CD4⁺ T Cell Depletion in an Untreated HIV Type 1-Infected Human Leukocyte Antigen-B*5801-Positive Patient with an Undetectable Viral Load

Abstract

We report a case of a patient infected with human immunodeficiency virus type 1 (HIV-1) for 20 years who has experienced CD4⁺ T cell depletion in spite of maintaining undetectable viral loads. Our data suggest that immune activation can cause CD4⁺ T cell depletion even when HIV-1 replication appears to be controlled by host factors.

Long-term nonprogressors (LTNPs) are HIV-1-infected patients who maintain normal CD4⁺ T cell counts without antiretroviral therapy. A subset of these patients, termed “elite suppressors,” have undetectable viral loads. The HLA-B*57 allele is overrepresented in elite suppressors [1], and the closely related HLA-B*5801 allele is associated with lower viral loads in untreated patients [2], suggesting that CD8⁺ T cells play an important role in the control of viral replication. Here, we describe an HLA-B*5801-positive male patient who received a serodiagnosis of HIV-1 infection in 1987 who has maintained an undetectable viral load, despite not being treated with antiretrovirals since 2000. In an effort to understand the mechanisms involved in his CD4⁺ T cell depletion, we performed phenotypic analysis of his T lymphocytes, and characterized his HIV-1 isolate. The results have implications for the treatment of patients with similar clinical courses.

Patient, materials, and methods

The patient’s clinical course over the past 10 years is outlined in figure 1. He tested positive for HIV in 1987 but did not present for care until 1995, at a time when quantitative viral load testing was not available. His initial regimen consisted of zidovudine monotherapy, and he was later administered regimens that included didanosine, lamivudine, and ritonavir. He discontinued each regimen within 1 month of starting therapy and has not been taking any antiretrovirals since 2000. Of 12 ultrasensitive HIV-1 RNA assays completed since 2000, his viral load was <50 copies/mL in 9 independent measurements. He had HIV-RNA levels of 155, 111, and 84 copies/mL in 2001, 2002, and 2006 respectively. Viral loads obtained 3 months after the positive values in 2001 and 2006 were <50 copies/mL, suggesting that these viral load peaks lasted for short periods of time.

Figure 1.

CD4⁺ T cell count and viral loads of the study subject over the past 10 years. Standard viral load testing with a limit of detection of 400 copies/mL was performed until 2000. Ultrasensitive viral load testing (limit of detection, 50 copies/mL) was performed after that time point. The unshaded points listed at 400 or 50 copies/mL were less than the limit of detection of the respective assays. The boxes above the graph denote the medications the patient was prescribed at a given time. He reports very poor adherence with all his medications. AZT, zidovudine; ddI, didanosine; RTV, ritonavir; 3TC, lamivudine.

An HIV-1 isolate was obtained from 12 million resting CD4⁺ T cells and plasma as previously described [3]. The nef gene was amplified and sequenced as described elsewhere [3]. The protocol was approved by the Johns Hopkins University School of Medicine Institutional Review Board (Baltimore, MD). Informed consent was obtained before phlebotomy.

HLA typing was performed as described elsewhere [4]. The patient’s genotype was A*33, A*66, B*5301, and B*5801.

Cryopreserved PBMCs were thawed and stained with fluorescent antibodies, and 500,000 events were analyzed on a Facscan machine (BD). Quantitative assays for lipopolysaccharide (LPS), LPS binding protein (LBP), soluble CD14, and endotoxin-core antibody (EndoCAb) were all performed as recently described [5].

Enzyme-linked immunosorbent spot analysis was performed as previously described [4]. Overlapping peptides representing the clade B HIV-1 proteome were obtained from the National Institutes of Health AIDS Research and Reference Reagent Program repository (Rockville, MD). A positive response was defined as a mean value that was >50 spot-forming cells per 10⁶ PBMCs. Unstimulated PBMCs routinely had <10 spot-forming cells per 10⁶ PBMCs.

Results

A single HIV-1 isolate was cultured from 12 million CD4⁺ T cells. It is possible that the inability to detect virus in the patient’s plasma was due to infection with an isolate with mutations in primer binding sites that interfered with amplification of the virus during PCR. The Roche 1.5 assay was thus simultaneously performed on plasma and culture supernatant from the isolate obtained from the latent reservoir. A total of 316,000 HIV RNA copies/mL were amplified from the culture supernatant, whereas the plasma viral load was undetectable. This demonstrates that the isolate was capable of replicating vigorously in vitro and could be efficiently amplified by the Roche 1.5 assay. However, we cannot rule out the presence of a second strain with mutations in the primer binding sites. Sequence analysis of the nef gene from the replication-competent isolate determined that it was a clade B virus. No deletions or insertions were present in this gene (data not shown; GenBank accession number EU410053).

Although the magnitude of an individual’s viral load usually determines the rate of progression of HIV-1 disease [6], the degree of immune activation has been shown to correlate even better with disease progression [7, 8]. Thus, we compared HLA-DR and CD38 expression on CD4⁺ and CD8⁺ T lymphocytes from this patient with lymphocytes from 10 elite suppressors with CD4⁺ T cell counts >600 cells/μL and a ratio of CD4⁺ to CD8⁺ cells >1.0. The 2 activation markers were coexpressed on 24% of the patient’s CD8⁺ T cells and 3.6% of his CD4⁺ T cells. In contrast, both activation markers were present on a mean of just 3.8% of CD8⁺ T cells (range, 1.2%-6.9%) and 0.6% of CD4⁺ T cells (range, 0.2%-1.6%) in elite suppressors with normal CD4⁺ T cell counts (figure 2A and B). Interestingly, HIV-1-negative control subjects had significantly lower frequencies of activated CD8⁺ and CD4⁺ T cells than elite suppressors. This is consistent with a recent report on CD8⁺ T cell activation in elite suppressor patients [9].

Figure 2.

Comparison of activation markers on CD8⁺ T cells (A) and CD4⁺ T cells (B) isolated from HIV-1-seronegative controls, HIV-1-infected elite suppressors (ES) and the study subject. A 2 sided t-test was used to compare the values obtained from HIV-seronegative controls and ES. Levels of plasma lipopolysaccharide (LPS; C), LPS-binding protein (LBP; D), soluble CD14 (sCD14; E) and endotoxin-core antibody (EndoCAb; F) in ES and the study subject. The horizontal lines depict the median values of the 10 ES studied.

Immune activation has recently been linked to microbial translocation in HIV-1-infected patients [5]. We thus measured LPS levels, LBP levels, and titers of endoCAb, the antibody that binds to LPS and soluble CD14, a protein secreted by monocytes in response to LPS stimulation. Brenchley et al. [5] found significantly higher levels of LPS, LBP and soluble CD14 in viremic HIV-1-infected patients with CD4⁺ T cell depletion, compared with elite suppressors. In agreement with that report, we found very low levels of these molecules in our cohort of elite suppressors, but, surprisingly, the patient with CD4⁺ T cell depletion also had low levels of all 3 markers (figure 2C-E). The endoCAb titer in this patient was also comparable to that seen in elite suppressors with normal CD4⁺ T cell counts (figure 2F). In contrast, we have observed higher levels of these markers of microbial translocation in HIV-1-infected patients with CD4⁺ counts of <350 cells/μL, compared with HIV-1-seronegative subjects (data not shown). Thus, in this patient, mechanisms other than microbial translocation may have been responsible for the observed immune activation.

CD8⁺ T lymphocytes play an important role in the control of viremia in elite suppressors. Twenty-four percent of this patient’s CD8⁺ T cells had an activated phenotype. Thus, we looked at the breadth and magnitude of the HIV-1-specific immune responses. The patient had a dramatic IFN-γ response to Gag peptides, and significant responses were seen to 76 of the 123 overlapping peptides tested (data not shown) (table 1). The patient also had immune responses to multiple, nonoverlapping epitopes in Pol, Env, Vpr, Nef, and Vif proteins (table 1).

Table 1.

Results of whole proteome IFN-γ enzyme-linked immunosorbent spot analysis. The frequency of spot-forming cells (SFCs) produced in response to each epitope is shown

Peptide	Protein	Amino acid range	SFCs/106 PBMCs
MGARASVLSGGELDR	Gag	1-15	65
LDRWEKIRLRPGGKK	Gag	13-27	230
HIVWASRELERFAVN	Gag	33-47	1920
GLLETSEGCRQILGQ	Gag	49-63	225
TVATLYCVHQRIEVK	Gag	81-95	120
LEKIEEEQNKSKKKA	Gag	101-115	145
DTGNSSQVSQNYPIV	Gag	121-135	105
QMVHQAISPRTLNAW	Gag	141-154	1265
EKAFSPEVIPMFSAL	Gag	162-175	55
EGATPQDLNTMLNTV	Gag	177-191	105
NTVGGHQAAMQMLKE	Gag	189-203	65
LKETINEEAAEWDRL	Gag	201-215	950
DRLHPVHAGPIAPGQ	Gag	213-227	175
REPRGSDIAGTTSTL	Gag	229-243	800
EQIGWMTNNPPIPVG	Gag	245-259	1900
PVGEIYKRWIILGLN	Gag	257-271	175
GLNKIVRMYSPTSIL	Gag	269-283	120
SILDIRQGPKEPFRD	Gag	281-295	2405
FRDYVDRFYKTLRAE	Gag	293-307	1120
RAEQASQEVKNWMTE	Gag	305-319	2405
MTETLLVQNANPDCK	Gag	317-331	160
ILKALGPAATLEEMM	Gag	333-347	240
EMMTACQGVGGPGHK	Gag	345-359	135
RVLAEAMSQVTNSAT	Gag	361-375	1070
SATIMMQRGNFRNQR	Gag	373-387	225
FNCGKEGHIAKNCRA	Gag	393-404	610
DCTERQANFLGKIWP	Gag	425-439	185
IWPSHKGRPGNFLQS	Gag	437-451	105
APPEESFRFGEETTT	Gag	457-471	55
KIATESIVIWGKTPK	Reverse transcriptase	374-388	155
KLPIQKETWEAWWTE	Reverse transcriptase	390-404	60
KVYLAWVPAHKGIGG	Reverse transcriptase	530-544	60
PAETGQETAYFLLKL	Integrase	90-104	85
MNKELKKIIGQVRDQ	Integrase	154-168	150
EHLKTAVQMAVFIHN	Integrase	170-184	125
WKSLVKHHMYISGKA	Vif	21-35	215
EAVRHFPRIWLHSLG	Vpr	29-43	75
HIYETYGDTWAGVEA	Vpr	45-59	85
VKIEPLGVAPTKAKR	Env	489-503	110
NNLLRAIEAQQHLLQ	Env	553-567	65
NMTWMEWEREIDNYT	Env	625-639	115
IVTRIVELLGRRGWE	Env	777-791	50
WVYHTQGYFPDWQNY	Nef	113-127	135

Discussion

The magnitude of viral load has been shown to correlate with the rate of progression of HIV-1 disease [6]. This patient was highly unusual in that he had progressive CD4⁺ T cell depletion while he maintained an undetectable viral load. A similar case of gradual CD4⁺ T cell depletion in spite of an undetectable viral load has been reported in a patient who was infected with an isolate that contained a large deletion in the nef gene [10]. A similar slow decrease in CD4⁺ T cell counts despite low viral loads was observed in some patients in the Sydney Blood Bank Cohort who were also infected with isolates containing a defective nef gene [11]. In contrast to those cases, no deletions or insertions were present in nef in this isolate, and vigorous replication was observed in vitro.

Furthermore, although it is possible that patients with undetectable viral loads and CD4⁺ T cell depletion are infected with isolates that are not amplified by the Roche 1.5 assay, we isolated replication-competent virus from the latent reservoir and showed that it could be effectively amplified by the commercial assay.

A recent study has suggested that microbial translocation may be responsible for immune activation in chronically HIV-1-infected patients with progressive disease but not in patients whose viral replication is controlled without antiretroviral therapy [5]. Given that the low levels of LPS, LBP and soluble CD14 in this patient were comparable to those in elite suppressors with normal CD4⁺ T cell counts in this study, it is unlikely that microbial translocation was the cause of immune activation and CD4⁺ T cell depletion. Other chronic infections or medical conditions [12] may have played a role in CD4⁺ T cell depletion at different time points. The patient was also infected with hepatitis C virus, but this alone is unlikely to explain his extensive immune activation and CD4⁺ T cell depletion, because 6 of the 10 elite suppressors with normal CD4⁺ T cell counts were also coinfected with the hepatitis C virus.

It is noteworthy that the patient was HLA-B*5801 positive, because this allele is closely related to the protective HLA-B*57 allele and has been associated with slow progression of disease [2]. In contrast to HLA-B*5701-[1] and HLA-B*5703-positive [4] elite suppressors who target few epitopes in Gag, this patient had a very broad IFN-γ response and targeted at least 29 distinct epitopes in Gag alone, each of which overlapped by no more than 3 amino acids. Immune responses were also detected to epitopes in Pol, Env, and Nef proteins. The total of 44 targeted epitopes targeted is higher than that seen in any patient in a recent study using the same techniques [13].

It is possible that, although this broad immune response led to the control of viral replication, it also may have precipitated extensive immune activation, which is an independent correlate of disease progression. The level of CD4⁺ and CD8⁺ T cell activation in this patient was much higher then that seen in 10 elite suppressors who maintained normal CD4⁺ T cell counts. Viral replication in sequestered regions such as the lymph nodes or the gastrointestinal tract is another possible explanation for the patient’s CD4⁺ T cell depletion. This replication could have triggered the strong HIV-specific immune response observed in the patient and partially controlled the virus, resulting in the absence of peripheral viremia. However, this phenomenon has never been described in an untreated HIV-1-infected patient, and it has recently been shown that some LTNPs maintain normal CD4⁺ T cell counts in spite of having high viral loads [14]. These LTNPs have low levels of T cell activation, which emphasizes the important role this process plays in disease progression [7, 8].

The results suggest that some untreated patients with undetectable viral loads may benefit from HAART. It would obviously be beneficial if the decrease in the CD4⁺ T cell count was directly due to extensive viral replication in a sequestered site. It could also be effective if the CD4⁺ T cell decrease was indirectly related to low-level viral replication. By decreasing the amount of antigen—and, thus, the HIV-1-specific IFN-γ effector response [15]—treatment with HAART could potentially result in diminished immune activation and reduced CD4⁺ T cell depletion.