Abstract
Objectives We aim to investigate whether circulating microparticle (CMPs) subsets were raised amongst people living with HIV in Malawi, and whether they associated with arterial stiffness.
Methods ART-naïve adults with a new HIV diagnosis and CD4 <100 cells/µL had microparticle characterisation and carotid femoral Pulse Wave Velocity (cfPWV) at 2 weeks post ART initiation. HIV uninfected controls were matched on age, systolic BP and diastolic BP in a 1:1 ratio. Circulating microparticles were identified from platelet poor plasma and stained for endothelial, leucocyte, monocyte and platelet markers.
Results The median (IQ) total CMP count was 12-fold higher in participants with HIV compared to those without (p<0.0001) and was associated with arterial stiffness (spearman rho 0.42, p<0.001). In adjusted analysis, every log increase in circulating particles showed a 20% increase in cfPWV (95% CI 4 – 40%, p=0.02).
In terms of subsets, endothelial and platelet derived microparticles were most strongly associated with HIV. Endothelial derived E-selectin+ CMPs were 22.5-fold higher and platelet derived CD42a+ CMPs were 12-fold higher (both p<0.0001). Endothelial and platelet derived CMPs also correlated most closely with arterial stiffness [spearman rho: E-selectin+ 0.57 and CD42a 0.56, both p<0.0001).
Conclusions Circulating microparticles associate strongly with HIV and arterial stiffness amongst adults in Malawi. Endothelial damage and platelet activation are of particular importance and future translational studies should consider prioritising this pathway.
Introduction
People living with HIV (PLWH) are at increased risk of cardiovascular diseases (1-4). Chronic immune activation has been implicated in vascular dysfunction but the underlying pathogenetic mechanisms have not been fully elucidated (2). Multiple pathogenic drivers such as microbial translocation, latent co-infections and low level HIV viraemia likely co-exist, making chronic inflammation in PLWH a heterogenous process (5). These mechanisms are even more complex in the low-income sub-Saharan Africa setting (SSA), where other infectious drivers of chronic inflammation, such as tuberculosis, CMV and recurrent acute infections, are prevalent (6).
During the inflammatory response leucocytes adhere to the endothelial cells via VCAM, ICAM, and undergo transcytosis leading to the development of foam cells (7). Pro-thrombotic pathways attract platelets to the resulting plaque and subsequent migration of smooth muscle cells into the intima leads to the formation of a fibrous cap (8). During this process, elastin is degraded by enzymes such as MMPs and is replaced by collagen, increasing arterial stiffness (8).
Arterial stiffness has been used as a measure of vascular dysfunction in HIV studies as well as other chronic inflammatory conditions (9, 10). We have previously shown that arterial stiffness, as measured by carotid femoral pulse wave velocity (cfPWV), is increased in PLWH during the first 3 months of ART and in people experiencing unstructured treatment interruption and that first line ART failure is common (6, 11, 12). cfPWV is a gold standard measurement of arterial stiffness and has been validated as a physiological biomarker of cardiovascular events and mortality (13, 14).
Circulating Microparticles (CMPs) are released into the circulation following activation or apoptosis of the affected cells (15). Through a process of blebbing, microparticles are formed from the originating host cell’s outer membrane (e.g. leucocytes, platelets and endothelial cells). During this process annexin V molecules, which are normally located on the inner membrane of a cell, are flipped round to become expressed on the microparticle surface (16). These microparticles also express the markers expressed on the surface of the cell of origin and so microparticle subsets are an indication of which cells are undergoing stress. An established body of evidence supports CMPs as biomarkers for atherosclerosis and endothelial damage amongst people with cardiovascular disease (17). In the Framingham Offspring cohort each tertile increase in Framingham risk score corresponded to a 9% log increase in endothelial MPs (18). The European Society of Cardiology released a position paper in 2018 calling for standardisation of microparticle identification techniques and additional cohort studies to help identify biomarker and therapeutic potential for microparticles in cardiovascular disease (19). Previous studies in HIV have assessed the presence and function of CMPs amongst PLWH in high resource settings and have separately identified tissue factor expression, imbalanced endothelial progenitor cell proportions and platelet activation (20-23). However, few studies have simultaneously characterised different circulating microparticle subsets amongst PLWH. Hijmans et al recently demonstrated evidence of in vitro endothelial cell stress, apoptosis and senescence induced by leucocyte, platelet and endothelial microparticle subsets from patients with HIV(24). Here, we aim to characterise circulating microparticle subsets amongst PLWH in a low-income sub-Saharan Africa setting, and examine their relationship with arterial stiffness, a physiological biomarker of cardiovascular risk.
Methods
Study cohort
The Study into HIV, Immune activation and EndotheLial Dysfunction (SHIELD) cohort recruited ART-naïve adults with a new HIV diagnosis and CD4 <100 cells/µL from the ART clinic and voluntary HIV testing clinic at Queen Elizabeth Central Hospital, Blantyre, Malawi, along with adults without HIV infection and without evidence of infection within the previous 3 months (full details of recruitment in (6)). In brief, participants underwent a detailed clinical assessment, blood draw for plasma storage and cfPWV at 2 weeks post ART initiation. All participants provided informed written consent and ethical approval was granted by the College of Medicine Research and Ethics Committee (COMREC), University of Malawi (P.09/13/1464) and the University of Liverpool Research and Ethics Committee (UoL000996).
SHIELD participants with and without HIV infection were ordered according to cfPWV values from lowest to highest. A convenience sample of 36 PLWH and 36 HIV matched controls without HIV infection were chosen from across the spectrum of cfPWV values. Participants with a cfPWV in the highest quartile (>9 m/s) were chosen randomly in a 2:1 ratio to patients with a cfPWV below the highest quartile from the cohort of patients with HIV. This was to enrich the number of potential CMPs for subset analysis (we hypothesised they would be higher at higher ranges of cfPWV) whilst also capturing a range of values to analyse associations between total CMP count and cfPWV. Participants without HIV were then matched to participants with HIV infection on age, systolic BP and diastolic BP in a 1:1 ratio.
Statistical analysis
Wilcoxon Ranksum and spearman rho were used to test association between CMPs and categorical or continuous variables respectively. To allow for multiple comparisons characterising 18 types of CMPs, a p value of less than 0.003 was used as significant. Linear regression was used to examine the effect of total CMPs on cfPWV which was log transformed for normality. The model was adjusted for confounders (age and sex) and mediators (blood pressure, haemoglobin and HIV) as identified previously (6).
Laboratory procedures
Platelet poor plasma (PPP) was isolated from thawed plasma and stained for Annexin V (supplementary material). Endothelial, leucocyte, monocyte and platelet microparticles were identified by flow cytometry on a CyAn ADP 9 colour flow cytometer (Beckman Coulter). FITC stained for Annexin V (BD Pharmingen), PE stained for VCAM, ICAM, E-selectin, 66b or CD16 (BD Pharmingen) and APC-Cy7 stained for PECAM or CD14 (BD Pharmingen). To identify tissue factor expression on monocytes, PE stained Annexin V, APC-Cy7 stained CD14 (BD Pharmingen) and FITC stained tissue factor (Sekisui Diagnostics). Single stain samples were also acquired for the purposes of compensation and isotype controls were analysed for gating. After identification of singlets, the microparticle pellet was gated on forward scatter and AnV (FITC) to identify the microparticle population which was less than 3µm in size and expressing AnV (see Figure 1Error! Reference source not found.). Gates were applied using thresholds provided by isotype controls.
Cell surface immune phenotyping was carried out for CD4 and CD8 T cell activation (HLA-DR+/CD38), exhaustion (PD1+) and senescence (CD57+), as well as monocyte subsets (classical CD14 CD16, intermediate CD14 ++ CD16 and nonclassical CD14 ++ CD16). The T-cell panel consisted of CD3 BV510, CD4 V450, CD38 PE Cy7, HLA-DR AF700, PD1 APC, and CD57 FITC (all from BD Biosciences) and CD8 PE (Biolegend). The monocyte panel consisted of HLA-DR AF700, CD14 PE Cy7 and CD16 PE (all from BD Biosciences).
A. 3µm beads were used to find forward scatter gate for microparticle identification B. Microparticle pellet was obtained from platelet poor plasma and stained for Annexin V. Total circulating microparticle count was measured by identifying microparticles Annexin V positive and less than 3µm forward scatter as shown. C. Total circulating microparticles in a patient with a high count compared to B. D. Isotope controls identified gates for PE and APC-Cy7 stains for each subset panel. E. Leucocyte microparticles F. Endothelial or platelet microparticles (see table 4).
Results
CMP data were available for 36 PLWH and 33 HIV uninfected participants. Sufficient plasma samples were not available for three of the selected HIV uninfected participants. For the 69 participants with available CMP data, median (IQR) age was 41 years (36 – 49) and 19 (40%) were female. 26 (54%) had an education level of at least primary school completion. Median (IQR) blood pressure was 132/78 (116/74 – 136/85) and eight (17%) had a history of smoking. Characteristics according to HIV status are shown in Table 1. For 36 PLWH, median (IQR) CD4 and HIV viral load were 35 cells/µL (28 – 52) and 1.1 × 105 copies/mL (0.6 – 2.5).
Relationship between absolute CMP counts, HIV and arterial stiffness
Total CMP counts were significantly associated with arterial stiffness (spearman rho 0.42, p<0.001; figure 2), as well as faster heart rate and higher creatinine (spearman rho: 0.30; p=0.01 and 0.31; p=0.01). The median (IQ) total CMP count was 12-fold higher in participants with HIV compared to those without [5100 × 106 particles/mL (2000 × 106 – 18000 × 106) versus 410 × 106 particles/mL (200 × 106 – 6000 × 106); p<0.0001]. When adjusted for a priori identified mediators and confounders (age,sex, haemoglobin and blood pressure), higher log total CMPs predicted increased cfPWV [fold change 1.20m/s, 95% CI 1.04 – 1.40; p=0.02]. When additionally adjusted for HIV status, the effect of log total circulating microparticle count on cfPWV was attenuated [fold change 1.07m/s (95% CI 1.23– 1.01), p=0.046. Table 2].
Total CMP counts also correlated directly with proportions of CD4 and CD8 T cells expressing markers of activation, exhaustion and senescence [spearman rho (p value): CD4 activated 0.44 (0.071), exhausted 0.27 (0.0007), senescent 0.44 (0.01); CD8 activated 0.49 (0.0004), exhausted 0.39 (0.041), senescent 0.44 (0.0006)]. However, there was no association between total CMPs and monocyte subsets [spearman rho (p value): classical 0.08 (0.57), intermediate −0.018 (0.90), nonclassical −0.07 (0.63)].
Relationship between CMP subsets, HIV and arterial stiffness
The largest elevations in CMPs were seen with endothelial and platelet derived microparticles. Endothelial derived E-selectin+ CMPs were 22.5-fold higher [median (IQR) 4.5 × 106 particles/mL (2.6 – 16.0) vs 0.20 × 106 (0.073 – 0.38); p<0.0001]. The increase in CMPs expressing PECAM but not E-selectin was less significant (Figure 3). Platelet derived CD42a+ CMPs were 12-fold higher amongst PLWH [median (IQR) CD42a 3.5 × 106 particles/mL (0.88 – 12.7) vs 0.3 × 106 (0.11 – 1.0); p>0.0001]. Amongst leucocyte derived CMPs, those positive for CD66b were 2.6-fold higher and those positive for CD16 but CD14 negative were elevated 1.6 fold [median (IQR) CD66b: 0.13 × 10 particles/mL (0.70 – 3.4) versus 0.05 (0.01 – 0.1), p<0.0001. CD16+CD14-: 0.48 × 106 (0.8 – 1.1) vs 0.3 × 106 (0.11 – 1.0), p<0.0001]. For CD14+ CMPs, only those expressing tissue factor were significantly higher [2.2-fold; median (IQR) 0.2 × 106 particles/mL (0.04 – 0.80) vs 0.09 × 10 (0.003 – 0.28), p<0.0001].
PLWH=People living with HIV. No HIV= People living without HIV. Red lines represent median microparticles count. Grey Asterix represent p value: ns denotes p>0.003, * denotes p<0.003 but >0.0001, **denotes p>0.0001, ***denotes p<0.0001. TF=Tissue Factor. CD=Cluster of Differentiation. VCAM=vascular cell adhesion molecule 1. ICAM=intracellular cell adhesion molecule 1. PECAM=platelet endothelial cell adhesion molecule.
CMPs of endothelial and platelet origin associated closely with cfPWV [spearman rho: E-selectin+ 0.57, p<0.0001 and CD42a 0.56, p<0.0001; Table 3). CD16+CD14-derived CMPs also correlated significantly with cfPWV (spearman rho 0.69, p<0.0001). For the 36 people in the HIV group, the proportion of activated CD8 T cells correlated indirectly with CD14+ CD16+ microparticles (rho −0.9, p=0.037).
Discussion
Here we show that total CMPs correlate closely with arterial stiffness and are markedly increased amongst people who present with HIV and advanced immune suppression in Malawi. Microparticles originating from endothelium and platelets were especially high amongst people with HIV and correlated strongly with arterial stiffness; suggesting that these pathways might be prioritised for future studies seeking to reduce inflammation driven cardiovascular disease amongst PLWH.
People living with HIV demonstrated 20 times more CMPs than people without HIV, and each log increase in CMPs demonstrated a 20% higher cfPWV. In adjusted models, the association between CMPs and arterial stiffness was attenuated by HIV, suggesting that CMPs may have a role in mediating the effect of HIV on endothelial damage. A relationship between total CMPs and activated T cells was also demonstrated, in keeping with previous data suggesting a role of activated and exhausted T cells in increased arterial stiffness amongst PLWH (6).
CMPs are induced in response to cellular stress in the form of activation, apoptosis or physical sheer stress (25). A similar case control study of 15 PLWH established on ART in the USA, showed significantly elevated levels of all types of microparticles compared to 15 participants without HIV infection(24), where microparticles from PLWH directly impaired endothelial cell function. A larger study of PLWH from the Czech Republic also found an increase in microparticles compared to participants without HIV, but found no difference amongst those on ART compared to those who were at their first presentation (26). However, none of these studies assessed the relationship between CMPs and arterial stiffness. By presenting associations with a validated physiological biomarker, we provide evidence for the use of CMPs as a clinically relevant tool to support characterisation of inflammation amongst PLWH.
As well as acting as biomarkers of inflammation, microparticles have been implicated in mediating inflammation induced pathogenesis. They may interact with endothelial cell surface adhesion molecules and cause endothelial damage through the production of nitrous oxide and pro-inflammatory cytokines (27). CMPs have been shown to transfer important material such as miRNA and lipids to other cells (28). In particular, the transfer of the CCR5 receptor to endothelial cells from leucocyte derived microparticles potentially renders them permissible to direct HIV infection (29). Circulating microparticles produced in response to HIV could therefore lead to endothelial damage, and therefore increased risk of cardiovascular disease, through multiple mechanisms including direct endothelial adhesion and activation, transfer of cytotoxic viral proteins, and propagation of low level viral replication.
We discovered that microparticles of endothelial and platelet origin are particularly expanded amongst PLWH in this low income SSA setting. It is plausible to hypothesis that HIV induced cytotoxic CD8 T cells may directly interact with platelets leading to cellular apoptosis and formation of microparticles that may then go on to perpetuate the cycle of endothelial damage. This potential mechanism requires further investigation. We show that E-selectin expression correlated particularly closely with both HIV status and arterial stiffness. Defining endothelial microparticles as CD51+, da Silva et al previously reported endothelial microparticles 20 times higher in PLWH compared to people without HIV (20). Endothelial microparticles promote expression of adhesion molecules on the endothelial cell surface, generating thrombosis, platelet activation and recruitment of inflammatory cells (23). Platelet driven thrombosis in combination with endothelial activation may take advantage of a compromised endothelial barrier and lead to decreased elasticity and smooth muscle reactivity within the arterial wall (30). Inflammatory cytokine release in response to ongoing cellular recruitment, as well as from direct effects of the HIV virus, can also activate and increase the proliferation and migration of vascular smooth muscle cells, and thus induce arterial stiffness (31). Wheway et al have previously found endothelial microparticles to form conjugates with T cell subsets, with increased binding to those cells that were pre-stimulated (32). Binding was VCAM and ICAM dependent and endothelial microparticles were able to stimulate in vitro proliferation of T cells.
Although this study examines a small convenience sample, we highlight demonstrate clinical relevant of quantification and characterisation of circulating microparticles and also confirm that this applies in the low income sub-Saharan Africa setting. We also elucidate the platelet – endothelial axis as an interesting pathway worthy of further investigation. In addition, our cohort of people living with HIV had experienced advanced immune suppression and so factors other than the direct effect of HIV itself (e.g. CMV, TB, cryptococcal disease) may be contributing. Further, findings are not generalisable to cohorts with more robust CD4 counts and therefore this needs to be established.
Overall, the characterisation of microparticles in this study lends weight to a model where active and significant immune activation amongst people living with HIV is strongly related to endothelial damage and may involve both CD8 T cell and platelet activation. Further research should investigate whether CMPs might represent translational targets to reduce inflammation driven cardiovascular risk amongst people living with HIV.
Data Availability
All relevant data is included in the published manuscript or supporting material