Residual Central Nervous System Immune Activation Is Not Prevented by Antiretroviral Therapy Initiated During Early Chronic HIV Infection

Abstract

Background

Antiretroviral therapy (ART) initiated during acute infection can potentially impact the central nervous system (CNS) reservoir, but the differential long-term effects of ART initiation during early or late chronic infection are unknown.

Methods

We included neuroasymptomatic people with human immunodeficiency virus (HIV) with suppressive ART initiated during chronic (>1 year since transmission) HIV with archived cerebrospinal fluid (CSF) and serum samples after 1 and/or ≥3 years of ART from a cohort study. CSF and serum neopterin was measured using a commercial immunoassay (BRAHMS, Germany).

Results

In total, 185 people with HIV (median, 79 [interquartile range, 55–128] months on ART) were included. A significant inverse correlation was found between CD4⁺ T-cell count and CSF neopterin only at baseline (r = −0.28, P = .002), but not after 1 (r = −0.026, P = .8) or ≥3 (r −0.063, P = .5) years of ART. No significant differences were seen in CSF or serum neopterin concentrations between different pretreatment CD4⁺ T-cell strata after 1 or ≥3 (median, 6.6) years of ART.

Conclusions

In people with HIV initiating ART during chronic infection, occurrence of residual CNS immune activation was not correlated with pretreatment immune status, even when treatment was initiated at high CD4⁺ T-cell counts, suggesting that the CNS reservoir, once established, is not differentially affected by the timing of ART initiation during chronic infection.

Human immunodeficiency virus type 1 (HIV) is a neurotropic virus that infects the central nervous system (CNS) shortly after transmission of infection [1, 2]. While a stable systemic viral reservoir is established very early in the course of infection [3], the timing of the establishment of an independent CNS reservoir is less certain [4]. The main cellular targets of HIV in the CNS are microglia and perivascular macrophages, where infection leads to local inflammation measurable in cerebrospinal fluid (CSF) by neopterin, a stable and sensitive pteridine biomarker of cellular activation [5, 6]. Although antiretroviral therapy (ART) usually prevents progressive CNS injury in people with HIV (PWH), residual CNS immune activation remains common even after long-term suppression of viral replication, and CSF neopterin concentrations above normal levels have been found in a substantial proportion of PWH even after long-term (>10 years) viral suppression, indicating a residual chronic immune activation in a subset of individuals despite apparent suppression of viral replication by ART [7, 8]. Interestingly, a recent study reported that CSF neopterin concentrations were essentially normalized after 96 weeks of treatment in individuals where ART was initiated very early during acute HIV infection [9]. These findings are important and indicate that there may be a delay in the establishment of a stable, permanent reservoir of infection in CNS resident cells. However, an outstanding question is whether the timing of ART initiation beyond the very early phase of acute infection may have an impact on the chronic residual immune activation frequently seen in PWH with virally suppressive ART. The aim of this study was to evaluate residual CNS immune activation measured as CSF neopterin concentrations in PWH initiating ART during chronic HIV infection across a spectrum of high to low CD4 cell counts.

METHODS

Participants and Study Design

Since 1985, PWH monitored at Sahlgrenska University Hospital, Gothenburg, Sweden, have been continuously included in the longitudinal Gothenburg HIV CSF Study Cohort [8, 10]. From this cohort, we retrospectively identified neuroasymptomatic, treatment-naive or off-treatment >1 year PWH who started ART during chronic HIV infection, were continuously treated with ART with systemic viral suppression during the study period, and who had CSF measurements after 1 and/or after ≥3 years of ART. Successful treatment was defined as HIV RNA <50 copies/mL after ≥6 months of ART. Study participants were stratified according to pretreatment CD4⁺ T-lymphocyte count: <50, 50–199, 200–349, 350–499, or ≥500 cells/µL. The cutoffs were chosen to reflect the important clinical stages of untreated HIV infection, as described previously [11]. Patients were defined as neuroasymptomatic if they were without symptoms or signs of neurological or cognitive disability on clinical examination at follow-up. Any treatment interruption was considered as end of therapy regarding time of follow-up. A maximum of 1 transient increase in plasma viral load (“blip”) defined as a single plasma HIV RNA value <500 copies/mL preceded and followed by a plasma RNA <50 copies/mL, was allowed during the study period [12].

In individuals with >1 available sample after ≥3 years of ART, the most recent was used. All ART regimens were allowed including changes in therapy, as long as plasma viral suppression was achieved. Study visits included CSF and venous blood sampling and clinical examination [8].

Patient Consent Statement

The research protocol and study design was approved by the Research Ethics Committee of the University of Gothenburg, and all subjects provided written informed consent.

CSF Sampling and Laboratory Methods

Cerebrospinal fluid white blood cell (WBC) and blood CD4⁺ T-lymphocyte counts were measured using routine methods. HIV RNA levels in plasma and CSF were quantified using the Roche Amplicor Monitor version 1.5 or the Roche TaqMan version 1, 2, or 6800 reverse-transcription polymerase chain reaction assays, according to the manufacturer's instructions. The corresponding lower limits of quantification were 50 (Amplicor), 40 (TaqMan v1), and 20 (TaqMan v2 and 6800) HIV RNA copies/mL.

After centrifugation, paired cell-free samples of CSF and plasma were immediately stored at −80°C until later analysis. Archived CSF and plasma samples were subsequently used for all study analyses.

CSF and serum neopterin concentrations were measured using a commercially available immunoassay (BRAHMS, Henningsdorf, Germany) with an upper normal reference value of 5.8 nmol/L in CSF and 9.1 nmol/L in serum [6, 13]. Immunoglobulin G (IgG) concentrations were measured by immunoturbidimetry on a Cobas instrument (Roche Diagnostics, Penzberg, Germany). IgG index was calculated as previously described [14]. CSF β2-microglobulin (β₂M) was measured using the N Latex β₂M kit on the Atellica NEPH 630 System (Siemens Healthcare GmbH. Erlangen, Germany).

Statistical Analysis

Descriptive statistics were performed using SPSS (IBM SPSS version 28, Chicago, Illinois) or Prism (version 9, GraphPad, La Jolla, California). Numerical variables were log₁₀-transformed where appropriate for the tests used. Comparisons of CSF biomarker concentrations between patient groups were analyzed using analysis of variance with least significant difference post hoc test. Correlations between CSF neopterin and pretreatment CD4⁺ T-cell count were analyzed using Pearson correlation. All tests were 2-sided, and a P value of <.05 was considered statistically significant.

RESULTS

Study Population

A total of 183 PWH (39% female) who initiated ART between 1996 and 2020 were identified and included in the study. Study participants were stratified according to pretreatment CD4⁺ T-lymphocyte counts. The distribution of participants between study groups and baseline characteristics at ART initiation is shown in Table 1.

Table 1.

Characteristics of Study Participants

Group Number	Pretreatment CD4+ T-Cell Count, Cells/µL					Total (N = 183)
	≥500 (n = 20)	350–499 (n = 23)	200–349 (n = 63)	50–199 (n = 51)	<50 (n = 26)
Pre-ART CD4+ count, cells/µL	630 (560–840)	390 (370–435.5)	260 (239–301.75)	130 (84–150)	25 (10–40)	240 (120–332.3)
Female sex, No. (%)	4 (18)	10 (43)	27 (42)	19 (50)	10 (26)	71 (38)
Age at ART initiation, y	35.9 (30.2–51.1)	45 (31.7–51.8)	38.2 (31.8–42.8)	38.3 (34.3–50.1)	42.9 (37.1–47.4)	39.8 (32.9–48.4)
Pre-ART Plasma-HIV RNA, log10 copies/mL	4.16 (3.66–4.85) n = 16	4.53 (4.16–4.88) n = 14	4.65 (4.28–5.46) n = 36	5.16 (4.63–5.70) n = 32	5.39 (5.20–5.63) n = 22	4.9 (4.3–5.6) n = 120
Pre-ART CSF HIV RNA, log10 copies/mL	3.67 (2.04–3.71) n = 16	3.65 (3.39–4.27) n = 14	4.08 (3.60–4.41) n = 36	4.19 (3.59–4.77) n = 32	2.89 (2.16–3.68) n = 22	3.77 (3.16–4.39) n = 120
Pre-ART CSF WBC	4.5 (0.75–6.75) n = 18	7.00 (4–8.8) n = 14	6.00 (2–12) n = 37	4.00 (1–8) n = 30	1.00 (0–1) n = 20	4.00 (1–8) n = 122
Months on ART at ≥3 y	44.5 (39–52) n = 12	67 (57.5–103) n = 18	87 (59–126) n = 50	100 (76–161) n = 35	79 (51–161) n = 15	79 (55–127.5) n = 133
CSF neopterin, nmol/L
Pre-ART	9.6 (6.5–20.7) n = 19	15.3 (13.9–22.0) n = 14	16.2 (11.3–27.7) n = 37	21.0 (10.7–34.55) n = 31	22.5 (12.2–34.48) n = 22	17.0 (10.4–24.55) n = 123
1 y	5 (4.85–8.3) n = 20	6.25 (4.38–7.58) n = 12	4.9 (4.4–7.36) n = 30	5.6 (4.5–8.2) n = 33	5.8 (4.53–8.08) n = 18	5.6 (4.5–7.8) n = 113
≥3 y	5.3 (4.8–6.7) n = 13	6 (4.8–7.28) n = 18	5.4 (4.55–8.05) n = 51	7 (4.8–8.45) n = 36	6.40 (4.75–8.05) n = 16	5.7 (4.8–8.1) n = 134
Serum neopterin, nmol/L
Pre-ART	9.6 (8.4–19.5) n = 19	15.85 (9.63–19.6) n = 14	16.9 (9.2–22.8) n = 37	21.8 (13–27.4) n = 31	32.95 (25.25–53.8) n = 22	18.5 (9.7–24.95) n = 123
1 y	6.1 (5.8–7.78) n = 20	6.1 (4.9–8.2) n = 12	8.9 (4.85–11.05) n = 30	7.3 (5.5–9.9) n = 33	8.5 (5.9–13.1) n = 18	7.3 (5.5–9.9) n = 113
≥3 y	5.3 (4.8–6.7) n = 13	7.15 (5.2–9.4) n = 18	6.8 (5.2–9.7) n = 51	8.45 (6.35–10.5) n = 36	7.85 (5.0–8.9) n = 16	7.35 (5.2–9.7) n = 134

Characteristics of included people with HIV, and CSF and serum neopterin concentrations, are shown for the whole study population and stratified by pretreatment CD4⁺ T-cell count. All values are shown as median (interquartile range) unless otherwise stated.

Abbreviations: ART, antiretroviral therapy; CSF, cerebrospinal fluid; HIV, human immunodeficiency virus; WBC, white blood cell count.

Pretreatment blood and CSF samples were available in 120 individuals. Pretreatment plasma viral load increased in parallel with declining nadir CD4⁺ T-lymphocyte counts (Table 1). CSF viral RNA copy numbers followed a similar pattern, with the exception of the patient group with the lowest pretreatment CD4⁺ count <50 cells/µL (Table 1). Additionally, patients with a CD4⁺ T-cell count <50 cells/µL had a lower CSF WBC count compared to the other patient groups. Median time on ART at the latest (≥3 years) study visit also varied between groups, with the shortest total time on ART found in the >500 CD4⁺ T cells/µL group and the longest in the 50–199 CD4⁺ T cells/µL group, with a median of 45 (interquartile range [IQR], 39–52) and 100 (IQR, 76–161) months, respectively (Table 1).

In total, 12 patients (6 after 1 year, and 6 at ≥3 years of ART) had a plasma viral “blip” during the study period, with a median viral load of 112 (range, 56–489) and 58 (range, 52–63) copies/mL after 1 and ≥3 years, respectively. No individuals had >1 plasma viral “blip,” and viral suppression was maintained <50 copies/mL at all additional visits throughout the study period.

In CSF, after 1 year of ART, 7 of 100 (7%) PWH (CD4⁺ T-cell count 50–199 cells/µL: n = 4; 350–499 cells/µL: n = 0; other groups: n = 1) had CSF but not plasma HIV RNA >20 copies/mL with a median of 43 (range, 21–172) copies/mL, constituting asymptomatic CSF escape. Correspondingly, at ≥3 years of ART, 8 of 121 (7%) patients (CD4⁺ T-cell count ≥500 cells/µL: n = 1; 350–499 cells/µL: n = 1; 200–349 cells/µL: n = 3; 50–199 cells/µL: n = 3; <50 cells/µL: n = 0) had CSF but not plasma HIV RNA >20 copies/mL, with a median of 31 (range, 21–46) log₁₀ copies/mL.

CSF Neopterin Analyses

CSF neopterin concentrations at baseline and after 1 and ≥3 years of ART in the different patient groups are shown in Table 1. A significant inverse correlation was found between pretreatment CD4 T-lymphocyte count and CSF neopterin concentration at baseline (r = −0.28, P = .002, n = 120), while no correlation between CSF neopterin and pretreatment CD4⁺ T-cell count was found after 1 year (r = −0.026, P = .8, n = 110) or ≥3 years (r −0.063, P = .5, n = 131) of ART. The relationships between pretreatment CD4⁺ T-cell count and CSF neopterin concentration at baseline and after 1 year and ≥3 years of suppressive ART are shown in Figure 1.

Figure 1.

Correlation between pretreatment CD4⁺ cell count and cerebrospinal fluid (CSF) neopterin. Relationship between pretreatment CD4⁺ cell count and CSF neopterin concentrations at baseline and after 1 and ≥3 years of antiretroviral therapy (ART). A significant inverse correlation between pretreatment CD4⁺ cell count and CSF neopterin was found at baseline (r = −0.28, P = .002), but not after 1 (r = −0.026, P = .8) or ≥3 (r = −0.063, P = .5) years of ART. The number of patients with normalized CSF neopterin was 57 of 110 (52%) after 1 year of ART and 67 of 134 (50%) after ≥3 years of ART. The dotted line indicates the upper normal reference value for CSF neopterin (5.8 nmol/L). Correlation between CSF neopterin and pretreatment CD4⁺ cell count was estimated using Pearson correlation.

At baseline, patients with pretreatment CD4⁺ T-cell count <50 cells/µL had significantly higher median CSF neopterin (22.5 [IQR, 12.2–34.5] nmol/L) concentrations compared with the 350–499 cells/µL (15.3 [IQR, 13.9–22.0] nmol/L; P = .006) and ≥500 cells/µL (9.6 [IQR, 6.5–20.7] nmol/L; P = .003) groups (Table 1, Figure 2A). However, no significant differences in CSF neopterin concentrations were seen between the patient groups after 1 and ≥3 years of ART (Figure 2A).

Figure 2.

Cerebrospinal fluid (CSF) neopterin concentrations at baseline and after 1 and ≥3 y of antiretroviral therapy (ART). Patients were stratified according to pretreatment CD4⁺ cell counts. Boxes indicate the median and interquartile range, whiskers indicate the 10th–90th percentiles, and the mean is indicated as “+.” Statistically significant differences are represented by P values and all statistical comparisons of neopterin concentrations between groups were analyzed using analysis of variance with least significant difference post hoc test. The upper normal reference values for neopterin in CSF (5.8 nmol/L) are indicated by dotted lines. Abbreviations: ART, antiretroviral therapy; CSF, cerebrospinal fluid; ns, not significant.

CSF neopterin was increased above the upper normal reference concentration (5.8 nmol/L) in nearly all PWH at baseline (Figure 1). While CSF neopterin concentrations decreased significantly after ART initiation, a substantial proportion of individuals still had CSF concentrations above the upper normal reference at follow-up. The number of patients with normalized CSF neopterin was 57 of 110 (52%) after 1 year, and 67 of 134 (50%) after ≥3 years of ART (Figure 1). Importantly, a substantial proportion of patients in all groups still had CSF neopterin concentrations above the upper normal reference after 1 and ≥3 years of ART, regardless of pretreatment CD4⁺ T-cell count.

A correlation was seen between baseline and year 1 CSF neopterin (r = 0.37, P = .0001), but not between 1 year and ≥3 years of ART or between baseline and ≥3 years. Additionally, at year 1, individuals who did not normalize CSF neopterin had 64% higher baseline CSF neopterin compared to individuals who had CSF neopterin below the upper normal reference (P < .001). No significant correlation was seen with CSF neopterin concentrations after ≥3 years of ART.

We detected no impact of potential covariates (year of assessment, duration of ART, age, and sex) on CSF neopterin in a multivariable analysis. Additionally, no differences were seen between individuals with or without available baseline CSF neopterin (data not shown).

Serum Neopterin Analyses

Serum neopterin concentrations in the study groups are shown in Table 1. Overall, a similar pattern was seen in serum and CSF. At baseline, significantly higher serum neopterin concentrations were seen in the groups with the lowest pretreatment CD4⁺ T-cell counts compared to the groups where ART was initiated earlier in the disease course. No significant differences in serum neopterin concentrations were seen between the patient groups after 1 and ≥3 years of ART.

The proportion of patients with serum neopterin concentrations below the upper normal reference (9.1 nmol/L) was also similar across groups; overall, 72 of 110 (65%) patients had normalized serum neopterin concentrations after 1 year, and 87 of 132 (66%) after ≥3 years of ART.

Additional Biomarker Analyses

To compare CSF neopterin results with other markers of immune activation, we also analyzed CSF β2M and IgG index in relation to pretreatment CD4⁺ T-cell count. No correlation was found between either biomarker and pretreatment CD4⁺ T-cell count after 1 and ≥3 years of ART (Supplementary Figure 1).

DISCUSSION

In this study of PWH initiating ART across different stages of HIV infection, we found that a substantial proportion of patients initiating ART during chronic HIV infection still had CSF biomarker evidence of ongoing immune activation after several (median, 6.6) years of effective ART with systemic viral suppression, regardless of pretreatment immune status. While a significant inverse correlation was found between pretreatment CD4⁺ T-cell count and CSF neopterin concentrations at baseline, no correlation was found during effective ART, indicating that ART initiated during chronic HIV infection does not fully prevent residual CNS inflammation, even when therapy is started at high CD4⁺ T-cell counts early in the course of chronic infection. Similarly, although significant differences in baseline CSF neopterin concentrations could be detected between the most immunosuppressed patient group and the groups with maintained immune function, no differences in CSF neopterin concentrations could be detected between the patient groups after 1 or ≥3 years of therapy.

Studies have shown an overall benefit in reducing the risk of serious AIDS-related as well as non-AIDS-related events in patients initiating ART at high CD4⁺ T-cell counts [15]. This benefit may partly be explained by a reduction in systemic immune activation, since increased concentrations of inflammatory biomarkers have been associated with morbidity and mortality in PWH [16, 17]. However, early ART does not reduce all comorbidities, and residual immune activation is seen despite suppressive therapy [3]. Residual immune activation may also be of potential importance in the CNS, where mild cognitive impairment has been associated with increased concentrations of CSF neopterin [18].

It has been well established in several cohort studies that ART initiated during chronic infection does not fully inhibit inflammatory responses within the CNS of PWH [19, 20]. While ART is often effective in inhibiting overt viral replication in CSF as well as in plasma, previous studies where a large proportion of included PWH initiated ART at low CD4⁺ cell counts have found CSF immune biomarker evidence of residual inflammatory activity in the CNS even after long-term viral suppression by ART [7, 8]. It has been presumed that an independent CNS reservoir that is not directly reachable by ART, consisting mainly of long-lived cells of microglial/macrophage lineage, plays an important part in this inflammatory process [20, 21]. This hypothesis is supported by the apparent lack of effect on residual CNS immune activation by treatment intensification studies [22, 23].

However, the time frame for the establishment of the CNS reservoir is not fully known. While the primary systemic viral reservoir in CD4⁺ T cells is established within days of infection [3], studies of CSF and plasma antibody responses as a surrogate marker for HIV antigen load and reservoir size indicate that the CNS reservoir is established later in the course of infection [4]. In patients initiating ART during chronic infection, effects on antibody levels in CSF and serum were limited, whereas CSF antibody levels were substantially reduced over time in patients who initiated ART during early infection [4]. In a seminal study, Hellmuth et al showed that individuals who initiated ART early during acute infection normalized all CSF biomarkers of immune activation after 24 and 96 weeks of therapy, while several plasma markers remained elevated compared to healthy controls [9]. Although the study did not include control patients initiating ART during chronic infection, no differences in CSF biomarker concentrations were seen after 96 weeks compared to healthy controls, suggesting that very early ART can potentially have an impact on the establishment and size of an independent CNS reservoir.

In contrast, in this study only 52% and 50% of patients initiating ART during chronic infection normalized CSF neopterin after 1 year and ≥3 years of effective ART, and normalization of residual CSF inflammation was independent of pretreatment CD4 T-cell count. These results confirm observations made in previous studies where ART was often initiated later in the disease course and suggest that the CNS reservoir, once established, cannot be easily reduced or eliminated by current ART [9, 24–26].

Overall, our results indicate that ART initiated even during the early stages of chronic HIV infection with high CD4⁺ T-cell count and conserved immune function does not prevent residual CNS immune activation, despite several years of suppressive ART. These findings are in contrast with observations made in patients starting ART during acute infection, where very early ART may have an influence on the establishment and size of an independent CNS reservoir. A reduction in reservoir size may in turn reduce the main driver of proinflammatory activity in the CNS [9]. Together, these findings have important potential implications, if the establishment of the CNS reservoir may be subject to intervention by very early initiation of ART, while ART initiated during the chronic phase can mitigate, but not eliminate, CNS inflammatory activity.

Strengths and Limitations

Strengths of the study include the inclusion of PWH across all immunological stages of untreated HIV infection, and the prospective character of the cohort from which included subjects were identified. All patients were recruited and subsequently followed and monitored at a single site throughout the study duration, and time on suppressive ART at the latest included study visit was long, and in many cases considerably longer than the inclusion criteria (≥3 years). Although all patients had successful ART without treatment failure, we cannot entirely exclude brief loss of viral suppression between assessments. Overall, the number of included patients was comparatively large for a study of CSF biomarkers. Although patients were clinically neuroasymptomatic, neurocognitive testing was not consistently performed, prohibiting an evaluation of the prevalence or potential influence of subclinical or asymptomatic neurocognitive impairment in the study volunteers. Additionally, no HIV-uninfected controls were included. However, we utilized the availability of well-characterized normal reference values of the studied biomarkers, which have previously been established in a large cohort of individuals, to evaluate differences between the study groups [6].

In conclusion, we found that in patients initiating ART during the chronic phase of HIV infection, pretreatment immune status was not correlated with the occurrence of residual immune activation during suppressive ART, even when treatment was initiated at high CD4⁺ T-cell numbers and after several years of effective ART. While pretreatment CD4⁺ T-cell count was correlated to CNS inflammation at baseline, no correlation was found during effective ART, indicating that even when initiated early in the course of chronic infection, ART does not fully prevent residual CNS inflammation in all individuals. In contrast to a previous study of patients with acute infection, ART had no apparent impact on residual CNS immune activation (measured as CSF neopterin, β2-microglobulin, and IgG index) regardless of when in chronic infection therapy was initiated. The clinical importance of residual CNS immune activation and the potential benefit of reduction of reservoir size by very early ART remain unclear, and potential long-term benefits need further investigation.

Supplementary Data

Supplementary materials are available at Open Forum 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.

Notes

Financial support. This work was supported by the Swedish state, under an agreement between the Swedish government and the county councils (agreement on medical education and research (ALF) agreement ALFGBG-965885; ALFGBG-966347).