Circulating platelet-natural killer cell complexes are elevated in people living with human immunodeficiency virus on antiretroviral therapy Authors: Brita Ostermeier, Amber Douglass, Tomas Raul Wiche Salinas, Sanchita Pandey, Cheryl Clarkson-Paredes, Gregory M Cresswell, Anastas Propratiloff, Marc Siegel, Conrad Russell Y Cruz, Mirko Paiardini, Sanjay B Maggirwar Published: 2026 ABSTRACT AbstractNatural killer (NK) cells stand as a critical pillar in the innate immune defense against viral infections, particularly human immunodeficiency virus (HIV). However, during HIV infection, NK cells demonstrate profound dysfunction, undermining effective immune control and accelerating disease progression along with viral persistence. Understanding the mechanisms driving NK cell dysfunction is essential for the development of innovative therapeutic strategies. This study reveals a significant increase in circulating platelet-NK cell complexes (PNKCs) in individuals living with HIV (PWH) on antiretroviral therapy (ART) compared to those without HIV. PWH exhibited distinct NK cell phenotypes within PNKCs, characterized by elevated levels of CD25, interferon-gamma (IFN-γ), and natural killer group 2, member D (NKG2D), alongside decreased NKp46, relative to NK cells not engaged in complexes with platelets (non-PNKCs). Moreover, the sustained presence of platelets, or PNKCs, during in vitro NK cell expansion resulted in a heightened population of NK cells exhibiting an activated phenotype. In simian immunodeficiency virus–infected Rhesus macaques, PNKC levels were positively correlated with both plasma and cerebrospinal fluid viral loads, underscoring the role of PNKCs in disease progression. Notably, this correlation diminished following monoclonal antibody treatment with inclacumab, which inhibits P-selectin-dependent platelet interactions. We propose that the chronic activation of platelets during HIV infection fosters platelet interactions with NK cells in PWH on ART, ultimately contributing to NK cell dysfunction. Introduction Platelets are increasingly recognized as multifaceted mediators in both hemostasis and immune regulation. In the context of human immunodeficiency virus (HIV) infection, platelets undergo significant activation, closely linked to elevated plasma viral loads and persistent immune activation, even among individuals undergoing antiretroviral therapy (ART)1–5 This activation is driven by direct interactions between HIV virions or their proteins and platelet surface receptors, resulting in the upregulation of activation markers such as CD62P (P-selectin) and the release of soluble CD40 ligand (sCD40L)1–6 The expression of CD62P facilitates the physical interaction between platelets and immune cells, forming cellular complexes7–9 (also referred to as aggregates3,10–17) through its binding to P-selectin glycoprotein ligand-1 (PSGL-1) on leukocytes,7 including T cells, monocytes, and potentially natural killer (NK) cells3–5 These complexes are not only biomarkers of immune activation but may also contribute to HIV pathogenesis by promoting inflammation and viral dissemination.3–5,18 In people with HIV (PWH), the frequency of platelet-immune cell complexes remains elevated despite effective ART,7,11,18 reflecting ongoing platelet activation and immune dysregulation3–5 Specifically, platelet-T-cell complexes (PTCs) and platelet-monocyte complexes (PMCs) have been associated with increased immune activation,7,11,18 altered monocyte and dendritic cell function,7,19 and accelerated disease progression3–5 Notably, platelets can harbor replication-competent HIV, acting as a transient viral reservoir and potentially facilitating viral spread to immune cells, even under conditions of viral suppression.5,18,20 Thus, therapeutic strategies targeting platelet activation, such as anti-CD62P antibodies, show promise in disrupting these pathogenic interactions and may provide valuable adjuncts to ART.3,18,21 The interplay between platelets and NK cells is an emerging area of research. Recent studies have documented increased platelet-NK cell aggregates in HIV-infected individuals, correlating with disease progression and immune exhaustion.3 These aggregates display features of over-activation and functional impairment, including elevated expression of exhaustion markers (CD39, lymphocyte-activation gene 3(LAG-3), programmed cell death protein 1), increased levels of apoptosis, and reduced cytotoxic capacity marked by diminished interferon-gamma (IFN-γ), granzyme B, and perforin production.3 Platelet-derived signals, potentially mediated by the SLAMF5 receptor, may further disrupt NK cell signaling pathways, exacerbating immune dysfunction (Deng et al.3). Critically, ART does not completely restore normal levels of platelet-NK cell aggregates or counteract NK cell exhaustion, underscoring the need for targeted interventions.,22,23 NK cell dysfunction in HIV is characterized by altered subset distributions, particularly the expansion of CD56− CD16+ NK cells,24 alongside impaired effector functions such as reduced degranulation, cytokine secretion, and antibody-dependent cellular cytotoxicity22,25–32 CD56−CD16+ NK cells also express higher levels of inhibitory receptors and lower levels of natural cytotoxicity receptors.33,34 Persistent immune activation, coupled with the presence of inhibitory cytokines like interleukin (IL)-10 and transforming growth factor-beta, further hampers NK cell activity, increasing susceptibility to opportunistic infections and malignancies.22,30,35 PWH (-ART) have elevated surface expression of NK activating receptors,36 and inhibitory receptor expression has been reported to be conserved in one cohort,36 while some are elevated in a different cohort.37 ART initiation aids in at least partial NK reconstitution, primarily by reducing viremia38 and by CD56−CD16+ NK subset expansion reversal in some cohorts,25,39,40 allowing partial recovery of NK cell phenotypes.36 However, another study found elevated activation (CD69) and reduced activating receptor expression during ART.41 While ART can partially restore NK cell phenotype and function, residual defects remain, particularly amidst chronic inflammation and ongoing platelet activation,22,23,30,32,33,39,41–44 as such, PWH may be at risk for other viral infections or cancers.45 Collectively, these findings underscore the central role of platelet-immune cell interactions in the immunopathogenesis of HIV. The formation of platelet-NK cell complexes (PNKCs) represents a novel mechanism of NK cell dysfunction, with significant implications for disease progression and therapeutic intervention. This study aims to elucidate the molecular and functional consequences of PNKC formation that are subserved by HIV infection. Materials and methods Ethics statement The Research Subjects Review Board at the George Washington University approved studies involving human samples under Institutional Review Board protocol number 021848. All blood samples were collected from adults following written informed consent, in accordance with the Declaration of Helsinki. Human samples and PBMC isolation Approximately 40 mL of blood samples were collected in acid-citrate dextrose vacutainers from either people living with HIV receiving ART (PWH [ART+]) or people living without HIV (PWOH) (Tables 1 and 2). Information on peripheral blood mononuclear cell (PBMC) processing can be found in the Materials and Methods. PBMCs were stained for Human NK Phenotype Panels (Table 1) and Human PNKC Panels (Table 2) immediately following isolation from peripheral blood. For more in-depth analysis, t-distributed stochastic neighbor embedding (t-SNE) (learning configuration: auto, iterations: 1000, perplexity: 30, K-nearest neighbors algorithm: exact [vantage point], gradient algorithm: Barnes–Hut) and FlowSOM (size of meta clusters: 3) were run on the concatenated NK cell populations from the Human NK Phenotype Panel using FlowJo v10. Sample sizes differ depending on flow panels, as panels were altered throughout the time of data collection. Table 1Demographic characteristics of PWOH and PWH for human NK phenotype panels.Human NK phenotype panelCD61+ human NK phenotype panelPWOHPWHPWOHPWH N 2315 (1 donor with no demographic info)137Mean age (SD), y33.87 (±15.02)54.21 (±15.68)35.23 (±17.04)47.14 (±14.77)Age, N (%) 18 to 2510 (43.48%)0 (0.00%)6 (46.15%)0 (0.00%) 26 to 407 (30.43%)4 (26.67%)3 (23.08%)3 (42.86%) 41 to 604 (17.39%)4 (26.67%)3 (23.08%)3 (42.86%) 60+2 (8.70%)6 (40.00%)1 (7.69%)1 (14.29%)Race, N (%) Asian5 (21.74%)0 (0.00%)1 (7.69%)0 (0.00%) Black/AA0 (0.00%)6 (40.00%)0 (0.00%)4 (57.14%) Black/AA/Asian0 (0.00%)1 (6.67%)0 (0.00%)1 (14.29%) White17 (73.91%)6 (40.00%)12 (92.31%)2 (28.57%) White/AA0 (0.00%)0 (0.00%)0 (0.00%)0 (0.00%) White/Asian0 (0.00%)0 (0.00%)0 (0.00%)0 (0.00%) Unspecified1 (4.35%)1 (6.67%)0 (0.00%)0 (0.00%)Biological sex, N (%) Male10 (43.48%)13 (86.67%)7 (53.85%)6 (85.71%) Female13 (56.52%)1 (6.67%)6 (46.15%)1 (14.29%)AA, African American. Demographic characteristics of PWOH and PWH for human NK phenotype panels. AA, African American. Table 2Demographic characteristics of PWOH and PWH for CD62P+ and CD61+ human PNKC panels.CD62P+ human PNKC panelCD61+ human PNKC panelPWOHPWHPWOHPWH N 24201513Mean age (SD), y34.22 (±14.36)51.8 (±15.66)35.133 (±15.51)46.69 (±14.45)Age, N (%) 18 to 257 (29.17%)0 (0.00%)4 (26.67%)0 (0.00%) 26 to 4011 (45.83%)7 (35.00%)7 (46.67%)6 (46.15%) 41 to 604 (16.67%)6 (30.00%)3 (20.00%)5 (38.46%) 60+2 (8.33%)7 (35.00%)1 (6.67%)2 (15.38%)Race, N (%) Asian7 (29.17%)0 (0.00%)4 (26.67%)0 (0.00%) Black/AA0 (0.00%)11 (55.00%)0 (0.00%)8 (61.54%) Black/AA/Asian0 (0.00%)1 (5.00%)0 (0.00%)1 (7.69%) White15 (62.50%)6 (30.00%)10 (66.67%)3 (23.08%) White/AA0 (0.00%)1 (5.00%)0 (0.00%)1 (7.69%) White/Asian1 (4.17%)0 (0.00%)1 (6.67%)0 (0.00%) Unspecified1 (4.17%)1 (5.00%)0 (0.00%)0 (0.00%)Biological sex, N (%) Male10 (41.67%)17 (85.00%)7 (46.67%)10 (76.92%) Female14 (58.33%)3 (15.00%)8 (53.33%)3 (23.08%)AA, African American. Demographic characteristics of PWOH and PWH for CD62P+ and CD61+ human PNKC panels. AA, African American. Human platelet isolation and activation Platelets were isolated from buffy coats (Gulf Coast Regional Blood Center, Houston, TX) by diluting 1:4 in Tyrode's Salts (#T2397-1L, Sigma-Aldrich) and processed as previously described.7,18,19 Platelets were activated at 100 × 106 cells/mL in Tyrode's Salts with 0.1 U/mL thrombin (#T6884-100UN, Sigma-Aldrich) for 30 min at 37 °C. Platelet supernatants were frozen at −80 °C. Platelets were washed 3 times with Tyrode's Salts to remove residual treatments before establishing cocultures. Platelet activation status was verified via flow cytometry using the Human Platelet Activation Antibody Panel. NK cell isolation from buffy coat Human PBMCs were obtained from buffy coats using Lymphoprep (STEMCELL). Platelet depletion washes were performed for some donors (−/−) and not others (+/+) (Table 3). NK cells were isolated using the negative selection EasySep Human NK Isolation Kit (#17955, STEMCELL). Baseline NK receptor expression was determined using the Human NK Phenotype Antibody Panel. (+/+) NK cells were directly expanded using the ImmunoCult NK Expansion Kit (STEMCELL Technologies, #100-0711). (−/−) NK cells were sorted for CD61−CD62P− using a Sony SH800Z Cell Sorter before coculture. Table 3Age and biological sex demographic characteristics of buffy coat donors.(−/−)(+/+) N 119Mean age ± SD, y43.45 ± 14.7840.22 ± 9.31Age, N (%) 18 to 251 (9.09%)1 (11.11%) 26 to 404 (36.36%)3 (33.33%) 41 to 605 (45.45%)4 (44.44%) 60+1 (9.09%)0 (0.00%)Biological sex, N (%) Male9 (81.82%)6 (66.67%) Female2 (18.18%)3 (33.33%) Age and biological sex demographic characteristics of buffy coat donors. Coculture of platelets and NK cells Autologous human platelets and NK cells were cultured at a ratio of 20:1 for 20 h, 7 days, and 14 days using the ImmunoCult NK Expansion Kit (STEMCELL Technologies, #100-0711). NK cells were also separately treated with platelet factors: 1 μg/mL sCD40L, 10 μg/mL platelet factor-4 (PF-4), 0.05× activated platelet supernatant, and 0.05× unactivated platelet supernatant, with factors being refreshed on days 3, 7, and 10. At 20 h, NK cells were collected for the 20 h PNKCs Antibody Panel. On days 7 and 14, NK cells were collected for the Expansion NK Phenotype Antibody Panel. Day 14 NK cells were cocultured with target cells for killing assays. Correlative light and electron microscopy Freshly isolated NK cells from buffy coats were plated with autologous activated platelets on poly-D-lysine and collagen-treated etched indium-tin-oxide coverslips (#06486-AB, 2spi). After 2 h of coculture, the cells were fixed with 4% paraformaldehyde (PFA) for confocal microscopy and stained with CD41a (1:75, #14-0419-82, Thermo Fisher [TF]), antimouse IgG AF633 (#A-21052, TF), and 4′,6-diamidino-2-phenylindole (DAPI) for confocal microscopy. Grids were hand-etched on the coverslips before imaging to help with orientation. Samples were imaged on a Carl Zeiss Cell Observer spinning disk confocal microscope with a 40× lens. Samples were then postfixed with 4% PFA and 2% glutaraldehyde in 0.1 M cacodylate buffer for 1 h. Next, samples were prepared for topological scanning electron microscope (SEM) by postfixing in 1% osmium tetroxide in 0.1 M cacodylate buffer for 1 h. Samples were dehydrated by gradual increases in ethanol and finished with a drying step using hexamethyldisilazane (low surface agent) before mounting the coverslips on aluminum stubs. Samples were then imaged on a FEI Teneo LV FEG SEM. Images from the confocal microscope were overlaid on the SEM images using MAPS 3.22 software (TF). PNKCs were quantified using ImageJ software. Live imaging Autologous human platelets and NK cells derived from buffy coats were cocultured at a 20:1 ratio in Roswell Park Memorial Institute 1640 (TF Scientific, #61870036) with the ImmunoCult NK Cell Expansion Kit Supplement (StemCell, #100-0715). Cells were plated in poly-D-lysine-treated 35 mm dishes with a 170-μm glass coverslip bottom (Ibidi, #81156). Live cell imaging was initiated 10 min after coculture using a Leica DMi8 microscope equipped with a 40× objective. Images were acquired every 30 s for a total duration of 8 h. Image processing and analysis were performed using Fiji (ImageJ) software. Platelet ligand expression on NK cells Freshly isolated NK cells were cultured with or without the ImmunoCult NK Cell Expansion Supplement (StemCell, #100-0715) for 20 h. NK cell expression of CD40L and PSGL-1 was determined before culture and at 20 h of culture via flow cytometry. NK cytotoxicity assay Cytotoxicity of day 14 expanded NK cells was tested against uninfected (Jurkat, U937, C20, A2780, H1299, K562) and HIV latently infected (JLAT 10.6, U1, HC69) cell lines, using the DELFIA EuTDA Cytotoxicity Detection Kit (#AD0116, Revvity). Before coculture, HIV latently infected cell lines (HC69, U1, and JLAT 10.6) were activated with 10 ng/mL TNF-α for 18 h, 100 nM ingenol for 24, and 48 h, respectively. The day of coculture, all target cell lines (Jurkat, JLAT 10.6, U937, U1, C20, HC69, A2780, H1299, K562) were prepared for the DELFIA EuTDA Cytotoxicity Detection Kit (#AD0116, Revvity) by incubating with EU Labeling Buffer (0.02% Pluronic F-127 in 1× phosphate-buffered saline (PBS)) + 1 μL bis(acetoxymethyl) 2,2′:6′,2″-terpyridine-6,6″-dicarboxylate for 20 min at 37 °C. Then, the target cells were washed with target wash buffer (20 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid, 2.5 mM probenecid, in 1× PBS) for the number of washes previously determined (Table S1). Seventy-seven milligrams of probenecid (#P36400, TF) was resuspended in 1 mL 1× Hank′s Balanced Salt Solution. Day 14 expanded NK cells were cocultured with the target cells, with a maximum of 100,000 target cells per well in a 96-well plate. According to the manufacturer's instructions, minimum lysis and europium-only controls were used. Maximum lysis controls were obtained by adding 1% Triton X-100 in NK media. Time-resolved fluorescence was read at 612 nm on a FLUOstar Omega (BMG LABTECH). Cytotoxicity was determined as %specific lysis = 100 × ([experimental data−spontaneous cell death]/[maximum cell death−spontaneous cell death]). In cases where maximum cell death was higher in the experimental condition, it was used as the maximum control for that run. Rhesus macaque PBMC samples Male Rhesus macaques (RMs) were treated with either vehicle (n = 5) or 4 mg/kg inclacumab (INC, n = 5) at days −7 p.i. and 14 p.i. RMs treated with vehicle had an average age of 41.6 ± 6.31 mo old, while RMs treated with INC had an average age of 42.8 ± 9.68 mo old. RMs were infected with SIVsm804E-CL757 S-S-P at day 0 p.i.46 Plasma viral load and cerebrospinal fluid (CSF) viral loads were quantified using a hybrid real-time quantitative polymerase chain reaction assays as previously described.47 PBMCs were cryopreserved and thawed for 18 h before staining for the RM NK Phenotype Antibody Panel or functional assays. PBMCs were cocultured with K562 cells at E:T ratios of 20:1, 10:1, and 1:1 for 4 h. Cytotoxicity was determined via flow cytometry using the RM NK Cytotoxicity Antibody Panel. Supernatants from cocultures were frozen at −80 °C and later processed for the Human CD8/NK LEGENDplex Multi-Analyte Flow Assay Kit (#741187, Biolegend). Antibody panels and flow cytometry All flow cytometry samples were acquired on a BD LSR X10 Fortessa, BD Celesta, or Cytek Aurora. First, viability staining was conducted. Then, PBMCs were treated with Human TruStain FcX FC Blocker (BioLegend), followed by surface staining. After surface staining, the cells were fixed with 4% PFA or prepared for intracellular staining using BD Cytofix/Cytoperm (#554722, BD), washed with 1× permeabilization buffer (#554723, BD), and then stained intracellularly. For the Human PNKCs panel, one subpanel contained CD61-stained samples while the others did not; therefore, the sample size (n) differs between the CD61 and CD62P measurements shown in Figs. 2B and C. Similarly, for the Human NK Phenotype Panel, 1 subpanel included CD61 staining, whereas the remaining subpanels did not. The sample size reported for the percentages includes all subpanels, whereas the sample size reported for the MFIs includes only the subpanel with CD61 (to account for differences in MFIs due to alterations in the panel). For Fig. 2D to N, PNKC and non-PNKC populations were identified using only the Human PNKC Phenotype subpanel that contained CD61; samples from subpanels without CD61 were not used for those comparisons. Human PNKC Panel. CD3 (HIT3a), CD19 (HIB19), CD14 (M5E2), CD56 (HCD56), CD16 (B73.1), CD61 (RUU-PL7F12), and CD62P (AK4). 20 h PNKC Panel. CD56 (HCD56), CD16 (B73.1), CD61 (RUU-PL7F12), and CD62P (AK4). For some of these samples, CD40L (24-31) or PSGL-1 (KPL-1) was added to the panel. Human NK Phenotype Panel. CD3 (HIT3a), CD19 (HIB19), CD14 (M5E2), CD56 (B159), CD16 (3G8), CD61 (RUU-PL7F12), natural killer group 2, member D (NKG2D) (1D11), DNAM-1 (11A8), NKp46 (9E2), KIRs (HP-MA4), NKG2A (131411), TIGIT (MBSA43), NKp44 (p44-8), NKp30 (AF29-4D12), CD69 (FN50), CD25 (BC96), CD107a (H4A3), IFN-γ (B27), Granzyme B (QA16A02), and Perforin (dG9). Human Platelet Activation Panel. Platelets were first fixed with 4% PFA and then stained with surface CD61 (VI-PL2) and CD62P (AK4) antibodies. Samples were acquired on a BD LSR X10 Fortessa or BD Celesta. Expansion NK Phenotype. CD56 (B159), CD16 (3G8), CD61 (RUU-PL7F12), NKG2D (1D11), DNAM-1 (11A8), NKp46 (9E2), KIR2DL1/S1/S3/S5+ (referenced generally as KIRs) (HP-MA4), NKG2A (131411), TIGIT (MBSA43), NKp44 (p44-8), NKp30 (AF29-4D12), CD69 (FN50), CD25 (BC96), CD107a (H4A3), IFN-γ (B27), granzyme B (QA16A02), and perforin (dG9). RM NK Phenotype Panels. CD3 (SP34-2), CD20 (2H7), CD14 (M5E2), CD56 (B159), CD16 (3G8), IFN-γ (B27), CD107a (H4A3), NKG2D (1D11), DNAM-1 (DX11), Nkp46 (BAB281), KIRs (HP-MA4), NKG2A (Z199), TIGIT (MBSA43), NKp44 (2.29), NKp30 (AF29-4D12), CD61 (VI-PL2), CD69 (FN50), CD25 (BC69), and CD62P (AC1.2). RM NK Cytotoxicity Antibody Panel. CD3 (SP34-2), CD20 (2H7), CD14 (M5E2), CD56 (B159), CD16 (3G8), IFN-γ (B27), CD107a (H4A3), CD61 (VI-PL2), CD62P (AC1.2), granzyme B (GB11), perforin, and NKG2A (A199). Statistics For experiments with 1 comparison, a single Mann–Whitney test (*≤0.05, **≤0.01, ***≤0.001, ****≤0.0001) was used. For experiments with multiple conditions, multiple Mann–Whitney tests (*≤0.05, **≤0.01, ***≤0.001, ****≤0.0001) were used. Specific statistics for each figure are listed in the figure legends. Results Phenotypic characterization of NK cells derived from PWH on ART To determine whether NK cell dysfunction persists during ART, we recruited people without HIV (PWOH, n = 24) and PWH on ART (PWH [ART+], n = 20) for 2 flow cytometry panels (Tables 1 and 2). NK cells were gated from PBMC populations using flow cytometry panels (Fig. S1A). Consistent with previous reports,41,48 NK cells derived from PWH exhibited reduced NKG2A (%, p = 0.0139), NKp46 (%, P = 0.0008, median fluorescent intensity [MFI], P = 0.0024), and increased CD69 (%, P = 0.0180, MFI, P = 0.0047) compared to those isolated from PWOH (Fig. 1A to C). Interestingly, our analyses also revealed decreased levels of DNAM-1 (%, P = 0.0106, MFI, P = 0.0034) and increased NKp44 (MFI, P = 0.0236) and TIGIT expression (MFI, P = 0.0297) in NK cells isolated from PWH (ART+) (Fig. 1D to F). These alterations are also visualized via dimensionality reduction (t-SNE) analyses (Fig. 1G). Other NK phenotypic markers were not altered (Fig. S1B to G). Similarly, in contrast to a previous report,40 we did not find alterations in the CD16/CD56 NK subsets in PWH (ART+) (Fig. S2A). NK cells from PWH (ART+) are phenotypically different from NK cells from PWOH. PBMCs were isolated from PWOH and PWH on ART (PWH) and stained for NK phenotypic markers via flow cytometry. Total NK cells were defined as CD3−, CD19−, and CD14− lymphocytes consisting of CD16+CD56−, CD16+CD56dim, and CD56bright populations. A to F) Total NK expression of NKG2A, NKp46, CD69, DNAM-1, NKp44, and TIGIT in PWOH and PWH. The first column of graphs indicates the percentage of NK cells expressing respective markers in PWOH (n = 23) and PWH (n = 15). The next column of graphs indicates the MFI for respective markers on NK cells in PWOH (n = 13) and PWH (n = 7). Mann–Whitney test (*≤0.05, **≤0.01, ***≤0.001). G) t-SNE plot with multicolor overlays for concatenated NK cells from PWOH (n = 13) and PWH (n = 7). H) FlowSOM clustering (3 clusters) of NK cells from PWOH (n = 13) and PWH (n = 7). To better understand the features of NK cells isolated from PWH, clustering analyses (FlowSOM) were conducted (Fig. 1H). Both groups have all 3 main subsets of NK cells; however, PWH appear to have a more homogeneous total NK population (Fig. 1H). In particular, the CD16+CD56dim population, which highly expresses CD61, is notably altered in PWH (Fig. 1H). The expression of CD61 (platelet-specific integrin β3) on leukocytes is widely used as a reliable marker for identifying platelet-leukocyte complexes (PLKCs),7,18,49 so we next investigated PNKCs in PWH. Elevated levels of PNKCs in ART-treated PWH Platelet markers such as CD61, CD62P, and CD42a can be efficiently transferred to immune and tumor cells through direct interactions, primarily via microvesicles and possibly trogocytosis. This transfer alters recipient cell phenotypes, with implications for tumor progression and immune modulation50–54 Previously, our lab determined that PMCs are elevated in PWH.7 Similarly, in this study, we found that CD61+ platelet-lymphocyte complexes (PLCs), PLKCs, PMCs, PTCs, and platelet-B-cell complexes (PBCs) are all elevated in PWH (ART+) (Fig. S2B to F). Specifically, we find elevated PNKCs in PWH (ART+) via surface staining of CD61+ (P = 0.041), CD62P+ (P = 0.0098), or CD61+ CD62P+ (P = 0.025) on NK cells (Fig. 2A, B). Notably, a CD61 bright population appears in 6/13 PWH and 1/15 PWOH, exemplified in Fig. 2A. The %CD62P+ PNKCs remain elevated across all NK subsets in PWH (Fig. 2C). CD16+CD56dim NK cells appear to express high levels of CD61 as determined by FlowSOM clustering (Fig. 1H). Indeed, when compared to other subsets, the %PNKCs are highest in the CD16+CD56dim subset for both CD61+ PNKCs and CD62P+ PNKCs (Fig. 2C). Importantly, we observed that the PNKC (CD62P+) formation is independent of age and biological sex among the PWH and PWOH (Fig. S2G and H). Furthermore, when comparing just the biological male sexes in both groups, PNKCs remain significantly elevated in PWH (P = 0.0009), suggesting that differences in biological sex distributions between PWH and PWOH do not account for enhanced PNKC formation (Fig. S2H). PNKCs are elevated in PWH and are phenotypically different from PNKCs from PWH and non-PNKCs. PBMCs were isolated from PWOH and PWH on ART (PWH [ART+]) and stained for platelet and NK activating and inhibitory markers via flow cytometry. Total NK cells are defined as CD3−, CD19−, and CD14− lymphocytes consisting of CD56bright, CD16+CD56+, and CD16+CD56dim populations. PNKCs were defined as CD61+, CD62P+, or CD61+CD62P+. A) PNKC gating strategy in PWOH and PWH. B) PNKCs in PWOH (CD61 n = 15, CD62P n = 24) versus PWH (CD61 n = 13, CD62P n = 20) (flow, %). C) % of total PNKCs (CD61+ and CD62P+) in various NK subsets from PWOH (CD61+ n = 15, CD62P+ n = 24) versus PWH (CD61+ n = 13, CD62P+ n = 20) (flow, %). Multiple Wilcoxon test (*≤0.05, **≤0.01, ***≤0.001). D to M) NK phenotypic markers: CD61+ NK cells and CD61− NK cells from PWOH (n = 13) and PWH (n = 7) (flow, MFI). Mann–Whitney test (*≤0.05, **≤0.01, ***≤0.001). N) t-SNE plot for CD61+ NK cells from PWOH (n = 13) and PWH (n = 7). We investigated the PNKC (CD61+) population by comparing PNKCs (CD61+) to non-PNKCs (CD61−) (Fig. 2). Regardless of infection status, PNKCs have increases in an NK activation marker (CD25, PWOH, P = 0.0002, PWH, P = 0.0156) and a marker of their effector function (IFN-γ, PWOH, P = 0.0002, PWH, P = 0.0156), and a decrease in activating receptor NKp46 levels (PWOH, P = 00005, PWH, P = 0.0156) compared to non-PNKCs (Fig. 2D to F). For PWOH, PNKCs have elevated markers of NK effector function (CD107a, P = 0.0266) and activating receptor (NKp44, P = 0.0005) and reduced activating receptor expression (NKp30, P = 0.0012) (Fig. 2G to I). We also investigated differences between PNKCs from PWOH versus PWH. Compared to PWOH, PNKCs from PWH have elevated NK activating marker expression (CD69, P = 0.0581, Fig. 2J) and reduced NK activating receptor expression (NKp46, P = 0.0369, Fig. 2F). However, compared to PWOH, non-PNKCs from PWH have elevated NKp44 (P = 0.0085), CD69 (P = 0.0011), TIGIT (P = 0.0236), and reduced DNAM-1 (P = 0.0008) and NKG2D (P = 0.0024) expression (Fig. 2H and J to M). Similarly, PNKCs clustered distinctly between PWOH and PWH groups (Fig. 2N). Together, these observations suggest that platelet-NK interactions impact NK phenotypes, with HIV infection itself having additional effects. This notion is consistent with the recent report.3 Morphological characteristics of PNKCs To confirm and morphologically characterize the complex formation between platelets and NK cells, we used correlative light and electron microscopy (CLEM) in an in vitro system. Specifically, autologous NK cells and platelets were purified from buffy coats. Before cocultures, we sorted for the CD61−CD62P− NK cell population, later defined as (−/−), to obtain an NK population “naive” to platelet complexing and to remove as many free platelets and PNKCs as possible before coculture. We confirmed successful platelet and PNKC removal on day 0 (P = 0.007937), which remained the case throughout the expansion (day 7 P = 0.004329, day 14 P = 0.009524) (Fig. S3D and E). Activated platelets and (−/−) NK cells were cocultured together for 2 h and imaged using DAPI to identify NK cells and CD42a to identify platelets (Fig. 3A to C). Two types of PNKCs were formed during these coculture experiments. Type 1 PNKCs consist of one or more platelets directly interacting with the NK cells (Fig. 3D). In contrast, Type 2 PNKCs (Fig. 3E) are NK cells to which there were no platelets physically attached but have clusters of platelet markers present on their surface (Fig. 3F). Due to the high surface expression of PSGL-1 on NK cells (Fig. 3G), we predicted that these interactions occur through the CD62P-PSGL-1 axis. While it is also possible that platelet-derived extracellular vesicles could be involved in the transfer of platelet markers to NK cells, we hypothesize that the Type 1 PNKCs progressively transition into Type 2 PNKCs based on previous understanding of PMCs.7 To investigate the temporal dynamics of PNKCs, we conducted time-lapse imaging of platelet-NK cell interactions over 8 h. At early time points (0 to 30 min), NK cells were observed making initial contacts with individual platelets, with many interactions appearing transient (Fig. 3H). By 1 to 2 h, a distinct subset of NK cells began to engage multiple platelets simultaneously (Fig. 3H). These multiplatelet interactions persisted for 8 h and, in some cases, 48 h (data not shown), with minimal dissociation, suggesting a more stable mode of binding than previously thought. In contrast, some NK cells exhibited short-lived platelet interactions while others showed no detectable platelet associations (Fig. 3H). The noninteracting NK cells remained spatially separated from platelets across all time points. Collectively, these observations suggest heterogeneity in platelet binding behavior within the NK cell population (Fig. 3H). Platelets form complexes with NK cells in vitro. Autologous NK cells and platelets were isolated from buffy coats from PWH. Activated platelets were cocultured with NK cells for 2 h and then prepared for CLEM. Confocal images were taken prior to SEM and overlaid on SEM images using MAPS 3.22 software. A) Platelet-NK coculture methods. Created with BioRender.com. B) SEM imaging of a PNKC using a FEI Teneo LV FEG SEM microscope. C) CLEM area for quantification with DAPI (blue) and CD41a-AF633 (green). D to E) Type 1 and type 2 PNKCs. F) CLEM quantification using ImageJ software. G) %PSGL-1 expression on NK cells (n = 4) (%, flow). H) Live cell time-lapse images of platelet-NK cell interactions at 0 to 8 h post coculture (scale is 5 um). A yellow arrow indicates an NK cell that interacts with multiple platelets simultaneously, and blue arrows indicate NK cells that have transient platelet associations or do not interact with platelets. Molecular mechanisms underlying PNKC formation We then explored 2 receptor-ligand interactions known to promote PNKCs: PSGL-1 (expressed on lymphocytes) binding to CD62P (aka P-selectin, expressed on platelets) and CD40L (expressed on some lymphocytes) binding to CD40 (expressed on platelets). After 20 h, most freshly isolated NK cells express PSGL-1 with low expression of CD40L (Fig. S3A). NK cells are reported to express CD40L upon stimulation with IL-2;55 however, when we added ImmunoCult NK Cell Expansion Supplement for 24 h, we did not observe an increase in either CD40L or PSGL-1 (Fig. S3B). Due to the constitutively high expression of PSGL-1 (Fig. 3H), we expect that NK cells may interact with platelets primarily through the engagement of this receptor. In support of this notion, we also confirmed increased expression of CD62P following platelet activation by thrombin (P < 0.0001) (Fig. S3C). After 20-h cocultures, activated platelets, but not unactivated platelets, formed a significantly higher number of PNKCs compared to the NK alone condition as quantified by CD61+CD62P+ double expression (P = 0.0005) (Fig. 4A). We also incubated subsets of cocultures with INC (which blocks CD62P binding to PSGL-156) and found that INC does not significantly prevent PNKC formation (Fig. 4A). We reasoned that the CD62P-PSGL-1 axis is involved in the formation of PNKCs; however, additional receptor-ligands, including CD40L-CD4057 and αIIbβ3-CD11b,58 may not be disrupted by INC.56 Effects of platelets and platelet factors on NK phenotypes and functions. Autologous NK cells and platelets were isolated from buffy coats from PWOH. Activated (ACT PLT) and unactivated (PLT) platelets were cocultured with NK cells, in presence of INC or nonimmune IgG control, and collected for flow at 20 h, 7 d, and 14 d of expansion. Subset of these cocultures was further incubated for 7 and 14 d with platelet factors, PF-4, sCD40L, or supernatants (0.05×) from PLT or ACT PLT. After expansion, NK cell cytotoxicity assays were conducted by coculturing NK cells with target cell lines for 2 to 4 h using a europium-based assay. A) %PNKCs (CD61+CD62P+) after 20 h platelet-NK coculture (flow, %) (n = 6 to 12). B and C) NK phenotypes on days 7 and 14 (n = 6 to 12) (flow, %). Multiple Wilcoxon test (*≤0.05, **≤0.01, ***≤0.001). D) Representative flow plot of CD25 expression on day 7. E) NK cytotoxicity of U937 and U1 cell lines (n = 6 to 10). Multiple Wilcoxon test (*≤0.05, **≤0.01, ***≤0.001). F) NK cytotoxicity of Jurkat and JLAT 10.6 cell lines (n = 6 to 10). Wilcoxon test (*≤0.05, **≤0.01, ***≤0.001). Next, we investigated the direct (platelet coculture) and “long-term” indirect effects (platelet factors) of platelets on NK phenotypes. We found that NK cells cocultured with activated platelets had a significantly reduced percentage of CD25+ cells on day 7 compared to NK alone (P = 0.0155) (Fig. 4B and D). This could indicate that activated platelets may transiently enhance NK cell activation. However, no other NK phenotypes were significantly altered by either direct or indirect effects of platelets on days 7 and 14 (Fig. 4B and C). This could indicate that platelets and platelet factors do not alter NK phenotypes; however, this does not support our ex vivo findings. As expected, CD62P MFI is not significantly different between the platelet-NK coculture conditions since platelets were only added at the beginning of the coculture (Fig. S3F). It is possible that the singular addition of platelets does not properly mimic in vivo exposure of NK cells to platelets and platelet factors, or that a longer coculture is needed to induce platelet and NK interactions and the PNKC phenotypes we see ex vivo. Effect of platelets on NK cell effector functions To further investigate the effects of platelets, we verified the effector functions of NK cells cocultured with either platelets or platelet factors using europium-based killing assays on day 14 of NK expansion (Fig. 4E and F). Chronic PF-4 treatment reduced NK cytotoxicity against U937 cells (P = 0.0156) but had no significant effect on U1 cells (Fig. 4E). Unactivated platelet coculture enhanced NK cytotoxicity against both JLAT 10.6 (P = 0.0039) and Jurkat cells (P = 0.0234) (Fig. 4F). Neither direct platelet cocultures nor platelet factors had any significant effect on the killing of other cell lines tested (Fig. S4A to E). Interestingly, there is a trend of increased NK lysis of HIV-infected microglia (HC69) cell lines compared to uninfected microglia (C20) cell lines (Fig. S4F). While slight alterations in effector functions are apparent, both direct platelet cocultures and platelet factor treatments do not appear to significantly affect NK phenotypes and functions during expansion. Characterization of PNKCs in simian immunodeficiency virus–infected RMs We next investigated the formation of PNKCs in an in vivo model of SIV infection in RMs that mimics continuous platelet exposure of NK cells in the context of HIV infection. Furthermore, in this model, we treated the RMs with INC once before infection (day −7) and once postinfection (day 14 p.i., Fig. 5A). Platelet inhibitor, INC, treatment does not prevent PNKC formation and has some effects on NK phenotypes. RMs were either treated with vehicle (n = 5) or with INC (n = 5) on weeks −1 and 2 p.i. On week 0 p.i., RMs were infected with SIVsm804E-CL747 S-S-P. Plasma viral load (PVL) and CSF viral loads were taken on days 4 p.i., 7 p.i., 14 p.i., 21 p.i., 28 p.i., and 35 p.i. PVL and CSF viral loads were quantified as SIV RNA copies/mL. PBMCs were collected on days −28 p.i., 7 p.i., 21 p.i., and necropsy (Nx) and cryopreserved. PBMCs were thawed for 18 h before staining for a flow cytometry including NK activating and inhibitory receptors. The remaining PBMCs were cocultured with K562 cells for 4 h, and %specific killing was determined via flow cytometry. A) Method schematic. Created with BioRender.com. B) %PNKCs (CD61 and CD62P) before and after SIV infection (flow, %). Wilcoxon test. C) Fold %PNKCs (CD61 and CD62P) for vehicle (black) or INC (red) treated RMs. Mann–Whitney test. D) %PNKCs (CD61+ and CD62P+) versus log10 SIV PVL copies/mL. Simple linear regression. E) %PNKCs (CD61+ and CD62P+) versus log10 SIV CSF viral load copies/mL. Simple linear regression. F) Phenotypes of NK cells in the vehicle (VEH) and INC treated groups (flow, %). Multiple Mann–Whitney test. G) Fold change in specific killing of K562 cells by RM PBMCs comparing Nx with day −28 p.i. (flow, FC). Mann–Whitney test. H) %specific killing of K562 cells by RM PBMCs on day −28 p.i. versus Nx (flow, FC). Wilcoxon test. SIV infection did not alter PNKC formation, although there was a trend of increased PNKCs on day 7 p.i. (Fig. 5B). Additionally, analogous to our in vitro coculture data (Fig. 4), in vivo INC treatment failed to prevent PNKCs (Fig. 5C). Interestingly, higher %PNKCs (CD61+ or CD62P+) on day 7 p.i. were positively correlated with increased plasma viral load at day 35 p.i. (CD61+ R2= 0.8200, P = 0.0343) (CD62P+ R2 = 0.9928, P = 0.003) (Fig. 5D). Additionally, the %PNKCs (CD62P+) on day 7 p.i. was also correlated with CSF viral load at day 35 p.i. (R2 = 0.7805, P = 0.0469) (Fig. 5E). INC treatment appears to disrupt the association between PNKCs and viral loads (Fig. 5D and E). Furthermore, the %PNKCs at other time points do not correlate with either plasma viral load or CSF viral load on day 35 p.i. (Fig. S5A and B). Notably, there are no significant correlations between the total PLCs, PTCs, or PBCs and plasma or CSF viral load, suggesting that the association with viremia is specific for PNKCs (Fig. S5C and D). Despite these correlations, no alterations in NK phenotypes were observed in RMs that were either infected or treated with INC following SIV infection (Fig. 5F). To further investigate the activity of NK cells from RMs, we developed a killing assay in which we cocultured RM PBMCs, isolated from the same animals before infection (day −28 p.i.) and at necropsy following SIV infection, with K562 cells for 4 h (Fig. S6A). Neither INC treatment (Fig. 5G) nor SIV infection itself (Fig. 5H) significantly affected NK cell killing cytotoxicity against K562 cells. However, there is a trend of impaired NK cytotoxicity during SIV infection (Fig. 5H). Similarly, levels of NK factors do not appear to be significantly altered following INC treatment or SIV infection (Fig. S6B and C). PNKC characteristics following depletion of platelets and their preformed cellular complexes We then sought to investigate the features of human PNKCs using an alternate in vitro method. During buffy coat isolation and NK cell negative selection, we noted incomplete removal of platelets, which is why we did platelet depletion washes and NK sorting in our previous in vitro experiments. Since we hypothesize that chronic platelet exposure may be driving PNKC phenotypes, we decided to compare the expansion of platelet-depleted and sorted NK cells (−/−) to nonplatelet-depleted and nonsorted NK cells (+/+) (Fig. 6A). PNKCs play important roles in total NK phenotypes before, during, and after expansion in vitro. NK cells were isolated from buffy coats. Some NK cells were sorted for the CD61−CD62P− NK cells (−/−) using a cell sorter to remove preformed PNKCs and residual platelets. Other NK cells were not sorted and the preformed PNKCs and residual platelets remained (+/+). Both groups were expanded for up to 14 days using the ImmunoCult NK Expansion Kit (STEMCELL Technologies). A) Example of platelet and PNKC depletion. B) Representative t-SNE analyses of NK expansion on days 0, 7, and 14. C to H) CD69, NKp44, IFN-γ, NKp30, NKp46, and NKG2D expression for platelet and PNKC depletion (−/−) (n = 5 to 6) versus no depletion (+/+) conditions (n = 7 to 8) (flow, MFI/%). Multiple Mann–Whitney test (*≤0.05). First, we observe via t-SNE that (+/+) NK cells have different phenotypes from (−/−) NK cells at all time points (Fig. 6B). Immediately after NK cell sorting, the (−/−) NK cells displayed reduced NK expression of CD69 (P = 0.0025) and increased levels of NKp44 (P = 0.0025) and IFN-γ (P = 0.0164) (Fig. 6C to E). It is expected that these phenotypic changes are due to PNKC removal since the cells were analyzed for phenotypic changes directly after isolation & depletion. While PNKCs are a small population, they significantly alter total NK phenotypes. To investigate the long-term effects of platelets on NK cell phenotypes, we expanded the NK cells after platelet and PNKC depletion for 14 days and conducted a multiparameter flow panel for NK phenotypic markers on days 0, 7, and 14. CD69 expression remains significantly lower for (−/−) NK cells on day 7 (P = 0.013), but not on day 14 (Fig. 6C). IFN-γ expression remained elevated in (−/−) NK cells on day 7 (P = 0.00067) and day 14 (P = 0.0012) (Fig. 6E). By day 14, (−/−) NK cells have enhanced NKp30 (P = 0.0012), and reduced NKp46 (P = 0.0012) and NKG2D (P = 0.0012) expression (Fig. 6F to H). Other NK phenotypic markers do not differ among the conditions on days 7 to 14 (Fig. S7A to F). Collectively, these experiments suggest that the phenotypic changes in NK cells were secondary to the long-term exposure to platelets. Discussion Platelet activation is increasingly recognized as a core feature of HIV pathogenesis and a driver of chronic immune dysregulation, even under suppressive ART.59 Multiple studies show that platelets form complexes with diverse leukocyte subsets and shape their phenotype and function,3,7–17 supporting a model in which early platelet-NK engagement could imprint later viral and inflammatory trajectories. During HIV infection, platelets are consistently activated, with elevated P-selectin, CD40L, and activation markers on platelets and platelet-derived products. Platelet activation markers correlate positively with plasma viral load and negatively with CD4 counts, underscoring their tight association with disease progression and immune activation.11,12,60 In this study, we present novel findings showing that the percentage of PNKCs (%PNKCs) formed in SIV-infected RMs on day 7 postinfection correlates with both plasma and CSF viral loads that were observed on day 35 postinfection. This supports the concept of early formation of PNKCs, as observed in SIV-infected RMs, which can serve as a mechanistic bridge between viral replication (in plasma and CSF) and long-term NK remodeling in PWH. This situates the percentage of PNKCs as a biologically plausible predictor and effector of subsequent viral and immune outcomes. This notion prompted investigations into the phenotypic changes occurring in NK cells among PWH. Previous research has assessed NK cell phenotypes in both PWH not on ART and those receiving ART,34,36,37,41 while Deng et al.3 recently explored the changes in NK phenotypes in PWH before and after ART. Persistent reshaping of NK cell phenotypes during suppressive ART is now well documented, with several studies showing only partial normalization of NK subsets and receptor expression in PWH. These changes are linked to suboptimal CD4+ T-cell restoration and ongoing immune dysfunction, supporting the interpretation that ART alone often fails to fully reconstitute innate immunity. PNKCs are reported to have increased activation markers and functional exhaustion, supporting a state of chronic hyperactivation followed by dysfunction.3 Consistent with previous findings, we observed a reduction in NKp46 expression and an increase in CD69 expression in NK cells from PWH on ART when compared to PWOH. Notably, we also identified previously unreported markers in PWH on ART, including reduced levels of NKG2A and DNAM-1, coupled with increased levels of NKp44 and TIGIT. These alterations indicate that NK cell phenotypes may not fully recover during ART, reinforcing similar observations made in viremic individuals.34,36,37 This warrants further investigation as ART duration is shown to influence NK cell restoration, and full normalization is rarely achieved even after many years61–65 Early ART initiation improves outcomes, but persistent deficits in NK cell function and phenotype are common, especially with longer ART duration.23,62 This highlights the need for adjunctive therapies to optimize immune recovery in HIV. In agreement with Deng et al.,3 we observed elevated ex vivo frequencies of PNKCs in PWH on ART, which were not confined to specific NK subsets defined by CD56/CD16 markers. Instead, in our studies, these PNKCs exhibited altered phenotypes likely stemming from inherent characteristics or being directly linked to HIV infection. Notably, CD69—an early marker of activation—was increased not only in PNKCs but also in other NK cell populations. This parallels studies in acute and chronic HIV where higher CD69+ NK frequencies correlated with higher viral load, lower CD4 counts, and poorer immune reconstitution, underscoring a tight link between CD69 upregulation and ongoing inflammation.32,65,66 In contrast, PNKCs displayed generally low NKp46 on all PNKCs, with additional decreases seen in PWH, consistent with chronic HIV-associated down-modulation of this activating receptor, persisting despite viral suppression.32,63,67 In vitro PNKC depletion experiments confirmed that alterations in PNKC levels can influence NK cell phenotypes throughout various stages of expansion. With this experiment, we can begin to understand the way PNKCs alter total NK phenotypes before, during, and after NK expansion, albeit in an HIV-free setting. Following PNKC removal, the phenotypic differences induced by platelets in NK cells were distinctly revealed. On Day 0, (+/+) higher levels of active NK cells (based on CD69 expression) were found to be predominant, with reduced NKp44 and IFN-γ expression. This supports our earlier observation indicating that PNKCs are phenotypically distinct from non-PNKCs. By Day 7, CD69 remains elevated and IFN-γ remains at lower levels in (+/+) NK cells compared to (−/−) NK cells. By day 14, NKG2D and NKp30 are elevated and NKp46 and IFN− γ are reduced in (+/+) NK cells compared to (−/−) NK cells. Throughout the expansion, we see 2 phenotypic markers, CD69 and NKp46, exhibit changes similar to what we see ex vivo in PWH. These observations have profound implications for understanding NK cell expansion dynamics during acute HIV infection. Together, our ex vivo and in vitro findings support a model in which HIV-driven and inflammation-linked phenotypic changes in PNKCs contribute to the persistently skewed, highly activated and functionally dysfunctional NK compartment observed in PWH on ART. Despite our findings, we faced notable limitations in studying the effects of chronic platelet exposure both in vitro and in vivo. For in vitro studies, platelets’ short lifespan and expression of major histocompatibility complex I complicated efforts to obtain fresh autologous platelets continuously over a 2-wk NK expansion period. Our in vitro coculture experiments involving platelets and NK cells did not yield substantial alterations in NK phenotypes or NK cytotoxicity, even with the addition of chronic platelet factors. This disparity may arise from our inability to adequately simulate prolonged platelet exposure in vitro. As another model of continuous platelet factor exposure, we conducted in vivo studies using simian immunodeficiency virus (SIV) infection in RMs. However, neither SIV infection nor INC treatment significantly altered PNKCs, NK phenotypes, or NK cytotoxicity during the study. The small sample size of 5 and the brief course of the infection model may not accurately reflect chronic infection dynamics, particularly since PWH on ART may have been infected for many years. We expect that both PNKCs and PNKC phenotypes are representatives of subtle changes that amplify over time during chronic infection with or without ART. Additionally, another important limitation of our study is the representation of female PWH. In our study, we only had 3 female PWH, with a majority of our PWH participants being male. This is an important limitation to consider as there may be sex differences in platelet biology. For example, 1 study noted that male PWH exhibit higher likelihood of thrombocytopenia,68 which could have implications for immune complex formation. While we did not find statistically significant sex differences in platelet complex formations, future studies with larger cohorts should consider addressing this. This study raises many questions and future avenues of research. First, the direct interactions between platelets and NK cells need to be investigated. If more than just CD62P and PSGL-1 are involved, additional drug treatments to complement INC may be needed to fully prevent formation of PNKCs or PLCs from forming. Second, the connection between NK dysfunction and chronic platelet exposure remains to be elucidated. Investigation into whether PNKCs are formed due to chronic HIV infection, ART therapy, or both, will provide insights into the mechanisms behind NK dysfunction in PWH. Third, it remains unknown whether platelets harboring HIV can disseminate the virus to NK cells, although observed correlation of in vivo % PNKCs with the viral load may lend support to this notion. Our lab previously established that platelets can harbor HIV and transmit it to T cells.18 If platelets transmit HIV to NK cells, then more PNKC interactions could mean more dissemination of HIV. However, if this were the case, it is interesting that we do not see the same association between PTCs and viral loads. Future studies might investigate what it is about PNKCs that may negatively impact HIV clearance. In summary, our research indicates that NK cell phenotypes remain dysfunctional even under suppressive ART, particularly among the subset of NK cells defined by their interactions with platelets (PNKCs), which are elevated in PWH on ART. This PNKC subset exhibits a unique phenotype that may exacerbate during chronic HIV infection, despite receiving effective ART. Understanding these dynamics is crucial for defining the role of PNKCs in the health outcomes of PWH and may lead to improved therapeutic strategies in the future. Supplementary Material