Increased Cell-free Mitochondrial DNA is a Marker of Ongoing Inflammation and Better Neurocognitive Function in Virologically Suppressed HIV-infected Individuals

Abstract

Cell-free Mitochondrial DNA (mtDNA) is a highly immunogenic molecule that is associated with several inflammatory conditions and with neurocognitive impairment during untreated HIV infection. Here, we investigate how cell-free mtDNA in cerebrospinal fluid (CSF) is associated with inflammation, neuronal damage and neurocognitive functioning in the context of long-term suppressive antiretroviral therapy (ART). We quantified the levels of cell-free mtDNA in the CSF from 41 HIV-infected individuals with completely suppressed HIV RNA levels in blood plasma (<50 copies/mL) by droplet digital PCR. We measured soluble CD14, soluble CD163, IP-10, MCP-1, IL-6, IL-8, TNF-α, neopterin, and neurofilament-light (NFL) by immunoassays in CSF supernatant or blood plasma.

Higher levels of mtDNA in CSF were associated with higher levels of MCP-1 (r=0.56, p<0.01) in CSF, and TNF-α (r=0.43, p<0.01), and IL-8 (r=0.44, p<0.01) in blood plasma. Subjects with a previous diagnosis of AIDS showed significantly higher levels of mtDNA (p<0.01) than subjects without AIDS. The associations between mtDNA and MCP-1 in CSF, and TNF-α in blood, remained significant after adjusting for previous diagnosis of AIDS (p<0.01). Additionally, higher levels of mtDNA were associated with a lower CD4 nadir (r=−0.41, p<0.01), and lower current CD4% (r=−0.34, p=0.03). Paradoxically, higher levels of mtDNA in CSF were significantly associated with better neurocognitive performance (r=0.43, p=0.02) and with less neuronal damage (i.e. lower NFL). Higher cell-free mtDNA is associated with inflammation during treated HIV-infection but the impact on neurocognitive functioning and neuronal damage remain unclear and may differ in the setting of suppressive ART.

Introduction

Persistent immune activation is the hallmark of human immunodeficiency virus (HIV) infection, and it is both associated with disease progression and can persist despite long-term suppressive antiretroviral therapy (ART) [1]. Immune activation is detectable shortly after infection in the brain and cerebrospinal fluid (CSF) of HIV-infected individuals [1]. During the course of infection, HIV enters the central nervous system (CNS) and replicates in permissive cells [2], resulting in an immune activation cascade that may contribute to neurodegeneration [3, 4], and neurocognitive impairment [5]. Apart from active replication, HIV viral proteins released from infected cells can be pro-inflammatory, resulting in further toxicity to bystander cells [4]. As a consequence of neurodegenerative processes, cellular components are also released into the cerebrospinal fluid (CSF). Some of these components, such as mitochondria, contain damage associated molecular patterns (DAMPs), which can stimulate the innate immune system and initiate a non-infectious inflammatory response [6]. These mitochondrial DAMPs include cell-free mitochondrial DNA (mtDNA), formyl peptides, and others [7]. Higher levels of cell-free mitochondrial DNA (mtDNA) in CSF are associated with poor outcome in the setting of traumatic brain injuries [7–9]. In HIV infection the release of mitochondrial DAMPs from dying cells may further contribute to inflammation and immune activation in the CNS.

Recently, we described associations between higher levels of mtDNA in CSF and increased markers of inflammation and neurocognitive deficits in a cohort of 31 HIV-infected individuals [10]. This previous cohort was limited by variable uptake of antiretroviral therapy (ART) and most individuals were not virally suppressed in their blood plasma and CSF at the time of sampling [10]. The ongoing viral replication may have confounded our ability to use mtDNA as a marker of neurodegeneration. To better understand the role of cell-free mtDNA in CSF as a biomarker of CNS inflammation and neurocognitive impairment, we evaluated the relationship between cell-free mtDNA in CSF, inflammation, and neurocognitive function in a well-characterized cohort of individuals on long-term ART with sustained viral suppression in blood plasma.

Methods

Ethics Statement

All adult subjects provided their written informed consent. No children were included in this study. The Office of Human Research Protections Program of the University of California, San Diego approved the study.

Study population and sample

This was a retrospective study of 41 HIV-infected individuals from the HIV Neurobehavioral Research Center (HNRC)’s Prospective Memory cohort [11]. At the time of sampling, all participants were on ART with undetectable levels of HIV RNA in blood (<50 copies/μL of plasma; Amplicor HIV Monitor Test, Roche Molecular Systems Inc.). Blood lymphocyte profiles were obtained by flow-cytometry (CLIA certified local laboratories). Epidemiological, behavioral risk and clinical data were also collected from participants [12]. Individuals were deemed to have AIDS if they met Centers for Disease Control criteria (http://www.aidsmap.com/CDC-case-definitions/page/1391604/). The estimated duration of infection (EDI) was determined using results of serologic and virologic tests as described previously [13]. Paired stored CSF and blood samples were retrospectively selected and used for measurements of CSF and blood plasma markers of inflammation and cell-free mitochondrial DNA.

CSF DNA Extraction and quantification

CSF was collected by lumbar puncture, centrifuged at 250g × 15 minutes to separate the supernatant from cells, and the supernatant was stored at −80°C. DNA was extracted from CSF supernatant samples using QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) per manufacturer’s protocol. Levels of mtDNA were measured by droplet digital PCR (ddPCR, Bio-Rad, Hercules, CA) using primer-probes combinations targeting the mitochondrial NADH dehydrogenase 2 gene (MT-ND2, Integrated DNA Technologies, IA) and the human genomic DNA (gDNA) by targeting the Ribonuclease P protein subunit p30 gene (RPP30, Integrated DNA Technologies, IA), using ZEN quenched probes [14]. Each sample was run in triplicate in a 20μL of reaction, which consisted of 10μL of 2× Bio-Rad supermix for probes, 1μL of either 20× Primer/FAM-Zen ND2 mix or 20× Primer/HEX-Zen RPP30 mix, 4μL of molecular grade water, and 5μL of total DNA, using the following conditions: (1) an initial activation of 95°C for 10 minutes, (2) 55 cycles of 94°C for 30 seconds and 60°C for 1 min, (3) enzyme inactivation at 98°C for 10 minutes, and 4°C hold. The end-point PCR reactions were read and analyzed using the Bio-Rad droplet reader and the QuantaSoft (Bio-Rad, CA) analysis software [15]. Levels of mtDNA were expressed in log₁₀ copies/mL of CSF.

Markers of innate immune activation, inflammation and neurofilament-light chain

The levels of selected markers of monocyte activation (neopterin, soluble CD163 [sCD163] and soluble CD14 [sCD14]), inflammatory cytokines (tumor necrosis factor-α [TNF-α], interleukin 6 [IL-6], interleukin 8 [IL-8]) and chemokines (monocyte chemoattractant protein 1 [MCP-1], interferon-γ-induced protein 10 [IP-10]) and, axonal damage (neurofilament-light chain [NFL]) were measured in all participants. Enzyme-linked immunosorbent assay (ELISA) was used to quantify the levels of sCD163 (Trillium Diagnostics, Brewer, ME, USA), sCD14 Quantikine ELISA Human sCD14 R&D Systems, MN, USA and neopterin (BRAHMS Neopterin EIA GmbH, Hennigsdorf, Germany) from blood plasma and CSF, and NFL in CSF (Uman Diagnnostics, Sweden). An electrochemiluminescence multiplex assay (Meso Scale Diagnostics, Rockville, MD, USA) was used to quantify the levels of TNF-α, IL-6, IL-8 and MCP-1, and IP-10 in CSF and blood plasma. All assessments were performed according to the manufacturer’s procedures.

Neuropsychological performance assessments

Individuals underwent neurocognitive testing using a standardized clinical battery testing seven ability areas consistent with Frascati recommendations for neuroAIDS research [16]. The details of this battery can be found in de Oliveira et al. [11]. All raw neurocognitive test scores were converted to demographically adjusted T-scores and averaged to create a summary T-score, which was used as our primary measure of neurocognitive performance.

Statistical analyses

Normality of the variables was assessed using a Shapiro test with a significance threshold of p ≤ 0.05. A two-tailed t-test was used to detect differences in the levels of cell-free mtDNA between AIDS vs. non-AIDS study groups. Univariate and multivariate associations between cell-free mtDNA in CSF supernatant and clinical and immunological variables were assessed using the Pearson correlation test and fixed effects regression analysis, respectively. All statistical analyses were performed using R statistical software [17]. Transformations and non-parametric analyses were performed if variables did not follow a normal distribution.

Results

Study participants characteristics

Our study cohort consisted of 41 HIV-infected subjects (31 males and 10 females), of which 66% were Caucasian (27/41), and with a median age of 51 years and a median CD4+ T cell count of 669 cells/μL at the time of sampling. Most subjects (40/41) were on ART but all had suppressed HIV RNA levels in blood plasma. A summary of the clinical characteristics of the subjects is provided in Table 1.

Table 1.

Characteristics of our study participants.

	Study participants (n=41)
Sex (M:F)1	31:10
Race (C:NC)2	27:14
Age	51 (33–58)
EDI (months)	170.9 (93.2–210.3)
CD4 Absolute	669 (482–893)
CD4 Percent	33 (24.4–39.7)
CD8 Absolute	918 (718–1195)
CD8 Percent	43.9 (35.1–52.5)
CD4 Nadir	237 (100–330)
CD4/CD8	0.83 (0.51–0.99)
mtDNA levels	3.99 (3.77–5.67)

¹M=Male and F=Female.

²C=Caucasian and NC=Non-Caucasian. Median with and interquartile range values are shown.

Cell-free mtDNA in CSF, inflammation and Immune Dysfunction

We first evaluated whether higher levels of cell-free mtDNA in CSF supernatant were associated with increased inflammation and with evidence of greater immune dysfunction (as defined by CD4 Nadir, or previous diagnoses of AIDS) in HIV-infected subjects on suppressive ART. All samples showed detectable levels of mtDNA in CSF (median: 3.99 log₁₀ copies/mL, Table 1). Univariate correlations demonstrated that higher levels of free mtDNA were associated with increased levels of MCP-1 in CSF (r = 0.56, p < 0.01, Figure 1A), and TNF-α (r = 0.43, p < 0.01, Figure 1B) and IL-8 (r = 0.44, p <0.01, Figure 1C) in peripheral blood plasma. Additionally, higher levels of mtDNA were associated with lower CD4+ T cell nadir (r = −0.41, p < 0.01, Figure 1D), and lower current CD4% (r = −0.34, p = 0.03, Figure 1E). Individuals with a previous diagnosis of AIDS had significantly higher levels of mtDNA (mean ± sd: 5.83 ± 1.56 log₁₀ copies/mL) when compared to those without AIDS (mean ± sd: 3.93 ± 0.35 log₁₀ copies/mL, p < 0.01, Figure 1F).

Figure 1. Mitochondrial DNA, inflammation and immunosuppression.

Higher levels of mtDNA in CSF were associated with higher levels of (a) MCP-1 in CSF, (b) TNF-α and (c) IL-8 in blood plasma. Additionally higher levels of mtDNA in CSF were associated with a lower (d) CD4 Nadir, and less CD4%. (e) Individuals with a previous AIDS diagnosis had higher levels of mtDNA in CSF than those without AIDS diagnosis.

In a multivariate model, only the relationships between mtDNA and CSF MCP-1 and TNF-α remained significant after adjusting for a previous diagnosis of AIDS (p <0.01 for both models, Table 2).

Table 2.

Multivariate associations of clinical variables while adjusting for AIDS status

	Clinical Variable p-value	AIDS p-value	R2
MCP-1 CSF	0.016	<0.001	0.52
TNF-α Plasma	0.021	<0.001	0.52
Mean T Score	0.037	<0.001	0.54
IP-10 CSF	0.131	<0.001	0.46
CD8 Percentage	0.145	<0.001	0.48
IL-8 Plasma	0.164	<0.001	0.46
Neopterin Plasma	0.185	<0.001	0.45
IP-10 Plasma	0.239	<0.001	0.47
Neopterin CSF	0.271	<0.001	0.47
CD4/CD8	0.335	<0.001	0.46
CD4 Absolute Count	0.347	<0.001	0.46
IL-6 CSF	0.376	0.002	0.39
CD8 Absolute Count	0.450	<0.001	0.46
sCD14 CSF	0.450	<0.001	0.46
NFL	0.515	<0.001	0.41
TNF-α CSF	0.565	0.032	0.74
MCP-1 Plasma	0.583	<0.001	0.44
IL-8 CSF	0.633	<0.001	0.45
sCD163 CSF	0.646	<0.001	0.45
sCD14 Plasma	0.708	<0.001	0.45
CD4 Percentage	0.717	<0.001	0.45
IL-6 Plasma	0.799	<0.001	0.47
CD4.Nadir	0.802	<0.001	0.45
sCD163 Plasma	0.998	<0.001	0.45

Although age was not directly associated with the levels of cell-free mtDNA in CSF, age has been previously associated with inflammation [18, 19]. In a multivariate regression analysis, levels of mtDNA remained associated with inflammation markers MCP-1 in CSF, and TNF-α and IL-8 in blood plasma (p < 0.01) after correcting for age.

Cell-free mtDNA in CSF and Neurocognitive Performance

We next evaluated if a higher level of cell-free CSF mtDNA was associated with worse neurocognitive impairment, as previously described [10]. In contrast to our previous work, and similar to other studies conducted in subjects with Alzheimer’s and Parkinson’s disease [20, 21], in this cohort of virally suppressed individuals better neurocognitive performance, as measured by the summary T-score, was correlated with higher cell-free mtDNA levels within the CSF supernatant (r = 0.43, p = 0.02; Figure 2). In a multivariate analysis higher levels of mtDNA remained associated with better neurocognitive performance after adjusting for AIDS (p=0.04, Table 2).

Figure 2. Associations of mtDNA and inflammation with neurocognitive function.

Higher levels of mtDNA in CSF were associated with better neuropsychological performance.

Cell-free mtDNA in CSF and Markers of Brain Damage

Next, we investigated the association between mtDNA and NFL (a marker of axonal neurodegeneration). We found that higher levels of NFL were associated with the lower levels of mtDNA in CSF even after adjusting for age (p=0.04), but this association was no longer significant after adjusting for AIDS status.

Discussion

In this cross-sectional study, we investigated the potential usefulness of cell-free mtDNA as a biomarker of HIV-associated CNS inflammation and neurocognitive impairment in the context of viral suppression with ART. We hypothesized that levels of free mtDNA in the CSF would reflect the levels of overall cell death in the CNS, and potentially contribute to the ongoing inflammatory response during suppressive ART.

Consistent with this hypothesis, we found that cell-free mtDNA in CSF samples from our cohort was strongly associated with systemic and CNS inflammation (MCP-1 in CSF, and TNF-α and IL-8 in blood plasma) during suppressive ART. These associations remained significant after adjusting for a previous diagnosis of AIDS. This was similar to our previous work [10], where we observed associations between mtDNA and inflammation in a heterogeneous cohort of HIV-infected individuals both on and off therapy. Thus, mtDNA may be a potential marker of ongoing CSF inflammation in HIV-infected individuals regardless of detectable viral replication.

In HIV infected individuals, persistent immune activation is common and associated with disease progression [22–24], even after long-term suppressive ART [1]. Persistent T-cell activation is associated with a blunted CD4+ T cell recovery during ART, and is a marker of progression to AIDS [21]. The cell death associated with the inflammatory process may lead to an increased release of mtDNA. In fact, we observed that individuals with a previous diagnosis of AIDS had significantly higher levels of cell-free mtDNA in the CSF compared to people without any AIDS-defining illness. Consistently, acutely infected HIV individuals and late presenters taking ART both have higher plasma levels of cell-free mtDNA when compared to healthy individuals or long-term non-progressors [22]. Similarly to bacterial DNA, mitochondrial DNA includes repeated CpG motifs, and can induce potent inflammatory reactions [7]. Cell-free mtDNA released during inflammatory cell death may induce further inflammation, leading to a persistent immune activation.

Next, we evaluated the associations between free mtDNA in CSF, neurodegeneration and neurocognitive performance. Contrary to our previous study [10], the current data show that higher levels of cell-free CSF mtDNA level was associated with less neurocognitive impairment in this cohort of virally suppressed subjects. This is despite the fact that mtDNA was associated with inflammation, and inflammation has been previously shown to be associated with NCI [2]. Previous studies have examined the levels of cell-free mtDNA in the CSF of individuals with neurodegenerative diseases in non-acute settings, as in Parkinson disease (PD) and Alzheimer disease (AD) [20, 21, 25]. Interestingly, in the individuals with PD or AD, the levels of cell-free mtDNA were lower than those in healthy individuals [20, 21], which is consistent with our findings. Additionally, we also found that decreased CSF mtDNA was associated with higher levels of NFL, another marker of neuronal degeneration. In the setting of ART suppressed HIV infection, there might be alterations in cellular and mitochondrial physiology, similar to other neurodegenerative diseases such as AD and PD, which leads to mtDNA depletion, neuronal damage (as suggested by the association with NFL), and worse neurocognitive outcomes. This supports the idea that in the context of suppressive ART, the lower levels of mtDNA maybe a marker of neurodegeneration. This association was not observable in non-suppressed individuals [10], where active viral replication and associated inflammation likely resulted in higher levels of released mtDNA, masking the above described relationship.

We recognize that this study has several limitations, including a small sample size and a lack of uninfected controls, which limit our conclusions. Further, future longitudinal studies are needed to better understand the interplay between free mtDNA and neurocognitive outcomes during HIV suppression. Finally, ex vivo and in vitro models approaches will be helpful to understand the mechanisms in which mitochondria and mtDNA may play a role on the development of neurocognitive dysfunction.

Despite these limitations, our provocative results suggest that despite being associated with increased inflammation, lower levels of cell-free mtDNA are associated with greater neurocognitive impairment and neurodegeneration during suppressive ART.