) package in R software (3.0.2 GUI 1.62). Expression plots of samples, expression plot overviews, and identification of outliers were performed using this software, and principle component analysis and ANOVA were used to identify differentially expressed genes (DEGs) between participant groups and between cell subsets of the same group. Fold change was determined using the mean Et for each group. Automated paired analysis between unstimulated and stimulated samples from the same individual was carried out using the Student t test. Correlation matrix analysis and multivariate linear-regression models were performed using MAST and Stats packages, also in R software. The Kolmogorov–Smirnov normality test was performed on single-gene data sets and on total 96-gene sets to select statistical tests to perform for correlation matrix analysis. The Pearson correlation coefficient was used if data were normally distributed, and the Spearman correlation was used if a nonparametric test was needed. Linear regression plots were generated using GraphPad software (Prism 6.0).

We measured the serologic response to influenza vaccination in 22 HIV perinatally infected youth (Table I) by HAI assay with each influenza strain in the vaccine in serum samples collected at T0 and T1. Higher baseline titers against H1N1 were observed in HIV-infected patients compared with noninfected age-matched controls, indicative of repeated vaccination during the preceding years (Fig. 1A). Baseline titers against the other strains (H3N2 and B) were not different between the HIV-infected and uninfected groups. Increases in Ab titer following vaccination were more robust to H1N1 and H3N2 (Fig. 1B), consistent with poor immunogenicity of the B strain (23). Serum titers to H1N1 increased ≥4-fold in 90% (9/10) of HIV uninfected individuals and in only 50% (11/22) of HIV-infected individuals. Participants who demonstrated good Ab responses, as measured by HAI titer, tended to have greater responses to the other vaccine Ags as well (Supplemental Fig. 1A). Vaccinees were characterized as Rs if they exhibited a serum H1N1 titer ≥1:40 and a 4-fold increase above baseline at T1; they were classified as nonresponders (NRs) if they did not meet these criteria. Response to H1N1 was confirmed by memory B cell ELISPOT assay performed at T1, which also showed a greater response in Rs compared with NRs (Supplemental Fig. 1B).

Table I.
Characteristics of study participants
HIVHIV+ RsHIV+ NRs
Age (y; mean [SEM]) 14.3 (3.4) 13.7 (2.4) 15.2 (2.2) 
Participants (n [female]) 10 (5) 11 (6) 11 (8) 
Lymphocytes/mm3 (SEM) 3063 (428) 2961 (363) 2494 (279) 
CD4+ T cells (%; mean [SEM]) 29.8 (6.3) 32.5 (6.1) 38.0 (4.9) 
HIV RNA < 50 copies/ml (nNA 10 11 
Prior influenza vaccinations between 2009 and 2012 (1, 2, or 3 yearly vaccinations) Unknown 5/4/2 3/6/3 
HIVHIV+ RsHIV+ NRs
Age (y; mean [SEM]) 14.3 (3.4) 13.7 (2.4) 15.2 (2.2) 
Participants (n [female]) 10 (5) 11 (6) 11 (8) 
Lymphocytes/mm3 (SEM) 3063 (428) 2961 (363) 2494 (279) 
CD4+ T cells (%; mean [SEM]) 29.8 (6.3) 32.5 (6.1) 38.0 (4.9) 
HIV RNA < 50 copies/ml (nNA 10 11 
Prior influenza vaccinations between 2009 and 2012 (1, 2, or 3 yearly vaccinations) Unknown 5/4/2 3/6/3 

NA, not applicable.

FIGURE 1.

Serological response to influenza vaccine strains. (A) Baseline serum titers from HIV-infected (n = 22) and noninfected (n = 10) individuals for each of the influenza strains contained in the 2012–2013 vaccine, as determined by HAI assay. Each points represents one person; horizontal lines show the mean. (B) Circos plot (49) showing fold change in serum titer, as determined by HAI assay for each participant and influenza strain. Ribbons originate from a node representing each participant and connect to individual vaccine strains (green, H1N1; blue, H3N2; red, B). Ribbon thickness represents the fold increase from T0 to T1. Study participants were clustered into three groups, HIV uninfected, HIV Rs, and HIV NRs, as indicated by black arcs outside of the Circos plot. *p < 0.05, Student t test.

FIGURE 1.

Serological response to influenza vaccine strains. (A) Baseline serum titers from HIV-infected (n = 22) and noninfected (n = 10) individuals for each of the influenza strains contained in the 2012–2013 vaccine, as determined by HAI assay. Each points represents one person; horizontal lines show the mean. (B) Circos plot (49) showing fold change in serum titer, as determined by HAI assay for each participant and influenza strain. Ribbons originate from a node representing each participant and connect to individual vaccine strains (green, H1N1; blue, H3N2; red, B). Ribbon thickness represents the fold increase from T0 to T1. Study participants were clustered into three groups, HIV uninfected, HIV Rs, and HIV NRs, as indicated by black arcs outside of the Circos plot. *p < 0.05, Student t test.

Close modal

Phenotypic analysis of CD4 T cells was performed by flow cytometry, following stimulation or not with H1N1 Ag, to determine subset frequencies in the participant groups (Fig. 2A). Comparable frequencies were observed for CD4 T cells (Table I) and maturation subsets, with the exception of the effector/terminally differentiated subset (CD45ROnegCCR7neg), which was significantly higher in NRs compared with Rs (p = 0.034, Fig. 2B), although it represented <2% of the CD4 population. Within the T central memory compartment (TCM; CD45RO+CCR7+), pTfh cells were characterized as CXCR5+ and exhibited similar frequencies in HIV+ Rs, HIV+ NRs, and healthy controls.

FIGURE 2.

Characterization of T0 CD4 T cell subset frequencies in study participants. (A) Representative dot plots of flow cytometry gating scheme for identifying T cell subsets and pTfh cells. Colored boxes designate populations of interest in sorting experiments. (B) Summary data for frequencies of CD4 T cell subsets from study participant groups. Frequencies shown are out of live CD3+CD4+ cells. *p <0.05, Student t test. HC, healthy control.

FIGURE 2.

Characterization of T0 CD4 T cell subset frequencies in study participants. (A) Representative dot plots of flow cytometry gating scheme for identifying T cell subsets and pTfh cells. Colored boxes designate populations of interest in sorting experiments. (B) Summary data for frequencies of CD4 T cell subsets from study participant groups. Frequencies shown are out of live CD3+CD4+ cells. *p <0.05, Student t test. HC, healthy control.

Close modal

Next, we examined the ability of pTfh cells from HIV+ study participants to respond to in vitro Ag stimulation. Paired analysis of CXCR5 expression on TCM cells showed a significant increase following in vitro stimulation with H1N1 Ag overnight (16 h) only in HIV+ Rs (Fig. 3). These data suggest that memory CD4 from HIV+ NRs might harbor functional defects that could impair vaccine responsiveness.

FIGURE 3.

Central memory cells from vaccine Rs induce CXCR5 expression in response to H1N1. Frequency of CD4 TCM cells (CD45RO+CCR7+) expressing CXCR5 with (H1N1) and without (−) overnight in vitro H1N1 Ag (5 μg/ml) stimulation in each study group. Lines connect measurements obtained from the same individual. *p < 0.05, paired t test.

FIGURE 3.

Central memory cells from vaccine Rs induce CXCR5 expression in response to H1N1. Frequency of CD4 TCM cells (CD45RO+CCR7+) expressing CXCR5 with (H1N1) and without (−) overnight in vitro H1N1 Ag (5 μg/ml) stimulation in each study group. Lines connect measurements obtained from the same individual. *p < 0.05, paired t test.

Close modal

To evaluate potential T cell–derived molecular biomarkers for predicting seasonal influenza vaccine responses and to dissect transcriptional mechanisms driving the upregulation of CXCR5 expression in H1N1-stimulated CD4 T cells, we performed targeted gene-expression analysis in purified CD4 T cells, pTfh cells, and non-pTfh cells (TCM CXCR5neg) (Supplemental Fig. 2A). The Fluidigm Biomark platform was used to analyze expression of a custom panel of genes involved in immune activation and inflammation, Tfh cell function, TCR signaling, cytokine and chemokine signaling, and antiviral and IFN-inducible factors (Supplemental Table I). Using this gene panel, cell subsets (PBMC, CD4 T, pTfh, and non-pTfh) distinctly segregated from one another by principle component analysis, although clustering was not evident among subsets from HIV+ and HIV individuals (Supplemental Fig. 2B).

For analysis of gene-expression levels in T0 samples, we directly compared R and NR groups using ANOVA and identified genes exhibiting differential expression between the two groups for each cell type. In the absence of any in vitro stimulation, more DEGs were identified in CD4 T cells (Rs versus NRs) compared with purified subsets (pTfh and non-pTfh cells), and few were overlapping between total CD4 T cells and subsets (Fig. 4A). This observation demonstrates that changes in gene expression from purified CD4 T cell subsets are not detectable in the analysis of total CD4 T cells. This concept was also revealed when evaluating DEGs in PBMCs compared with CD4 T cells and/or subsets; only two genes (BCL6 and IL6ST) overlapped between PBMCs and CD4 T cells, and they showed enrichment in different groups (Supplemental Fig. 3A). Differential gene expression in PBMCs between Rs and NRs was dominated by genes with higher expression in NRs, whereas analyses of CD4 T cells revealed that a higher number of DEGs were expressed in Rs, highlighting the divergent functions for different cell types (Fig. 4A). Few DEGs reached significance (p < 0.05) in pTfh and non-pTfh cells, and the top five DEGs (according to p value) were used for the next analysis to explore the relationship between baseline gene expression in CD4 TCM subsets and vaccine response.

FIGURE 4.

Differential gene-expression analysis in T0 CD4, pTfh, and non-pTfh cells between HIV+ flu vaccine Rs and NRs. (A) Bar graph shows the average fold change in gene expression between Rs and NRs for CD4 T cells, pTfh cells, and non-pTfh cells (bottom) using a fold change cutoff >1.5. Red bars correspond to gene expression that was higher in Rs compared with NRs, and green bars indicate higher gene expression in NRs compared with Rs. (B) Gene networks were generated using IPA software and combined DEGs from the three cell types in (A) as input. Colors in the network are defined as follows: red indicates upregulation in Rs versus NRs, green indicates upregulation in NRs versus Rs, and white indicates that the molecule was added from the Ingenuity Knowledge Base. Solid lines indicate direct interaction, and dashed lines indicate indirect interaction. (C) The top five upstream regulators were determined using IPA software and combined DEGs from the three cell types in (A) as input. The overlap p values show statistically significant overlap by the Fisher exact test between dataset genes and the genes that are known to be regulated by a particular molecule.

FIGURE 4.

Differential gene-expression analysis in T0 CD4, pTfh, and non-pTfh cells between HIV+ flu vaccine Rs and NRs. (A) Bar graph shows the average fold change in gene expression between Rs and NRs for CD4 T cells, pTfh cells, and non-pTfh cells (bottom) using a fold change cutoff >1.5. Red bars correspond to gene expression that was higher in Rs compared with NRs, and green bars indicate higher gene expression in NRs compared with Rs. (B) Gene networks were generated using IPA software and combined DEGs from the three cell types in (A) as input. Colors in the network are defined as follows: red indicates upregulation in Rs versus NRs, green indicates upregulation in NRs versus Rs, and white indicates that the molecule was added from the Ingenuity Knowledge Base. Solid lines indicate direct interaction, and dashed lines indicate indirect interaction. (C) The top five upstream regulators were determined using IPA software and combined DEGs from the three cell types in (A) as input. The overlap p values show statistically significant overlap by the Fisher exact test between dataset genes and the genes that are known to be regulated by a particular molecule.

Close modal

Network analysis was performed on enriched genes identified in CD4 T cells and subsets from Fig. 4A using Ingenuity Pathway Analysis (IPA) (Fig. 4B). IPA generated two networks for the enriched genes: one contained mostly genes enriched in NRs, and the other had mostly genes enriched in Rs. IL-12 complex, STAT5, and ERK1/2 were major nodes in the network displaying NR-enriched genes. Multiple markers of immune activation, such as ICOS, IL2RA, CD69, and IFNG, were enriched in NRs in the absence of stimulation. The genes enriched in Rs contained multiple transcription factors, such as RUNX3, STAT1, STAT4, BCL6, and inhibitor of DNA binding (ID) 2. ID2 and ID3 were shown to be involved in T lineage developmental progression (24). ID2 was strongly enriched in CD4 cells (34-fold) and pTfh cells (19-fold) from vaccine Rs, whereas ID3 was not differentially expressed in these samples. The top five upstream regulators of enriched genes were IL21, CD40LG, IL15, IL4, and TCR (Fig. 4C).

In the next set of analyses, we sought to evaluate the activation potential of immune cells from T0 samples. We compared gene expression from CD4 T, pTfh, and non-pTfh cells from Rs and NRs following in vitro stimulation with H1N1 to identify DEGs that may be directly related to vaccine response (Fig. 5A). As expected, the IL21 gene was expressed in pTfh cells and was not detectable in non-pTfh or CD4 T cells. Importantly, IL21 expression was higher in pTfh cells from Rs following in vitro stimulation, together with upregulation of other critical Tfh cell functional markers (CD40LG, ICOS, and IL6ST). Non-pTfh cells exclusively exhibited DEGs that were expressed more highly in NRs, suggesting an immune-inhibitory role for IFNG, DUSP4, RUNX3, and IL21R in this cell type. In CD4 T cells, expression of the chemokine receptor CXCR4 was strongly associated with NR status. Increased expression of HIV coreceptors on CD4 T cells is a concern in HIV-infected individuals, because they could enhance de novo infection of target cells. Similar analyses in PBMCs revealed three DEGs between Rs and NRs following H1N1 stimulation: IL21R, IRF4, and IL2 (Supplemental Fig. 3B). Of these genes, only IL21R was observed in DEG analyses following in vitro stimulation in non-pTfh cells. Network analysis of DEGs from H1N1-stimulated CD4 T, pTfh, and non-pTfh cells from the two vaccine response groups revealed two gene networks: the first related to TCR signaling with CD4, ZAP70, and NFAT complex as major nodes, and the second related to IL21 signaling (Fig. 5B). Increased CXCR5 gene expression in CD4 T cells following H1N1 stimulation supports the flow cytometry data showing an increase in CXCR5 expression in central memory T cells (Fig. 3). ID2 expression was maintained at high levels in CD4 and pTfh cells from Rs compared with NRs, suggesting that its expression is not related to activation status but rather is related to the differentiation or maturation state of the cell. Transcription regulators ID2 and ID3 were included in the top five upstream regulators for the DEGs identified, along with IL21, CD40LG, and TCR (Fig. 5C).

FIGURE 5.

Differential gene-expression analysis in H1N1-stimulated T0 CD4, pTfh, and non-pTfh cells between HIV+ flu vaccine Rs and NRs. (A) Bar graph show the average fold change in gene expression between Rs and NRs for CD4 T cells, pTfh cells, and non-pTfh cells using ANOVA p value < 0.05 and fold change >1.5 as cutoffs. Red bars correspond to gene expression that was higher in Rs compared with NRs; green bars indicate higher gene expression in NRs compared with Rs. (B) Gene networks were generated using IPA software and combined DEGs from the three cell types in (A) as input. Colors in the network are defined as follows: red indicates upregulation in Rs versus NRs, green indicates upregulation in NRs versus Rs, and white indicates that the molecule was added from the Ingenuity Knowledge Base. Solid lines indicate direct interaction and dashed lines indicate indirect interaction. (C) The top five upstream regulators were determined using IPA software and combined DEGs from the three cell types in (A) as input. The overlap p values show statistically significant overlap by the Fisher exact test between dataset genes and the genes that are known to be regulated by a particular molecule.

FIGURE 5.

Differential gene-expression analysis in H1N1-stimulated T0 CD4, pTfh, and non-pTfh cells between HIV+ flu vaccine Rs and NRs. (A) Bar graph show the average fold change in gene expression between Rs and NRs for CD4 T cells, pTfh cells, and non-pTfh cells using ANOVA p value < 0.05 and fold change >1.5 as cutoffs. Red bars correspond to gene expression that was higher in Rs compared with NRs; green bars indicate higher gene expression in NRs compared with Rs. (B) Gene networks were generated using IPA software and combined DEGs from the three cell types in (A) as input. Colors in the network are defined as follows: red indicates upregulation in Rs versus NRs, green indicates upregulation in NRs versus Rs, and white indicates that the molecule was added from the Ingenuity Knowledge Base. Solid lines indicate direct interaction and dashed lines indicate indirect interaction. (C) The top five upstream regulators were determined using IPA software and combined DEGs from the three cell types in (A) as input. The overlap p values show statistically significant overlap by the Fisher exact test between dataset genes and the genes that are known to be regulated by a particular molecule.

Close modal

To further dissect gene-expression changes following H1N1 stimulation in CD4 T cells and pTfh cells, we performed paired analyses using data from each participant to examine how gene expression changed on an individual level following H1N1 stimulation. This analysis revealed 12 common genes whose expression was significantly induced in CD4 T cells from Rs and NRs in response to H1N1 stimulation. These genes consisted of IFN-inducible and antiviral genes (OAS1, IFIT2, STAT1, MX1, FAS, IRF4, BST2) (Fig. 6A), immune activation genes (CD74, ICOS, IL2RA), and transcription factors (FOXP3, BCL6). Nineteen differentially induced genes (DIGs) were identified in CD4 T cells: 16 were upregulated exclusively in Rs, and 3 were upregulated exclusively in NRs. DIGs in Rs included chemokine signaling molecules (CCR2, CXCL10, CCR7) and classical activation markers (CD69, CD38). The DIGs observed only in NRs were IL2, STAT5A, and IL21R.

FIGURE 6.

Gene induction in response to H1N1 stimulation in T0 CD4 and pTfh cells in HIV+ Rs and NRs to influenza vaccine. Dual-column heat maps show average fold change in gene expression from H1N1-stimulated cells compared with unstimulated cells in CD4 T cells (A) and pTfh cells (B). Values were derived by calculating the Δ for each study participant, such that (Et_H1N1) − (Et_Unstim) = Δ, if the group of deltas was found to be significant (p < 0.05) for the group (R or NR), the average fold change was calculated and displayed in the heat map as a color based on the scale provided to the right of each heat map.

FIGURE 6.

Gene induction in response to H1N1 stimulation in T0 CD4 and pTfh cells in HIV+ Rs and NRs to influenza vaccine. Dual-column heat maps show average fold change in gene expression from H1N1-stimulated cells compared with unstimulated cells in CD4 T cells (A) and pTfh cells (B). Values were derived by calculating the Δ for each study participant, such that (Et_H1N1) − (Et_Unstim) = Δ, if the group of deltas was found to be significant (p < 0.05) for the group (R or NR), the average fold change was calculated and displayed in the heat map as a color based on the scale provided to the right of each heat map.

Close modal

Similar to analyses in CD4 T cells, DIG expression patterns in pTfh cells showed induction of IFN-inducible genes (IFIT2, IRF4, MX1, OAS1, STAT1) that were common in Rs and NRs (Fig. 6B). However, genes that exhibited significant changes only in Rs, such as ABCB1, DUSP4, BST2, and IL21, were mostly different from what was observed in our analysis of CD4 T cells. Induction of SAMHD1 was observed in CD4 and pTfh cells from Rs. IL21 is the signature cytokine for Tfh cell function required for high-affinity Ab production from B cells; thus, IL21 induction following Ag stimulation in pTfh cells from Rs supports the notion that pTfh cells can exhibit a function similar to lymph node Tfh cells. ABCB1 and DUSP4 were the only genes that showed reduced expression following H1N1 stimulation, suggesting regulatory roles for these genes. ABCB1 (MDR1) encodes P-glycoprotein, a membrane transporter enriched in pathogenic Th17 cells (25). DUSP4 is a phosphatase and negative regulator of the MAPK superfamily and was shown to contribute to dampened CD4 T cell memory responses in the elderly (26). FAS, CCR5, LAG3, and IL2 genes were induced in NRs in response to H1N1 stimulation. CCR5 is a chemokine receptor that is upregulated in response to inflammation and plays a role in homing to sites of inflammation; it is also a coreceptor for HIV entry. LAG3 is an immune checkpoint receptor that dampens T cell activation by competing with CD4 for MHC binding (27, 28). Finally, IL-2 is a crucial cytokine for immune cell development and function that is produced primarily by activated CD4+ T cells; however, IL-2 signaling was shown to reduce Ag-specific humoral immunity by directly suppressing Tfh cell formation (29). IL2 induction was observed in PBMCs, CD4 cells, and pTfh cells from NRs, pointing to a global increase in IL2 being associated with poor vaccine responses. Overall, these data show that in vitro H1N1-stimulated induction of gene expression in pTfh cells at baseline can produce a distinct molecular signature in vaccine Rs.

To identify the most important variables in our analysis of Ab responses to the influenza vaccine in HIV-infected children and adolescents, we performed pairwise Spearman correlations between the aforementioned observations (54 DEGs and 45 DIGs) and immunological correlates of response to the vaccine (i.e., H1N1 serum Ab titer at T1, fold change Ab titer increase [T1/T0], and memory B cell H1N1 ELISPOT data at T1). This analysis allowed us to reduce the number of observations from 99 to 33 DEGs and DIGs. The significantly correlated observations were subjected to cluster analysis to identify highly correlative gene-expression patterns in our data set (Fig. 7). A cluster of DEGs from H1N1-stimulated CD4 T cells revealed correlations among RORA, RUNX3, PIK3C2B, STAT3, STAT4, and ICOS following Ag stimulation in the HIV+ vaccinees. All of these genes were expressed more highly in Rs compared with NRs, suggesting a functional relationship between gene expression in CD4 T cells and vaccine response. A second cluster contained IFN-inducible genes expressed in pTfh cells which were significantly induced in both groups with H1N1 stimulation, suggesting coregulation of transcription of these genes, as would be anticipated. Interestingly, IL21 gene induction (DIG) and IL21R gene expression (DEG) correlated with postvaccination vaccine-response parameters (H1N1 titer and ELISPOT).

FIGURE 7.

Cluster analysis of DEGs and DIGs with serological and immunological correlates of Influenza vaccine response. Heat map showing pairwise Spearman correlations for a group of parameters including 33 DEGs and DIGs. Nine serological and immunological correlates were calculated against each other using R statistical software. Spearman correlations that resulted in p values > 0.05 were assigned an r value of 0 (white), whereas significant correlations were clustered and visualized as a heat map depicting r values on a scale of −1 (pink) to +1 (green). The parameters tested are color-coded as follows: blue type designates parameters from non-pTfh cells, green type designates parameters from CD4 T cells, red type designates parameters from pTfh cells, and black type designates serum titer and ELISPOT results for each vaccine Ag.

FIGURE 7.

Cluster analysis of DEGs and DIGs with serological and immunological correlates of Influenza vaccine response. Heat map showing pairwise Spearman correlations for a group of parameters including 33 DEGs and DIGs. Nine serological and immunological correlates were calculated against each other using R statistical software. Spearman correlations that resulted in p values > 0.05 were assigned an r value of 0 (white), whereas significant correlations were clustered and visualized as a heat map depicting r values on a scale of −1 (pink) to +1 (green). The parameters tested are color-coded as follows: blue type designates parameters from non-pTfh cells, green type designates parameters from CD4 T cells, red type designates parameters from pTfh cells, and black type designates serum titer and ELISPOT results for each vaccine Ag.

Close modal

We performed univariate linear regression analysis to further evaluate the relationship between IL21 signaling molecules at baseline and memory B cell responses after vaccination (Fig. 8). We observed a strong relationship between H1N1-induced IL21 expression and H1N1 ELISPOT T1 (r2 = +0.504, p = 0.0003) and a similar association with H1N1 titer T1 (r2 = +0.432, p = 0.0012) (Fig. 8A). Conversely, T0 IL21R expression in pTfh cells (in absence of stimulation) showed a significant negative linear relationship with the same parameters (Fig. 8B). Linear relationships were not observed for IL21 DIG or IL21R DEG parameters with the H1N1 titer fold change (T1/T0) (Fig. 8A, 8B). Given the strong association of IL21 and IL21R gene expression with immunological correlates of vaccine response, we measured plasma levels of IL21 in the study participants before and after vaccination. No differences were observed between Rs and NRs at either time point (T0 or T1), nor did we observe a significant change in plasma IL21 in response to the vaccine in any group (Fig. 8C), suggesting that gene expression of IL21 is more informative as a biomarker for response to flu vaccination than are plasma levels of IL21 cytokine.

FIGURE 8.

Gene expression from IL21 signaling pathway in T0 pTfh cells correlates with postvaccination serological and immunological responses to influenza vaccination. xy plots show linear regression of IL21R expression (without stimulation) (B) and Δ IL21 expression in pTfh cells (as determined by Δ= [Et_H1N1] − [Et_Unstim]) (A) with H1N1 spots T1 per million PBMC (left panels), H1N1 titer T1 (middle panels), and H1N1 titer T1/T0 (right panels). R2 and p values are shown. (C) IL-21 plasma levels before (T0) and after (T1) influenza vaccination in HIV+ Rs and NRs.

FIGURE 8.

Gene expression from IL21 signaling pathway in T0 pTfh cells correlates with postvaccination serological and immunological responses to influenza vaccination. xy plots show linear regression of IL21R expression (without stimulation) (B) and Δ IL21 expression in pTfh cells (as determined by Δ= [Et_H1N1] − [Et_Unstim]) (A) with H1N1 spots T1 per million PBMC (left panels), H1N1 titer T1 (middle panels), and H1N1 titer T1/T0 (right panels). R2 and p values are shown. (C) IL-21 plasma levels before (T0) and after (T1) influenza vaccination in HIV+ Rs and NRs.

Close modal

Because we were unable to identify a single observation (DEG or DIG) with a strong relationship to the fold change in H1N1 titer, we performed multivariate regression analyses to explore whether the combined effect of any of these transcriptional parameters could be governing vaccine responses. Using the results from the cluster analysis in Fig. 7, we reduced the number of observations for multivariate analysis from 33 to 18 by retaining a single observation from each cluster of correlated observations. The best-fit model following a bidirectional stepwise regression contained the pTfh cell IL21 DIG parameter, with a significant positive effect on the fold change in serum titer response to H1N1 (Table II). Other significant parameters in the model were baseline CXCL10 expression in CD4 T cells after H1N1 stimulation (DEG) and the pTfh cell STAT1 DIG with negative effects, whereas CD4 IL6ST DEG exhibited a positive relationship. Taken together, our findings support a model in which Tfh cell–related gene expression by memory CD4 T cells in the peripheral blood is an indicator of vaccine response, whereas Th1-related gene expression (IFN-γ–induced genes CXCL10 and STAT1) is an indicator of poor responses.

Table II.
Prevaccination biomarkers associated with H1N1 titer fold change (T1/T0)
BiomarkerEstimateSEPr (>|t|)
MX1 DIG (pTfh) 44.3948 23.6437 0.081419 
IL6ST DEG (CD4) 100.6927 26.95345 0.002215 
STAT1 DIG (pTfh) −73.1667 28.01986 0.020523 
IL21 DIG (pTfh) 29.34022 11.29875 0.02111 
ICOS H1N1 DEG (CD4) −39.9732 21.93624 0.089845 
CXCL10 H1N1 DEG (CD4) −6.67798 2.538514 0.019761 
KLRG1 DIG (CD4) −13.8312 8.429431 0.123101 
BiomarkerEstimateSEPr (>|t|)
MX1 DIG (pTfh) 44.3948 23.6437 0.081419 
IL6ST DEG (CD4) 100.6927 26.95345 0.002215 
STAT1 DIG (pTfh) −73.1667 28.01986 0.020523 
IL21 DIG (pTfh) 29.34022 11.29875 0.02111 
ICOS H1N1 DEG (CD4) −39.9732 21.93624 0.089845 
CXCL10 H1N1 DEG (CD4) −6.67798 2.538514 0.019761 
KLRG1 DIG (CD4) −13.8312 8.429431 0.123101 

Multivariable regression model at baseline with H1N1 titer fold change as the dependent variable including DEG and DIG variables. Multiple R2: 0.6191; adjusted R2: 0.4287; p value: 0.02878. Bold text refers to variables with a significant association with H1N1 titer fold change.

Tfh cells residing in lymphoid follicles play a critical role in B cell maturation and the formation of robust Ab responses (1), but the relationship between Tfh cells and circulating Tfh-like cells (pTfh cells) remains unclear. However, the data presented in this article with regard to pTfh cells are in line with two recent studies showing the value of circulating IL-21–producing CD4 T cells as biomarkers for vaccine-induced responses (4, 30). Our study is novel in that it demonstrates the importance of IL21 gene expression in response to Ag stimulation in memory CD4 T cells, even at T0 time points, rather than days to weeks after vaccination. We used the Fluidigm Biomark platform to interrogate gene expression of a curated panel of genes from several cell populations, including total CD4 T cells and pTfh cells (CXCR5+ CD4 TCM) isolated from HIV-infected virologically suppressed children and adolescents on ART prior to seasonal flu vaccination. This approach, although biased compared with whole-transcriptome analyses (e.g., RNA sequencing or microarray), allowed us to obtain a large amount of transcriptional information on highly relevant cell types (e.g., pTfh CD4 T cells) from relatively low cell numbers (1–2 million PBMCs) obtained from pediatric samples (31). Moreover, because this patient group was vaccinated on a yearly basis and had high pre-existing Ab titers, we used overnight stimulation with H1N1 Ag and compared gene expression in cells, with or without stimulation, to gain insight into the functional capabilities of CD4 T cells T0. Ex vivo stimulation with H1N1 led to highly correlative IL21 gene expression in pTfh cells from Rs, which was not detectable in total PBMCs or CD4 T cells. In contrast, NRs exhibited increased IL2 gene expression from pTfh cells, total PBMCs, and CD4 T cells. These data, coupled with the observation that pTfh cell frequencies are equivalent between Rs and NRs, further suggest that the quality of pTfh cells is important for Ab responses to flu vaccination.

IL-21 is a pleiotropic cytokine with effects on lymphoid and myeloid cells. Dysregulation of IL-21 signaling can lead to detrimental immunosuppression (cancer) or immune stimulation (autoimmunity), depending on the biological context (32). Based on the immunomodulatory effects of IL-21 on cytotoxic T cells, it was studied previously as an immunotherapeutic modality for the treatment of several cancers and SIV (model of HIV) (3335). CD4 T cells express IL-21R at low levels constitutively that increase upon TCR activation. Interestingly, in our study, higher expression of IL21R in the absence of Ag stimulation on pTfh cells, non-pTfh cells, and PBMCs was always associated with vaccine nonresponsiveness, pointing to a necessity for tightly regulated IL-21 signaling for productive vaccine responses.

Our study revealed important insights into the biology of pTfh cell and immune responses. ID2 was highly expressed in CD4 T cells (and pTfh cells) in vaccine Rs with and without in vitro stimulation. ID2 was shown to play roles in T cell differentiation and immune responses in murine models (24, 36, 37), and overexpression of ID2 in T cell lymphomas and other cancers (38, 39) suggests a role in the regulation of cell survival and proliferation. Further studies are required to explore the role of ID2 in Ab-driven responses in humans. In addition to the classic Th1 cytokine, IL2, IFNG was associated with poor responses in our cohort. IFNG expression was enriched in total CD4 T cells (6-fold) and pTfh cells (4-fold) from NRs; excessive IFN-γ production by T cells was associated with autoimmunity and is attributed to impaired quality of the T cell response, accumulation of Tfh cells, and overactive GC responses (40). The increased transcription of IL2 and STAT5A in CD4 T cells from vaccine NRs supports the hypothesis that IL-2 production by CD4 T cells after Ag exposure may antagonize pTfh cell function. In fact, it is known that the IL2 signaling pathway inhibits Tfh cell differentiation through STAT5-mediated expression of PRDM1 (Blimp-1) (41, 42).

Tfh cells are of intense interest in HIV, because they harbor latent and active HIV reservoirs during chronic HIV infection, even in patients under virological control that is achieved through natural immunity or ART (4346). pTfh cells have increased susceptibility to HIV infection in vitro (47). We did not observe differences in pTfh cell frequency or gene expression between HIV-infected and uninfected individuals; however, it was evident that gene induction in pTfh cells was variable within HIV-infected subjects following in vitro stimulation and led to differential responses to influenza vaccination. The question remains whether HIV infection of pTfh cells could play a role in this context and needs further investigation.

Overall, we noted that differential gene expression at baseline (in the absence of stimulation) was skewed toward enrichment of deleterious immune activation/inflammation genes in NRs. Participants classified as vaccine Rs were characterized by gene expression indicating low baseline immune activation, whereas PBMC and CD4 TCM subsets (pTfh and non-pTfh cells) from vaccine NRs displayed molecular signatures (CD69, IL2RA, IFNG, ICOS) consistent with a state of chronic immune activation, despite being under stable combination ART and virological control. The results for ICOS gene expression in CD4 and pTfh cells were in agreement with a previous report showing higher baseline levels of ICOS in elderly individuals not infected with HIV who had poor responses to influenza vaccination and demonstrated stunted ICOS induction at 7 d postvaccination (7).

Statistical data integration revealed significant and biologically meaningful correlations between gene expression from H1N1-stimulated pTfh cells T0 and H1N1-specific Ab production postvaccination. The definitions imposed on influenza vaccinees for R status, although accepted in the field, can be problematic when individuals have a protective titer at baseline. Higher baseline titers result in lower fold increases following vaccination (48). Thus, by integrating gene-expression results with the postvaccination Ab production, rather than fold change, we were able to identify features that may be predictive for influenza vaccine responses in HIV-infected individuals. The finding of IL21 expression as a highly correlative marker of Ab responses postvaccination, supported by recent findings by Schultz et al. (4) with regard to HIV-specific IL-21–producing CD4 T cells in the RV144 HIV vaccine trial, strongly implicate IL21 as a biomarker for Ab-driven vaccine responses. Additionally, identification of molecular targets prior to vaccination, as demonstrated in this study, opens up the possibility for therapeutic intervention to improve immune function. This approach has great potential to advise studies on immune responses to other vaccine Ags, particularly therapeutic vaccines in which vaccine efficacy involves stimulating a memory response (e.g., HIV, cancer).

We would like to acknowledge all patients and guardians who consented to participate in the study. We thank Rajendra Pahwa for helpful comments on this project and Melanie Weiss, Jennifer Faudella, and Quita Nimrod for administrative assistance. We also thank Mario Roederer and Pratip Chattopadhyay for useful discussions and suggestions during the preliminary phase of the study.

This work was supported by grants obtained by the Children's Hospital Bambino Gesù, Rome, Italy in 2015 and 2016 and by the National Institutes of Health Grants R01AI108472 (to S. Pahwa) and P30AI073961 to the Miami Center for AIDS Research at the University of Miami, Miller School of Medicine.

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

The online version of this article contains supplemental material.

Abbreviations used in this article:

ART

antiretroviral therapy

B

B/Wisconsin/1/2010

Ct

cycle threshold

DEG

differentially expressed gene

DIG

differentially induced gene

Et

expression threshold

GC

germinal center

HA

hemagglutinin

HIA

hemagglutination inhibition

H1N1

influenza A/California/7/2009

H3N2

A/Victoria/361/2011

ID

inhibitor of DNA binding

IPA

Ingenuity Pathway Analysis

NR

nonresponder

pTfh

peripheral Tfh

R

responder

TCM

T central memory compartment

Tfh

T follicular helper

T0

prevaccination

T1

21 d postvaccination.

1
Schaerli
P.
,
Willimann
K.
,
Lang
A. B.
,
Lipp
M.
,
Loetscher
P.
,
Moser
B.
.
2000
.
CXC chemokine receptor 5 expression defines follicular homing T cells with B cell helper function.
J. Exp. Med.
192
:
1553
1562
.
2
Bryant
V. L.
,
Ma
C. S.
,
Avery
D. T.
,
Li
Y.
,
Good
K. L.
,
Corcoran
L. M.
,
de Waal Malefyt
R.
,
Tangye
S. G.
.
2007
.
Cytokine-mediated regulation of human B cell differentiation into Ig-secreting cells: predominant role of IL-21 produced by CXCR5+ T follicular helper cells.
J. Immunol.
179
:
8180
8190
.
3
Shulman
Z.
,
Gitlin
A. D.
,
Weinstein
J. S.
,
Lainez
B.
,
Esplugues
E.
,
Flavell
R. A.
,
Craft
J. E.
,
Nussenzweig
M. C.
.
2014
.
Dynamic signaling by T follicular helper cells during germinal center B cell selection.
Science
345
:
1058
1062
.
4
Schultz
B. T.
,
Teigler
J. E.
,
Pissani
F.
,
Oster
A. F.
,
Kranias
G.
,
Alter
G.
,
Marovich
M.
,
Eller
M. A.
,
Dittmer
U.
,
Robb
M. L.
, et al
.
2016
.
Circulating HIV-specific interleukin-21(+)CD4(+) T cells represent peripheral Tfh cells with antigen-dependent helper functions.
Immunity
44
:
167
178
.
5
Pallikkuth
S.
,
Parmigiani
A.
,
Silva
S. Y.
,
George
V. K.
,
Fischl
M.
,
Pahwa
R.
,
Pahwa
S.
.
2012
.
Impaired peripheral blood T-follicular helper cell function in HIV-infected nonresponders to the 2009 H1N1/09 vaccine.
Blood
120
:
985
993
.
6
Pallikkuth
S.
,
Pilakka Kanthikeel
S.
,
Silva
S. Y.
,
Fischl
M.
,
Pahwa
R.
,
Pahwa
S.
.
2011
.
Upregulation of IL-21 receptor on B cells and IL-21 secretion distinguishes novel 2009 H1N1 vaccine responders from nonresponders among HIV-infected persons on combination antiretroviral therapy.
J. Immunol.
186
:
6173
6181
.
7
Herati
R. S.
,
Reuter
M. A.
,
Dolfi
D. V.
,
Mansfield
K. D.
,
Aung
H.
,
Badwan
O. Z.
,
Kurupati
R. K.
,
Kannan
S.
,
Ertl
H.
,
Schmader
K. E.
, et al
.
2014
.
Circulating CXCR5+PD-1+ response predicts influenza vaccine antibody responses in young adults but not elderly adults.
J. Immunol.
193
:
3528
3537
.
8
George
V. K.
,
Pallikkuth
S.
,
Parmigiani
A.
,
Alcaide
M.
,
Fischl
M.
,
Arheart
K. L.
,
Pahwa
S.
.
2015
.
HIV infection worsens age-associated defects in antibody responses to influenza vaccine.
J. Infect. Dis.
211
:
1959
1968
.
9
Pallikkuth
S.
,
Kanthikeel
S. P.
,
Silva
S. Y.
,
Fischl
M.
,
Pahwa
R.
,
Pahwa
S.
.
2011
.
Innate immune defects correlate with failure of antibody responses to H1N1/09 vaccine in HIV-infected patients.
J. Allergy Clin. Immunol.
128
:
1279
1285
.
10
Parmigiani
A.
,
Alcaide
M. L.
,
Freguja
R.
,
Pallikkuth
S.
,
Frasca
D.
,
Fischl
M. A.
,
Pahwa
S.
.
2013
.
Impaired antibody response to influenza vaccine in HIV-infected and uninfected aging women is associated with immune activation and inflammation.
PLoS One
8
:
e79816
.
11
Sasaki
S.
,
He
X. S.
,
Holmes
T. H.
,
Dekker
C. L.
,
Kemble
G. W.
,
Arvin
A. M.
,
Greenberg
H. B.
.
2008
.
Influence of prior influenza vaccination on antibody and B-cell responses.
PLoS One
3
:
e2975
.
12
Frasca
D.
,
Diaz
A.
,
Romero
M.
,
Blomberg
B. B.
.
2016
.
The generation of memory B cells is maintained, but the antibody response is not, in the elderly after repeated influenza immunizations.
Vaccine
34
:
2834
2840
.
13
Rinaldi
S.
,
Zangari
P.
,
Cotugno
N.
,
Manno
E. C.
,
Brolatti
N.
,
Castrucci
M. R.
,
Donatelli
I.
,
Rossi
P.
,
Palma
P.
,
Cagigi
A.
.
2014
.
Antibody but not memory B-cell responses are tuned-down in vertically HIV-1 infected children and young individuals being vaccinated yearly against influenza.
Vaccine
32
:
657
663
.
14
Nakaya
H. I.
,
Wrammert
J.
,
Lee
E. K.
,
Racioppi
L.
,
Marie-Kunze
S.
,
Haining
W. N.
,
Means
A. R.
,
Kasturi
S. P.
,
Khan
N.
,
Li
G. M.
, et al
.
2011
.
Systems biology of vaccination for seasonal influenza in humans.
Nat. Immunol.
12
:
786
795
.
15
Tan
Y.
,
Tamayo
P.
,
Nakaya
H.
,
Pulendran
B.
,
Mesirov
J. P.
,
Haining
W. N.
.
2014
.
Gene signatures related to B-cell proliferation predict influenza vaccine-induced antibody response.
Eur. J. Immunol.
44
:
285
295
.
16
Nakaya
H. I.
,
Hagan
T.
,
Duraisingham
S. S.
,
Lee
E. K.
,
Kwissa
M.
,
Rouphael
N.
,
Frasca
D.
,
Gersten
M.
,
Mehta
A. K.
,
Gaujoux
R.
, et al
.
2015
.
Systems analysis of immunity to influenza vaccination across multiple years and in diverse populations reveals shared molecular signatures.
Immunity
43
:
1186
1198
.
17
Memoli
M. J.
,
Shaw
P. A.
,
Han
A.
,
Czajkowski
L.
,
Reed
S.
,
Athota
R.
,
Bristol
T.
,
Fargis
S.
,
Risos
K.
,
Powers
J. H.
, et al
.
2016
.
Evaluation of antihemagglutinin and antineuraminidase antibodies as correlates of protection in an influenza A/H1N1 virus healthy human challenge model.
MBio
7
:
e00417
e00516
.
18
Cotugno
N.
,
Douagi
I.
,
Rossi
P.
,
Palma
P.
.
2012
.
Suboptimal immune reconstitution in vertically HIV infected children: a view on how HIV replication and timing of HAART initiation can impact on T and B-cell compartment.
Clin. Dev. Immunol.
2012
:
805151
.
19
Cagigi
A.
,
Cotugno
N.
,
Giaquinto
C.
,
Nicolosi
L.
,
Bernardi
S.
,
Rossi
P.
,
Douagi
I.
,
Palma
P.
.
2012
.
Immune reconstitution and vaccination outcome in HIV-1 infected children: present knowledge and future directions.
Hum. Vaccin. Immunother.
8
:
1784
1794
.
20
Cagigi
A.
,
Rinaldi
S.
,
Di Martino
A.
,
Manno
E. C.
,
Zangari
P.
,
Aquilani
A.
,
Cotugno
N.
,
Nicolosi
L.
,
Villani
A.
,
Bernardi
S.
, et al
.
2014
.
Premature immune senescence during HIV-1 vertical infection relates with response to influenza vaccination.
J. Allergy Clin. Immunol.
133
:
592
594
.
21
Cagigi
A.
,
Pensieroso
S.
,
Ruffin
N.
,
Sammicheli
S.
,
Thorstensson
R.
,
Pan-Hammarström
Q.
,
Hejdeman
B.
,
Nilsson
A.
,
Chiodi
F.
.
2013
.
Relation of activation-induced deaminase (AID) expression with antibody response to A(H1N1)pdm09 vaccination in HIV-1 infected patients.
Vaccine
31
:
2231
2237
.
22
Dominguez
M. H.
,
Chattopadhyay
P. K.
,
Ma
S.
,
Lamoreaux
L.
,
McDavid
A.
,
Finak
G.
,
Gottardo
R.
,
Koup
R. A.
,
Roederer
M.
.
2013
.
Highly multiplexed quantitation of gene expression on single cells.
J. Immunol. Methods
391
:
133
145
.
23
DiazGranados
C. A.
,
Denis
M.
,
Plotkin
S.
.
2012
.
Seasonal influenza vaccine efficacy and its determinants in children and non-elderly adults: a systematic review with meta-analyses of controlled trials.
Vaccine
31
:
49
57
.
24
Miyazaki
M.
,
Miyazaki
K.
,
Chen
S.
,
Itoi
M.
,
Miller
M.
,
Lu
L. F.
,
Varki
N.
,
Chang
A. N.
,
Broide
D. H.
,
Murre
C.
.
2014
.
Id2 and Id3 maintain the regulatory T cell pool to suppress inflammatory disease.
Nat. Immunol.
15
:
767
776
.
25
Ramesh
R.
,
Kozhaya
L.
,
McKevitt
K.
,
Djuretic
I. M.
,
Carlson
T. J.
,
Quintero
M. A.
,
McCauley
J. L.
,
Abreu
M. T.
,
Unutmaz
D.
,
Sundrud
M. S.
.
2014
.
Pro-inflammatory human Th17 cells selectively express P-glycoprotein and are refractory to glucocorticoids.
J. Exp. Med.
211
:
89
104
.
26
Yu
M.
,
Li
G.
,
Lee
W. W.
,
Yuan
M.
,
Cui
D.
,
Weyand
C. M.
,
Goronzy
J. J.
.
2012
.
Signal inhibition by the dual-specific phosphatase 4 impairs T cell-dependent B-cell responses with age.
Proc. Natl. Acad. Sci. USA
109
:
E879
E888
.
27
Workman
C. J.
,
Cauley
L. S.
,
Kim
I. J.
,
Blackman
M. A.
,
Woodland
D. L.
,
Vignali
D. A.
.
2004
.
Lymphocyte activation gene-3 (CD223) regulates the size of the expanding T cell population following antigen activation in vivo.
J. Immunol.
172
:
5450
5455
.
28
Huard
B.
,
Prigent
P.
,
Pagès
F.
,
Bruniquel
D.
,
Triebel
F.
.
1996
.
T cell major histocompatibility complex class II molecules down-regulate CD4+ T cell clone responses following LAG-3 binding.
Eur. J. Immunol.
26
:
1180
1186
.
29
Ballesteros-Tato
A.
,
León
B.
,
Graf
B. A.
,
Moquin
A.
,
Adams
P. S.
,
Lund
F. E.
,
Randall
T. D.
.
2012
.
Interleukin-2 inhibits germinal center formation by limiting T follicular helper cell differentiation.
Immunity
36
:
847
856
.
30
Spensieri
F.
,
Siena
E.
,
Borgogni
E.
,
Zedda
L.
,
Cantisani
R.
,
Chiappini
N.
,
Schiavetti
F.
,
Rosa
D.
,
Castellino
F.
,
Montomoli
E.
, et al
.
2016
.
Early rise of blood T follicular helper cell subsets and baseline immunity as predictors of persisting late functional antibody responses to vaccination in humans.
PLoS One
11
:
e0157066
.
31
Cotugno
N.
,
De Armas
L.
,
Pallikkuth
S.
,
Rossi
P.
,
Palma
P.
,
Pahwa
S.
.
2015
.
Paediatric HIV infection in the 'omics era: defining transcriptional signatures of viral control and vaccine responses
.
J. Virus Erad.
1
:
153
158
.
32
Leonard
W. J.
,
Wan
C. K.
.
2016
.
IL-21 signaling in immunity.
F1000 Res.
5
:
224
.
33
White
L.
,
Krishnan
S.
,
Strbo
N.
,
Liu
H.
,
Kolber
M. A.
,
Lichtenheld
M. G.
,
Pahwa
R. N.
,
Pahwa
S.
.
2007
.
Differential effects of IL-21 and IL-15 on perforin expression, lysosomal degranulation, and proliferation in CD8 T cells of patients with human immunodeficiency virus-1 (HIV).
Blood
109
:
3873
3880
.
34
Strbo
N.
,
de Armas
L.
,
Liu
H.
,
Kolber
M. A.
,
Lichtenheld
M.
,
Pahwa
S.
.
2008
.
IL-21 augments natural killer effector functions in chronically HIV-infected individuals.
AIDS
22
:
1551
1560
.
35
Pallikkuth
S.
,
Micci
L.
,
Ende
Z. S.
,
Iriele
R. I.
,
Cervasi
B.
,
Lawson
B.
,
McGary
C. S.
,
Rogers
K. A.
,
Else
J. G.
,
Silvestri
G.
, et al
.
2013
.
Maintenance of intestinal Th17 cells and reduced microbial translocation in SIV-infected rhesus macaques treated with interleukin (IL)-21.
PLoS Pathog.
9
:
e1003471
.
36
Choi
Y. S.
,
Gullicksrud
J. A.
,
Xing
S.
,
Zeng
Z.
,
Shan
Q.
,
Li
F.
,
Love
P. E.
,
Peng
W.
,
Xue
H. H.
,
Crotty
S.
.
2015
.
LEF-1 and TCF-1 orchestrate T(FH) differentiation by regulating differentiation circuits upstream of the transcriptional repressor Bcl6.
Nat. Immunol.
16
:
980
990
.
37
Lin
Y. Y.
,
Jones-Mason
M. E.
,
Inoue
M.
,
Lasorella
A.
,
Iavarone
A.
,
Li
Q. J.
,
Shinohara
M. L.
,
Zhuang
Y.
.
2012
.
Transcriptional regulator Id2 is required for the CD4 T cell immune response in the development of experimental autoimmune encephalomyelitis.
J. Immunol.
189
:
1400
1405
.
38
Lasorella
A.
,
Boldrini
R.
,
Dominici
C.
,
Donfrancesco
A.
,
Yokota
Y.
,
Inserra
A.
,
Iavarone
A.
.
2002
.
Id2 is critical for cellular proliferation and is the oncogenic effector of N-myc in human neuroblastoma.
Cancer Res.
62
:
301
306
.
39
Yu
S.
,
Zhou
X.
,
Steinke
F. C.
,
Liu
C.
,
Chen
S. C.
,
Zagorodna
O.
,
Jing
X.
,
Yokota
Y.
,
Meyerholz
D. K.
,
Mullighan
C. G.
, et al
.
2012
.
The TCF-1 and LEF-1 transcription factors have cooperative and opposing roles in T cell development and malignancy. [Published erratum appears in 2014 Immunity 40: 166.]
Immunity
37
:
813
826
.
40
Lee
S. K.
,
Silva
D. G.
,
Martin
J. L.
,
Pratama
A.
,
Hu
X.
,
Chang
P. P.
,
Walters
G.
,
Vinuesa
C. G.
.
2012
.
Interferon-γ excess leads to pathogenic accumulation of follicular helper T cells and germinal centers.
Immunity
37
:
880
892
.
41
Nurieva
R. I.
,
Podd
A.
,
Chen
Y.
,
Alekseev
A. M.
,
Yu
M.
,
Qi
X.
,
Huang
H.
,
Wen
R.
,
Wang
J.
,
Li
H. S.
, et al
.
2012
.
STAT5 protein negatively regulates T follicular helper (Tfh) cell generation and function.
J. Biol. Chem.
287
:
11234
11239
.
42
Johnston
R. J.
,
Choi
Y. S.
,
Diamond
J. A.
,
Yang
J. A.
,
Crotty
S.
.
2012
.
STAT5 is a potent negative regulator of TFH cell differentiation.
J. Exp. Med.
209
:
243
250
.
43
Connick
E.
,
Folkvord
J. M.
,
Lind
K. T.
,
Rakasz
E. G.
,
Miles
B.
,
Wilson
N. A.
,
Santiago
M. L.
,
Schmitt
K.
,
Stephens
E. B.
,
Kim
H. O.
, et al
.
2014
.
Compartmentalization of simian immunodeficiency virus replication within secondary lymphoid tissues of rhesus macaques is linked to disease stage and inversely related to localization of virus-specific CTL.
J. Immunol.
193
:
5613
5625
.
44
Connick
E.
,
Mattila
T.
,
Folkvord
J. M.
,
Schlichtemeier
R.
,
Meditz
A. L.
,
Ray
M. G.
,
McCarter
M. D.
,
Mawhinney
S.
,
Hage
A.
,
White
C.
,
Skinner
P. J.
.
2007
.
CTL fail to accumulate at sites of HIV-1 replication in lymphoid tissue.
J. Immunol.
178
:
6975
6983
.
45
Perreau
M.
,
Savoye
A. L.
,
De Crignis
E.
,
Corpataux
J. M.
,
Cubas
R.
,
Haddad
E. K.
,
De Leval
L.
,
Graziosi
C.
,
Pantaleo
G.
.
2013
.
Follicular helper T cells serve as the major CD4 T cell compartment for HIV-1 infection, replication, and production.
J. Exp. Med.
210
:
143
156
.
46
Banga
R.
,
Procopio
F. A.
,
Noto
A.
,
Pollakis
G.
,
Cavassini
M.
,
Ohmiti
K.
,
Corpataux
J. M.
,
de Leval
L.
,
Pantaleo
G.
,
Perreau
M.
.
2016
.
PD-1(+) and follicular helper T cells are responsible for persistent HIV-1 transcription in treated aviremic individuals.
Nat. Med.
22
:
754
761
.
47
Pallikkuth
S.
,
Sharkey
M.
,
Babic
D. Z.
,
Gupta
S.
,
Stone
G. W.
,
Fischl
M. A.
,
Stevenson
M.
,
Pahwa
S.
.
2015
.
Peripheral T follicular helper cells are the major HIV reservoir within central memory CD4 T cells in peripheral blood from chronically HIV-infected individuals on combination antiretroviral therapy.
J. Virol.
90
:
2718
2728
.
48
Cagigi
A.
,
Cotugno
N.
,
Rinaldi
S.
,
Santilli
V.
,
Rossi
P.
,
Palma
P.
.
2016
.
Downfall of the current antibody correlates of influenza vaccine response in yearly vaccinated subjects: toward qualitative rather than quantitative assays.
Pediatr. Allergy Immunol.
27
:
22
27
.
49
Krzywinski
M.
,
Schein
J.
,
Birol
I.
,
Connors
J.
,
Gascoyne
R.
,
Horsman
D.
,
Jones
S. J.
,
Marra
M. A.
.
2009
.
Circos: an information aesthetic for comparative genomics.
Genome Res.
19
:
1639
1645
.

The authors have no financial conflicts of interest.

Supplementary data