Susceptibility to opportunistic infections in HIV-infected patients with increased CD4 T-cell counts on antiretroviral therapy may be predicted by markers of dysfunctional effector memory CD4 T cells and B cells

Abstract

Objectives

HIV-infected patients responding to combination antiretroviral therapy (ART) after experiencing severe immunodeficiency may exhibit persistent immune defects and occasionally experience opportunistic infections (OIs) despite increased CD4 T-cell counts. The investigation of immune defects in such patients was examined in this study.

Methods

CD4 effector memory T-cell (T_em-cell) function [assessed by blood cytomegalovirus (CMV) interferon-γ (IFN-γ) enzyme-linked immunosorbent spot-forming cell assay (ELISPOT) counts] and B-cell dysregulation [assessed by serum immunoglobulin A (IgA) and IgE levels] were examined in 27 patients with increased CD4 T-cell counts after receiving ART for over 2 years. Two of these patients and one other had developed OIs on ART and are described in detail.

Results

Serum levels of IgA and IgE were higher than reference intervals (P<0.001) and CMV IFN-γ ELISPOT counts were lower than those in non-HIV-infected controls (P<0.001) in the HIV-infected patients. Low CMV IFN-γ ELISPOT counts were associated with high IgA levels (r=−0.5, P=0.01, Spearman's correlation test) and segregated with high IgE levels (P=0.06, Fisher's test). CMV IFN-γ ELISPOT counts and serum IgA and IgE levels did not change significantly over a median time of 35 (range 8–60) months after the first measurement, whereas CD4 T-cell counts increased. All three patients who experienced OIs had repeatedly low CMV IFN-γ ELISPOT counts and increased serum levels of IgA and/or IgE.

Conclusion

Low CD4 T_em-cell function and B-cell dysregulation are immune defects that may persist independently of changes in the CD4 T-cell count in HIV-1-infected patients responding to ART and are associated with an increased risk of developing an OI.

Introduction

HIV-1-infected patients receiving combination antiretroviral therapy (ART) who had a nadir CD4 T-cell count below 50 cells/μL may have persistent immune dysfunction despite increased CD4 T-cell counts [1–4]. A small proportion of such patients experience opportunistic infections (OIs) [5–19]. Case studies of such patients with cytomegalovirus (CMV) retinitis implicate low pathogen-specific CD4 T-cell responses [6–8] including loss of CD4 T cells with an effector immunophenotype [9]. It is unclear how to monitor patients for such immune defects. We demonstrated that blood CMV interferon-γ (IFN-γ) enzyme-linked immunosorbent spot-forming cell assay (ELISPOT) responses remain low after 3 years of ART in patients without CMV disease, despite increased CD4 T-cell counts [20]. As we and others have shown that IFN-γ ELISPOT responses to CMV lysate are produced by CD4 T cells [20,21], these findings suggest a deficiency of CD4 effector memory T-cell (T_em-cell) numbers and/or function. We also associated low levels of IFN-γ mRNA in unstimulated CD4 T cells with low levels of interleukin (IL)-23 mRNA in adherent mononuclear leucocytes [22], suggesting a defect of antigen-presenting cells.

The immune dysfunction caused by chronic HIV infection is characterized by increased production of immunoglobulin G (IgG), IgA and IgE in addition to CD4 T-cell depletion. Elevated serum levels of IgA and IgE are predictors of progression to AIDS in untreated patients [23,24]. High serum IgE levels are not associated with the production of allergen-specific IgE antibodies [25,26] and in some patients are associated with an immunodeficiency syndrome resembling hyper-IgE syndrome, where serum IgA levels may also be high [27]. Antiretroviral therapy only partly corrects the increased immunoglobulin production [28,29]. High serum IgA levels in HIV-infected patients receiving ART are associated with B-cell activation [29], but the fundamental cause and significance of these abnormalities remain unclear.

Here, we describe the clinical and immunological findings in three HIV-infected patients who experienced recurrent or de novo Aspergillus fumigatus and/or Mycobacterium avium complex (MAC) infection whilst receiving long-term ART even though total CD4 T-cell counts had increased. We provide evidence that these patients had low CD4 T_em-cell function associated with increased production of IgA and/or IgE. Two of these patients were enrolled in a longitudinal study of immune reconstitution in a group of 27 patients who had increased CD4 T cells on ART and a previous nadir count below 50 cells/μL. Consequently, we examined the relationship between CMV IFN-γ ELISPOT counts and serum immunoglobulin levels in all 27 HIV-infected patients.

Study subjects and methods

Patients and controls

The patient study population comprised 27 HIV-1-infected patients of median age 41 (range 32–65) years, attending out-patient clinics at Royal Perth Hospital. All had nadir CD4 T-cell counts below 50 cells/μL, had received ART for at least 2 years and were enrolled in a longitudinal study of immune reconstitution. Two of these patients (patients 1 and 2) experienced localized MAC infection and/or A. fumigatus infection. In addition, another 59-year-old patient who had a nadir CD4 T-cell count of 99 cells/μL and experienced pulmonary Mycobacterium intracellulare infection after long-term ART (patient 3) was also included in the study. All HIV-infected patients were CMV-seropositive. Controls comprised 21 CMV-seropositive HIV-seronegative healthy individuals of median age 34 (range 21–67) years.

Determination of T-cell counts and plasma HIV RNA level

Whole blood CD4 and CD8 T-cell counts were determined by flow cytometry. Briefly, peripheral blood samples were collected in ethylenediaminetetraacetic acid (EDTA) anticoagulant and directly conjugated fluorescent antibodies (Coulter, Miami, FL) were added. Red cells were lysed using Multi-Q-Prep (Coulter), and cells were analysed using a Coulter EPICS XL-MCL Flow Cytometer (Coulter). Plasma HIV RNA levels were assayed by a quantitative reverse transcriptase polymerase chain reaction method [Roche Amplicor Version 1.0 (standard protocol, 400–750000 HIV-1 RNA copies/mL) for samples collected between 1996 and 1998, or Version 1.5 (ultrasensitive protocol, 50–75000 copies/mL) for samples collected from 1998 onwards; Roche Diagnostic Systems, Branchburg, NJ].

Assay of serum immunoglobulin levels

Serum levels of IgG and IgA were assayed on a Hitachi 917 nephelometer (Behringer Mannheim GmbH, Mannheim, Germany) and serum IgE levels were assayed on an Immulite 2000 Analyser (Diagnostics Products Corporation (DPc), Los Angeles, CA) in the Clinical Biochemistry and Core Services Laboratory, Royal Perth Hospital. Reference intervals are IgG 5.8–13.7 g/L, IgA 0.64–3.4 g/L and IgE <210 kU/mL.

ELISPOT assay of IFN-γ-producing cells

Interferon-γ ELISPOT assays were undertaken with cryopreserved peripheral blood mononuclear cells (PBMC) stimulated by a whole CMV lysate or phytohaemagglutinin, as described previously [20].

Skin testing

Skin testing was undertaken by the intradermal injection of 0.1 mL of M. avium sensitin (Commonwealth Serum Laboratories, Melbourne, Australia) or A. fumigatus antigen (Ebos, Sydney, Australia). Skin test sites were read at 48 h and also at 20 min for the A. fumigatus antigen.

Statistics

Continuous variables describing patients and controls were compared with Wilcoxon's rank sum test. Serum IgE and IgA levels were compared with the normal range for the Western Australian population using the exact binomial test. Correlations between continuous variables were evaluated using Spearman's test. In addition, numbers of values falling above and below the median were compared with Fisher's exact test. Multivariable regression analyses were undertaken on the log₁₀ scale to improve conformity to model assumptions. In all cases, P<0.05 was accepted as defining statistical significance and 0.05<P<0.10 was noted as indicating a trend.

Results

Mycobacterium avium complex and/or Aspergillus fumigatus infection in HIV-infected patients with increased CD4 T-cell counts on long-term ART

Patient 1 had invasive pulmonary A. fumigatus infection of the upper lobe of the right lung, demonstrated by culture of a needle aspirate, on presentation to our unit with HIV infection when the CD4 T-cell count was 44 cells/μL (4%). This resolved after several months of therapy for the A. fumigatus infection, followed by maintenance itraconazole therapy, and commencement of ART. Following 34 months of ART, when the CD4 T-cell count was 390 cells/μL (30%), the patient experienced the first of three major haemoptyses related to pulmonary cavitation at the site of the previous aspergillosis. Recurrent invasive aspergillosis was suspected but A. fumigatus was not isolated. However, after 63 months of ART when the CD4 T-cell count was 783 cells/μL (27%), investigation of further haemoptyses demonstrated extensive consolidation in the upper lobe of the right lung with a computed tomography (CT) scan showing infiltration of the hilum and retrobronchial space. Embolization of bronchial vessels was undertaken to control bleeding. Pathogens were not isolated from samples taken at bronchoscopy but the consolidation resolved after intravenous amphotericin therapy.

The serum IgE level was assayed on five occasions between 40 and 70 months of ART and ranged from 790 to 1600 kU/L. The IgA level was 9.6 g/L at the time of the second relapse. The serum level of A. fumigatus-specific IgE was<0.35 kU/L and A. fumigatus precipitins were not detected. However, there was a 14-mm skin test response 20 min after an intradermal injection of A. fumigatus antigen and 11 mm of induration at 48 h. A biopsy of the induration demonstrated changes of a leucocytoclastic vasculitis, consistent with a type-III hypersensitivity immune response [30].

Patient 2 had focal MAC cerebritis after 23 months of ART following cessation of maintenance therapy for disseminated MAC infection when the CD4 T-cell count was 294 cells/μL. A cutaneous (Delayed-type hypersensitivity DTH) response and whole blood T-cell IFN-γ response to M. avium antigens were negative. Further details are given elsewhere [15]. This patient also developed A. fumigatus sinusitis 69 months after commencing ART. A CT scan demonstrated considerable mucosal thickening of both maxillary antra and sections of tissue taken during antrostomies contained large aggregates of fungal elements with septate hyphae and numerous birefringent crystals consistent with an Aspergillus sp. Samples were not taken for culture but serum A. fumigatus-specific IgE antibody was present at a very high level of 26.6 kU/L (reference <0.35 kU/L). The sinusitis resolved following the antrostomies and use of topical corticosteroids. The serum total IgE level was also markedly increased at 6400 kU/L and subsequently remained >1500 kU/L for at least 3 years. Serum IgG and IgA levels were within the reference interval. At the time of the sinusitis the patient had several complications of long-term ART including diabetes associated with severe insulin resistance, a dorso-cervical fat pad and osteonecrosis of several joints.

Patient 3 developed M. intracellulare infection of the upper lobes of both lungs after 104 months of ART when the CD4 T-cell count was 400 cells/μL. The infection remained active after 6 months of appropriate antimycobacterial therapy. There was a previous history of Pneumocystis jiroveci pneumonitis, but never mycobacterial infection. The patient had emphysema of the lungs caused by smoking cigarettes and long-standing severe antiretroviral lipodystrophy syndrome. Cutaneous DTH responses to M. avium sensitin and tuberculin were absent. The serum IgE level was increased at 330 and 400 kU/L on both occasions it was assayed during the 5 months after diagnosis of the M. intracellulare infection, but IgG and IgA levels were within the reference interval.

None of the patients had a history of atopic disease and the serum level of IgE antibody to a grass pollen mix was <0.35 kU/L in all patients. The serum level of IgE antibody to house dust/mite mix was 1.16 kU/L (moderate positive) and 9.66 kU/L (high positive) in patients 1 and 2, respectively. Patient 2 had continuously undetectable plasma HIV RNA, whereas patients 1 and 3 had intermittent HIV viraemia related to poor adherence or cessation of ART following adverse effects, respectively.

CMV IFN-γ ELISPOT counts remained low in patients with Mycobacterium avium complex and/or Aspergillus fumigatus infection

CMV IFN-γ ELISPOT counts were lower in the HIV-infected patients [n=27; median 16 (interquartile range (IQR) 5–27) spots/2 × 10⁵ cells] than in non-HIV-infected controls (n=21; median 64 (IQR 28–135) spots/2 × 10⁵ cells) (P<0.001). The 95th percentile value for HIV-negative controls was 13 spots/2 × 10⁵ cells. All seven available samples from patient 1 and five of seven samples from patient 2 generated values below the 95th percentile for controls, even though total CD4 T-cell counts had increased on ART (Fig. 1). In patient 3, two of three CMV IFN-γ ELISPOT counts during the 6 months after the M. intracellulare infection was diagnosed (12, 10 and 17 spots/2 × 10⁵ cells) were below the 95th percentile of controls.

Figure 1.  Serial measurements of cytomegalovirus (CMV) interferon (IFN)-γ enzyme-linked immunosorbent spot-forming cell assay (ELISPOT) counts and CD4 T-cell counts in patients 1 and 2. The median (long-dashed line) and lower 95th percentile values for non-HIV-infected controls (short-dashed line) are shown. ART, antiretroviral therapy.

High serum IgA and IgE levels were associated with low CMV IFN-γ ELISPOT counts

Serum levels of IgG, IgA and IgE were above the reference interval in 48, 52 and 37% of the 27 patients in the immune reconstitution study, respectively (P<0.001). As the three patients with OIs had high serum IgE levels, we correlated serum IgE levels with CMV IFN-γ ELISPOT counts in all 27 patients from the immune reconstitution study. There was a weak negative correlation between serum IgE levels and CMV IFN-γ ELISPOT counts (r=–0.27, P=0.2) and a trend for high serum IgE levels to segregate with low CMV IFN-γ ELISPOT counts (P=0.06, Fisher's test) (Fig. 2a). There was also a clear negative correlation between serum IgA levels and CMV IFN-γ ELISPOT counts (r=–0.5, P=0.01) (Fig. 2b) but no association between serum IgG levels and CMV IFN-γ ELISPOT counts (P=0.5) (Fig. 2c). CD4 T-cell counts and plasma HIV RNA levels were not significantly correlated with CMV IFN-γ ELISPOT counts (P>0.2). The strong negative association between serum IgA levels and CMV IFN-γ ELISPOT counts remained in multivariable analyses adjusting for IgE levels and the other covariates described (P<0.01). Serum IgE and IgA levels did not correlate (r=–0.06, P=0.7).

Figure 2.  Relationship between serum levels of (a) immunoglobulin E (IgE), (b) IgA, (c) IgG and cytomegalovirus (CMV) interferon (IFN)-γ enzyme-linked immunosorbent spot-forming cell assay (ELISPOT) counts in 27 HIV-1-infected patients with nadir CD4 T-cell counts below 50 cells/μL who had received antiretroviral therapy (ART) for over 2 years. For IgE and IgG levels vs CMV IFN-γ ELISPOT counts, Fisher's test was carried out using the number of values in each quadrant defined by the respective median values.

Serum IgA and IgE levels and CMV IFN-γ ELISPOT counts were consistent over time

To determine if markers of immune dysfunction changed over time, a second analysis was undertaken in 15 patients at a median time of 35 (range 8–60) months after the first measurement (Table 1). Serum immunoglobulin levels and CMV IFN-γ ELISPOT counts at the two time-points correlated strongly, indicating that these markers of immune dysfunction were consistent over time. In contrast, CD4 T-cell counts increased (P=0.04) and there was no correlation between CD4 T-cell counts at the two time-points.

Table 1.   Correlation between markers of immune function measured at two time-points in 15 patients who had received antiretroviral therapy for over 2 years

Parameter	R-value	P-value
CMV, cytomegalovirus; ELISPOT, enzyme-linked immunosorbent spot-forming cell assay; IFN, interferon; Ig, immunoglobulin.
Current CD4 T-cell count (cells/μL)	0.316	0.252
Current CD8 T-cell count (cells/μL)	0.178	0.525
HIV-RNA (copies/mL)	0.109	0.698
CMV IFN-γ ELISPOTs	0.747	0.001
IgG (g/L)	0.690	0.004
IgA (g/L)	0.773	0.001
IgE (kU/L)	0.923	< 10^−6

Discussion

The blood CD4 T-cell count or percentage is a valuable marker of OI risk in HIV-1-infected patients who are not receiving ART, or who are considering cessation of maintenance antimicrobial therapy following an increase in their CD4 T-cell count on ART [31,32]. However, some patients who were severely immunodeficient before commencing ART remain susceptible to OIs despite an increased CD4 T-cell count. Koletar et al. [5] estimated the risk of an AIDS-defining illness in patients who had previously had a nadir CD4 T-cell count below 50 cells/μL to be 1.75 events per 100 person-years over a median follow-up time of 184 weeks (range 8–216 weeks). Over 40% of events occurred at a higher than expected CD4 T-cell count. Infections with MAC or A. fumigatus in this situation are often localized [10–14], as in the patients described here.

In this study, we have demonstrated that HIV-infected patients on ART who developed A. fumigatus and/or MAC infection at a higher than expected CD4 T-cell count had low CD4 T_em-cell function and high serum levels of IgA and/or IgE. Skin testing demonstrated an absent DTH response to M. avium sensitin in both patients with MAC infection and a type III hypersensitivity immune response to A. fumigatus antigens in the patient with invasive aspergillosis, providing further evidence of cellular immunodeficiency and arguing against immune restoration disease [33]. The immunoregulatory disorder resulting in increased serum IgE levels in these patients appears to be acquired, because none of the patients gave a history of atopic disease and serum IgE antibody to grass pollens was not detected. IgE antibody to house dust/mite in the patients with the highest serum IgE levels might reflect polyclonal B-cell activation.

We found an inverse relationship between CMV IFN-γ ELISPOT counts and serum levels of IgA and to a lesser extent IgE in HIV-infected patients undergoing immune reconstitution. Furthermore, CMV IFN-γ ELISPOT counts and serum levels of IgE and IgA were stable over time. Weinberg et al. have also demonstrated stability of CMV IFN-γ ELISPOT counts over a period of 1 year [21]. These findings suggest a persistent immune defect that induces CD4 T_em-cell dysfunction and B-cell dysregulation.

Increased serum levels of IgG, IgA and IgE are a characteristic of advanced HIV-1 infection. Serum IgA and IgE levels correlate best with disease progression, possibly because high serum IgG levels reflect IgG1 antibody responses to HIV antigens rather than immune activation [34]. Here, only IgE levels were above the reference interval in all three patients with OIs, whereas serum IgA levels showed the strongest inverse correlation with CMV IFN-γ ELISPOT counts in the 27 patients in the immune reconstitution study. This might reflect differences in the regulation of IgA and IgE production [35]. Longitudinal studies of serum IgA and IgE levels in severely immunodeficient HIV-infected patients commencing ART are needed to address this issue. Indeed, the importance of obtaining longitudinal data was recently demonstrated in a study of the immunomodulatory effects of cyclooxygenase-2 inhibitors in HIV-infected patients receiving ART, where reductions in serum IgA levels were observed even though levels were within the reference interval [36].

Increased production of immunoglobulins in HIV-1-infected patients reflects the effects of multiple cytokines and costimulatory molecules that activate naïve and memory B cells and impair memory B-cell function [37,38]. The inverse relationship between immunoglobulin levels and IFN-γ responses in HIV-infected patients on ART cannot simply be explained by a bias towards type (T) 2 cytokine production, because we have shown that IgE levels correlated inversely with mitogen-induced production of both IL-5 (T2) and IFN-γ (T1) mRNA [39]. Weak positive correlations were also evident between levels of IFN-γ, IL-4δ2 (both T1), IL-4 and IL-5 (both T2) mRNA after mitogenic stimulation. An alternative explanation is that high immunoglobulin production and low IFN-γ responses reflect persistent plasmacytoid dendritic cell (pDC) dysfunction in HIV-infected patients receiving ART [40] because pDCs regulate B-cell differentiation and immunoglobulin class switching [41,42]. Furthermore, increased expression of pro-apoptotic molecules such as tumour necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) on pDCs [43] might impair the function of CMV-specific CD4 T cells [44].

In summary, we describe three patients who developed recurrent or de novo A. fumigatus and/or MAC infection associated with low CD4 T_em-cell function and increased serum levels of IgA and/or IgE and provide evidence that low CD4 T_em-cell function and B-cell dysregulation might be manifestations of HIV-induced immune dysfunction that is independent of CD4 T-cell counts and not corrected by ART. We suggest that assays of CD4 T_em-cell function and serum IgE and IgA levels should be further evaluated as tests of immune function in severely immunodeficient HIV-infected patients responding to ART.

Acknowledgements

Financial support for this study was received from the National Health and Medical Research Council of Australia.