Loss of CCR2 Expressing Non-Classical Monocytes are Associated with Cognitive Impairment in Antiretroviral Therapy-Naïve HIV-Infected Thais

Abstract

HIV DNA in monocytes has been linked to HIV-associated neurocognitive disorders (HAND), however, characterization of monocyte subsets associated with HAND remains unclear. We completed a prospective study of antiretroviral therapy-naïve, HIV-infected Thais, with varying degrees of cognitive impairment, compared to HIV-uninfected controls. Monocyte subsets’ CCR2, CCR5 and CD163 expression were profiled and inflammatory markers in plasma and cerebrospinal fluid (CSF), measured. Lower numbers of CCR2⁺non-classical monocytes were associated with worse neuropsychological test performance (r=0.43,p=0.024). CCR2⁺non-classical monocyte count inversely correlated with CSF neopterin (r=−0.43,p=0.035) and plasma TNF-α levels (r=−0.40,p=0.041). These data benchmark CCR2⁺non-classical monocytes as an independent index of cognitive impairment.

Graphical Abstract

Lower numbers of CCR2 expressing non-classical monocytes (CCR2⁺Mono3^lo) correlated with worse neuropsychological test performance. These data highlight a new promising cellular biomarker of neurocognitive disease.

1. INTRODUCTION

The prevalence of HIV-associated neurocognitive disorder (HAND) remains high (estimated 30-50%) and persists despite plasma HIV RNA suppression with potent combination antiretroviral therapy (cART) (Tozzi et al. , 2009). These cognitive dysfunctions range in severity, spanning from the milder deficits of asymptomatic neurocognitive impairment (ANI) and mild neurocognitive disorder (MND) to more severe deficits in HIV-associated dementia (HAD). These cognitive deficits are not only widespread but impact everyday functioning, and increase morbidity and mortality with critical public health effects (Ellis et al. , 1997, Heaton et al. , 2004, Ickovics et al. , 2001). To date, there are no clinically proven therapies for HAND for individuals already on stable, virologically suppressive anti-HIV regimens (Shapshak et al. , 2011).

Our studies and those of others have highlighted the association between monocytes and HIV neuropathogenesis. We have shown a direct association between cognitive impairment and a greater number of cells harboring HIV DNA selected from cellular populations enriched with monocytes (CD14⁺) (Kusao et al. , 2012). Recent studies from others further document that in vitro, cytosolic HIV RNA in monocytes is mandatory for subsequent microglial activation and neurotoxicity (Faissner et al. , 2014). Studies in an SIV non-human primate model of AIDS, underlined the monocyte/macrophage trafficking into the CNS as a driver neuronal injury (Campbell et al. , 2014).

Monocyte nomenclature can be categorized by CD14 and CD16 expression into distinct populations (Gordon and Taylor, 2005, Zawada et al. , 2011): classical monocytes (Mono1; CD14⁺⁺CD16⁻), intermediate monocytes (Mono2; CD14⁺⁺CD16⁺), non-classical monocytes (Mono3; CD14⁺CD16⁺) (Ziegler-Heitbrock et al. , 2010) and a recently described 4^th ‘transitional’ monocyte subset (Mono4; CD14⁺CD16⁻) (Jalbert et al. , 2013a). These monocyte subsets have different inflammatory profiles (Jalbert, Crawford, 2013a) as well as migratory capacities (D’Antoni et al; unpublished data) and can be further subdivided according to additional cell surface receptor expression patterns (Wong et al. , 2012).

During the pre-cART era, high levels of several soluble factors were detected in circulation among subjects with HIV encephalitis and HIV dementia. These include monocyte chemoattractant protein-1 (MCP-1) (Bernasconi et al. , 1996, Conant et al. , 1998, Kelder et al. , 1998), neopterin (Brew et al. , 1990, Fuchs et al. , 1989, Sonnerborg et al. , 1989), tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β) (Sippy et al. , 1995, Tyor et al. , 1992) and interleukin-6 (IL-6) (Perrella et al. , 1992, Rusconi et al. , 2013). In patients on virally suppressive cART regimens, the levels of neopterin, TNF-α, MCP-1 and IL-6 have been documented to remain elevated and persist when compared to HIV uninfected subjects or those not on cART (Abdulle et al. , 2002, Cassol et al. , 2013, Probasco et al. , 2010, Sevigny et al. , 2004). Most of these soluble mediators, which can be secreted by monocytes and macrophages, have been associated with neurocognitive impairment in HIV-infected individuals on suppressive cART (Cassol, Misra, 2013, Sevigny, Albert, 2004).

The identification of new biomarkers for HIV-infected individuals with milder forms of HAND and among subjects who are suppressed on cART would afford an opportunity for mechanistic insight into the pathogenesis of HAND. Chemokine receptors have been suggested to play a role in HAND. We have shown in a single arm trial that intensification with a C-C chemokine receptor type 5 (CCR5) inhibitor improved cognition in cART suppressed cognitively impaired HIV-infected subjects (Ndhlovu et al. , 2014). Recent studies identified CCR2 on monocytes as a correlate to cognitive impairment in HIV-infected individuals (Williams et al. , 2014). CD163, a scavenger receptor, is emerging as a potential marker of activation and inflammation in HIV and a correlate of neurological injury. The frequency of CD163⁺ CD14⁺ CD16⁺ monocytes are increased in viremic HIV-infected individuals compared to HIV-uninfected individuals or ART suppressed aviremic HIV subjects (Fischer-Smith et al. , 2008).

Here we seek to discover new correlates to cognitive impairment in participants with HAND by utilizing multi-flow cytometry to evaluate distinct monocyte subsets, using CCR2, CCR5 and CD163 expression profiling among treatment naïve chronically HIV-infected subjects from Bangkok.

2. MATERIAL AND METHODS

2.1 Participant Characteristics

Twenty-nine participants were selected from a larger study designed to investigate HIV DNA as a marker of HAND among individuals who were cART-naïve and met Thai Ministry of Health guidelines to initiate therapy (SEARCH 011, clinicaltrials.gov: NCT00782808), as previously described (Valcour et al. , 2013). For this study, we randomly selected individuals with both high (≥ 1000 copies/10⁶ cells) and low (< 1000 copies/10⁶ cells) HIV DNA. The high HIV DNA included 14 participants (median=2137 copies/10⁶ cells; interquartile range (1255, 2658)) and the low HIV DNA, 15 participants (median=615 copies/10⁶ cells; interquartile range (294, 668)). HAND diagnosis: Participants were diagnosed by consensus conference as (1) cognitively normal (HIV NL), (2) ANI, (3) MND, or (4) HAD using established guidelines(Antinori et al. , 2007). Neuropsychological testing was performed on all participants. Age-matched HIV-uninfected Thai individuals were evaluated as controls.

NPZ Global Score: Subjects completed a 1-hour neuropsychological (NP) testing battery. All raw scores were transformed to standard z-scores using Thai normative data as previously described. A measure of global performance (NPZ-global) was calculated as the arithmetic mean of all tests with domain scores similarly calculated from domain-specific components of the battery. The neuropsychological battery included tests of attention and concentration, working memory and executive function, verbal and visual learning and memory, psychomotor and manual dexterity, motor, language, and visuospatial domains, as previously described (Schifitto et al. , 2007).

2.2 Monocyte Phenotyping

Blood was separated within eight hours of the blood draw and peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll-Hypaque density gradient centrifugation (Pfizer, Inc., New York, NY). Separated PBMCs and plasma were cryopreserved for future phenotypic analyses of cells and plasma cytokines. Cryopreserved PBMCs were thawed using RPMI 1640 (Hyclone, Logan, UT), supplemented with 10% fetal bovine serum, and placed in 96 well polypropylene round bottom plates. Cells were stained with Live/Dead® Fixable Aqua Dead Cell Stain for 15 min at room temperature followed by two, ten minute sequential incubations with anti-CCR2 (AlexaFluor647) and CCR5 (Phycoerythrin (PE)) at 37 °C (Jalbert et al 2013a, b). Cells were then stained at room temperature for 30 minutes with anti-CD3 (V500), CD14 (Qdot605), CD16 (AlexaFluor700), CD56 (PE-Cy7), CD19 (PE-Cy7), CD20 (PE-Cy7), human leukocyte antigen-D related (HLA-DR) (Allophycocyanin (APC)-H7) and CD163 (AlexaFluor488). Cells were fixed with 1% PFA and data were acquired on a custom 4-laser BD LSRFortessa (BD Biosciences, San Jose, CA). Compensation and gating analyses were performed using FlowJo (Treestar Inc, Ashland, OR). All antibodies were purchased from BD Bioscience (San Jose, CA), except for CD14-Qdot605 and yellow Live/Dead® Fixable Aqua Dead Cell Stain (Life Technologies, Grand Island, NY) and CD163-AlexaFluor488 (R&D Systems, Minneapolis, MN).

The gating strategy for identification of monocytes, their subsets, and receptor expression is shown in Figure 1. Monocytes were identified from the PBMC population by the exclusion of dead cells, lymphocytes (CD3 positive cells), natural killer (NK; CD56 positive) and B cells (CD19 or CD20 positive cells). Monocytes, positive for HLA-DR, were subset by CD14 and CD16 expression: Classical monocytes (Mono1; CD14⁺⁺CD16⁻), intermediate monocytes (Mono2; CD14⁺⁺CD16⁺), non-classical monocytes (Mono3^hi; CD14⁺CD16⁺ and Mono3^lo; CD14^loCD16⁺) and transitional monocytes (Mono4; CD14⁺CD16⁻). Monocyte populations were further assessed for CCR2, CCR5 and CD163 expression (Supplemental Figure 1). Monocyte numbers were calculated by multiplying monocyte number (Mono/μl) from the white blood cell count obtained from the clinical laboratory values and the percent frequency generated by flow cytometry.

Figure 1. Monocyte Identification and Phenotype Gating Strategy.

Top row: Singlets were identified using forward scatter (FSC) height (H) and area (A). Dead cells, staining positive for Yellow Reactive Amine Dye (YARD), were excluded and PBMCs were identified by size and granularity (side scatter (SSC) vs FSC). Middle row: CD3 positive cells (lymphocytes) were excluded, along with CD19-20-56 positive cells (B cells and natural killer cells). Cells positive for HLA-DR were identified as monocytes. Bottom row: HLA-DR positive cells were subset according to CD14 and CD16 expression. Mono1 are characterized as being CD14⁺⁺CD16⁻; Mono2, CD14⁺⁺CD16⁺; Mono3^hi, CD14⁺CD16⁺; Mono3^lo, CD14^loCD16⁺ and Mono4, CD14⁺CD16⁻.

2.3 Plasma Soluble Biomarkers

We evaluated the following circulating inflammatory biomarkers: neopterin, TNF-α, IL-6, IL-1β and MCP-1 in the plasma and CSF in this cohort. TNF-α, IL-6, IL-1β and MCP-1 were quantified in triplicate as part of a custom multiplex ELISA array according to the manufacturer's protocol (Quansys Biosciences, Logan, UT). Data were captured on the Odyssey infrared imaging system (Li-Cor Biosciences, Lincoln, NE) and analyzed using Quansys Q-view Plus software (Quansys Biosciences). Single-analyte ELISA was performed in duplicate to detect levels of neopterin (GenWay Biotech, San Diego, CA) and analyzed using SoftMax Pro (Molecular Devices, Sunnyvale, CA).

2.4 CD14 ⁺ monocyte isolation and HIV DNA quantification

Cryopreserved PBMCs were thawed in warm media (RPMI 1640 supplemented with 20 % fetal bovine serum), washed once, and resuspended in RoboSep buffer (StemCell Technologies, Vancouver, BC, Canada). Samples were placed in a RoboSep automated cell separator (StemCell Technologies) and CD14⁺ cells were purified through magnetic separation using the EasySep human monocyte enrichment kit without CD16 depletion (StemCell Technologies). DNA was extracted from CD14⁺ monocytes or total PBMCs using the QIAamp DNA Micro Extraction kit (Qiagen, Valencia, CA) and quantified using the ND-2000 spectrophotometer (NanoDrop Technologies, Wilmington, DE) as previously described (Shiramizu et al. , 2012). Determination of HIV DNA content was assessed using multiplex real-time PCR with HIV gag and β-globin primer pairs to amplify respective regions that were detected with FAM™-labeled HIV gag and VIC®-labeled β-globin probes. Using standard reference plasmids with one copy of the β-globin housekeeping gene and one copy of the HIV gag gene and appropriate positive/negative controls, samples were run in triplicate on StepOnePlus Real-Time PCR System and analyzed using the StepOne software (Applied Biosystems, Foster City, CA). The copy numbers of each sample gene were analyzed against the standard curves and used to calculate HIV DNA copy number per 1×10⁶ cells.

2.5 Statistical Analyses

The demographic and clinical information of participants were summarized by percentage for gender and by median and interquartile range (IQR) for continuous variables, e.g. CD4 count. The gender comparisons were examined by Fisher's exact tests. For continuous variables, Kruskal-Wallis tests were performed to compare multiple neuropsychological diagnosis groups. Wilcoxon-Mann-Whitney tests were conducted to compare two different groups, e.g., monocyte levels between HIV-uninfected and infected subjects, or between HIV NL subjects and HIV HAND subjects, without multiple comparison adjustments. The association analyses between two continuous variables were evaluated by Spearman correlation coefficients, e.g. between NPZ global score and a monocyte subset. The association between NPZ global score and CCR2⁺ Mono3^lo subpopulation was further explored by multiple linear regression, adjusting for Mono3^lo, HIV DNA from CD14⁺ cells, and plasma viral load. Model statistical assumptions were checked. Statistical analyses were performed using SAS software version 9.3 (SAS Institute Inc., Cary, NC). A two-sided p-value < 0.05 was regarded as statistically significant.

3. RESULTS

3.1 Clinical Cohort

We evaluated 29 HIV-infected Thai participants who were classified by Frascati HAND criteria to have normal cognition (NL; n=17), Asymptomatic Neurocognitive Impairment (ANI; n=6), Mild Neurocognitive Disorder (MND; n=1) and HIV-associated Dementia (HAD; n=5) (Table 1). Within the HIV-infected participants, there were no differences between the HAND and the non-HAND groups in terms of education, plasma viral load, CD4 count and CD8 T cell activation (Table 1). NPZ global score and HIV DNA were significantly different between non-HAND and HAND (ANI and MND+HAD) groups (p=0.0002 and p=0.016, respectively; Table 1). Both the HIV-infected and uninfected individuals were gender and aged matched (Table 1). The frequency of activated CD8⁺ T cells (HLA-DR⁺ CD38⁺) was significantly greater in the HIV–infected individuals compared to the HIV-uninfected group (p<0.0001; Table 1).

Table 1.

Clinical and Demographic Characteristics of Participants (n = 73)

	HIV-Uninfected		HIV-Infected
	n = 44	P-value* (HIV− vs. HIV+)	Normal	HAND (HIV-associated neurocognitive disorders)		P-value** (Comparing 3 groups within HIV+)
			n = 17	n = 6 ANI = 6	n = 6 MND = 1 HAD = 5
Gender (% male)	45.5%	p = 0.48	64.7%	50.0%	33.3%	p = 0.46
Age (years)	35.1 (32.2, 38.2)	p = 0.10	33.2 (28.9, 36.5)	33.8 (26.4, 39.6)	33.3 (29.3, 37.1)	p = 0.97
Education (years)	-	-	11 (9, 15)	14 (12, 19)	6 (2, 14)	p = 0.11
NPZ global score	-	-	0.277 (0.086, 0.569)	−0.280 (−0.451, 0.026)	−1.562 (−1.620, −1.466)	p = 0.0002
Log10 Plasma Viral Load (copies/ml)	-	-	4.998 (4.350, 5.469)	4.914 (4.405, 5.161)	5.565 (4.760, 5.589)	p = 0.33
Log10CD14+ Total HIV DNA	-	-	1.342 (1.255, 1.568)	1.831 (1.663, 2.199)	2.090 (2.025, 2.152)	p = 0.016
CD4 count (cells/μL)	-	-	239 (159, 328)	331 (261, 397)	190 (30, 310)	p = 0.14
CD8 T cell activation (%)	9.1 (6.8, 12.2)	p < 0.0001	35.9 (31.5, 42.3)	44.7 (38.9, 48.5)	36.8 (35.2, 53.5)	p = 0.60

Data are shown as median (interquartile range), except for gender.

^*P-values calculated by Fisher's exact test or Wilcoxon-Mann-Whitney test

^**P-values calculated by Kruskal-Wallis tests

3.2 Monocyte subsets and HAND diagnosis

Monocytes were categorized into five different subsets based on CD14 and CD16 expression (Figure 1). Cell surface expression for the chemokine receptors, CCR2 and CCR5 and, the scavenger receptor, CD163 were measured on all monocyte subsets (Supplemental Figure 1). By monocyte subset, the HIV-infected HAND group had lower Mono1 and CCR2 expressing Mono1 (CCR2⁺ Mono1) numbers compared to the HIV-uninfected group (p = 0.0013 and p = 0.0040, respectively; Figure 2A, B). The group of HIV-infected participants with normal cognition had the lowest number of Mono1 and CCR2⁺ Mono1, which differed from that of the HIV-uninfected group (p < 0.001 and p < 0.001, respectively) and the HIV-infected HAND group (p = 0.034 and p = 0.048, respectively) (Figure 2A, B).

Figure 2. Monocyte subset numbers and HAND diagnosis.

(A-J) Monocyte subset numbers (left column) and the number of CCR2⁺ monocytes sorted by subset (right column), expressed as cell number per μl, were compared in HIV-uninfected individuals with normal cognition, HIV-infected individuals with normal cognition (normal) and HIV-infected individuals with HAND.

We observed a significantly lower number of Mono2s in the HAND group compared to the HIV-uninfected participants (p = 0.0018; Figure 2C). The numbers of Mono2 and CCR2⁺ Mono2 cells in the HIV-infected normal group were significantly lower than the HIV-uninfected group (p < 0.001 and p = 0.020, respectively; Figure 2C, D). No differences were seen between Mono2 and CCR2⁺ Mono2 numbers between the HIV-infected normal and HIV-infected HAND groups (Figure 2C, D).

As seen with the previous subsets and compared to the HIV-uninfected individuals, we noted a lower number of Mono3^hi and CCR2⁺ Mono 3^hi in the HIV-infected normal group (p < 0.001 and p = 0.036, respectively; Figure 2E, F) and in the HIV-infected HAND group (p = 0.0057 and p = 0.0065, respectively; Figure 2E, F). No differences were seen in the numbers of Mono3^hi and CCR2⁺ Mono 3^hi between the HIV-infected HAND and HIV-infected normal groups (Figure 2E, F). There were no significant differences between the Mono3^lo populations across groups, however, the HIV-infected HAND group had a significantly lower number of CCR2⁺ Mono 3^lo subsets compared to the HIV-infected normal group (p = 0.031) and the HIV-uninfected group (p = 0.0056 (Figure 2G, H). The HIV-infected normal group also had significantly lower CCR2⁺ Mono 3^lo numbers than that of the HIV-uninfected group (p = 0.032; Figure 2H). There were no significance differences observed with the Mono4 cells however the HIV-infected HAND group had lower CCR2⁺ Mono4 numbers compared to the HIV-uninfected individuals (p = 0.025; Figure 2I, J).

3.3 Relationship between neuropsychological function and circulating monocyte subsets

The NPZ global score directly correlated with the number of the non-classical monocyte subset, Mono3^lo, (r = 0.43, p = 0.024), as well as the number of CCR2⁺ Mono3^lo (r = 0.43, p = 0.024) (Figure 3A) among HIV-infected participants. We did not find other correlations between the NPZ global scores and any other monocyte subset expressing the various receptors we measured (data not shown). The number of Mono3^lo cells directly correlated with the NPZ-psychomotor score (r = 0.44, p = 0.022; Figure 3B) and a trend was noted with the NPZ-motor score (r = 0.37, p = 0.060; data not shown). The number of CCR2 expressing Mono3^lo directly correlated with the NPZ-psychomotor score (r = 0.47, p = 0.014; Figure 3B) and the NPZ-motor score (r = 0.38, p = 0.050; Figure 3C). On the corollary, CCR2⁻ Mono3^lo numbers, correlated with both the NPZ global and NPZ-psychomotor score (r = 0.43, p = 0.025 and r = 0.44, p = 0.022, respectively, data not shown).

Figure 3. Correlations between Mono3^lo subsets and neuropsychological testing data.

Correlations between Mono3^lo and CCR2⁺Mono3^lo subsets, expressed as cell number per μl, and NPZglobal score (A), NPZ-psychomotor score (B) and NPZ-motor score (C), were assessed.

3.4 Correlations between plasma and cerebrospinal fluid (CSF) inflammatory markers, monocyte HIV DNA content and monocyte subsets

Statistically significant negative correlations between the inflammatory markers, CSF neopterin and plasma TNF-α levels, and the number of CCR2⁺ Mono3^lo cells were observed (r = −0.43, p = 0.035 and r = −0.40, p = 0.041; Figure 4A, B). No significant correlations were observed between CSF neopterin or plasma TNF- α levels and the number of monocyte subsets, including Mono3^lo and CCR2⁻ Mono3^lo populations (data not shown). Plasma TNF-α levels also inversely correlated with the number of CD163⁺ Mono3^locells (r = −0.41, p = 0.035; Figure 4C). There was also significant correlation between CSF IL-6 levels and Mono3^hi numbers (r = −0.45, p = 0.027) (data not shown). There were no correlations found between any of the monocyte subsets evaluated and CSF and plasma IL-1β or MCP-1 levels (data not shown). There were no significant correlations however we did observe a positive trend between the total number of classical monocytes (Mono1) and total HIV DNA within the CD14⁺ bead selected monocytes (r = 0.34, p = 0.067) (data not shown).

Figure 4. Correlations between Mono3^lo subsets and inflammatory marker levels.

Correlations between the CCR2⁺Mono3^lo subset, expressed as cell number per μl, and neopterin CSF levels (pg/μl; A); and TNF-α plasma levels (pg/μl; B) were assessed. Correlations between CD163⁺Mono3^lo number and plasma TNF-α (pg/μl; C) were also assessed.

3.5 Multiple linear regression analysis

A multiple linear regression analysis was performed to further explore the relationship between neuropsychological testing performance (NPZ global) and the number of monocyte populations and subpopulations, HIV DNA from CD14⁺ enriched peripheral blood mononuclear cells (PBMCs) and plasma viral load. In this model, the number of CCR2⁺ Mono3^lo cells was the only independent predictor of NPZ global (p = 0.043; Table 2), which was consistent with the earlier simple regression analysis shown in Figure 3B.

Table 2.

Multiple linear regression analysis

R-Square	Coeff Var	Root MSE	NPZglobal Mean
0.328	−708.2	0.661	−0.093343

Parameter	Estimate	Standard Error	t Value	p-value
Intercept	1.166	0.956	1.22	0.2356
CCR2 Mono6 count	27.497	12.809	2.15	0.0431
Mono6 count	−0.010	0.014	−0.73	0.4752
Log10CD14+ HIV DNA	−0.477	0.311	−1.53	0.1399
Log10 plasma viral load	−0.145	0.194	−0.75	0.4627

4. DISCUSSION

Monocytes are thought to contribute to HAND pathogenesis by mediating HIV-neuroinvasion, which ultimately leads to HIV in the brain tissue, neuro-inflammation and neuronal damage. Many studies have focused on soluble factors, such as cytokines and chemokines, which might contribute to pathogenesis. Data regarding receptors are still lacking, as assessing multiple cell surface chemokine receptors can be technically challenging. We have developed a flow cytometry based assay to measure several cell surface receptors simultaneously (including chemokine and scavenger receptors) on cryopreserved PBMCs.

When examining the different receptors found on the monocyte subsets in this cohort, our data show unexpectedly, that lower numbers of a unique subset of non-classical monocytes, with low CD14 and high CD16 expression and more specifically, those expressing CCR2 in the periphery were linked to cognitive impairment among treatment naïve individuals. The loss of this population in circulation may represent an increased transmigration of this subset into tissue sites. Since CCR2⁺ Mono3^lo prominently correlated with clinical performance and given the patrolling nature of non-classical monocytes (Wong, Yeap, 2012) one may hypothesize that the CCR2⁺ Mono3^lo subset, may be important in mediating neuroinvasion into the CNS compartment.

Others have identified CCR2 as being an important predictor of HAND. In an in vitro blood brain barrier (BBB) model, Williams et al (2014) demonstrated that CCR2 mean fluorescence intensity (MFI) was significantly higher on the intermediate monocyte subset population from participants with HAND compared to unimpaired HIV-individuals. The functional consequence to this difference was an increase in migration across the BBB toward an MCP-1 gradient. The discrepancies in the specific monocyte subset identified in relation to HAND may reflect differences in the cohort and experimental parameters examined between studies. Interestingly, a recent study by Uleri et al document, using an in vitro model system, that HIV Tat may promote HIV neuroinvasion in vivo through increases in CD16 and CCR2 on monocytes (Uleri et al. , 2014). It would be interesting in future studies to determine the HIV content in these unique minor monocytoid subsets as triggers for neuroinvasion and neuroinflammation.

A clear consensus on human monocyte nomenclature has not yet been achieved including agreement on how to assess and quantify this circulating population using single cell flow cytometric techniques (Abeles et al. , 2012, Tallone et al. , 2011, Wong, Yeap, 2012). A number of studies have documented the frequency of monocytes populations in limited depth in HAND (Ancuta et al. , 2008, Pulliam et al. , 1997). When calculated by number of monocytes, our data document changes in the proportion of monocyte subsets in HAND, a finding that differs from other studies. This could be due to the distinct populations assessed but also on the design and depth of the immunophenotyping presented (Abeles, McPhail, 2012, Ndhlovu, Umaki, 2014). We have observed that to accurately assess chemokines on monocytes due to their recycling nature sequential and coordinated methods of staining are required (Jalbert et al. , 2013b). Our recent studies presented here have assessed monocytes using a strategy adopted by several groups using HLA-DR gating to obtain better resolution of these populations in the setting of disease (Abeles, McPhail, 2012).

Neopterin is a soluble factor that reflects the intensity of monocyte/macrophage activation and has shown utility as a biomarker of HIV-neuropathology (Hagberg et al. , 2010, Valcour, Ananworanich, 2013). We found that CCR2 expressing monocytes were associated with several measured soluble biomarkers. There was an inverse correlation between the number of CCR2⁺Mono3^lo and CSF neopterin and plasma TNF-α levels. That is the fewer CCR2⁺Mono3^lo in circulation, the greater the peripheral inflammatory profile. Given these data, one might conclude that the greater presence of CCR2⁺ Mono3^lo may serve a protective role in cases where HAND exists. Although their presence is associated with less inflammation for cytokines important to HAND, in this study, they also correlate with disease severity (based on NPZ global score). Indeed CCR2 expressing monocytes have been shown to have greater sensitivity in entry to the CNS (Williams et al. , 2013). On the corollary however, a lower frequency in the circulation of these subsets may suggest increased seeding into the brain through a CCR2-CCL2 driven mechanism resulting in a greater inflammatory environment as revealed by higher levels of neopterin and TNF-α. Recent studies have identified a novel biomarker, 6-Sulfo LacNAc (SLAN) that is linked to the secretion of TNF-α by Mono3 non-classical monocytes (Dutertre et al. , 2012). Given our data, it appears that the entire Mono3^lo population may be an important marker of neurocognitive impairment but that the CCR2⁺ Mono3^lo cells may be particularly noteworthy as these cells correlate with the inflammation associated with cognitive impairment. Whether this unique subset is responsible for the release of these inflammatory mediators warrants further investigation. Taken together our data suggest that CCR2⁺Mono3^lo may serve a novel biomarker for assessing the degree of HAND.

The strength of this study is the ability to correlate multiple monocyte cellular subpopulations and clinical neurocognitive parameters in subjects with HAND. Although the cohort is small, it has highlighted new promising cellular biomarkers of neurocognitive disease. Moreover, it is interesting to note that the HIV-infected individuals analyzed in this study were not on cART and may benefit from additionally from cART and chemokine mediated blockade targeting peripheral monocytes.

Highlights.

• Monocyte HIV DNA correlates with HAND, which persists despite cART.

• Monocyte subset characterization in the context of HAND remains incomplete.

• Lower CCR2⁺ non-classical monocyte number associated with worse NPZ testing.

• CCR2⁺ non-classical monocyte count inversely correlated with neopterin and TNF-α.

• Loss of CCR2⁺ non-classical monocytes may serve as a biomarker for HAND.

Abbreviations

CCR2

C-C chemokine receptor type 2

CSF

cerebrospinal fluid

HAND

HIV-associated neurocognitive disorders

TNF-α

tumor necrosis factor-α

A

area

AIDS

acquired immune deficiency syndrome

APC

Allophycocyanin

ANI

asymptomatic neurocognitive impairment

BBB

blood brain barrier

cART

combination antiretroviral therapy

CCR2

C-C chemokine receptor type 2

CCR5

C-C chemokine receptor type 5

CNS

central nervous system

CSF

cerebrospinal fluid

ELISA

enzyme-linked immunosorbent assay

FSC

forward scatter

H

height

HAD

HIV-associated dementia

HAND

HIV-associated neurocognitive disorders

HIV

human immunodeficiency virus

HLA-DR

human leukocyte antigen-D related

IL

interleukin

IQR

interquartile range

MCP-1

monocyte chemoattractant protein-1

MFI

mean fluorescence intensity

MND

mild neurocognitive disorder

Mono

monocyte

NK

natural killer

NL

cognitively normal

NPZ

neuropsychological performance

PBMC

peripheral blood mononuclear cell

PCR

polymerase chain reaction

PE

Phycoerythrin

SIV

Simian immunodeficiency virus

SSC

side scatter

SLAN

Sulfo LacNAc

TNF-α

tumor necrosis factor-α

YARD

Yellow Reactive Amine Dye