α4⁺β7^hiCD4⁺ memory T cells harbor most Th-17 cells and are preferentially infected during acute SIV infection

Abstract

Human immunodeficiency virus/simian immunodeficiency virus (HIV/SIV) infections are believed to infect minimally activated CD4⁺ T cells after viral entry. Not much is known about why SIV selectively targets these cells. Here we show that CD4⁺ T cells that express high levels of the α4β7 heterodimer are preferentially infected very early during the course of SIV infection. At days 2–4 post infection, α4⁺β7^hiCD4⁺ T cells had ∼5× more SIV-gag DNA than β7⁻CD4⁺ T cells. α4⁺β7^hiCD4⁺ T cells displayed a predominantly central memory (CD45RA⁻CD28⁺CCR7⁺) and a resting (CD25⁻CD69⁻HLA-DR⁻Ki-67⁻) phenotype. Although the expression of detectable CCR5 was variable on α4⁺β7^hi and β7⁻CD4⁺ T cells, both CCR5⁺ and CCR5⁻ subsets of α4⁺β7^hi and β7⁻CD4⁺ T cells were found to express sufficient levels of CCR5 mRNA, suggesting that both these subsets could be efficiently infected by SIV. In line with this, we found similar levels of SIV infection in β7⁻CD4⁺CCR5⁺ and β7⁻CD4⁺CCR5⁻ T cells. α4β7^hiCD4⁺ T cells were found to harbor most T helper (Th)-17 cells that were significantly depleted during acute SIV infection. Taken together, our results show that resting memory α4⁺β7^hiCD4⁺ T cells in the blood are preferentially infected and depleted during acute SIV infection, and the loss of these cells alters the balance between Th-17 and Th-1 responses, thereby contributing to disease pathogenesis.

INTRODUCTION

Acute human immunodeficiency virus (HIV) and simian immunodeficiency virus (SIV) infections are characterized by a massive loss of memory CD4⁺ T cells,^{1, 2, 3, 4, 5, 6, 7} a process that is most pronounced in mucosal tissues, as these tissues are enriched in memory CD4⁺ T cells. As in non-natural hosts, severe depletion of CD4⁺ T cells has been reported in natural hosts, such as sooty mangabeys and African green monkeys, during acute SIV infection.^{8, 9} Surprisingly, the kinetics of infection in memory CD4⁺ T cells that reach a peak ∼10 days after SIV challenge appear to be independent of the route of infection.^{4, 5}

The factors determining which target cells that are infected initially, and those responsible for the delay between initial infection and the explosive levels of viral replication observed 10 days later, remain unclear. Studies⁵ have shown that before day 7 post infection (pi) <5 to 10% of CD4⁺ T cells in either the mucosal tissues or periphery are infected, whereas by day 10 pi, most memory CD4⁺ T cells are infected and carry viral DNA. Li et al.⁴ showed that a few founder populations of CD4⁺ T cells in the gut-associated lymphoid tissue (GALT) are infected initially during SIV infection. Interestingly, these founder cells were minimally activated CD4⁺ T cells (CD25⁻CD69⁻Ki-67⁻) and had a predominantly resting phenotype. It is not clear why SIV targets these CD4⁺ T cells during the early phase of viral infection.

Recent in vitro studies¹⁰ have shown that CD4⁺ T cells that express high levels of the α4β7 (α4⁺β7^hi) mucosal homing receptor can bind to HIV and SIVsmm with high affinity making these cells more susceptible to infection. Arthos et al.¹⁰ suggested that the expression of high levels of α4β7 on CD4⁺ T cells likely offers a selective advantage to the virus; hence, these cells may be the earliest cells targeted and infected by HIV. The loss of these mucosal homing T helper (Th) cells severely compromises the ability of the mucosal immune system. In line with this, recent studies have shown that mucosal Th-17 CD4⁺ T cells have a critical role in maintaining the integrity of the mucosal barrier.¹¹ Others have shown that alterations in the Th-17 and Th-1 responses contribute to disease progression.^{12, 13, 14}

We hypothesized that if HIV selectively targeted the α4⁺β7^hiCD4⁺ T cells during the early phase of infection, then these cells would be preferentially infected very early during the course of infection. We tested our hypothesis in the SIVmac251-infected rhesus macaque model by quantifying the cell-associated SIV DNA in α4⁺β7^hiCD4⁺ T cells from peripheral blood during the early phase of SIV infection and compared them with SIV DNA levels in CD4⁺ T cells that either lacked (β7⁻) or expressed intermediate levels of α4β7 (α4⁺β7^int). Previous studies¹⁵ have shown that HIV-specific CD4⁺ T cells that were preferentially infected carried 2–5× more viral DNA than Epstein–Barr virus- or cytomegalovirus-specific CD4⁺ T cells. To determine whether the differentiated and activated nature of α4⁺β7^hiCD4⁺ T cells contributes to their susceptibility to SIV infection, we evaluated the expression of phenotypic (CD28, CD45RA, and CCR7) and activation (CD25, CD69, HLA-DR, and Ki-67) markers on these cells. In addition, to evaluate whether preferential infection and depletion of α4⁺β7^hiCD4⁺ T cells led to alterations in the balance between Th-17- and Th1-type responses, we evaluated the expression of interleukin (IL)-17 and interferon (IFN)γ in α4⁺β7^hiCD4⁺ T cells and compared them with β7⁻CD4⁺ T cells.

RESULTS

α4⁺β7^hiCD4⁺ T cells are a minor population of peripheral CD4⁺ T cells

We used an antibody against the β7 receptor to identify different subsets of CD4⁺ T cells. As β7 can form a heterodimer with α4 (CD49d) or αE (CD103),¹⁶ we first evaluated the expression of these receptors on β7 subsets in peripheral blood of healthy rhesus macaques. Our results showed that ∼50% of peripheral blood CD4⁺ T cells expressed the β7 receptor, whereas the rest of the 50% lacked the β7 expression. Approximately 20% of CD4⁺ T cells that expressed the β7 receptor did so at high levels and had a β7^hi phenotype, whereas ∼30% of them expressed intermediate levels of β7 receptor and had a β7^int phenotype (Figure 1a). More than 95% of β7^hi and ∼65% of β7^int CD4⁺ T cells co-expressed the α4 receptor (Figure 1b), hereafter these are referred to as α4⁺β7^hi and α4⁺β7^int subsets. Studies have shown that cells expressing high levels of the α4β7 efficiently traffic to mucosal tissues.¹⁶

Figure 1

Phenotypic analysis of CD3⁺CD4⁺ T cells in the peripheral blood from healthy rhesus macaques (n=8). (a) Representative contour plots and (b) relative proportions of CD95⁺β7^hi (α4⁺β7^hi), CD95⁻ β7^int (α4⁺β7^int), and CD95⁺β7⁻ (β7⁻) subsets of CD4⁺ T cells. (c) Expression of α4 and αE integrins on α4⁺β7^hi, α4⁺β7^int, and β7⁻CD4⁺ T-cell subsets. (d) Relative proportions and (e) mean fluorescence intensity (MFI) of α4 expression on α4⁺β7^hi, α4⁺β7^int, and β7⁻CD4⁺ T-cell subsets. MFI was determined using FlowJo 8.6 software after gating on each subset.

PowerPoint slide

On the other hand, CD4⁺ T cells that lacked β7 expression contained a mix of α4⁺ (∼75%) and α4⁻ (∼25%) cells (Figure 1c), and these subsets together are referred to as β7⁻CD4⁺ T-cell subsets. α4 associates with β1 to form the α4β1 heterodimer and binds to VCAM-1 (vascular cell adhesion molecule-1) on activated endothelial cells, which is involved in trafficking of T cells to peripheral lymphoid tissues.^{17, 18, 19, 20} Interestingly, β7⁻CD4⁺ T-cell subsets expressed α4 integrin at a significantly higher density than α4⁺β7^hiCD4⁺ T-cell subsets (Figure 1e). Our results suggest that most of the β7⁻CD4⁺ T-cell subsets likely home to peripheral lymphoid tissues.

All of the α4⁺β7^hiCD4⁺ and β7⁻CD4⁺ T-cell subsets expressed CD95, indicating that these were memory CD4⁺ T cells. On the other hand, α4⁺β7^intCD4⁺ T cells lacked the CD95 expression, which is indicative of a naive phenotype. Previous studies have shown that rhesus macaque T cells expressing CD95 were memory T cells, whereas T cells that lacked CD95 were naive T cells.²¹ As our analysis was confined to peripheral blood, it is possible that the pattern of α4β7 expression may vary from that seen in GALT.

α4⁺β7^hiCD4⁺ T cells express a central memory phenotype

Next, we evaluated the phenotype of α4⁺β7^hiCD4⁺ T-cell subsets based on the co-expression of α4 and β7 integrins. Memory CD4⁺ T cells that express the α4β7 heterodimer appear on the diagonal, as the antibodies to α4 and β7 integrins bind to the same heterodimerized molecule, and co-labels with the Act-1 clone that recognizes an heterodimerized epitope of α4 and β7 integrins²² (Supplementary Figure S1a online).

Three major subsets of memory CD4⁺ T cells could be delineated based on the differential expression of these two integrin molecules, namely α4⁺β7^hi subset, and two subsets of β7⁻CD4⁺ T cells, namely α4⁺β7⁻ and α4⁻β7⁻ subsets (Figure 2a, b). Similar proportions of the three subsets are found within human memory CD4⁺ T cells (M. Roederer, personal communication; Supplementary Figure S1e online).

Figure 2

α4⁺β7^hiCD4⁺ T-cell subsets show a predominantly central memory phenotype (n=8). Co-staining for α4 and β7 identifies memory CD4⁺ T cells that express the α4β7 heterodimer along a diagonal. (a) Representative dot plots and (b) relative proportions of α4⁺β7^hi, α4⁺β7⁻, and α4⁻β7⁻ subsets of CD4⁺ T cells. (c) Representative dot plots and (d) relative proportions of α4⁺β7^hi, α4⁺β7⁻ and α4⁻β7⁻ subsets of CD4⁺ T cells expressing CD45RA, CD28, and CCR7. (e) Representative dot plots and relative proportions of α4⁺β7^hi, α4⁺β7⁻, and α4⁻β7⁻ subsets of CD4⁺ T cells expressing CCR5. Most α4⁺β7^hi subsets express detectable levels of surface CCR5. (f) Relative expression of CCR5 mRNA in CCR5⁺ and CCR5⁻ subsets of α4⁺β7^hi and β7⁻CD4⁺ T cells (n=6).

PowerPoint slide

Majority of the α4⁺β7^hiCD4⁺ T-cell subsets expressed a central memory phenotype (C28⁺CCR7⁺CD45RA⁻), whereas most of the effector memory CD4⁺ T cells (CD28⁻CCR7⁻CD45RA⁺) are found within the α4⁺β7⁻ subsets (Figure 2c, d).

Interestingly, most (95%) of the α4⁺β7^hiCD4⁺ T-cell subsets expressed detectable levels of CCR5 on their surface (Figure 2e), whereas only a few (<20%) of the β7⁻CD4⁺ T-cell subsets expressed detectable levels of CCR5. To determine whether these subsets expressed variable levels of CCR5 mRNA, we sorted CCR5⁺ and CCR5⁻ subsets of α4⁺β7^hi and β7⁻CD4⁺ T cells from uninfected animals (Supplementary Figure S1b online) and quantified the levels of CCR5 mRNA in these subsets using a relative quantitative PCR (qPCR) assay (Figure 2f). Our results show that although a majority of the β7⁻CD4⁺ T-cell subsets expressed little CCR5 on their surface, they had significant levels of CCR5 mRNA that did not differ from that of the other memory CD4⁺ T-cell subsets.

α4⁺β7^hiCD4⁺ T cells are preferentially infected during acute SIV infection

To determine whether the differential expression of α4 and β7 integrins on CD4⁺ T-cell subsets made them variably susceptible to infection and subsequent loss, we evaluated the dynamics of α4⁺β7^hiCD4⁺ T-cell subsets in peripheral blood and compared them with α4⁺β7^int and β7⁻CD4⁺ T-cell subsets.

We observed a significant loss of both α4⁺β7^hi and β7⁻CD4⁺ T-cell subsets in peripheral blood at day 63 pi compared with day 10 pi (Figure 3a, b), whereas there was no major difference in either the frequency or absolute numbers of α4⁺β7^intCD4⁺ T-cell subsets or CD8⁺ T-cell subsets (Figure 3a–d). The loss of both α4⁺β7^hi and β7⁻CD4⁺ T-cell subsets was not unexpected, as previous studies⁵ have reported a significant loss of memory CD4⁺ T cells in peripheral blood during acute SIV infection; both α4⁺β7^hi and β7⁻CD4⁺ T-cell subsets have a memory phenotype. However, we observed a significantly higher loss of α4⁺β7^hiCD4⁺ T-cell subsets by day 63 pi compared with the β7⁻CD4⁺ T-cell subsets (Figure 3e), suggesting that α4⁺β7^hiCD4⁺ T-cell subsets were preferentially depleted during the course of acute SIV infection.

Figure 3

Dynamics of α4β7 subsets of CD4⁺ T cells in peripheral blood during acute simian immunodeficiency virus (SIV) infection. Significant loss of α4β7^hi and β7⁻ CD4⁺ T cells is observed during acute SIV infection. Both frequency and absolute numbers are shown for (a, b) CD4⁺ and (c, d) CD8⁺ T cells (n=8). (e) Higher numbers of α4⁺β7^hiCD4⁺ T cells were destroyed by day 63 pi (post infection) compared with β7⁻ CD4⁺ T cells. The percentage (%) change in absolute numbers of α4⁺β7^hi and β7⁻CD4⁺ T cells was calculated using each animal's day 63 values relative to its day 0 values with day 0 values as 100%.

PowerPoint slide

To determine whether the preferential loss of α4⁺β7^hiCD4⁺ T-cell subsets was due to a higher level of viral infection in these subsets, we measured the level of SIV-gag DNA in α4⁺β7^hiCD4⁺ T-cell subsets and compared them with α4⁺β7^int and β7⁻CD4⁺ T-cell subsets. α4⁺β7^hi, α4⁺β7^int, and β7⁻CD4⁺ T-cell subsets, discriminated based on the expression of β7 and CD95²¹ (Supplementary Figure S1c online), were sorted from peripheral blood and used in a qPCR assay to measure SIV-gag DNA. We have shown earlier that this assay was highly sensitive for measuring viral infection in CD4⁺ T cells.⁵ Measuring the level of SIV DNA in target cells allows us to quantify both productively and non-productively infected cells; hence, it is a more sensitive assay for quantifying viral infection in cells than RNA-based assays that measure only productively infected cells. Previous studies have shown that most of the memory CD4⁺ T cells carried viral DNA during acute SIV infection, whereas only about 7–20% of them were productively infected.^{4, 5}

We found ∼5× more SIV-gag DNA in α4⁺β7^hiCD4⁺ T-cell subsets at days 2–4 pi compared with β7⁻CD4⁺ T-cell subsets, suggesting that α4⁺β7^hiCD4⁺ T-cell subsets were preferentially infected very early during the course of infection (Figure 4b, c). By day 10 pi, there was a significant increase in the number of SIV-gag DNA copies in both subsets compared with days 2–4 pi. However, as at days 2–4 pi, α4⁺β7^hiCD4⁺ T-cell subsets harbored ∼3× more SIV-gag DNA than β7⁻CD4⁺ T-cell subsets (Figure 4b, c).

Figure 4

α4⁺β7^hiCD4⁺ T cells are preferentially infected during acute simian immunodeficiency virus (SIV) infection. (a) Plasma viral loads. Limit of detection is <30 copies per ml of plasma. (b) Levels of SIV-gag DNA were determined using qPCR assay for SIV-gag in sorted subsets of α4⁺β7^hi, α4⁺β7^int, and β7⁻ CD4⁺ T cells. α4⁺β7^hiCD4⁺ T cells are preferentially infected at days 2–4 and day 10 pi (post infection). (c) Ratio of SIV-gag DNA levels in α4⁺β7^hi:β7⁻CD4⁺ T cells at days 2–4 pi, 10 pi, and 63 pi. The expression of (d) HLA-DR, (e) CD69, (f) CD25, and (g) Ki-67 on α4⁺β7^hi and β7⁻CD4⁺ T cells during the early phase of SIV infection.

PowerPoint slide

By day 63 pi, however, the number of SIV-gag copies decreased significantly in both α4⁺β7^hi and β7⁻CD4⁺ T-cell subsets (Figure 4b), indicating that most of the cells that were infected at day 10 pi were destroyed by day 63 pi. Interestingly, unlike days 2–4 and 10 pi, both α4⁺β7^hi and β7⁻CD4⁺ T-cell subsets at day 63 pi had similar levels of SIV-gag DNA (Figure 4c).

Few α4⁺β7^hiCD4⁺ T-cell subsets expressed HLA-DR (Figure 4d) that did not change after infection. Previous reports²³ have shown that there was no significant change in the expression of HLA-DR on memory CD4⁺ T cells during the early phase of acute SIV infection. Similar to HLA-DR, few α4⁺β7^hiCD4⁺ T-cell subsets expressed CD25 (<0.5%), CD69 (<0.5%), or Ki-67 (∼3%) before infection that did not change significantly at day 4 pi (Figure 4d–g). Previous studies have shown that there was little or no change in the expression of HLA-DR or Ki-67 on memory CD4⁺ T cells at days 4, 7, and 10 pi compared with pre-infection values,²³ whereas others have shown that very few peripheral blood CD4⁺ T cells express CD25, CD69, or Ki-67 during the acute phase of SIV infection.²⁴

Unlike α4⁺β7^hiCD4⁺ T-cell subsets, higher proportions of β7⁻CD4⁺ T-cell subsets expressed HLA-DR before infection, yet these cells were not preferentially infected. Similarly, β7⁻CD4⁺ T-cell subsets were found to express significantly higher levels of Ki-67 at day 4 pi compared with uninfected animals. Taken together, these findings indicate that preferential infection during the early phase of SIV infection is confined to α4⁺β7^hiCD4⁺ T-cell subsets that display a predominantly resting phenotype. This is in line with previous reports by Li et al.,⁴ who showed that SIV primarily infects resting memory CD4⁺ T cells in the GALT that express a CD25⁻CD69⁻Ki-67⁻ phenotype. It is important to point out that our results only represent snap shots of the changes occurring in vivo. As such it is difficult to rule out if α4⁺β7^hiCD4⁺ T-cell subsets were transiently activated at the time of infection.

To confirm whether SIV preferentially targeted CD4⁺ T-cell subsets that expressed the α4β7 heterodimer during the early phase of SIV infection, we evaluated the levels of SIV-gag DNA in α4⁺β7^hi, α4⁺β7⁻, and α4⁻β7⁻ subsets of memory CD4⁺ T cells (Figure 5a) using peripheral blood samples (days 2–3 pi) that were obtained from an unrelated study.

Figure 5

CD4⁺ T cells that express the α4β7 heterodimer are preferentially infected in vivo. Memory CD4⁺ T cells were sorted based on the co-expression of α4 and β7 integrins by flow cytometry, and used in the qPCR assay for simian immunodeficiency virus (SIV)-gag DNA. (a) Sorting strategy and purity of sorted subsets are shown along with the level of SIV infection in α4⁺β7^hi, α4⁺β7⁻, and α4⁻β7⁻ subsets at days 2–3 pi (post infection). α4⁺β7^hi subsets harbor significantly higher levels of SIV DNA at days 2–3 pi (n=4; P=0.0244) compared to α4⁺β7⁻ and α4⁻β7⁻ subsets. Statistical analysis was performed using Kruskal–Wallis one way analysis of variance (ANOVA) (b) There is no significant difference in the level of SIV-gag DNA between β7⁻CD4⁺CCR5⁺ and β7⁻CD4⁺CCR5⁻ T-cell subsets of memory CD4⁺ T cells from the peripheral blood at days 7 and14 pi (day 7: n=2, day 14: n=2). β7⁻ subsets were gated based on the lack of β7 expression and CD95, followed by sorting of CCR5⁺ and CCR5⁻ subsets.

PowerPoint slide

Our results showed that during the first few days after infection, the virus is largely confined to memory CD4⁺ T cells that express the α4β7 heterodimer; α4⁺β7^hiCD4⁺ T-cell subsets had significantly higher levels of SIV-gag DNA than either α4⁺β7⁻ or α4⁻β7⁻CD4⁺ T-cell subsets (Figure 5a).

Taken together, these results support the in vitro findings reported by Arthos et al.,¹⁰ and show that SIVmac251 selectively target CD4⁺ T cells that express high levels of the α4β7 receptor.

Both β7⁻CD4⁺CCR5⁺ and β7⁻CD4⁺CCR5⁻ T-cell subsets are equally infected during acute SIV infection

To determine whether detectable surface expression of CCR5 makes CD4⁺ T cells subsets more susceptible to SIV infection, we evaluated the levels of SIV infection in β7⁻CD4⁺CCR5⁺ T-cell subsets and compared them with β7⁻CD4⁺CCR5⁻ T-cell subsets. Krzysiek et al.²⁵ showed that CCR5⁺CD4⁺ Th cells with non-lymphoid homing potential are preferentially depleted during HIV infection despite early treatment. Similarly, a number of studies have shown that CCR5⁺CD4⁺ Th cells were preferentially depleted during HIV and SIV infection, whereas others have shown that both CCR5⁺ and CCR5⁻CD4⁺ T-cell subsets were infected during acute SIV infection.^{1, 5, 26, 27}

As most (>95%) of the α4⁺β7^hiCD4⁺ T-cell subsets express detectable levels of CCR5 on their surface, we did not have sufficient samples to sort enough α4⁺β7^hiCD4⁺CCR5⁻ T-cell subsets for the qPCR analysis. Hence, we restricted our analysis to β7⁻CD4⁺ T-cell subsets. We sorted β7⁻CD4⁺CCR5⁺ and β7⁻CD4⁺CCR5⁻ T-cell subsets from peripheral blood of SIV-infected animals (days 7 and 14 pi) obtained from an unrelated study and quantified the levels of SIV-gag DNA by qPCR.

Our results (Figure 5b) show that both β7⁻CD4⁺CCR5⁺ and β7⁻CD4⁺CCR5⁻ T-cell subsets had similar levels of SIV-gag DNA, which did not differ significantly, suggesting that preferential infection was not directly dependent on the surface level of CCR5 expression. As shown in Figure 2f, CCR5⁻ T-cell subsets express sufficient levels of CCR5 message that is likely sufficient to make these cells susceptible to SIV infection. These results support previous studies,⁵ showing that there was no difference in the level of SIV infection in memory CD4⁺CCR5⁺ and memory CD4⁺CCR5⁻ T-cell subsets at days 3, 7, 10, and 14 pi.

α4⁺β7^hiCD4⁺ T-cell subsets harbor most Th-17 cells and are lost during acute SIV infection

Previous studies have shown that mucosal CD4⁺ T cells are major producers of IL-17, and HIV infection was associated with alterations in the balance between Th17- and Th1-type responses.^{11, 12, 13, 14} To determine whether the preferential loss of α4⁺β7^hiCD4⁺ T-cell subsets was accompanied by changes in the profile of Th17- vs. Th1-type responses, we evaluated the production of IL-17 and IFNγ by α4⁺β7^hiCD4⁺ T-cell subsets and compared them with β7⁻CD4⁺ T-cell subsets.

Our results showed that α4⁺β7^hiCD4⁺ T-cell subsets harbored significantly higher frequencies of Th17-type cells, whereas β7⁻CD4⁺ T-cell subsets harbored higher proportions of IFNγ-producing Th1-type cells (Figure 6a, b). Most of the Th-17 cells were found to be CD28⁺ memory CD4⁺ T cells (Supplementary Figure S1d online). The ratio of IFNγ:IL-17-producing cells was ∼2 within the α4β7^hiCD4⁺ T-cell subsets (Figure 6c), whereas it was >10 within the β7⁻CD4⁺ T-cell subsets, indicating that α4β7^hiCD4⁺ T-cell subsets were the primary source of Th-17 responses. SIV infection was found to significantly skew the ratio to a predominantly Th-1 phenotype, as there was a significant depletion of Th-17 cells during early SIV infection.

Figure 6

α4⁺β7^hiCD4⁺ T cells harbor significantly higher frequencies of Th-17 cells that are lost after simian immunodeficiency virus (SIV) infection. α4⁺β7^hiCD4⁺ T-cell subsets have a significantly higher capacity to produce interleukin (IL)-17 compared with β7⁻CD4⁺ T-cell subsets. Expression of IL-17 and interferon (IFN)γ was determined by flow cytometry using peripheral blood mononuclear cells (PBMCs) from uninfected (n=6) and SIV-infected animals (n=8) after stimulation with phorbol myristate acetate (PMA)/ionomycin for 4 h. (a) Representative dotplots showing IL-17 and IFNγ production by α4⁺β7^hi and β7⁻CD4⁺ T-cell subsets at days 0 and 77 pi (post infection). (b) IL-17 and IFNγ production by α4⁺β7^hi and β7⁻CD4⁺ T-cell subsets at days 0 (n=6) and 77 pi (n=8). (c) The ratio of IFNγ:IL-17 producing α4⁺β7^hi and β7⁻CD4⁺ T-cell subsets at days 0 and 77 pi.

PowerPoint slide

DISCUSSION

Early SIV infection have been shown to target founder populations of CD4⁺ T cells.⁴ These cells are minimally activated, and have a central role in early viral replication and dissemination. Not much is known about why SIV specifically targets these founder populations of CD4⁺ T cells. Given the highly activated microenvironment in mucosal tissues, and the enrichment of target cells in these tissues, one would expect massive infection and replication to occur after viral entry. However, there appears to be a substantial delay of about 7–8 days before infection explodes out of the early founder population of cells to infect and destroy most of the memory CD4⁺ T-cell compartment.^{4, 5} Arthos et al.¹⁰ using in vitro studies showed that HIV and SIVsmm selectively targeted CD4⁺ T cells that expressed the α4β7 heterodimer, suggesting that these cells could be earliest targets for viral infection in vivo. It is not known whether HIV preferentially targets memory CD4⁺ T cells that express the α4β7 receptor in vivo.

Using the SIVmac251 infection model, we show that memory CD4⁺ T cells that expressed high levels of the α4β7 heterodimer are preferentially infected during the early phase of viral infection; α4⁺β7^hiCD4⁺ T cells harbored ∼5× more SIV-gag DNA compared with β7⁻CD4⁺ T cells. Surprisingly, α4⁺β7^hiCD4⁺ T cells were found to be predominantly resting T cells (CD25⁻CD69⁻HLA-DR⁻Ki67⁻) with a central memory (CD28⁺CD45RA⁻CCR7⁺) phenotype that did not significantly change during the early phase of viral infection. Li et al.⁴ reported that during the early phase of infection, SIV infection in the GALT is confined to resting CD4⁺ T cells that were CD25⁻CD69⁻Ki67⁻.

Both CCR5⁺ and CCR5⁻ subsets of α4⁺β7^hiCD4⁺ and β7⁻CD4⁺CCR5⁺ T cells were found to have significant levels of CCR5 mRNA, suggesting that all the memory CD4⁺ T cells subsets have sufficient CCR5 for infection by SIV. In line with this, both β7⁻CD4⁺CCR5⁺ and β7⁻CD4⁺CCR5⁻ T cells had similar levels of SIV infection. It is likely that once SIV binds to the cells, the high levels of CCR5 expression enables SIV to more efficiently infect α4⁺β7^hiCD4⁺ T cells, especially during the early phase of infection, when viral loads are very low, as seen at days 2–4 pi.

By days 7–10 pi, there was a significant increase in the extent of infection in both α4⁺β7^hi and β7⁻CD4⁺ T cells, most likely as a consequence of the massive viral replication occurring during this period. However, α4⁺β7^hiCD4⁺ T-cell subsets were found to harbor ∼3× more SIV DNA than β7⁻CD4⁺ T-cell subsets.

Interestingly, α4⁺β7^hiCD4⁺ T cells were found to contain ∼3.5 to 4 copies of SIV per cell at day 10 pi, suggesting that there were likely more than 1 virion per cell. The assay used to quantify SIV infection in cells is stringently standardized using a non-expressing SIV-infected cell line that carries only a single copy for SIV DNA.⁵ As such, we do not think that these high copy numbers are assay related. It is more likely that the expression of the α4β7 receptor along with CD4 and high levels of CCR5 makes α4⁺β7^hiCD4⁺ T cells more susceptible to super infection with SIV. Previous studies have reported multiply infected CD4⁺ T cells in HIV-infected patients. Jung et al.²⁸ using fluorescent in situ hybridization assays showed that CD4⁺ T cells in the spleen of HIV-infected patients carried 4–6 proviruses per infected cell. Future studies will attempt to address this question in more detail.

By day 63 pi, the extent of infection had equalized between the two subsets, suggesting that there was a significantly higher loss of infected α4⁺β7^hiCD4⁺ T cells after peak viral infection. In line with this conclusion, a significantly greater number of α4⁺β7^hiCD4⁺ T-cell subsets were lost between days 0 and 63 pi compared with the β7⁻CD4⁺ T-cell subsets.

A number of factors likely contribute to the amplification of infection in β7⁻CD4⁺ T-cell subsets. Acute SIV infection has been shown to be associated with acute immune activation,^{23, 24} leading to the production of numerous pro-inflammatory mediators such as IL-15.²³ Similarly, acute HIV infection is associated with an increase in numerous pro-inflammatory cytokines, such as TNFα (tumor necrosis factor-α), IFNγ, IL-6, IL-8, MCP-1 (monocyte chemotactic protein-1), IFNα.²⁹ These factors likely have a role in making CD4⁺ T cells more susceptible to infection. Eberly et al.²³ showed that there was a significant increase in the production of IL-15 between days 7 and 14 pi, which was found to correlate with infection in memory CD4⁺ T cells during the early stages of infection. Others have shown that efficiency of infection was greatly influenced by the density of CD4, but not CCR5, expression on the surface of target cells.^{30, 31, 32, 33}

It is also possible that during periods of high viremia, as seen during the ramp-up phase of acute viral infection, the threshold for efficient infection is lower than during the pre-ramp phase, when viral loads are very low. Thus, although β7⁻CD4⁺ T-cell subsets inherently expressed low levels of CCR5, it was likely sufficient for SIV to efficiently infect these cells between days 7 and 14 pi when plasma viral loads are as high as 7–8 logs per ml.

Previous studies^{4, 5} have shown that <5% of memory CD4⁺ T cells were infected before day 7 pi, whereas by days 10–14 pi, there was a significant increase in infection of these cells. Although these memory CD4⁺ T cells lacked detectable CCR5 on their surface, they were found to have 20× more CCR5 mRNA compared with naive CD4⁺ T cells, which was sufficient for them to be infected by SIV. In line with these studies, we found significant levels of CCR5 mRNA in all the subsets of α4⁺β7^hi and β7⁻CD4⁺ T cells. Others have shown that, minimal expression of CCR5 was sufficient for infection of CD4⁺ T cells.^{30, 32, 33} Eberly et al.²³ showed that acute SIV infection was associated with little or no upregulation of CCR5 expression on memory CD4⁺ T cells during the first 10 days of infection, yet there was a significant increase in the level of SIV infection in these cells between days 7 and 10 pi. Taken together, these studies indicate that when viral loads are high, minimal levels of CCR5 is likely sufficient to support efficient infection of memory CD4⁺ T-cell subsets.

Interestingly, α4⁺β7^hiCD4⁺ T-cell subsets harbored a significantly higher potential to produce IL-17, suggesting that most of the Th-17 cells are found within these mucosal homing T-cell subsets. The preferential loss of these subsets during acute SIV infection likely cripples the ability of the immune system to maintain homeostasis in mucosal tissues. On the other hand, most of the IFNγ-producing Th-1 cells were found to be β7⁻CD4⁺ T-cell subsets, and fewer of these cells were lost, suggesting a skewing of the Th cells response to a Th-1 phenotype. Recent studies¹¹ have shown that Th-17 responses have a central role in protecting the mucosal barrier function, and the loss of Th-17 cells was associated with systemic dissemination of Salmonella in SIV-infected rhesus macaques. Cecchinato et al.¹³ showed that altered balance between Th-17 and Th-1 cells in mucosal sites predicted AIDS progression in SIV-infected rhesus macaques. On the other hand, Favre et al.¹⁴ showed that pathogenic SIV infection was associated with altered balance between Th-17 and T-regulatory populations. Similarly, Brenchley et al.¹² showed that Th-17 cells were differentially depleted in pathogenic vs. non-pathogenic lentiviral infections. Macal et al.³⁴ showed that restoration of mucosal CD4 T cells was associated with enhanced Th-17 responses in HIV-infected patients. Taken together, these studies suggest that preferential infection and depletion of α4⁺β7^hiCD4⁺ T-cell subsets significantly skews Th cell responses toward a Th-1 phenotype, thereby contributing to disease progression.

In conclusion, our results provide in vivo evidence that, during the early phase of SIV infection, SIVmac251 selectively targets α4⁺β7^hiCD4⁺ T-cell subsets. These cells are predominantly central memory T cells with a resting phenotype and resemble the founder populations of CD4⁺ T cells described by Li et al.⁴ The inherently high level of CCR5 expression likely allows SIV to more efficiently infect the α4⁺β7^hiCD4⁺ T-cell subsets very early during the course of viral infection. Development of therapeutic strategies targeted at preventing the early infection of memory CD4⁺ T-cell subsets that express high levels of α4β7 heterodimer may significantly aid in the early control of viral infection and contribute to the maintenance of homeostasis in mucosal tissues.

METHODS

Animals and infection. Mamu*A01^neg rhesus macaques (Macaca mulatta) of Indian origin (n=8) housed at Advanced Bioscience Laboratories (MD) were used in this study. The animals were housed in accordance with the American Association for Accreditation of Laboratory Animal Care guidelines and were sero-negative for SIV, simian retrovirus, and simian T-cell leukemia virus type-1. All the animals were infected with 100 animal-infectious doses of uncloned pathogenic SIVmac251 intravenously. Peripheral blood was collected at different time points (days 0, 2, 4, 10, and 63) after challenge, and complete blood count data were obtained at days 0 and 63 pi. No complete blood count data were collected at other time points; hence, absolute counts for various subsets could only be determined at days 0 and 63 pi. In addition, peripheral blood samples were also obtained from 14 SIVmac251-infected rhesus macaques at days 2–3 (n=4) and 7–14 pi (n=10), 77 (n=8), and from uninfected animals (n=6–8).

Tissue samples. Peripheral blood mononuclear cells (PBMCs) were isolated by density gradient centrifugation, as previously described.⁵ Plasma viral loads were determined by real-time PCR (ABI Prism 7700 Sequence Detection System, Applied Biosystems, Foster City, CA) using reverse-transcribed viral RNA as templates using methods, as previously described³⁵ (limit of detection <30 copies per ml of plasma).

Antibodies and flow cytometry. All antibodies, except for CD103, used in this study were obtained from BD Biosciences (San Diego, CA), and titrated using rhesus macaque PBMC. For phenotypic analysis and sorting of CD4⁺ T-cell subsets, cells were labeled simultaneously with the following combinations of antibodies: CD3-Cy7APC, CD8-Alexa-700, CD4-APC, CD95-FITC, CCR5-PE, HLA-DR–Texas Red-PE, and Integrin β7 (FIB504 clone)-Cy5-PE. FIB504 clone has been shown to cross-react with rhesus macaques.^{10, 21} β7 Integrin forms a heterodimer with α4 (CD49d) or αE (CD103) integrins, and has a central role in homing of T cells to mucosal tissues.^{36, 37, 38, 39} To delineate α4⁺β7⁺CD4⁺ T-cell subsets from αEβ7⁺ CD4⁺ T-cell subsets, cells were labeled with CD3-Cy7APC, CD8-Alexa-700, CD4-Qdot-605, CD95-APC, CD49d-PE, CD103-FITC (clone 2G5 from Beckman Coulter Fullerton, CA), and Integrin β7-Cy5-PE. Clone 2G5 has been shown to cross-react with rhesus macaque mucosal homing T cells, and expressed by intraepithelial lymphocytes.^{40, 41} To evaluate the phenotype of α4β7 subsets of CD4⁺ T cells, the cells were labeled with a panel of CD3-Cy7APC, CD8-Qdot-605, CD4-APC, CD95-FITC, CD49d-PE and β7-Cy5-PE, CD45RA-TR-PE, CD28-Cy7-PE, and CCR7-Alexa-680. The expression of activation markers was determined by using a panel of CD3-Cy7APC, CD8-Alexa-700, CD4-TR-PE, CD95-FITC, CD49d-PE, β7-Cy5-PE, CD69, or CD25-APC.

To determine which α4β7 subsets were preferentially infected, α4β7 subsets of CD4⁺ T cells (discriminated based on the expression of β7 and CD95²¹) were sorted, and subjected to qPCR assay for measuring SIV-gag DNA, as described previously.^{5, 15} To determine whether CD4⁺ T cells that expressed the α4β7 heterodimer were preferentially infected, PBMCs were labeled with CD3-Cy7APC, CD8-Alexa-700, CD4-APC, CD95-FITC, CD49d-PE, and β7-Cy5-PE. In addition, KI-67 expression was evaluated using a panel of CD3-Cy7APC, CD8-Alexa-700, CD4-APC, CD95-PE, and β7-Cy5-PE. After surface labeling, the cells were fixed/permeabilized using the Cytofix/Perm Kit (BD Biosciences) and labeled with Ki-67-FITC.

To evaluate the ability of α4β7 subsets of CD4⁺ T cells to produce IL-17 and IFNγ, PBMCs were stimulated with phorbol myristate acetate at 10 ng ml⁻¹ and ionomycin at 500 ng ml⁻¹ in the presence of Brefeldin-A (1 μM) for 4 h. The cells were harvested and labeled with CD3-Cy7APC, CD8-Alexa-700, CD4-Qdot-605, CD95-APC, and β7-Cy5-PE. After surface labeling, the cells were fixed/permeabilized using the Cytofix/Perm Kit and labeled with IL-17-PE (e-Biosciences, San Diego, CA) and IFNγ-FITC (BD Biosciences).

Labeled cells were fixed in 0.5% paraformaldehyde and analyzed using a modified Becton Dickinson Aria sorter.

qPCR assay. T-cell-associated viral DNA was measured by a qPCR assay for SIV gag using a Perkin-Elmer ABI 7500 instrument, as previously described^{5, 15} using SIV gag primers and probe as described by Lifson et al.⁴² Cell numbers were quantified simultaneously using rhesus macaque albumin-specific primers and probe using previously described protocols.⁵ The assay was calibrated using a cell line that carried a single copy of proviral SIV DNA as shown previously.⁵ Sorted samples were lysed in Proteinase-K and used for PCR analysis. We have previously shown that this assay was highly sensitive and can be used to successfully measure SIV-DNA in single cells.⁵

CCR5 mRNA levels were determined using primer/probes as described previously,⁵ and normalized to β2M levels in sorted subsets. Data are shown relative to α4⁺β7^intCD4⁺ (naive) T cells.

Data analysis. Flow cytometric data were analyzed using FlowJo version 8.6 (Tree Star, Ashland, OR). Statistical analysis was carried out using GraphPad Prism Version 4.0 software (GraphPad Prism Software, San Diego, CA) using non-parametric tests. As the data did not exhibit normal distribution due to small sample size, comparisons between groups were carried out using Mann–Whitney U-test, and multiple comparisons between groups was carried out using the Kruskal–Wallis one way analysis of variance (ANOVA). P<0.05 was considered significant.