Lack of Compartmentalization of HIV-1 Quasispecies Between the Gut and Peripheral Blood Compartments

Abstract

Compartmental differences in human immunodeficiency virus type 1 (HIV-1) between the gut and peripheral blood and within the gut were examined. Biopsy specimens from the colon and ileum and peripheral blood samples were collected from chronically HIV-1–infected individuals. HIV-1 envelope sequences were examined from cell-associated DNA and RNA and virion RNA. Phylogenetic analysis revealed no evidence of compartmentalization of HIV-1 between the gut and peripheral blood and within the gut (colon and ileum). HIV-1 sequences detected in the gut were transcriptionally active and were also found in peripheral blood from matching time points, providing evidence of ongoing virus production in the gut and equilibrium of HIV-1 between the gut and peripheral blood compartments.

The gut mucosa is a prominent part of the mucosa-associated lymphoid tissues in the body [1]. It is in continuous contact with a great number of food and microbial antigens that enter the host through the intestine. Because of its proximity to the external environment, which results in a constant state of physiological inflammation characterized by intrinsically high levels of proinflammatory and human immunodeficiency virus (HIV)–stimulatory cytokines [2–4], the immune response in the gut can be independently regulated and distinct from the systemic immune response [5, 6]. The presence of baseline physiologic inflammation, the abundance of CD4⁺ T cells that express CCR5 chemokine receptor [7–9], and the marked enrichment of CD4⁺ T cells with an activated, memory phenotype render the gut mucosa extremely permissive to HIV-1 infection and supportive of HIV-1 replication. Our understanding of the contribution of the gut mucosa to sustaining HIV-1 replication and dissemination is still largely limited, despite the recognition of its importance in the HIV-1 pathogenesis from the early days of the HIV-1 infection epidemic [10–12]. In an effort to address the potential role that the gut mucosa plays as a viral reservoir, studies were undertaken to compare HIV-1 sequences derived from 2 different sites of the gut (colon and terminal ileum) and the peripheral blood of patients with chronic HIV-1 infection.

METHODS

A total of 20 HIV-infected individuals were studied. The study participants were chronically HIV-infected with plasma HIV RNA levels of <50 copies/mL for >1 year (group 1; n = 10) and with plasma HIV RNA levels of ≥50 copies/mL (group 2; n = 10) (Table 1). All specimens were obtained in accordance with protocols approved by the institutional review board of the National Institute of Allergy and Infectious Diseases, National Institutes of Health.

Table 1.

Clinical Characteristics of Study Participants With Chronic HIV-1 Infection

Subject ID	HIV RNA level in plasma, copies/mL	Total CD4+ cell count, cells/μL	Duration of pVL of <50 copies/mL prior to biopsy, years	Estimated duration of infection, years
					No. of HIV-1 sequences obtained
					Colon	Colon	Ileum	Ileum	PBMCs	PBMCs	Plasma
					DNA	RNA	DNA	RNA	DNA	RNA	RNA	Antiretoviral regimen
Group 1 (n = 10)
1	<50	456	6.7	7.4	10	NS	8	NS	7	8	NS	Atazanavir, lamivudine, zidovudine
2	<50	701	3.9	14.2	8	2	7	6	8	2	NS	Lamivudine, nevirapine, stavudine
3	<50	357	1.2	17.6	11	NS	11	1	12	3	NS	Abcavir, atazanavir, lamivudine, ritonavir
4	<50	818	2.8	13.6	4	NS	9	NS	9	2	NS	Abcavir, atazanavir, emtricitabine, ritonavir
5	<50	570	5.9	8.6	8	NS	8	NS	NS	NS	NS	Abcavir, lamivudine, nevirapine
6	<50	971	1.0	16.1	2	6	2	NS	ND	ND	ND	Atazanavir, didanosine, lamivudine, ritonavir, zidovudine
7	<50	478	1.2	5.0	10	NS	8	NS	10	NS	NS	No ART
8	<50	539	3.8	15.1	10	2	10	4	6	9	NS	Abcavir, lamivudine, nelfinavir, zidovudine
9	<50	1390	1.0	11.9	10	3	11	NS	9	11	NS	Abcavir, efavirenz, tenofovir disoproxil fumerate
10	<50	698	2.6	23.3	7	NS	1	NS	7	5	2	Didanosine, lamivudine, lopinavir, ritonavir, zidovudine
Median (IQR)	<50 (NA)	634 (493–789)	2.7 (1.2–3.9)	13.9 (9.4–15.9)
Group 2 (n = 10)
11	30,141	509	NA	8.6	12	4	ND	ND	12	17	15	No ART
12	12,540	341	NA	5.0	7	6	ND	ND	9	6	NS	Didanosine, lamivudine, nelfinavir, zalcitabine
13	2082	198	NA	19.0	7	9	2	10	10	7	10	No ART
14	2577	774	NA	10.4	8	8	NS	7	NS	3	2	No ART
15	18,010	381	NA	4.9	9	9	10	10	11	8	18	No ART
16	83,187	181	NA	17.9	9	11	8	9	5	7	16	Emtricitabine, fosamprenavir, ritonavir, tenofovir disoproxil fumerate
17	7052	169	NA	8.5	10	7	11	12	12	4	3	Atazanavir, emtricitabine, ritonavir, tenofovir disoproxil fumerate, zidovudine
18	13,994	36	NA	0.1	10	9	ND	ND	ND	ND	ND	No ART
19	204,202	101	NA	1.1	9	9	10	NS	ND	ND	ND	No ART
20	347,674	204	NA	6.3	6	10	3	7	10	NS	13	No ART
Median (IQR)	16,002 (8424–69,926)	201 (172–371)	NA	7.4 (4.9–9.9)

NOTE. Patients in group 1 had human immunodeficiency virus (HIV) RNA levels of <50 copies/mL, and patients in group 2 had HIV RNA levels of ≥50 copies/mL. The duration of infection is an estimate of the time between the date of the first positive HIV-1 test and the sampling time point. Serial colonic biopsies were performed for subjects 18 and 19 at months 0, 3, and 12. ART, antiretroviral therapy; IQR, interquartile range; NA, not applicable; ND, not determined because of limited sample availability; NS, amplification of HIV was not successful; PBMC, peripheral blood mononuclear cell; pVL, plasma viral load.

Biopsy samples from the colon and terminal ileum were weighed and homogenized in RLT buffer (RNeasy mini kit; Qiagen) for HIV RNA or in cell lysis buffer (Gentra Puregene tissue kit; Qiagen) for HIV DNA, using a mini mortar and pestle. The Siemens Quantiplex HIV-1 3.0 assay was used to measure HIV RNA copy number. Levels of HIV DNA in the colonic and terminal ileum tissues as well as peripheral blood mononuclear cells (PBMCs) were measured by quantitative real-time polymerase chain reaction (PCR) as described elsewhere [13]. The detection limit of the assay was 10 copies of HIV DNA per PCR reaction. DNA and RNA results are expressed as copies per milligram of tissue or copies per 1 × 10⁶ cells.

Extraction of viral RNA from plasma, complementary DNA synthesis, and nested PCR amplification of the HIV-1 gene were performed as described elsewhere [14]. Single molecules of a 425-bp fragment, encompassing the C2-V3 region of the HIV-1 envelope gene, obtained through limiting dilution, were PCR-amplified and cloned into the pCR2.1-TOPO vector (Invitrogen) for sequence analysis of individual clones. The fragments of the C2-V3 region of HIV-1 env gene were sequenced from cell-associated DNA, cell-associated RNA, and cell-free virion RNA. On average, 8 independent PCR clones per sample (range, 2–18) and a total of 42 sequences per subject (range, 10–75) were analyzed.

Phylogenetic relationships among the HIV-1 envelope sequences were estimated, by use of the neighbor-joining method, with the PAUP* program (version 4.0b10; Sinauer Associates). Statistical support for various nodes in the neighbor-joining tree was obtained by 1000 replications of the bootstrap procedure. The level of diversity of the HIV envelope sequences was estimated, by use of uncorrected average pairwise distances, with the MEGA program [15], using a pairwise-deletion of gap option. All sequences were derived from independent PCR reactions; they have been deposited in GenBank and are available under the accession numbers HQ451076–HQ451869.

Statistical analysis was performed with WinStat (R. Fitch Software). Significant differences in comparing 2 groups were determined by using the 2-tailed unpaired nonparametric Mann-Whitney U test.

RESULTS

A total of 20 patients with chronic HIV infection (19 men and 1 woman; median age, 50 years) were divided into 2 groups on the basis of plasma viral load at the time of the gut biopsy (Table 1). Group 1 included 10 patients who had been receiving suppressive antiretroviral therapy (ART) for >1 year prior to the gut biopsy (except subject 8, who was not receiving ART at the time of the study). HIV DNA was present in all gut and PBMC samples analyzed for the study, except 1 ileum sample from subject 14. By use of a nested PCR technique, HIV RNA could be detected in the majority of samples (94% of the gut samples and 88% of the PBMC samples) from the patients in group 2. However, detection of HIV RNA was restricted to 20%–40% of the gut and 80% of the PBMC samples in the patients in group 1. We found no association between the ability to detect HIV RNA by PCR in the gut (colon and ileum) and the duration of suppression of plasma viremia to levels of <50 copies/mL.

The levels of HIV DNA in the gut (colon and ileum) and PBMCs were similar in groups 1 and 2 (median HIV DNA level in the gut in group 1 vs group 2, 71 vs 105 copies/mg; P = .405; median HIV DNA level in PBMCs in group 1 vs group 2, 157 vs 189 copies/10⁶ cells; P = .563). Within the gut, no regional difference was observed in the levels of HIV DNA and HIV RNA when comparing the colon with the ileum (P > .2). An exception was a trend seen in HIV DNA level in group 1, where there was ∼10 times more HIV DNA on average in the ileum than in the colon (HIV DNA level in the ileum vs the colon, 92 vs 13 copies/mg). However, the difference did not reach statistical significance (P = .267).

To better understand the relationship between HIV-1 circulating in the blood and HIV-1 present in gut mucosa, we sequenced a region of the HIV-1 envelope gene, using both DNA and RNA as template, from the gut and peripheral blood samples from both groups of patients (Figure 1). All HIV-1 sequences obtained from an individual patient formed a monophyletic group in the phylogenetic tree and were intermixed and distributed throughout the main trunk of the phylogenetic tree for the given individual. Analysis of HIV-1 sequences from viremic patients in group 2 permitted a detailed assessment of HIV-1 sequence comparison between the gut and peripheral blood. In doing so, we were able to demonstrate that HIV-1 variants detected in the gut were genetically indistinguishable from those in the plasma from matching time points (Figure 1B). Furthermore, HIV-1 sequences detected in the colon and ileum were genetically indistinguishable, which indicates that there is no compartmentalization of HIV-1 within the gut (colon and ileum).

Figure 1.

Phylogenetic analysis of the human immunodeficiency virus type 1 (HIV-1) envelope sequences obtained from gut and peripheral blood compartments. Two representative patterns are shown for group 1 (patients with HIV RNA levels of <50 copies/mL) (A) and group 2 (patients with HIV RNA levels of ≥50 copies/mL) (B ). C, Phylogenetic analysis of HIV for the 2 subjects who underwent serial colonic biopsies prior to and 3 and 12 months after starting highly active antiretroviral therapy. Phylogenetic relationships among the C2-V3 region of the env gene were estimated using the neighbor-joining method. All the sequences presented in the phylogenetic trees come from independent polymerase chain reaction (PCR) runs that utilized samples obtained by limiting dilution. Bootstrap percentile values from 1,000 replications are shown at nodes defining major groupings of sequences. Statistical support of ≥50% is shown. The charts shown depict the color and shape codes used to identify different genetic materials derived from the colon, terminal ileum, peripheral blood mononuclear cells (PBMCs), and plasma (A, B ) or different sampling time points (C ). Gray symbols in the charts indicate that PCR amplification of the HIV-1 env gene was unsuccessful. An additional phylogenetic tree analysis, which involved the envelope sequences obtained from all 10 patients, revealed that isolates from each patient were well confined within distinct patient-specific clusters that were divergent from each other and from known laboratory strains of HIV-1, confirming the absence of cross-patient contamination (data not shown). HIV-1 U455 was used as an outgroup.

Sequential sampling of HIV DNA sequences from the colon was available for 2 patients in group 2 from whom biopsy specimens were obtained prior to and 3 and 12 months after the successful initiation of highly active ART (HAART) (Figure 1C). Despite the substantial decrease in plasma viremia seen in these 2 patients during the sampling period, HIV DNA sequences continued to be detected in colon biopsy specimens. The diverse provirus population present at baseline (0 months) persisted throughout the 12-month sampling period. This is illustrated by the phylogenetic trees in Figure 1C, where colon-derived HIV DNA sequences are randomly distributed irrespective of sampling time points. In addition, there is no indication of evolution of HIV-1 in the gut while these patients were receiving suppressive ART.

In line with the earlier results obtained from the phylogenetic analyses, the average within-sample nucleotide variations were similar between the gut and the peripheral blood compartments. The genetic diversity of HIV DNA was 4.09% and 6.57% (P = .222) for the gut and peripheral blood, respectively, in group 1; and it was 2.21% and 3.75% (P = .061) for the gut and peripheral blood, respectively, in group 2. The genetic diversity of HIV RNA was 2.42% and 3.57% (P = .350) for the gut and peripheral blood, respectively, in group 1; and it was 2.19% and 3.39% (P = .071) for the gut and peripheral blood, respectively, in group 2. This was also true for the genetic diversity within the gut: The genetic diversity of HIV DNA was 4.82% and 3.13% (P = .369) for the colon and ileum, respectively, in group 1; and it was 2.21% and 1.95% (P > .999) for the colon and ileum, respectively, in group 2. The genetic diversity of HIV RNA was 2.42% and 2.12% (P = .814) for the colon and ileum, respectively, in group 1; and it was 2.05% and 2.28% (P = .914) for the colon and ileum, respectively, in group 2. These results further support a notion of equilibrium of HIV-1 between the gut and peripheral blood compartments.

DISCUSSION

In the present study, through sequence analysis of the env C2-V3 region, we clearly demonstrated a lack of compartmentalization of HIV-1 quasispecies between the gut and peripheral blood. HIV-1 sequences detected in the gut compartment were genetically indistinguishable from those present in PBMCs and plasma from matching time points, which indicates equilibrium of HIV-1 quasispecies between the gut and peripheral blood. We have also found that HIV-1 variants detected in the gut compartment were present in both proviral DNA and cell-associated RNA forms and that these transcriptionally active forms of HIV-1 in the gut were also found in PBMCs and plasma from matching time points.

Different regions of the HIV genome display different degrees of variability due to virus adaptation to strong selective pressures (eg, immune surveillance or drug-selective pressure). Therefore, further studies may be necessary to definitely address questions regarding (1) the compartmentalization between the gut and peripheral blood, perhaps by analyzing other parts of the HIV genome in addition to the C2-V3 region of the env gene, and (2) the archival nature of HIV in the gut by performing a longitudinal follow-up study. In addition, due to the limited number of sequences available for some samples, it is possible that unsampled HIV variants with low frequency might have existed in the gut.

Given the unique immunological environment including a preponderance of CCR5⁺CD4⁺ T cells in the gut mucosa [7–9], one could hypothesize that selective pressures differ between the gut and peripheral blood compartments and that, as a consequence, the gut mucosa harbors different quasispecies than those in peripheral blood. However, our data demonstrated that distribution patterns of HIV-1 sequences were similar between colon and ileal regions of the gut and PBMCs. These data indicate that the gut mucosa does not appear to serve as a sanctuary site for HIV-1 replication and that, rather, free exchange of HIV-infected cells takes place between the gut and peripheral blood compartments. Our current findings further extend the observation made by Chun et al [16], who found evidence of cross-infection between the gut and blood in HIV-infected individuals with plasma HIV RNA levels of <50 copies/mL. However, due to technical limitations of the gut biopsy process, we cannot completely exclude the possibility that contamination of the biopsy specimens with small amounts of blood—the limitation common to all studies of this type—might have played a role in the presence of similar HIV-1 variants in the gut.

Sequential sampling of HIV DNA sequences from the colonic region of the gut was available from 2 HIV-infected individuals. In these patients, we detected the persistence of HIV DNA in the gut (colon) compartment even after prolonged suppression of plasma viremia was achieved for up to 1 year; notably, we found no demonstrable sequence evolution. One could argue that the HIV DNA detected in the gut mostly represents replication-defective variants. Alternatively, it is possible that long-lived HIV-infected cells exist in the various compartments and these cells were producing HIV-1 but not propagating infection in the local areas.

Given the importance of the mucosal compartment in HIV-1 pathogenesis along with the observation of the persistence of HIV-1 in the gut and the peripheral blood of patients receiving suppressive ART, our findings may have important implications with regard to treatment. Insights into the replenishment and maintenance of HIV-1 replication in the gut will help the development of novel therapeutic strategies, in conjunction with current HAART regimens, aimed at lowering of the viral set point and conceivably eliminating the low levels of viral replication that potentially takes place in the gut.

Funding

This work was supported by the National Cancer Institute, National Institutes of Health (NIH; contract HHSN261200800001E); the National Institute of Allergy and Infectious Disease (NIAID), NIH (contract HHSN261200800001E); and the Intramural Research Program of NIAID and the NIH Clinical Center.