Measurement of Human Immunodeficiency Virus p24 Antigen in Human Cerebrospinal Fluid With Digital Enzyme-Linked Immunosorbent Assay and Association With Decreased Neuropsychological Performance

Abstract

New tools are needed to understand human immunodeficiency virus central nervous system involvement. Testing 15 cerebrospinal fluid (CSF) samples for p24 antigen, using a high-sensitivity assay, we found a strong correlation trend between CSF p24 concentration and worse neuropsychological performance.

To best understand and eventually cure human immunodeficiency virus (HIV), a fuller understanding of all anatomic viral reservoirs is needed. There is building evidence that the central nervous system (CNS) is one such reservoir, based on reports of CNS virologic escape and the presence of low-level HIV in the CNS despite combination antiretroviral therapy (cART) [1, 2]. At the same time, HIV RNA concentrations may not provide a complete understanding of HIV in the CNS and other sites [3]. Therefore, more research is needed on other markers that could be used to measure viral persistence as well as reactivation.

The HIV p24 antigen is the most abundant HIV protein and is essential for assembly of the capsid that encases HIV genetic material. Detection of the HIV p24 antigen is used clinically to diagnose early HIV infection, during which HIV-specific antibodies are not yet detectable. However, previous p24 assays have been hampered by low sensitivity and thus far have not been shown to be valuable in monitoring HIV over the course of infection. High-sensitivity p24 assays are being developed, including one that has been referred to as “digital” enzyme-linked immunosorbent assay (ELISA) based on microscopic well capture and digital camera platform (Quanterix). This assay allows for the detection of very low-level (femtomolar) concentrations of p24. Specifically, the lowest limit of quantification is 0.01 pg/mL and the lowest limit of detection is 0.003 pg/mL. The assay has been shown to be very sensitive for the detection of early HIV infection and is also being used to measure in vitro HIV latency reactivation [4, 5]. We undertook an exploratory study in which cerebrospinal fluid (CSF) and matched plasma concentrations of HIV p24 were measured using this digital ELISA.

METHODS

Adult outpatients with chronic HIV infection were enrolled between March 2011 and October 2016 at the Emory University Center for AIDS Research clinical core site in Atlanta as part of ongoing neurocognitive studies. Individuals were excluded from the studies for the following conditions: (1) neurologic disease known to affect memory (including stroke, cancer involving the brain, traumatic brain injury, and AIDS-related opportunistic infection of the CNS); (2) current ongoing substance use reported (marijuana use in the last 7 days or cocaine, heroin, methamphetamine, or other nonmarijuana illicit drug use in the last 30 days); (3) heavy alcohol consumption (defined as >7 drinks per week for women and >14 drinks per week for men); or (4) serious mental illness (depression was not excluded if well controlled with treatment).

Participants with a history of treated syphilis and a persistently positive rapid plasma regain (RPR) titer of 1:8 or less were eligible for the study if they had a ≥4-fold decrease in RPR 6 months after treatment and had no neurologic symptoms at initial syphilis presentation.

A neuropsychological battery was administered that included the following 9 tests used commonly in studies of cognition and HIV infection: Trail Making Test, parts A and B; Hopkins Verbal Learning Test, total recall and delayed recall; Grooved Pegboard Test, dominant and nondominant hands; Stroop Color Naming and Color and Word tests; and Letter Fluency (Controlled Oral Word Association Test). Scores were adjusted for demographic characteristics using published norms, and score adjustment for practice effects was made for longitudinal visits by using median practice effect data from previous work [6]. A composite global neuropsychological score (NPT-9) was then calculated by averaging individual T scores. The Global Deficit Score (GDS), a validated measure of neurocognitive impairment in HIV infection based on T scores, was also calculated. A GDS ≥0.5 is consistent with overall impairment, and a GDS ≥1.5 is suggestive of dementia [7].

HIV RNA concentrations from plasma and CSF were measured at the Emory Center for AIDS Research virology core using the Abbott laboratories m2000 Real Time HIV-1 assay system (lowest limit of detection, 40 copies/mL). For the purposes of this analysis, samples were selected to represent a broad range of CSF HIV RNA levels as well as the presence of CNS virologic escape during cART. CNS virologic escape was defined as either (1) detectable CSF HIV RNA in the setting of plasma HIV RNA <40 copies/mL or (2) CSF HIV RNA >1 log₁₀ copies/mL higher than plasma HIV RNA [8].

Laboratory staff performing p24 testing were blinded to the HIV RNA concentrations in the samples. Statistical analysis were performed using SAS JMP 13 software. Normality was assessed with the Shapiro-Wilk test. Neither HIV RNA nor p24 met criteria for normality, and correlation statistics are thus reported with Spearman ρ test. The study was approved by the Emory University Institutional Review Board and written consent was obtained from all participants.

RESULTS

Findings were analyzed in a total of 15 participants ( Table 1). The median estimated duration of HIV infection was 5 years. Four of 15 participants were currently treated with 3-drug cART, all of whom had been receiving treatment for ≥6 months. Two of these qualified for CNS virologic escape. Of the remaining 11 participants, 4 were cART naive, and the other 7 were cART experienced but had been off therapy for ≥6 months (maximum time off therapy, 30 months).

Table 1.

Demographics and Laboratory Results, Including CSF p24 Concentrations

Participant ID No.a	Age, y	Sex	Race	CD4 Cells/µL	CD4 Cells, %	CSF WBCs/ µL	CSF RBCs/ µL	CSF Protein, mg/dL	CSF HIV RNA, Copies/mL	CSF p24, pg/mL	Plasma HIV RNA, Copies/mL	Plasma p24, pg/mL
233	51	M	AA	237	10	23	0	57	463798	3.7176	656108	0.3520
245	40	M	AA	415	24	40	5	101	323081	0.8732	25017	UD
703	50	M	AA	58	10	12	1	95	177187	0.995	182300	UD
737	48	M	AA	529	26	11	6	118	36886	UD	17653	UD
705	52	M	W	29	7	5	0	36	32623	0.9335	170276	0.0739
701	56	M	AA	12	1	0	3	35	9884	0.8698	62028	1.7597
211b	48	M	AA	171	18	0	0	67	5643	0.6831	51	UD
226	33	M	AA	212	12	11	1	54	3687	0.0964	1118111	UD
721	36	M	AA	44	3	4	0	38	2144	0.0234	234384	Detected at <0.01
251	57	F	W	41	5	4	2	30	1691	0.4189	1976241	>28
723	50	M	AA	10	4	0	0	40	1091	Detected at <0.01	52597	0.0895
707	52	M	AA	14	4	0	0	50	907	UD	657500	UD
717	53	F	A	229	27	31	2	65	623	UD	79	UD
754 b	40	M	NAm	250	20	0	1	30	43	UD	<40	Detected at <0.01
704	46	M	W	329	21	0	1	74	<40	UD	<40	UD
Median (IQR) or No. (%)	50 (40–52)	M, 13 M (86.7)	AA, 11 AA (73.3)	171 (29–250)	10 (4–21)	4 (0–12)	1 (0–2)	54 (36–74)	3687 (907–36886)	0.0964 (UD–0.8732)	62028 (79-656108)	UD (UD–0.0895)

Abbreviations: AA, African American; CSF, cerebrospinal fluid; F, female; HIV, human immunodeficiency virus; ID, identification; IQR, interquartile range; M, male; NAm, Native American; RBCs, red blood cells; UD, undetectable; W, white; WBCs, white blood cells.

^aIndividuals listed in descending order of CSF HIV RNA levels.

^bCentral nervous system virologic escape case.

CSF p24 was detected in 10 of 11 participants (91%) with CSF HIV RNA levels >1000 copies/mL, but plasma p24 was detected in only 6 of 11 (54.5%) with plasma HIV RNA levels >1000 copies/mL. CSF p24 concentrations correlated strongly with CSF HIV RNA concentrations (ρ = 0.804; P < .001) but not with plasma HIV RNA (ρ = 0.393; P = .15). However, plasma p24 did not correlate significantly with either plasma HIV RNA (ρ = 0.35; P = .21) or CSF p24 (ρ = 0.34; P = .22). The median GDS was 0.75, and only 2 participants had a GDS ≥1.5. The mean NPT-9 score was 41.3 (standard deviation, of 9.4). Although the correlation between CSF HIV RNA and NPT-9 was low (ρ = −0.25; P = .37), there was a strong correlation trend between CSF p24 and NPT-9 (ρ = −0.51; P = .055).

DISCUSSION

Given the persistence of HIV despite cART, there is a need to more fully understand all reservoirs in order to achieve eradication of the virus. In this brief report, we demonstrate that HIV p24 antigen is quantifiable at low levels in CSF using a novel digital ELISA. CSF p24 more closely correlated with worse neuropsychological performance than did CSF HIV RNA. These findings support the longstanding interest in measuring viral proteins in the CNS during HIV infection.

In 2 seminal studies from the pre-cART era using older assays, CSF p24 was detectable in <15% of infected individuals [9, 10]. However, it was more likely to be detectable among individuals with advanced dementia in these studies. Specifically, some participants with significant dementia (defined as a Memorial Sloan Kettering AIDS dementia score ≥2) had markedly elevated CSF p24 concentrations of >100 pg/mL. Participants in our study had milder impairment (only 2 had GDS consistent with dementia), and most had a relatively recent history of cART. These factors probably explain why none of out participants had markedly high CSF p24 concentrations.

The presence of p24 may be an indication of “translationally competent” HIV infection of cells in the CNS. The fact that CSF p24 in this study was more closely correlated than CSF HIV RNA levels with worse neuropsychological performance supports this idea. Specifically, if p24 is being produced in the CNS then it is likely that other viral proteins are also being produced, including those directly associated with neurotoxicity such as tat, gp120, and vpr [11].

We acknowledge the limitations of this study, including the small exploratory sample size. Some participants had high CSF protein concentrations, which could reflect blood-brain barrier impairment. Therefore, it is possible that the CSF p24 observed in this study could come from the systemic circulation. However, the close correlation with CSF HIV RNA concentration (as opposed to the nonsignificant correlation with plasma HIV RNA concentration) provides some evidence that HIV p24 in the CSF is locally produced. CSF p24 was undetectable in 1 specimen with a relatively high CSF HIV RNA concentration. The measurement of p24 from blood has been hampered in the past by the formation of antigen-antibody complexes [12], and this is a possible explanation for p24 being undetectable in this CSF sample as well as several plasma p24 samples.

Despite the sensitivity of the p24 assay, all of the samples with CSF HIV RNA levels <1000 copies had undetectable CSF p24, including 1 case of low-level CNS virologic escape. Therefore, further research is needed to determine whether the assay could be optimized for maximum sensitivity (including dissociation steps to account for antigen-antibody complexes in the CSF). Given the preliminary nature of this small study, more comprehensive research will be needed to further investigate CSF p24 by high-sensitivity assay as a marker for neurocognitive impairment during poorly controlled HIV infection and as a marker for HIV CNS reactivation during latency reversal therapy.

Notes

Acknowledgments. We acknowledge Michele Wolfe, Linan Song, and Bryan Klekota at Quanterix for their assistance with this project.

Financial support. This work was supported by the National Institutes of Health (grants K23MH095679 to A. M. A., K24MH097673 to S. L. L., and P30AI050409 to Emory University for the Center for AIDS Research) and the Emory Medical Care Foundation.

Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.