Proteomic Signature of HIV-Associated Subclinical Left Atrial Remodeling and Incident Heart Failure

word count: 350 Manuscript text word count: 4945 Tables/Figures: 3 Tables, 5 Figures + 17 Supplemental Tables, 7 Supplemental Figures

50 Corresponding Author: Tess E Peterson, PhD MPH [email protected] T32 Postdoctoral Research Fellow Division of Cardiology Department of Medicine Johns Hopkins University (612) 807-6786 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 3

Background: People living with HIV (PLWH) are at higher risk of heart failure (HF) and preceding subclinical cardiac abnormalities, including left atrial dilation, compared to people without HIV (PWOH). Hypothesized mechanisms include premature aging linked to chronic immune activation. We leveraged plasma proteomics to identify potential novel contributors to HIV-associated differences in indexed left atrial volume (LAVi) among PLWH and PWOH and externally validated identified proteomic signatures with incident HF among a cohort of older PWOH.

We performed proteomics (Olink Explore 3072) on plasma obtained concurrently with cardiac magnetic resonance imaging among PLWH and PWOH in the United States. Proteins were analyzed individually and as agnostically defined clusters. Cross-sectional associations with HIV and LAVi were estimated using multivariable regression with robust variance. Among an independent general population cohort, we estimated associations between identified signatures and LAVi using linear regression and incident HF using Cox regression.

Among 352 participants (age 55±6 years; 25% female), 61% were PLWH (88% on ART; 73% with undetectable HIV RNA) and mean LAVi was 29±9 mL/m2. Of 2594 analyzed proteins, 439 were associated with HIV serostatus, independent of demographics, hepatitis C virus infection, renal function, and substance use (FDR<0.05). We identified 73 of these proteins as candidate contributors to the independent association between positive HIV serostatus and higher LAVi, enriched in tumor necrosis factor (TNF) signaling and immune checkpoint proteins regulating T cell, B cell, and NK cell activation. We identified one protein cluster associated with LAVi and HIV regardless of HIV viral suppression status, which comprised 42 proteins enriched in TNF signaling, ephrin signaling, and extracellular matrix (ECM) organization. This protein cluster and 30 of 73 individual proteins were associated with incident HF among 2273 older PWOH (age 68±9 years; 52% female; 8.5±1.4 years of follow- up).

Proteomic signatures that may contribute to HIV-associated LA remodeling were enriched in immune checkpoint proteins, cytokine signaling, and ECM organization. These signatures were also associated with incident HF among older PWOH, suggesting specific markers of chronic immune activation, systemic inflammation, and fibrosis may identify shared pathways in HIV and aging that contribute to risk of HF. Key Words: proteomics; HIV; left atrial size; cardiovascular magnetic resonance imaging; biomarkers, heart failure, aging, inflammation All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 4

1 With improved access to early and effective antiretroviral therapy (ART), HIV has become a 2 chronic condition marked by higher risk and earlier onset of aging-related diseases, including 3 cardiovascular disease (CVD).1 People living with HIV (PLWH), even when virologically 4 suppressed due to ART, are at higher risk of myocardial diseases,2 such as incident heart 5 failure (HF)3-5 and atrial fibrillation (AF).6 Advanced HIV disease has been classically associated 6 with left ventricular (LV) systolic dysfunction and dilated cardiomyopathy, but with effective ART, 7 this phenotype has transitioned to one characterized by subclinical abnormalities in cardiac 8 structure and function that may presage clinical HF, particularly HF with preserved ejection 9 fraction (EF).7 We previously demonstrated PLWH have higher risk of diastolic dysfunction and 10 related structural abnormalities compared to people without HIV (PWOH), including subclinical 11 left atrial (LA) enlargement,8,9 which has been supported by other data across world regions.10-13 12 Indexed left atrial volume (LAVi) reflects the cumulative effect of chronically elevated LV filling 13 pressure and is an independent predictor of diastolic dysfunction and clinical HF.14 14 Mechanisms contributing to HIV-associated myocardial disease are hypothesized to be 15 multifactorial, mediated in large part by chronic immune activation and dysfunction that persists 16 despite ART-induced viral suppression. Hypothesized contributors include direct immune 17 activation and exhaustion due to HIV viral persistence and co-pathogens, metabolic dysfunction 18 due to both HIV and use of certain antiretrovirals, higher mucosal permeability and microbial 19 translocation, and higher prevalence of risk factors with known inflammatory effects, such as 20 substance use.7 However, it remains unclear to what degree mechanisms underlying this risk 21 differ among PLWH compared to PWOH, among whom aging-related immunosenescence and 22 inflammaging may similarly contribute to clinical HF risk. Moreover, therapeutic and preventive 23 strategies among PLWH, particularly for HF, are largely guided by data from clinical trials 24 conducted among PWOH. 25 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 5 Proteomics provides a high-throughput discovery approach to generate novel mechanistic 26 hypotheses that could lead to improved risk prediction, disease characterization, and novel 27 targeted pharmacologic therapies. We performed proteomic analyses using the Olink Explore 28 3072 platform on stored plasma obtained concurrently with cardiovascular magnetic resonance 29 (CMR) among largely ART-treated PLWH and sociodemographically similar PWOH enrolled in 30 the Subclinical Myocardial Abnormalities in HIV (SMASH) study,9 the discovery cohort. Our 31

was to identify plasma proteomic signatures that may represent novel contributors to 32 HIV-associated atrial remodeling. We then sought to evaluate if identified proteomic signatures 33 were associated with LAVi, incident AF, and incident adjudicated clinical HF among a large 34 independent external validation cohort of older PWOH in the Multi-Ethnic Study of 35 Atherosclerosis (MESA). Such a validation approach could provide support for a premature 36 aging hypothesis among PLWH and identify common and potentially novel mechanistic 37 pathways that contribute to higher risk of clinical HF among older PWOH. A full study overview 38 is depicted in FIGURE 1. 39

40 Discovery Study Population (SMASH) 41 Participants in the SMASH study were recruited from the Multicenter AIDS Cohort Study 42 (MACS), the Women’s Interagency HIV Study (WIHS), and the AIDS Linked to the IntraVenous 43 Experience (ALIVE) study and underwent CMR with concurrent blood collection and storage 44 between March 2015 and February 2018. All three cohorts include PLWH and PWOH. The 45 WIHS and MACS are multicenter observational longitudinal cohort studies composed of women 46 or men who have sex with men, respectively.15,16 The ALIVE study is a community-based cohort 47 of men and women with a history of injection drug use.17 See the Supplemental Methods for 48 additional cohort details. The SMASH study enrolled active participants 40-70 years of age from 49 the Chicago and Baltimore/Washington, DC MACS sites; the Washington, DC WIHS site; and 50 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 6 the ALIVE study, which is based in Baltimore. Exclusion criteria included estimated glomerular 51 filtration rate (eGFR, Chronic Kidney Disease Epidemiology Collaboration 2021) 52 350 pounds, known claustrophobia or contrast allergy, and 53 contraindications to CMR. Proteomics was performed on a subset of total SMASH participants 54 with complete CMR and adequate stored plasma volume (Supplemental Figure S1A). The 55 Institutional Review Boards of Johns Hopkins University, Georgetown University, and 56 Northwestern University approved this study, and all participants signed informed consent. 57 Given the sensitive nature of data collected for this study, requests for data sharing from 58 researchers certified in human subject confidentiality can be submitted to and will be reviewed 59 by the parent cohorts: https://statepi.jhsph.edu/mwccs/ and 60 https://www.jhsph.edu/research/affiliated-programs/aids-linked-to-the-intravenous-experience/. 61 Sociobehavioral and Clinical Data in SMASH 62 Participants completed an interviewer-administered structured questionnaire and biological 63 measures concurrent with CMR. Prescribed medications, history of CVD, and substance use 64 during the preceding 5 years were queried. These data were supplemented with data collected 65 through the parent cohorts, including demographics and CVD risk factors. See Supplemental 66

for a full list and definitions of covariates. Among PLWH, measures of HIV disease 67 activity included current plasma HIV RNA concentrations (Roche ultrasensitive assay), current 68 and nadir CD4+ T cell counts/µL, history of AIDS-defining malignancy or opportunistic infection, 69 and use of ART, including regimen. 70 Cardiovascular Magnetic Resonance in SMASH 71 CMR was performed at Johns Hopkins Hospital (Baltimore) or Northwestern Memorial Hospital 72 (Chicago) on 1.5 T Siemens Avanto or Aera scanners (Erlangen, Germany) using a 73 standardized protocol that included gadolinium enhancement, detailed previously.9 Briefly, short 74 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 7 and long-axis cines (30 phases/cardiac cycle) were acquired with a steady-state free precession 75 sequence. Multimodality Tissue Tracking software (MTT; version 6.0, Toshiba, Japan) was used 76 to quantify LA volumes and other standard metrics,18 blinded to participant characteristics. 77 Maximum LA volume was measured using the LA volume curve generated by the Simpson’s 78

from four-chamber and two-chamber views and was indexed to body surface area 79 (LAVi). 80 Proteomics in SMASH 81 We performed proteomics on EDTA plasma stored at -80°C using the Olink Explore 3072 82 platform at the Laboratory of Clinical Investigation at the National Institute on Aging, NIH 83 (Baltimore, MD) using a standardized protocol. Olink technology has been described 84 previously19 and is outlined in the Supplemental Methods alongside details on data quality 85 control, normalization, and calibration. Values are reported as log2-transformed relative plasma 86 abundances, which were winsorized to 5 standard deviations (SD) to minimize the effect of 87 outliers and standardized for comparability. Following quality control, no samples were 88 excluded, and 2594 proteins were analyzed with final mean intra- and inter-assay coefficients of 89 variation of 9% and 15%, respectively. 90 Weighted Gene Co-expression Network Analysis 91 High-dimensional proteomic data often contain informative patterns lost when analyzing proteins 92 individually. To address this, we derived unique clusters of highly correlated proteins using 93 weighted gene co-expression network analysis (WGCNA) implementing the WGCNA package 94 in R (version 4.2). See Supplemental Methods for details. 95 Statistical Modeling 96 Complete case analyses were performed to estimate the cross-sectional associations between 97 plasma protein abundances and both HIV serostatus and continuous LAVi. Proteins were 98 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 8 analyzed individually (single-analyte approach) and as clusters agnostically defined using 99 WGCNA (multi-analyte approach). Inference was made using multivariable linear regression 100 with Huber-White robust variance estimators to account for likely heteroscedasticity. Analyses 101 were corrected for multiple comparisons using a step-down false discovery rate (FDR) 102 procedure defined by Benjamini and Hochberg and a threshold for significance of FDR<0.05. 103 Mean standardized differences in plasma protein abundances by HIV serostatus were first 104 estimated adjusting for age, sex, race/ethnicity, educational attainment (dichotomized at high 105 school diploma), current hazardous alcohol use, pack-years of smoking in prior 5 years, 106 stimulant use in prior 5 years, opioid use in prior 5 years, hepatitis C virus infection (HCV), and 107 eGFR. Mean differences in LAVi were then estimated per SD increment in individual plasma 108 protein abundances, restricting to those proteins with HIV associations (at FDR<0.05) and 109 adjusting for covariates above, as well as HIV serostatus, body mass index (BMI), systolic blood 110 pressure (SBP), hypertension medication use, dyslipidemia, and diabetes. Proteins of interest 111 were defined as those with significant independent associations with both HIV serostatus and 112 LAVi. We further analyzed these proteins in subgroup analyses excluding PLWH with detectable 113 plasma HIV RNA (>50 copies/mL). We assessed the degree to which they may reflect the 114 independent association between positive HIV serostatus and higher LAVi. We evaluated 115 differences in their associations with LAVi by HIV serostatus and age using multiplicative 116 interaction terms. We estimated protein associations with basic clinical and HIV-related 117 characteristics. All analyses were conducted using R (version 4.2). 118 Functional Annotation Analysis 119 Annotation-based overrepresentation analyses were performed using the Gene Ontology (GO): 120 Biological Processes annotation database and the clusterProfiler package in R (version 4.2).20 121 This employs a Fisher’s exact test comparing the proportion of total significantly associated 122 proteins mapping to a given functional annotation to the proportion of total analyzed proteins 123 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 9 mapping to the same annotation. The background used in these analyses was proteins 124 measured on Olink Explore 3072 that passed quality control. Relationships between proteins of 125 interest were further explored using interaction networks generated using a database of known 126 and predicted protein-protein interactions, STRING.21 127 External Validation Cohort (Older Persons without HIV in MESA) 128 External validation was performed in MESA, a prospective community-based cohort initiated in 129 2000 to study the characteristics and progression of subclinical CVD among a diverse 130 population in the United States, age 45-84 years without clinical CVD at baseline.22 We used 131 data from exam 5 (years 2010-2012) when participants were age 53-94 years due to use of the 132 same CMR protocol23 and proteomics platform (Olink Explore 3072) as SMASH at that time 133 point. MESA participants were also, on average, over 10 years older than SMASH participants 134 at exam 5, allowing more extensive evaluation of the impact of older age on the identified 135 proteomic signature among PWOH. See Supplemental Methods for study population and data 136 acquisition details. 137 We first performed cross-sectional analyses of LAVi among MESA participants with proteomics 138 and CMR using the same approach to statistical modeling as described above, with slight 139 variation in covariates due to differences in data collection or prevalences among study 140 populations—specifically, we adjusted for field center, age, sex, race/ethnicity, education level, 141 smoking, BMI, SBP, blood pressure-lowering therapy, dyslipidemia, diabetes, and eGFR. 142 We additionally assessed associations of identified proteomic signatures with time to incident 143 AF and incident clinical HF. AF was defined by the presence of ≥1 inpatient or outpatient 144 International Classification of Diseases, Ninth Revision codes for AF. HF criteria included 145 symptomatic, physician-diagnosed HF as well as ≥1 of the following: (i) pulmonary 146 edema/congestion by chest x-ray; (ii) dilated LV or poor LV function by echocardiography or 147 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 10 ventriculography; and/or (iii) evidence of LV diastolic dysfunction. Inpatient and/or outpatient 148 records were independently reviewed by two physicians for HF classification and event date 149 assignment. See Supplemental Methods for detailed ascertainment and adjudication 150 procedures. Hazard ratios (HR) were estimated per SD increment in plasma protein level using 151 Cox proportional hazards models adjusting for the same covariates as cross-sectional analyses. 152 Participants with prevalent disease (AF or HF) at the start of follow-up were excluded from those 153 analyses, and time at risk was calculated as the time from the date of MESA exam 5 to an 154 incident event, death, loss to follow-up, or the end of follow-up, whichever occurred first. 155 Lastly, we estimated associations between proteins validated with at least one outcome and 156 basic clinical characteristics independent of field center. 157

158 SMASH Participant Characteristics 159 Of 468 persons initially enrolled in SMASH, 352 were selected for plasma proteomics based on 160 availability of complete CMR data, including gadolinium enhancement, and adequate stored 161 plasma volume (Figure S1A). Selected participants had lower representation of females and 162 lower prevalence of hypertension and HCV compared to unselected participants (Table S1). 163 Demographic, clinical, and CMR characteristics of selected participants are summarized in 164 TABLE 1 by HIV serostatus and by parent cohort in Table S2. Participants were mean (SD) age 165 55 (6) years, 25% female, 61% PLWH, and had predominantly normal LVEF. LAVi was higher 166 among PLWH compared to PWOH—29.7 (11.3) and 27.7 (8.3) mL/m2, respectively (p=0.02 for 167 difference, adjusting for demographics, substance use, and traditional CVD risk factors). 168 Proteomic Signature of HIV Serostatus 169 Of 2596 proteins analyzed, plasma abundances of 439 proteins were significantly associated 170 with HIV serostatus following multivariable adjustment (415 positively associated with positive 171 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 11 serostatus and 24 inversely associated, FDR<0.05) (FIGURE 2). These proteins were 172 statistically enriched for many biological processes, including T cell activation, differentiation, 173 and migration; B cell activation; natural killer (NK) cell mediated cytotoxicity; regulation of the 174 mitogen-activated protein kinase (MAPK) and extracellular signal-regulated kinase 1/2 175 (ERK1/ERK2) cascades; tumor necrosis factor (TNF) signaling; response to interleukin (IL)-1; 176 IL-10 production; and interferon-γ (IFN-γ) production. The strongest observed association with 177 HIV was that of cytotoxic and regulatory T cell molecule (CRTAM). On average, the adjusted 178 plasma abundance of CRTAM was 1.01 SD higher among PLWH compared to PWOH (95% 179 confidence interval [CI]: 0.841.17; FDR<1.0×10-18). Upon exclusion of PLWH with detectable 180 plasma HIV RNA, 414 of these 439 proteins remained significantly associated with positive HIV 181 serostatus (390 positively, 24 inversely); one of the strongest of which was CRTAM 182 (FDR<1.0×10-18). See Table S3 for results for all evaluated protein associations with HIV 183 serostatus and Table S4 for enrichment details. 184 Many classical markers reflecting chronic immune activation and systemic inflammation among 185 PLWH were significantly elevated among PLWH compared to PWOH, including the monocyte 186 activation markers soluble CD14 (FDR=2.60×10-10) and CD163 (FDR=0.009), as well as 187 CXCL10 (FDR=1.73×10-8) and TNFR1 (FDR=0.014). Notably, IL-6 levels did not differ by HIV 188 serostatus (p=0.19, FDR=0.42), consistent with previously reported immunoassay results 189 among this study population.24 These inferences were unchanged following exclusion of PLWH 190 with detectable plasma HIV RNA. 191 Of six protein clusters agnostically identified using WGCNA (details in Figure S2 and Tables 192 S5-S6), higher plasma abundance of one cluster (Brown) was independently associated with 193 positive HIV serostatus (p=3.99×10-6), and this association remained following exclusion of 194 PLWH with detectable plasma HIV RNA (p=1.02×10-4) (Table S7A-B). This cluster comprised 195 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 12 42 proteins enriched for regulation of T cell proliferation, TNF signaling, ephrin signaling, cell-196 cell adhesion, and extracellular matrix (ECM) organization (Table S8). 197 HIV-Associated Proteomic Signature of Left Atrial Remodeling 198 Next, we identified proteome features significantly associated with both HIV serostatus and LAVi 199 that may reflect potential contributors to HIV-associated LA remodeling. Of 439 proteins 200 associated with HIV serostatus, 78 were also associated with LAVi (74 positively, 4 inversely), 201 adjusting for demographics, substance use, HIV, HCV, and renal function. Upon further 202 adjustment for traditional CVD risk factors, 73 remained associated (69 positively, 4 inversely) 203 (FIGURE 3A, Table S9), all of which were associated with concordant directionality compared 204 to positive HIV serostatus (FIGURE 3B). In subgroup analyses, 72 (99%) of these proteins 205 differed in mean plasma level between PLWH with undetectable plasma HIV RNA and PWOH. 206 These proteins were statistically enriched for involvement in T cell activation, differentiation, and 207 proliferation; B cell activation; NK cell-mediated cytotoxicity; cell-cell adhesion; and TNF 208 signaling, amongst others (FIGURE 3C, Table S10). 209 The strongest observed protein association with LAVi was that of known HF biomarker, N-210 terminal pro-B type natriuretic peptide (NT-proBNP). On average, LAVi was 3.09 mL/m2 higher 211 (11% of mean LAVi) per SD increment in plasma NT-proBNP (95% CI: 2.164.03; 212 FDR=7.75×10-9). Notably, among the top 5 strongest associations with LAVi was CRTAM, which 213 also had the strongest association with HIV serostatus; LAVi was 2.37 mL/m2 higher on average 214 per SD (95% CI: 1.273.47; FDR=3.40×10-4). The 73 proteins associated with LAVi exhibited 215 pairwise correlations in plasma abundances that ranged widely from weak to strong (Figure S3) 216 with several moderate confidence protein-protein interactions detected (Figure S4). 217 Next, we determined what proportion of the multivariable adjusted association between HIV and 218 LAVi was accounted for by further adjustment for each identified proteome feature. When 219 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 13 further adjusting the HIV-LAVi analysis for each of these 73 proteins separately (in addition to 220 demographics, substance use, HCV, eGFR, and CVD risk factors), 38 proteins reduced the 221 independent HIV association by >30%, and 8 reduced it by >50%—those being CRTAM, CD27, 222 CD79B, CD160, TNF receptor superfamily (TNFRSF) 8, TNFRSF9, TNFRSF17, and IL18BP 223 (Table S11). Notably, adjustment for CRTAM reduced the association between positive HIV 224 serostatus and higher LAVi by 82%, suggesting it may substantially account for that observed 225 association. 226 We evaluated these 73 proteins for differences in their adjusted LAVi associations by HIV 227 serostatus and found no significant protein × HIV interactions (FDR<0.05). Given type II error is 228 likely high for these multiplicative tests, we present interactions with higher significance 229 thresholds to broaden hypothesis-generating potential. Five proteins exhibited associations with 230 LAVi that differed by HIV serostatus at FDR<0.10 and were present only among PLWH: NT-231 proBNP; CD80, CD83, and TNFRSF9 (CD137), all important immune checkpoint co-stimulatory 232 molecules;25-27 and folate receptor β (FOLR2), an important folate binding protein expressed by 233 myeloid cells, including infiltrating M2-like macrophages28 (FIGURE 3D). See Table S12 for 234

of all 73 individual protein interaction tests alongside modeling results stratified by HIV 235 serostatus. We detected no protein × age interactions (Table S13). 236 In our multi-analyte approach, we found higher plasma level of the one agnostically defined HIV-237 associated protein cluster (Brown) was also associated with higher LAVi, adjusting for age, sex, 238 race/ethnicity, eGFR, HCV, and HIV serostatus (2.03 mL/m2 higher per SD increment in Brown 239 cluster plasma level, 95% CI: 0.853.21, p=0.001, FIGURE 4). Following further adjustment for 240 substance use and traditional CVD risk factors, the estimate of association was only mildly 241 attenuated (1.96 mL/m2 higher per SD, 95% CI: 0.753.16, p=0.002). We found adjustment for 242 this protein cluster reduced the independent association between positive HIV serostatus and 243 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 14 higher LAVi by 36%. Moreover, this association differed by HIV serostatus (p=0.006) and was 244 only present among PLWH. We did not detect an interaction with age (p=0.18). 245 Eleven proteins were identified in both single- and multi-analyte approaches—specifically 246 butyrophilin subfamily 2 member A1 (BTN2A1), C-type lectin domain containing 14A 247 (CLEC14A), delta-like canonical notch ligand 1 (DLL1), ephrin A2 (EPHA2), endothelial cell 248 adhesion molecule (ESAM), T cell immunoglobulin and mucin-domain containing-3 (TIM-3 or 249 HAVCR2), junctional adhesion molecule 2 (JAM2), layilin (LAYN), TNFRSF1B, TNFRSF14, 250 and TNFRSF21. 251 Associations Between Proteomic Signature and Clinical Variables in SMASH 252 Many components of the identified protein signature were associated with age (p<0.05), 253 including the Brown cluster and 26 of 73 individual proteins. Notably, higher plasma CRTAM 254 was not associated with traditional CVD risk factors (age, BMI, hypertension, dyslipidemia, 255 diabetes) but was associated with substance use (smoking and stimulant use) and many HIV-256 related factors (HCV, detectable plasma HIV RNA, lower current and nadir CD4+ T cell counts, 257 not receiving ART, and shorter ART duration). Similarly, the Brown protein cluster was largely 258 only associated with substance use and HIV-related clinical factors—smoking, opioid use, HCV, 259 detectable plasma HIV RNA, and lower current and nadir CD4+ T cell counts. See Figure S5 260 for results of all evaluated protein associations with demographic and clinical characteristics. 261 External Validation of Protein Signatures with Indexed Left Atrial Volume in MESA 262 A total of 1242 MESA participants had both CMR imaging and Olink Explore 3072 proteomic 263 data (Figure S1B), demographic and clinical characteristics of whom are summarized in 264 TABLE 2. Participants were mean (SD) age 67 (9) years, 52% female, and had predominantly 265 normal LVEF. 266 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 15 We tested each feature of the HIV-associated proteomic signature of LA remodeling that we 267 identified in SMASH for associations with LAVi in MESA. Among this cohort of older PWOH, 8 268 of 73 plasma proteins were cross-sectionally associated with LAVi following multivariable 269 adjustment (TABLE 3, Table S14). Following NT-proBNP, the strongest validated association 270 was that of CLEC14A (mean difference: 1.36 mL/m2 higher per SD, 95% CI: 0.652.07, 271 FDR=0.006), followed by collagen type IV alpha 1 chain (COL4A1) (1.20 mL/m2 higher per SD, 272 95% CI: 0.561.85, FDR=0.006). Moreover, 28 (of 73 proteins) had LAVi associations that 273 differed by age (interaction FDR<0.05, 41 proteins FDR<0.10) (Table S15). Sixteen of these 274 proteins that differed by age had significant associations with LAVi only among a subgroup of 275 MESA participants older than median age at exam 5 (≥67 years), and only one (NT-proBNP) 276 was associated among participants age <67 years (TABLE 3). Furthermore, plasma level of the 277 Brown protein cluster was only associated with LAVi among older MESA participants (age ≥67 278 years, mean difference in LAVi per SD: 1.23 mL/m2, 95% CI: 0.032.43, p=0.04; age <67 years: 279 -0.72, 95% CI: -1.750.32, p=0.17; interaction p=0.005). 280 External Validation of Protein Signatures with Incident Clinical Events in MESA 281 Incident clinical event analyses did not require CMR data and were performed in a larger 282 subgroup of the MESA cohort to maximize statistical power (Figure S1C). Distributions of 283 demographic and clinical characteristics were similar to those in cross-sectional analyses 284 presented above (TABLE 2). 285 Among 2185 MESA participants without known AF at exam 5, there were 252 incident AF 286 events over a median [IQR] follow-up of 7.8 [0.9] years. Only NT-proBNP (of 73 proteins) was 287 associated with incident AF (Table S16 and Figure S6) following multivariable adjustment (HR: 288 1.54 per SD increment in NT-proBNP, 95% CI: 1.331.77, FDR=2.08×10-7). Plasma abundance 289 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 16 of the Brown protein cluster was also associated with incident AF (HR: 1.18 per SD increment, 290 95%CI: 1.01 to 1.37, p=0.036). No interactions with baseline age were detected (all FDR>0.10). 291 Among 2273 MESA participants without known HF at exam 5, there were 54 incident 292 adjudicated clinical HF events over a median [IQR] follow-up of 8.9 [0.8] years. Thirty of 73 293 proteins were associated with incident HF (FIGURE 5) following adjustment, the strongest of 294 which was TNFRSF1B (HR: 1.81 per SD increment, 95% CI: 1.412.33, FDR=2.58×10-4). Five 295 of these proteins were also associated with LAVi in MESA when restricting to participants above 296 the median age—CLEC14A, cysteine rich transmembrane BMP regulator 1 (CRIM1), NT-297 proBNP, thrombospondin-2 (THBS2), and tissue inhibitor matrix metalloproteinase-1 (TIMP1). 298 See Table S17 and Figure S7 for results of all evaluated protein associations with time to 299 incident HF. The hazard of HF was also significantly greater with higher plasma level of the 300 Brown protein cluster (HR: 1.79 per SD, 95% CI: 1.32 to 2.43, p=1.82×10-4). No interactions with 301 age were detected (all FDR>0.10). Notably, 9 of the 11 proteins identified in both single- and 302 multi-analyte discovery analyses in SMASH were associated with incident HF in MESA—303 BTN2A1, CLEC14A, DLL1, ESAM, HAVCR2, JAM2, LAYN, TNFRSF1B, and TNFRSF14. 304 Associations Between Validated Proteomic Signature Features and Clinical Variables 305 Every component of the protein signature validated with clinical HF among MESA participants 306 was positively and strongly associated with age (all p<0.001) (FIGURE 5). Many components 307 were also strongly associated with hypertension (28 proteins and Brown cluster), diabetes (25 308 proteins and Brown cluster), and higher BMI (22 proteins and Brown cluster)—which remained 309 associated when adjusted for eGFR. 310

311 In this discovery study, we identified a plasma proteomic signature representing potential novel 312 biological contributors to the association between HIV infection and left atrial remodeling—a 313 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 17 marker of subclinical heart disease. Importantly, this signature was strongly associated with age 314 and higher risk for incident clinical HF among a large independent cohort of PWOH who were 315 over 10 years older. This proteomic signature comprised 73 individual proteins and one 316 agnostically-defined cluster of 42 proteins, largely comprising immune checkpoint proteins 317 (ICPs), cytokine signaling, ephrin signaling, and ECM organization pathways. Our findings 318 support the premise that persistent immune activation, higher systemic inflammation, and 319 fibrosis contribute to HIV-associated atrial remodeling, even among PLWH on ART with viral 320 suppression. Furthermore, these same mechanisms may also contribute to aging-related HF 321 among PWOH. 322 Aging, HIV and Heart Failure Converge on Inflammation and Immune Activation 323 It is well-known that PLWH exhibit elevated systemic inflammation and immune activation 324 compared to PWOH, characterized by T cell exhaustion, macrophage activation, and elevated 325 levels of circulating pro-inflammatory cytokines that persist even among virally suppressed 326 PLWH on ART. Immunosenescence is a well-described consequence of aging, and 327 maladaptively activated innate and adaptive immune responses have also been implicated in 328 acute and chronic HF.29 This is the prevailing hypothesized mechanism through which PLWH 329 experience excess risk of aging-related diseases, including HF, and our findings support and 330 expand upon those previously put forward. 331 Proteins Associated with HIV, Left Atrial Remodeling, and Incident Clinical HF 332 We identified 30 proteins associated with HIV serostatus and LAVi in SMASH and incident 333 clinical HF among older PWOH in MESA, many of which are novel. Five of these proteins were 334 associated with LAVi among PWOH in MESA when restricting to participants above median 335 age. Interestingly, this list also included several proteins exhibiting especially high potential for 336 mediating and/or moderating effects on the association between HIV and atrial remodeling as 337 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 18 well as 9 proteins identified in both single- and multi-analyte discovery analyses. This overlap in 338 identification is important, because these approaches are both agnostic and independent of 339 each other; identification by both approaches therefore strengthens the plausibility of the 340 identified proteins as potential contributors to both HIV- and aging-related LA remodeling and 341 HF. 342 Many of these proteins of especially high interest are important ICPs—specifically CD27, 343 TNFRSF8 (CD30), CD79B, CD80, TNFRSF9 (CD137), HAVCR2 (TIM-3), TNFRSF14 (HVEM), 344 TIGIT, TNFRSF1B (TNFR2), and less extensively studied BTN2A1, CD83, IL18BP, and 345 TNFRSF17. Many of the remaining proteins—CLEC14A, CRIM1, DLL1, ESAM, JAM2, LAYN, 346 THBS2, and TIMP1—are involved broadly in cell migration and tissue homeostasis. 347 Immune Checkpoint Proteins 348 Activation of T cells, B cells, and NK cells is regulated by several costimulatory and inhibitory 349 ICPs. Elevated expression of ICPs is a known consequence of HIV infection that persists 350 despite ART and contributes to immune dysfunction, HIV disease progression, and viral 351 persistence, detailed recently.30 Many of the ICPs we identified belong to the TNF receptor 352 superfamily, upregulation of which is both a hallmark of HIV-associated inflammation and 353 immune activation31 and has been extensively observed in HF.32 Higher soluble CD27 and 354 CD80 have also recently been identified as correlates of risk for non-AIDS adverse events 355 among PLWH receiving ART.33 Therapeutic blockade of ICPs has, however, proven 356 controversial and dependent on whether target molecules serve co-stimulatory or co-inhibitory 357 functions. Taken together with the unique features of HIV-associated immune dysfunction, this 358 highlights the need for focused future investigation of ICPs in HIV-associated cardiac 359 remodeling and HF. 360 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 19 Cell Migration and Tissue Homeostasis 361 In this study, the one agnostically defined protein cluster of interest included four proteins 362 involved in ephrin signaling, one of which was also identified in single-analyte discovery 363 analyses (EPHA2). Ephrin signaling mediates a wide range of cellular processes involved in 364 tissue homeostasis and response to injury, including angiogenesis and activation/migration of T 365 cells, B cells, and macrophages.34 Though ephrin signaling has not been well-studied among 366 PLWH, it has been linked to other viral infections35 as well as cardiomyocyte development, 367 systolic and diastolic function, and TGF-β mediated cardiac fibrosis.36-38 368 Several of these higher interest proteins are related to tissue growth and development, 369 angiogenesis, and ECM remodeling—including CRIM1, THBS2, and TIMP1. CRIM1 modulates 370 activity of bone morphogenic proteins, members of the TGFβ superfamily that are known to play 371 roles in cardiac development and disease pathophysiology in HF, pulmonary arterial 372 hypertension, and atherosclerosis.39 Both THBS2 and TIMP1 are critical regulators of tissue 373 homeostasis with roles in ECM remodeling, angiogenesis, and cell growth/differentiation, and 374 have been extensively implicated in HF pathogenesis,40,41 including in recent plasma proteome 375 studies.42-44 376 Other proteome features of particularly high interest have been linked to migration of various 377 immune cell subsets in tissue homeostasis—specifically DLL1 (a ligand of Notch receptors), 378 ESAM (endothelial cell adhesion molecule), JAM2 (involved in cell migration), and LAYN 379 (involved in focal adhesion). Immune cell recruitment plays an important role in myocardial 380 response to injury;29 however, further investigation is needed to clarify the role of these proteins 381 in shared mechanisms of HIV- and aging-related cardiac remodeling and HF. 382 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 20 CRTAM 383 We identified potential biological contributors to the association between HIV and LA 384 remodeling. Abundance of one protein—CRTAM—reduced that association by 82%, markedly 385 more than any other protein, and was strongly associated with several clinical factors related to 386 HIV disease severity. CRTAM is an immune activation marker expressed by NK cells, CD8+ T 387 cells and a subset of CD4+ T cells with enhanced cytotoxic activity.45 Measures of CRTAM 388 expression have previously been shown to associate with positive HIV serostatus, HIV viral 389 reservoir size, and viral rebound kinetics.46,47 It is also believed critical to shaping the gut 390 microbiome and Th17 responses at the mucosa,48 increased permeability of which is a 391 consequence of Th17 HIV-susceptibility and a contributor to chronic inflammation among 392 PLWH.49 CRTAM was not associated with LAVi or incident HF among older PWOH in MESA, 393 consistent with our hypothesis that it is an HIV-specific contributor to CVD. Taken together, our 394

and prior work suggest CRTAM may represent a novel HIV-specific marker of 395 cardiovascular risk worthy of further investigation into its prognostic and therapeutic utility. 396

397 There are limitations to this study. Protein levels measured with Olink are semi-quantitative, 398 limiting comparison to other assays. Previously published data demonstrated many proteins 399 assessed by Olink Explore 3072 were associated with cis-protein quantitative trait loci (pQTLs), 400 suggesting specificity of protein identification.50 Nonetheless, all proteins of interest require 401 validation using other methods. 402 Type II error may be inflated due to the method used for multiple testing correction, which 403 assumes independence of tests (i.e., biological independence of proteins) and is therefore 404 conservative. For example, given macrophage activation is a hallmark of chronic immune 405 activation among PLWH, we hypothesized it would be detected in our enrichment analyses. It 406 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 21 was indeed positively associated with HIV and LAVi in SMASH but not following multiple testing 407 correction (p=0.03; FDR=0.36). 408 Discovery analyses are cross-sectional, resulting in temporal ambiguity. There are also 409

of incident event ascertainment in MESA, particularly for AF which relied on 410 hospitalizations and claims data alone, possibly leading to misclassification bias. Finally, though 411 the study populations are sociodemographically diverse, results may not be translatable to 412 unrepresented or underrepresented subpopulations. 413

414 In conclusion, we discovered a novel plasma proteomic signature that may in part reflect or 415 contribute to HIV-associated left atrial remodeling and also predicted incident clinical HF among 416 a large independent cohort of older PWOH. This signature was enriched in pathways of immune 417 activation, cytokine signaling, and extracellular matrix organization. Our findings suggest 418 pathways underlying risk of HF among PLWH may also contribute to HF pathogenesis among 419 older PWOH. If successfully validated in other external populations, these proteins may help 420 refine current HF risk prediction models and could represent novel therapeutic targets for HF 421 among PLWH and PWOH. 422 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 22 ACKNOWLEDGMENTS The authors gratefully acknowledge the contributions of the study participants and the dedication of the MWCCS, ALIVE, SMASH, and MESA-TOPMed investigators and staff. Full MWCCS acknowledgement may be found at https://statepi.jhsph.edu/mwccs/acknowledgements/. A full list of participating MESA investigators and institutes can be found at http://www.mesa-nhlbi.org. We would also like to thank Jinshui Fan and Giovanna Fantoni for their help in the Olink proteomic analysis. SOURCES OF FUNDING This study was supported in part by the National Institutes of Health (NIH: R01 HL126552 (SMASH study, KCW and WSP); U01-DA036297 (ALIVE data collection); and U01-HL146201, U01-HL146193, U01-HL146240, and U01-HL146205 (MACS, WIHS data collection). Proteomics were supported by U01HL146201, Johns Hopkins University Center for AIDS Research (P30AI094189), the Intramural Research Program of the National Institute on Aging, NIH (ZIAAG000297), and a generous gift from Dr. Nancy Grasmick. TEP is supported by NIH/NHLBI T32 HL007227. TTB is supported in part by K24 AI120834. VSH is supported by NIH NHLBI 1K23HL166770-01 (Bethesda, MD) and Sarnoff Scholar Award 138828 (McLean, VA). Support for the Multi-Ethnic Study of Atherosclerosis (MESA) projects are conducted and supported by the National Heart, Lung, and Blood Institute (NHLBI) in collaboration with MESA investigators. Support for MESA is provided by contracts 75N92020D00001, HHSN268201500003I, N01-HC-95159, 75N92020D00005, N01-HC-95160, 75N92020D00002, N01-HC-95161, 75N92020D00003, N01-HC-95162, 75N92020D00006, N01-HC-95163, 75N92020D00004, N01-HC-95164, 75N92020D00007, N01-HC-95165, N01-HC-95166, N01- HC-95167, N01-HC-95168, N01-HC-95169, UL1-TR-000040, UL1-TR-001079, UL1-TR- 001420, UL1TR001881, DK063491, and R01HL105756. Molecular data for the Trans-Omics in Precision Medicine (TOPMed) program was supported by the NHLBI. Proteomics was supported by TOPMed MESA Multi-Omics (HHSN2682015000031, HSN26800004, HHSN268201600034I).Core support including centralized genomic read mapping and genotype calling, along with variant quality metrics and filtering were provided by the TOPMed Informatics Research Center (3R01HL-117626-02S1; contract HHSN268201800002I). Core support including phenotype harmonization, data management, sample-identity QC, and general program coordination were provided by the TOPMed Data Coordinating Center (R01HL- 120393; U01HL-120393; contract HHSN268201800001I). DISCLOSURES TTB has served as a consultant to Gilead Sciences, Merck, ViiV Healthcare, and Janssen. FJP has served as a consultant and/or on the Speakers Bureau for Gilead Sciences, Merck, ViiV Healthcare, and Janssen. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 23 SUPPLEMENTAL MATERIALS Supplemental Methods Tables S1–S17 (Tables S3, S5, and S9 in Excel file due to size) Figure S1–S7

1. Feinstein MJ, Hsue PY, Benjamin LA, Bloomfield GS, Currier JS, Freiberg MS, Grinspoon SK, Levin J, Longenecker CT, Post WS. Characteristics, Prevention, and Management of Cardiovascular Disease in People Living With HIV: A Scientific Statement From the American Heart Association. Circulation. 2019;140:e98-e124. doi: 10.1161/CIR.0000000000000695 2. Wu KC, Woldu B, Post WS, Hays AG. Prevention of heart failure, tachyarrhythmias and sudden cardiac death in HIV. Curr Opin HIV AIDS. 2022;17:261-269. doi: 10.1097/COH.0000000000000753 3. Alonso A, Barnes AE, Guest JL, Shah A, Shao IY, Marconi V. HIV Infection and Incidence of Cardiovascular Diseases: An Analysis of a Large Healthcare Database. J Am Heart Assoc. 2019;8:e012241. doi: 10.1161/JAHA.119.012241 4. Freiberg MS, Chang CH, Skanderson M, Patterson OV, DuVall SL, Brandt CA, So- Armah KA, Vasan RS, Oursler KA, Gottdiener J, et al. Association Between HIV Infection and the Risk of Heart Failure With Reduced Ejection Fraction and Preserved Ejection Fraction in the Antiretroviral Therapy Era: Results From the Veterans Aging Cohort Study. JAMA Cardiol. 2017;2:536-546. doi: 10.1001/jamacardio.2017.0264 5. Chen Y, Gao Y, Zhou Y, Li X, Wang H, Polonsky TS, Vermund SH, Qian HZ, Qian F. Human Immunodeficiency Virus Infection and Incident Heart Failure: A Meta-Analysis of Prospective Studies. J Acquir Immune Defic Syndr. 2021;87:741-749. doi: 10.1097/QAI.0000000000002629 6. Sardana M, Hsue PY, Tseng ZH, Vittinghoff E, Nah G, Dewland TA, Marcus GM. Human Immunodeficiency Virus Infection and Incident Atrial Fibrillation. J Am Coll Cardiol. 2019;74:1512-1514. doi: 10.1016/j.jacc.2019.07.027 7. Sinha A, Feinstein M. Epidemiology, pathophysiology, and prevention of heart failure in people with HIV. Prog Cardiovasc Dis. 2020;63:134-141. doi: 10.1016/j.pcad.2020.01.002 8. Doria de Vasconcellos H, Post WS, Ervin AM, Haberlen SA, Budoff M, Malvestutto C, Magnani JW, Feinstein MJ, Brown TT, Lima JAC, et al. Associations Between HIV Serostatus and Cardiac Structure and Function Evaluated by 2-Dimensional Echocardiography in the Multicenter AIDS Cohort Study. J Am Heart Assoc. 2021;10:e019709. doi: 10.1161/JAHA.120.019709 9. Wu KC, Haberlen SA, Plankey MW, Palella FJ, Piggott DA, Kirk GD, Margolick JB, Post WS. Human immunodeficiency viral infection and differences in interstitial ventricular fibrosis and left atrial size. Eur Heart J Cardiovasc Imaging. 2021. doi: 10.1093/ehjci/jeab037 10. Hsue PY, Hunt PW, Ho JE, Farah HH, Schnell A, Hoh R, Martin JN, Deeks SG, Bolger AF. Impact of HIV infection on diastolic function and left ventricular mass. Circ Heart Fail. 2010;3:132-139. doi: 10.1161/CIRCHEARTFAILURE.109.854943 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 24 11. Reinsch N, Kahlert P, Esser S, Sundermeyer A, Neuhaus K, Brockmeyer N, Potthoff A, Erbel R, Buck T, Neumann T. Echocardiographic findings and abnormalities in HIV- infected patients: results from a large, prospective, multicenter HIV-heart study. Am J Cardiovasc Dis. 2011;1:176-184. 12. Shuldiner SR, Wong LY, Peterson TE, Wolfson J, Jermy S, Saad H, Lumbamba MAJ, Singh A, Shey M, Meintjes G, et al. Myocardial Fibrosis Among Antiretroviral Therapy- Treated Persons With Human Immunodeficiency Virus in South Africa. Open Forum Infect Dis. 2021;8:ofaa600. doi: 10.1093/ofid/ofaa600 13. Zanni MV, Awadalla M, Toribio M, Robinson J, Stone LA, Cagliero D, Rokicki A, Mulligan CP, Ho JE, Neilan AM, et al. Immune Correlates of Diffuse Myocardial Fibrosis and Diastolic Dysfunction Among Aging Women With Human Immunodeficiency Virus. J Infect Dis. 2020;221:1315-1320. doi: 10.1093/infdis/jiz184 14. Nagueh SF, Smiseth OA, Appleton CP, Byrd BF, Dokainish H, Edvardsen T, Flachskampf FA, Gillebert TC, Klein AL, Lancellotti P, et al. Recommendations for the Evaluation of Left Ventricular Diastolic Function by Echocardiography: An Update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. Eur Heart J Cardiovasc Imaging. 2016;17:1321-1360. doi: 10.1093/ehjci/jew082 15. Kaslow RA, Ostrow DG, Detels R, Phair JP, Polk BF, Rinaldo CR. The Multicenter AIDS Cohort Study: rationale, organization, and selected characteristics of the participants. Am J Epidemiol. 1987;126:310-318. doi: 10.1093/aje/126.2.310 16. Barkan SE, Melnick SL, Preston-Martin S, Weber K, Kalish LA, Miotti P, Young M, Greenblatt R, Sacks H, Feldman J. The Women's Interagency HIV Study. WIHS Collaborative Study Group. Epidemiology. 1998;9:117-125. 17. Vlahov D, Anthony JC, Munoz A, Margolick J, Nelson KE, Celentano DD, Solomon L, Polk BF. The ALIVE study, a longitudinal study of HIV-1 infection in intravenous drug users: description of methods and characteristics of participants. NIDA Res Monogr. 1991;109:75-100. 18. Tao S, Ashikaga H, Ciuffo LA, Yoneyama K, Lima JAC, Frank TF, Weiss RG, Tomaselli GF, Wu KC. Impaired left atrial function predicts inappropriate shocks in primary prevention implantable cardioverter-defibrillator candidates. J Cardiovasc Electrophysiol. 2017;28:796-805. doi: 10.1111/jce.13234 19. Ebai T, Kamali-Moghaddam M, Landegren U. Parallel protein detection by solid-phase proximity ligation assay with real-time PCR or sequencing. Curr Protoc Mol Biol. 2015;109:20.10.21-20.10.25. doi: 10.1002/0471142727.mb2010s109 20. Wu T, Hu E, Xu S, Chen M, Guo P, Dai Z, Feng T, Zhou L, Tang W, Zhan L, et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. Innovation (Camb). 2021;2:100141. doi: 10.1016/j.xinn.2021.100141 21. Szklarczyk D, Kirsch R, Koutrouli M, Nastou K, Mehryary F, Hachilif R, Gable AL, Fang T, Doncheva NT, Pyysalo S, et al. The STRING database in 2023: protein-protein association networks and functional enrichment analyses for any sequenced genome of interest. Nucleic Acids Res. 2023;51:D638-D646. doi: 10.1093/nar/gkac1000 22. Bild DE, Bluemke DA, Burke GL, Detrano R, Diez Roux AV, Folsom AR, Greenland P, Jacob DR, Kronmal R, Liu K, et al. Multi-Ethnic Study of Atherosclerosis: objectives and design. Am J Epidemiol. 2002;156:871-881. doi: 10.1093/aje/kwf113 23. Ambale-Venkatesh B, Hong SY, Habibi M, Lim D, Wu E, Jorgensen N, Hundley WG, Shea S, Liu K, Gomes AS, et al. Left Atrial Remodeling Assessed by Serial Longitudinal All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 25 Cardiac MRI in MESA. JACC Cardiovasc Imaging. 2021;14:1678-1680. doi: 10.1016/j.jcmg.2021.02.014 24. Peterson TE, Landon C, Haberlen SA, Bhondoekhan F, Plankey MW, Palella FJ, Piggott DA, Margolick JB, Brown TT, Post WS, et al. Circulating biomarker correlates of left atrial size and myocardial extracellular volume fraction among persons living with and without HIV. BMC Cardiovasc Disord. 2022;22:393. doi: 10.1186/s12872-022-02835-y 25. Simons KH, de Jong A, Jukema JW, de Vries MR, Arens R, Quax PHA. T cell co- stimulation and co-inhibition in cardiovascular disease: a double-edged sword. Nat Rev Cardiol. 2019;16:325-343. doi: 10.1038/s41569-019-0164-7 26. Yousif LI, Tanja AA, de Boer RA, Teske AJ, Meijers WC. The role of immune checkpoints in cardiovascular disease. Front Pharmacol. 2022;13:989431. doi: 10.3389/fphar.2022.989431 27. Grosche L, Knippertz I, König C, Royzman D, Wild AB, Zinser E, Sticht H, Muller YA, Steinkasserer A, Lechmann M. The CD83 Molecule - An Important Immune Checkpoint. Front Immunol. 2020;11:721. doi: 10.3389/fimmu.2020.00721 28. Puig-Kröger A, Sierra-Filardi E, Domínguez-Soto A, Samaniego R, Corcuera MT, Gómez-Aguado F, Ratnam M, Sánchez-Mateos P, Corbí AL. Folate receptor beta is expressed by tumor-associated macrophages and constitutes a marker for M2 anti- inflammatory/regulatory macrophages. Cancer Res. 2009;69:9395-9403. doi: 10.1158/0008-5472.CAN-09-2050 29. Adamo L, Rocha-Resende C, Prabhu SD, Mann DL. Reappraising the role of inflammation in heart failure. Nat Rev Cardiol. 2020;17:269-285. doi: 10.1038/s41569- 019-0315-x 30. Gubser C, Chiu C, Lewin SR, Rasmussen TA. Immune checkpoint blockade in HIV. EBioMedicine. 2022;76:103840. doi: 10.1016/j.ebiom.2022.103840 31. Wang C, Watts TH. Maintaining the balance: costimulatory TNFRs and control of HIV. Cytokine Growth Factor Rev. 2012;23:245-254. doi: 10.1016/j.cytogfr.2012.05.011 32. Murphy SP, Kakkar R, McCarthy CP, Januzzi JL. Inflammation in Heart Failure: JACC State-of-the-Art Review. J Am Coll Cardiol. 2020;75:1324-1340. doi: 10.1016/j.jacc.2020.01.014 33. Premeaux TA, Moser CB, McKhann A, Hoenigl M, Yeung ST, Pang APS, Corley MJ, Lederman MM, Landay AL, Gianella S, et al. Monitoring Circulating Immune Checkpoint Proteins as Predictors of Non-AIDS Morbid Events in People With HIV Initiating Antiretroviral Therapy. Open Forum Infect Dis. 2022;9:ofab570. doi: 10.1093/ofid/ofab570 34. Darling TK, Lamb TJ. Emerging Roles for Eph Receptors and Ephrin Ligands in Immunity. Front Immunol. 2019;10:1473. doi: 10.3389/fimmu.2019.01473 35. de Boer ECW, van Gils JM, van Gils MJ. Ephrin-Eph signaling usage by a variety of viruses. Pharmacol Res. 2020;159:105038. doi: 10.1016/j.phrs.2020.105038 36. Luxan G, Stewen J, Diaz N, Kato K, Maney SK, Aravamudhan A, Berkenfeld F, Nagelmann N, Drexler HC, Zeuschner D, et al. Endothelial EphB4 maintains vascular integrity and transport function in adult heart. Elife. 2019;8. doi: 10.7554/eLife.45863 37. Karsenty C, Guilbeau-Frugier C, Genet G, Seguelas MH, Alzieu P, Cazorla O, Montagner A, Blum Y, Dubroca C, Maupoint J, et al. Ephrin-B1 regulates the adult diastolic function through a late postnatal maturation of cardiomyocyte surface crests. Elife. 2023;12. doi: 10.7554/eLife.80904 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 26 38. Su SA, Xie Y, Zhang Y, Xi Y, Cheng J, Xiang M. Essential roles of EphrinB2 in mammalian heart: from development to diseases. Cell Commun Signal. 2019;17:29. doi: 10.1186/s12964-019-0337-3 39. Ye D, Liu Y, Pan H, Feng Y, Lu X, Gan L, Wan J, Ye J. Insights into bone morphogenetic proteins in cardiovascular diseases. Front Pharmacol. 2023;14:1125642. doi: 10.3389/fphar.2023.1125642 40. Zhang K, Li M, Yin L, Fu G, Liu Z. Role of thrombospondin‑1 and thrombospondin‑2 in cardiovascular diseases (Review). Int J Mol Med. 2020;45:1275-1293. doi: 10.3892/ijmm.2020.4507 41. Moore L, Fan D, Basu R, Kandalam V, Kassiri Z. Tissue inhibitor of metalloproteinases (TIMPs) in heart failure. Heart Fail Rev. 2012;17:693-706. doi: 10.1007/s10741-011- 9266-y 42. Egerstedt A, Berntsson J, Smith ML, Gidlöf O, Nilsson R, Benson M, Wells QS, Celik S, Lejonberg C, Farrell L, et al. Profiling of the plasma proteome across different stages of human heart failure. Nat Commun. 2019;10:5830. doi: 10.1038/s41467-019-13306-y 43. Wells QS, Gupta DK, Smith JG, Collins SP, Storrow AB, Ferguson J, Smith ML, Pulley JM, Collier S, Wang X, et al. Accelerating Biomarker Discovery Through Electronic Health Records, Automated Biobanking, and Proteomics. J Am Coll Cardiol. 2019;73:2195-2205. doi: 10.1016/j.jacc.2019.01.074 44. Regan JA, Truby LK, Tahir UA, Katz DH, Nguyen M, Kwee LC, Deng S, Wilson JG, Mentz RJ, Kraus WE, et al. Protein biomarkers of cardiac remodeling and inflammation associated with HFpEF and incident events. Sci Rep. 2022;12:20072. doi: 10.1038/s41598-022-24226-1 45. Hermans D, van Beers L, Broux B. Nectin Family Ligands Trigger Immune Effector Functions in Health and Autoimmunity. Biology (Basel). 2023;12. doi: 10.3390/biology12030452 46. Corley MJ, Pang APS, Rasmussen TA, Tolstrup M, Søgaard OS, Ndhlovu LC. Candidate host epigenetic marks predictive for HIV reservoir size, responsiveness to latency reversal, and viral rebound. AIDS. 2021;35:2269-2279. doi: 10.1097/QAD.0000000000003065 47. Sperk M, Zhang W, Nowak P, Neogi U. Plasma soluble factor following two decades prolonged suppressive antiretroviral therapy in HIV-1-positive males: A cross-sectional study. Medicine (Baltimore). 2018;97:e9759. doi: 10.1097/MD.0000000000009759 48. Perez-Lopez A, Nuccio SP, Ushach I, Edwards RA, Pahu R, Silva S, Zlotnik A, Raffatellu M. CRTAM Shapes the Gut Microbiota and Enhances the Severity of Infection. J Immunol. 2019;203:532-543. doi: 10.4049/jimmunol.1800890 49. Renault C, Veyrenche N, Mennechet F, Bedin AS, Routy JP, Van de Perre P, Reynes J, Tuaillon E. Th17 CD4+ T-Cell as a Preferential Target for HIV Reservoirs. Front Immunol. 2022;13:822576. doi: 10.3389/fimmu.2022.822576 50. Eldjarn GH, Ferkingstad E, Lund SH, Helgason H, Magnusson OT, Gunnarsdottir K, Olafsdottir TA, Halldorsson BV, Olason PI, Zink F, et al. Large-scale plasma proteomics comparisons through genetics and disease associations. Nature. 2023;622:348-358. doi: 10.1038/s41586-023-06563-x All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 27 TABLE 1. SMASH participant characteristics by HIV serostatus (n=352) CHARACTERISTIC Median [IQR] or % (n) PLWH (n=214) PWOH (n=138) Study site --- --- MACS Baltimore 75 (35%) 46 (33%) MACS Chicago 28 (13%) 17 (12%) WIHS 39 (18%) 21 (15%) ALIVE 72 (34%) 54 (39%) Demographics Age, years 55 [51, 58] 55 [52, 58] Female sex at birth 55 (26%) 34 (25%) Race and ethnicity --- --- Black, non-Hispanic 151 (71%) 94 (68%) White, non-Hispanic 52 (24%) 34 (25%) Hispanic 11 (5%) 10 (7%) Education level ≥ high school 156 (73%) 108 (78%) Substance Use Smoking status --- --- Current 114 (53%) 57 (41%) Former 56 (26%) 57 (41%) Never 44 (21%) 24 (17%) Pack-years of smoking a 0.99 [0.00, 2.82] 0.21 [0.00, 2.16] Hazardous alcohol use (AUDIT score >8) 23 (11%) 20 (14%) Opioid use a 61 (29%) 47 (34%) Stimulant use a 87 (41%) 50 (36%) Clinical Factors History of cardiovascular disease 14 (7%) 8 (6%) Body mass index, kg/m2 25.7 [23.0, 29.3] 26.9 [24.1, 31.0] Hypertension b 108 (51%) 74 (54%) Systolic blood pressure, mmHg 124 [118, 134] 128 [121, 135] Blood pressure-lowering medication use 74 (35%) 49 (36%) Dyslipidemia c 130 (61%) 77 (56%) Total cholesterol, mg/dL 172 [149, 195] 177 [151, 208] High density lipoprotein cholesterol, mg/dL 51 [41, 63] 58 [46, 69] Lipid-lowering medication use 54 (25%) 31 (23%) Diabetes d 27 (13%) 16 (12%) Diabetes medication use 20 (9%) 12 (9%) eGFR, CKD-EPI, mL/min/1.73m2 89 [74, 103] 93 [78, 106] Hepatitis C diagnosis 50 (23%) 28 (20%) All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 28 Cardiovascular Magnetic Resonance LV ejection fraction, % 72 [67, 76] 72 [68, 76] LV ejection fraction <50% 4 (2%) 0 (0%) LV mass indexed by BSA, mg/m2 61.8 [56.8, 68.6] 61.7 [55.8, 67.0] LV end-diastolic volume indexed by BSA, mL/m2 67.6 [58.0, 77.0] 64.4 [54.3, 76.2] LA volume indexed by BSA, mL/m2 28.1 [22.1, 35.8] 26.7 [22.6, 32.9] LA volume indexed by BSA ≥40 mL/m2 32 (15%) 8 (6%) LA volume indexed by BSA ≥53 mL/m2 6 (3%) 2 (1%) Late gadolinium enhancement 81 (38%) 46 (33%) Extracellular volume fraction, % 28.9 [26.5, 31.2] 28.2 [26.1, 29.9] HIV-Related Factors HIV viral load detectable (>50 RNA copies/mL) 55 (26%) --- CD4+ T cell count, cells/µL 605 [400, 815] --- CD4+ nadir T cell count, cells/µL 271 [149, 386] --- History of AIDS diagnosis 29 (20%) --- ART use 188 (88%) --- Duration of ART, years 12.8 [5.4, 16.0] --- Protease inhibitor base 75 (35%) --- Non-nucleoside reverse transcriptase inhibitor base 57 (27%) --- Integrase strand inhibitor base 53 (25%) --- Other ART base 3 (1%) --- a Reported in the five years preceding cardiovascular magnetic resonance imaging (CMR). b Hypertension is defined as use of antihypertensive medication or systolic blood pressure≥140 mmHg or diastolic blood pressure ≥90 mmHg averaged over the preceding 5 years when available or at the time of CMR study visit. c Dyslipidemia is defined as use of lipid-lowering medication or fasting total cholesterol level ≥200 mg/dL or low-density lipoprotein cholesterol level ≥130 mg/dL or high-density lipoprotein cholesterol level <40 mg/dL or serum triglyceride level ≥150 mg/dL closest to the time of CMR study visit. d Diabetes is defined as use of hypoglycemic medication or fasting serum glucose levels ≥126 mg/dL closest to the time of CMR study visit. Hemoglobin A1C level <6.5% was used to exclude diabetes if fasting glucose levels were not available. PLWH=people living with HIV; PWOH=people without HIV; IQR=interquartile range, reported as (25th, 75th) percentiles; AUDIT=Alcohol Use Disorders Identification Test; eGFR=estimated glomerular filtration rate; LV=left ventricular; LA=left atrial; BSA=body surface area; ART=antiretroviral therapy. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 29 TABLE 2. MESA external validation participant characteristics (2010-2012) CHARACTERISTIC Median [IQR] or % (n) Cross-Sectional LAVi Analysis (n=1242) Incident Event Analysis (n=2273) Demographics Age, years 67 [60, 74] 67 [61, 75] Female sex at birth 641 (52%) 1182 (52%) Race and ethnicity --- --- Black, non-Hispanic 282 (23%) 550 (24%) White, non-Hispanic 546 (44%) 950 (42%) Hispanic 230 (18%) 444 (19%) Chinese 184 (15%) 329 (15%) Education level ≥ high school 1105 (89%) 2002 (88%) Clinical Factors Smoking status --- --- Current 82 (7%) 156 (7%) Former 540 (44%) 1016 (45%) Never 620 (50%) 1101 (48%) Pack-years of smoking 0 [0, 11] 0 [0, 12] Body mass index, kg/m2 27.3 [24.3, 30.7] 27.5 [24.4, 31.3] Hypertension a 665 (54%) 1251 (55%) Systolic blood pressure, mmHg 119 [108, 135] 119 [109, 136] Diastolic blood pressure, mmHg 69 [62, 75] 69 [62, 75] Blood pressure-lowering medication use 622 (50%) 1157 (51%) Dyslipidemia b 911 (73%) 1687 (74%) Total cholesterol, mg/dL 183 [159, 208] 182 [158, 208] HDL-cholesterol, mg/dL 53 [44, 63] 53 [44, 64] Lipid-lowering medication use 454 (37%) 865 (38%) Diabetes c 236 (19%) 466 (21%) Diabetes medication use 15 (13%) 29 (12%) eGFR, CKD-EPI, mL/min/1.73m2 84 [71, 94] 83 [70, 94] Cardiovascular Magnetic Resonance LV ejection fraction, % 62.7 [58.1, 66.9] --- LV ejection fraction <50% 41 (3%) --- LV mass indexed by BSA, mg/m2 64.2 [56.3, 73.7] --- LV end-diastolic volume indexed by BSA, mL/m2 65.0 [57.1, 73.2] --- LA volume indexed by BSA, mL/m2 33.8 [27.2, 41.1] --- LA volume indexed by BSA ≥40 mL/m2 357 (29%) --- LA volume indexed by BSA ≥53 mL/m2 75 (6%) --- Late gadolinium enhancement d 70 (9%) --- Extracellular volume fraction, % e 26.8 [24.6, 29.0] --- All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 30 a Hypertension is defined as use of antihypertensive medications or systolic blood pressure≥140 mmHg or diastolic blood pressure ≥90 mmHg. b Dyslipidemia is defined as use of lipid-lowering medications or fasting total cholesterol level ≥200 mg/dL or low-density lipoprotein cholesterol level ≥130 mg/dL or high-density lipoprotein cholesterol level <40 mg/dL or serum triglyceride level ≥150 mg/dL. c Diabetes is defined as use of hypoglycemic medication or fasting serum glucose levels ≥126 mg/dL. d Measured in a subgroup who received late gadolinium enhancement (n=762). e Assessed in a subgroup who received T1 mapping (n=258). IQR=interquartile range, reported as (25th, 75th) percentiles; LV=left ventricular; LA=left atrial; BSA=body surface area; HDL=high-density lipoprotein; eGFR=estimated glomerular filtration rate. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 31 TABLE 3. Summary of cross-sectional associations between identified HIV-associated proteomic signature of left atrial size and indexed left atrial volume by age among people without HIV in MESA (n=1242) Analysis Sample Mean (SD) Age Proteins FDR<0.05 All participants 68 (9) CKB, CLEC14A, CNTN3, COL4A1, IGFBP2, NOTCH3, NT-proBNP, PDGFRA Participants ≥ median age 75 (5) B3GNT7, CLEC14A, COL4A1, CRIM1, CX3CL1, DCTPP1, EPHA2, LTBP2, NOTCH3, NT-proBNP, PDGFRA, SCARF2, THBS2, THBS4, TIMP1, VCAM1, Brown protein cluster Participants < median age 60 (4) NT-proBNP Bold indicates protein × continuous age interaction at FDR<0.05 evaluated at exam 5 (2010- 2012, median age 67 years). Brown protein cluster comprises 42 proteins and is described in Figure 4. Complete modeling results can be found in Supplemental Tables S14-S15. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 32 FIGURE 1. Study overview Images generated using BioRender. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 33 FIGURE 2. Proteomic signature of HIV seropositivity among people living with and without HIV in the United States (n=352). (A) Volcano plot of mean differences in standardized plasma protein abundance comparing PLWH to PWOH vs. -log10 false discovery rate (FDR; Benjamini-Hochberg) estimated using linear regression with robust variance, adjusting for age, sex, race/ethnicity, education, pack-years of smoking, hazardous alcohol use, stimulant use in the prior 5 years, opioid use in the prior 5 years, hepatitis C infection, and estimated glomerular filtration rate. Threshold for significance is indicated by gray dotted line, FDR<0.05; 415 of 2596 proteins positively associated with positive HIV serostatus and 24 proteins inversely associated; comparing suppressed PLWH vs. PWOH, 414 of these 439 total proteins were associated with suppressed positive HIV serostatus (not depicted). Complete modeling results can be found in Supplemental Table S3. (B) Enriched biological processes among proteins associated with HIV serostatus, estimated using Fisher’s exact test, Gene Ontology: Biological Processes reference database, and a threshold for significance of FDR<0.05. Protein ratio=proportion of proteins significantly associated with HIV serostatus (n=439) that map to given annotation. Proteins mapping to each annotation can be found in Supplemental Table S4. PLWH=persons living with HIV; PWOH=persons without HIV; suppressed HIV viral load=HIV RNA<50 copies/µL All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 34 FIGURE 3. HIV-associated proteomic signature of left atrial size among people living with and without HIV in the United States (n=352). All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 35 (A) Volcano plot of mean differences in left atrial volume indexed for body surface area (LAVi) per standard deviation (SD) increment in plasma abundance of 2594 individual proteins vs. -log10 false discovery rate (FDR; Benjamini-Hochberg) estimated using linear regression with robust variance. Purple indicates proteins only associated when adjusting for age, sex, race/ethnicity, HIV, hepatitis C (HCV) infection, and estimated glomerular filtration rate (eGFR). Orange indicates proteins associated with further adjustment for education, body mass index (BMI), systolic blood pressure (SBP), anti-hypertensive medication, diabetes, dyslipidemia, current hazardous alcohol use, pack-years of smoking in prior 5 years, stimulant use in prior 5 years, and opioid use in prior 5 years. The threshold for significance is indicated by gray dotted line, FDR<0.05. Most saturated model (orange) yielded 69 proteins positively associated with LAVi and 4 proteins inversely associated. Complete modeling results can be found in Supplemental Table S9. (B) Enriched biological processes among proteins associated with both positive HIV serostatus and higher LAVi, same directionality, estimated using Fisher’s exact test, Gene Ontology: Biological Processes reference database, and a threshold for significance of FDR<0.05. Protein ratio=proportion of total HIV- and LAVi-associated proteins (n=73) that map to given annotation. Proteins mapping to each annotation can be found in Supplemental Table S10. TNF=tumor necrosis factor; NK=natural killer; PDGF=platelet derived growth factor. (C) Beta-beta plot of mean differences in LAVi per SD increment in individual plasma protein abundances vs. mean differences in standardized protein abundances comparing persons living with vs. without HIV (PLWH, PWOH). All associations were estimated using linear regression with robust variance, and proteins depicted (n=73) are restricted to those significantly associated with both parameters with a FDR<0.05. HIV point estimates (y-axis) were adjusted for age, sex, race/ethnicity, education, current hazardous alcohol use, pack- years of smoking in prior 5 years, stimulant use in prior 5 years, opioid use in prior 5 years, HCV, and eGFR. LAVi point estimates (x-axis) were adjusted for the same covariates, as well as BMI, SBP, anti-hypertensive medication, dyslipidemia, and diabetes. (D) Mean differences in LAVi per SD increment in individual plasma protein abundances among strata of PLWH and PWOH, estimated using linear regression with robust variance adjusting for age, sex, race/ethnicity, education, BMI, SBP, anti-hypertensive medication, dyslipidemia, diabetes, current hazardous alcohol use, pack-years of smoking in prior 5 years, stimulant use in prior 5 years, opioid use in prior 5 years, HCV, and eGFR. Proteins depicted are limited to those with HIV×protein multiplicative interactions with a FDR<0.10 (n=73 proteins tested). All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 36 FIGURE 4. Relationship between an HIV-associated, agnostically defined cluster of plasma proteins and left atrial size among people living with and without HIV in the United States (n=352). All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 37 (A) Mean difference in indexed left atrial volume (LAVi) per standard deviation (SD) increment in ‘Brown’ protein cluster plasma abundance among all participants, among strata of participants living with and without HIV, and among strata of participants above and below median age in SMASH (55 years), estimated using linear regression with robust variance adjusting for age, sex, race/ethnicity, education, body mass index, systolic blood pressure, anti-hypertensive medication, dyslipidemia, diabetes, current hazardous alcohol use, pack-years of smoking in prior 5 years, stimulant use in prior 5 years, opioid use in prior 5 years, hepatitis C infection, and estimated glomerular filtration rate. (B) Enriched biological processes among 42 proteins comprising the ‘Brown’ protein cluster, estimated using Fisher’s exact test, Gene Ontology: Biological Processes reference database, and a threshold for significance of Benjamini-Hochberg false discovery rate (FDR)<0.05. Protein ratio=proportion of total ‘Brown’ cluster proteins (n=42) that map to given annotation. Proteins mapping to each annotation can be found in Supplemental Table S8. (C) Interaction network of 42 proteins comprising the ‘Brown’ protein cluster generated using STRING, a public database of known and predicted protein-protein interactions. Interactions include direct (physical) and indirect (functional) associations derived from computational prediction, knowledge transfer between organisms, and interactions aggregated from other databases. Line thickness indicates strength of data support. Szklarczyk D, et al. The STRING database in 2023: protein–protein association networks and functional enrichment analyses for any sequenced genome of interest. Nucleic Acids Res. 2023;51(D1):D638-64 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 38 FIGURE 5. Association between identified HIV-associated proteomic signature of left atrial size and time to incident adjudicated clinical heart failure among people without HIV in MESA (n=2273). a Estimated using Cox proportional hazards regression, adjusting for study site, age, sex, race/ethnicity, education, body mass index, systolic blood pressure, anti-hypertensive medication, dyslipidemia, diabetes, smoking, and estimated glomerular filtration rate. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 39 b Estimated among those with vs. without dichotomous characteristic or per standard deviation (SD) increment in continuous characteristic using linear regression, adjusting for study site. Proteome features displayed are limited to those with multivariable adjusted associations with HIV serostatus and LAVi in SMASH and time to incident clinical heart failure in the MESA. Complete modeling results can be found in Supplemental Tables S17 and Supplemental Figure S7. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 1 Proteomic Signature of HIV-Associated Subclinical Left Atrial Remodeling and Incident Heart Failure Supplemental Materials Tess E Peterson, Virginia S Hahn, Ruin Moaddel, Min Zhu, Sabina A Haberlen, Frank J Palella, Michael Plankey, Joel S Bader, Joao AC Lima, Robert E Gerszten, Jerome I Rotter, Stephen S Rich, Susan R Heckbert, Gregory D Kirk, Damani A Piggott, Luigi Ferrucci, Joseph B Margolick, Todd T Brown, Katherine C Wu, Wendy S Post Table of Contents Table/Figure Page Detailed Methods. 3 TABLE S1. SMASH participant characteristics by inclusion vs. exclusion from proteomics study. 6 TABLE S2. SMASH participant characteristics by parent cohort. 8 TABLE S3. Cross-sectional associations between plasma protein abundances and HIV serostatus among (A) PLWH and PWOH and (B) PLWH with undetectable plasma HIV RNA and PWOH in SMASH. 10 TABLE S4. Proteins cross-sectionally associated with HIV serostatus corresponding to statistically over-represented biological processes. 11 TABLE S5. Individual proteins within each of six clusters agnostically defined using weighted gene co-expression network analysis. 12 TABLE S6. Enriched biological processes and pathways within protein clusters agnostically defined using weighted gene co-expression network analysis. 13 TABLE S7A. Cross-sectional associations between agnostically defined clusters of proteins and HIV serostatus among PLWH and PWOH in SMASH. 16 TABLE S7B. Cross-sectional associations between agnostically defined clusters of proteins and HIV serostatus among PLWH with undetectable plasma HIV RNA and PWOH in SMASH. 16 TABLE S8. Proteins in Brown cluster defined using weighted gene co-expression network analysis corresponding to statistically over-represented biological processes. 17 TABLE S9. Cross-sectional associations between HIV-associated plasma protein abundances and indexed left atrial volume among PLWH and PWOH in SMASH. 17 TABLE S10. Proteins cross-sectionally associated with HIV serostatus and indexed left atrial volume corresponding to statistically over-represented biological processes. 18 TABLE S11. Percent difference in association between HIV serostatus and indexed left atrial volume (LAVi) with adjustment for plasma abundance of 73 individual candidate protein contributors in SMASH. 19 TABLE S12. Difference in association between plasma abundance of 73 individual proteins of interest and indexed left atrial volume (LAVi) by HIV serostatus in SMASH. 21 TABLE S13. Difference in association between plasma abundance of 73 individual proteins of interest and indexed left atrial volume (LAVi) by age group in SMASH. 23 TABLE S14. Cross-sectional associations between proteins of interest and indexed left atrial volume (LAVi) in the Multi-Ethnic Study of Atherosclerosis (median age 67 years). 25 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 2 TABLE S15. Cross-sectional associations between proteins of interest and indexed left atrial volume (LAVi) by age group in the Multi-Ethnic Study of Atherosclerosis (median age 67 years). 28 TABLE S16. Association between proteins of interest and incident atrial fibrillation in the Multi- Ethnic Study of Atherosclerosis (median age 67 years at the beginning of follow-up). 30 TABLE S17. Association between proteins of interest and incident clinical heart failure in the Multi-Ethnic Study of Atherosclerosis (median age 67 years at the beginning of follow-up). 33 FIGURE S1A. Flow diagram of SMASH study participants included in analysis sample. 36 FIGURE S1B. Flow diagram of MESA study participants included in cross-sectional analysis sample. 37 FIGURE S1C. Flow diagram of MESA study participants included in longitudinal analysis samples. 38 FIGURE S2. Clusters of proteins agnostically defined using weighted gene co-expression network analysis. 39 FIGURE S3. Spearman’s correlation between plasma abundances of 73 proteins independently associated with both HIV seropositivity and incremental indexed left atrial volume, same directionality, among PLWH and PWOH in SMASH. 40 FIGURE S4. Interaction network of 73 proteins independently associated with both HIV seropositivity and incremental indexed left atrial volume, same directionality in SMASH. 41 FIGURE S5. Association between identified HIV-associated proteomic signature of left atrial size and clinical characteristics in SMASH. 42 FIGURE S6. Volcano plot of association between identified HIV-associated proteomic signature of left atrial size and time to incident atrial fibrillation in the Multi-Ethnic Study of Atherosclerosis. 44 FIGURE S7. Volcano plot of associations between identified HIV-associated proteomic signature of left atrial size and time to incident adjudicated clinical heart failure in the Multi- Ethnic Study of Atherosclerosis. 45 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 3 SUPPLEMENTAL METHODS SMASH (Discovery) Parent Cohort Details The Multicenter AIDS Cohort Study (MACS) enrollment began in 1984 and occurred at 4 U.S. sites (Baltimore, MD/Washington, DC; Chicago, IL; Pittsburgh, PA; and Los Angeles, CA) over 4 enrollment waves: 1984-1985, 1987-1991, 2001-2003 and 2010-2018. The Women’s Interagency HIV Study (WIHS) enrollment began in 1994 and recruited women with and without HIV in 4 waves: 1994-1995, 2001-2002, and 2011-2012 from 10 cities (6 original cohort sites were in Brooklyn, NY; the Bronx/Manhattan, NY; Washington, DC; Chicago, IL; San Francisco, CA; and Los Angeles, CA; 4 southern sites were added in 2013: Chapel Hill, NC; Atlanta, GA; Birmingham, AL/Jackson, MS; and Miami, FL. The Los Angeles site discontinued active follow-up in 2013). MACS and WIHS have since become the MACS/WIHS Combined Cohort Study (https://statepi.jhsph.edu/mwccs/). AIDS Linked to the Intravenous Experience (ALIVE) enrollment began in 1988. Additional cohort recruitment occurred during 1994–1995, 1998, 2000, and 2005–2008. SMASH (Discovery) Covariate Definitions Body mass index: calculated as weight in kilograms divided by the square of height in meters. Chronic active hepatitis C virus infection: positive antibody and detectable hepatitis C virus ribonucleic acid (RNA). Diabetes: use of hypoglycemic medications or fasting serum glucose levels ≥126 mg/dL closest to the time of CMR study visit. Hemoglobin A1C level <6.5% was used to exclude diabetes if fasting glucose levels were not available. Dyslipidemia: use of lipid-lowering medications or fasting total cholesterol level ≥200 mg/dL or low- density lipoprotein (LDL) cholesterol level ≥130 mg/dL or high-density lipoprotein (HDL) cholesterol level <40 mg/dL or serum triglyceride level ≥150 mg/dL closest to the time of CMR study visit. Educational level: did or did not attain a high school diploma. Estimated glomerular filtration rate (eGFR): calculated using the CKD Epi (2021) equation. Hazardous alcohol use: a score of >8 on the Alcohol Use Disorders Identification Test (AUDIT) assessed at the time of CMR. History of cardiovascular disease: prior myocardial infarction, angioplasty, stent, coronary artery bypass surgery, other heart surgery, or heart failure confirmed by medical records. Hypertension: use of antihypertensive medications or systolic blood pressure≥140 mmHg or diastolic blood pressure ≥90 mmHg averaged over the preceding 5 years when available or at time of cardiac magnetic resonance imaging (CMR) study visit. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 4 Opioid use: included any of the following by any administration route over the 5 years preceding CMR: methadone (including both prescribed and non-prescribed), heroin, speedball, opioid-based pain medications, prescribed and non-prescribed. Pack-years of smoking: assessed over the 5 years preceding CMR. Stimulant use: included any of the following by any administration route over the 5 years preceding CMR: cocaine including crack, speedball, crystal methamphetamine, Phenmetrazine (Preludin), benzedrine, methadrine, uppers, speed, methylphenidate (Ritalin), (dextroamphetamine) Dexedrine, amphetamine/dextroamphetamine (Adderall). MESA (External Validation) Cohort Details and Methods Elements of the primary analysis were externally validated in the Multi-Ethnic Study of Atherosclerosis (MESA) (https://www.mesa-nhlbi.org/). The MESA is a prospective population- based cohort initiated in 2000 to study the characteristics and progression of subclinical cardiovascular disease (CVD) among a diverse population in the United States. MESA recruited a total of 6814 participants from six geographic areas across the U.S.—Baltimore City and Baltimore County, Maryland; Chicago, Illinois; Forsyth County, North Carolina; Los Angeles County, California; Northern Manhattan and the Bronx, New York; and St. Paul, Minnesota. At the time of enrollment, participants were 45-84 years of age and had no history of clinical CVD—including coronary artery disease, peripheral vascular disease, cerebrovascular disease, and heart failure.1 The study protocol was approved by Institutional Review Boards of Columbia University, Johns Hopkins University, Northwestern University, University of California Los Angeles, University of Minnesota, and Wake Forest University; and all participants signed informed consent. The present study utilized data on a subset of MESA participants from cohort exam 5 (2010-2012) with follow-up observed through December 31, 2019. CMR imaging was performed at exam 5 using 1.5-T scanners (Magnetom Avanto and Magnetom Espree, Siemens Medical Systems, Erlangen, Germany) with six-channel anterior and posterior phased-array torso coil elements. The protocol has been described in detail previously (doi:10.1016/j.jcmg.2021.02.014) and included acquisition of one cine horizontal long-axis section (four-chamber view), at least 12 cine short-axis sections from the atria to the cardiac apex, and one cine vertical long-axis section (two-chamber view) using a steady-state free precession pulse sequence. Proteomics was performed using the Olink Explore 3072 assay on EDTA plasma stored at -80°C using the same standardized laboratory protocol and data quality control, normalization, and calibration methods described in the primary methods and in more detail below. This data was generated through the Trans-Omics for Precision Medicine (TOPMed) program in the laboratory of Dr. Robert Gerzsten. Information on hospital admissions, outpatient diagnoses, and deaths were ascertained at follow-up telephone interviews conducted every 9-12 months with the participant or a proxy. Medical records and International Classification of Disease (ICD) diagnosis codes were obtained for inpatient and outpatient events, and death certificates were obtained for all deaths. Medical records for reported heart failure (HF) events were independently reviewed by two physicians. The probable diagnosis of HF required a physician diagnosis, signs or symptoms of HF, and medical treatment for HF. The definite diagnosis of HF also required ≥1 criterion such as All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 5 pulmonary congestion or edema by chest X-ray; reduced left ventricular (LV) function or dilated LV by echocardiography or ventriculography; or evidence of LV diastolic dysfunction. We included both probable and definite diagnosis of HF. End of follow-up for HF event ascertainment was December 31, 2019. Incident cases of atrial fibrillation (AF) during the follow-up period were identified through MESA event surveillance and, for participants enrolled in fee-for-service Medicare, from inpatient and outpatient Medicare claims data. AF was documented as present if an International Classification of Diseases, Ninth Revision diagnosis code 427.31 (AF) or 427.32 (atrial flutter) was present. End of follow-up for AF event ascertainment was December 31, 2018. Olink Proximity Extension Assay and Data Quality Control The Olink Explore 3072 assay was used in both SMASH and MESA. Olink technology has been described in detail previously (doi:10.1371/journal.pone.0095192). Briefly, DNA oligonucleotide- labeled antibody pairs bind target antigen in solution and, when bound pairwise, hybridize and are extended by a DNA polymerase. This forms a new DNA barcode, which is then amplified and quantified by microfluidic quantitative PCR. This technology differs from standard immunoassays in that it does not rely on spectrometry for quantification, and only matched DNA reporter pairs will be amplified, leading to high specificity by mitigating cross-reactive binding. Data quality control, normalization, and calibration were carried out using the Normalized Protein eXpression (NPX) software at probe, sample, and plate levels. Specifically, data was normalized using multiple internal assay controls used to monitor immunoreaction, extension, and readout steps, and inter-plate sample controls were used to further normalize for inter-plate variation. The Olink Explore 3072 platform assays 2924 total proteins. Proteins were excluded from analysis if they did not meet Olink batch release standards (n=54), did not pass internal or external protein quality control standards (n=37), or were detected in <50% of samples (n=239). Weighted Gene Co-expression Network Analysis (WGCNA) We derived unique clusters of highly correlated proteins using the WGCNA package in R (version 4.2) (doi:10.1002/0471142727.mb2010s109, doi:10.1186/1471-2105-9-559). Briefly, unsigned weighted networks of proteins were constructed using pairwise bi-weight midcorrelations of plasma abundance values and a soft threshold transformation that produced an approximately scale-free topology. Next, the topological overlap matrix was calculated and used to hierarchically cluster proteins. A summary plasma abundance measure for each cluster, the eigenprotein, was then derived from the within-network expression correlation matrix using principal component analysis and was used in downstream modeling. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 6 SUPPLEMENTAL TABLE S1. SMASH participant characteristics by inclusion vs. exclusion from proteomics study. CHARACTERISTIC Median [IQR] or % (n) p OLINK (n=352) No OLINK (n=116) Demographics Age, years 55 [51, 58] 55 [50, 59] 0.671 Female sex at birth 25.3% (89) 43.1% (50) 0.0004 Race and ethnicity --- --- Black, non-Hispanic 69.6% (245) 77.6% (90) 0.190 White, non-Hispanic 24.4% (86) 16.4% (19) Hispanic 6% (21) 6% (7) Education level ≥ high school 75% (264) 57% (66) 0.0003 Annual income ≥$10,000 60.8% (205) 48.2% (53) 0.026 Substance Use Current smoker 48.6% (171) 47.4% (55) 0.912 Pack-years of smoking a 0.77 [0.00, 2.67] 0.81 [0.00, 2.72] 0.849 Hazardous alcohol use (AUDIT score >8) 12.2% (43) 9.5% (11) 0.528 Opioid use a 30.7% (108) 37.9% (44) 0.183 Stimulant use a 38.9% (137) 42.2% (49) 0.600 Clinical Factors History of CVD 6.2% (22) 6.0% (7) 1.000 Body mass index, kg/m2 26.1 [23.5, 30.6] 28.2 [24.1, 31.8] 0.014 Hypertension 51.7% (182) 72.4% (84) 0.0001 Systolic blood pressure, mmHg 126 [119, 135] 126 [120, 137] 0.311 BP-lowering medication use 34.9% (123) 44.8% (52) 0.072 Dyslipidemia 58.8% (207) 58.6% (68) 1.000 Total cholesterol, mg/dL 174 [150, 198] 173 [152, 198] 0.975 HDL-cholesterol, mg/dL 53 [43, 66] 52 [42, 65] 0.598 Lipid-lowering medication use 24.1% (85) 21.6% (25) 0.656 Diabetes 12.2% (43) 15.5% (18) 0.449 Diabetes medication use 9.1% (32) 14.7% (17) 0.128 eGFR, CKD-EPI, mL/min/1.73m2 93 [76, 105] 85 [68, 100] 0.003 Hepatitis C diagnosis 22.4% (78) 39.3% (44) 0.0007 Cardiovascular Magnetic Resonance LV ejection fraction, % 72.3 [67.6, 75.7] 73.1 [69.0, 76.3] 0.194 LV mass indexed by BSA, mg/m2 61.8 [56.4, 68.3] 59.8 [54.3, 65.2] 0.077 LV end-diastolic volume indexed by BSA, mL/m2 65.8 [56.9, 76.6] 66.6 [54.4, 75.1] 0.323 LA volume indexed by BSA, mL/m2 27.3 [22.3, 35.1] 29.5 [23.1, 37.6] 0.200 LA volume indexed by BSA ≥40 mL/m2 11.4% (40) 18.2% (16) 0.127 Late gadolinium enhancement 36.1% (127) 16% (8) 0.008 Extracellular volume fraction, % 28.6 [26.1, 30.4] 29.2 [27.4, 31.5] 0.570 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 7 HIV-Related Factors HIV diagnosis 60.8% (214) 68.1% (79) 0.194 HIV viral load detectable (>50 RNA copies/mL) 25.7% (55) 35.4% (28) 0.135 CD4+ T cell count, cells/µL 605 [400, 815] 613 [443, 813] 0.644 CD4+ nadir T cell count, cells/µL 271 [149, 386] 261 [141, 483] 0.929 History of AIDS diagnosis 20.4% (29) 23.9% (11) 0.768 ART use 88.3% (188) 87.3% (69) 0.990 Duration of ART, years 12.8 [5.4, 16.0] 9.7 [3.6, 15.0] 0.134 Protease inhibitor base 35.2% (75) 36.7% (29) 0.921 NNRTI base 26.8% (57) 25.3% (20) 0.921 Integrase strand inhibitor base 24.9% (53) 25.3% (20) 1.000 Other ART base 1.4% (3) 0% (0) 0.684 a Reported in the five years preceding cardiovascular magnetic resonance imaging. PLWH=persons living with HIV; PWOH=persons without HIV; IQR=interquartile range, reported as (25th, 75th) percentiles; AUDIT=Alcohol Use Disorders Identification Test; BSA=body surface area; eGFR=estimated glomerular filtration rate; ART=antiretroviral therapy. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 8 SUPPLEMENTAL TABLE S2. SMASH participant characteristics by parent cohort (n=352). CHARACTERISTIC Median [IQR] or % (n) MACS (n=166) WIHS (n=60) ALIVE (n=126) Demographics Age, years 55 [51, 59] 54 [51, 57] 55 [52, 58] Female sex at birth 0% (0) 100% (60) 23% (29) Race/ethnicity --- --- --- Black, non-Hispanic 42% (71) 88% (53) 96% (121) White, non-Hispanic 48% (79) 7% (4) 2% (3) Hispanic 10% (16) 5% (3) 2% (2) Education level ≥ high school 92% (152) 85% (51) 48% (61) Annual income ≥$10,000 81% (124) 59% (34) 38% (47) Substance Use Smoking status --- --- --- Current 31% (51) 52% (31) 71% (89) Former 43% (71) 23% (14) 22% (28) Never 26% (44) 25% (15) 7% (9) Pack-years of smoking a 0.00 [0.00, 1.11] 0.75 [0.00, 1.74] 2.58 [0.79, 3.67] Hazardous alcohol use (AUDIT score >8) 12% (20) 12% (7) 13% (16) Opioid use a 16% (26) 8% (5) 61% (77) Stimulant use a 31% (52) 25% (15) 56% (70) Clinical Factors History of cardiovascular disease 4% (6) 15% (9) 6% (7) Body mass index, kg/m2 25.9 [23.8, 28.9] 28.6 [25.0, 33.4] 25.5 [22.6, 29.9] Hypertension 45% (75) 48% (29) 62% (78) Systolic blood pressure, mmHg 127 [120, 135] 121 [114, 126] 127 [120, 136] Diastolic blood pressure, mmHg 79 [74, 83] 77 [72, 79] 84 [78, 89] Blood pressure-lowering medication use 31% (51) 35% (21) 40% (51) Dyslipidemia 69% (115) 55% (33) 47% (59) All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 9 Total cholesterol, mg/dL 178 [155, 203] 175 [152, 206] 166 [144, 192] HDL-cholesterol, mg/dL 51 [42, 61] 60 [49, 70] 54 [42, 70] Lipid-lowering medication use 33% (55) 20% (12) 14% (18) Diabetes 11% (18) 17% (10) 12% (15) Diabetes medication use 8% (13) 12% (7) 10% (12) eGFR, CKD-EPI, mL/min/1.73m2 97 [77, 105] 83 [69, 100] 91 [75, 106] Hepatitis C diagnosis 10% (16) 9% (5) 46% (57) Cardiovascular Magnetic Resonance LV ejection fraction, % 71.8 [66.6, 74.6] 75.5 [70.2, 78.6] 72.2 [67.3, 75.4] LV mass indexed by BSA, mg/m2 63.0 [59.3, 69.1] 51.4 [46.9, 57.5] 63.9 [57.6, 68.7] LV end-diastolic volume indexed by BSA, mL/m2 67.8 [58.4, 77.0] 59.8 [51.3, 69.0] 67.0 [59.6, 79.2] LA volume indexed by BSA, mL/m2 26.9 [21.8, 34.0] 26.9 [22.8, 34.9] 28.5 [22.9, 35.8] LA volume indexed by BSA ≥40 mL/m2 9% (15) 10% (6) 15% (19) Late gadolinium enhancement 37.3% (62) 31.7% (19) 36.5% (46) Extracellular volume fraction, % 27.7 [25.5, 29.7] 29.7 [28.3, 32.2] 29.0 [26.8, 30.8] HIV-Related Factors HIV diagnosis 62% (103) 65% (39) 57% (72) HIV viral load detectable (>50 RNA copies/mL) 14% (14) 13% (5) 50% (36) CD4+ T-cell count, cells/µL 694 [492, 944] 631 [476, 819] 420 [286, 641] CD4+ nadir T-cell count, cells/µL 335 [226, 510] 237 [130, 318] 220 [116, 311] History of AIDS diagnosis 14% (14) 39% (15) NA ART use 88% (91) 92% (36) 86% (61) Duration of ART, years 13.2 [6.1, 15.7] 11.0 [4.6, 16.3] NA Protease inhibitor base 32% (33) 23% (9) 46% (33) Non-nucleoside reverse transcriptase inhibitor base 32% (33) 33% (13) 16% (11) Integrase strand inhibitor base 24% (25) 36% (14) 20% (14) Other ART base 0% (0) 0% (0) 4% (3) a Reported in the five years preceding cardiovascular magnetic resonance imaging. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 10 PLWH=persons living with HIV; PWOH=persons without HIV; IQR=interquartile range, reported as (25 th, 75th) percentiles; AUDIT=Alcohol Use Disorders Identification Test; BSA=body surface area; eGFR=estimated glomerular filtration rate; ART=antiretroviral therapy. SUPPLEMENTAL TABLE S3. Cross-sectional associations between plasma protein abundances and HIV serostatus among (A) PLWH and PWOH and (B) PLWH with undetectable plasma HIV RNA and PWOH in the United States. See Supplemental Excel file ‘Supplemental Table S3.xlsx’ Mean differences in standardized protein abundance (A) comparing PLWH to PWOH; and (B) comparing PLWH with plasma HIV RNA <50 copies/mL to PWOH, estimated using linear regression with robust variance adjusting for age, sex, race/ethnicity, education, pack-years of smoking, hazardous alcohol use, stimulant use in the prior 5 years, opioid use in the prior 5 years, hepatitis C, and estimated glomerular filtration rate. FDR=Benjamini-Hochberg false discovery rate. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 11 SUPPLEMENTAL TABLE S4. Proteins cross-sectionally associated with HIV serostatus corresponding to statistically over-represented biological processes. Gene Ontology: Biological Process Enrichment FDR a Protein Count Proteins Cytokine production 0.005 67 ACE2, ADGRG1, AXL, B2M, BST2, BTN2A1, BTN3A2, CCN4, CD14, CD160, CD244, CD274, CD28, CD4, CD40, CD46, CD6, CD74, CD80, CD83, CLEC4A, COL3A1, CRTAM, CX3CL1, CXCL6, DLL1, EPHA2, F11R, F2R, FCGR2B, FLT4, HAVCR2, HGF, IFNL1, IL12RB1, IL18, IL1R1, IL1RL2, IL6ST, IL7, INHBB, KLRK1, LAG3, LGALS9, LILRA2, LILRB4, LY9, MDK, NOS2, PDCD1LG2, PGLYRP2, PLA2G10, PLA2G1B, PTPRC, SEMA7A, SLAMF1, SLAMF6, SPON2, TIGIT, TNFRSF14, TNFRSF1B, TNFRSF21, TNFRSF8, TRIM21, VSIG4, VSIR, XCL1 T cell activation, differentiation, and migration 6.23E-05 58 ADA, B2M, CCL2, CD160, CD27, CD274, CD28, CD300A, CD4, CD46, CD48, CD6, CD7, CD74, CD80, CD83, CD8A, CLEC4A, CRTAM, CTSL, FCGR2B, FGL1, HAVCR2, HLA-DRA, ICAM1, IFNL1, IGFBP2, IL12RB1, IL18, IL1RL2, IL2RA, IL6ST, IL7, ITGAL, KLRK1, LAG3, LGALS9, LILRB4, LY9, MDK, PDCD1LG2, PTPRC, SIRPB1, SLAMF1, SLAMF6, SLAMF7, TGFBR2, TIGIT, TNFRSF14, TNFRSF1B, TNFRSF21, TNFRSF4, TNFRSF9, TNFSF13B, VCAM1, VSIG4, XCL1, VSIR Regulation of MAPK cascade 0.029 56 ACE2, ADAM9, AIDA, ARHGEF5, BMPER, CCL11, CCL14, CCL15, CCL17, CCL2, CCL20, CCL23, CCL25, CCL4, CCL8, CD27, CD300A, CD4, CD40, CD74, CDH2, CX3CL1, EDA2R, EPHA2, F2R, FAS, FCGR2B, FCRL3, FLT4, GDF15, GH1, GPR37, HAVCR2, HGF, ICAM1, IGFBP3, IGFBP4, LGALS9, LILRB4, LTBR, MARCO, MYDGF, NOTCH1, PDGFRA, PLA2G1B, PTPRC, REN, RNF149, ROBO1, SEMA7A, SLAMF1, SMPD1, TNFRSF11A, TNFRSF19, VEGFA, XCL1 Angiogenesis 0.029 52 ACVRL1, ADGRG1, AMOT, ANGPT2, B4GALT1, BMPER, BSG, CCL11, CCL2, CCN3, CD160, CD40, CDH5, CLEC14A, COL18A1, COL4A1, CX3CL1, CXCL10, CXCL13, CXCL8, DLL1, EPHA2, EPHB4, FLT4, GRN, HGF, HS6ST1, HSPG2, HYAL1, IL18, MDK, MYDGF, NOS3, NOTCH1, NOTCH3, NRCAM, NRP2, PDGFRA, PGF, PTPRM, ROBO1, RSPO3, SCG2, SMOC2, TGFBR2, THBS2, THBS4, TIE1, TNFRSF12A, TYMP, VEGFA, VEGFC Mononuclear cell migration 1.69E-05 37 ARHGEF5, CCL11, CCL14, CCL15, CCL17, CCL2, CCL20, CCL23, CCL25, CCL27, CCL4, CCL8, CCN3, CRTAM, CSF1, CX3CL1, CXCL10, CXCL11, CXCL12, CXCL13, CXCL16, F11R, ICAM1, ITGAL, ITGB7, JAM2, KLRK1, LGALS9, LGMN, MDK, MSTN, NBL1, SLAMF1, SLAMF8, TNFRSF11A, TNFRSF14, XCL1 Response to tumor necrosis factor 4.34E-04 34 ADAM9, ASAH1, CCL11, CCL14, CCL15, CCL17, CCL2, CCL20, CCL23, CCL25, CCL4, CCL8, CD14, CD40, CX3CL1, CXCL16, CXCL8, EDA2R, FAS, HYAL1, IL18BP, KRT18, SMPD1, TNFRSF11A, TNFRSF14, TNFRSF17, TNFRSF19, TNFRSF1A, TNFRSF1B, TNFRSF21, TNFRSF4, TNFSF13B, VCAM1, XCL1 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 12 Granulocyte migration 1.16E-04 32 ADGRE2, BSG, CCL11, CCL14, CCL15, CCL17, CCL2, CCL20, CCL23, CCL25, CCL4, CCL8, CD300A, CD74, CSF1, CX3CL1, CXCL10, CXCL11, CXCL13, CXCL6, CXCL8, CXCL9, IL1R1, ITGB2, MDK, MSTN, PLA2G1B, SCG2, SLAMF1, SLAMF8, THBS4, XCL1 Viral life cycle 0.035 30 ACE2, ATG16L1, AXL, BSG, BST2, CCL2, CCL8, CD28, CD4, CD46, CD74, CD80, CTSL, CXCL8, EPHA2, F11R, HAVCR1, ICAM1, ITGB7, LGALS9, NECTIN2, NOTCH1, P4HB, SCARB2, SIGLEC1, SLAMF1, SMPD1, TNFRSF14, TNFRSF4, TRIM21 Regulation of ERK1/ERK2 cascade 0.037 27 BMPER, CCL11, CCL14, CCL15, CCL17, CCL2, CCL20, CCL23, CCL25, CCL4, CCL8, CD4, CD74, CX3CL1, F2R, FLT4, HAVCR2, ICAM1, LGALS9, MARCO, NOTCH1, PDGFRA, PTPRC, SEMA7A, SLAMF1, TNFRSF11A, XCL1 Response to IL-1 4.03E-04 23 CCL11, CCL14, CCL15, CCL17, CCL2, CCL20, CCL23, CCL25, CCL4, CCL8, CD38, CD40, CX3CL1, CXCL8, HYAL1, IL1R1, IL1RL2, INHBB, LGALS9, SMPD1, TNFRSF11A, XCL1, ZBP1 αβ T cell activation and migration 0.001 23 ADA, CD160, CD274, CD28, CD300A, CD80, CD83, CLEC4A, CRTAM, CTSL, HLA-DRA, IL12RB1, IL18, IL2RA, LGALS9, LILRB4, LY9, PTPRC, SLAMF6, TGFBR2, TNFRSF14, VSIR, XCL1 Interferon-γ production 0.001 21 AXL, BTN3A2, CD14, CD160, CD244, CD274, CRTAM, HAVCR2, IFNL1, IL12RB1, IL18, IL1R1, KLRK1, LGALS9, LILRB4, PDCD1LG2, PGLYRP2, SLAMF1, SLAMF6, VSIR, XCL1 B cell activation 0.007 18 ADA, CD27, CD28, CD300A, CD38, CD40, CD74, FCGR2B, FCRL3, IGLC2, IL7, MZB1, PTPRC, SLAMF8, TNFRSF13B, TNFRSF21, TNFRSF4, TNFSF13B IL-10 production 0.007 14 CD274, CD28, CD46, CD83, DLL1, FCGR2B, HGF, LGALS9, LILRB4, PDCD1LG2, TIGIT, TNFRSF21, XCL1, VSIR Natural killer cell mediated cytotoxicity 0.030 13 CD160, CRTAM, HAVCR2, IL18, KLRD1, KLRK1, LAG3, LGALS9, NCR1, NECTIN2, SH2D1A, SLAMF6, SLAMF7 Treg cell differentiation 0.020 8 CD28, CD46, HLA-DRA, LAG3, LGALS9, LILRB4, MDK, VSIR a Estimated using Fisher’s exact test, Gene Ontology Biological Processes reference database, and a threshold for significance of Benjamini-Hochberg false discovery rate (FDR)<0.05. MAPK=mitogen-activated protein kinases; ERK=extracellular signal-regulated kinases. SUPPLEMENTAL TABLE S5. Individual proteins within each of six clusters agnostically defined using weighted gene co-expression network analysis. See Supplemental Excel file ‘Supplemental Table S5.xlsx’ All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 13 SUPPLEMENTAL TABLE S6. Enriched biological processes and pathways within protein clusters agnostically defined using weighted gene co-expression network analysis. CLUSTER PROTEIN COUNT ENRICHED PROCESSES Turquoise 499  cellular response to stress (98)  vesicle-mediated transport (93)  cytoskeleton organization (80)  regulation of cell cycle (51)  actin filament-based process (51)  small GTPase mediated signal transduction (39)  autophagy (36)  regulation of protein serine/threonine kinase activity (34)  response to virus (33)  cellular response to chemical stress (31)  Ras protein signal transduction (28)  I-kappaB kinase/NF-kappaB signaling (27)  intrinsic apoptotic signaling pathway (27)  regulation of small GTPase mediated signal transduction (23)  platelet activation (20)  establishment or maintenance of cell polarity (17)  type I interferon signaling pathway (10)  endocytic recycling (9)  Fc-gamma receptor signaling pathway (7)  histamine production involved in inflammatory response (4) Blue 167  organonitrogen compound catabolic process (36)  organophosphate metabolic process (31)  protein-containing complex organization (31)  nucleotide metabolic process (26)  cell cycle process (21)  response to oxidative stress (16)  generation of precursor metabolites and energy (14)  protein folding (11)  cellular detoxification (9)  glutathione metabolic process (6)  NADP metabolic process (5)  ERAD pathway (5)  NADPH regeneration (4)  glucose 6-phosphate metabolic process (4)  AMP and GMP metabolic processes (4) All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 14 Brown 42  negative regulation of multicellular organismal process (15)  cell-cell adhesion (15)  viral entry into host cell (7)  regulation of T cell proliferation (7)  tumor necrosis factor-mediated signaling pathway (5)  TNFR2 non-canonical NFkB pathway (5)  ephrin receptor signaling pathway (4)  lymphangiogenesis (3)  regulation of extracellular matrix organization (3) Yellow 26  cell migration (18)  regulation of protein phosphorylation (11)  regulation of cellular component organization (11)  leukocyte migration (10)  cellular response to cytokine stimulus (10)  regulation of cell differentiation (10)  circulatory system development (9)  regulation of transferase activity (8)  regulation of protein kinase activity (8)  cellular response to growth factor stimulus (8)  response to wounding (8)  response to growth factor (8)  regulation of MAPK cascade (8)  regulation of cell adhesion (8)  negative regulation of cell death (8)  regulation of ERK1 and ERK2 cascade (7)  angiogenesis (7)  chemokine-mediated signaling pathway (6)  neutrophil and granulocyte chemotaxis (6)  coagulation (5)  regulation of smooth muscle cell migration (4)  platelet activation (4)  negative regulation of apoptotic signaling pathway (4)  protein kinase B signaling (4) Green 23  carboxylic acid metabolic process (15)  organonitrogen compound biosynthetic process (8)  lipid metabolic process (7)  fatty acid metabolic process (4)  alcohol metabolic process (4)  serine family amino acid catabolic process (3)  glycine metabolic process (3)  olefinic compound metabolic process (3) All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 15 Red 12  muscle contraction (7)  muscle structure development (6)  actin cytoskeleton organization (6)  supramolecular fiber organization (6)  striated muscle cell differentiation (5)  sarcomere organization (4)  myofibril assembly (4)  regulation of heart contraction (3) Enriched biological processes among proteins composing each cluster defined by weighted gene co-expression network analysis, estimated using Fisher’s exact test, Gene Ontology Biological Processes reference database, and a threshold for significance of Benjamini- Hochberg false discovery rate (FDR)<0.05. Number of proteins mapping to given annotation noted in parentheses. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 16 SUPPLEMENTAL TABLE S7A. Cross-sectional associations between agnostically defined clusters of proteins and HIV serostatus among PLWH and PWOH in the United States (n=352). CLUSTER MODEL 1 a MODEL 2 b Estimate (95% CI) p-value Estimate (95% CI) p-value Blue 0.14 (-0.08 to 0.35) 0.208 0.13 (-0.09 to 0.35) 0.238 Brown 0.45 (0.25 to 0.65) 1.28E-05 0.48 (0.28 to 0.67) 3.99E-06 Green 0.15 (-0.05 to 0.34) 0.152 0.17 (-0.03 to 0.38) 0.088 Red 0.05 (-0.16 to 0.26) 0.622 0.05 (-0.17 to 0.26) 0.661 Turquoise 0.10 (-0.11 to 0.32) 0.336 0.09 (-0.13 to 0.31) 0.421 Yellow 0.23 (0.03 to 0.43) 0.027 0.24 (0.03 to 0.45) 0.023 Mean difference in standardized protein cluster plasma abundance comparing persons living with HIV (PLWH) to persons without HIV (PWOH), estimated using linear regression with robust variance. a Adjusted for age, sex, race/ethnicity, estimated glomerular filtration rate, and hepatitis C. b Further adjusted for education, current hazardous alcohol use, pack-years of smoking in prior 5 years, stimulant use in prior 5 years, and opioid use in prior 5 years. SUPPLEMENTAL TABLE S7B. Cross-sectional associations between agnostically defined clusters of proteins and HIV serostatus among PLWH with undetectable plasma HIV RNA and PWOH in the United States (n=297). CLUSTER MODEL 1 a MODEL 2 b Estimate (95% CI) p-value Estimate (95% CI) p-value Blue 0.12 (-0.10 to 0.34) 0.296 0.10 (-0.12 to 0.33) 0.366 Brown 0.36 (0.15 to 0.57) 6.97E-04 0.42 (0.21 to 0.64) 1.02E-04 Green 0.20 (-0.01 to 0.41) 0.067 0.22 (0.00 to 0.44) 0.046 Red 0.02 (-0.21 to 0.25) 0.845 0.03 (-0.21 to 0.27) 0.816 Turquoise 0.10 (-0.12 to 0.33) 0.370 0.10 (-0.13 to 0.34) 0.396 Yellow 0.21 (-0.01 to 0.42) 0.061 0.22 (0.00 to 0.45) 0.047 Mean difference in standardized protein cluster plasma abundance comparing persons living with HIV (PLWH) with undetectable plasma HIV RNA (<50 copies/mL) to persons without HIV (PWOH), estimated using linear regression with robust variance. a Adjusted for age, sex, race/ethnicity, estimated glomerular filtration rate, and hepatitis C. b Further adjusted for education, current hazardous alcohol use, pack-years of smoking in prior 5 years, stimulant use in prior 5 years, and opioid use in prior 5 years. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 17 SUPPLEMENTAL TABLE S8. Proteins in Brown cluster defined using weighted gene co- expression network analysis corresponding to statistically over-represented biological processes. Gene Ontology: Biological Process Enrichment FDR Protein Count Proteins Cell-cell adhesion 0.022 15 B2M, BSG, CD46, CD93, DSC2, ESAM, FSTL3, HAVCR2, JAM2, NECTIN2, NECTIN4, TGFBR2, TNFRSF14, TNFRSF21, VSIG4 Cytokine production 0.044 13 ACVRL1, IFNGR1, IL10RB, LTBR, RELT, TGFBR2, TNFRSF10B, TNFRSF11A, TNFRSF14, TNFRSF19, TNFRSF1A, TNFRSF1B, TNFRSF21 Viral entry into host cell 0.022 7 BSG, CD46, EPHA2, NECTIN2, NECTIN4, SCARB2, TNFRSF14 T cell proliferation 0.030 7 CD46, HAVCR2, TGFBR2, TNFRSF14, TNFRSF1B, TNFRSF21, VSIG4 TNF signaling 0.021 5 TNFRSF11A, TNFRSF14, TNFRSF19, TNFRSF1A, TNFRSF1B TNFR2 non-canonical NFkB pathway 0.024 5 LTBR, TNFRSF11A, TNFRSF14, TNFRSF1A, TNFRSF1B Ephrin signaling 0.022 4 EFNA4, EPHA2, EPHB4, EPHB6 Lymphangiogenesis 0.022 3 ACVRL1, CLEC14A, EPHA2 Extracellular matrix organization 0.022 3 CST3, TNFRSF1A, TNFRSF1B a Estimated using Fisher’s exact test, Gene Ontology Biological Processes reference database, and a threshold for significance of Benjamini-Hochberg false discovery rate (FDR)<0.05. TNF=tumor necrosis factor. SUPPLEMENTAL TABLE S9. Cross-sectional associations between HIV-associated plasma protein abundances and indexed left atrial volume among PLWH and PWOH in the United States (n=352). See Supplemental Excel file ‘Supplemental Table S9.xlsx’ Mean differences in left atrial volume indexed for body surface area (LAVi) per standard deviation (SD) increment in individual plasma protein abundances, estimated using linear regression with robust variance. Model 1 adjusts for age, sex, and race/ethnicity; Model 2 further adjusts for HIV serostatus, hepatitis C infection, and estimated glomerular filtration rate; and Model 3 further adjusts for education, body mass index, systolic blood pressure, anti-hypertensive medication, dyslipidemia, diabetes, current hazardous alcohol use, pack-years of smoking in prior 5 years, stimulant use in prior 5 years, and opioid use in prior 5 years. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 18 SUPPLEMENTAL TABLE S10. Proteins cross-sectionally associated with HIV serostatus and left atrial size corresponding to statistically over-represented biological processes. Gene Ontology: Biological Process Enrichment FDR a Protein Count Proteins Cell-cell adhesion 1.20E-04 26 CD160, CD27, CD74, CD80, CD83, CDH1, CDH17, CRTAM, CX3CL1, ESAM, HAVCR2, IGFBP2, IL1RL2, ITGB7, JAM2, KLRK1, LAG3, PDGFRA, PODXL2, SCARF2, SIGLEC1, THBS4, TIGIT, TNFRSF14, TNFRSF21, VCAM1 T cell activation 2.60E-05 20 CD160, CD27, CD48, CD74, CD80, CD83, CRTAM, HAVCR2, IGFBP2, IL1RL2, KLRK1, LAG3, SLAMF7, TIGIT, TNFRSF14, TNFRSF1B, TNFRSF21, TNFRSF4, TNFRSF9, VCAM1 Cytokine production 0.002 20 AXL, BTN2A1, CD160, CD74, CD80, CD83, COL3A1, CRTAM, CX3CL1, DLL1, EPHA2, HAVCR2, IL1RL2, KLRK1, LAG3, TIGIT, TNFRSF14, TNFRSF1B, TNFRSF21, TNFRSF8 Leukocyte cell-cell adhesion 7.26E-05 18 CD160, CD27, CD74, CD80, CD83, CRTAM, HAVCR2, IGFBP2, IL1RL2, ITGB7, JAM2, KLRK1, LAG3, PODXL2, TIGIT, TNFRSF14, TNFRSF21, VCAM1 T cell proliferation 0.007 10 CD80, CRTAM, HAVCR2, IGFBP2, TNFRSF14, TNFRSF1B, TNFRSF21, TNFRSF4, TNFRSF9, VCAM1 TNFR2 non-canonical NFĸB pathway 0.003 8 CD27, TNFRSF1B, TNFRSF4, TNFRSF8, TNFRSF9, TNFRSF13B, TNFRSF14, TNFRSF17 B cell activation 0.019 9 CD27, CD74, CD79B, CDH17, DLL1, TNFRSF13B, TNFRSF21, TNFRSF4, VCAM1 Cellular response to tumor necrosis factor 0.022 9 CX3CL1, FAS, IL18BP, TNFRSF1B, TNFRSF4, TNFRSF21, TNFRSF14, TNFRSF17, VCAM1 Viral entry into host cell 0.029 8 AXL, CD74, CD80, EPHA2, ITGB7, SIGLEC1, TNFRSF4, TNFRSF14 T cell differentiation 0.045 8 CD27, CD74, CD80, CD83, CRTAM, IL1RL2, LAG3, TNFRSF9 Lymphocyte migration 0.023 7 CRTAM, CX3CL1, JAM2, CXCL10, TNFRSF14, ITGB7, KLRK1 Natural killer cell mediated cytotoxicity 0.009 6 CD160, CRTAM, HAVCR2, KLRK1, LAG3, SLAMF7 Platelet derived growth factor signaling 0.027 5 COL3A1, COL4A1, PDGFRA, THBS2, THBS4 a Estimated using Fisher’s exact test, Gene Ontology Biological Processes reference database, and a threshold for significance of Benjamini-Hochberg false discovery rate (FDR)<0.05. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 19 SUPPLEMENTAL TABLE S11. Percent difference in association between HIV serostatus and indexed left atrial volume (LAVi) with adjustment for plasma abundance of 73 individual candidate protein contributors (n=352). Protein Mean Difference in LAVi, PLWH vs. PWOH (95% CI) p-value FDR Reduction in HIV Estimate with Protein Adjustment, % CRTAM 0.45 (-2.32 to 3.23) 0.75 0.75 82.0 TNFRSF8 0.84 (-1.26 to 2.94) 0.43 0.44 66.4 CD27 0.96 (-1.11 to 3.03) 0.37 0.38 61.7 TNFRSF17 1.13 (-0.90 to 3.16) 0.28 0.29 54.8 CD160 1.14 (-1.51 to 3.78) 0.40 0.41 54.5 IL18BP 1.20 (-0.96 to 3.36) 0.28 0.29 51.9 TNFRSF9 1.21 (-0.96 to 3.38) 0.27 0.29 51.5 CD79B 1.24 (-0.82 to 3.30) 0.24 0.27 50.3 PDCD1 1.30 (-0.86 to 3.47) 0.24 0.27 47.8 DLL1 1.32 (-0.76 to 3.41) 0.21 0.25 47.0 TNFRSF14 1.33 (-0.75 to 3.41) 0.21 0.25 46.8 CD74 1.35 (-0.81 to 3.51) 0.22 0.26 46.0 KLRK1 1.35 (-0.97 to 3.67) 0.25 0.28 45.9 VCAM1 1.40 (-0.79 to 3.59) 0.21 0.25 43.9 EFEMP1 1.44 (-0.66 to 3.54) 0.18 0.24 42.5 SLAMF7 1.44 (-0.69 to 3.58) 0.18 0.24 42.2 EPHA2 1.46 (-0.71 to 3.64) 0.19 0.24 41.5 TNFRSF1B 1.48 (-0.63 to 3.60) 0.17 0.24 40.6 CD48 1.51 (-0.61 to 3.64) 0.16 0.23 39.5 CLEC14A 1.54 (-0.61 to 3.70) 0.16 0.23 38.3 FABP2 1.55 (-0.87 to 3.98) 0.21 0.25 37.9 LAG3 1.56 (-0.86 to 3.99) 0.21 0.25 37.4 CHRDL1 1.57 (-0.43 to 3.56) 0.12 0.21 37.4 PODXL2 1.58 (-0.49 to 3.64) 0.13 0.21 36.9 CXCL10 1.58 (-0.78 to 3.95) 0.19 0.24 36.7 SIGLEC1 1.58 (-0.58 to 3.75) 0.15 0.23 36.6 FOLR2 1.59 (-0.69 to 3.87) 0.17 0.24 36.4 BTN2A1 1.61 (-0.52 to 3.74) 0.14 0.21 35.6 TIMP1 1.66 (-0.46 to 3.78) 0.13 0.21 33.7 JAM2 1.69 (-0.46 to 3.85) 0.12 0.21 32.3 CDH1 1.70 (-0.56 to 3.96) 0.14 0.21 32.1 CSF1 1.70 (-0.38 to 3.78) 0.11 0.21 32.0 TIGIT 1.70 (-0.52 to 3.93) 0.13 0.21 31.8 LTBP2 1.73 (-0.46 to 3.93) 0.12 0.21 30.7 THBS4 1.74 (-0.42 to 3.90) 0.11 0.21 30.4 NOTCH3 1.74 (-0.46 to 3.95) 0.12 0.21 30.3 OGN 1.75 (-0.44 to 3.93) 0.12 0.21 30.1 TNFRSF4 1.75 (-0.44 to 3.94) 0.12 0.21 30.0 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 20 CNTN3 1.76 (-0.35 to 3.87) 0.10 0.21 29.6 HAVCR2 1.76 (-0.42 to 3.94) 0.11 0.21 29.5 TNFRSF13B 1.77 (-0.31 to 3.85) 0.09 0.21 29.0 NT-proBNP 1.79 (-0.55 to 4.13) 0.13 0.21 28.4 CD80 1.81 (-0.34 to 3.97) 0.10 0.21 27.5 TGFBR3 1.83 (-0.32 to 3.98) 0.10 0.21 26.7 SCARF2 1.85 (-0.42 to 4.12) 0.11 0.21 25.9 ITGB7 1.86 (-0.45 to 4.17) 0.12 0.21 25.7 COL4A1 1.87 (-0.30 to 4.04) 0.09 0.21 25.3 THBS2 1.87 (-0.39 to 4.13) 0.10 0.21 25.1 CDH17 1.87 (-0.40 to 4.15) 0.11 0.21 25.0 TFPI2 1.88 (-0.35 to 4.10) 0.10 0.21 24.9 OXT 1.88 (-0.34 to 4.10) 0.10 0.21 24.7 LAYN 1.89 (-0.30 to 4.08) 0.09 0.21 24.6 CRIM1 1.90 (-0.32 to 4.11) 0.09 0.21 24.1 ESAM 1.92 (-0.23 to 4.06) 0.08 0.21 23.4 CD83 1.92 (-0.27 to 4.11) 0.09 0.21 23.3 LAMA4 1.92 (-0.20 to 4.04) 0.08 0.21 23.3 S100G 1.92 (-0.25 to 4.09) 0.08 0.21 23.1 IGFBP2 1.93 (-0.25 to 4.12) 0.08 0.21 22.7 CKB 1.94 (-0.24 to 4.12) 0.08 0.21 22.5 TNFRSF21 1.96 (-0.25 to 4.17) 0.08 0.21 21.6 FAS 1.98 (-0.22 to 4.18) 0.08 0.21 20.9 DCTPP1 1.98 (-0.28 to 4.23) 0.09 0.21 20.8 ROR1 2.00 (-0.21 to 4.21) 0.08 0.21 20.0 PDGFRA 2.01 (-0.33 to 4.36) 0.09 0.21 19.4 COL3A1 2.02 (-0.22 to 4.26) 0.08 0.21 19.2 CX3CL1 2.05 (-0.24 to 4.33) 0.08 0.21 18.1 ITGBL1 2.05 (-0.37 to 4.47) 0.10 0.21 18.0 AXL 2.07 (-0.40 to 4.53) 0.10 0.21 17.3 IL1RL2 2.08 (-0.24 to 4.40) 0.08 0.21 16.8 FBLN2 2.08 (-0.28 to 4.45) 0.08 0.21 16.7 PROS1 2.08 (-0.52 to 4.69) 0.12 0.21 16.7 KAZALD1 2.17 (-0.25 to 4.60) 0.08 0.21 13.1 B3GNT7 2.19 (-0.20 to 4.58) 0.07 0.21 12.4 Estimated using linear regression with robust variance, adjusting for age, sex, race/ethnicity, education, body mass index, systolic blood pressure, anti-hypertensive medication, dyslipidemia, diabetes, current hazardous alcohol use, cumulative pack-years of smoking in prior 5 years, stimulant use in prior 5 years, opioid use in prior 5 years, hepatitis C, estimated glomerular filtration rate, and indicated protein. CI=confidence interval; FDR=false discovery rate (Benjamini-Hochberg). All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 21 SUPPLEMENTAL TABLE S12. Difference in association between plasma abundance of 73 individual proteins of interest and indexed left atrial volume (LAVi) by HIV serostatus in SMASH (n=352). Protein Mean Difference in LAVi per SD Increment in Plasma Protein, PLWH (95% CI) Mean Difference in LAVi per SD Increment in Plasma Protein, PWOH (95% CI) Interaction p-value Interaction FDR AXL 2.19 (0.81 to 3.57) -0.08 (-1.68 to 1.51) 0.048 0.195 B3GNT7 0.99 (-0.39 to 2.37) 1.52 (0.15 to 2.88) 0.983 0.995 BTN2A1 2.20 (0.81 to 3.60) 0.59 (-1.11 to 2.29) 0.057 0.198 CD160 2.07 (0.68 to 3.45) 2.15 (0.35 to 3.95) 0.679 0.775 CD27 2.05 (0.77 to 3.32) 1.13 (-0.65 to 2.91) 0.223 0.418 CD48 2.13 (0.78 to 3.48) 0.40 (-1.75 to 2.56) 0.086 0.252 CD74 2.12 (0.36 to 3.88) 0.38 (-1.67 to 2.44) 0.129 0.268 CD79B 2.03 (0.68 to 3.38) 0.58 (-1.37 to 2.54) 0.109 0.262 CD80 2.25 (0.83 to 3.67) -0.35 (-2.05 to 1.36) 0.0066 0.096 CD83 2.92 (1.46 to 4.38) 0.21 (-1.41 to 1.84) 0.0021 0.076 CDH1 1.85 (0.46 to 3.24) 0.60 (-1.12 to 2.32) 0.241 0.419 CDH17 1.04 (-0.28 to 2.36) 1.38 (-0.25 to 3.02) 0.944 0.971 CHRDL1 1.77 (0.29 to 3.25) 0.48 (-1.25 to 2.20) 0.268 0.439 CKB 0.78 (-0.73 to 2.28) 0.66 (-0.76 to 2.08) 0.913 0.956 CLEC14A 2.94 (1.37 to 4.51) 0.61 (-0.68 to 1.90) 0.039 0.187 CNTN3 -2.11 (-3.70 to -0.53) -1.55 (-3.58 to 0.49) 0.601 0.730 COL3A1 1.26 (0.17 to 2.35) 1.01 (-0.80 to 2.82) 0.672 0.775 COL4A1 2.28 (0.95 to 3.60) 1.67 (0.23 to 3.10) 0.482 0.628 CRIM1 2.04 (0.74 to 3.35) -0.47 (-2.26 to 1.32) 0.016 0.148 CRTAM 2.43 (1.08 to 3.79) 1.56 (-0.31 to 3.43) 0.261 0.439 CSF1 2.33 (0.82 to 3.85) -0.13 (-2.09 to 1.83) 0.021 0.152 CX3CL1 2.00 (0.46 to 3.54) -0.42 (-2.16 to 1.32) 0.084 0.252 CXCL10 2.57 (1.05 to 4.10) 0.44 (-1.34 to 2.21) 0.016 0.148 DCTPP1 1.53 (0.27 to 2.80) 1.16 (-0.13 to 2.45) 0.299 0.469 DLL1 2.04 (0.53 to 3.54) 0.66 (-0.97 to 2.28) 0.128 0.268 EFEMP1 2.45 (1.08 to 3.81) 1.89 (0.53 to 3.24) 0.181 0.347 EPHA2 2.00 (0.37 to 3.63) 1.03 (-0.43 to 2.49) 0.163 0.323 ESAM 2.11 (0.75 to 3.48) -0.14 (-1.87 to 1.59) 0.117 0.267 FABP2 1.83 (0.53 to 3.13) 0.60 (-0.88 to 2.08) 0.239 0.419 FAS 2.20 (0.77 to 3.64) 0.67 (-0.81 to 2.15) 0.123 0.268 FBLN2 1.48 (0.09 to 2.87) 0.87 (-0.47 to 2.22) 0.364 0.519 FOLR2 2.89 (1.50 to 4.28) 0.21 (-1.63 to 2.06) 0.0006 0.055 HAVCR2 2.35 (0.78 to 3.91) 0.20 (-1.42 to 1.82) 0.011 0.137 IGFBP2 1.70 (0.18 to 3.23) 0.56 (-0.90 to 2.02) 0.626 0.737 IL18BP 2.79 (1.32 to 4.25) 0.77 (-1.17 to 2.71) 0.057 0.198 IL1RL2 -1.72 (-3.21 to -0.23) -0.67 (-2.15 to 0.82) 0.313 0.476 ITGB7 0.91 (-0.54 to 2.36) 1.68 (-0.18 to 3.53) 0.61 0.730 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 22 ITGBL1 1.82 (0.37 to 3.26) 0.32 (-1.26 to 1.90) 0.1 0.262 JAM2 2.74 (1.29 to 4.19) 1.01 (-0.39 to 2.41) 0.104 0.262 KAZALD1 1.48 (0.13 to 2.83) 1.09 (-0.30 to 2.47) 0.724 0.801 KLRK1 1.45 (-0.29 to 3.19) 1.18 (-0.50 to 2.87) 0.331 0.493 LAG3 1.34 (-0.08 to 2.76) 1.11 (-1.33 to 3.55) 0.532 0.669 LAMA4 1.99 (0.67 to 3.30) 0.44 (-1.33 to 2.22) 0.107 0.262 LAYN 2.00 (0.64 to 3.36) 1.71 (0.40 to 3.02) 0.823 0.884 LTBP2 2.11 (0.70 to 3.51) 0.26 (-1.39 to 1.91) 0.046 0.195 NOTCH3 2.67 (1.31 to 4.03) 0.96 (-0.65 to 2.56) 0.045 0.195 NT-proBNP 3.88 (2.65 to 5.11) 1.12 (-0.22 to 2.46) 0.0033 0.080 OGN 2.09 (0.75 to 3.44) 1.02 (-0.68 to 2.73) 0.09 0.254 OXT -1.08 (-2.54 to 0.38) -1.32 (-2.75 to 0.12) 0.425 0.564 PDCD1 2.28 (0.96 to 3.61) 0.16 (-1.58 to 1.91) 0.039 0.187 PDGFRA 1.65 (0.35 to 2.95) 1.32 (-0.53 to 3.17) 0.369 0.519 PODXL2 1.85 (0.48 to 3.23) -0.78 (-2.80 to 1.23) 0.074 0.247 PROS1 -1.84 (-3.01 to -0.68) -1.90 (-3.95 to 0.15) 0.422 0.564 ROR1 1.60 (0.39 to 2.82) 1.79 (-0.09 to 3.67) 0.801 0.873 S100G 1.16 (-0.15 to 2.47) 1.36 (-0.12 to 2.83) 0.704 0.790 SCARF2 1.92 (0.52 to 3.32) 0.86 (-0.69 to 2.42) 0.36 0.519 SIGLEC1 2.03 (0.38 to 3.68) 0.07 (-1.77 to 1.90) 0.019 0.152 SLAMF7 1.02 (-0.26 to 2.31) 1.75 (-0.09 to 3.59) 0.917 0.956 TFPI2 1.25 (-0.17 to 2.67) 1.08 (-0.40 to 2.55) 0.42 0.564 TGFBR3 1.99 (0.64 to 3.33) 1.15 (-0.25 to 2.56) 0.229 0.418 THBS2 1.97 (0.73 to 3.22) 0.28 (-1.44 to 2.00) 0.027 0.152 THBS4 1.92 (0.71 to 3.12) 1.53 (-0.26 to 3.31) 0.494 0.632 TIGIT 1.86 (0.55 to 3.16) 1.26 (-0.01 to 2.54) 0.27 0.439 TIMP1 2.30 (0.75 to 3.84) -0.08 (-2.47 to 2.31) 0.082 0.252 TNFRSF13B 1.55 (0.38 to 2.73) 1.26 (-0.81 to 3.32) 0.577 0.714 TNFRSF14 2.32 (0.89 to 3.74) -0.27 (-2.25 to 1.71) 0.024 0.152 TNFRSF17 1.04 (-0.24 to 2.32) 1.59 (-0.61 to 3.78) 0.995 0.995 TNFRSF1B 2.44 (0.94 to 3.94) 0.18 (-1.60 to 1.96) 0.026 0.152 TNFRSF21 2.16 (0.69 to 3.63) 0.73 (-0.80 to 2.25) 0.111 0.262 TNFRSF4 2.08 (0.57 to 3.58) 0.54 (-0.98 to 2.07) 0.133 0.269 TNFRSF8 2.07 (0.72 to 3.41) 1.29 (-0.80 to 3.37) 0.302 0.469 TNFRSF9 3.31 (1.99 to 4.63) 0.83 (-0.78 to 2.45) 0.0053 0.096 VCAM1 2.55 (1.08 to 4.03) 0.21 (-1.74 to 2.15) 0.051 0.195 Estimated using linear regression models with robust variance among subgroups, adjusting for age, sex, race/ethnicity, education, body mass index, systolic blood pressure, anti-hypertensive medication, dyslipidemia, diabetes, current hazardous alcohol use, cumulative pack-years of smoking in prior 5 years, stimulant use in prior 5 years, opioid use in prior 5 years, hepatitis C, estimated glomerular filtration rate, and indicated protein. Multiplicative protein × HIV interaction term tested in model of all participants, further adjusting for HIV serostatus. PLWH=persons living with HIV (n=214); PWOH=persons without HIV (n=138); SD=standard deviation; CI=confidence interval; FDR=false discovery rate (Benjamini-Hochberg). All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 23 SUPPLEMENTAL TABLE S13. Difference in association between plasma abundance of 73 individual proteins of interest and indexed left atrial volume (LAVi) by age group in SMASH (n=352). Protein Mean Difference in LAVi per SD Increment in Plasma Protein, Age ≥55 (95% CI) Mean Difference in LAVi per SD Increment in Plasma Protein, Age <55 (95% CI) Interaction p-value Interaction FDR AXL 1.71 (0.17 to 3.25) 1.07 (-0.26 to 2.40) 0.703 0.994 B3GNT7 2.16 (0.84 to 3.49) 0.42 (-0.80 to 1.64) 0.277 0.994 BTN2A1 1.73 (0.26 to 3.20) 1.26 (-0.26 to 2.78) 0.935 0.994 CD160 2.70 (1.04 to 4.36) 1.62 (0.07 to 3.17) 0.756 0.994 CD27 1.89 (0.38 to 3.39) 1.50 (0.02 to 2.99) 0.559 0.994 CD48 1.28 (-0.32 to 2.88) 2.08 (0.55 to 3.62) 0.748 0.994 CD74 0.92 (-0.89 to 2.73) 2.24 (0.49 to 3.98) 0.373 0.994 CD79B 1.65 (0.08 to 3.21) 1.31 (-0.21 to 2.84) 0.225 0.994 CD80 1.77 (0.20 to 3.35) 0.94 (-0.63 to 2.51) 0.527 0.994 CD83 1.55 (-0.15 to 3.25) 1.65 (-0.08 to 3.38) 0.994 0.994 CDH1 1.54 (0.18 to 2.90) 1.27 (-0.31 to 2.85) 0.536 0.994 CDH17 1.43 (0.03 to 2.83) 1.06 (-0.43 to 2.54) 0.643 0.994 CHRDL1 1.45 (-0.17 to 3.06) 0.98 (-0.62 to 2.58) 0.917 0.994 CKB 0.32 (-1.08 to 1.73) 1.59 (-0.06 to 3.24) 0.504 0.994 CLEC14A 2.28 (0.89 to 3.67) 1.37 (-0.19 to 2.93) 0.914 0.994 CNTN3 -2.45 (-4.24 to -0.67) -0.91 (-2.54 to 0.72) 0.304 0.994 COL3A1 0.90 (-0.57 to 2.37) 1.48 (0.29 to 2.67) 0.210 0.994 COL4A1 2.83 (1.42 to 4.25) 0.83 (-0.58 to 2.24) 0.053 0.994 CRIM1 1.77 (0.35 to 3.19) 0.77 (-0.75 to 2.28) 0.249 0.994 CRTAM 2.42 (0.77 to 4.08) 2.45 (0.82 to 4.09) 0.627 0.994 CSF1 1.45 (-0.40 to 3.31) 1.27 (-0.19 to 2.73) 0.663 0.994 CX3CL1 1.18 (-0.36 to 2.72) 1.33 (-0.49 to 3.16) 0.771 0.994 CXCL10 0.80 (-0.97 to 2.56) 3.15 (1.50 to 4.79) 0.494 0.994 DCTPP1 1.29 (-0.13 to 2.72) 1.35 (0.06 to 2.64) 0.203 0.994 DLL1 2.03 (0.45 to 3.60) 0.94 (-0.59 to 2.47) 0.530 0.994 EFEMP1 2.49 (1.14 to 3.84) 1.72 (0.32 to 3.12) 0.260 0.994 EPHA2 1.73 (0.12 to 3.33) 1.15 (-0.38 to 2.68) 0.943 0.994 ESAM 1.79 (0.23 to 3.34) 0.51 (-0.88 to 1.90) 0.730 0.994 FABP2 1.79 (0.23 to 3.35) 1.01 (-0.45 to 2.46) 0.716 0.994 FAS 1.27 (-0.27 to 2.81) 1.18 (-0.18 to 2.54) 0.925 0.994 FBLN2 1.79 (0.19 to 3.40) 0.55 (-0.73 to 1.83) 0.691 0.994 FOLR2 2.11 (0.62 to 3.61) 1.75 (0.15 to 3.36) 0.409 0.994 HAVCR2 1.78 (0.27 to 3.30) 0.94 (-0.76 to 2.64) 0.979 0.994 IGFBP2 1.27 (-0.25 to 2.78) 1.66 (0.09 to 3.22) 0.981 0.994 IL18BP 1.81 (0.22 to 3.40) 2.01 (0.34 to 3.68) 0.853 0.994 IL1RL2 0.17 (-1.54 to 1.88) -2.44 (-3.70 to -1.18) 0.105 0.994 ITGB7 0.90 (-0.98 to 2.78) 1.28 (-0.23 to 2.78) 0.431 0.994 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 24 ITGBL1 2.48 (0.76 to 4.21) 0.33 (-0.97 to 1.64) 0.985 0.994 JAM2 2.41 (0.84 to 3.97) 1.39 (0.09 to 2.69) 0.797 0.994 KAZALD1 2.34 (0.88 to 3.80) 0.96 (-0.29 to 2.20) 0.294 0.994 KLRK1 1.09 (-0.61 to 2.79) 1.28 (-0.30 to 2.85) 0.601 0.994 LAG3 1.47 (-0.38 to 3.32) 1.78 (0.24 to 3.32) 0.708 0.994 LAMA4 1.66 (0.12 to 3.21) 0.55 (-1.03 to 2.14) 0.351 0.994 LAYN 2.22 (1.05 to 3.40) 1.25 (-0.23 to 2.72) 0.994 0.994 LTBP2 2.67 (1.14 to 4.19) -0.07 (-1.47 to 1.33) 0.202 0.994 NOTCH3 2.12 (0.73 to 3.51) 2.18 (0.66 to 3.70) 0.205 0.994 NT-proBNP 2.79 (1.51 to 4.07) 3.34 (1.91 to 4.77) 0.569 0.994 OGN 1.63 (0.18 to 3.07) 1.51 (-0.13 to 3.15) 0.537 0.994 OXT -0.89 (-2.46 to 0.68) -1.52 (-2.81 to -0.22) 0.226 0.994 PDCD1 1.73 (0.11 to 3.35) 1.63 (0.10 to 3.15) 0.565 0.994 PDGFRA 2.25 (0.80 to 3.71) 0.83 (-0.65 to 2.31) 0.238 0.994 PODXL2 0.93 (-0.71 to 2.57) 1.46 (-0.12 to 3.04) 0.811 0.994 PROS1 -1.25 (-2.90 to 0.40) -2.43 (-3.66 to -1.20) 0.781 0.994 ROR1 1.43 (0.01 to 2.86) 1.78 (0.34 to 3.21) 0.591 0.994 S100G 1.85 (0.50 to 3.21) 0.24 (-1.12 to 1.60) 0.421 0.994 SCARF2 2.01 (0.62 to 3.41) 1.12 (-0.31 to 2.55) 0.914 0.994 SIGLEC1 1.08 (-0.59 to 2.75) 1.67 (-0.04 to 3.38) 0.992 0.994 SLAMF7 1.15 (-0.31 to 2.62) 1.49 (-0.01 to 3.00) 0.260 0.994 TFPI2 1.70 (0.36 to 3.05) 0.63 (-0.99 to 2.25) 0.359 0.994 TGFBR3 1.79 (0.56 to 3.03) 1.40 (-0.22 to 3.01) 0.863 0.994 THBS2 2.24 (0.59 to 3.88) 0.95 (-0.25 to 2.15) 0.923 0.994 THBS4 1.92 (0.37 to 3.46) 2.14 (0.70 to 3.58) 0.706 0.994 TIGIT 2.14 (0.91 to 3.36) 0.64 (-0.73 to 2.01) 0.976 0.994 TIMP1 2.56 (0.58 to 4.55) 0.29 (-1.45 to 2.03) 0.446 0.994 TNFRSF13B 1.33 (-0.31 to 2.97) 1.51 (-0.05 to 3.07) 0.709 0.994 TNFRSF14 1.82 (0.24 to 3.39) 0.68 (-0.98 to 2.35) 0.974 0.994 TNFRSF17 0.86 (-0.99 to 2.71) 1.43 (0.04 to 2.81) 0.848 0.994 TNFRSF1B 1.96 (0.27 to 3.65) 0.78 (-0.94 to 2.51) 0.524 0.994 TNFRSF21 1.40 (-0.04 to 2.84) 1.42 (-0.05 to 2.89) 0.994 0.994 TNFRSF4 0.90 (-0.63 to 2.43) 1.96 (0.49 to 3.44) 0.199 0.994 TNFRSF8 1.78 (0.20 to 3.37) 1.69 (0.15 to 3.24) 0.794 0.994 TNFRSF9 2.26 (0.78 to 3.74) 2.35 (0.82 to 3.88) 0.824 0.994 VCAM1 2.58 (0.86 to 4.29) 1.49 (-0.01 to 2.99) 0.285 0.994 Estimated using linear regression models with robust variance among subgroups, adjusting for age, sex, race/ethnicity, education, body mass index, systolic blood pressure, anti-hypertensive medication, dyslipidemia, diabetes, current hazardous alcohol use, cumulative pack-years of smoking in prior 5 years, stimulant use in prior 5 years, opioid use in prior 5 years, HIV, hepatitis C, estimated glomerular filtration rate, and indicated protein. Multiplicative protein × age interaction term tested in model of all participants. Subgroups defined by dichotomization at median age in SMASH (55 years); SD=standard deviation; CI=confidence interval; FDR=false discovery rate (Benjamini-Hochberg). All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 25 SUPPLEMENTAL TABLE S14. Cross-sectional associations between proteins of interest and indexed left atrial volume (LAVi) in the Multi-Ethnic Study of Atherosclerosis (median age 67 years, n=1242) Protein Model 1 Model 2 Mean Difference in LAVi per SD Increment in Plasma Protein (95% CI) p-value FDR Mean Difference in LAVi per SD Increment in Plasma Protein (95% CI) p-value FDR AXL 0.71 (0.11 to 1.31) 0.02 0.08 0.48 (-0.12 to 1.09) 0.11 0.31 B3GNT7 0.54 (-0.06 to 1.14) 0.08 0.19 0.59 (-0.01 to 1.19) 0.06 0.19 BTN2A1 0.41 (-0.31 to 1.13) 0.26 0.37 0.30 (-0.41 to 1.01) 0.40 0.52 CD160 0.62 (-0.06 to 1.30) 0.07 0.19 0.49 (-0.18 to 1.16) 0.15 0.32 CD27 0.36 (-0.31 to 1.03) 0.29 0.39 0.32 (-0.35 to 0.98) 0.35 0.52 CD48 0.66 (0.05 to 1.27) 0.04 0.12 0.46 (-0.15 to 1.08) 0.14 0.32 CD74 0.52 (-0.18 to 1.23) 0.15 0.26 0.51 (-0.21 to 1.23) 0.16 0.32 CD79B 0.46 (-0.18 to 1.10) 0.16 0.27 0.42 (-0.22 to 1.06) 0.20 0.37 CD80 0.29 (-0.30 to 0.87) 0.34 0.44 0.19 (-0.39 to 0.77) 0.52 0.61 CD83 0.29 (-0.37 to 0.94) 0.39 0.46 0.29 (-0.38 to 0.95) 0.40 0.52 CDH1 -0.49 (-1.07 to 0.10) 0.10 0.21 -0.49 (-1.09 to 0.10) 0.10 0.29 CDH17 0.06 (-0.54 to 0.67) 0.83 0.88 0.11 (-0.49 to 0.70) 0.73 0.79 CHRDL1 -0.01 (-0.65 to 0.63) 0.98 0.98 0.01 (-0.62 to 0.64) 0.97 0.97 CKB 0.91 (0.24 to 1.58) 0.008 0.06 1.15 (0.45 to 1.84) 0.001 0.02 CLEC14A 1.45 (0.73 to 2.18) 8.2E-05 0.003 1.36 (0.65 to 2.07) 1.8E-04 0.01 CNTN3 -0.79 (-1.43 to -0.16) 0.01 0.07 -1.05 (-1.73 to -0.38) 0.002 0.03 COL3A1 0.77 (0.13 to 1.40) 0.02 0.07 0.71 (0.09 to 1.34) 0.03 0.12 COL4A1 1.21 (0.57 to 1.86) 2.3E-04 0.006 1.20 (0.56 to 1.85) 2.5E-04 0.01 CRIM1 0.91 (0.25 to 1.57) 0.007 0.06 0.83 (0.19 to 1.48) 0.01 0.09 CRTAM 0.6 (-0.04 to 1.24) 0.07 0.18 0.51 (-0.13 to 1.15) 0.12 0.31 CSF1 0.6 (-0.07 to 1.27) 0.08 0.19 0.39 (-0.29 to 1.06) 0.26 0.42 CX3CL1 0.88 (0.25 to 1.50) 0.006 0.06 0.82 (0.20 to 1.44) 0.01 0.08 CXCL10 0.52 (-0.09 to 1.13) 0.10 0.21 0.35 (-0.28 to 0.98) 0.27 0.43 DCTPP1 0.73 (0.11 to 1.36) 0.02 0.08 0.64 (0.02 to 1.27) 0.04 0.17 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 26 DLL1 0.31 (-0.40 to 1.03) 0.39 0.46 0.21 (-0.51 to 0.92) 0.57 0.66 EFEMP1 0.32 (-0.36 to 0.99) 0.36 0.45 0.26 (-0.41 to 0.92) 0.45 0.56 EPHA2 0.85 (0.16 to 1.54) 0.02 0.07 0.76 (0.08 to 1.45) 0.03 0.13 ESAM 0.69 (-0.02 to 1.39) 0.06 0.16 0.54 (-0.15 to 1.23) 0.13 0.31 FABP2 -0.06 (-0.71 to 0.58) 0.84 0.88 0.10 (-0.54 to 0.73) 0.77 0.83 FAS -0.01 (-0.62 to 0.60) 0.97 0.98 -0.07 (-0.67 to 0.53) 0.82 0.86 FBLN2 0.42 (-0.22 to 1.06) 0.20 0.29 0.28 (-0.35 to 0.91) 0.39 0.52 FOLR2 0.55 (-0.10 to 1.21) 0.10 0.21 0.39 (-0.26 to 1.05) 0.24 0.40 HAVCR2 0.48 (-0.24 to 1.20) 0.19 0.29 0.34 (-0.39 to 1.06) 0.36 0.52 IGFBP2 0.99 (0.33 to 1.65) 0.003 0.04 1.26 (0.56 to 1.97) 4.4E-04 0.008 IL18BP 0.46 (-0.22 to 1.13) 0.18 0.29 0.35 (-0.32 to 1.02) 0.31 0.47 IL1RL2 -0.16 (-0.78 to 0.46) 0.62 0.67 -0.14 (-0.75 to 0.48) 0.66 0.74 ITGB7 0.40 (-0.20 to 1.00) 0.19 0.29 0.35 (-0.24 to 0.94) 0.24 0.40 ITGBL1 0.56 (-0.10 to 1.22) 0.09 0.21 0.49 (-0.16 to 1.15) 0.14 0.32 JAM2 0.60 (-0.13 to 1.33) 0.11 0.21 0.52 (-0.20 to 1.24) 0.16 0.32 KAZALD1 0.66 (0.03 to 1.29) 0.04 0.13 0.53 (-0.09 to 1.16) 0.09 0.29 KLRK1 0.44 (-0.15 to 1.03) 0.15 0.26 0.38 (-0.20 to 0.96) 0.20 0.37 LAG3 0.04 (-0.60 to 0.68) 0.90 0.93 0.01 (-0.64 to 0.66) 0.97 0.97 LAMA4 0.88 (0.20 to 1.55) 0.01 0.06 0.79 (0.13 to 1.45) 0.02 0.11 LAYN 0.56 (-0.18 to 1.29) 0.14 0.25 0.56 (-0.16 to 1.29) 0.13 0.31 LTBP2 0.76 (0.10 to 1.43) 0.03 0.09 0.70 (0.04 to 1.36) 0.04 0.16 NOTCH3 1.05 (0.39 to 1.72) 0.002 0.03 0.95 (0.30 to 1.61) 0.004 0.04 NT-proBNP 3.63 (2.93 to 4.33) <1.0E-20 <1.0E-20 3.41 (2.70 to 4.11) <1.0E-20 <1.0E-20 OGN 0.26 (-0.54 to 1.07) 0.52 0.58 0.17 (-0.63 to 0.98) 0.67 0.74 OXT -0.44 (-1.03 to 0.14) 0.14 0.25 -0.60 (-1.20 to 0.00) 0.05 0.18 PDCD1 0.42 (-0.21 to 1.04) 0.19 0.29 0.44 (-0.17 to 1.06) 0.16 0.32 PDGFRA 1.07 (0.43 to 1.71) 0.001 0.02 0.94 (0.32 to 1.57) 0.003 0.03 PODXL2 0.37 (-0.29 to 1.02) 0.27 0.38 0.30 (-0.36 to 0.96) 0.38 0.52 PROS1 -0.78 (-1.40 to -0.16) 0.01 0.07 -0.72 (-1.33 to -0.10) 0.02 0.12 ROR1 0.77 (0.08 to 1.46) 0.03 0.10 0.65 (-0.03 to 1.34) 0.06 0.20 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 27 S100G -0.29 (-0.94 to 0.36) 0.38 0.46 -0.24 (-0.87 to 0.40) 0.46 0.56 SCARF2 0.95 (0.23 to 1.67) 0.01 0.06 0.88 (0.18 to 1.59) 0.01 0.09 SIGLEC1 0.21 (-0.40 to 0.83) 0.50 0.56 -0.05 (-0.69 to 0.58) 0.87 0.90 SLAMF7 0.28 (-0.34 to 0.90) 0.38 0.46 0.14 (-0.46 to 0.74) 0.66 0.74 TFPI2 0.56 (-0.12 to 1.24) 0.11 0.21 0.45 (-0.23 to 1.13) 0.20 0.37 TGFBR3 0.53 (-0.06 to 1.12) 0.08 0.19 0.49 (-0.10 to 1.07) 0.10 0.29 THBS2 0.63 (-0.08 to 1.34) 0.08 0.19 0.46 (-0.26 to 1.18) 0.21 0.37 THBS4 0.95 (0.33 to 1.58) 0.003 0.03 0.76 (0.11 to 1.42) 0.02 0.12 TIGIT -0.07 (-0.71 to 0.57) 0.82 0.88 -0.08 (-0.71 to 0.55) 0.80 0.85 TIMP1 0.84 (0.15 to 1.52) 0.02 0.07 0.70 (0.01 to 1.38) 0.05 0.17 TNFRSF13B 0.32 (-0.31 to 0.94) 0.32 0.43 0.36 (-0.27 to 0.99) 0.27 0.42 TNFRSF14 0.39 (-0.32 to 1.10) 0.28 0.39 0.25 (-0.45 to 0.96) 0.48 0.58 TNFRSF17 0.27 (-0.38 to 0.92) 0.41 0.48 0.28 (-0.37 to 0.93) 0.40 0.52 TNFRSF1B 0.45 (-0.26 to 1.16) 0.21 0.31 0.32 (-0.39 to 1.03) 0.38 0.52 TNFRSF21 0.33 (-0.35 to 1.01) 0.34 0.44 0.25 (-0.42 to 0.92) 0.46 0.56 TNFRSF4 0.48 (-0.23 to 1.19) 0.18 0.29 0.44 (-0.26 to 1.13) 0.22 0.38 TNFRSF8 0.27 (-0.39 to 0.94) 0.42 0.48 0.30 (-0.37 to 0.97) 0.38 0.52 TNFRSF9 0.67 (-0.01 to 1.34) 0.05 0.16 0.71 (0.04 to 1.39) 0.04 0.16 VCAM1 0.62 (-0.05 to 1.28) 0.07 0.19 0.49 (-0.17 to 1.15) 0.14 0.32 Estimated using linear regression with robust variance. Model 1 adjusts for study site, age, sex, race/ethnicity, and estimated glomerular filtration rate. Model 2 further adjusts for education, body mass index, systolic blood pressure, anti-hypertensive medication, hyperlipidemia, diabetes, and smoking. SD=standard deviation; CI=confidence interval; FDR=false discovery rate (Benjamini-Hochberg). All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 28 SUPPLEMENTAL TABLE S15. Cross-sectional associations between proteins of interest and indexed left atrial volume (LAVi) by age group in the Multi-Ethnic Study of Atherosclerosis (median age 67 years, n=1242) Protein Mean Difference in LAVi per SD Increment in Plasma Protein, Age ≥67 (95% CI) Mean Difference in LAVi per SD Increment in Plasma Protein, Age <67 (95% CI) Interaction p-value Interaction FDR AXL 1.00 (0.05 to 1.95) 0.01 (-0.72 to 0.75) 0.07 0.12 B3GNT7 1.38 (0.46 to 2.29) -0.16 (-0.92 to 0.61) 0.001 0.01 BTN2A1 1.03 (-0.01 to 2.07) -0.58 (-1.49 to 0.34) 0.001 0.01 CD160 0.49 (-0.55 to 1.53) 0.45 (-0.38 to 1.28) 0.41 0.50 CD27 0.69 (-0.32 to 1.69) -0.17 (-1.04 to 0.71) 0.02 0.05 CD48 0.77 (-0.19 to 1.73) 0.18 (-0.56 to 0.92) 0.35 0.43 CD74 1.10 (-0.04 to 2.25) -0.25 (-1.12 to 0.62) 0.006 0.03 CD79B 0.37 (-0.65 to 1.40) 0.45 (-0.34 to 1.24) 0.53 0.61 CD80 0.29 (-0.69 to 1.28) -0.01 (-0.70 to 0.68) 0.14 0.19 CD83 0.83 (-0.20 to 1.87) -0.30 (-1.11 to 0.51) 0.04 0.08 CDH1 -0.26 (-1.23 to 0.70) -0.56 (-1.29 to 0.17) 0.24 0.30 CDH17 -0.21 (-1.11 to 0.70) 0.55 (-0.19 to 1.28) 0.77 0.81 CHRDL1 0.41 (-0.56 to 1.38) -0.28 (-1.09 to 0.53) 0.20 0.26 CKB 1.18 (0.11 to 2.26) 0.91 (0.05 to 1.77) 0.17 0.22 CLEC14A 2.28 (1.24 to 3.32) 0.38 (-0.55 to 1.31) 4.3E-04 0.01 CNTN3 -0.58 (-1.60 to 0.44) -1.43 (-2.33 to -0.52) 0.09 0.14 COL3A1 0.93 (-0.01 to 1.88) 0.35 (-0.42 to 1.11) 0.01 0.03 COL4A1 2.10 (1.12 to 3.08) 0.30 (-0.48 to 1.07) 0.001 0.01 CRIM1 1.47 (0.52 to 2.42) 0.16 (-0.66 to 0.99) 0.006 0.02 CRTAM 1.06 (0.02 to 2.09) 0.04 (-0.72 to 0.81) 0.09 0.13 CSF1 0.56 (-0.49 to 1.60) 0.03 (-0.77 to 0.83) 0.07 0.12 CX3CL1 1.62 (0.66 to 2.59) -0.06 (-0.84 to 0.71) 0.006 0.03 CXCL10 0.67 (-0.33 to 1.67) -0.04 (-0.82 to 0.74) 0.01 0.04 DCTPP1 1.71 (0.78 to 2.64) -0.25 (-1.04 to 0.53) 0.001 0.01 DLL1 1.08 (0.05 to 2.10) -0.67 (-1.60 to 0.26) 0.005 0.02 EFEMP1 0.80 (-0.19 to 1.79) -0.31 (-1.16 to 0.55) 0.04 0.08 EPHA2 1.42 (0.40 to 2.44) -0.11 (-0.98 to 0.76) 0.002 0.01 ESAM 0.94 (-0.05 to 1.93) -0.01 (-0.96 to 0.94) 0.04 0.07 FABP2 0.93 (-0.07 to 1.93) -0.69 (-1.51 to 0.13) 0.02 0.05 FAS 0.16 (-0.73 to 1.05) -0.22 (-0.98 to 0.54) 0.09 0.13 FBLN2 0.65 (-0.36 to 1.66) -0.24 (-0.98 to 0.51) 0.007 0.03 FOLR2 0.49 (-0.44 to 1.41) 0.16 (-0.76 to 1.08) 0.44 0.52 HAVCR2 0.63 (-0.50 to 1.75) -0.04 (-1.00 to 0.92) 0.01 0.04 IGFBP2 1.40 (0.28 to 2.52) 0.95 (0.09 to 1.80) 0.15 0.21 IL18BP 1.03 (-0.01 to 2.08) -0.45 (-1.24 to 0.35) 0.006 0.02 IL1RL2 -0.23 (-1.18 to 0.73) 0.08 (-0.70 to 0.87) 0.42 0.51 ITGB7 0.83 (-0.03 to 1.70) -0.07 (-0.85 to 0.71) 0.10 0.15 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 29 ITGBL1 1.01 (-0.06 to 2.07) -0.04 (-0.80 to 0.71) 0.09 0.14 JAM2 1.23 (0.22 to 2.25) -0.24 (-1.25 to 0.77) 0.03 0.06 KAZALD1 1.12 (0.12 to 2.12) -0.08 (-0.83 to 0.68) 0.008 0.03 KLRK1 0.39 (-0.52 to 1.29) 0.34 (-0.43 to 1.11) 0.07 0.11 LAG3 0.00 (-1.03 to 1.02) 0.06 (-0.70 to 0.82) 0.53 0.61 LAMA4 1.25 (0.24 to 2.26) 0.31 (-0.52 to 1.13) 0.04 0.08 LAYN 1.33 (0.26 to 2.40) -0.38 (-1.31 to 0.54) 0.005 0.02 LTBP2 1.53 (0.55 to 2.52) -0.12 (-0.94 to 0.70) 0.02 0.04 NOTCH3 1.56 (0.61 to 2.52) 0.29 (-0.57 to 1.14) 0.02 0.05 NT-proBNP 4.17 (3.08 to 5.26) 2.58 (1.68 to 3.47) 0.01 0.04 OGN 0.85 (-0.34 to 2.04) -0.49 (-1.56 to 0.58) 0.02 0.05 OXT -0.46 (-1.43 to 0.51) -0.68 (-1.41 to 0.05) 0.90 0.92 PDCD1 0.67 (-0.19 to 1.52) 0.11 (-0.73 to 0.95) 0.15 0.21 PDGFRA 1.36 (0.36 to 2.36) 0.39 (-0.38 to 1.16) 0.02 0.05 PODXL2 0.70 (-0.26 to 1.67) -0.14 (-1.01 to 0.73) 0.07 0.12 PROS1 -0.49 (-1.46 to 0.47) -0.79 (-1.57 to -0.01) 0.62 0.67 ROR1 1.09 (0.07 to 2.11) 0.23 (-0.68 to 1.13) 0.02 0.05 S100G 0.33 (-0.67 to 1.33) -0.82 (-1.59 to -0.05) 0.03 0.05 SCARF2 1.61 (0.57 to 2.64) 0.21 (-0.74 to 1.15) 0.02 0.04 SIGLEC1 0.00 (-1.10 to 1.10) -0.05 (-0.76 to 0.65) 0.54 0.61 SLAMF7 -0.09 (-1.02 to 0.84) 0.48 (-0.26 to 1.21) 0.79 0.82 TFPI2 0.54 (-0.49 to 1.57) 0.30 (-0.58 to 1.18) 0.13 0.19 TGFBR3 0.55 (-0.34 to 1.45) 0.48 (-0.28 to 1.25) 0.20 0.26 THBS2 1.49 (0.37 to 2.60) -0.55 (-1.37 to 0.27) 0.002 0.01 THBS4 1.81 (0.82 to 2.81) -0.13 (-0.95 to 0.69) 0.003 0.02 TIGIT -0.27 (-1.26 to 0.72) 0.11 (-0.72 to 0.93) 0.70 0.75 TIMP1 1.49 (0.44 to 2.55) -0.13 (-0.99 to 0.74) 0.003 0.02 TNFRSF13B 0.35 (-0.62 to 1.32) 0.44 (-0.36 to 1.24) 0.92 0.92 TNFRSF14 1.00 (-0.03 to 2.03) -0.64 (-1.59 to 0.32) 0.005 0.02 TNFRSF17 0.25 (-0.80 to 1.29) 0.39 (-0.40 to 1.17) 0.89 0.92 TNFRSF1B 1.03 (0.01 to 2.06) -0.59 (-1.55 to 0.37) 0.001 0.01 TNFRSF21 0.78 (-0.22 to 1.79) -0.34 (-1.21 to 0.53) 0.007 0.03 TNFRSF4 0.98 (0.03 to 1.94) -0.33 (-1.24 to 0.58) 0.002 0.01 TNFRSF8 0.25 (-0.72 to 1.22) 0.18 (-0.74 to 1.10) 0.63 0.69 TNFRSF9 1.26 (0.26 to 2.25) -0.03 (-0.88 to 0.83) 0.02 0.05 VCAM1 1.36 (0.31 to 2.40) -0.39 (-1.14 to 0.35) 0.001 0.01 Estimated using linear regression with robust variance, adjusting for study site, age, sex, race/ethnicity, education, body mass index, systolic blood pressure, anti-hypertensive medication, hyperlipidemia, diabetes, smoking, and estimated glomerular filtration rate. Subgroups defined by dichotomization at median age in MESA cross-sectional analysis sample (67 years). SD=standard deviation; CI=confidence interval; FDR=false discovery rate (Benjamini-Hochberg). All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 30 SUPPLEMENTAL TABLE S16. Association between proteins of interest and incident atrial fibrillation in the Multi-Ethnic Study of Atherosclerosis (median age 67 years at the beginning of follow-up, n=2185) Protein Model 1 Model 2 HR per SD Increment in Plasma Protein (95% CI) p-value FDR HR per SD Increment in Plasma Protein (95% CI) p-value FDR AXL 1.07 (0.94 to 1.21) 0.31 0.45 1.06 (0.93 to 1.20) 0.38 0.55 B3GNT7 1.05 (0.93 to 1.19) 0.46 0.61 1.03 (0.91 to 1.16) 0.66 0.81 BTN2A1 1.15 (0.99 to 1.32) 0.06 0.18 1.12 (0.97 to 1.29) 0.13 0.31 CD160 1.14 (0.99 to 1.30) 0.06 0.18 1.11 (0.97 to 1.27) 0.13 0.31 CD27 1.19 (1.04 to 1.37) 0.01 0.10 1.18 (1.02 to 1.36) 0.02 0.19 CD48 1.12 (0.98 to 1.27) 0.10 0.21 1.10 (0.96 to 1.25) 0.17 0.35 CD74 1.19 (1.04 to 1.36) 0.01 0.10 1.17 (1.02 to 1.34) 0.03 0.19 CD79B 1.15 (1.01 to 1.31) 0.03 0.15 1.13 (0.99 to 1.29) 0.06 0.22 CD80 1.06 (0.93 to 1.21) 0.35 0.50 1.05 (0.92 to 1.20) 0.47 0.67 CD83 1.05 (0.92 to 1.21) 0.46 0.61 1.03 (0.90 to 1.19) 0.65 0.81 CDH1 1.10 (0.97 to 1.24) 0.14 0.27 1.07 (0.95 to 1.21) 0.29 0.46 CDH17 1.02 (0.90 to 1.17) 0.74 0.84 1.01 (0.89 to 1.15) 0.87 0.96 CHRDL1 1.10 (0.97 to 1.25) 0.15 0.28 1.09 (0.96 to 1.25) 0.17 0.35 CKB 1.11 (0.96 to 1.29) 0.16 0.28 1.17 (1.00 to 1.36) 0.05 0.22 CLEC14A 1.21 (1.05 to 1.39) 0.01 0.10 1.19 (1.03 to 1.37) 0.02 0.19 CNTN3 1.02 (0.90 to 1.17) 0.73 0.84 0.99 (0.87 to 1.14) 0.93 0.97 COL3A1 1.09 (0.96 to 1.24) 0.19 0.33 1.06 (0.93 to 1.21) 0.36 0.54 COL4A1 1.03 (0.91 to 1.17) 0.64 0.77 1.03 (0.91 to 1.18) 0.62 0.81 CRIM1 1.10 (0.97 to 1.25) 0.12 0.25 1.09 (0.96 to 1.24) 0.16 0.35 CRTAM 1.04 (0.91 to 1.18) 0.61 0.74 1.03 (0.9 to 1.18) 0.64 0.81 CSF1 1.12 (0.98 to 1.28) 0.09 0.20 1.09 (0.95 to 1.25) 0.20 0.39 CX3CL1 1.15 (1.00 to 1.32) 0.04 0.17 1.14 (1.00 to 1.31) 0.05 0.22 CXCL10 1.16 (1.03 to 1.31) 0.02 0.10 1.15 (1.02 to 1.30) 0.03 0.19 DCTPP1 1.09 (0.95 to 1.24) 0.21 0.36 1.08 (0.95 to 1.23) 0.23 0.41 DLL1 1.18 (1.02 to 1.36) 0.02 0.10 1.15 (0.99 to 1.32) 0.06 0.22 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 31 EFEMP1 1.13 (0.99 to 1.29) 0.06 0.18 1.11 (0.97 to 1.26) 0.13 0.31 EPHA2 1.15 (1.00 to 1.33) 0.06 0.18 1.11 (0.96 to 1.29) 0.16 0.35 ESAM 1.13 (0.99 to 1.30) 0.08 0.19 1.12 (0.97 to 1.28) 0.12 0.31 FABP2 1.04 (0.91 to 1.18) 0.60 0.74 1.03 (0.91 to 1.17) 0.65 0.81 FAS 1.05 (0.93 to 1.20) 0.44 0.61 1.03 (0.90 to 1.17) 0.67 0.81 FBLN2 1.13 (0.99 to 1.30) 0.07 0.18 1.11 (0.97 to 1.27) 0.13 0.31 FOLR2 1.02 (0.89 to 1.16) 0.79 0.87 1.00 (0.87 to 1.14) 0.99 1.00 HAVCR2 1.13 (0.98 to 1.30) 0.08 0.19 1.09 (0.95 to 1.25) 0.23 0.42 IGFBP2 1.09 (0.94 to 1.26) 0.27 0.42 1.14 (0.98 to 1.34) 0.09 0.29 IL18BP 1.13 (0.99 to 1.29) 0.08 0.19 1.11 (0.97 to 1.27) 0.13 0.31 IL1RL2 0.96 (0.85 to 1.10) 0.59 0.74 0.96 (0.84 to 1.10) 0.56 0.76 ITGB7 1.01 (0.89 to 1.15) 0.83 0.90 1.00 (0.88 to 1.14) 1.00 1.00 ITGBL1 1.12 (0.99 to 1.28) 0.08 0.19 1.09 (0.95 to 1.24) 0.21 0.40 JAM2 1.17 (1.02 to 1.35) 0.02 0.10 1.16 (1.01 to 1.34) 0.03 0.20 KAZALD1 1.01 (0.88 to 1.15) 0.93 0.97 1.01 (0.89 to 1.15) 0.88 0.96 KLRK1 1.02 (0.90 to 1.16) 0.74 0.84 1.01 (0.89 to 1.15) 0.87 0.96 LAG3 1.17 (1.03 to 1.33) 0.02 0.10 1.16 (1.02 to 1.32) 0.02 0.19 LAMA4 1.10 (0.96 to 1.27) 0.16 0.28 1.10 (0.95 to 1.26) 0.19 0.38 LAYN 1.20 (1.04 to 1.39) 0.01 0.10 1.19 (1.03 to 1.37) 0.02 0.19 LTBP2 1.13 (0.99 to 1.29) 0.06 0.18 1.13 (0.99 to 1.29) 0.06 0.22 NOTCH3 1.00 (0.88 to 1.15) 0.95 0.97 1.01 (0.88 to 1.15) 0.93 0.97 NT-proBNP 1.57 (1.37 to 1.81) 1.8E-10 1.3E-08 1.54 (1.33 to 1.77) 2.9E-09 2.1E-07 OGN 1.15 (0.99 to 1.34) 0.06 0.18 1.12 (0.96 to 1.30) 0.14 0.33 OXT 1.24 (1.09 to 1.40) 0.001 0.03 1.21 (1.06 to 1.37) 0.005 0.18 PDCD1 1.04 (0.92 to 1.18) 0.51 0.66 1.04 (0.92 to 1.19) 0.53 0.73 PDGFRA 1.16 (1.02 to 1.33) 0.02 0.10 1.14 (1.00 to 1.30) 0.04 0.22 PODXL2 1.00 (0.87 to 1.14) 0.97 0.97 1.02 (0.89 to 1.17) 0.73 0.85 PROS1 1.00 (0.88 to 1.14) 0.97 0.97 0.99 (0.87 to 1.13) 0.92 0.97 ROR1 0.98 (0.85 to 1.12) 0.75 0.84 0.97 (0.84 to 1.12) 0.67 0.81 S100G 1.01 (0.89 to 1.14) 0.93 0.97 1.02 (0.90 to 1.15) 0.80 0.92 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 32 SCARF2 1.13 (0.99 to 1.30) 0.07 0.18 1.12 (0.98 to 1.28) 0.10 0.29 SIGLEC1 1.19 (1.05 to 1.36) 0.007 0.10 1.16 (1.01 to 1.32) 0.03 0.20 SLAMF7 1.08 (0.95 to 1.22) 0.26 0.41 1.07 (0.94 to 1.22) 0.29 0.46 TFPI2 1.12 (0.97 to 1.28) 0.13 0.25 1.11 (0.96 to 1.28) 0.15 0.33 TGFBR3 1.08 (0.95 to 1.24) 0.25 0.40 1.08 (0.95 to 1.24) 0.24 0.42 THBS2 1.11 (0.98 to 1.26) 0.09 0.20 1.08 (0.95 to 1.22) 0.25 0.42 THBS4 1.07 (0.94 to 1.22) 0.30 0.45 1.05 (0.92 to 1.19) 0.48 0.68 TIGIT 1.01 (0.88 to 1.15) 0.91 0.97 1.00 (0.88 to 1.14) 0.97 1.00 TIMP1 1.17 (1.03 to 1.33) 0.01 0.10 1.14 (1.01 to 1.30) 0.04 0.22 TNFRSF13B 1.14 (1.00 to 1.29) 0.05 0.18 1.14 (1.00 to 1.29) 0.05 0.22 TNFRSF14 1.24 (1.08 to 1.42) 0.003 0.06 1.20 (1.04 to 1.38) 0.01 0.19 TNFRSF17 1.13 (0.99 to 1.29) 0.07 0.18 1.13 (0.99 to 1.29) 0.08 0.27 TNFRSF1B 1.16 (1.02 to 1.33) 0.03 0.13 1.14 (0.99 to 1.31) 0.06 0.22 TNFRSF21 1.08 (0.94 to 1.24) 0.25 0.40 1.07 (0.93 to 1.23) 0.33 0.51 TNFRSF4 1.05 (0.91 to 1.20) 0.52 0.66 1.03 (0.90 to 1.18) 0.69 0.81 TNFRSF8 1.06 (0.93 to 1.21) 0.37 0.52 1.07 (0.94 to 1.22) 0.29 0.46 TNFRSF9 1.20 (1.05 to 1.36) 0.007 0.10 1.19 (1.05 to 1.36) 0.009 0.19 VCAM1 1.07 (0.94 to 1.22) 0.30 0.45 1.06 (0.94 to 1.21) 0.34 0.52 Estimated using Cox proportional hazards. Model 1 adjusts for study site, age, sex, race/ethnicity, and estimated glomerular filtration rate. Model 2 further adjusts for education, body mass index, systolic blood pressure, anti-hypertensive medication, dyslipidemia, diabetes, and smoking. N=252 incident atrial fibrillation events. HR=hazard ratio; SD=standard deviation; CI=confidence interval; FDR=false discovery rate (Benjamini-Hochberg). All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 33 SUPPLEMENTAL TABLE S17. Association between proteins of interest and incident clinical heart failure in the Multi-Ethnic Study of Atherosclerosis (median age 67 years at the beginning of follow-up, n=2273) Protein Model 1 Model 2 HR per SD Increment in Plasma Protein (95% CI) p-value FDR HR per SD Increment in Plasma Protein (95% CI) p-value FDR AXL 1.19 (0.91 to 1.55) 0.206 0.269 1.20 (0.92 to 1.57) 0.181 0.240 B3GNT7 1.09 (0.83 to 1.43) 0.514 0.560 1.04 (0.80 to 1.37) 0.756 0.800 BTN2A1 1.75 (1.30 to 2.34) 1.93E-04 0.002 1.68 (1.25 to 2.25) 0.001 0.005 CD160 1.35 (1.02 to 1.78) 0.037 0.069 1.32 (0.99 to 1.75) 0.056 0.105 CD27 1.70 (1.29 to 2.25) 1.98E-04 0.002 1.67 (1.25 to 2.22) 0.001 0.005 CD48 1.49 (1.16 to 1.93) 0.002 0.008 1.49 (1.15 to 1.93) 0.003 0.012 CD74 1.53 (1.17 to 2.00) 0.002 0.008 1.49 (1.12 to 1.98) 0.007 0.022 CD79B 1.52 (1.20 to 1.92) 0.001 0.004 1.50 (1.18 to 1.91) 0.001 0.007 CD80 1.43 (1.10 to 1.84) 0.007 0.019 1.44 (1.10 to 1.88) 0.007 0.022 CD83 1.58 (1.19 to 2.10) 0.002 0.007 1.53 (1.15 to 2.04) 0.004 0.014 CDH1 1.37 (1.09 to 1.72) 0.008 0.020 1.29 (1.01 to 1.66) 0.041 0.085 CDH17 1.27 (0.97 to 1.67) 0.084 0.133 1.20 (0.91 to 1.59) 0.191 0.240 CHRDL1 1.30 (1.00 to 1.69) 0.052 0.089 1.31 (1.01 to 1.71) 0.042 0.086 CKB 1.15 (0.84 to 1.58) 0.376 0.428 1.34 (0.96 to 1.88) 0.082 0.134 CLEC14A 1.84 (1.38 to 2.44) 3.11E-05 4.50E-04 1.85 (1.37 to 2.50) 5.68E-05 0.001 CNTN3 1.13 (0.84 to 1.52) 0.433 0.487 1.02 (0.75 to 1.39) 0.902 0.902 COL3A1 1.21 (0.93 to 1.59) 0.159 0.235 1.14 (0.87 to 1.50) 0.348 0.403 COL4A1 1.16 (0.89 to 1.51) 0.277 0.331 1.23 (0.94 to 1.61) 0.131 0.191 CRIM1 1.36 (1.06 to 1.76) 0.017 0.040 1.37 (1.07 to 1.77) 0.014 0.036 CRTAM 1.22 (0.92 to 1.61) 0.166 0.237 1.23 (0.93 to 1.64) 0.150 0.210 CSF1 1.48 (1.14 to 1.91) 0.003 0.011 1.40 (1.07 to 1.83) 0.014 0.036 CX3CL1 1.19 (0.88 to 1.59) 0.260 0.316 1.22 (0.91 to 1.64) 0.183 0.240 CXCL10 1.17 (0.89 to 1.53) 0.256 0.316 1.17 (0.89 to 1.53) 0.270 0.326 DCTPP1 1.20 (0.91 to 1.59) 0.195 0.264 1.24 (0.94 to 1.64) 0.128 0.191 DLL1 1.70 (1.28 to 2.26) 9.25E-04 0.002 1.62 (1.22 to 2.16) 0.001 0.007 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 34 EFEMP1 1.32 (1.01 to 1.72) 0.041 0.073 1.29 (0.99 to 1.69) 0.063 0.109 EPHA2 1.51 (1.12 to 2.03) 0.007 0.019 1.42 (1.05 to 1.92) 0.024 0.055 ESAM 1.54 (1.17 to 2.03) 0.002 0.009 1.51 (1.15 to 1.98) 0.003 0.012 FABP2 1.50 (1.17 to 1.93) 0.002 0.007 1.45 (1.14 to 1.86) 0.003 0.012 FAS 1.29 (1.01 to 1.66) 0.041 0.073 1.25 (0.97 to 1.60) 0.079 0.130 FBLN2 1.24 (0.93 to 1.66) 0.149 0.227 1.22 (0.91 to 1.63) 0.189 0.240 FOLR2 1.35 (1.04 to 1.77) 0.026 0.054 1.30 (1.01 to 1.69) 0.046 0.090 HAVCR2 1.55 (1.16 to 2.08) 0.003 0.011 1.46 (1.08 to 1.96) 0.013 0.034 IGFBP2 1.19 (0.85 to 1.65) 0.311 0.366 1.41 (1.00 to 1.99) 0.049 0.094 IL18BP 1.56 (1.20 to 2.04) 0.001 0.006 1.56 (1.18 to 2.05) 0.002 0.009 IL1RL2 1.14 (0.86 to 1.51) 0.367 0.426 1.14 (0.85 to 1.53) 0.380 0.421 ITGB7 1.20 (0.92 to 1.57) 0.187 0.258 1.16 (0.89 to 1.51) 0.272 0.326 ITGBL1 1.34 (1.03 to 1.75) 0.029 0.059 1.25 (0.96 to 1.63) 0.095 0.147 JAM2 1.71 (1.30 to 2.25) 1.36E-04 0.002 1.71 (1.29 to 2.27) 1.79E-04 0.002 KAZALD1 1.09 (0.83 to 1.45) 0.529 0.568 1.14 (0.86 to 1.53) 0.356 0.406 KLRK1 1.07 (0.81 to 1.42) 0.636 0.673 1.07 (0.81 to 1.41) 0.649 0.697 LAG3 1.22 (0.92 to 1.60) 0.161 0.235 1.19 (0.91 to 1.56) 0.208 0.257 LAMA4 1.31 (0.98 to 1.74) 0.068 0.113 1.28 (0.96 to 1.70) 0.088 0.139 LAYN 1.67 (1.23 to 2.26) 0.001 0.006 1.64 (1.21 to 2.21) 0.001 0.009 LTBP2 1.24 (0.94 to 1.64) 0.121 0.188 1.28 (0.98 to 1.69) 0.073 0.123 NOTCH3 1.11 (0.83 to 1.49) 0.465 0.514 1.16 (0.87 to 1.54) 0.310 0.365 NT-proBNP 1.99 (1.47 to 2.68) 6.68E-06 2.18E-04 1.94 (1.43 to 2.62) 1.66E-05 0.001 OGN 1.57 (1.14 to 2.16) 0.005 0.015 1.48 (1.07 to 2.03) 0.018 0.043 OXT 1.34 (1.02 to 1.76) 0.033 0.066 1.21 (0.91 to 1.59) 0.187 0.240 PDCD1 1.30 (1.02 to 1.65) 0.035 0.066 1.33 (1.03 to 1.70) 0.026 0.057 PDGFRA 1.00 (0.75 to 1.33) 1.000 1.000 0.97 (0.73 to 1.28) 0.824 0.859 PODXL2 0.99 (0.74 to 1.32) 0.946 0.959 1.10 (0.82 to 1.48) 0.529 0.577 PROS1 0.99 (0.76 to 1.29) 0.929 0.955 0.97 (0.74 to 1.28) 0.845 0.869 ROR1 1.21 (0.89 to 1.64) 0.223 0.281 1.26 (0.92 to 1.72) 0.144 0.207 S100G 1.29 (1.03 to 1.62) 0.026 0.054 1.33 (1.07 to 1.67) 0.012 0.033 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 35 SCARF2 1.21 (0.90 to 1.62) 0.200 0.265 1.22 (0.91 to 1.62) 0.176 0.240 SIGLEC1 1.39 (1.06 to 1.82) 0.016 0.040 1.30 (0.99 to 1.71) 0.058 0.106 SLAMF7 1.05 (0.80 to 1.38) 0.726 0.757 1.02 (0.78 to 1.35) 0.868 0.880 TFPI2 1.38 (1.05 to 1.83) 0.023 0.053 1.38 (1.05 to 1.82) 0.023 0.053 TGFBR3 1.22 (0.91 to 1.64) 0.185 0.258 1.27 (0.95 to 1.69) 0.109 0.165 THBS2 1.60 (1.29 to 1.97) 1.31E-05 2.39E-04 1.53 (1.22 to 1.91) 1.84E-04 2.24E-04 THBS4 1.19 (0.90 to 1.57) 0.214 0.274 1.13 (0.86 to 1.48) 0.374 0.420 TIGIT 1.44 (1.12 to 1.86) 0.004 0.012 1.41 (1.09 to 1.82) 0.008 0.025 TIMP1 1.54 (1.21 to 1.97) 4.87E-04 0.004 1.49 (1.16 to 1.92) 0.002 0.009 TNFRSF13B 1.43 (1.14 to 1.78) 0.002 0.007 1.40 (1.12 to 1.76) 0.003 0.012 TNFRSF14 1.85 (1.41 to 2.43) 9.13E-06 2.22E-04 1.74 (1.33 to 2.29) 6.61E-05 0.001 TNFRSF17 1.51 (1.15 to 1.98) 0.003 0.011 1.49 (1.13 to 1.97) 0.005 0.016 TNFRSF1B 1.80 (1.42 to 2.28) 1.23E-06 8.95E-05 1.81 (1.41 to 2.33) 3.53E-06 2.58E-04 TNFRSF21 1.35 (1.01 to 1.80) 0.042 0.073 1.35 (1.01 to 1.80) 0.040 0.085 TNFRSF4 1.52 (1.18 to 1.97) 0.001 0.007 1.53 (1.17 to 2.01) 0.002 0.010 TNFRSF8 1.34 (1.04 to 1.72) 0.024 0.053 1.41 (1.08 to 1.84) 0.011 0.031 TNFRSF9 1.44 (1.12 to 1.84) 0.004 0.012 1.49 (1.15 to 1.93) 0.003 0.012 VCAM1 1.28 (0.97 to 1.68) 0.081 0.132 1.29 (0.99 to 1.69) 0.062 0.109 Estimated using Cox proportional hazards. Model 1 adjusts for study site, age, sex, race/ethnicity, and estimated glomerular filtration rate. Model 2 further adjusts for education, body mass index, systolic blood pressure, anti-hypertensive medication, dyslipidemia, diabetes, and smoking. N=54 incident adjudicated clinical heart failure events. HR=hazard ratio; SD=standard deviation; CI=confidence interval; FDR=false discovery rate (Benjamini-Hochberg). All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 36 SUPPLEMENTAL FIGURE S1A. Flow diagram of SMASH study participants included in analysis sample. Selection for Olink in SMASH included complete cardiovascular magnetic resonance (CMR) with late gadolinium enhancement for the purpose of additional analyses and cost effectiveness. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 37 SUPPLEMENTAL FIGURE S1B. Flow diagram of MESA study participants included in cross- sectional analysis samples. Reflects data availability as of January 2024. LAVi=indexed left atrial volume. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 38 SUPPLEMENTAL FIGURE S1C. Flow diagram of MESA study participants included in longitudinal analysis samples. Reflects data availability as of January 2024. HF=heart failure; AF=atrial fibrillation. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 39 SUPPLEMENTAL FIGURE S2. Clusters of proteins agnostically defined using weighted gene co-expression network analysis. (A) Dendrogram of protein clusters; gray indicates unclustered proteins; (B) Spearman’s correlations between cluster eigenprotein values, i.e., first principal component of cluster-specific proteins. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 40 SUPPLEMENTAL FIGURE S3. Spearman’s correlation between plasma abundances of 73 proteins independently associated with both HIV seropositivity and incremental indexed left atrial volume, same directionality, among PLWH and PWOH in the United States (n=352). All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 41 SUPPLEMENTAL FIGURE S4. Interaction network of 73 proteins independently associated with both HIV seropositivity and incremental indexed left atrial volume, same directionality. Generated using STRING, a public database of known and predicted protein-protein interactions. Interactions include direct (physical) and indirect (functional) associations derived from computational prediction, knowledge transfer between organisms, and interactions aggregated from other databases. Proteins depicted are limited to those with high confidence association(s) with at least one other protein. Szklarczyk D, et al. The STRING database in 2023: protein–protein association networks and functional enrichment analyses for any sequenced genome of interest. Nucleic Acids Res. 2023;51(D1):D638-646. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 42 SUPPLEMENTAL FIGURE S5. Association between identified HIV-associated proteomic signature of left atrial size and clinical characteristics in SMASH (n=352) All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 43 SUPPLEMENTAL FIGURE S5 Continued. Association between identified HIV-associated proteomic signature of left atrial size and clinical characteristics in SMASH (n=352) Mean standardized difference in plasma abundance among participants with vs. without dichotomous characteristic or per standard deviation increment in continuous characteristic estimated using linear regression. Number displayed is p-value where “0” indicates <0.001. HS=high school; CVD=cardiovascular disease; BMI=body mass index; eGFR=estimated glomerular filtration rate; ART=antiretroviral therapy; PI=protease inhibitor; NNRTI=non-nucleoside reverse transcriptase inhibitor; INSTI=integrase strand inhibitor. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 44 SUPPLEMENTAL FIGURE S6. Volcano plot of association between identified HIV-associated proteomic signature of left atrial size and time to incident adjudicated atrial fibrillation in the Multi-Ethnic Study of Atherosclerosis (n=2185) Hazard ratios estimated using Cox proportional hazards adjusting for study site, age, sex, race/ethnicity, education, body mass index, systolic blood pressure, anti-hypertensive medication, dyslipidemia, diabetes, smoking, and estimated glomerular filtration rate. N=252 incident adjudicated atrial fibrillation events. Purple indicates p<0.05, Orange indicates FDR<0.05. Complete modeling results can be found in Supplemental Table S16. SD=standard deviation; FDR=false discovery rate (Benjamini-Hochberg). All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint 45 SUPPLEMENTAL FIGURE S7. Volcano plot of associations between identified HIV-associated proteomic signature of left atrial size and time to incident adjudicated clinical heart failure in the Multi-Ethnic Study of Atherosclerosis (n=2273) Hazard ratios estimated using Cox proportional hazards adjusting for study site, age, sex, race/ethnicity, education, body mass index, systolic blood pressure, anti-hypertensive medication, dyslipidemia, diabetes, smoking, and estimated glomerular filtration rate. N=54 incident adjudicated heart failure events. Purple indicates p<0.05, Orange indicates FDR<0.05. Complete modeling results can be found in Supplemental Table S17. SD=standard deviation; FDR=false discovery rate (Benjamini-Hochberg). All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprintthis version posted February 14, 2024. ; https://doi.org/10.1101/2024.02.13.24302797doi: medRxiv preprint