Enhancing antiretroviral delivery to secondary lymph nodes by targeting CD169 + macrophages with HIV-mimicking nanoparticles

Abstract

Despite significant progress in combinatorial antiretroviral therapy, people living with HIV remain susceptible to co-morbidities and increased risks of mortality. Chronic inflammation and immune activation are correlated with viral persistence in reservoir sites such as secondary lymph nodes and are postulated to be a driver of exaggerated risk of HIV-associated co-morbidities. Previous studies have revealed that low and heterogeneous penetration of antiretrovirals (ARVs) in lymph nodes can contribute to viral persistence. In addition, sub-optimal adherence to daily oral ARVs can lead to the development of antiviral resistance and viral rebound from these sanctuary sites. To overcome these deficiencies, we developed membrane-wrapped poly-lactic acid nanoparticles expressing the ganglioside, GM3 (GM3-NPs) and incorporating dual ARVs, Rilpivirine and Cabotegravir (CAB), for targeted delivery to lymph nodes. We have previously shown that GM3-CD169 mediated uptake of NPs results in their prolonged retention in CD169⁺ macrophage-associated nanoparticle-containing compartments (NPCCs) in vitro, resembling non-degradative virus (HIV-1)-containing compartments. Here, we demonstrate that these NPCCs are surface-accessible and can promote GM3-NP transmission to CD4⁺ T cells upon initiation of cell-to-cell contacts. Consequently, dual ARV-loaded GM3-NPs persisted in CD169⁺ NPCCs and promoted sustained inhibition of both cis- and trans-infection of macrophages and CD4⁺ T cells, respectively. Importantly, GM3-NPs specifically targeted and persisted in lymph node-resident CD169⁺ macrophages in mice. These findings suggest that GM3 incorporation facilitates targeted delivery of NPs to CD169⁺ myeloid cells in lymph nodes and might increase ARV distribution and persistence in lymphatic tissue reservoirs of HIV-1.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-29002-5.

Introduction

As of 2024, there were 40.8 million people living with HIV (PWH), 1.3 million new infections, and approximately 630,000 deaths due to AIDS-related illnesses¹. Despite great advances in combination antiretroviral therapy (ART), HIV remains a global health crisis. Lymph nodes are HIV reservoir sites that are central to viral pathogenesis and remain a major obstacle to HIV cure strategies as they harbor 36% of HIV RNA-positive cells². A major challenge to achieving a functional cure is viral persistence in tissue reservoir sites, such as lymph nodes, even in ART-suppressed PWH^3–8. Besides sub-optimal concentrations of antiretrovirals (ARVs) in lymph nodes, which has been postulated to contribute to viral persistence in these sanctuary sites^9–11, additional challenges include the requirement of daily pills, that result in poor treatment adherence and eventual viral rebound from secondary lymphoid tissues^12–15. Hence, there remains an unmet need to design long-acting formulations that will overcome ARV insufficiency in secondary lymph nodes by increasing ARV delivery and retention in these virus reservoir sites.

Over the years, there has been a tremendous effort to transition towards long-acting nanoformulations to overcome barriers to drug adherence^16,17. Cabenuva (Rilpivirine and Cabotegravir)¹⁸ and Sunlenca (Lenacapavir)¹⁹ are FDA-approved long-acting (LA) injectables for HIV suppressive therapy. Dapivirine vaginal ring²⁰ and Apretude (LA Cabotegravir)²¹ are approved LA therapies for pre-exposure prophylaxis (PrEP). Additional promising LA therapeutics are currently under investigation including nanoformulated prodrug injectables, pills, subdermal implants, hydrogels and microarray patches²². Overall, the biodistribution and, thus, the efficacy of these therapies depends on the physico-chemical properties of ARVs, which include low aqueous solubility, high lipophilicity, high potency and low pKa^11,23–25. In general, there remains a need to develop vehicles that will encapsulate ARVs for targeted delivery to the lymph nodes, irrespective of their physico-chemical characteristics. In this study, we describe the development of long-acting lipid polymeric nanoparticles (GM3-NPs) as a platform to increase dual ARV delivery and retention in lymph nodes and facilitate homogenous exposure of both ARVs to tissue-resident cells. GM3 is a ganglioside expressed on the viral membrane of HIV-1 particles that specifically binds to the lectin receptor CD169 on macrophages and dendritic cells^26–29. Importantly, CD169 is highly expressed in myeloid cells, especially in tissue-resident subcapsular sinus (SCS) macrophages and was demonstrated to be critical for HIV-1 accumulation and dissemination in secondary lymph nodes^30–32. Furthermore, CD169-GM3-dependent uptake of HIV-1 in macrophages and dendritic cells triggered the formation of non-degradative virus-containing compartments (VCCs)^33,34. VCCs are plasma membrane invaginations that remain surface-accessible and independent of the endolysosomal pathway^34–38. Viral particles retained within these non-acidic compartments were shown to be protected from degradation with extended infectivity^34,35. Additionally, VCCs mediate HIV-1 trans-infection of CD4 + T cells through virological synapses, an efficient strategy for cell-associated viral spread ^38–41.

Our previous studies characterized GM3-encapsulated nanoparticles with a biodegradable poly-lactic acid (PLA) core (GM3-NPs) designed to mimic the virion structure of HIV-1, which, upon binding to CD169 were trafficked to VCC-like compartments that we termed nanoparticle containing compartments (NPCCs)^42–44. Building on this, we describe development of lipid polymeric GM3-NPs encapsulating two ARVs, Rilpivirine (RPV) and Cabotegravir (CAB), that are designed to traffic to surface-accessible NPCCs upon capture by CD169⁺ macrophages. We hypothesized that CD169-mediated trafficking of GM3-NPs to NPCCs could be leveraged for efficient ARV transfer from CD169⁺ macrophages to CD4⁺ T cells during macrophage-T cell interactions, thereby suppressing HIV-1 infection of T cells. We further extended this approach to evaluate the potential of GM3-NPs for targeted ARV delivery to lymph node–resident CD169⁺ macrophages in vivo, with the goal of enhancing drug delivery and persistence within these key HIV reservoir sites. Here, we report localization of GM3-NPs within NPCCs in CD169 + macrophages, not only mediated sustained suppression of HIV-1 infection in macrophages, but also in co-cultured CD4 + T cells. Further, GM3-NPs extensively co-localized with CD169 + sub-capsular sinus (SCS) macrophages in lymph nodes of mice and persisted in lymph nodes for up to 35 days post-inoculation. Our findings introduce a novel strategy for the safe and efficient targeted delivery of ARVs to HIV tissue reservoirs for sustained viral suppression.

Results

Macrophage uptake of ARV-encapsulating GM3-NPs leads to ARV transfer to T cells

We synthesized GM3 expressing lipid-wrapped polymeric NPs that mimic the lipid membrane of HIV-1 as previously described⁴³, to facilitate NP delivery to CD169 + macrophages. We chose polymeric NPs with a biodegradable PLA core because of their high drug-loading capacity and slow-release properties that render them ideal candidates for high-dose drug delivery⁴⁵. The combination of RPV and CAB as ARVs was chosen due to the need for combination therapy in ARV regimens, their favorable nanoencapsulation properties, and lack of drug-drug interaction^46,47. GM3-NPs were also engineered with a lipid layer that confers a protein-repellent surface, to decrease non-specific interactions and protect polymeric nanocarriers from corona formation that can trigger rapid degradation^43,44. These NPs were synthesized through a one-step nanoprecipitation process as depicted in Fig. 1a. NPs were also fluorescently labeled with Cyanine 5 (1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-(Cyanine 5), Cy5) for ease of visualization by immunofluorescence. We also synthesized NPs with externalized phosphatidylserine (PS-NPs), which serves as an “eat-me” signal, phosphatidylserine^48–50. In CD169+ macrophages these PS-NPs are trafficked to endolysosomal compartments, instead of NPCCs⁴³.

Fig. 1.

Long-term HIV-1 suppression by RPV and CAB loaded GM3-NPs in CD169 + MDMs and co-cultured CD4 + T cells. (a) Representative scheme of RPV & CAB loaded NPs. Lipid membrane is composed of cholesterol (dark grey), DPPC (blue), GM3 or PS (red). PLA is used as a polymeric core (light grey) and ARVs, RPV (red) and CAB (yellow), are highlighted. (b) CD169 + MDMs were exposed to GM3-NPs, PS-NPs, or free ARVs at an initial concentration of 1 µM RPV for 3 h prior to infection with HIV-1Δenv/luc-G, (10 ng p24^Gag) on days 1, 7, 14, 21, 28, and 35 post-drug treatment. Infection levels were assessed 3 days post-infection and normalized against untreated cells (set at 100%). Each data point represents CD169 + MDMs derived from an independent donor. (c) MDMs were pre-treated with GM3-NPs, PS-NPs or free drug mixture at an initial 1 µM RPV concentration prior to initiation of co-cultures at indicated days. Co-cultures were infected with HIV-1Δenv/GFP-G (MOI 0.2) and cells harvested for flow cytometry at day 3 post infection. Each data point represents Jurkat cells co-cultured with MDMs from 4 different donors. The data is represented as %GFP + cells (mean ± SEM). Statistical significance was determined using one-way ANOVA followed by Tukey’s multiple comparisons test. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, and ****p ≤ 0.0001.

The average hydrodynamic diameter of drug loaded NPs, as determined by dynamic light scattering (DLS), was 345 ± 33 nm for GM3-NPs and 309 ± 15 nm for PS-NPs. The concentration of RPV in GM3 and PS NP formulations was 58.3 ± 7.9 μg/mL and 61.4 ± 70 μg/mL, respectively, while CAB concentrations were calculated to be 311.5 ± 47.2 μg/mL and 299.8 ± 90.3 μg/mL, respectively. These values correspond to encapsulation efficiencies of 28.0 ± 3.4% for RPV and 3.0 ± 0.4% for CAB in GM3-NPs, and 32.9 ± 7.0% and 3.3 ± 1.6% for RPV and CAB in PS-NPs.

Since constitutive CD169 expression in monocyte-derived macrophages (MDMs) is lower than in tissue-resident macrophages^43,51, we used IFN-λ to stimulate CD169 expression on MDMs (Fig. S2a, b), as described previously⁴³. We initially assessed the long-acting antiviral efficacy of NPs upon treatment of CD169 + MDMs with RPV/CAB loaded GM3-NPs, PS-NPs or unencapsulated soluble free RPV/CAB (free ARVs) as previously described⁴³. MDMs were exposed to soluble or NP-associated RPV/CAB (1 µM RPV-based input), returned to culture, and subsequently infected with a luciferase expressing VSV G pseudotyped single cycle HIV-1 (HIV-1Δenv/luc/G) at 1, 7, 14, 21 and 28 days post ARV exposure. Complete viral inhibition was achieved up to day 7 post-drug exposure across all treatments (Fig. 1b). However, differences in anti-viral efficacy emerged starting day 14, with GM3-NPs maintaining inhibition of HIV infection up to 28 days post-treatment (mean ± SEM) (18.3 ± 4.0%) compared to PS-NPs (45.8 ± 11.3%) or free ARVs (59.9 ± 4.0%) (Fig. 1b). These results are consistent with our previously published studies⁴³ and suggest that establishment of cell-associated ARV depots in GM3-NP-exposed CD169 + MDMs, can promote sustained inhibition of HIV-1 infection in MDMs.

We have previously demonstrated that drug depots within GM3-NP-exposed CD169⁺ MDMs can facilitate release of soluble ARVs into the extracellular space⁴³. Consequently, we sought to investigate whether extracellular ARV released from GM3-NP-exposed MDMs promotes sustained inhibition of HIV-1 infection in bystander CD4 + T cells. We established a direct co-culture system with pre-treated CD169⁺ MDMs and Jurkat T cells at a 1:4 ratio. Following drug exposure and extensive washes, MDMs were co-cultured with T cells for an additional 24 h. Subsequently, MDM-T cell co-cultures were infected with a GFP-expressing single cycle HIV-1 (HIV-1Δenv/GFP/G, MOI 0.2), and harvested 3 days post-infection. GFP expression in CD4 + T cells was determined via flow cytometry to measure virus infection establishment in T cells (Fig. 1c). This process was repeated at each time point (Day 1, 7, 14, 21, & 28 post drug treatment). On day 1, Jurkat T cells co-cultured with CD169 + MDMs exposed to ARVs for 1 day demonstrated similar levels of infection across the different drug modalities (mean ± SEM: 6.6% ± 2.0% for GM3-NPs, 9.0 ± 2.7% for PS-NPs and 5.0 ± 1.8% for free ARVs). However, differences in anti-viral efficacy between GM3-NPs, PS-NPs and free ARVs became apparent over time, with exposure to GM3-NPs resulting in significantly lower HIV-1 infection efficiency in Jurkat T cell/MDM co-cultures (% GFP + T cells was 14.4 ± 2.5% on day 7, 25.3 ± 0.7% on day 14, 36.5 ± 0.4% on day 21, and 49.1 ± 1.2% on day 28) compared to the no drug control cultures (Fig. 1c). In contrast, the anti-viral efficacy of PS-NPs and free ARVs was significantly attenuated by day 7 in Jurkat/MDM co-cultures (%GFP + T cells was 42.7 ± 2.3% and 44.9% ± 0.8%) compared to GM3-NP treated co-cultures (Fig. 1c). An additional loss in anti-viral efficacy was observed at days 14, 21 and 28 in Jurkat T/MDM co-cultures cells, treated with PS-NPs and free ARVs. Overall, these findings suggest that GM3-NPs retained within CD169 + MDMs resulted in sustained inhibition of HIV-1 infection in both macrophages and CD4 + T cells.

We next investigated how capture and retention of GM3-NPs within CD169⁺ MDMs promotes ARV transfer to bystander CD4 + T cells. Since our previously published findings⁴³ demonstrated that GM3-NPs localize within CD169 + , CD81 + non-lysosomal NPCCs, while PS-NPs co-localizes within Lamp1 + late endosomes, we hypothesized that the non-degradative nature of NPCCs facilitates extracellular release and transfer of GM3-NPs to CD4 + T cells. To that end, we first quantified the uptake of Cy5-labelled GM3-NPs and PS-NPs (6.8 × 10^4 NPs per cell) in CD169 + MDMs via flow cytometry and observed similar capture of both GM3- and PS-NPs by MDMs (Fig. 2a–c). Subsequently, to assess NP transfer, we established a direct co-culture system using CD169 + MDMs and primary CD4 + T cells from autologous donors. CD169 + MDMs pre-treated with GM3-NPs or PS-NPs were washed extensively and co-cultured with autologous CellTracker Blue-labeled CD4 + T cells. Transfer of NPs to T cells was determined via immunofluorescence microscopy 24 h post co-culture (Fig. 2d). Interestingly, despite the equivalent uptake of GM3 and PS-NPs in CD169 + MDMs (Fig. 2a–c), only GM3-NPs were easily detected in co-cultured CD4 + T cells (white arrows in Fig. 2d) as opposed to PS-NPs whose transfer to T cells was undetectable. Additionally, flow cytometry analysis for CD169^- CD3⁺ T cells harvested 24 h post initiation of co-cultures (Fig. S1), demonstrated association of GM3-NPs with T cells (mean ± SEM: 8.9 ± 1.8%) which was significantly higher than that observed in co-cultures with PS-NP-incubated MDMs (2.0 ± 0.34%) (Fig. 2e, f). Interestingly, GM3-NP transfer to CD4 + T cells was abrogated upon indirect co-culture (Transwell system) (Fig. 2g), suggesting that direct cell-to-cell contact was required for efficient GM3-NP exchange from MDMs to T cells.

Fig. 2.

Efficient transfer of GM3-NPs from CD169 + MDMs to CD4 + T cells. (a) Epifluorescence images depicting Cy5-labeled GM3 or PS-NPs (red) capture by CD169 + MDMs. Scale bar = 100 µm. (b) Flow cytometric analysis of GM3-NP and PS-NP uptake by CD169 + MDMs. (c) Mean fluorescence intensity (MFI) of NP uptake in MDMs. The data is represented as MFI (mean ± SEM) of GM3-NPs or PS-NPs uptake by MDMs from 5 independent donors. (d) Cy5-labeled GM3 or PS-NPs (magenta) transfer from CD169 + MDMs to CD4 + T cells (blue). Representative images of live MDM-CD4 + T cell co-cultures treated with Cy5-GM3-NPs (top) or PS-NPs (bottom) at 24 h post co-culture. CD4 + T cells with NPs identified by white arrows. Scale bar = 20 µm. (e, f) Representative flow cytometry analysis of CD4 + T cells 24 h post-culture with MDMs pre-treated with GM3 NPs (e) or PS-NPs (f). (g) Quantification of GM3-NPs and PS-NPs acquisition by CD4 + T cells upon co-culture with NP-pulsed MDMs. (h) Quantification of GM3-NP acquisition by CD4 + T cells upon (direct) or indirect (Transwell) co-culture with NP-pulsed MDMs. Each data point represents CD4 + T cells and MDMs from 5-6 different donors. Statistical significance was determined using paired t test. *p ≤ 0.05 **p  ≤ 0.01 ***p  ≤ 0.001 ****p  ≤ 0.0001

Persistence of GM3-NPs in NPCCs in macrophages facilitates NP transfer to CD4 + T cells

NP containing compartments (NPCCs) are non-degradative compartments similar to HIV virus-containing compartments (VCCs)^43,44. Since VCCs are surface-accessible plasma membrane invaginations in macrophages that facilitate long-term persistence of infectious HIV-1 particles and transfer to CD4 + T cells across synaptic junctions³³, we hypothesized that NPCCs could also enable sustained release of intact NPs to bystander cells, similar to transfer of MDM-associated infectious HIV-1 particles from VCCs to T cells. To assess the surface accessibility and temporal dynamics of NPCCs, CD169 + MDMs were exposed to GM3 or PS-NPs, extensively washed and returned to culture to allow NPCC formation. Subsequently, dextran (10 kDa), a fluid phase marker known to permeate through open intracellular compartments like VCCs³⁵, was added and confocal microscopy was used to assess the co-localization of NPs with open (dextran positive) compartments (Fig. 3a, b). There was extensive co-localization of GM3-NPs with dextran (Manders correlation coefficient, MCC 0.31 ± 0.09) compared to that observed with PS-NPs (MCC 0.10 ± 0.04) in CD169 + MDMs, suggesting that GM3-NPs are trafficked to surface-accessible plasma membrane-associated NPCCs.

Fig. 3.

Retention of GM3-NPs in surface accessible NPCCs in CD169 + MDMs. (a) Confocal microscopy images of CD169 + MDMs exposed to Cy5-labelled (red) GM3-NPs (top) or PS-NPs (bottom) for 30 min prior to LMW dextran (green) addition. (b) Mander’s correlation coefficient (M1) co-localization analysis of GM3-NPs or PS-NPs with dextran. Each data point represents analysis from a single Sect. (8-9 fields) with a total of 15-20 cells per field with cells derived from 3 independent donors. Statistical analysis was performed using paired t test. *p ≤ 0.05 **p≤ 0.01 ***p≤ 0.001 ****p ≤ 0.0001. (c) Representative confocal microscopy images of GM3-NPs (red) and dextran (green) in MDMs on Days 1, 7, 14, and 21 post-NP addition. Scale bar = 5 µm.

We subsequently investigated the persistence of these surface-accessible GM3-NPs containing NPCCs over time at days 1, 7, 14, and 21 post NP addition in CD169 + MDMs (Fig. 3c). Dextran was added to NP-harboring MDMs and co-localization of GM3-NPs and dextran was determined by confocal microscopy. Remarkably, even after this extended culture period, co-localization of GM3-NPs and dextran remained evident (Fig. 3c), indicating sustained retention of GM3-NPs within surface-accessible plasma membrane compartments in CD169 + MDMs.

Targeting CD169 + SCS macrophages increases penetration of NPs in secondary lymph nodes

To assess the targeted delivery potential of GM3-NPs to LNs, we compared the distribution of GM3-NPs, PS-NPs or BLK-NPs (ligand-deficient) in LNs upon sub-cutaneous administration in BALB/c mice. The size of these NPs, as determined by dynamic light scattering (DLS), was slightly smaller than their drug loaded counterparts: 216 nm ± 17.7 nm for GM3-NPs, 194.6 nm ± 5.9 nm for PS-NPs, and 216.9 ± 4.4 nm for BLK-NPs (mean ± SD; Fig. 4a), because of the absence of encapsulated cargo. The polydispersity index (PDI), which measures the size distribution heterogeneity (0–1 range), was found to be 0.2 for all NPs, indicating that the NP preparations were relatively homogeneous. Subsequently, NPs were administered subcutaneously (80-100µL, ~ 7 × 10^11 NPs) into the upper leg of BALB/c mice and draining inguinal lymph nodes were collected 24 h post-inoculation (Fig. 4b).

Fig. 4.

Subcutaneous injection of GM3-NPs in BALB/C mice results in dissemination from the infection site. (a) Size distribution of GM3-NPs (left) PS-NPs (middle) and BLK-NPs (right) measured by dynamic light scattering (DLS). (b) Schematic of experimental flow. (c) IVIS imaging of BALB/c mice on day 0 (top) and day 1 (bottom). The fluorescence signal is displayed as radiance efficiency. (d) Quantification of total radiance efficiency in (c) and is represented as mean ± SEM from 5 independent injections per NP. Statistical analysis performed by 2-way ANOVA with Sidak’s multiple comparisons test. *p ≤ 0.05, **p ≤ 0.01, ***p≤ 0.001, ****p≤ 0.0001.

To assess NP dissemination in mice, fluorescent NP presence in tissues and total radiance efficiency was quantified by IVIS (Fig. 4c, d). While fluorescence signal was predominantly restricted to the injection site in the upper leg on day 0 (Fig. 4c), likely due to the high and potentially saturating concentration of NPs present at the injection site, GM3-NPs appeared to migrate away from the injection site more extensively compared to PS-NPs and BLK-NPs (Fig. 4c, d, 2-fold decrease in total radiance efficiency at the injection site for GM3-NPs). These findings suggested that GM3 incorporation in the lipid membrane of NPs facilitates their drainage from the interstitium into circulation.

Subsequently, mice were euthanized 24 h post-injection, and the draining inguinal LNs were harvested and processed for immunohistochemistry (IHC) to visualize the spatial distribution of NPs in LNs (Fig. 5a, b, Fig. S3). GM3-NPs extensively co-localized with CD169 + SCS macrophages at 24 h post NP inoculation (Fig. 5a). Conversely, PS-NPs and BLK-NPs were found infrequently in draining LNs and co-localized with cells expressing low to minimal CD169 expression. Based on their localization, PS-NPs appeared to potentially accumulate in parenchymal regions of LNs (Fig. 5b), while BLK-NPs seemed to accumulate in sub-capsular region of LN, though they were not associated with CD169 + macrophages (Fig. 5b). Quantitative analysis of multiple lymph node sections using the Vectra Polaris Quantitative Pathology Imaging System and HALO software revealed a significantly higher presence of GM3-NPs (mean ± SEM, 8.5 ± 2.6%) in LNs compared to PS-NPs (0.5 ± 0.4%) and BLK-NPs (1.1% ± 1.3%) at 24 h post-inoculation (Fig. 5c). Additionally, we quantified accumulation of NPs in the popliteal LNs by flow cytometry and observed selective enrichment of GM3-NPs at 24 h post inoculation (Fig. S3d). To further characterize anatomical distribution of GM3-NPs in LNs, draining popliteal LNs were harvested 24 h following s.c. footpad injection with NPs and stained for CD169 (macrophages), CD3 (T cells) and B220 (B cells) expression. Accumulation of GM3-NPs in popliteal LNs was greater (mean ± SEM, 39.8% ± 7.1%) than that observed for PS-NPs (4.6% ± 2.3%) or BLK-NPs (5.0% ± 0.7%) (Fig. 6a, b). In addition to the predominant GM3-NP localization within the SCS region, GM3-NPs could also be identified within B cell follicles and paracortex, regions characterized by the intense B220 and CD3 staining, respectively (Figs. 6c, S4). Thus, incorporation of GM3 in lipid-polymeric NPs led to enhanced targeted delivery of CD169 + SCS macrophages in secondary lymph nodes.

Fig. 5.

Co-localization of GM3-NPs with CD169 + subcapsular macrophages in lymph nodes. (a) Representative images of inguinal LNs from mice 24-h post- NP inoculation. Fluorescent GM3-NPs (top), PS-NPs (middle) and BLK-NPs (bottom) are represented in red. Tissue cryosections were stained for CD169 (green) and DAPI (blue). (b) Magnification of NP-rich regions from the respective LN sections in (a). (c) Percentage of NP positive cells relative to the total number of DAPI positive cells in LNs and is represented as mean ± SEM. Each data point represents quantification of NP positive cells in LNs from one mouse (n = 5). Statistical analysis performed by one-way ANOVA followed by Tukey’s multiple comparison test. *p  ≤ 0.05, **p ≤ 0.01, ***pp ≤ 0.001, ****p ≤ 0.0001.

Fig. 6.

Anatomical distribution of GM3-NPs within lymph nodes. (a) Representative images of popliteal LNs from mice 1-day post-injection with GM3, PS and BLK-NPs. (b) Quantitative analysis of NP distribution in popliteal LNs. Data points represent 3 mice per treatment group (n = 3). Statistical analysis performed by one-way ANOVA followed by Tukey’s multiple comparison test. *p ≤ 0.05, **p 0.01, ***p≤ 0.001, ***p ≤ 0.0001 (c) Representative images of popliteal LN from a mouse inoculated with GM3-NPs (red), and stained with anti-CD3 (green) and anti-B220 (blue) antibodies. Bottom panels represent LN sections (from top panel) at a higher magnification.

Persistence of GM3-NPs in lymph node resident CD169 + macrophages

Given the potential of GM3-NPs to persist in CD169 + MDMs in vitro, we next sought to investigate whether long-term co-localization of GM3-NPs with CD169 + SCS macrophages could be recapitulated in vivo. To this end, Balb/c mice received a s.c injection in the upper leg and the distribution of GM3-NPs in inguinal lymph nodes was assessed over 35 days (Fig. 7a). For each timepoint (Day 1, 7, 14, 21 and 35), tissue sections were stained for CD169 and DAPI to assess the retention of GM3-NPs in the lymph nodes (Fig. 7b). While a temporal decrease in the intensity of NP fluorescence was observed, strikingly, GM3-NPs persisted in the lymph nodes for over 35 days (Fig. 7c), with the initial percentage of GM3-NPs present at 9.1 ± 1.9% on day 0, gradually decreasing to 1.8 ± 0.5% by day 35. Conversely, NP fluorescence in Balb/c mice inoculated with PS-NPs or BLK-NPs was undetectable by day 7 indicating rapid PS-NP or BLK-NP degradation or clearance (Fig. S5). We assessed the persistence of GM3-NPs in CD169 + SCS macrophages by determining the percentage of NP-colocalization with CD169 + macrophages over time. This was done using multiplex fluorescent immunohistochemistry and subsequent quantitative analysis was performed with the Halo area quantification software (Fig. 7d). On day 1 post-injection, a significant proportion of NPs co-localized with CD169 + SCS macrophages (46.5 ± 5.8%). Remarkably, GM3-NP co-localization with CD169 + macrophages remained consistently high until day 35 (80.4 ± 1.5%) (Fig. 7b, e). These findings suggest that GM3-NPs not only successfully localized to LNs but also persisted in CD169 + macrophages over time. Importantly, these findings demonstrate effective targeting and prolonged retention of GM3-NPs in SLT-resident macrophages.

Fig. 7.

Persistent association of GM3-NPs with lymph node-resident CD169 + macrophages. (a) Schematic of experimental workflow to assess persistence of GM3-NPs in LNs over 35 days. (b) For each timepoint, LN sections were stained for CD169 (green) and DAPI (blue). GM3-NPs are shown in red. (c) Quantification of GM3-NPs in LNs over 35 days. (d) Illustration of the analysis strategy performed using Halo software. All images were acquired via whole-slide scanning and analyzed with Halo software. (e) Quantification of GM3-NP co-localization with CD169 + macrophages, expressed as %CD169 + NP + cells (CD169 + cells containing NPs relative to the total number of NP-positive cells) at indicated days post NP injection. Data points represent 2–3 mice per timepoint (mean ± SEM). Scale bar = 100 μm.

Discussion

In this study, we leveraged the previously well-characterized CD169-GM3 dependent HIV-1 capture and trans infection mechanism²⁶ to develop and characterize ARV-loaded GM3-NPs that mimic CD169-mediated trafficking of HIV, thereby enhancing ARV delivery and retention in NPCCs in macrophages and promoting subsequent transfer to bystander CD4 + T cells. We show that NPCCs are accessible to membrane-impermeable tracers such as fluorescent dextran, confirming their extracellular connectivity, like VCCs. Additionally, our results demonstrate that prolonged suppression of HIV-1 infection in macrophages also extends to bystander CD4 + T cells through efficient drug and NP transfer from macrophages to T cells, upon initiation of cell-to-cell contacts. While we did not directly image NP trafficking from NPCCs into CD4 + T cells, the detection of NPs in co-cultured CD4 + T cells, combined with the known open nature of NPCCs, suggests that NPCCs may facilitate transfer to neighboring cells via membrane contact or synapse-like structures akin to HIV trans-infection.

Our findings in mice support enhanced delivery and prolonged retention of GM3-NPs in LNs via successful targeting of CD169 + subcapsular sinus (SCS) macrophages. Given the preponderance of CD169 + macrophages in secondary lymphoid tissues, notably in secondary LNs, these results suggest that GM3-NPs could improve the dissemination and retention of ARVs in LNs, supporting their potential utility in long-acting therapy. Interestingly, GM3-NPs infiltrated follicular and paracortical regions in the LN, though the mechanism underlying their distribution to parenchymal regions of LNs remains unclear. Sewald et al. previously demonstrated that the capture of murine leukemia virus (MLV) and HIV-1 particles by SCS macrophages in lymph nodes through GM3-CD169 interactions resulted in virus particle concentration within VCCs of CD169 + myeloid cells, some of which extended to B cell follicles for viral delivery⁵². In our study, we observed the formation of NPCCs in CD169 + MDMs in vitro, but have not assessed their presence in SCS macrophages in vivo. However, our IHC analysis demonstrated prolonged retention of GM3-NPs in SCS macrophages with visible membrane protrusions into the B cell follicles where NPs were subsequently localized (Fig. S4). Based on these observations, we hypothesize that GM3-NPs could potentially be transferred to follicular resident T cells via surface-accessible CD169 + NPCCs.

Lymph node B cell follicles are thought to be one of the major sites of HIV-1 persistence in chronic infection, and while frequency of HIV RNA + cells in LNs and B cell follicles is reduced upon ART initiation, RNA + cells are still detected in B cell follicles during ART in both HIV and SIV infection ^53–56. CD4 + T follicular helper cells (Tfh) are the predominant CD4 T cell memory population harboring replication competent virus during ART⁵⁷. While enhanced susceptibility of CD4 Tfh to HIV infection and their close proximity to follicular dendritic cells displaying infectious HIV-1 particles has been hypothesized to contribute to virus persistence in B cell follicles^58–60 , suboptimal immunological and pharmacological control due to incomplete CTL and ARV penetrance can also contribute to follicular concentration of HIV RNA + cells ^61–63. Some studies have suggested that more than 90% of viral RNA-positive cells remain unexposed to ARVs⁶², creating pharmacological sanctuaries that can promote chronic inflammation and viral rebound upon treatment interruption^64–68. Consequently, achieving effective ARV delivery to B cell follicles via GM3-NPs and increased ARV presence throughout the lymph nodes including follicular and paracortical regions of LNs, might be an effective strategy to overcome pharmacological recalcitrance of anatomical reservoirs of HIV.

CABENUVA is currently the only combinatorial LA formulation approved for HIV-1 treatment. It consists of two separate intramuscular injections of RPV and CAB administered simultaneously. However, mass spectrometry imaging (MSI) studies have demonstrated that the s.c administrations of multiple ARVs simultaneously does not guarantee their co-delivery to tissue reservoirs, including LNs, spleen, gut and brain^69–71. This limitation, which fails to address the heterogeneous ARV distribution, can contribute to the development of HIV drug resistance^72–74, ultimately leading to treatment failure. In this work, we developed a platform for dual encapsulation of RPV and CAB into a single NP, potentially facilitating the co-delivery of ARVs to anatomical sites and increasing their spatial co-localization in target cells. Targeting tissue-associated myeloid cells promises a long retention time in LNs, which remains difficult to achieve with conventional delivery strategies. While we acknowledge that ARV concentrations in LNs of mice inoculated with RPV/CAB-loaded GM3-NPs need to be assessed prior to establishing this targeting modality for therapeutic relevance, the detection of GM3-NPs in follicular areas, along with the potential for efficient macrophage-T cell NP transfer, might overcome existing barriers to achieving homogenous ARV concentrations across secondary lymphoid tissues.

Materials and methods

Ethics statement

This research has been determined to be exempt by the Institutional Review Board of the Boston University Medical Center since it does not meet the definition of human subjects research, since all human samples were collected in an anonymous fashion and no identifiable private information was collected.

Fabrication of ARV-loaded lipid-coated polymer NPs

Drug loaded lipid-coated polymer NPs were fabricated via one-step nanoprecipitation procedure as described previously⁴³ . Briefly, 20 μg of Rilpivirine (RPV, MedKoo Biosciences, Inc., Durham, NC) from a stock solution of 100 μg/mL in Acetonitrile (ACN, Sigma-Aldrich, St Louis, MO) and 1 mg of Cabotegravir (CAB, Selleckchem Houston, TX) from a stock solution of 10 mg/mL in dimethyl sulfoxide (DMSO, Sigma-Aldrich, St Louis, MO) were mixed with 1 mg of Poly(D,L-lactide) (PLA, Resomer R207S, 209 kDa, ester terminated polymer) solution in ACN (2.5 mg/mL stock solution in ACN) and vortexed.

All lipids were dissolved in chloroform except for GM3, which was dissolved in a chloroform and methanol mix at a ratio of 3:1 (v:v). Using a lipid/polymer weight ratio of 15%, GM3-NPs were prepared by adding 0.15 mg of lipid mixture containing 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC, 57 mol%, 25 mg/mL), cholesterol (40 mol%, 25 mg/mL), and GM3 Ganglioside (Milk, Bovine-Ammonium Salt) (3 mol%, 2 mg/mL) to 4 ml sterile, cell culture grade water (Gibco Distilled Water, Thermo Fisher Scientific, Waltham, MA). To fabricate negatively charged polymer NP controls (PS-NPs), 10 mol% 1,2-dioleoyl-sn-glycero-3-phospho-L-serine (DOPS, 10 mg/mL) was used instead of GM3 and DPPC mol% was adjusted to 50 mol%. For experiments involving fluorescence microscopy or flow cytometry, fluorescent marker 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-(Cyanine 5) (Cy5, 1 mol%, 1 mg/ml) was added to the lipid mixture. The DPPC mol% was then adjusted to 56 mol% for GM3-NPs and 49 mol% for PS-NPs. All lipids were purchased from Avanti Polar Lipids (Alabaster, AL). Next, drug containing PLA solution in acetonitrile was added dropwise to the aqueous solution containing lipid mixture. The final mixture was sonicated in a bath sonicator (Branson 5510, Branson Ultrasonics, Danbury, CT) for 5 min. Following this, the drug loaded lipid-coated polymer NPs were purified via 3 washing cycles (4000 g–15 min) using an Amicon Ultra-4 centrifugal filter (Millipore Sigma, Burlington, MA) with a molecular weight cutoff of 10 kDa to remove organic solvent, free drug, and lipid molecules. For synthesis of ARV-free NPs (ARV-free), fluorescently labeled lipid-coated polymeric NPs for in vivo applications were fabricated following a similar nanoprecipitation strategy as outlined above for the drug-loaded NPs. For size measurement, dynamic light scattering (DLS, Zetasizer Nano ZS90, Malvern, Worcestershire, UK) was used while concentration was determined by measuring absorbance on UV–Vis (FLAME-T-UV–VIS, FLMT08476, Ocean Insight, Orlando, FL) and applying Beer’s law as previously described⁴³. Briefly, the concentration of dye molecules per NP preparation was determined by extinction measurements using the known extinction coefficient of Cy5. Using the mol composition of the membrane, we then determined the total number of lipids in each preparation. Assuming a lipid cross-section of 0.6 nm^2, we calculated the total surface area of the lipid coating. Finally, this area was divided by the surface area per NP (the nanoparticle diameter was determined by DLS) to determine the total number of NPs.

Quantification of drug concentrations

To quantify the amount of encapsulated drug (RPV and CAB) in GM3 and PS NPs, particles were dissolved in ACN, and the resulting free drugs from particles were measured using an Agilent HPLC with a diode array detector (DAD). RPV was detected at the wavelength detection of 290 nm at an average retention time of 7.6 min, while CAB was detected at 254 nm at an average retention time of 2.9 min. Standard solutions for the calibration curve were run with the samples for each HPLC run. The encapsulation efficiency percentages were determined using the following formula.

Encapsulation efficiency (%) =  × 100.

Cells

HEK293T (ATCC Manassas, VA, USA; Cat# 3216) and TZM-bl (BEI Resources Manassas, VA, USA, Cat# ARP-8129) cells were cultured in DMEM (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) containing 10% FBS (Gibco, Thermo Fisher Scientific, Waltham, MA, USA, Cat# 16,000,044) and 1% pen/strep (Gibco, Thermo Fisher Scientific, Waltham, MA, USA, Cat# 15,070,063). Jurkat E6 T cells (BEI Resources Cat# ARP-177) were maintained in RPMI-1640 (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) containing 10% heat-inactivated FBS (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) and 1% pen/strep (Gibco, Thermo Fisher Scientific, Waltham, MA, USA). All cell lines were tested for mycoplasma contamination and confirmed negative. Human MDMs were derived from positively isolated CD14 + peripheral blood monocytes by culturing in RPMI-1640 containing 10% heat-inactivated human AB serum (Sigma-Aldrich, St. Louis, MO, USA) and recombinant human macrophage colony stimulating factor (M-CSF) (20 ng/mL; PeproTech, Cranbury, NJ, USA) for 5 days, as described previously⁴³. CD169 expression on MDMs was induced upon treatment with IFN-λ (5 ng/ml) for 24–48 h. Cell surface expression of CD169 was confirmed by FACS using an Alexa 647-conjugated mouse anti-CD169 antibody (BioLegend, San Diego, CA, USA). Autologous CD4 + cells were isolated from CD14- flowthrough and positively selected using anti-CD4 antibody-coated beads (Miltenyi Biotec, Bergisch Gladbach, Germany; Cat# 130–050-201). CD4 + T cells were activated with 2% Phytohemagglutinin (PHA; Thermo Fisher Scientific, Waltham, MA, USA; Cat# 10,576–015) in R10 media supplemented with recombinant human interleukin-2 (IL-2, 50 µg/ml) for 48 h, and then maintained in R10 media supplemented with IL-2 (50 µg/ml).

Viruses

Single-round replication-competent HIV-1 particles pseudotyped with VSV-G and expressing either luciferase (HIV-1Δenv/luc-G) or GFP (HIV-1Δenv/GFP-G) reporters, were produced in HEK293T cells by transient co-transfection of HIV-1Δenv/luc or HIV-1Δenv/GFP proviral plasmids, along with a VSV-G expression plasmid, as described previously³⁹. Virus-containing cell supernatants were harvested 2 days post-transfection, cleared of cell debris by centrifugation (300 × g, 5 min), passed through 0.45 µm filters, and concentrated by ultracentrifugation on a 20% sucrose cushion (1000,000 × g for 2 h with a SW28 rotor (Beckman Coulter, Brea, CA, USA). The virus pellets were resuspended in PBS (Invitrogen, Waltham, MA, USA; Cat# 14,190–250), aliquoted and stored at − 80 °C until use. The capsid content of viral stocks was determined by a p24^gag ELISA³⁹ and virus titer was measured on TZM-bl cells.

Infections

CD169 + MDMs were seeded in a 96-well plate at a density of 5 × 10⁴ cells per well. Soluble drugs or ARV-loaded polymer nanoparticles were administered to the cells at a concentration equivalent to 1 µM of Rilpivirine (RPV), while Cabotegravir (CAB) concentration ranged from 6 to 33 µM⁴⁵. Cells were incubated at 37 °C for 3 h with ARVs, washed extensively to remove unbound drug, before returning to culture. ARV-treated cells were infected with HIVΔenv/luc-G via spinocculation on days 1, 7, 14, 21, 28, and 35 post-drug treatment. Cells (MDMs) were lysed 3 days post-infection, and luciferase expression in the cell lysates was quantified using the Bright-Glo luciferase expression system (Promega Corporation, Madison, WI, USA), as previously described⁴³. Alternatively, ARV-exposed MDMs (of 5 × 10⁴ cells) were co-cultured with Jurkat T cells (2 × 10⁵ cells) overnight on days 1, 7, 14, 21, 28, and 35 post-drug treatment. Subsequent to overnight MDM-Jurkat co-culture, cells were infected HIVΔenv/GFP-G (MOI 0.2) in the presence of polybrene (10 μg/mL final, MilliporeSigma, Burlington, MA, USA; Cat# TR-1003-G)). Co-cultures were harvested 3 days post-infection stained with PE-conjugated anti-CD3 antibody (Fisher Scientific, Hampton, NH, USA; Cat# 347,347) and processed for flow cytometry analysis.

Assessing NP transfer from CD169 + MDMs to CD4 + T cells

CD169 + MDMs were seeded in a 24-well plate at a density of 1.5 × 10⁵ cells per well and incubated with GM3-NPs and PS-NPs (1.0 × 10¹⁰ NPs per well) for 30 min for analysis by flow cytometry or 1 h for immunofluorescence imaging. Cells were washed extensively to remove excess unbound NPs. Uptake of NPs by MDMs was assessed by flow cytometry. To determine NP transfer from MDMs to T cells, autologous primary CD4 + T cells were added to MDMs (direct culture) or seeded on a Transwell membrane (Corning Transwell, Fisher Scientific, Hampton, NH, USA; Cat# 07–200-147) for indirect co-cultures, at a density of 4.5 × 10⁵ cells per well and cultured overnight. Cells were stained with PE-conjugated anti-CD3 (Fisher Scientific; Cat# 347,347) and Alexa 647-conjugated mouse anti-CD169 antibody (BioLegend, San Diego, CA, USA; Cat# 346,006) and processed for flow cytometry analysis. Transfer of NPs was assessed using the gating strategy depicted in Fig. S1. For imaging studies, live cell co-cultures were imaged using the EVOS microscope.

Immunofluorescence microscopy

CD169 + MDMs, (5 × 10⁵ cells on a glass bottom 35 mm culture dish (Fisher Scientific, Waltham, MA, USA; Cat# 150,680) were incubated with Cy5-labelled GM3-NPs or PS-NPs (1.0 × 10¹⁰ NPs) for 1 h at 37 °C. Cells were washed extensively with cold PBS and incubated with Dextran-AF488 (100 µg/mL, Fisher Scientific, Waltham, MA, USA; Cat# D22910) for 15 min at 4 °C. Cells were fixed, and counter-stained with DAPI for nuclear staining. For visualizing long-term GM3-NP co-localization within NPCCs, R10 media was added post-wash and cells were returned to culture at 37 °C. At each time point, cells were washed with cold PBS, stained with Dextran-AF488 and DAPI, and processed for confocal microscopy. Images were acquired using a Leica SP5 confocal microscope and analyzed using ImageJ software and the JACoP plugin was used to measure the co-localization of Cy5-NPs and dextran through Mander’s coefficient analysis.

Animals

Female BALB/c mice (17–20 g, 6–8 weeks) were purchased from Jackson Laboratories (Bar Harbor, ME, USA). Mice were housed under a 12 h light/dark cycle in a pathogen-free environment with a standard chow diet provided ad libitum. All animal experimental procedures were approved by the Boston University Institutional Animal Care and Use (IACUC), and experiments were performed in accordance with institutional (Boston University IACUC) guidelines and regulations. All methods are reported per ARRIVE guidelines.

Nanoparticle injections

For studies assessing NP distribution, mice under isoflurane anesthesia were administered a single subcutaneous (s.c) injection of 6.5–7 × 10¹¹ NPs of GM3-NPs, PS-NPs and BLK-NPs on the right upper leg, or 2.8 × 10¹¹ GM3-NPs for footpad injections using a syringe with a 26G × 3/8 needle (McKesson, Richmond, VA, USA; Cat# 1,057,162). For studying long-term retention of GM3-NPs in LNs, mice received a single s.c injection of 10¹⁷ NPs.

In vivo fluorescence imaging (IVIS)

Distribution of NPs was first determined by whole-body imaging using IVIS. At each time point, mice were shaved on the leg to increase sensitivity of signal. Subsequently, they were put under isoflurane anesthesia and imaged using excitation and emission filters at 640 and 680 nm, respectively. All images were acquired under the same field of view using the autoexposure parameters which corresponded to binning factor 4 and f-stop 8. Analysis was then performed using the Living Image version 4.4.7 software. Fluorescence signal was manually adjusted to remove background, and quantification was done by drawing region of interests (ROI) around areas with detectable fluorescence intensities. Measurements were reported as total radiant efficiency [p/s/sr]/[µW/cm2].

Immunohistochemistry

Inguinal or popliteal lymph nodes (LNs) were excised, washed twice with PBS, fixed with 4% paraformaldehyde (PFA, Boston bioproducts, Boston BioProducts, MA, USA; Cat# BM-155) for 30 min, washed once prior to overnight incubation in 20–30% sucrose solution at 4 °C. Tissues were embedded in OCT on a based mold (Fisher Scientific, MA, USA; Cat# 22–363-555), snap frozen on dry ice and cryo-sectioned. Tissue sections were blocked with Dako Protein Block, Serum-Free (Agilent Technologies, CA, USA; #X090930-2) for 30 min at room temperature. Sections were stained with FITC-conjugated rat anti-CD169 (1:100, Bio-Rad, CA, USA; Cat# MCA947F), Alexa594-conjugated rat anti-CD45R/B220 antibody (1:500, BioLegend, CA, USA; Cat# 103,254), and Alexa488-conjugated rabbit anti-CD3 (1:100, Fisher Scientific, MA, USA; Cat# 53-0032-82) overnight at 4 °C, prior to staining with DAPI (200 ng/ml, Sigma-Aldrich, MO, USA). Sections were mounted with Fluoromount-G (SouthernBiotech, AL, USA; Cat# OB100-01). Fluorescent images were acquired by confocal microscopy (Leica SP5) or whole-slide imaging.

Whole slide imaging

Imaging of tissue containing slides was performed using the Vectra Polaris Quantitative Pathology Imaging System (Akoya Biosciences, USA). Unstained slides were used as negative controls for background fluorescence. Exposure for all fluorophores was adjusted based on regions of the lymph nodes with the strongest signal, to maximize signal specificity. Generated images were analyzed using the HALO software (HALO v3.64234.263 & HALO AI v3.64234, Indica Labs, Albuquerque, NM, USA). First, using the view settings, fluorescence signals were optimized by adjusting minimal intensities based on background fluorescence from unstained slides. Following this, each LN section was annotated by delineating the area of analysis using the flood fill annotation tool. For fluorescence quantification, the immunophenotyping Highplex FL v4.2.3 module was used to assess the number of NP + cells and quantify the percentage of NP + cells that were CD169 + . To this end, membrane and cytoplasm detection thresholds were optimized using the real-time tuning feature (Fig. 7d). Analysis results from the algorithm were reported as total number of cells in the area, based on DAPI staining.

Statistical analysis

All statistical analysis was performed using GraphPad Prism 9. Unless stated otherwise, P-values were calculated using t-tests, one-way or two-way ANOVA, followed by the Tukey multiple comparison post-test, or Sidak’s multiple comparisons test, respectively. Symbols represent *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. No symbol or ns: not significant (p ≥ 0.05).

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

J.F., B.M.R. and S.G. designed the experiments. S.Z., and H.Z. performed nanoparticle synthesis and physical characterization (DLS and HPLC) for all the NPs used throughout the studies. T.C. performed flow cytometry to detect NPs in lymph nodes. J.F conducted all other experiments and data analysis. J.F, B.M.R., and S.G. wrote the manuscript, S.Z and H.Z reviewed and provided valuable insight.

Funding

National Institute of Allergy and Infectious Diseases, F31AI172625, R01AI132111, R01AI132111.

Data availability

All data generated or analyzed during this study are included in this published article (and its Supplementary Information files) and are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.