HIV Promotes Atherosclerosis via Circulating Extracellular Vesicle MicroRNAs

People living with HIV (PLHIV) are at a higher risk of having cerebrocardiovascular diseases (CVD) compared to HIV negative (HIVneg) individuals. The mechanisms underlying this elevated risk remains elusive. We hypothesize that HIV infection results in modified microRNA (miR) content in plasma extracellular vesicles (EVs), which modulates the functionality of vascular repairing cells, i.e., endothelial colony-forming cells (ECFCs) in humans or lineage negative bone marrow cells (lin− BMCs) in mice, and vascular wall cells. PLHIV (N = 74) have increased atherosclerosis and fewer ECFCs than HIVneg individuals (N = 23). Plasma from PLHIV was fractionated into EVs (HIVposEVs) and plasma depleted of EVs (HIV PLdepEVs). HIVposEVs, but not HIV PLdepEVs or HIVnegEVs (EVs from HIVneg individuals), increased atherosclerosis in apoE−/− mice, which was accompanied by elevated senescence and impaired functionality of arterial cells and lin− BMCs. Small RNA-seq identified EV-miRs overrepresented in HIVposEVs, including let-7b-5p. MSC (mesenchymal stromal cell)-derived tailored EVs (TEVs) loaded with the antagomir for let-7b-5p (miRZip-let-7b) counteracted, while TEVs loaded with let-7b-5p recapitulated the effects of HIVposEVs in vivo. Lin− BMCs overexpressing Hmga2 (a let-7b-5p target gene) lacking the 3′UTR and as such is resistant to miR-mediated regulation showed protection against HIVposEVs-induced changes in lin− BMCs in vitro. Our data provide a mechanism to explain, at least in part, the increased CVD risk seen in PLHIV.


Introduction
The advent of antiretroviral therapy (ART) has changed the clinical course of HIV infection from a fatal disease to a chronic illness by preventing the development of acquired immunodeficiency syndrome (AIDS) through long-lasting viral suppression [1]. However, ART does not fully eradicate the virus, and people living with HIV (PLHIV) develop non-AIDS-related comorbidities such as diabetes mellitus, respiratory diseases, hepatic diseases, and cerebrocardiovascular disease (CVD) [2,3]. The persistent low-level viremia, detectable only by ultrasensitive PCR, coupled with the chronic state of immune activation and inflammation, represent the most significant factor contributing to the elevated risk for the development of CVD [4]. Indeed, epidemiological studies have shown

PLHIV Have Increased Carotid Intima-Media Thickness (cIMT) and Decreased ECFC Numbers
To study the effects of HIV infection on atherogenesis, we recruited 74 PLHIV on ART with well-controlled viral loads (undetectable HIV RNA by RT-PCR) and 23 HIV neg individuals for this study (Supplemental Table S1). To minimize the confounding of other risk factors for atherosclerosis and for circulating ECFC levels, we excluded individuals with a history of myocardial infarction, stroke, diabetes, malignancy, pregnancy, smoking, and illicit drug use to allow us to evaluate the degree of atherosclerosis and ECFC reduction due to long-term HIV infection. PLHIV had a higher atherosclerosis burden, as evidenced by increased cIMT compared to HIV neg individuals ( Figure 1A). The number of circulating ECFCs (formerly known as late endothelial progenitor cells or EPCs) was significantly decreased in PLHIV, compared to the HIV neg individuals as measured by CFUs ( Figure 1B). These results are consistent with previous work showing increased atherosclerosis burden and decreased ECFCs levels in PLHIV, including those with well controlled viral loads and no differences in traditional risk factors relative to the HIV neg individuals [7]. These data support a role for HIV-specific factors in promoting atherogenesis and reducing vascular repair.
Int. J. Mol. Sci. 2023, 24, x FOR PEER REVIEW 3 controlled viral loads) produces EVs with altered cargo that promote atherogen providing a plausible mechanism for the increased CVD risk seen in these individua

PLHIV Have Increased Carotid Intima-Media Thickness (cIMT) and Decreased ECFC Numbers
To study the effects of HIV infection on atherogenesis, we recruited 74 PLHI ART with well-controlled viral loads (undetectable HIV RNA by RT-PCR) and 23 H individuals for this study (Supplemental Table S1). To minimize the confounding of o risk factors for atherosclerosis and for circulating ECFC levels, we excluded individ with a history of myocardial infarction, stroke, diabetes, malignancy, pregnancy, sm ing, and illicit drug use to allow us to evaluate the degree of atherosclerosis and E reduction due to long-term HIV infection. PLHIV had a higher atherosclerosis burde evidenced by increased cIMT compared to HIV neg individuals ( Figure 1A). The numb circulating ECFCs (formerly known as late endothelial progenitor cells or EPCs) wa nificantly decreased in PLHIV, compared to the HIV neg individuals as measured by C ( Figure 1B). These results are consistent with previous work showing increased ath sclerosis burden and decreased ECFCs levels in PLHIV, including those with well trolled viral loads and no differences in traditional risk factors relative to the HIV neg viduals 7 . These data support a role for HIV-specific factors in promoting atheroge and reducing vascular repair.

HIV posEVs Promote Atherogenesis in apoE −/− Mice
To investigate the potential mechanisms underlying the increased atheroscle burden seen in PLHIV, we focused on circulating EVs from peripheral blood. To tes hypothesis that EVs from PLHIV carry the signals that modulate atherogenesis, a pr erogenic animal model was used in which 3-week-old apoE −/− mice fed a high-fat diet injected weekly for 12 weeks with: (1) circulating EVs from PLHIV (HIV posEVs ), (2) PL plasma depleted of EVs (HIV PL depEVs ), (3) circulating EVs from HIV neg subjects (HIV n or (4) PBS. These mice were sacrificed following 12 weeks of treatment for pathobiolo analysis. H&E and Oil Red O (ORO) staining were performed to evaluate the athero rosis burden. Mice treated with HIV posEVs had a higher atherosclerosis burden, as m ured by both H&E and ORO staining, than animals that were treated with PBS (Figu Interestingly, no increase in atherosclerosis was seen in mice injected with either . cIMT in PLHIV is higher than that in HIV neg subjects (* p < 0.05). (B). Circulating ECFC levels as determined by CFUs in PLHIV are lower than in HIV neg individuals (*** p < 0.001). PLHIV (N = 74) and HIV neg subjects (N = 23). Statistical test used: student's t-test.

HIV posEVs Promote Atherogenesis in apoE −/− Mice
To investigate the potential mechanisms underlying the increased atherosclerosis burden seen in PLHIV, we focused on circulating EVs from peripheral blood. To test our hypothesis that EVs from PLHIV carry the signals that modulate atherogenesis, a proatherogenic animal model was used in which 3-week-old apoE −/− mice fed a high-fat diet were injected weekly for 12 weeks with: (1) circulating EVs from PLHIV (HIV posEVs ), (2) PLHIV plasma depleted of EVs (HIV PL depEVs ), (3) circulating EVs from HIV neg subjects (HIV negEVs ), or (4) PBS. These mice were sacrificed following 12 weeks of treatment for pathobiological analysis. H&E and Oil Red O (ORO) staining were performed to evaluate the atherosclerosis burden. Mice treated with HIV posEVs had a higher atherosclerosis burden, as measured by both H&E and ORO staining, than animals that were treated with PBS ( Figure 2). Interestingly, no increase in atherosclerosis was seen in mice injected with either HIV PL depEVs or HIV negEVs compared to the PBS treated mice. Analysis of cholesterol levels showed that there were no significant differences in total cholesterol, triglycerides, HDL-C, or LDL-C across all mouse groups (HIV posEVs , HIV PL depEVs , HIV negEVs , or PBS). However, HIV posEVs -treated mice had increased serum levels of C-reactive protein compared to the other treatment groups (Supplemental Figure S1). These data indicate that EVs, rather than other fractions of plasma, serve as carriers of the pro-atherosclerotic signals in PLHIV.
In addition, HIV posEVs seem to work via factors other than circulating lipids to promote accelerated atherosclerosis in apoE −/− mice.
PL depEVs or HIV negEVs compared to the PBS treated mice. Analysis of cholesterol levels showed that there were no significant differences in total cholesterol, triglycerides, HDL-C, or LDL-C across all mouse groups (HIV posEVs , HIV PL depEVs , HIV negEVs , or PBS). However, HIV posEVs -treated mice had increased serum levels of C-reactive protein compared to the other treatment groups (Supplemental Figure S1). These data indicate that EVs, rather than other fractions of plasma, serve as carriers of the pro-atherosclerotic signals in PLHIV. In addition, HIV posEVs seem to work via factors other than circulating lipids to promote accelerated atherosclerosis in apoE −/− mice.

HIV posEVs Cause Vascular Cell Senescence and Apoptosis In Vivo
To further investigate HIV posEVs -mediated atherogenesis, we asked if HIV posEVs treatment in apoE −/− mice could cause vascular senescence and apoptosis-cellular processes associated with atherogenesis. We found that there were significantly higher numbers of senescent and apoptotic cells in the aortic wall, including cells in the endothelial layer, from HIV posEVs -treated apoE −/− mice compared with the aortas isolated from mice treated with HIV PL depEVs , HIV negEVs , or PBS ( Figure 3; enlargements of representative images are available in Supplemental Figure S2). These results indicate that treatment of the apoE −/− mice with HIV posEVs , but not HIV PL depEVs or HIV negEVs , induces cell senescence and apoptosis in ECs and SMCs in the vascular wall-phenotypic changes associated with accelerated cellular aging that promote atherosclerosis.

HIV posEVs Cause Vascular Cell Senescence and Apoptosis In Vivo
To further investigate HIV posEVs -mediated atherogenesis, we asked if HIV posEVs treatment in apoE −/− mice could cause vascular senescence and apoptosis-cellular processes associated with atherogenesis. We found that there were significantly higher numbers of senescent and apoptotic cells in the aortic wall, including cells in the endothelial layer, from HIV posEVs -treated apoE −/− mice compared with the aortas isolated from mice treated with HIV PL depEVs , HIV negEVs , or PBS ( Figure 3; enlargements of representative images are available in Supplemental Figure S2). These results indicate that treatment of the apoE −/− mice with HIV posEVs , but not HIV PL depEVs or HIV negEVs , induces cell senescence and apoptosis in ECs and SMCs in the vascular wall-phenotypic changes associated with accelerated cellular aging that promote atherosclerosis.

HIV posEVs Affect lin − BMC Senescence and Functionality
To elucidate the mechanisms whereby HIV posEVs mediated the effects of HIV infection on atherogenesis, decreased ECFC levels, and vascular senescence, we examined the effects of HIV posEVs treatment on lin − BMC senescence and functionality. The lin − BMC fraction is enriched for EPCs that contribute to ECFCs, and previous work has demonstrated that lin − BMCs promote vascular repair and angiogenesis [12,13,26]. Similar to what was seen with the vascular cells, we observed a relatively high percentage of β-gal positive lin − BMCs in the PBS control group, which is consistent with premature aging seen in apoE −/− mice [27]. Treatment with HIV posEVs further accelerated lin − BMC senescence ( Figure 4A,B). There was no significant difference in the percentage of β-gal positive cells in the HIV PL depEVs and HIV negEVs treatment groups compared to the PBS control treatment. In addition, lin − BMCs from the HIV posEVs -treated animals had decreased migratory capacity, as determined by the wound healing (scratch) assay ( Figure 4C,D), as well as diminished proliferation, as measured by the MTT assay ( Figure 4E). This decreased migratory capacity and proliferation were not seen in the HIV negEVs or HIV PL depEVs treatment groups. To further assess the impact of HIV posEVs on lin − BMC health and functionality, lin − BMCs isolated from 3-week old apoE −/− mice were cultured in vitro in the presence of HIV posEVs , HIV PL depEVs , HIV negEVs , or PBS. The in vitro treatment of lin − BMCs with HIV posEVs for 96 h showed significantly elevated levels of cell senescence compared to lin − BMCs treated with either HIV PL depEVs , HIV negEVs , or PBS, as measured by β-gal staining ( Figure 4F,G). Therefore, both in vivo and in vitro experiments support that exposure to HIV posEVs , but not HIV PL depEVs or HIV negEVs , increases lin − BMC senescence with decreased cellular functionality. Since at least some of the ECFCs are believed to be derived from the EPCs encompassed in lin − BMCs, it is reasonable to assume that HIV posEVs -induced lin − BMC senescence may have contributed to the decreased ECFCs in PLHIV and the vascular aging process in apoE −/− mice treated with HIV posEVs .

HIV posEVs Impairs EC Functionality
Next, we determined if mature vascular ECs would also be affected in the same manner as lin − BMCs by HIV posEVs . To that end, we cultured ECs isolated from 3-week-old apoE −/− mice for 96 h with HIV posEVs , HIV PL depEVs , HIV negEVs , or PBS. Similar to lin − BMCs, ECs treated with HIV posEVs , but not HIV PL depEVs or HIV negEVs , showed impaired migration ( Figure 5A,B) and reduced proliferation ( Figure 5C) compared to PBS-treated ECs. These results indicate that HIV posEVs not only affect lin − BMC health and functionality but also directly target ECs. It is plausible that the combined action of the HIV posEVs on lin − BMCs and ECs may have contributed to the accelerated vascular senescence, injury, and atherogenesis.

HIV posEVs and HIV negEVs Show Similar Physical Characteristics and Surface Markers
Our data support the role of circulating EVs in mediating, at least in part, the effects of HIV infection on atherosclerosis. To investigate the effect of HIV infection on circulating EVs, we first compared the physical properties of EVs isolated from PLHIV and HIV neg individuals using the Nanoparticle Tracking Analyses (NTA). EVs from both groups showed similar biophysical characteristics, including size, protein concentration, as well as particle/protein, particle/RNA, and particle/lipid ratios (Supplemental Figure  S3A-E). The absence of multiple peaks in the size distribution histograms suggests that both HIV posEVs and HIV negEVs are composed of a uniform distribution of vesicle sizes (Supplemental Figure S3C,D). Staining of the HIV posEVs and HIV negEVs with antibodies for EV surface markers-CD63, CD81, CD9, and Hsp70-showed that EVs from both groups contained similar levels of these well-established EV surface markers as measured by flow cytometry (Supplemental Figure S3F,G). To evaluate if the EVs prepared from the plasma of PLHIV were not contaminated by viral particles, HIV posEVs and HIV negEVs samples were analyzed for HIV p24 levels by ELISA. There was minimal p24 expression in both groups, and there was no difference between groups (Supplemental Figure S3H). To evaluate if the EV preparations were contaminated with lipoproteins that could influence the atherosclerotic phenotype, we assessed ApoB 100 and ApoA1 levels by Western blot analysis. This analysis showed that there was no lipoprotein contamination in the EVs (Supplemental Figure S3I). These results indicate that HIV infection does not significantly alter the biophysical characteristics and surface marker expression patterns of plasma EVs.  ( Figure 5A,B) and reduced proliferation ( Figure 5C) compared to PBS-treated ECs. These results indicate that HIV posEVs not only affect lin − BMC health and functionality but also directly target ECs. It is plausible that the combined action of the HIV posEVs on lin − BMCs and ECs may have contributed to the accelerated vascular senescence, injury, and atherogenesis.

HIV posEVs and HIV negEVs Show Similar Physical Characteristics and Surface Markers
Our data support the role of circulating EVs in mediating, at least in part, the effects of HIV infection on atherosclerosis. To investigate the effect of HIV infection on circulating EVs, we first compared the physical properties of EVs isolated from PLHIV and HIV neg individuals using the Nanoparticle Tracking Analyses (NTA). EVs from both groups

HIV posEVs Have Differential microRNA Cargo Compared to HIV negEVs
Next, we examined whether HIV infection altered the composition of EV cargo, which, in turn, could mediate the effects of HIV infection on atherosclerosis. Since we have previously shown that miRs play important regulatory roles in lin − BMC senescence and CVD [12,13,28], we compared the miR profiles in EVs from PLHIV with the largest plaque areas (N = 5) and HIV neg subjects with no plaques (N = 5), representing the extremes of the disease spectrum. All individuals were male, non-smokers, and had an average age of 47 and 46 years, respectively. Overall, 415 mature miRs were shown to be present in the EVs in both groups. Among these miRs, 126 miRs were upregulated and 49 miRs were downregulated by at least 2-fold in HIV posEVs relative to HIV negEVs (Supplemental Figure S4A). Among the 126 upregulated miRs in HIV posEVs , 43 miRs had ≥100 counts per million reads mapped (CPM) in the HIV posEVs samples. Of these 43 miRNAs, let-7b-5p, miR-16-5p, miR-423-5p, miR-103a-3p, and miR-107 were the most abundant miRs in the HIV posEVs . The average CPM of these five miRs in each group is shown in Supplemental Figure S4B.
Validation of differential expression of these miRs in HIV posEVs versus HIV negEVs was performed using qRT-PCR in an independent sample set (N = 5 samples in each group). Four of these miRNAs: let-7b-5p, miR-16-5p, miR-423-5p, and miR-103a-3p, were validated by qRT-PCR (Supplemental Figure S4C). Putative mRNA targets were identified for each of the 4 validated EV-miRs using Ingenuity Pathway Analysis Software (IPA) (Qiagen) (Supplemental Table S3). This analysis showed that the target genes of the four validated EV-miRs were enriched in biological processes involved in cardiac enlargement, cardiac fibrosis, and cardiac cell necrosis/death (Supplemental Figure S4D). A review of literature on these miRs shows that the predicted targets of Let-7b-5p are involved in cardiovascular pathologies such as cardiac fibrosis, dilated cardiomyopathy (DCM), myocardial infarction (MI), arrhythmia, angiogenesis, atherosclerosis, and hypertension, while the targets of miR-16-5p and miR-103a-3p are involved in EPC senescence and impairment of cellular functions, such as proliferation (Supplemental Table S2). Collectively, these results indicate that HIV posEVs have higher levels of miRs that could be involved in molecular pathways that contribute to atherosclerosis and lin − BMC dysfunction. These analyses suggest that the candidate EV-miRs had the potential to mediate the effects of HIV posEVs and hence HIV infection, on the cardiovascular system.

Discussion
It is well documented that PLHIV have an elevated risk of developing a wide of pathologies, including CVD [31]. Several mechanisms have been implicated in t velopment of CVD in PLHIV, including oxidative stress, ER stress, NLRP3 inflamma activation, and autophagy inhibition (reviewed in [32]). Indeed, CVD and non-AID fining cancers are the main causes of morbidity and mortality for PLHIV, particula the US, where most PLHIV achieve adequate viral suppression [33,34]. We rec PLHIV and HIV neg individuals with similar cardiovascular risk factors, and exclude tors that could affect ECFC levels, including malignancy, pregnancy, smoking, and drug use, to determine the atherosclerosis burden and ECFC reduction due to long HIV infection. We found that cIMT measurement, the surrogate marker for atheroscl burden, was significantly increased in PLHIV compared with HIV neg individuals. A

Discussion
It is well documented that PLHIV have an elevated risk of developing a wide range of pathologies, including CVD [31]. Several mechanisms have been implicated in the development of CVD in PLHIV, including oxidative stress, ER stress, NLRP3 inflammasome activation, and autophagy inhibition (reviewed in [32]). Indeed, CVD and non-AIDSdefining cancers are the main causes of morbidity and mortality for PLHIV, particularly in the US, where most PLHIV achieve adequate viral suppression [33,34]. We recruited PLHIV and HIV neg individuals with similar cardiovascular risk factors, and excluded factors that could affect ECFC levels, including malignancy, pregnancy, smoking, and illicit drug use, to determine the atherosclerosis burden and ECFC reduction due to long term HIV infection. We found that cIMT measurement, the surrogate marker for atherosclerosis burden, was significantly increased in PLHIV compared with HIV neg individuals. Atherosclerosis is an inflammatory process that develops in response to endothelial injury [35,36]. Multiple lines of evidence indicate that ECFCs play an important role in repairing injured vessels and in atherogenesis [8]. Conflicting data, however, exist regarding ECFC changes in PLHIV and their role in HIV-associated CVD development [37]. In this study, we report reduced ECFC levels in PLHIV compared to HIV neg individuals, supporting the role of ECFCs in HIV-associated atherosclerosis.
Several factors, including HIV infection itself, chronic inflammation, and persistent immune activation, are implicated in the increased CVD risk seen in PLHIV [31]. Evidence suggests that sustained monocyte/macrophage activation is a key factor in chronic inflammation that would favor the development of HIV-associated atherogenesis [38]. We hypothesized that CVD risk factors, particularly HIV-specific factors, affect the composition of the cargo of plasma EVs, specifically their miR content. This alteration in EV-miR composition, in turn, promotes the development of CVD. Virtually all cells in the body secrete EVs, and they are present in all biofluids. However, circulating blood cells and endothelial cells lining the inner lumen of the blood vessels are directly affected by the low-level HIV viremia and/or chronic inflammation/immune activation seen in PLHIV, including those with well-controlled viral loads [33,34]. As a result, they are more likely to produce EVs with modified characteristics/cargo reflecting the persistent inflammatory state associated with HIV infection [35]. Thus, analysis of plasma EVs may provide important mechanistic insights about how HIV infection increases the risk for atherogenesis. To study the effects of HIV posEVs in mediating atherosclerosis development, we chose the well-established apoE −/− mouse model fed a high fat/high cholesterol diet. Injection of HIV posEVs exacerbated atherosclerosis development in these mice. In contrast, HIV PL depEVs or HIV negEVs had no effect on atherosclerosis burden, showing similar plaque sizes to those seen in the PBS treated mice. These results show, for the first time, that the effects of HIV infection on CVD are mediated, at least in part, by EVs present in the peripheral blood. This model of HIV posEVs /apoE −/− mice also represents a novel model system to study atherosclerosis associated with HIV infection. The use of HIV posEVs and HIV PL depEVs allowed us to effectively determine the impact of the factors encompassed in the plasma on atherogenesis. The observed differences in atherosclerosis burden between HIV posEVs -and HIV negEVs -treated apoE −/− mice, together with the lack of differences in atherosclerosis burden between HIV PL depEVs -and PBS-treated apoE −/− mice, suggest that EVs, rather than other factors in the plasma such as lipid proteins, serve as the primary carriers for the atherogenic signals associated with HIV infection. The similar lipid protein profile between HIV posEVs -and HIV negEVs -treated apoE −/− mice (with different atherosclerosis burdens) also suggests that cholesterol levels are not a determining factor for the increased atherogenesis associated with HIV posEVs injection in apoE −/− mice.
To determine the potential mechanisms underlying the effects of HIV posEVs /HIV infection on atherosclerosis, we focused on ECFCs, lin − BMCs, and vascular wall cells. Senescence and dysfunction of ECFCs have been associated with atherosclerosis initiation and progression. Furthermore, injection of BM cells ameliorated atherosclerosis development in apoE −/− mice [36]. Several groups, including our laboratory, have shown that lin − BMCs can induce angiogenesis and vascular repair [12,13]. In this study, we have observed that injection of HIV posEVs into apoE −/− mice resulted in lin − BMC senescence, as well as impaired proliferation and migration, whereas HIV PL depEVs and HIV negEVs showed no discernible effects on these cells compared to PBS treated mice. β-galactosidase and TUNEL staining revealed increased cellular senescence and apoptosis in the aortas from mice treated with HIV posEVs compared with mice treated with HIV PL depEVs , HIV negEVs , or PBS. We also studied the effects of HIV posEVs on lin − BMCs and ECs in vitro. Similar to the in vivo observations, exposure of lin − BMCs in vitro to HIV posEVs , but not HIV PL depEVs or HIV negEVs , resulted in elevated lin − BMC senescence. Furthermore, HIV posEVs caused decreased EC migration and proliferation in vitro. Thus, the increased vascular cell senescence and apoptosis observed in the apoE −/− mice treated with HIV posEVs may be the result of both deficient vascular repair and direct insult to vascular cells. The lack of differences in the levels of HDL-C and LDL-C cholesterol levels between apoE −/− mice treated with HIV posEVs and controls (Supplemental Figure S1) provide further support for lin − BMC and vascular cell senescence and apoptosis as possible mechanisms, instead of cholesterol levels, that underlie the effects of HIV infection-specific factors on CVD mediated by HIV posEVs .
To study how HIV infection impacted EVs, we first studied the biophysical properties and EV surface markers and found no significant differences between the two groups [39]. We then focused on the cargo content by specifically analyzing the miR composition of the EVs. Each miR has the potential to regulate multiple target genes and pathways, and in that way exerting greater biological effects than individual proteins and mRNAs. Furthermore, we have previously demonstrated that miR dysregulation plays a key role in lin − BMC senescence, functional impairment, and their angiogenic potential [12,13,28]. Comprehensive analyses and prioritization of the sRNA-seq data obtained in HIV posEVs and HIV negEVs led to the identification and validation of four EV-miRs that are elevated in HIV posEVs relative to HIV negEVs (≥100 copies and ≥2-fold different between groups). Importantly, these EV-miRs and their target genes are implicated in a variety of pathobiological processes consistent with CVD development, including cell senescence, DNA damage, cellular processes, chronic inflammation, hypertension, cardiomyopathy, and atherosclerosis.
To investigate the role of these candidate EV-miRs in regulating lin − BMC and vascular cell senescence and in affecting the cardiovascular system, we selected two of candidates for further analysis: let-7b-5p and miR-103a-3p. Let-7b-5p has been shown to regulate neural stem cell (NSC) senescence, has been found in myocardial infarction, and is associated with pulmonary arterial hypertension [40][41][42]. MiR-103a-3p has been previously shown to be upregulated in senescent BM stem cells [43]. Furthermore, circulating miR-103a-3p has been shown to contribute to angiotensin II-induced renal inflammation and fibrosis via a SNRK/NF-κB/p65 regulatory axis [44]. Using the approach of generating tailored EVs (TEVs) through genetic modification of MSCs first described by our group [12], we produced TEVs loaded with individual antagomirs for let-7b-5p or miR-103a-3p. Remarkably, these miRZip-let-7b-5p and miRZip-103a-3p TEVs were able, at least partially, to abrogate the effects of HIV posEVs in inducing lin − BMC senescence.
Due to logistical difficulties, we used plasma EVs isolated from previously frozen plasma samples. Freezing and thawing plasma samples could activate platelet and increase the chance of contamination of EVs by EV-like particles produced by platelets activation [49]. However, all samples used in this study from both groups (PLHIV and HIV negative) underwent the same treatment and storage procedures. Therefore, the difference between the effects of HIV negEVs and HIV posEVs could not be accounted for by contamination from EV-like particles produced by platelet activation. To address any potential impact of freezing and thawing on EVs, we performed thorough characterization of the plasma EVs used in this study for their protein membrane profile per the guidelines of the International Society for Extracellular Vesicles (ISEV) [39]. We further analyzed the presence of platelet-derived growth factor (PDGF), which is expressed by platelets, in the plasma EV preparations), and were unable to detect any evidence of levels of PDGF in our EV preparations, arguing against contamination of the EVs by platelet activation.
Our results are consistent with a previous study showing that plasma EVs from PLHIV with controlled viral loads induce necrosis in HUVECs [50]. In this work, we take an additional step in exploring the possible molecular mechanism underlying the effects of plasma EVs derived from PLHIV on vascular endothelial cells. Our current work focused on plasma EVs-miRNAs and identified several candidate EV-miRNAs that are overexpressed in EVs from PLHIV that account for, partially, the effect of HIV posEVs on atherogenesis. In addition to EV-miRNAs, other EV cargos have been investigated by others. Recently, it has been reported that HIV virally encoded proteins, e.g., nef, can be detected in EVs from bronchoalveolar lavage (BAL) fluid in patients with undetectable viral loads [51]. The Nef carrying EVs induces endothelial cell apoptosis. As such, it is imperative to further investigate different classes of molecules in future studies to fully elucidate the mechanisms at play in the effects of PLHIV EVs on the cardiovascular system.
Although the results presented here show that miRs derived from EVs play a central role in mediating the effect of HIV infection and ART on the development of CVD, we can not rule out that additional molecules delivered by the EVs may also contribute to the development of atherosclerosis in PLHIV. In addition to miRs, EVs contain long non-coding RNAs, mRNAs, proteins, and other classes of molecules that make up EV cargo. In addition, we have shown that PLHIV have decreased numbers of ECFCs compared to HIV negative individuals with similar CVD risk factors. However, we have not shown that this decrease in ECFCs is due directly to exposure to EVs. In future work, we will assess the impact of EVs derived from PLHIV on human ECFCs from uninfected individuals. This will confirm that human ECFCs respond in the same fashion as mouse lin − BMCs to HIV posEVs . Furthermore, a comparison between the plasma EVs-miRs found in PLHIV with atherosclerotic plaques and the plasma EVs-miRs from HIV neg individuals with atherosclerotic plaques could refine our understanding of which EV-miRs are specific for CVD development in PLHIV.
In summary, we have shown that PLHIV have an increased atherosclerosis burden and decreased ECFC levels, relative to HIV neg subjects. We have demonstrated that HIV posEVs recapitulate the effects of HIV infection on atherosclerosis with accompanied changes in lin − BMC senescence and functionality in apoE −/− mice. Importantly, we have identified candidate EV-miRs and the involvement of the let-7b-5p-Hmga2 axis in mediating the HIV effects. Experiments are ongoing in our laboratory to delineate the molecular pathways used by other candidate EV-miRs that underlie the HIV effects, with the goal of identifying therapeutic targets that may prevent the exacerbated atherosclerosis observed in PLHIV.

Materials and Methods
Laboratory health and safety procedures have been complied with throughout the experimental procedures as described.

Human Subjects
Human subjects who participated in this study were recruited following a protocol approved by the Institutional Review Board (IRB) of the University of Miami (IRB protocol #20140333). All participants gave informed consent prior to their inclusion in this study. PLHIV on ART and HIV neg individuals over the age of 18, from both sexes, were enrolled. Exclusion criteria were a history of CVD events (myocardial infarction and stroke), diabetes, malignancy, pregnancy, smoking (conditions shown to affect ECFC levels and functions), or individuals who lacked the capacity to consent. High-resolution B-mode carotid ultrasound was performed according to the standardized and validated scanning and reading protocols previously detailed [52]. Carotid intima-media thickness (cIMT) was measured per the consensus documents [53] using the automated computerized edge tracking software M'Ath (Intelligence in Medical Technologies, Inc., Paris, France) [54]. The cIMT protocols yield measurements of the distance between the lumen-intima and media-adventitia. Total cIMT is calculated as a composite measure of the means of the near and far wall IMT of all carotid sites (the common carotid artery, the internal carotid artery, and the bifurcation) from both sides of the neck. Atherosclerotic plaques were defined as the area of focal wall thickening that is 50% greater than the surrounding wall thickness. Peripheral blood was collected from each participant in K2E EDTA tubes by venipuncture. Plasma was isolated by centrifugation at 1000× g for 10 min at room temperature, collected, aliquoted, and stored at −80 • C until needed.

Isolation & Characterization of Plasma EVs
As defined by the International Society for Extracellular Vesicles (ISEV) [39], EVs, or "extracellular vesicles," are the lipid bilayer particles that are unable to replicate. EVs from human plasma were isolated by differential centrifugation, as previously described [55]. Briefly, 5 mL of plasma samples underwent a series of centrifugation (2000× g for 30 min followed by 12,000× g for 45 min) and ultra-centrifugation (110,000× g 2 h) steps at 4 • C. The pellet was resuspended in 15 mL of PBS, filtered through a 0.22 µm filter, and centrifuged at 110,000× g 70 min. The pellet was resuspended in 15 mL of PBS and centrifuged again at 110,000× g 70 min. Finally, the EV pellets were resuspended in 200 µL of PBS for downstream analyses. Protein concentration of EV samples was determined using the Pierce BCA Protein Assay Kit (ThermoFisher Scientific, Waltham, MA, USA). Nanoparticle tracking analyses (NTA) were used to examine the size distribution and particle number of EVs. For NTA, the isolated EVs were diluted 100 times and analyzed on the Nanosight LM10 system with Nanosight NTA 2.3 Analytical Software (Malvern Instruments Ltd., Malvern, UK). For this system, the range of diameter size detection limits is 10 nm~1000 nm. To evaluate the surface markers of EVs, EVs were incubated with anti-CD63 antibodycoated magnetic beads (Dynabeads) (ThermoFisher Scientific). The bead bound EVs were then incubated with rabbit anti-human antibodies against EV surface markers CD63, CD9, CD81, or Hsp70 (ExoAB Antibody Kit, System Biosciences, Palo Alto, CA, USA). The percentage of expression of each of these EV markers were assessed by flow cytometry (CytoFlex Flow Cytometer, Beckman Coulter, Brea, CA, USA). A sulfophosphovanilin colorimetric assay was used to assess the concentration of total lipids in EV preparations as previously described [56]. Briefly, 50 µL of EV suspension were mixed with 250 µL of 96% sulfuric acid in chloroform-pretreated tubes and incubated at 90 • C for 20 min with the tube lid open. Thereafter, 220 µL of samples were transferred into a 96-well polystyrene plate, cooled to room temperature, and incubated with 110 µL of 0.2 mg/mL vanillin in 17% phosphoric acid for 10 min at room temperature. Absorbance was measured at 540 nm using a spectrophotometric plate reader (Spectra Max M5, Molecular Devices, San Jose, CA, USA).

EVs with Modified miR Contents
To generate EVs with modified miR contents, including tailored extracellular vesicles (TEVs) loaded with antagomirs for candidate EV-miRs, we used genetically modified bone marrow mesenchymal stromal cells (MSCs)-following an approach that was previously described by our group [12]. Lentiviral plasmids expressing antagomirs or miRs that are under investigation were purchased from System Biosciences). Lentiviral particles were generated using HEK 293T cells co-transfected with the antagomir/miR-expressing plasmid, the packaging plasmid psPAX2 (Addgene #12259), and the envelope glycoprotein expressing plasmid, pCMV-VSV-G (Addgene #8454). The culture supernatant containing the viral particles was collected 48 h post transfection and the viral particles were concentrated using the Lenti-X™ Concentrator (Takara, Kusatsu, Shiga, Japan) and quantified using the Lenti-X™ qRT-PCR Titration Kit (Takara). MSCs isolated from 3-month-old C57BL/6 apoE −/− mice were transduced with each construct to make EVs with modified miR content. After 48 h of transduction, cells were cultured in the presence of 3 µg/mL puromycin for 48 h. Surviving cells were washed with PBS and cultured in serum-free media for 24 h. EVs with modified miR contents were isolated from serum-free media using differential centrifugation as previously described [55]. Briefly, 80 to 90 mL of cell culture media underwent a series of centrifugation (300× g 10 min followed by 2000× g 10 min and 12,000× g 30 min) and ultra-centrifugation (100,000× g 70 min) steps at 4 • C. The pellet was carefully resuspended in 15 mL of PBS and centrifuged again 100,000× g 70 min. The final pellet was resuspended in 200 µL of PBS for further analysis.

Animal Model
To determine the role of plasma EVs in mediating the effects of HIV infection on atherosclerosis, we used the atherosclerosis-prone apoE −/− mouse model. All procedures were performed in accordance with a protocol approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Miami (Protocol # 18-171-LF) and the Guide for the care and use of laboratory animals [57]. Animals were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). Four-week-old male apoE −/− mice were divided into four groups based on their weekly tail vein injection regimens: (1) circulating EVs from PLHIV (HIV posEVs ), (2) PLHIV plasma depleted of EVs (HIV PL depEVs ), (3) circulating EVs from HIV negative subjects (HIV negEVs ) and (4) PBS. All animals from all groups were fed a high fat/high cholesterol diet. Animals were assigned to groups using simple randomization. There were five animals in each group. Inclusion criteria were sex, age, and overall good health assessment. The exclusion criteria during the course of the experiments were the presence of animal stress (behaviors such as hiding, pressing the head against cage walls), weight loss, or decreased food or water intake. No animals had to be excluded from the study. Mice were injected weekly for 12 weeks (12 injections total). The total protein amount of 500 µg was used for each injection in a total volume of 100 µL to maintain uniformity across all treatments. One week after the final injection, animals were sacrificed, and the aortas and bone marrow were collected. Euthanasia was conducted according to the AVMA Guidelines for the Euthanasia of Animals, with CO 2 exposure as the primary method, followed by quick cervical dislocation done by trained personnel as a secondary method when needed.

Atherosclerosis Measurement
The aorta was cut into thoracic and abdominal segments and embedded in Tissue-Tek OCT media. Multiple 8-µm sections were made for histological and immunofluorescent staining. Oil Red O (ORO), hematoxylin and eosin (H&E), and TUNEL staining were performed in alternating frozen sections, such that a survey of a substantial portion of each of the proximal thoracic and abdominal segments was obtained. In an effort to capture the aspect of the extension of the aortic lesions, more than one tissue section per animal was included in the final analysis. For H&E staining, aortic frozen sections were fixed for one minute with a 95% ethanol solution and stained with 0.1% Mayers hematoxylin for 1 min, then washed with water for 2 min. Next, slides were dipped in 0.5% Eosin solution for 30 s, washed in 95% ethanol solution and in absolute ethanol, followed by two xylene washes. For ORO staining, slides were fixed with 10% formalin for 10 min, rinsed with PBS for 1 min and then 60% Isopropanol for 15 s. Next, sections were stained with ORO Working Solution (Sigma Aldrich, St. Louis, MO, USA) for 1 min and rinsed in 60% isopropanol for 15 s, followed by 3 washes in PBS. Counterstaining for the nuclei was performed by mounting the sections with Prolong Diamond Antifade Mountant (DAPI, ThermoFisher Scientific). Tissue sections were viewed and photographed on a Zeiss Observer Z1 microscope. Atherosclerotic burden was calculated as the ratio of lesion area/total lumen area (ImageJ software, 1.46r, NIH). At least 5 ORO-staining histological sections for each segment were used to quantify the atherosclerosis burden.

Quantification of Endothelial Colony-Forming Cells (ECFCs) from Humans
Circulating endothelial colony-forming cells, or ECFCs [7], formerly known as late endothelial progenitor cells, were quantified by colony-forming units (CFUs) using an established protocol [58]. Briefly, human peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll-Paque reagent (Amersham Biosciences, Amersham, UK) from whole blood. Cells were washed twice in PBS with 2% FBS and plated on a 24-well plate coated with fibronectin at a density of 10 4 cells/well in 500 µL endothelial basal medium-2 (EBM-2, Lonza, Basel, Switzerland) supplemented with EGM-2MV single aliquots and 10% FBS. In the first two days, the medium was changed daily to remove nonadherent cells and debris, and then every other day. After one week of cultivation, the cells were dissociated with trypsin and reseeded onto fibronectin-coated plates for further cultivation. After 2-3 weeks of culturing, colonies were stained with 6.0% glutaraldehyde and 0.5% crystal violet solution (Sigma-Aldrich) and counted using a stereomicroscope (OM2300S-GX4 3.5X-45X Zoom Stereo Boom Microscope, Omano, China).

Lipid Panel and C Reactive Protein ELISA Assays
Biochemical analyses were performed using a Vitros 5600 analyzer (Ortho Clinical Diagnostics, Raritan, NJ, USA). Total cholesterol, triglycerides, and HDL-C were directly determined, and LDL-C was calculated using the Friedewald equation [59] An enzymelinked immunosorbent assay (ELISA) was used to quantify C-reactive protein in mouse serum according to the manufacturer's protocol (Abcam, Cambridge, UK).

Cell Migration "Scratch" Assay
To assess cell migration capacity, lin − BMCs were seeded in 24-well plates (1 × 10 5 cells per well). EV treatment was performed with a concentration of 500 µg of total EV protein per 1 mL of culture medium. Cell monolayers were scratched using a sterile pipette tip. The scratch areas were photographed and measured at 0 h, 24 h and 48 h post scratching using the EVOS FL Auto Fluorescence Inverted Microscope Imaging System (ThermoFisher Scientific). The area of scratches was analyzed with ImageJ software (NIH).

Apoptosis Assay
Apoptosis in the aortic wall was detected and quantified using the TUNEL assay. TUNEL reactions were performed with the In Situ Cell Detection Kit, TMR red (Roche, Basel, Switzerland), according to the manufacturer's protocol. In an effort to capture the true extent of the aortic lesions, the tissue was cut into multiple sections, with alternating sections being used for Oil Red O (ORO), hematoxylin and eosin (H&E), and TUNEL staining. Tissue from multiple mice was used and treated in the same manner.

Senescence Assay
Cellular senescence was evaluated by β-galactosidase assays. For cultured cells, cells were stained with the β-galactosidase staining kit (Cell Signaling Technology, Danvers, MA, USA) according to the manufacturer's protocol. For frozen tissue sections, the Senescence Detection Kit (Abcam, Cambridge, MA, USA) was used according to the manufacturer's protocol. Images were acquired in 10 random microscopic fields per sample at 10× mag-nifications using the EVOS FL Auto Fluorescence Inverted Microscope Imaging System (ThermoFisher Scientific).

Cell Proliferation Assay
Cell proliferation was evaluated using the MTT (3-[4, 5-dimethylthiazol-2-yl]-2, 5diphenyl tetrazolium bromide) assay. Briefly, lin − BMCs or ECs were seeded in 96-well plates (5 × 10 3 cells per well) and allowed to grow for 48 h in EBM TM -2 Endothelial Cell Growth Basal Medium-2 (Lonza) complete media. Fresh media containing 1 mg/mL of MTT (Millipore Sigma) was added to the culture wells, according to the manufacturer's instructions. To quantify cell proliferation, MTT formazan crystals were eluted with isopropanol, and the absorbance of samples was assessed at 570 nm, using 690 nm as a reference wavelength, with a spectrophotometric ELISA plate reader (Spectra Max M5, Molecular Devices).

Small RNA Sequencing (sRNA-seq)
To identify EV-miRs that are associated with elevated atherosclerosis burden in PLHIV, we selected PLHIV with the largest plaque areas (N = 5) and HIV neg subjects with no plaques (N = 5), representing the extremes of the disease spectrum. All individuals were male, non-smokers, and had an average age of 47 and 46 years, for the PLHIV and HIV neg group, respectively. sRNA-seq of EVs from these samples was performed by System Biociences (SBI, Palo Alto, CA, USA). Briefly, EVs were isolated using the ExoQuick-ULTRA precipitation method, and a sRNA-seq library was prepared using the XRNA Exosome RNA-Seq Library Kit [60,61]. Sequencing was performed on the Illumina NextSeq with 1 × 75 bp single-end reads. On average, 5-10 million reads were generated for each library. After quality control (QC) using FastQC and end trimming, about 5 million reads per sample were mapped to the reference human genome using Bowtie [62]. SAMtools and Picard [63,64] were used for differential expression analyses, including read coverage, determination of small RNA abundance, and differential expression analysis across samples. Expression analyses were conducted based on normalized counts [counts per million (CPM)] of mature miR sequences. MiRs with a two or higher fold increase in the HIV posEV group compared to controls were considered for further analysis. Of the upregulated miR sequences, the top 5 miRs with the highest expression level were prioritized for further validation by qRT-PCR.

qRT-PCR Validation of Candidate EV-miRs
Small RNAs were extracted from EVs using the miRNeasy Mini Kit (Qiagen, Hilden, Germany). Reverse transcription and qPCR were performed using the specific miRNA primer and probe sets provided by the Taqman™ MicroRNA Assay kits (Thermo Fisher Scientific). The U6 small nucleolar RNA was used as an internal control [45,[65][66][67]. Taq-Man™ Gene Expression Assays were performed on the ABI PRISM 7300 system (Applied Biosystems, Waltham, MA, USA), and the relative expression was calculated using the 2 −∆∆Ct method. All values were normalized to U6 RNA levels.

Ingenuity Pathway Analyses (IPA)
IPA's Target Analysis Tool was utilized to identify gene targets of candidate EV-miRs. IPA's "Disease and Functions" tool was used to identify the diseases or cellular functions that were significantly enriched with the target genes of the miRs of interest. A FDR p-value < 0.01 was considered significant for the analyses. The p-value for each "Disease/Function category" is provided as a range composed of multiple p-values for each of the subcategories within the larger disease or function category (e.g., the cardiac enlargement category is made up of the categories of enlargement, ventricular hypertrophy, cardiomegaly, etc., each with its own p-value).

Transduction of Mouse lin − BMCs
To dissect the molecular mechanisms underlying the EV-miR effects, we focused on High Mobility Group AT-Hook 2 (Hmga2), a known target of let-7b-5p and a gene shown to regulate the senescence of neural stem cells and lin − BMCs [28,30]. To that end, we conducted a series of transduction experiments in lin − BMCs isolated from young wild type mice overexpressing either wild-type Hmga2 (wtHmga2) or Hmga2 lacking the 3 UTR (Hmga2-3 UTRdel, therefore resistant to miR regulation). Lentiviral transduction and cell selection were performed as described above.

In Vitro EV Treatment
Cells were seeded in 24-well plates (1 × 10 5 cells per well). EBM™-2 Endothelial Basal Medium-2 (Lonza) supplemented with EGM™-2 MV Micro-vascular Endothelial SingleQuotsTM Kit (Lonza) was mixed with EVs at a concentration of 500 µg protein per mL of culture media for 5 days (media with EVs was changed every 2-3 days).

Statistical Analyses
Kolmogorov-Smirnov and Shapiro-Wilks normality tests were used to assess normal distribution. The Kruskal-Wallis test was used to compare differences between treatment groups and the control group for non-parametric variables, followed by the Dunn's Multiple Comparison test. For continuous variables, Student's t test (t test) and analysis of variance (ANOVA) were used to compare differences between two groups and among ≥3 different groups, respectively. For the animal model, a total sample of 20 subjects achieves 80% power to detect differences with a 0.05 significance level and a relatively large effect size of 0.84. Although we have a relatively small sample, based on our preliminary study with an effect size above 2.5, we have enough power to detect the difference across 4 groups. Statistical analysis was performed using GraphPad Prism 7 (GraphPad, San Diego, CA, USA). p < 0.05 was considered statistically significant. For sRNA-seq, expression statistics were calculated and visualized using R 3.5.2.
Conflicts of Interest: Joshua M. Hare reports having a patent for cardiac cell-based therapy and holds equity in Vestion Inc. and maintains a professional relationship with Vestion Inc. as a consultant and member of the Board of Directors and Scientific Advisory Board. Vestion Inc. did not play a role in the design, conduct, or funding of the study. Dr. Joshua Hare is the Chief Scientific Officer, a compensated consultant and board member for Longeveron Inc. and holds equity in Longeveron. Dr. Hare is also the co-inventor of intellectual property licensed to Longeveron. Longeveron did not play a role in the design, conduct, or funding of the study. The University of Miami is an equity owner in Longeveron Inc., which has licensed intellectual property from the University of Miami.