Preferential Infection of α4β7+ Memory CD4+ T Cells During Early Acute Human Immunodeficiency Virus Type 1 Infection.

BACKGROUND: Establishment of persistent human immunodeficiency virus type 1 (HIV-1) reservoirs occurs early in infection, and biomarkers of infected CD4+ T cells during acute infection are poorly defined. CD4+ T cells expressing the gut homing integrin complex α4β7 are associated with HIV-1 acquisition, and are rapidly depleted from the periphery and gastrointestinal mucosa during acute HIV-1 infection.
METHODS: Integrated HIV-1 DNA was quantified in peripheral blood mononuclear cells obtained from acutely (Fiebig I-III) and chronically infected individuals by sorting memory CD4+ T-cell subsets lacking or expressing high levels of integrin β7 (β7negative and β7high, respectively). HIV-1 DNA was also assessed after 8 months of combination antiretroviral therapy (cART) initiated in Fiebig II/III individuals. Activation marker and chemokine receptor expression was determined for β7-defined subsets at acute infection and in uninfected controls.
RESULTS: In Fiebig I, memory CD4+ T cells harboring integrated HIV-1 DNA were rare in both β7high and β7negative subsets, with no significant difference in HIV-1 DNA copies. In Fiebig stages II/III and in chronically infected individuals, β7high cells were enriched in integrated and total HIV-1 DNA compared to β7negative cells. During suppressive cART, integrated HIV-1 DNA copies decreased in both β7negative and β7high subsets, which did not differ in DNA copies. In Fiebig II/III, integrated HIV-1 DNA in β7high cells was correlated with their activation.
CONCLUSIONS: β7high memory CD4+ T cells are preferential targets during early HIV-1 infection, which may be due to the increased activation of these cells.

CD4+ T cells are the main targets of human immunodeficiency virus type 1 (HIV-1) infection, and formation of early reservoirs of infected cells that persist despite years of suppressive therapy represents a major barrier to HIV-1 cure. CD4+ T cells vary in susceptibility to HIV-1 infection [1], and many cellular characteristics including activation, memory status, and homing potential contribute to permissiveness [2-8]. Efforts to better define initial viral targets will improve our understanding of the establishment of HIV-1 reservoirs, with implications for both prevention and cure strategies.

We previously showed that expression of the gut homing integrin complex α4β7 on CD4+ T cells was associated with HIV-1 acquisition and disease progression, and that these cells were rapidly depleted from the gastrointestinal mucosa during acute HIV-1 infection [9]. These data build on previous studies in nonhuman primates and in vitro systems suggesting increased susceptibility of α4β7+ cells to HIV-1 infection [10-12], which may contribute to the rapid gut CD4+ T-cell depletion observed during primary HIV-1 and simian immunodeficiency virus (SIV) infection. Preferential SIV infection of α4β7+ CD4+ T cells was observed within 10 days of infection in nonhuman primates, though not maintained after 2 months in untreated conditions [11]. Whether α4β7+ cells are preferentially infected during acute HIV-1 infection in humans and, if so, whether this preferential infection is maintained remains unclear. To better understand the role of α4β7 in establishing HIV-1 reservoirs, we measured HIV-1 DNA levels in α4β7+ memory CD4+ T cells in very early acute HIV-1 infection in the RV254 cohort in Thailand.

METHODS

Clinical Specimens

Eighteen participants from an acute HIV-1 infection cohort in Bangkok, Thailand (RV254/SEARCH010, ClinicalTrials.gov identifier NCT00796146 (Table 1) [13] were included in this study. Selected individuals were males in Fiebig stages I, II, and III (n = 6 each) [14]. Ten of the Fiebig II/III individuals with available specimens were also analyzed following 8 months of suppressive combination antiretroviral therapy (cART). One additional RV254 male participant in Fiebig stage III was analyzed by single-cell reverse-transcription quantitative polymerase chain reaction (RT-qPCR). Six chronically infected untreated age-matched Thai participants (RV304/SEARCH013, ClinicalTrials.gov identifier NCT01397669) were also included.

Table 1.

Demographic and Clinical Characteristics of Uninfected Study Participants and Participants With Human Immunodeficiency Virus Type 1 Infection

Characteristic	Fiebig I	Fiebig II/III	Chronic	Uninfected
No. of participants	6	12	6	6
Days since HIV exposure,a median (IQR)	15 (11–17)	12 (9–24)	ND	NA
Age, y, mean (SD)	26 (5)	28 (7)	35 (10)	33 (7)
Male sex, %	100	100	100	33
HIV-1 subtype	CRF01_AE	CRF01_AEb	ND	NA
Viral load, copies/mL, median (IQR), × 103	10.5 (4–135)	490 (266–1997)	147.5 (17–594)	NA
Post-cART analysis (8 mo of cART)	ND	Yes (n = 10)	ND	NA
CD4 count, cells/μL, median (IQR)	570 (489–900)	338 (266–382)	314 (170–353)	ND
CD8 count, cells/μL, median (IQR)	505 (356–767)	519 (249–672)	1782 (592–2025)	ND

All participants were of Thai origin. HIV-1–infected participants are grouped by infection stage at the time of HIV-1 diagnosis as follows: Fiebig stage I, Fiebig stage II/III, and chronic infection.

Abbreviations: cART, combination antiretroviral therapy; HIV, human immunodeficiency virus; IQR, interquartile range; NA, not applicable; ND, not determined; SD, standard deviation.

aTime since HIV-1 exposure was self-reported by participants and is therefore approximate.

bOne participant in Fiebig II/III was infected with subtype B.

Cell Surface Phenotyping and Sorting

Cryopreserved peripheral blood mononuclear cells (PBMCs) were analyzed by flow cytometry (Supplementary Methods). Cryopreservation did not substantially alter β7 surface expression (Supplementary Figure 1). CD4+ T cells were sorted on a BD FACS Aria II SORP (BD Biosciences) into CD45RA-defined and β7-defined populations (~104–106 cells each population) and frozen as dry cell pellets.

Quantification of Total and Integrated HIV-1 DNA

HIV-1 total and integrated DNA were quantified by nested real-time PCR (LTR and alu-LTR for total and integrated assays, respectively) as described previously [15] and in the Supplementary Methods.

Quantification of Cell-associated HIV-1 RNA

CD4+ T-cell viral gene expression was measured by qPCR as previously described [16, 17], using CRF01_AE HIV-1–specific assays for spliced and unspliced RNA (Supplementary Methods). Cells were sorted by flow cytometry either at 1 cell per well or in limiting dilution to estimate RNA-positive cell frequency.

Statistical Analysis

Statistical analysis was performed using GraphPad Prism software. Analysis of unpaired and paired measurements was performed by Mann-Whitney U test and Wilcoxon signed-rank test, respectively. Correlations were assessed by Spearman rank correlation test. P ≤ .05 was considered statistically significant.

RESULTS

β7high CD4+ T-cell Dynamics Across HIV-1 Infection Stages

To assess the dynamics of circulating α4β7+ CD4+ T cells during HIV-1 infection, we measured the expression of integrin β7 on CD45RA-negative (memory) CD4+ T cells in PBMCs in viremic individuals in Fiebig I (n = 6, pre–peak viremia), Fiebig II/III (n = 12, peak or shortly after peak viremia), and chronic infection (n = 6), as well as uninfected controls (n = 6) (Table 1; Figure 1A) [14]. High surface expression of β7 on memory CD4+ T cells (β7high) can be a surrogate for coexpression of α4 [11, 18–20]; therefore, we interpret β7high cells as heterodimeric α4β7 receptor positive. The median frequency of β7high memory CD4+ T cells was 12% (range, 7%–16%) in uninfected controls; 9% (range, 8%–17%) at Fiebig I, 9% (range, 6%–11%) at Fiebig II/III, and 10% (range, 6%–12%) in chronic infection (Figure 1B). The frequency did not differ between infection stages but was significantly lower at Fiebig II/III compared to uninfected controls (P = .009). Circulating β7high cells correlated inversely with plasma viremia across all infection stages (ρ = −0.52; P = .01; Figure 1C).

Figure 1.

Frequency of β7high CD4+ T cells during acute and chronic human immunodeficiency virus type 1 (HIV-1) infection. A, Flow cytometry gating strategy used to quantify and sort peripheral blood CD4+ T-cell subsets by integrin β7 expression. Sequential gating from left to right was used to isolate the 3 populations indicated in the far right plot: CD45RA− β7high, CD45RA− β7negative, and CD45RA+β7int. B, Frequency of peripheral blood β7high CD4+ T cells in uninfected, acute (Fiebig I, II/III), and chronic HIV-1–infected individuals. Median values, 95% confidence intervals, and the nonparametric Mann-Whitney test significant P value are shown (P < .05). C, β7high CD4+ T-cell frequency is plotted vs concurrent viral load. D, Longitudinal analysis of β7high CD4+ T-cell frequency in acute infection (Fiebig II/III) and 8 months after suppressive combination antiretroviral therapy (cART) initiation. Wilcoxon signed-rank test P value is shown; black bars depict median values. The colors in Figure 1C (and throughout the manuscript) reflect groups shown in Figure 1B.

To assess the impact of viral suppression by cART on β7high memory CD4+ T cells, β7high cells were measured longitudinally following 8 months of cART initiated at Fiebig II/III. Posttreatment specimens were only available for 10 of 12 individuals. Circulating β7high cell frequency increased modestly after therapy, from a median of 8% at acute infection to 10% (P = .006; Figure 1D). These data further support an inverse relationship between plasma viremia and β7high memory CD4+ T-cell frequency.

HIV-1 DNA Enrichment in β7high Cells

To determine whether β7high cells are preferentially targeted by HIV-1 during early acute infection, we measured integrated HIV-1 DNA in sorted β7high memory CD4+ T cells and their β7negative counterparts in peripheral blood (Figure 1A). During Fiebig I, integrated HIV-1 DNA burden was low in β7high cells (median, 87 copies per million [CPM] cells) and was similar between β7high and β7negative cells (Figure 2A). At Fiebig II/III, integrated HIV-1 DNA burden was considerably greater, with a median value of 760 CPM β7high cells. β7negative cells harbored less DNA; the median reduction relative to β7high cells was approximately 4-fold (P = .003). In chronic infection, β7high cells contained approximately 2500 integrated HIV-1 DNA CPM cells, which was approximately 2-fold greater than β7negative cells (P = .03). Comparing the ratio of integrated HIV-1 DNA in β7high and β7negative cells at each infection stage, the enrichment in β7high cells was greatest in Fiebig II/III (ie around peak viremia) (Figure 2B). Plasma viral load and absolute CD4 count did not correlate with this ratio.

Figure 2.

Human immunodeficiency virus type 1 (HIV-1) integrated DNA enrichment in integrin β7high CD4+ T cells in acute and chronic infection. A, Peripheral blood mononuclear cell CD4+ T-cell–associated integrated HIV-1 DNA in fluorescence-activated cell–sorted β7high and β7negative subsets in Fiebig I (left), Fiebig II/III (middle), and chronic (right) infection. Significant Wilcoxon test P values are shown (P < .05). B, Ratio of integrated HIV-1 DNA in β7high vs β7negative cells is plotted by infection stage. In 1 (Fiebig I) case, both quantitative polymerase chain reaction (qPCR) replicate values were below the limit of detection and the individual was excluded from analysis. Median values, 95% confidence intervals, and significant nonparametric Mann-Whitney test P values are shown. C, Spliced (env) viral RNA-positive cell frequency was estimated by limiting dilution cell sorting and reverse-transcription qPCR for β7high and β7negative CD4+ T cells from Fiebig II/III-infected individuals (n = 7) for whom values were within the range of assay detection. D, Longitudinal integrated HIV-1 DNA in β7-defined CD4+ T-cell subsets before and after 8 months of suppressive combination antiretroviral therapy (cART). Fiebig II/III and post-cART values are shown for β7high (left) and β7negative (middle) CD4+ T cells; fold-change before and after cART within β7high and β7negative CD4+ T cells is shown at right. E, β7high and β7negative CD4+ T-cell–associated HIV-1 DNA values in cART-suppressed individuals. F, Correlation between CD4+ T-cell–associated integrated HIV-1 DNA values during Fiebig II/III acute infection and 8 months post-cART in β7high (left) and β7negative (right) subsets. Negative qPCR measurements were assigned the cellular input limit of detection (LOD) value and are represented by open symbols (A, D, and E); calculations using an LOD value combined with a positive signal are also represented by open symbols (B and F).

To validate results from the integrated HIV-1 DNA assay, total HIV-1 DNA was assessed in the same cells from the Fiebig II/III donors. Additional controls included measurement in naive (CD45RA+) cells expressing intermediate surface β7 (β7int; Figure 1A) [9]. Similar to integrated HIV-1 DNA, total HIV-1 DNA was greatest in β7high cells (median, 6200 CPM cells; Supplementary Figure 2), whereas β7negative and β7int cells contained approximately 2- and 60-fold fewer copies (P = .03 and P = .0005, respectively). Since HIV-1 infection downregulates CD4 surface expression [21], β7high and β7negative T cells double-negative for CD4 and CD8 were also analyzed (Supplementary Figure 2). Median total HIV-1 DNA in β7high double-negative T cells was approximately 40-fold less than that of the β7high CD4+ T cells (160 CPM cells), and was similar between the β7high and β7negative subsets (Supplementary Figure 2). Thus, both total and integrated HIV-1 DNA are enriched in β7high vs β7negative memory CD4+ T cells, while infection is less frequent in T cells lacking surface CD4.

To determine whether viral transcription is also more common among β7high cells, we determined the frequency of cells expressing spliced viral RNA by qPCR for HIV-1 env in 7 of the Fiebig II/III donors [16]. The proportion of HIV-1 transcriptionally active β7high cells was approximately 400 per million cells and was 5-fold greater than that of β7negative CD4+ T cells (P = .03; Figure 2C). Therefore, β7high cells more commonly transcribe HIV-1 during acute infection.

To assess the maintenance of β7high cell preferential infection following viral suppression, we measured integrated HIV-1 DNA longitudinally in individuals who initiated cART during acute infection Fiebig stage II/III and remained on treatment for 8 months (n = 10). As expected [22], integrated HIV-1 DNA declined after treatment in both β7high and β7negative cells (Figure 2D). The median decrease was 49-fold in β7high cells (P = .002) and 6-fold in β7negative cells (P = .002). The decline in integrated DNA was significantly greater in β7high cells (P = .004), indicating potentially greater decay of infected β7high than β7negative cells during therapy. Integrated HIV-1 DNA burden was approximately 20 CPM β7high cells and did not differ between β7high and β7negative cells following treatment (Figure 2E). There was a strong correlation between integrated HIV-1 DNA before and after treatment, in both β7high and β7negative subsets, consistent with previous analyses of total CD4+ T cells [23] (Figure 2F).

Phenotypic Profile of β7high Memory CD4+ T Cells

To understand characteristics of β7high cells that may relate to their susceptibility to HIV-1 infection, we measured surface expression of T-cell activation markers (CD38, CD69, HLA-DR, PD-1, ICOS, OX40), memory markers (CCR7, CD27), and chemokine receptors (CCR5, CXCR5, CXCR3, CCR6) implicated in HIV-1/SIV infection, in 6 uninfected controls and the Fiebig II/III acutely infected participants (n = 12; Supplementary Figure 3) [6, 11, 18, 24–31].

In uninfected individuals, differences in the expression of activation markers between β7high and β7negative cells were relatively small and varied in direction. CD38 expression was higher among β 7high, while HLA-DR, ICOS, and OX40 were more common on β7negative cells (all P = .03; Figure 3A). Surface expression of PD-1 and CD69 did not differ between β7high and β7negative cells. As expected, CD4+ T-cell activation increased in acute infection relative to uninfected controls (Figure 3A), regardless of β7 profile. Comparing the β7-defined subsets during acute infection, CD38, CD69, and ICOS expression were greater among β7high than β7negative CD4+ T cells, while OX40 was higher for β7negative. PD-1 and HLA-DR did not differ between the subsets. Chemokine receptor expression also differed between β7high and β7negative cells in acutely infected individuals, with significantly greater CXCR5 and CXCR3 expression by β7high cells (both P < .001; Figure 3B). Surface expression of CCR5, CCR6, and CCR7 was similar between the β7-defined subsets. Similar trends were observed in uninfected controls. CD38+ HLA-DR+ double-positive cells, a more definitive measure of T-cell activation, were also increased among β7high cells in uninfected controls and at Fiebig II/III (P = .03 and P = .001, respectively; Figure 3C). Thus, β7high cells tend to display a more activated phenotype than β7negative cells, particularly during acute HIV-1 infection, potentially predisposing β7high cells to HIV-1 infection.

Figure 3.

Activation marker and chemokine receptor expression by β7high CD4+ T cells. Flow cytometric analysis of surface protein expression on β7negative and β7high CD4+ T cells from peripheral blood mononuclear cells of Fiebig II/III participants with human immunodeficiency virus type 1 infection (top) and uninfected (bottom) individuals. The frequency of cells positive for the indicated activation marker (A) and chemokine receptor (B) within β7high and β7negative subsets is plotted for each participant. C, Frequency of cells double-positive for the indicated combination of markers is plotted as in A and B. For each graph, nonparametric Wilcoxon signed-rank test significant and trending P values are shown. Abbreviations: n.s., not significant; Tcm, central memory T cells; Tem, effector memory T cells; Ttm, transitional memory T cells.

β7high cells were also enriched for peripheral T follicular helper (CXCR5+PD-1+) and Th1/Th17 (CXCR3+CCR6+) cells relative to β7negative cells, mirroring the distribution of cells positive for the individual markers, CXCR5 and CXCR3, and was consistent between infected and uninfected individuals (Figure 3C). Memory status did not differ between β7high and β7negative subsets among uninfected controls, with comparable central (CCR7+CD27+), transitional (CCR7–CD27+), and effector (CCR7–CD27–) memory frequencies. In contrast, in acute infection, the proportion of central and transitional memory cells among β7high cells was greater than that among β7negative cells, while effector memory predominated among β7negative cells. This suggests that effector memory β7high cells are preferentially depleted during acute infection, though acquisition or maintenance of a more long-lived memory phenotype by β7high cells relative to β7negative cells is also possible.

β7high CD4+ T-cell Activation and HIV-1 Burden

We explored the relationship between the extent of β7high CD4+ T-cell activation and the amount of β7high CD4+ T-cell–associated HIV-1 DNA in the Fiebig II/III acutely infected individuals. Integrated HIV-1 DNA in β7high cells was highly correlated with the proportion of β7high cells expressing CD38 (P = .006; Figure 4A). ICOS+ and PD-1+ β7high cells were also associated with integrated HIV-1 DNA copies (P = .05 and P = .03, respectively), while a similar trend was observed for HLA-DR+ and CD38+HLA-DR+ double-positive β7high cells (both P = .06; Figure 4A and data not shown). Therefore, the general activation state of β7high cells is highly linked with their infection burden.

Figure 4.

Relationship between β7high CD4+ T-cell activation and cell-associated virus. A, Correlation between the frequency of β7high CD4+ T cells expressing the indicated activation marker and β7high CD4+ T-cell–associated integrated human immunodeficiency virus type 1 (HIV-1) DNA in peripheral blood mononuclear cells (PBMCs) of Fiebig II/III acutely infected individuals. Spearman ρ and P values are shown. B, Schematic for single-cell HIV-1 gene and host protein analysis applied to PBMCs directly ex vivo. Surface protein expression recorded by flow cytometry for single cells deposited into multiwell plates is combined with reverse-transcription quantitative polymerase chain reaction (RT-qPCR) viral gene data for each cell by well position. C, Surface CD38 expression (fluorescence intensity) is plotted for β7negative and β7high CD4+ T cells that either lack or contain HIV-1 RNA. Graphs depict results from 2 Fiebig III participants with a high frequency of infected cells. Cells (dots) are colored by HIV-1 RNA status assessed by gag and LTR single-cell RT-qPCR. Surface CD38 values < 50 (autofluorescence of unstained cells, “negative”) are randomly distributed between 1 and 50 for visualization. Significant differences determined by nonparametric Mann-Whitney rank-sum test are indicated within like HIV-1 infection states. Median values and 95% confidence intervals are shown.

To determine whether the increased level of CD38 on β7high cells is related to HIV-1 infection at the individual cell level, single CD4+ T cells were assessed for combined surface activation marker expression and viral RNA. PBMCs from 2 Fiebig III donors with a high frequency of infected cells were surface stained and sorted by flow cytometry for single-cell HIV-1 RT-qPCR (Figure 4B) [16, 17]. Cells qualitatively positive for viral LTR or gag RNA by a single-copy assay (Supplementary Figure 4) were (1) broadly defined as infected, including preintegration stages; (2) approximately 10-fold more common than env-expressing cells; and, therefore, (3) most amenable to well-powered comparisons. β7high cells expressed more CD38 than β7negative cells regardless of HIV-1 RNA positivity (Figure 4C; n > 800 cells per donor). However, CD38 expression did not significantly differ between HIV-1+ and uninfected β 7high cells, though a trend of increased CD38 on HIV-1+ β7high was observed for 1 donor (P = .10). Furthermore, the frequency of gag/LTR+ infected cells was similar between β7high and β7negative subsets (3–4 × 105 per 106 cells, not shown), in contrast to env RNA and integrated DNA. Thus, increased activation among β7high cells is independent of cellular HIV-1 infection status defined by gag/LTR viral RNA.

Discussion

The initial stages of HIV-1 infection are dynamic and play an important role in establishing disease progression and formation of viral reservoirs. Therefore, understanding host–virus interactions in very early infection is relevant for HIV-1 prevention and treatment strategies. Given the importance of α4β7+ CD4+ T cells as HIV-1 targets, including at transmission, we hypothesized that these cells are preferentially infected during early acute HIV-1 infection. Indeed, most individuals in Fiebig II/III and chronic untreated HIV-1 infection exhibited greater infection of β7high cells compared to β7negative memory cells. Enrichment of HIV-1 DNA in β7high CD4+ T cells was not observed at Fiebig I or following suppressive cART. Spliced HIV-1 RNA, a marker of active virus transcription, was also increased in β7high cells during Fiebig II/III. High expression of β7 demarcated more activated memory CD4+ T cells during acute infection, and β7high cell activation levels were associated with the HIV-1 DNA burden. Collectively, HIV-1 DNA and RNA enrichment in β7high CD4+ T cells indicates a viral preference to infect and replicate in these cells, during both acute and chronic infection.

Our findings reproduce those of a rhesus macaque SIV model, in which α4β7+ cells were preferentially infected and partially depleted at peak viremia [11]. They also support our previous report of α4β7+ cell frequency predicting higher rates of HIV-1 acquisition and disease progression, and the rapid depletion of α4β7+ CD4+ T cells from the gastrointestinal tract of participants from the same RV254 cohort [9]. Consistent with Sivro et al, depletion of circulating β7high cells was apparent at Fiebig II/III, whereas depletion of colonic β7high cells was previously observed as early as in Fiebig I, suggesting a more rapid loss of β7high cells in the colon vs blood [9]. The preferential infection of α4β7+ CD4+ T cells in chronic, untreated HIV-1 is consistent with the enrichment of HIV-1 Gag protein expression among α4β7+ CD4+ T cells in chronic infection reported previously [20], but in contrast to the macaque model, where SIV DNA content in β7-defined subsets did not differ 3 months postinfection, indicating potential differences between SIV and HIV-1 infection [11].

One limitation of the current study is the cross-sectional nature of the comparison between uninfected and acutely HIV-1–infected individuals, particularly given the high variability in β7high CD4+ T-cell frequency among controls. In addition, measuring integrated HIV-1 DNA likely overestimates the size of the replication competent reservoir due to the inability of this assay to discriminate between intact and defective proviruses [32, 33]. However, defective virus is likely less abundant in this early acute cohort than during chronic or late acute HIV-1 given the limited duration of viral replication and the limited number of defects that accumulated in this relatively short period of time. Moreover, due to the exceptionally small size of viral reservoirs in this early acute infection cohort, proviral DNA measurements are the most robust and sensitive approach for quantifying infection [34]. Another possible limitation is that given the slight reduction in β7high CD4+ T cells in PBMCs following cryopreservation suggesting labile β7 expression, our results may underestimate HIV-1 DNA enrichment in these cells in vivo.

The mechanism behind this preferential targeting remains incompletely understood. In the present work, β7high CD4+ T cells of uninfected individuals overlapped with activated and peripheral T follicular helper cells, both of which are associated with susceptibility to HIV-1 [6, 30, 35]. During acute infection, activation markers were substantially upregulated on β7high cells, and generally to a greater extent than that observed for β7negative cells, indicating heightened activation of the β7high subset. The frequency of Th1/Th17 and transitional and central memory cells, all previously shown to be preferentially infected [28, 36], are also elevated among β7high cells. Another possibility is that β7high effector memory cells were selectively infected and depleted. Notably, of all surface markers tested, only markers of activation were correlated with integrated HIV-1 DNA within β7high cells, suggesting that activation state drives β7high cell HIV-1 susceptibility in acute infection. Increased activation of β7high cells was independent of early or abortive infection of these cells determined by the presence of LTR or gag RNA. It remains to be determined if increased activation of β7high cells is associated with transcriptionally active cellular infection. In the setting of chronic HIV-1, the mechanism(s) underlying HIV-1 DNA enrichment in β7high cells may differ from that of acute infection. For example, greater persistence of β7high infected cells may also contribute. Together these data support a model whereby α4β7+ CD4+ T cells are inherently more susceptible to HIV-1 compared to α4β7-negative cells and become preferentially activated and infected during the rise of HIV-1 viremia in early infection.

The low frequency of β7high CD4+ T cells in blood, 5%–15% of total CD4+ T cells, observed here and previously [9] indicates that despite virus enrichment in these cells compared to the other β7-defined populations, the majority of infected cells are present outside of this subset. This provides a potential explanation for lack of therapeutic efficacy of α4β7-specific antibody in the setting of treatment interruption reported recently in rhesus macaques [37-39]. However, mutations in SIV env observed following anti-α4β7 therapy are consistent with HIV/SIV preference for this receptor (personal communication, M. Roederer). The proportion of β7high CD4+ T cells in the gut of uninfected individuals is significantly higher (~30%) than that observed in peripheral blood, and further studies are required to evaluate the efficacy of α4β7 targeting in prevention of mucosal transmission [40].

In summary, our data add to a large body of evidence that α4β7+ CD4+ T cells are preferred HIV-1 targets. Infection of this gut-homing subset may have important implications for HIV-1 pathogenesis, including during treatment.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Click here for additional data file.

Click here for additional data file.

Click here for additional data file.

Click here for additional data file.

Click here for additional data file.

Click here for additional data file.