New Potential MicroRNA Biomarkers in Human Immunodeficiency Virus Elite Controllers, Human Immunodeficiency Virus Infections, and Coinfections with Hepatitis B Virus or Hepatitis C Virus

Hepatitis C virus (HCV), hepatitis B virus (HBV), and human immunodeficiency virus (HIV) have been considered the commonest blood-borne viruses. HCV and/or HBV infections, due to their similar transmission routes, constitute the most frequent comorbidities in HIV-infected subjects. It is estimated that from about 37 million HIV-infected patients in the world, 2–15% and 5–20% are coinfected with HCV and HBV, respectively [1]. In addition, the simultaneous presence of HIV infection with viral hepatitis can affect HIV disease management and, in return, HIV infection may accelerate the progression of liver disease [2, 3]. Another subset of the HIV-infected persons is elite controllers (ECs) which appear to occur in 0.15–1.5% of HIV-infected individuals [4‒7]. However, ECs have the ability to control the HIV replication without antiviral therapy and the viral load in these individuals often is at undetectable levels (<50 copies/mL) [8, 9]. Nevertheless, the mechanism(s) of control for HIV infections in ECs is not yet fully understood.

Numerous cellular factors are involved in the development of viral infections that can be used as both therapeutic targets and diagnostic biomarkers [9, 10]. One such important host factor is microRNA (miRNA) that belongs to a small (∼22 nucleotides long) noncoding RNA family [11]. MiRNAs play critical roles in many biological processes like proliferation or rapid growth, metastasis, apoptosis, tumorigenesis, immune responses, and inflammation by regulating the gene expression at the level of mRNA and proteins [11‒13]. Reportedly, some cellular miRNAs have been proposed to act as critical players in various viral diseases [9, 12‒15]. For example, the liver-specific miR-122 upregulates the HCV life cycle by binding to a 5′ untranslated region (UTR) of HVC RNA, while miR-122 act as a suppressor of HBV replication [16]. Put differently, viral proteins alter the expression of cellular miRNAs directly or indirectly to provide good conditions for the development of the viral disease [12]. Moreover, cellular restriction factors such as p21 and TASK1 are one of the major sets of genes suppressing HIV infections. In vitro studies have proven that HIV can lead to the downregulation of p21 and TASK1 expression by exploiting the cellular miRNAs (e.g., let-7c, miR-124a, and/or miR-34a) and thereby restricts the host’s innate inhibition mechanism [9, 17]. The CCL2 (monocyte chemoattractant protein-1, MCP-1) as the ligand for CCR1, CCR2, and CCR5 chemokines, has the potential to restrict HIV entry mediated by CD4/CCR5 [18]. It recently has been shown that miR-146a, by targeting CCL2, can maintain chronic inflammation associated with HIV in microglial cells [19]. Nevertheless, few studies have been done to recognize miRNAs and their target genes as therapeutic and diagnostic tools in different subsets in a HIV-infected population.

Some viral families encode miRNAs to contribute to viral infection and decrease the antiviral immune responses by regulating the transcription of cellular and viral genes [20]. Accumulating evidence has found that the different expression patterns of the viral miRNAs (v-miRNAs) during various phases of viral infection, suggesting the potential use of v-miRNAs as biomarkers [21, 22]. Reportedly, miR-N367, HIV-miR-H1, miR-H3, vmiR99, vmiR88, vmiR-TAR, miR-TAR-3p, and vmiRNA#1-5 are derived from both noncoding and coding regions of the HIV RNA genome [9]. It has been found that the HIV nef-derived miRNA, called miR-N367, decreases transcription of the virus through inhibition of nef expression and long terminal repeat (LTR) transcription [23]. Additionally, miR-N367 may be contributing to the maintenance of the virus at a latent stage by targeting poly(A)-binding protein cytoplasmic 4 (PABPC4) [24]. However, the role of this v-miRNA as a diagnostic biomarker and therapeutic target for ECs and other HIV-infected subsets has not been well studied.

Given that the virus interaction with the host immune system, during coinfections, is different from infection alone, it is possible that the expression profile of cellular and viral factors is different between coinfections of HIV-HCV and HIV-HBV with mono-HIV infection [2, 3, 25]. Thus, the investigation of specific cellular and viral factors may be applied to recognize novel biomarkers to differentiate between HIV-HBV and/or HIV-HCV coinfection from HIV monoinfection. It would also be beneficial to differentiate ECs from VP and might contribute to finding the design of novel anti-HIV drugs. In previous studies, the expression pattern of cellular miRNAs has only been studied among a limited group of HIV-infected subjects, but none assessed whether the expression profile of one or more cellular/viral miRNAs could be used as a potential biomarker to differentiate between various groups of the HIV-infected population. The present research attempts to evaluate the expression profile of the chosen cellular/viral miRNAs (miR-17-5p, -29a, -106a-5p, -125a, -140-3p, -191-5p, -200a-3p, -339-3p, -590-3p, and miR-N367), and cellular factors associated with HIV replication (p21, SDF-1, XCL1, CCL-2, and IL-2) in PBMCs of the HIV-HCV coinfection group, HIV/HBV coinfection group, ECs, HIV-infected patients receiving anti-retroviral treatment (HIV ART), as well as ART-naive HIV-infected cases (HIV naïve).

One hundred and twenty participants participated in the research, who were grouped into 6 categories: healthy control (N = 20), ECs (N = 20), ART-naïve HIV (N = 20), HIV ART (N = 20), coinfected cases (HIV-HCV (N = 20), and HIV-HBV (N = 20)). ECs have viral load <50 copy/mL (below the limit of detection) without any treatments and their CD4⁺ cell count were >500 cells/mm³ for determined follow-up (>6 years). We received all samples (except controls) from the HIV infected cases who went to the Iranian Research Center for HIV/AIDS (Tehran, Iran), the centers related to Tabriz University of Medical Sciences, as well as Kermanshah University of Medical Sciences. Furthermore, we excluded all patients with any concurrent diseases like cytomegalovirus and mycobacterium tuberculosis from the research. Except ART HIV cases, all patients did not receive any antiviral drugs and medications. The expression levels of miRNA-17, - 29, -106, -125, -140, -191, -200, -339, and -590 were compared between 5 patient groups and healthy subjects. Also, the expression level of selected miRNAs was compared between HIV Naïve, HIV ART, HIV/HCV, and HIV/HBV with ECs. Moreover, we obtained the peripheral blood mono-nuclear cells (PBMCs) from each participant and kept at a temperature of −80°C. Table 1 reports the clinical and demographic features of all the subjects. Then, we matched the control and patient groups with regard to gender, size, and age and received the ethical approval from Tabriz University of Medical Sciences (IR.TBZMED.REC.1400.149). Also, the consent forms were obtained from the participants prior to starting our research.

According to the present research, the monitoring of HIV, HCV, and HBV was conducted by routine laboratory tests. Two distinct fourth-generation commercial ELISA kits including INNOTEST-HIV Ab IV (provided from Innogenetics in Ghent, Belgium) and DIA.PRO (Milano in Italy) were applied for detecting and confirming the anti-HIV antibody (Ab). HBV infection test (hepatitis B virus antigen (HBsAg)) was carried out by anti-HBV antibody AlphaLISA Detection Kit (PerkinElmer, USA) and HCV was detected by real-time PCR method. HIV-HBV coinfection was defined as detecting of HIV-positive cases showing HBsAg positive and HIV-HCV coinfection were as HIV-positive subjects with anti‐HCV Ab and HCV RNA (with the GeneProof HCV PCR Kit [GeneProof, Czech Republic]). Moreover, EC subjects were defined as having a plasma viral load less than 50 copies/mL, as well as a CD4 count greater than 350/mL for determined follow-up time (6 years) [26].

RNA extraction was conducted by QIAamp DSP Virus Kit Procedure (Qiagen GmbH, Hilden, Germany) for HCV, HIV-HCV, as well as HIV (coinfected)-positive samples according to the manufacturer. DNA extraction kit (Qiagen GmbH, Hilden, Germany) was applied for HBV DNA extraction. For quality and quantity of the extracted RNA, ND-1000 Spectrophotometer (NanoDrop, Fisher Thermo, DE, USA) was used. Furthermore, the load of HIV was determined using the Artus HIV RG RT-PCR Kit. Also, HBV and HCV loads were measured by the Artus HBV PCR Kit and Artus HCV RG RT-PCR kit (Qiagen GmbH, Qiagen Strasse, Germany), respectively. The CD4+ T-cell count was carried out using a PIMA CD4 analyzer (Alere, Germany) for each participant.

Standard Ficoll-Paque gradient (Lympholyte-H, Cedarlane, Hornby: Canada) was used for isolating the PBMCs of the blood samples and then we obtained a 6-mL peripheral blood specimen from each participant in blood collection tubes (EDTA vacutainer tubes). Then, we performed centrifugation for separating the plasma (from whole blood samples) and PBMC isolation using Ficoll density gradient centrifugation. Additionally, PBS (phosphate-buffered saline [PBS] at a pH of 7.3 ± 0.1) was used for washing the isolated PBMCs three times. Finally, all PBMC samples were stored at −20°C (in the RNAlater storage solution [Ambion, Austin, TX, USA]).

We selected these miRNAs (miR-17-5p, -29a, -106a-5p, -125a, -140-3p, -191-5p, -200a-3p, -339-3p, and -590-3p) based on some strategies: (1) the role of these miRNAs in the essential cellular physiology process, (2) The role of these miRNAs in HIV life cycle and infection, (3) previous data in our research and review articles in miRNAs field, and (4) previous miRNA arrays. In addition, Trizol total RNA isolation reagent (provided from Roche Diagnostics in Belgium) was applied to total RNA extraction from PBMCs. CDNA synthesis was conducted using 5 μg of total RNA and with the miScript II RT Kit (Qiagen). MiRNA expression profiles were performed via miScript SYBR Green PCR kit (provided from Qiagen GmbH, Qiagen Strasse in Germany) based on the company’s directions. Two SNORDs (68 and 61) were used as the normalization control to relatively quantify the values. Also, we applied 2^−ΔΔCT method (Livak & Schmittgen, 2001) for analyzing the data of gene expression.

The levels of p21, SDF-1, XCL1, CCL2, and IL-2 were measured using Human P21 ELISA Kit Elisa Kit (Proteintech, Wuhan, China), Human XCL1/Lymphotactin ELISA Kit PicoKine™ (Boster Biological Technology, CA, USA), Human SDF-1 alpha Quantikine ELISA Kit (R&D Systems, Minneapolis, MN), Human high-sensitivity C-C motif chemokine 2 (CCL2) ELISA kit (Cusabio Biotech Co., Ltd., Wuhan, China), and Human IL-2 ELISA Kit (Abcam, Cambridge, MA, USA) based on the manufacturer’s directions.

GraphPad Prism 6.0 and SPSS 16.0 were employed for analyzing the data. Upon the pre-processing of the qPCR data, we analyzed differences in expression level of miRNA between the HIV-infected groups with the healthy people and between the ECs with the other HIV-infected groups using the Kruskal-Wallis H or one-way ANOVA test. Then, we implemented the receiver-operator characteristic (ROC) curve analysis for all miRNAs for identifying miRNA specificity and sensitivity as the biomarkers of HIV-infected groups. Likewise, the correlation of the expression levels of selected miRNA, viral loads, CD⁴⁺, as well as the selected cellular factors were analyzed by Spearman’s correlation coefficient. Finally, Benjamini and Hochberg’s method was followed for controlling the false discovery rate, and thus, the adjusted p values were computed.

Table 1 reports the basic demographic information and CD⁴⁺ count as well as the viral load of the cases. The samples were homogeneous for p value = 0.091 and age (p value = 0.18). Moreover, mean of the CD4+ level in all HIV-infected groups declined in comparison to the healthy group (p value <0.001).

According to the results, expression pattern of the cellular miRNAs in HIV-infected groups was remarkably different from the healthy group (Fig. 1). The mean fold change of miR-17-5p, -106a-5p, -125a, -140-3p, and -339-3p were statistically increased in comparison to the HIV/HBV- as well as HIV/HCV-infected people than in the controls. And, based on correlation analysis results, miR-106a-5p expression (r = 0.58, p < 0.001) and miR-125a (r = 0.64, p < 0.001) had a positive correlation with HCV viral load (Table 2).

Besides, the expression patterns of the miR-17-5p, -29a, -125a, -200a-3p, and -339-3p were statistically similar between the ECs and healthy groups; nevertheless, mean expression level of the miR-106a-5p, -191-5p, and -590-3p were significantly lower and the miR-140-3p levels enhanced in the ECs in comparison to the control group. Furthermore, a reverse association was found between the miR-29a expression level (r = −0.54, p value <0.001) with HIV viral load (Table 2).

Results of the HIV ART group in comparison with the healthy people showed that expression of miR-125a, -191-5p, 200a-3p, -339-3p was significantly increased and the miR-29a expression level was decreased (Fig. 1). Furthermore, in the ART-naïve HIV-infected group, expression levels of miR-17-5p, -200a-3p were up-regulated, while the levels of miR-29a, 140-3p were downregulated in comparison to the healthy group.

Based on the outputs from Spearman correlation test, miR-106a-5p (r = 0.6, p < 0.001) and miR-191-5p (r = 0.55, p < 0.001) positively correlated with the level of HIV-nef gene, as well as a remarkable negative correlation between the miR-29a expression level (r = −0.41, p value <0.01), and nef was observed (Table 2).

For further information of if the expression patterns of the chosen cellular and viral miRNAs differ between the EC group and other HIV-infected groups or not, the EC group was used as the control group. Outputs demonstrating expression profile of the miR-17-5p, -29a, -106a-5p, -125a, -191-5p, 339-3p, and -590-3p were significantly different between coinfection groups with the EC group. Additionally, mean fold changes in the miR-17-5p (5.38fold, p value <0.0001) as well as miR-191-5p (4.07fold, p value <0.0001) enhanced in HIV/HCV and HIV/HBV coinfection groups than that in the EC group (Fig. 2). In addition, the mean fold change of cellular miR-106a-5p, -125a, -191-5p, -200a-3p, and 339-3p in the HIV ART group was significantly upregulated in comparison with the controls. As given by Figure 2, the expression level of some miRNAs was remarkably different between the ART-naïve samples and the EC samples and we found the highest increase and decrease in expression for miR-17-5p (4.61fold, p value <0.0001) and miR-140-3p (−3.49fold, p value <0.0001).

It should also be pointed out that the ART-naïve HIV group was used as a control group for examining the differences in the profile of nef-derived miR-N367 expression between the HIV-infected groups. According to our results, no statistical differences in the mean level of miR-N367 were found among HIV/HBV and HIV/HCV coinfected groups with the ART-naïve HIV group. Whereas, the mean fold change of miR-N367 enhanced in the EC and HIV ART groups in comparison with the ART-naïve HIV-infected group (Fig. 2). In addition, based on correlation analysis, we observed a negative association in the expression level of miR-N367 (r = −0.47, p < 0.01) with HCV RNA viral load as well as between the level of miR-N367 (r = −0.42, p value <0.01) with HIV viral load (Table 2). Finally, we found a positive association of miR-N367 expression level with the HIV-nef gene level (r = 0.39, p value <0.05).

Based on ROC analysis using 2^−ΔΔCt values, some of selected cellular miRNAs can probably differentiate HIV-infected patients from healthy controls (Fig. 3a–e). According to the ROC results, miR-106-5p (AUC: 0.79; p value = 0.003), miR-140-3p (AUC: 0.7; p value = 0.02), miR-200c-3p (AUC: 0.73; p value = 0.01), and miR-339-3p (AUC: 0.93; p value <0.0001) have shown to be beneficial markers to discriminate the healthy controls from the HIV/HBV coinfected group. Furthermore, miR-125a (AUC: 0.75; p value = 0.02), miR-339-3p (AUC: 0.72; p value = 0.03), and miR-590-3p (AUC = 0.85; p value = 0.001) with optimal sensitivity and specificity were potentially useful biomarkers to differentiate EC group from healthy controls. According to ROC curve results, the PBMC level of miR-17-5p, miR-106-5p, and miR-191-5p were a useful marker for discriminating healthy control subjects from the HIV-naïve group, the AUC value of 0.82 (p value = 0.0003), 0.73 (p value = 0.01) and 0.68 (p value = 0.04), respectively (Fig. 3e). More information about the ROC curve analysis is given in Figure 3.

Also, to assess diagnostic value of the studied miRNAs in PBMC to discriminate EC from other HIV-infected groups, ROC curve analysis was performed. According to the result in Figure 4, miRNAs including miR-106-5p, miR-125a, miR-200c-3p, and -N367 with AUC: 0.72, p value = 0.04, AUC: 0.72, p value = 0.039, AUC: 0.74, p value = 0.03, and AUC: 0.79, p value = 0.007, respectively, can be used as tools to discriminate between ECs and ART-naïve HIV-infected groups (Fig. 4c). Moreover, the diagnostic power values of miR-590-3p in differentiating between ECs and HIV/HCV coinfected group were AUC: 0.88 (Fig. 4). In addition, according to the ROC curve analysis, miR-17-5p (AUC: 0.88; p value = 0.0005), miR-29a (AUC: 0.85; p value = 0.002), and miR-339-3p (AUC: 0.8; p value = 0.005) are useful markers for discriminating the EC group from the HIV/HBV coinfection group (AUC: 0.81; 95% CI = 0.65–0.93, p value = 0.004) (Fig. 4).

For investigating the patterns of expressing some cellular factors associated with HIV replication, the mean serum concentrations of p21, IL-2, SDF-1, XCL1, and CCL2 were examined by the ELISA method. According to the results obtained, the mean levels of p21 (1.42fold, p value <0.001), SDF-1 (1.48fold, p value <0.001), and XCL1 (1.3fold, p value <0.01) enhanced in ECs in comparison to the healthy group samples. Further, as shown in Figure 5a, a marked increase of CCL2 and SDF-1 levels were observed in the HIV/HBV and HIV/HCV coinfected groups versus the control group. In contrast, in HIV/HCV and HIV/HBV coinfected people, the levels of p21 were remarkably downregulated in comparison with the healthy controls.

No significant differences in the mean concentration of p21, IL-2, XCL1, and CCL2 were found between the HIV ART and healthy control groups. However, the mean level of SDF-1 in HIV ART samples was remarkably declined in comparison to the controls. And, among the studied groups, the level of IL-2 expression was significantly different only between ART-naïve HIV-infected group with the control group (Fig. 5a).

With regard to the outputs from the Spearman correlation analysis, we observed a negative association of miR-106a-5p to p21 (r = −0.63, p < 0.001), and between miR-200c-3p with SDF-1 (r = −0.54, p < 0.001) and between miR-590-3p with XCL1 (r = −0.49, p < 0.001) (Table 2). Assuming that the selected cellular factors may contribute to controlling the HIV infection, the expression levels of these factors in HIV ART and ART-naïve and coinfected groups were compared with the EC group (Fig. 5a). The results demonstrated that there were significant decreases in the mean concentration of p21 and SDF-1 in HIV-infected groups compared to the EC group. In addition, the mean level of CCL2 in the EC group was significantly higher than in ART-naïve HIV-infected, HIV/HBV, and HIV/HCV coinfection groups as well as the mean level of XCL1 was significantly decreased in HIV/HBV coinfection, HIV ART, and ART-naïve HIV-infected group compared to the EC group. Further information is shown in Figure 5b.

In this report, we investigate the different expression patterns of some miRNAs (miR-17-5p, - 29a, -106a-5p, -125a, -140-3p, -191-5p, -200c-3p, -339-3p, and -590-3p), cellular factors (CD4+ T-cell, p21, IL-2, SDF-1, XCL1, and CCL2), and viral factors (nef and miR-N367) in 5 patient groups and compare them to a healthy control group. Additionally, in four groups including HIV/HBV and HIV/HCV coinfection groups, ART HIV and ART-naïve HIV-infected groups are compared to ECs. For the first time, the correlation among the expression levels of mentioned miRNAs with cellular factors such as p21, IL-2, SDF-1, XCL1, CCL2, CD4+ T-cell, and viral factors such as HIV/HCV/HBV load and nef expression are evaluated. Understanding the mechanisms of controlling HIV replication and the immunovirological ways involved in this process in ECs can provide innovative ways toward diagnosis and controlling HIV infection. Therefore, in the current study, ECs are compared to other involved groups to identify a possible new candidate biomarker.

The viral infection or life cycle is affected by miRNAs by two main ways. In the first, more direct, way, the interaction between miRNAs and the 3′ UTR of viral mRNAs lead to the suppression and/or promotion of viral mRNA translation. In the second or indirect way, the regulation of host gene expression contributes to the viral replication or cellular physiology [27]. MiRNA profiles are considered as important factors in response to viral infections because it has been revealed that miRNAs contribute to different aspects of cell/viral metabolism such as immune response and viral infectivity/replication [28].

In the current study, we demonstrate the differences between miRNA expression pattern in HIV/HBV and HIV/HCV coinfection groups, ART HIV, and ART-naïve HIV-infected groups compared to the EC group (Fig. 2). Our results reveal an expression pattern of some selected cellular miRNAs in ECs that are almost similar to healthy and ART HIV samples, while they were shown to be different compared to other patient groups. Also, our findings indicated the mean fold change of some miRNAs were remarkably different between the ART-naïve HIV-infected group and the EC group and we found the highest increase and decrease in expression for miR-17-5p (4.61fold, p value <0.0001) and miR-140-3p (−3.49fold, p value <0.0001), respectively, although it was found that ECs develop anti-HIV antibodies and the HIV viral load in these individuals often is at undetectable levels. Therefore, there is a risk that ECs will be identified as HIV-negative individuals [29, 30]. However, there are no specific laboratory algorithms to identify ECs [31, 32]. This in turn proves that miRNAs potentially could be used for the diagnosis of ECs. In a recent research, Biswas et al. [33] demonstrated that some miRNAs such as miR-223-3p, miR-195-5p, miR-20b-5p, and miR-16-5p could distinguish between healthy controls and HIV infection (early stage). It has been reported that the expression level of miR-590, miR-339, miR-200c, miR-191, miR-140 decreases in ECs and healthy control groups compared to and viremic patients and HIV-infected patients under ART [28]. Similarly, in the current study, expression level of miR-339-3p, miR-200c-3p, and miR-191 in ECs was significantly lower than that of HIV-infected patients under the ART group (Fig. 2); however, there was no significant difference in expression level of miR-140-3p and miR-590-3p between ECs and HIV-infected patients under ART groups.

One of the most important aspects of ECs is the role of cellular factors in these patients. Our results revealed, the mean levels of host anti-HIV factors (p21 and SDF-1) significantly increased in ECs in comparison to the healthy samples (p21: 1.42fold, p value <0.001; SDF-1: 1.48fold, p value <0.001) (Fig. 5a). The expression levels of p21 and SDF-1 were considerably downregulated in all patient groups compared to ECs (Fig. 5b). Additionally, outputs from Spearman correlation analysis demonstrated a significant reverse association of miR-106a-5p and p21 (r = −0.63, p < 0.01), miR-200c-3p with SDF-1 (r = −0.54, p < 0.001), as well as miR-590-3p and XCL1 (r = −0.49, p < 0.001) (Table 2). P21 protein leads to HIV replication through inhibitor of cyclin-dependent kinases (CDKs; a group of cellular enzymes that have a role in regulating the cell cycle and required for replication of some viruses such as HIV) [34]. Among the selected miRNAs, miR-106a-5p and −590 were negatively correlated with p21 expression. Put differently, the level of miR-106a-5p and -590 expression is low in ECs. SDF-1 has different roles such as roles in B cell lymphopoiesis, angiogenesis, and several cancers [8, 35] and it has been shown that SDF-1 can restrict entry of HIV by blocking binding of the HIV gp120 glycoprotein to CXCR4 [36]. Furthermore, SDF-1 can inhibit the accumulation of HIV proviral DNA, which is an important step in the HIV replication cycle [36, 37]. XCL1 is a chemokine with a crucial effect on the immune system, as an activator of CD8+ T lymphocytes so that the natural killer cells secrete this chemokine in infections [38]. Moreover, levels of miR-17, -191, and −590 expression were negatively correlated with XCL expression. These results propose p21 and SDF-1 as important cellular factors involved in HIV replication inhibition in ECs and could be used for finding novel ways for HIV treatment. The levels of CCL2 expression in the HIV/HCV and HIV/HBV coinfection groups were remarkably higher than in ART-naïve HIV-infected groups and significantly higher than that of ECs and HIV ART. This pro-inflammatory chemokine (CCL2) is one of the essential agents in the induction of chronic inflammation. Previous researchers demonstrated a multifaceted effect of CCL2 in HIV pathogenesis and progression of HIV-induced disease [39]. CCL2 expression had a significantly inverse correlation to the expression levels of miR-125 and -N367. Tatro et al. [40] found that the expression level of miR-125a is overexpressed in HIV-infected patients and ectopic expression of this miRNA leads to inducing HIV replication by suppressing protein translation of interferon-induced transmembrane protein 3 (IFITM3) and soluble tumor necrosis factor receptor (sTNFR1A) [40]. Reportedly, HIV promotes viral protein translation mediated with PCAF by downregulating the expression level of miR-17-5p; in return, miR-17-5p can inhibit HIV replication by suppressing PCAF [41]. However, in our study, the miR-17-5p and miR-125a levels were upregulated in HIV-coinfected groups and HIV ART group compared to the EC group. Overall, due to the small sample size in the current study, the results are not reliable and require more studies to understand the performance of these miRNAs in ECs.

It has been found that miR-29a and -29b express in T cells and can suppress HIV replication and HIV nef expression by targeting the nef gene and 3′-UTR of the HIV genome, respectively [42‒44]. Thus, these miRNAs may play a critical role in the control of HIV infection, especially in EC subjects. However, in the current study, there was no significant difference in the expression pattern of miR-29a between the EC group and the healthy control group. Thus, more experimental studies are needed for understanding the effect of this miRNA on HIV infection in ECs.

Another highly important aspect of ECs is the differential diagnosis between ECs and other HIV patients (coinfection, ART-naïve HIV-infected, HIV ART) and even healthy people. As mentioned before, there are no testing algorithms specific to ECs and there is a risk for misdiagnosis of HIV infections. Therefore, there are potentials for miRNAs to be used for the diagnosis of ECs which we evaluate using ROC curve analysis. According to the ROC results, miR-125a, miR-339-3p, and miR-590-3p with optimal sensitivity and specificity were potentially useful biomarkers to differentiate the EC group from the healthy controls (Fig. 3c). According to the results, miRNAs (miR-106-5p, miR-125a, miR-200c-3p, and -N367) can be used as tools to discriminate between EC and ART-naïve HIV-infected groups. Moreover, the diagnostic power values of miR-590-3p in differentiating between the EC and HIV/HCV coinfected groups was with AUC: 0.88 (Fig. 4). In addition, according to the ROC curve analysis, miR-17-5p, miR-29a, and miR-339-3p are useful markers for discriminating the EC group from the HIV/HBV coinfection group (Fig. 4). Reportedly, EC status cannot prevent super-infection and acute retroviral syndrome following it [45]. Although limited information is available, some studies have suggested that the risk of some non-HIV-related diseases such as cardiovascular disease is high in ECs [46‒48]. Given that the mechanism of controlling HIV infection is not completely known yet in EC subjects, as well as there are not being any specific methods for distinctive different HIV-infected groups, it can be concluded that identifying suitable factor candidates as potential therapeutic and diagnostic biomarkers is needed. Here, for the first time, we assessed the differentially expressed selected immunological factors and cellular/viral miRNAs in ECs compared to ART-naïve HIV, HCV/HIV, HIV ART, and HBV/HIV coinfected cases. Given that some of the selected cellular factors and miRNAs maybe play roles in the regulation of host immune response and control of viral replication, thus, these factors can be introduced as novel therapeutic targets and/or biomarkers for the diagnosis of ECs and differentiation from other HIV-infected groups. In conclusion, deregulation of the assessed miRNAs and cellular factors may provide new ways toward selecting novel candidates for diagnosis and controlling HIV infection.

The small sample size was a limitation of our study and validation studies with larger sample sizes are needed to better understand the role of selected miRNAs as diagnostic or prognostic biomarkers for discriminating HIV-infected groups. In addition, investigating the role of these miRNAs in the HIV life cycle or their roles in regulating the antiviral immune responses requires more studies.

The research was directed ethically according to the World Medical Association Declaration of Helsinki. The study subjects have given their written informed consent, and the study protocol was approved by the Medical Ethics Committee of Tabriz University of Medical Sciences Institutional Review Board (IR.TBZMED.REC.1400.149).

The authors declare that they have no conflicts of interest.

This project was supported by Kashan University of Medical Sciences, Kashan, Iran.

Mohsen Moghoofei and Hossein Bannazadeh Baghi related to the concepts, design, as well as drafting of the manuscript; Bashdar Mahmud Hussen, Majid Noori, Babak Sayad, Maryam Ebadi Fard Azar, Hamed Mirzaei, Javid Sadri Nahand, Mobina Bayat, Farhad Babaei, Romina Karampour, and Farah Bokharaei-Salim contributed to data collection, statistical analyses, and drafting the manuscript. Each scholar verified the resulting version to submit the paper.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.