Secondary Logo

Journal Logo

Advanced Search
BASIC SCIENCE

Plasma microRNA profiling predicts HIV-associated neurocognitive disorder

Asahchop, Eugene L.; Akinwumi, Segun M.; Branton, William G.; Fujiwara, Esther; Gill, M. John; Power, Christopher

Author Information

aDepartment of Medicine (Neurology)

bDepartment of Mathematical and Statistical Sciences

cDepartment of Psychiatry, University of Alberta, Edmonton, Alberta

dSouthern Alberta Clinic

eDepartment of Medicine, University of Calgary, Calgary

fDepartment of Medical Microbiology & Immunology, University of Alberta, Edmonton, Alberta, Canada.

Correspondence to Dr Christopher Power, Division of Neurology, Department of Medicine, 6–11 Heritage Medical Research Centre Edmonton, AB, Canada T6G2S2. Tel: +1 7804071938; fax: +1 7804071984; e-mail: [email protected]

Received 23 December, 2015

Revised 4 May, 2016

Accepted 8 May, 2016

Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's Website (http://www.AIDSonline.com).

doi: 10.1097/QAD.0000000000001160

Abstract

Introduction

HIV type 1 (HIV-1) enters the central nervous system (CNS) soon after primary infection and infects macrophage lineage and astrocytic cells within the CNS [1,2]. This process establishes a viral reservoir within the brain early in the course of the disease [3]. As the infection progresses, a subset of persons manifest a neurodegenerative syndrome, termed HIV-associated neurocognitive disorder (HAND) that is caused by HIV-1 infection [4]. HAND is a progressive condition in the absence of treatment and is classified as asymptomatic neurocognitive impairment (ANI), minor neurocognitive disorder (MND) and HIV-associated dementia (HAD). HAND is common, affecting approximately 25% of HIV-infected patients, despite the availability of effective combination antiretroviral therapy (cART) [4–7]. The prevalence of HAND is reported to be higher in sub-Saharan Africa where access to treatment is limited (reviewed in [8]). The development of HAND among HIV-infected patients is a multifactorial process influenced by changes in host and viral gene expression [9–12], altered host immune responses [13,14] as well as cART (reviewed in [15]) leading to reduced treatment adherence, poor quality of life and worsened survival [16]. Specifically, HAND results from the production and release of HIV-1 viral proteins (e.g. gp120, tat and vpr) and host inflammatory molecules (e.g. cytokine, chemokines, protease and free radicals) from infected and activated glia and perivascular macrophages that can induce apoptosis and mitochondria DNA damage in neurons [17–20]. The induction of mitochondria DNA damage by antiretroviral (ARV) drugs such as nucleoside reverse transcriptase inhibitors has also been reported [21]. Mitochondrial DNA damage is a major cause of neurodegeneration in vascular and metabolic diseases such as type diabetes and cardiovascular diseases [22,23]. Thus, there are pathogenic mechanisms that are common to HAND and metabolic diseases, as a consequence of mitochondrial DNA damage.

microRNAs (miRNAs) belong to a class of noncoding RNA molecules (18–22 nucleotides) that participate in the regulation of host (and viral) gene expression [24]. Both host and viral miRNA profiling can yield insights into the underlying disease mechanisms [25–27] as well as serving as useful biomarkers for diagnosis, prognosis and response to treatment [28,29]. Recent studies have reported that miRNA profiles differ in brain between HIV-infected individuals with HAND and non-HAND and in blood from HIV-infected elite controllers and viremic patients [30,31]. Thus, host miRNAs might contribute to the pathogenesis of HAND but also represent biomarkers of diagnosis and prognosis. Recent reports have suggested that free miRNAs in biofluids [e.g. plasma, serum, urine, saliva and cerebrospinal fluid (CSF)] are stable and protected from endogenous RNase activity by carrier proteins or are packaged into microvesicles [32–35]. Circulating levels of individual miRNAs in plasma can be indicative of specific diseases when compared with miRNA levels in control patients [28,36–39]. The diagnosis of HAND is based on clinical assessment and neuropsychological testing in combination with neuroimaging and CSF analyses as indicated. Currently, there are no validated biological laboratory markers for establishing the diagnosis of HAND. As circulating miRNAs in plasma are stable, readily detected and also have the potential to exert gene-specific actions, they might represent useful biomarkers for HAND. Thus, we hypothesized that plasma miRNA profiles in HAND compared with non-HAND HIV/AIDS patients could be a distinguishing biomarker for HAND and perhaps also contribute to the development of the neurological syndrome. Plasma miRNA expression profiles in HIV/AIDS patients with and without HAND were investigated, which revealed differential expression of three miRNAs in plasma that distinguished HAND from non-HAND patients in two cohorts.

Materials and methods

Patient cohorts and HIV-associated neurocognitive disorder diagnosis

All patients were HIV-1 seropositive and receiving active care at the Southern Alberta Clinic, Calgary, Alberta, Canada [40–43]. The diagnosis of HAND was based on established criteria (neurological, neuroimaging and neuropsychological assessments) [44]. See Supplementary Methods for details of patient recruitment and HAND diagnosis, https://links.lww.com/QAD/A935.

Experimental design

In plasma samples from the Discovery Cohort (non-HAND, n = 25; HAND, n = 22: ANI, n = 0; MND, n = 11 and HAD, n = 11), mature miRNAs were measured using the Affymetrix 3.0 array hybridization chips (Affymetrix, Santa Clara, California, USA) (Supplementary Fig. 1, https://links.lww.com/QAD/A935) from which differentially expressed miRNAs were identified in HAND versus non-HAND groups. In the Validation Cohort (non-HAND, n = 12; HAND, n = 12: ANI, n = 0; MND, n = 6 and HAD n = 6), plasma samples were used for validation of differentially expressed miRNAs identified in the Discovery Cohort using the same methods. The miRNAs showing altered expression in both the Discovery and Validation Cohorts were selected for further analyses. In addition, the expression of the common miRNAs was also verified in the Validation Cohort samples by quantitative real-time reverse transcriptase PCR (qRT-PCR).

microRNA extraction from plasma samples

Total RNA was extracted from whole blood collected in EDTA tubes using miRNeasy serum/plasma kit (Qiagen, Toronto, Ontario, Canada) according to manufacturer's instruction with modification. Complete details of protocol are found in Supplementary Methods section, https://links.lww.com/QAD/A935.

microRNA microarray

The Affymetrix miRNA array 3.0 GeneChip was used for array hybridization. The 3.0 GeneChip contains human mature miRNA (n = 1733) and premiRNA (n = 1658) probes. The Affymetrix FlashTag Biotin highly sensitive and reproducible (HSR) RNA Labeling Kit was used to label samples for miRNA analysis. Briefly, 8 μl of undiluted RNA was poly-A tailed followed by biotin-HSR ligation. The labeled samples were hybridized to Affymetrix miRNA 3.0 GeneChip array at 48 °C for 24 h. Arrays were stained and washed on Affymetrix GeneChip Fluidics 450 following manufacturer's protocol and scanned with Affymetrix GeneChip Scanner 3000 7G System.

Quantitative real-time reverse transcription PCR

The expression of miRNAs, identified from array hybridization studies, was verified by qRT-PCR on a BioRad IQ 5 cycler (BioRad, Richmond, California, USA). A detailed description of the methodologies is provided in Supplementary Methods section, https://links.lww.com/QAD/A935.

Statistics and bioinformatics analyses

Two independent software packages including Affymetrix Expression and Transcription Console (ETC) and GeneSpring (version 12.6, Agilent Technologies) were used to normalize the data and determine differentially expressed miRNAs. The normalization in both software packages was based on the Robust Multiarray Average algorithm, in which data are background-corrected, log 2 transformed and quartile normalized. To identify differentially expressed miRNAs, the median of each probe set in the HAND or non-HAND patients was calculated and the significance of any differences determined by Mann–Whitney unpaired test on data subjected to Affymetrix and GeneSpring normalization, respectively. A cut-off fold change (≥2) in relative miRNA abundance was applied, and a P value of 0.05 or less was considered statistically significant. This was followed by a correction for multiple comparisons. See Supplementary Methods for complete statistics and bioinformatics analyses, https://links.lww.com/QAD/A935.

Results

Cohort features

In the present study, two cohorts of patients were examined including the Discovery Cohort (Table 1), which was derived from HIV/AIDS patients previously assessed and diagnosed with HAND or non-HAND (1998–2013), whereas in the Validation Cohort (Table 2 ), patients were recruited prospectively beginning in 2013 and diagnosed with either HAND or non-HAND. HAND and non-HAND patients in the Discovery Cohort were matched in terms of mean age, sex and education but differed significantly in mean peak viral load and nadir CD4+ T-cell counts (Table 1). In the Validation Cohort, all patients were prospectively (and randomly) recruited from the same clinic but HAND and non-HAND patients differed significantly with respect to ethnicity, mean education, nadir CD4+ T-cell count and CD4+ T-cell count at the time of assessment (Table 2). Of note, mean duration of HIV-1 seropositivity, frequency of hepatitis C Virus infection (including viremia) or substance abuse, HIV-1 subtypes, polypharmacy, mean Central Nervous System Penetration Effectiveness (CPE) ranking scores, antiretroviral count numbers and complete blood count results did not differ between HAND and non-HAND groups for both the Discovery and Validation Cohorts. The differences between cohorts reflected changes in the clinic's patient demographics during the study period (1998–2015).

T1-3
Table 1:
Sociodemographic and clinical features of HIV/AIDS patients in the Discovery Cohort (n = 47).
T2-3
Table 2:
Sociodemographic and clinical features of HIV/AIDS patients in the Validation Cohort (n = 24).

Differentially expressed microRNAs in HIV-associated neurocognitive disorder compared with non-HIV-associated neurocognitive disorder patients

For both Cohorts, plasma miRNA profiles were analyzed and compared using GeneSpring or Affymetrix ETC with similar findings. MiRNA profiles in plasma were analyzed, subjected to statistical assessment (a fold change ≥2.0 or ≤−2.0 with associated P value ≤0.05) and verified (Supplementary Fig. 1, https://links.lww.com/QAD/A935). In the Discovery Cohort, this comparative analysis identified nine miRNAs that were upregulated in HAND compared with the non-HAND group, whereas no miRNAs were downregulated at this cut-off level (Fig. 1a, Supplementary Tables 1 and 2, https://links.lww.com/QAD/A935). Hierarchical clustering analysis showed a distinct miRNA profile for HAND compared with the non-HAND patients for the nine differentially expressed miRNAs. Next, we corrected for multiple comparisons (Bonferroni) among the identified miRNAs. This analysis identified four miRNAs that passed the correction for multiple comparisons including miR-3196, miR-3656, miR-3940–5p and miR-4687–3p. To extend these analyses, miRNA profiles in the Validation Cohort were examined, revealing that the miRNA profiling based on GeneSpring or Affymetrix ETC were similar but 14 miRNAs were upregulated in HAND compared with the non-HAND patients (Fig. 1b, Supplementary Tables 3 and 4, https://links.lww.com/QAD/A935), whereas 39 miRNAs were downregulated in the HAND group (Fig. 1c, Supplementary Tables 3 and 4, https://links.lww.com/QAD/A935). The miRNA profiles in each cohort were compared with identify common upregulated miRNAs in both cohorts based on either GeneSpring or Affymetrix ETC. This comparison yielded three miRNAs that were increased significantly in plasma from HAND patients in both cohorts including miR-3665, miR-4516 and miR-4707–5p (mean fold change 2.4, 3.3 and 2.9, respectively) (Fig. 1d).

F1-3
Fig. 1:
Differentially expressed microRNAs in the Discovery and Validation Cohorts.(a) Hierarchical cluster analysis of upregulated microRNAs in HIV-associated neurocognitive disorder (HD) group compared with non-HIV-associated neurocognitive disorder (nonHD) in the Discovery Cohort. The upregulated microRNA profile clusters patients into HD and nonHD groups. The top dendrogram further suggests that two groups underlie the HD clustering. The side dendrogram shows three major clusters of microRNAs in the upregulated profile. The microRNAs, hsa-miR-663 and hsa-miR-4516, form a group that is most dissimilar from the others, whereas hsa-miR-3940–5p and hsa-miR-149–3p are the most similar followed by the pair hsa-miR-3196 and hsa-miR-1228–5p. (b) Hierarchical cluster analysis of upregulated microRNAs in HD group compared with nonHD in the Validation Cohort. The top dendrogram shows that there are three substantial clusters in the upregulated microRNA profile of the Discovery Cohort, one cluster consists of the HD patients and the two other clusters are a mix of HD and nonHD with nonHD being predominant. The side dendrogram shows that microRNA hsa-miR-3613–5p is the most dissimilar compared with the others. In contrast, hsa-miR-3960 and hsa-miR-4787–5p are the most similar followed by the pair hsa-miR-638 and hsa-miR-4487. (c) Hierarchical cluster analysis of downregulated microRNA in HD group compared with nonHD in the Validation Cohort. There are three clusters in the downregulated microRNA profile of the Validation Cohort, one cluster consists of the nonHD patients and the two other clusters are a mix of HD and nonHD with HD being predominant (top dendrogram). Four groups of clustered microRNAs are shown in downregulated microRNA profile of the Validation Cohort (side dendrogram). The most similar of the group consisted of the microRNAs from hsa-miR-433 and hsa-miR-134 inclusive and the most dissimilar group consist of the microRNAs from hsa-miR-451 to hsa-miR-652 inclusive. Cluster analyses were based on GeneSpring-derived values; spectral values: +4.0 (orange) to −4.0 (blue). (d) Comparison of relative expression of microRNAs associated with HIV-associated neurocognitive disorder in both cohorts based on GeneSpring and Affymetrix ETC.

Validation of HIV-associated neurocognitive disorder-associated microRNAs

As the above results were predicated on analyses of array-derived data, it was important to verify the expression levels using another technique. A qRT-PCR assay was used to measure the expression level of miRNAs that were associated with HAND status in the Validation Cohort. Analyses of miR-3196, miR-3656, miR-3940–5p and miR-4687–3p relative to miR-16–5p expression showed that these miRNAs were not upregulated in the Validation Cohort's HAND group. In contrast, primers for miR-3665 and miR-4516 amplified each miRNA with a median relative fold increases in expression for miR-3665 (1.8, P < 0.05) (Fig. 2a) and miR-4516 (2.4, P < 0.05) (Fig. 2b) in the HAND compared with the non-HAND group in the Validation Cohort. (The primers targeting miR-4707–5p was not sufficiently sensitive to amplify this miRNA.) The sensitivity and specificity of the two former miRNA-based expression levels measured by qRT-PCR in diagnosing HAND were assessed by receiver-operating characteristic (ROC) curve showing that each miRNA was able to predict HAND status with area under the curve (AUC) of 0.76 [95% confidence interval (CI), 0.54–98] for miR-3665 (Fig. 2c) and 0.79 (95% CI, 0.59–0.98) for miR-4516 (Fig. 2d). A combined AUC for miR-3665 and miR-4516 was 0.8 (95% CI, 0.69–0.9) (data not shown).

F2-3
Fig. 2:
microRNA expression and prediction in HIV-associated neurocognitive disorder by quantitative real-time reverse transcriptase PCR.The expression levels of miR-3665 (a) and miR-4516 (b) were profiled in plasma from HIV-associated neurocognitive disorder (HAND) and non-HIV-associated neurocognitive disorder (nonHAND) patients in the Validation Cohort by quantitative real-time reverse transcriptase PCR assay, as verification of the array results. microRNA expression was normalized to miR-16–5p expression. The receiver-operating characteristic curve analyses for miR-3665 (sensitivity = 90.0%, specificity = 63.6%) (c) and miR-4516 (sensitivity = 72.7%, specificity = 81.8%) (d) predicted the diagnosis of HIV-associated neurocognitive disorder (Mann–Whitney U test, P < 0.05).

Plasma microRNAs’ expression predicts HIV-associated neurocognitive disorder

The above findings indicated that specific miRNAs were associated with the diagnosis of HAND, prompting the comparison of miRNA expression levels with other patient variables associated with HAND. In a univariate analysis, using data from the Validation Cohort, the three miRNAs in the following order – miR-3665, miR-4516 and miR-4707–5p – displayed high odds ratios together with significant differences between the HAND and non-HAND groups (P < 0.05) (Supplementary Fig. 2A, https://links.lww.com/QAD/A935). Indeed, the predictive value of each of these miRNAs was better than other patient variables known to be associated with HAND (e.g. CD4+ T-cell count, viral load, CPE score). As CD4+ T-cell counts in blood are known to be predictive of HAND, linear univariate regression analyses were performed to determine if there was an association between the three identified miRNAs and CD4+ T-cell counts; there was a negative correlation between miR-3665 (r2 = −0.47; P = 0.01 or r2 = −0.52; P = 0.004), miR-4516 (r2 = −0.45, P = 0.013 or r2 = −0.55, P = 0.003) or miR-4707–5p (r2 = −0.54, P = 0.004 or r2 = −0.36, P = 0.044) and current or nadir CD4+ T-cell counts, respectively. In a multivariate analysis, there was a synergistic trend in predictive values for HAND among the combined miRNAs, a combination of miR-3665 and miR-4516 or with miR-3665 alone. Other combinations were less predictive than individual miR-3665 alone (Supplementary Fig. 2B, https://links.lww.com/QAD/A935). To assess the diagnostic value further of each miRNA (miR-3665, miR-4516 or miR-4707–5p) or a combination of these miRNAs for HAND, the data were subjected to analysis by ROC curves. Individually, each miRNA displayed a robust prediction for HAND status in the Validation Cohorts (Fig. 3); for each miRNA in the Validation Cohort, the AUC was 0.85 (CI, 0.78–0.93), 0.78 (CI, 0.69–0.88) and 0.81 (CI, 0.72–0.90), for miR-3665 (Fig. 3A), miR-4516 (Fig. 3B) and miR-4707–5p (Fig. 3C), respectively. A combination of all three miRNAs in the Validation Cohort showed AUC 0.87 (CI, 0.79–0.94) (Fig. 3D). These findings collectively underscored the association between individual miRNAs and the presence of HAND based on array analyses as well as qRT-PCR.

F3-3
Fig. 3:
Comparison of receiver-operating characteristic curves based on microRNA array hybridization.Receiver-operating characteristic curve analyses of individual microRNAs or averages of all three microRNAs in the prediction of HIV-associated neurocognitive disorder (HAND) in the Validation Cohort 2. (a) miR-3665 (sensitivity = 91.7%, specificity = 75.0%), (b) miR-4516 (sensitivity = 58.3%, specificity = 91.7%), (c) miR-4707–5p (sensitivity = 83.3%, specificity = 66.7%) and (d) averaged microRNAs (sensitivity = 83.3%, specificity = 75.0%).

Genes targeted by individual microRNAs

MiRNAs are known to regulate gene function by base pairing of the miRNA with the mRNA at the 3′ untranslated region (UTR) to reduce the expression of the targeted gene. To connect differentially expressed miRNAs in HAND with their putative mRNA targets, different bioinformatics tools were applied to the three miRNAs identified in the present studies. For each miRNA that predicted HAND status, the corresponding mRNA gene targets were considered if the genes were detected by at least two bioinformatics tools (Supplementary Tables 5–7, https://links.lww.com/QAD/A935). Among the filtered mRNAs, multiple genes were identified that are involved in brain development, immunity, inflammation, apoptosis, transcription, metabolism, axonal guidance, transport within the CNS and antiviral activity. Indeed, several of these genes have been implicated in the pathogenesis of HIV-1. To identify the functional classes of genes targeted by these three miRNAs, predicted gene targets (Supplementary Tables 5–7, https://links.lww.com/QAD/A935) were uploaded into the Database Annotation, Visualization and Integrated Discovery bioinformatics software. Gene families containing a minimum of three genes and a fold enrichment of at least 1.0 were considered for analysis. Among the gene families targeted by the three upregulated miRNAs in HAND, genes involved in CNS development ranked highest with 20% of genes involved (Fig. 4a). Other gene families targeted by the three miRNAs include cell signaling, cytokine signaling, regulation of apoptosis, epidermal growth factor (EGF) signaling and inflammation. CNS genes were further analyzed to identify the subclass within the CNS. This analysis identified several subclasses of gene families within the CNS (e.g. related to axon, synapse, neurogenesis, neuronal projection, behavior) that were targeted by these three miRNAs (Fig. 4b). Network analyses using the Validation Cohort data set highlighted genes involved in miRNA processing and insulin-like growth factor signaling (Supplementary Fig. 3A, https://links.lww.com/QAD/A935). Network analyses also showed that the three identified miRNAs had predicted targets associated with neural cell functions (Supplementary Fig. 3B, https://links.lww.com/QAD/A935). In summary, these bioinformatics analyses based on miRNA expression pointed to multiple genes connected to a range of cellular functions and viability that were relevant to HAND pathogenesis.

F4-3
Fig. 4:
Functional classes of genes targeted by HIV-associated neurocognitive disorder associated microRNAs.(a) Percentages of predicted classes of genes targeted by the three upregulated microRNAs (miR-3665, miR-4516 and miR-4707–5p) in patients with HIV-associated neurocognitive disorder based on bioinformatics analyses. (b) Central nervous system related gene families targeted by the three microRNAs.

Discussion

The current report is the first study to the best of our knowledge that demonstrates a distinctive miRNA profile in plasma that distinguishes HAND from non-HAND patients in different cohorts and using different methodologies. Using Discovery and Validation Cohorts, this study identified three differentially expressed miRNAs in both cohorts, although other miRNAs were identified with altered expression within individual cohorts. These findings were evident using different but complementary laboratory, bioinformatics and biostatistics methods. The three identified miRNAs in these studies targeted multiple genes implicated in neurological disease and HIV infection, which again was apparent using different bioinformatics tools. Furthermore, the same three miRNAs displayed robust predictive capacity for HAND individually as well as in concert, with miR-3665 showing the greatest association with HAND, followed by miR-4516 and lastly miR-4707–5p. Recent studies have shown that using a panel of miRNAs to predict or diagnose a disease improves the sensitivity compared with an individual miRNA [37], thus supporting the results presented herein. The fact that two of these miRNAs were validated using another technique, qRT-PCR, with resulting AUCs of 0.77 and 0.79 for miR-3665 and miR-4516, respectively, further support the notion that miRNAs might serve as biomarkers of HAND. The clinical application of these miRNAs to diagnose HAND using an easily accessible biospecimen (plasma) and a qRT-PCR assay platform further simplifies its use as a biomarker. The present results point to miRNAs as informative biomarkers for HAND as well as potential indicators of molecular mechanisms that contribute to HAND development.

Altered miRNA expression has been described previously in the brains of people with HAND and correlated with targeted transcript expression; in particular, three miRNAs (miR-137, 153 and 218) were found to be associated with genes involved in neurodegeneration [45]. Interestingly, in the study by Zhou et al.[45], the authors showed differential expression of miRNAs that targeted genes implicated in axonal guidance and insulin signalling. A recent study by Kadri et al.[46] using a qRT-PCR-based approach in a small single cohort of HIV/AIDS patients assessed a subset of human miRNAs (752 miRNAs) and identified several as being associated with neurocognitive impairment of uncertain cause. None of the miRNAs identified in this latter study were found to have altered expression in our study. In the present study, bioinformatics analyses revealed that altered expression of miRNAs was associated with IGF-1, a gene involved in neuronal viability and function and on network analyses with downregulation of mRNAs (SEMA6D, SEMA6B and ELV) on other bioinformatics analyses, which are involved in axonal guidance. MiRNA profiling of cerebrospinal fluid from patients with and without HIV infection showed that multiple miRNAs were differentially expressed with a strong association between induction of miRNA expression and the presence of HIV encephalitis [47]. Our group previously reported that miRNAs targeting cell-death-related genes in HIV-infected brains were downregulated with accompanying increased expression of caspase-6 in brains from patients with HIV encephalitis [30]. In brains from animals with simian immuno-deficiency virus (SIV), encephalitis-explicit miRNA profiles have been described that implicate different pathogenic mechanisms including neuronal dysfunction [48] and inflammatory gene expression [49]. Moreover, plasma miRNA measurement in animals with SIV encephalitis disclosed a distinct profile associated with brain disease [36], reflecting the observations described in the current study.

Several issues require clarification in the present observations including the divergent clinical and demographic features of the two cohorts. The cohorts differed in the time periods of the populations studied. These differences were evident in the overall duration of infection with shorter times in the Discovery Cohort; differences in immunosuppression levels between cohorts and also in the demographic aspects as the Validation Cohort showed increased numbers of patients reporting lower levels of education in the HAND group; this might reflect differing ethnicity composition in each cohort with more patients from sub-Saharan Africa in the Validation Cohort (as new immigrants to Canada). The plasma profiling results from the two cohorts also differed in that the Discovery Cohort did not show reduced differentially expressed miRNAs at the cut-off of 2.0 or less; however, if the cut-off was lowered to 1.5 or less, multiple downregulated miRNAs were detected in the Discovery Cohort's HAND plasma specimens, underscoring the restricted dynamic ranges in differential expression of miRNAs. Although the prediction of HAND by miR-3665, miR-4516 and miR-4707–5p in multivariate analyses or correcting for multiple comparison (Bonferroni) was NS in the Discovery Cohort, the univariate logistic regression analyses established each of these miRNAs as predictors of HAND. The small sample size might explain the lack of significance in multivariate analyses. It is important to note also that the miRNAs that were significantly upregulated in the Discovery Cohort after Bonferroni correction did not show significance by qRT-PCR in the Validation Cohort. The limited number of validated miRNAs measured in the array used in the present study was obviated to some extent by qRT-PCR verification, and the actual target genes identified in bioinformatics analyses remain to be validated fully. Multiple bioinformatics tools were applied to predict genes potentially modulated by miRNAs, from which the common targeted genes and gene families were identified. The fact that 20% of the genes targeted by the three miRNAs participate in CNS function suggests that gene expression could be selectively reduced in the brain, contributing to the development of HAND. Specifically, our network analyses (Supplementary Fig. 2B, https://links.lww.com/QAD/A935) revealed that both miR-3665 and miR-4516 targeted MOBP (myelin-associated oligodendrocyte basic protein). In a study by Borjabad et al.[50], the authors showed that MOBP is downregulated in HAND patients’ brains, thus supporting a regulatory role for these miRNAs in HAND. As the cellular sources of miRNAs observed in these data as well as their cellular targets remain to be discovered, future studies will need to confirm the source of individual miRNAs and their relationship with the CNS function. Other studies have shown that there is an altered miRNA profile in the peripheral blood of patients with CNS diseases such as schizophrenia and multiple sclerosis (MS) [51,52]. It has recently been reported that extracellular microvesicles (exosomes) can transport brain-specific miRNAs into the peripheral blood as part of their role in cell–cell communication [53]. Although the cellular sources are still to be confirmed, it is conceivable that the miRNAs identified in the present study might be secreted from brain cells and transported to the vascular circulation. As individual miRNAs can be associated with more than one disease condition, future studies will be required to confirm specificity of these miRNAs in HAND compared with other inflammatory neurodegenerative diseases such as MS and Alzheimer's disease. As increased neuroinflammation in HIV-1 infection is associated with the development of HAND [14] and ART improves outcomes in HAND [54], it will be important to evaluate the diagnostic efficacy of the present miRNA profile in HIV/AIDS patients with and without HAND in the settings of increased neuroinflammation during HIV-1 infection and in response to ART introduction. In a recent study by Gisslén et al.[55], the authors showed that neurofilament light protein (NFL), a marker of neuronal injury, detected in plasma and CSF correlated with each other and that higher levels are associated with HAD. Moreover, larger studies are required to study the association between plasma NFL levels as well as comorbidities such as diabetes, cardiovascular diseases and hypertension, together with the miRNA profile dysregulation in HAND.

miRNAs have proven to be valuable as both indicators of underlying disease mechanisms and biomarkers for disease diagnosis and prognosis as well as responses to treatment [28,29]. Their relative stability in plasma and associated properties permitting microvesicle-mediated transport between cells highlight the potential application of miRNAs as disease biomarkers and perhaps therapeutic agents (reviewed in [34,35,56]). The present study pointed to three miRNAs as potential biomarkers of HAND, but future studies will require a larger sample size to determine the associated subtype of HAND (ANI, MND or HAD) and perhaps its prognosis in terms of likelihood of progressing to more severe types. It is important to note that the present cohorts comprised only the MND and HAD subtypes of HAND, whereas none of the patients had ANI. Studies in larger and different HAND cohorts are required to substantiate these findings and perhaps provide a broader range of miRNAs pointing to the diagnosis of HAND and its multifactorial causative mechanisms.

Acknowledgements

The authors thank the staff (Noshin Koenig and Brenda Beckthold) at the Southern Alberta Clinic (SAC) for their support of the current studies and Drs Kenneth Witwer and David Broadhurst for helpful discussions. The authors appreciate the willingness of patients at SAC to participate in these studies.

Author's contribution: E.L.A. designed the study, performed experiments, analyzed data, performed bioinformatics analyses and wrote the first draft of manuscript. S.M.A. performed statistical analyses and contributed to manuscript preparation. W.G.B. analyzed data, performed bioinformatics analyses and contributed to writing the manuscript. E.F. designed neuropsychological testing battery and contributed to manuscript preparation. J.M.G. recruited study participants and contributed to manuscript preparation. C.P. conceived and designed the study, designed the neuropsychological testing battery, supervised experiments and contributed to manuscript preparation.

Funding: These studies were supported by a Canadian Institutes of Health Research (CIHR) Emerging Team grant (C.P., E.F. and J.M.G.). E.L.A. is supported by fellowships from CIHR and Alberta Innovates-Health Solutions. C.P. holds a Canada Research Chair in Neurological Infection and Immunity (Tier 1).

Conflicts of interest

There are no conflicts of interest.

References

1. Wright PW, Vaida FF, Fernandez RJ, Rutlin J, Price RW, Lee E, et al. Cerebral white matter integrity during primary HIV infection. AIDS 2015; 29:433–442.
2. Sturdevant CB, Joseph SB, Schnell G, Price RW, Swanstrom R, Spudich S. Compartmentalized replication of R5 T cell-tropic HIV-1 in the central nervous system early in the course of infection. PLoS Pathog 2015; 11:e1004720.
3. Desplats P, Dumaop W, Smith D, Adame A, Everall I, Letendre S, et al. Molecular and pathologic insights from latent HIV-1 infection in the human brain. Neurology 2013; 80:1415–1423.
4. Heaton RK, Franklin DR Jr, Deutsch R, Letendre S, Ellis RJ, Casaletto K, et al. Neurocognitive change in the era of HIV combination antiretroviral therapy: the longitudinal CHARTER study. Clin Infect Dis 2015; 60:473–480.
5. Heaton RK, Clifford DB, Franklin DR Jr, Woods SP, Ake C, Vaida F, et al. HIV-associated neurocognitive disorders persist in the era of potent antiretroviral therapy: CHARTER Study. Neurology 2010; 75:2087–2096.
6. Yusuf AJ, Hassan A, Mamman AI, Muktar HM, Suleiman AM, Baiyewu O. Prevalence of HIV-associated neurocognitive disorder (HAND) among patients attending a tertiary health facility in Northern Nigeria. J Int Assoc Providers AIDS Care 2014.
7. Milanini B, Wendelken LA, Esmaeili-Firidouni P, Chartier M, Crouch PC, Valcour V. The montreal cognitive assessment to screen for cognitive impairment in HIV patients older than 60 years. J Acquir Immune Defic Syndr 2014; 67:67–70.
8. Lekoubou A, Echouffo-Tcheugui JB, Kengne AP. Epidemiology of neurodegenerative diseases in sub-Saharan Africa: a systematic review. BMC Public Health 2014; 14:653.
9. Gill AJ, Kovacsics CE, Cross SA, Vance PJ, Kolson LL, Jordan-Sciutto KL, et al. Heme oxygenase-1 deficiency accompanies neuropathogenesis of HIV-associated neurocognitive disorders. J Clin Invest 2014; 124:4459–4472.
10. Vartak-Sharma N, Gelman BB, Joshi C, Borgamann K, Ghorpade A. Astrocyte elevated gene-1 is a novel modulator of HIV-1-associated neuroinflammation via regulation of nuclear factor-kappaB signaling and excitatory amino acid transporter-2 repression. J Biol Chem 2014; 289:19599–19612.
11. Fields J, Dumaop W, Elueteri S, Campos S, Serger E, Trejo M, et al. HIV-1 Tat alters neuronal autophagy by modulating autophagosome fusion to the lysosome: implications for HIV-associated neurocognitive disorders. J Neurosci 2015; 35:1921–1938.
12. Si Q, Kim MO, Zhao ML, Landau NR, Goldstein H, Lee S. Vpr- and Nef-dependent induction of RANTES/CCL5 in microglial cells. Virology 2002; 301:342–353.
13. Spudich S, Gisslen M, Hagberg L, Lee E, Liegler T, Brew B, et al. Central nervous system immune activation characterizes primary human immunodeficiency virus 1 infection even in participants with minimal cerebrospinal fluid viral burden. J Infect Dis 2011; 204:753–760.
14. Walsh JG, Reinke SN, Mamik MK, McKenzie BA, Maingat F, Branton WG, et al. Rapid inflammasome activation in microglia contributes to brain disease in HIV/AIDS. Retrovirology 2014; 11:35.
15. Nightingale S, Winston A, Letendre S, Michael BD, McArthur JC, Khoo S, et al. Controversies in HIV-associated neurocognitive disorders. Lancet Neurol 2014; 13:1139–1151.
16. Kusao I, Shiramizu B, Liang CY, Grove J, Agsalda M, Troelstrup D, et al. Cognitive performance related to HIV-1-infected monocytes. J Neuropsychiatry Clin Neurosci 2012; 24:71–80.
17. Rozzi SJ, Borelli G, Ryan K, Steiner JP, Reglodi D, Mocchetti I, et al. PACAP27 is protective against tat-induced neurotoxicity. J Mol Neurosci 2014; 54:485–493.
18. Fields JA, Serger E, Campos S, Divakaruni AS, Kim C, Smith K, et al. HIV alters neuronal mitochondrial fission/fusion in the brain during HIV-associated neurocognitive disorders. Neurobiol Dis 2016; 86:154–169.
19. Na H, Acharjee S, Jones G, Vivithanaporn P, Noorbakhsh F, McFarlane N, et al. Interactions between human immunodeficiency virus (HIV)-1 Vpr expression and innate immunity influence neurovirulence. Retrovirology 2011; 8:44.
20. Jones GJ, Barsby NL, Cohen EA, Holden J, Harris K, Dickie P, et al. HIV-1 Vpr causes neuronal apoptosis and in vivo neurodegeneration. J Neurosci 2007; 27:3703–3711.
21. Cui L, Locatelli L, Xie MY, Sommadossi JP. Effect of nucleoside analogs on neurite regeneration and mitochondrial DNA synthesis in PC-12 cells. J Pharmacol Exp Ther 1997; 280:1228–1234.
22. Fetterman JL, Holbrook M, Westbrook DG, Brown JA, Feeley KP, Breton-Romero R, et al. Mitochondrial DNA damage and vascular function in patients with diabetes mellitus and atherosclerotic cardiovascular disease. Cardiovasc Diabetol 2016; 15:53.
23. Lauritzen KH, Kleppa L, Aronsen JM, Eide L, Carlsen H, Haugen OP, et al. Impaired dynamics and function of mitochondria caused by mtDNA toxicity leads to heart failure. Am J Physiol Heart Circ Physiol 2015; 309:H434–H449.
24. Guo H, Ingolia NT, Weissman JS, Bartel DP. Mammalian microRNAs predominantly act to decrease target mRNA levels. Nature 2010; 466:835–840.
25. Porichis F, Hart MG, Griesbeck M, Everett HL, Hassan M, Baxter AE, et al. High-throughput detection of miRNAs and gene-specific mRNA at the single-cell level by flow cytometry. Nat Commun 2014; 5:5641.
26. Bernard MA, Zhao H, Yue SC, Anandaiah A, Koziel H, Tachado SD. Novel HIV-1 miRNAs stimulate TNFalpha release in human macrophages via TLR8 signaling pathway. PLoS One 2014; 9:e106006.
27. Noorbakhsh F, Ellestad KK, Maingat F, Warren KG, Han MH, Steinman L, et al. Impaired neurosteroid synthesis in multiple sclerosis. Brain 2011; 134:2703–2721.
28. Reynoso R, Laufer N, Hackl M, Skalicky S, Monteforte R, Turk G, et al. microRNAs differentially present in the plasma of HIV elite controllers reduce HIV infection in vitro. Sci Rep 2014; 4:5915.
29. Dubin PH, Yuan H, Devine RK, Hynan LS, Jain MK, Lee WM. Micro-RNA-122 levels in acute liver failure and chronic hepatitis C. J Med Virol 2014; 86:1507–1514.
30. Noorbakhsh F, Ramachandran R, Barsby N, Ellestad KK, LeBlanc A, Dickie P, et al. microRNA profiling reveals new aspects of HIV neurodegeneration: caspase-6 regulates astrocyte survival. FASEB J 2010; 24:1799–1812.
31. Witwer KW, Watson AK, Blankson JN, Clements JE. Relationships of PBMC microRNA expression, plasma viral load, and CD4+ T-cell count in HIV-1-infected elite suppressors and viremic patients. Retrovirology 2012; 9:5.
32. Weber JA, Baxter DH, Zhang S, Huang DY, Huang KH, Lee MJ, et al. The microRNA spectrum in 12 body fluids. Clin Chem 2010; 56:1733–1741.
33. Mitchell PS, Parkin RK, Kroh EM, Fritz BR, Wyman SK, Pogosova-Agadjanyan EL, et al. Circulating microRNAs as stable blood-based markers for cancer detection. Proc Natl Acad Sci USA 2008; 105:10513–10518.
34. Kosaka N, Iguchi H, Ochiya T. Circulating microRNA in body fluid: a new potential biomarker for cancer diagnosis and prognosis. Cancer Sci 2010; 101:2087–2092.
35. Weilner S, Schraml E, Redl H, Grillari-Voglauer R, Grillari J. Secretion of microvesicular miRNAs in cellular and organismal aging. Exp Gerontol 2013; 48:626–633.
36. Witwer KW, Sarbanes SL, Liu J, Clements JE. A plasma microRNA signature of acute lentiviral infection: biomarkers of central nervous system disease. AIDS 2011; 25:2057–2067.
37. Xiao B, Wang Y, Li W, Baker M, Guo J, Corbet K, et al. Plasma microRNA signature as a noninvasive biomarker for acute graft-versus-host disease. Blood 2013; 122:3365–3375.
38. Ng EK, Chong WW, Jin H, Lam EK, Shin VY, Yu J, et al. Differential expression of microRNAs in plasma of patients with colorectal cancer: a potential marker for colorectal cancer screening. Gut 2009; 58:1375–1381.
39. Zampetaki A, Kiechl S, Drozdov I, Willeit P, Mayr U, Prokopi M, et al. Plasma microRNA profiling reveals loss of endothelial miR-126 and other microRNAs in type 2 diabetes. Circ Res 2010; 107:810–817.
40. McCombe JA, Vivithanaporn P, Gill MJ, Power C. Predictors of symptomatic HIV-associated neurocognitive disorders in universal healthcare. HIV Med 2013; 14:99–107.
41. Maingat F, Halloran B, Acharjee S, van Marle G, Church D, Gill MJ, et al. Inflammation and epithelial cell injury in AIDS enteropathy: involvement of endoplasmic reticulum stress. FASEB J 2011; 25:2211–2220.
42. Vivithanaporn P, Heo G, Gamble J, Krentz HB, Hoke A, Gill MJ, et al. Neurologic disease burden in treated HIV/AIDS predicts survival: a population-based study. Neurology 2010; 75:1150–1158.
43. van Marle G, Rourke SB, Zhang K, Silva C, Ethier J, Gill MJ, et al. HIV dementia patients exhibit reduced viral neutralization and increased envelope sequence diversity in blood and brain. AIDS 2002; 16:1905–1914.
44. Antinori A, Arendt G, Becker JT, Brew BJ, Byrd DA, Cherner M, et al. Updated research nosology for HIV-associated neurocognitive disorders. Neurology 2007; 69:1789–1799.
45. Zhou L, Pupo GM, Gupta P, Liu B, Tran SL, Rahme R, et al. A parallel genome-wide mRNA and microRNA profiling of the frontal cortex of HIV patients with and without HIV-associated dementia shows the role of axon guidance and downstream pathways in HIV-mediated neurodegeneration. BMC Genomics 2012; 13:677.
46. Kadri F, LaPlante A, De Luca M, Doyle L, Velasco-Gonzalez C, Patterson JR, et al. Defining plasma microRNAs associated with cognitive impairment in HIV-infected patients. J Cell Physiol 2015; doi:10.1002/jcp.25131.
47. Pacifici M, Delbue S, Ferrante P, Jeansonne D, Kadri F, Nelson S, et al. Cerebrospinal fluid miRNA profile in HIV-encephalitis. J Cell Physiol 2013; 228:1070–1075.
48. Yelamanchili SV, Chaudhuri AD, Chen LN, Xiong H, Fox HS. microRNA-21 dysregulates the expression of MEF2C in neurons in monkey and human SIV/HIV neurological disease. Cell Death Dis 2010; 1:e77.
49. Liu J, Sisk JM, Gama L, Clements JE, Witwer KW. Tristetraprolin expression and microRNA-mediated regulation during simian immunodeficiency virus infection of the central nervous system. Mol Brain 2013; 6:40.
50. Borjabad A, Morgello S, Chao W, Kim SY, Brooks AI, Murray J, et al. Significant effects of antiretroviral therapy on global gene expression in brain tissues of patients with HIV-1-associated neurocognitive disorders. PLoS Pathog 2011; 7:e1002213.
51. Fan HM, Sun XY, Niu W, Zhao L, Zhang QL, Li WS, et al. Altered microRNA expression in peripheral blood mononuclear cells from young patients with schizophrenia. J Mol Neurosci 2015; 56:562–567.
52. Kacperska MJ, Jastrzebski K, Tomasik B, Walenczak J, Konarska-Krol M, Glabinski A. Selected extracellular microRNA as potential biomarkers of multiple sclerosis activity-preliminary study. J Mol Neurosci 2014; 56:154–163.
53. Haqqani AS, Delaney CE, Tremblay TL, Sodja C, Sandhu JK, Stanimirovic DB. Method for isolation and molecular characterization of extracellular microvesicles released from brain endothelial cells. Fluids Barriers CNS 2013; 10:4.
54. Gates TM, Cysique LA, Siefried KJ, Chaganti J, Moffat KJ, Brew BJ. Maraviroc-intensified combined antiretroviral therapy improves cognition in virally suppressed HIV-associated neurocognitive disorder. AIDS 2016; 30:591–600.
55. Gisslén M, Price RW, Andreasson U, Norgren N, Nilsson S, Hagberg L, et al. Plasma concentration of the neurofilament light protein (NFL) is a biomarker of CNS injury in HIV infection: a cross-sectional study. EBioMedicine 2016; 3:135–140.
56. Pegtel DM, Peferoen L, Amor S. Extracellular vesicles as modulators of cell-to-cell communication in the healthy and diseased brain. Philos Trans R Soc Lond B Biol Sci 2014; 369:201305516.
Keywords:

biomarker; HIV; HIV-associated neurocognitive disorder; microRNA; nervous system

Supplemental Digital Content

Copyright © 2016 Wolters Kluwer Health, Inc.