Army liposome formulation containing QS-21 render human monocyte-derived macrophages less permissive to HIV-1 infection by upregulating APOBEC3A

Monocyte-derived macrophages (MDM) are highly permissive to HIV-1 infection potentially due to the downregulation of innate factors during the differentiation process. The environmental milieu and innate anti-viral factors which are modulated during macrophage differentiation, have been associated with their increased permissiveness to HIV-1 infection. Here, we demonstrate that the Army Liposome Formulation containing MPLA, and QS-21 (ALFQ) activated MDM that are normally permissive to HIV-1 infection to generate a proinflammatory environment and upregulated anti-viral factors notably APOBEC3A. Induction of APOBEC3A by ALFQ decreased permissiveness to HIV-1 infection, while knockdown of APOBEC3A with APOBEC3AsiRNA resulted in a significant loss in the restriction of HIV-1 infectivity. The liposome formulation ALF55, with identical lipid composition but lacking QS-21 had no effect. Furthermore, the capacity of ALFQ to modulate MDM permissiveness to HIV-1 infection was predominantly mediated by large ALFQ liposomes. Our findings highlight a relationship between innate immune activation, proinflammatory milieu, and upregulation of anti-HIV proteins. Induction of these responses can switch the HIV-1 permissive MDM into a more refractory phenotype.

MDM. The characteristics of the liposome formulations are described in Table 1 and Fig. 1. ALF55 and ALFQsuv particles are small unilamellar nanovesicles (SUV) with a size range of 50-100 nm (Fig. 1a, c) whereas ALFQ is polydisperse and comprised both small and large unilamellar vesicles (Fig. 1b). The size of ALFQ vesicles ranged from 100 nm to several µm as shown in Fig. 1b, and from previously published data 24 . The MDM cultures were CD14 + CD11b + MHCII + and did not contain CD3 + T cells (Fig. 2a). A schematic of the pre-treatment of MDM with the formulations is shown in Fig. 2b. The cells were harvested and analyzed for apoptosis and necrosis by Annexin V and 7AAD staining. Culture supernatants were analyzed for the presence of lactate dehydrogenase (LDH) as an indicator of pyroptosis. The percentages of apoptotic and necrotic MDM as measured by Annexin V and 7AAD staining was relatively low in the three donors and was comparable between the untreated (no adjuvant) and treated MDM cultures (Fig. 2c, d). Pre-treatment with ALF55 or ALFQ did not induce LDH production (data not shown).

ALFQ induced higher levels of proinflammatory cytokines and chemokines.
Having determined that ALF55 and ALFQ did not induce apoptosis of MDM (Fig. 2), we evaluated the profile and levels of cytokines and chemokines that were induced in the culture supernatants following exposure of MDM to ALF55 and ALFQ. The radar plots (Fig. 3) show the changes in the cytokine and chemokine levels induced by ALF55 or ALFQ in relation to the unstimulated MDM cultures.
Although the MDM from the two donors exhibited variability in the magnitude of their responses, culturing MDM with ALF55 and ALFQ induced significantly higher levels of all the cytokines and chemokines, except for MCP-1, compared to the media control. Comparative analyses between ALF55 and ALFQ showed induction of cytokine responses that were distinct in both magnitude and profile. MDM cultured with ALFQ generally www.nature.com/scientificreports/ secreted higher levels of proinflammatory cytokines and chemokines than MDM cultured with ALF55 as reflected by the significantly higher levels of IL-12p70, IFNγ, GM-CSF, IL-1β, IL-18, TNF-α, IP-10 and MIP-1α (Fig. 3). ALF55 or ALFQ did not induce IFNα in any of the donors. These results indicate that MDM respond to different environmental stimuli during their differentiation process, and this determines the profile and magnitude of soluble mediators that are secreted in the supernatant.

ALFQ upregulated expression of MHC-II and CD86.
We characterized the effect of ALF55 and ALFQ on the expression of antigen presentation receptors (MHC-I, MHC-II), co-stimulatory receptors (CD80, CD86), phagocytosis-associated receptors (CD16, CD32) and HIV-capture associated receptor sialic acid-binding immunoglobulin-like lectin-1 (Siglec-1) on MDM. We previously demonstrated that Siglec-1 on MDM is an important attachment receptor for HIV-1 and has a profound influence on HIV-1 infectivity 25 . The surface expression of these receptors was evaluated by assessment of the mean fluorescent intensity (MFI) (Fig. 4) and the quantitation (Fig. 5) of the cellular receptors 24 h post-treatment with the adjuvants. ALF55 and ALFQ differentially influenced MHC receptors, co-stimulatory molecules and Fc-gamma receptors. ALFQ significantly increased the expression of MHC-II, CD80, and CD86 and correspondingly decreased the expression of CD16 when compared to ALF55 or the unstimulated (no adjuvant) controls (Fig. 5). Interestingly, ALF55 significantly decreased the expression of CD86 when compared to the unstimulated controls (Fig. 5). Although the number of CD32 receptors was comparable between MDM exposed to ALFQ or ALF55, they were significantly lower in the ALFQ-exposed MDM when compared to the unstimulated controls. The number of MHC-I receptors was comparable between ALF55-exposed MDM and the unstimulated controls. In contrast, MHC-I expression was significantly decreased in the ALFQ-exposed MDM when compared to ALF55-exposed or the unstimulated (no adjuvant) controls (Fig. 5). These results suggest that ALFQ could enhance the capacity of MDM to present antigen through the MHC-II antigen presentation pathway with a concomitant modulation in receptor-mediated phagocytosis. The observed effects may be a direct consequence of the immunomodulator(s) or an indirect result of the cytokines secreted in response to immunomodulators that influences the immune environment in which the cells are undergoing differentiation.

Pre-treatment with ALFQ modulates permissiveness of MDM to HIV-1 infection.
We previously demonstrated that MDM comprised two populations: a non-adherent (CD14 + CD4 + Siglec-1 hi ) population refractory to HIV-1 infection and an adherent (CD14 + CD4 -Siglec-1 l°) population which was highly permissive to HIV-1 infection 26 . Therefore, we assessed the influence of ALF55 and ALFQ on the permissiveness of adherent MDM to HIV-1 infection. MDM were either untreated (no adjuvant) and served as infection control or pre-treated with ALF55 or ALFQ for 24 h. Non-adherent MDM were removed, and the remaining adherent MDM were subsequently infected with US-1 as previously described 26 . MDM were harvested on day(s) 1 and 5 post-infection and analyzed for the presence of intracellular p24 by flow cytometry (Fig. 6a, b). When compared to the infection controls, pre-treatment with ALF55 or ALFQ did not modulate the permissiveness   Fig. 6a, b). Furthermore, whereas the percentage of HIV-1-infected MDM was not statistically different between day 1 and day 5 post-infection in the ALFQ-treated MDM, they were significantly increased between day 1 and day 5 post-infection in the untreated and ALF55 pre-treated MDM. These results indicate that exposure to ALFQ render adherent MDM significantly less permissive to HIV-1 infection. One probable reason for the lower number of infected cells in the ALFQ-treated cultures may be because infected MDM from the ALFQ-treated cultures became apoptotic instead of being productively infected. Although our data (Fig. 2) showed that ALFQ by itself did not induce apoptosis or pyroptosis, it is possible that ALFQ-treated cultures may be more susceptible to apoptosis when infected with HIV-1. However, as shown in Fig. 6c, the percentages of apoptotic and necrotic MDM were comparable between the untreated (no adjuvant) and adjuvant-treated MDM at each time-point and ranged between 7.3 and 13.6% in the untreated cultures, 8.4-11.9% for ALF55 and 6.1-9.2% for ALFQ at day 1 post infection., and 10. 6-14.2% in the untreated cultures, 10.6-17.7% for ALF55 and 10.6-18.7% for ALFQ at day 5 post infection. (Fig. 6c), suggesting that apoptosis and or necrosis of HIV-1 infected cells was not the cause of reduced infection in ALFQ treated cultures.
Another potential reason for the lower number of infected cells in the ALFQ-treated cultures may be because ALFQ modulated the number of viruses that was captured by the ALFQ-treated cultures during the initial HIV-1 exposure. Although our data (Fig. 5) showed that ALFQ did not modulate expression of Siglec-1, an important HIV-1 attachment receptor, we determined whether HIV-1 capture and entry was decreased in the ALFQtreated MDM. Cells were lysed 1 h post-infection of the MDM and virus capture and entry was determined by the presence of RNA gag using qRT-PCR 27 . The number of HIV-1 RNA copies per 10 6 MDM was comparable in the untreated and adjuvant-treated MDM (Fig. 6d) suggesting that pre-treatment with ALFQ did not modulate attachment and entry of HIV-1 in MDM, but influenced the HIV-1 replication capacity.
As shown in Fig. 1b, ALFQ is polydisperse and comprise both small and large unilamellar vesicles. In order to determine whether the decreased permissiveness to HIV-1 infection following exposure of MDM to ALFQ was influenced by liposome size, ALFQ was microfluidized to generate small ALFQ liposomes (ALFQsuv). MDM   www.nature.com/scientificreports/ were either untreated (no adjuvant) and served as infection control or pre-treated with ALF55 , ALFQ or ALFQsuv for 24 h. Non-adherent MDM were removed, and the remaining adherent MDM were subsequently infected with US-1 for 5 days. Notably, pre-treatment with ALFQ or ALFQsuv significantly decreased the percentage of HIV-1-infected MDM at day 5 post-infection when compared to the untreated or ALF55 pre-treated MDM ( Fig. 7a, b). When compared to the untreated MDM, pre-treatment with ALFQ decreased HIV-1 infectivity by sixfold and 23.4-fold, respectively for donors 118 and 170. However, in the presence of ALFQsuv, the decrease in infectivity was only 1.9-fold and 6.7-fold respectively, for the two donors. Furthermore, quantitative analyses of the HIV-1 DNA in the MDM at day 5 post-infection showed that pre-treatment with ALFQ resulted in lower levels of viral DNA when compared to MDM that were pre-treated with ALFQsuv or ALF55 (Fig. 7c). This suggests either that the capacity of ALFQ to modulate MDM permissiveness to HIV-1 infection was principally mediated by the large ALFQ liposomes per se, or by the greater number of QS-21 molecules embedded on the outer surfaces of the large ALFQ liposomes when compared to ALFQsuv that have a large fraction of QS-21 molecules embedded on the inner leaflets of the outer lamellae. The presence of QS-21 on the outer and inner lamellae of the vesicles is a topic of further research.
ALFQ up-regulates APOBEC3A gene expression and protein synthesis in adherent MDM. Antiviral cellular restriction factors modulate HIV-1 infection. To gain more insight into the possible reasons for the restricted permissiveness to HIV-1 infection in the adherent MDM from the ALFQ-treated cultures, the expression of several antiviral cellular restriction factors was compared in adherent MDM pre-treated with ALF55, ALFQ or without any adjuvant (Fig. 8). Based on our previous study where we observed a differential expression of HIV-1 restriction factors in MDM subsets using RNA-Seq 26 , we evaluated the gene expression of SAMHD1, IFI16, APOBEC family, TRIM5α and TRIM22 in lysates of the adherent MDM by qRT-PCR. The data are plotted as fold change in gene expression in adherent MDM pre-treated with ALF55 or ALFQ in relation to adherent MDM without any adjuvant treatment. Pre-treatment with ALF55 did not induce any increase in gene expression of the restriction factors tested. In contrast, ALFQ significantly upregulated the gene expression of SAMHD1, IFI16, APOBEC3 families, TRIM5α, and TRIM22 to varying degrees. The most noticeable effect of ALFQ was observed in approximately 30-to 1100-fold higher expression of APOBEC3A (Fig. 8).
In order to determine if the increased gene expression in adherent MDM from the ALFQ-treated MDM cultures was also accompanied by a concomitant increase in protein synthesis, lysates of adherent MDM from ALF55-treated, ALFQ-treated, and adjuvant-untreated cultures were subjected to SDS-PAGE followed by Western www.nature.com/scientificreports/ blotting with antibodies against SAMHD1, IFI16, APOBEC3A, and GAPDH. APOBEC3A protein was only detected in the lysates of the ALFQ-treated cultures. The APOBEC3A band was detected in all 4 donors. However, the APOBEC3A band was more prominent in the lysates from donors #170 and #118 when compared to donors #205 and #008. We believe that the observed differences could be due to the use of anti-APOBEC3A polyclonal antibodies from two different sources. The use of a different source for anti-APOBEC3A antibody was due to supply chain difficulties. IFI16 protein was detected under all conditions with donor #118 although the level of protein expression was higher in the lysates of the ALFQ-treated cultures. Interestingly, in donor #170, IFI16 protein was detected primarily in ALFQ-and not in ALF55-treated cultures. The difference in IFI16 levels between donors #118 and #170 suggests that ALF55 may have a differential effect on IFI16 expression in different donors. IFI16 protein was not detected with donors #205 and #008. SAMHD1 protein was detected under all conditions in all the donors and further analyses with an antibody against phosphorylated-Thr592-SAMHD1 (p-SAMHD1) revealed phosphorylation on Thr592 residue in all four donors under all conditions. ( Fig. 9a and S1-S5). Quantitation of the Western blot bands by densitometry analysis (Fig. 9b), and normalization of each band against GAPDH, confirmed the presence of APOBEC3A levels exclusively in ALFQ-treated cultures and an increase in IFI16 levels upon treatment with ALFQ. Quantitation of SAMHD1 bands showed a differential effect   16 . To further evaluate the immunomodulatory role of APOBEC3A in restriction of HIV-1 infection, MDM were untreated or treated with control siRNA or APOBEC3A siRNA (A3A siRNA) 24 h prior to ALFQ treatment ( Fig. 10 and S6). The MDM cultures were subsequently infected with HIV-1. Silencing was done www.nature.com/scientificreports/ using a mix of two human APOBEC3A siRNAs. There was no significant difference in APOBEC3A gene expression between ALFQ treated and control siRNA treated MDM. However, treatment with A3A siRNA resulted in a significant decrease in APOBEC3A gene expression (Fig. 10a), a concomitant decrease in APOBEC3A protein synthesis was also observed by Western blotting (Fig. 10b and S6) and densitometry (Fig. 10c) with an increase in HIV-1 infectivity (Fig. 10d, e). HIV-1 infection of MDM treated with A3AsiRNA resulted in a significant loss in the restriction of HIV-1 infectivity in all 3 donors tested (Fig. 10d), resulting in a similar percentage of p24 positive MDMs as untreated MDM, whereas control siRNA-treated MDM showed similar results (restriction of infection) as ALFQ-treated MDM (Fig. 10d, e). The data from Figs. 9 and 10, demonstrate that the induction of APOBEC3A by ALFQ restricts HIV-1 infection in ALFQ-treated MDM.

Discussion
Our data highlight interesting and novel differences on the effects of Army Liposome Formulations on MDM, particularly the ALFQ formulation which contains QS-21. Exposure of MDM to ALFQ which contains two immunostimulants synthetic monophosphoryl lipid A (3D-PHAD®) and QS-21, modulated permissiveness of the MDM to HIV-1 infection. This was concomitant with increased expression of various innate anti-HIV restriction genes and proteins particularly APOBEC3A, inflammasome activation, increased production of proinflammatory cytokines and chemokines, and increased surface expression of MHC-II and the co-stimulatory molecules CD80 and CD86. M1 macrophages are associated with increased CD86 expression 14,28 , and polarization towards this phenotype is largely dependent on environmental cues 29,30 . In contrast, exposure of MDM to ALF55 which contains only one immunostimulant, 3D-PHAD®, induced lower levels of proinflammatory mediators, had minimal or no effect on anti-HIV restriction genes and did not make MDM less permissive to HIV-1 infection. Our previous studies using RNA-seq 26 had demonstrated reduced expression of innate restriction factors in adherent MDM, which contributed to increased permissiveness to HIV-1. In the present study we evaluated if ALFQ would positively affect the expression of the innate restriction factors, which in turn would modulate permissiveness to HIV-1 infection. Our results demonstrate that exposure of MDM to ALFQ, but not ALF55 induced increased gene expression of APOBEC3A and IFI16 as determined by qRT-PCR with a concomitant increase in protein levels. Furthermore, exposure of MDM to ALFQ also elicited the secretion of proinflammatory effector mediators creating an inflammatory environment, thus polarizing the MDM into an inflammatory phenotype. The cumulative effects of these induced responses may be pivotal in rendering MDM significantly less permissive to HIV-1 infection. ALFQ and ALF55 are unilamellar liposomes with identical phospholipid and cholesterol concentrations. However, ALF55 comprises SUVs (0.1 µ), while ALFQ is polydisperse and contains both small and large liposomes. The size of the small liposomes ranges from 50 to 100 nm as opposed to the large liposomes which  . The outer surface of the ALFQ particle has a "sugar lawn" consisting of ten different sugars (two on MPLA, and eight on QS-21) which could serve as ligands for binding to cell receptors. In this regard, the larger ALFQ liposomes may have increased visibility compared to small ALFQ, which could influence the outcome of the MDM responses. Our study showed that small ALFQ liposomes (ALFQsuv) had a significantly reduced capacity to modulate permissiveness of MDM to HIV-1 infection when compared to ALFQ which indicates that either the liposome size, or the placement of QS-21 on the outer leaflet of the outer lamella (ALFQ), or on both the outer and inner leaflets of the outer lamella (ALFsuv), greatly influenced the effects of ALFQ on MDM. However, ALF55 which are also SUVs did not exhibit the same effects on MDM as ALFQsuv in modulating MDM permissiveness to HIV-1 infection. Therefore, size may not be the only critical factor. The major chemical difference between the 3 liposomes is the presence of QS-21 in ALFQ and ALFQsuv but not in ALF55. Free QS-21 is highly toxic, however when QS-21 is present in liposomes, it irreversibly binds to cholesterol in the liposomes and loses its toxic effect but maintains its adjuvanticity 33 . The precise role of QS-21 in ALFQ is poorly characterized. Free QS-21 in combination with MPLA (presumably micellar) has been shown to activate the ACS-NLRP3 inflammasome and cause the subsequent release of caspase-1 dependent proinflammatory cytokines IL-1β/IL-18 in in vitro studies with mouse APCs 23 . We observed similar results when both QS-21 and MPLA were formulated in liposomes as ALFQ.
IL-1β and IL-18 are the signature cytokines of inflammasome activation. Although both cytokines were induced by ALF55 and ALFQ, the levels were significantly higher in response to ALFQ thus indicating that the QS21-containing adjuvants are superior inflammasome activators. Although inflammasome activation is associated with pyroptotic cell death, neither ALF55 nor ALFQ elicited increased LDH production in the MDM cultures. Indeed, the amount of LDH production in the MDM cultures exposed to ALF55 or ALFQ was similar to the levels detected in the control cultures that were not exposed to any of the adjuvants. Our observation that MDM may secrete IL-1β or IL-18 and not undergo pyroptosis is in line with recent reports that highlight the capability of myeloid cells to secrete IL-1β in a GSDMD-independent manner 34,35 , and other studies in which viable monocytes, macrophages, and neutrophils have been shown to secrete IL-1β and not undergo pyroptosis 34,36 .
The difference in permissiveness to HIV-1 infection between MDM that were exposed to ALFQ compared to those that were exposed to ALF55 was independent of cellular expression of Siglec-1 or the capacity to capture HIV-1 virions. However, quantitative analyses of the HIV-1 DNA from ALFQ pre-treated MDM showed significantly lower levels of viral DNA when compared to MDM that were pre-treated with ALF55 (Fig. 7c), www.nature.com/scientificreports/ thereby indicating that although ALFQ did not inhibit the capture and entry of HIV-1 (Fig. 6c), ALFQ inhibited the replication phase. Earlier reports have shown that monocytes are largely refractory to HIV-1 infection 37 , due to the presence of innate anti-HIV-1 genes and proteins including APOBEC3 16 and SAMHD1 12 proteins. The differentiation process into MDM is associated with down-regulation of some of these innate anti-HIV-1 genes rendering the MDM highly permissive to HIV-1 infection 16,38 . In addition, although cells may express SAMHD1, the anti-HIV-1 activity of SAMHD1 is lost when it is phosphorylated at position T592. This results in a loss in the ability of the cell to restrict HIV-1 infection 39,40 . Our results show that there was a donor-to-donor variation in the levels of SAMHD1 as well as in p-SAMHD1 in ALFQ treated MDM. The level of p-SAMHD1 was decreased in some donors, while in others it was slightly increased compared to untreated MDMs. Thus, it is possible that more than one restriction factor is involved in modulating HIV-1 infection.
Our data further indicates a differential effect on innate anti-HIV-1 genes that are associated with modulation of HIV-1 infection in MDM. Innate sensors including IFI16 are necessary for cells such as macrophages to sense the intracellular presence of HIV-1 and subsequently trigger innate cellular processes necessary for control of the viral infection 3 . The innate cellular HIV-1 restriction factor APOBEC3A is a member of the APOBEC3 family of DNA cytidine deaminases, which restrict HIV-1 replication by hypermutating viral cDNA and inhibiting reverse transcription and integration 41 . APOBEC3A has also been recently highlighted in the maintenance of latency in HIV-1 infected CD4 + T cells 15 .
It has been proposed that differentiation of monocytes into MDM reduces expression of innate restriction factors, rendering the otherwise relatively refractory monocytes into an HIV-1 permissive MDM phenotype 16 . www.nature.com/scientificreports/ In this study, the dramatic increase in APOBEC3A gene expression and the corresponding protein synthesis of APOBEC3A in MDM that were treated with ALFQ, coincident with the ability of those MDM to restrict the HIV-1 infectivity suggest that ALFQ renders an HIV-1-permissive target into a more restrictive phenotype. Experiments with A3A siRNA and Western blotting demonstrated a significant decrease in APOBEC3A gene and protein expression with a concomitant loss of restriction of HIV-1 infectivity. Thus, our study demonstrates that the induction of APOBEC3A by ALFQ restricts HIV-1 infection in ALFQ-treated MDM. Interestingly, donor #118, showed enhanced gene and protein expression of APOBEC3A and IFI16, and decreased levels of pSAMHD1. These outcomes occur in the proinflammatory cytokine environment induced with ALFQ treatment. Although our study did not address the relationship between the inflammatory environment and synthesis of innate anti-HIV-1 proteins, a recent study 14 showed that the presence of APOBEC3A was necessary for induction of proinflammatory cytokines and upregulation of CD86 expression in macrophages. Our study also showed increased levels of proinflammatory cytokines, MHC-II and CD86 expression coincident with APOBEC3A synthesis in ALFQ-exposed MDM. Therefore it may be plausible to postulate that ALFQ induces the synthesis of APOBEC3A and possibly other anti-viral proteins which promote the induction of a proinflammatory environment that polarize MDM into an M1 phenotype, rendering them less permissive to HIV-1 infection 42,43 . This indicates a possibility that the outcome of the HIV-1/target cell relationship could be tilted in favor of the target cell. ALFQ as a potent adjuvant has shown great promise in several preclinical studies with parasitic (malaria), viral (HIV and SARS-CoV-2), and bacterial (Campylobacter) vaccine formulations 33,44 . Three ongoing phase 1 clinical trials which utilize ALFQ as the adjuvant have been recently completed. Two of these were with circumsporozoite malaria proteins (ClinicalTrials.gov Identifier: NCT04296279 and NCT04268420) and the third one was with SARS-CoV-2 spike-ferritin nanoparticle (ClinicalTrials.gov Identifier: NCT04784767). ALFQ will also be used in an ongoing recent phase 1 trial in Bangkok, Thailand with HIV envelope protein (NCT04658667). In addition to the use of ALFQ in HIV-1 vaccine studies, our data also highlights the potential use of ALFQ as a therapeutic agent since it targets the innate immune system. Virus purification. HIV-1 subtype B primary virus, US-1, was grown in PBMCs from the virus stock provided by Dr. Victoria Polonis (USMHRP). The primary virus was purified as previously described 45 . Infectivity and p24 concentrations were determined before and after purification to ensure that infectivity was not lost during the purification procedure. (Table 1) were prepared as described previously 31  www.nature.com/scientificreports/ was diluted 10 times with Sorensen's phosphate-buffered saline (SPBS). The microfluidizer was prepared by injecting 6 ml of 70% isopropanol thrice and 6 ml of SPBS five times at 30,000 psi. This was followed by the passage of 6 ml of the above diluted ALFQ sample through the microfluidizer at 30,000 psi. The process was repeated six times. A gel loading pipette tip was used to remove residual liquid from the intake and output ports of the microfluidizer. Finally, the microfluidized ALFQ was filter-sterilized through a 0.22 µm filter and stored at 4℃ until use. The concentration of liposomes was determined by the cholesterol assay 48 .

Adjuvant formulations. Adjuvant formulations
Cryo-electron microscopy. Cryo-electron microscopy was performed under a contract with University of Virginia (UVA). Briefly liposome samples (~ 3.5 μl) were applied to a glow-discharged, perforated carboncoated grid (2/1-3C C-Flat; Protochips, Raleigh, NC), manually blotted to near dryness with filter paper, and rapidly plunged into liquid ethane to freeze the sample. The grids were stored in liquid nitrogen, then transferred to a Gatan 626 cryo-specimen holder (Gatan, Warrrendale, PA) and maintained at ~ 180 °C. Low-dose images were collected on a Tecnai F20 Twin transmission electron microscope (FEI {now ThermoFisher Scientific}, Hillsboro, OR) operating at 120 kV 49 . The digital micrographs were recorded on a Gatan US4000 CCD or a Teitz XF416 camera.  45 . MDM were used on day 5 post-culture.

Generation of adjuvant-derived supernatants.
On day 4 post-culture, ALF55 or ALFQ were added to cultures of MDM (3-4 × 10 5 cells/well) and the cultures were incubated at 37 °C/5%CO 2 for 24 h, after which the cells and supernatants were harvested. MDM cultured in media not exposed to any adjuvant served as controls.
In each appropriate well, the final concentration of 3D-PHAD® was 0.5 μg/ml and QS-21 was 0.25 μg/ml. Cells and supernatants were harvested. The viability of the cells was determined by trypan blue exclusion as well as Annexin V/7AAD staining. Supernatants were clarified twice at 13,200 rcf for 20 min at 4 °C and transferred into clean tubes. Samples were stored at − 20 °C for subsequent analyses and assays.
Detection of cell surface molecules. Cells were washed in cold FACS buffer (PBS-containing 0.5% bovine serum albumin (BSA)) and blocked in FACS buffer containing 10% normal goat serum. Following incubation for 20 min at 4 °C with a cocktail containing 5-10 μg of specific mAb, the cells were washed in cold FACS buffer, fixed in PBS-containing 2% paraformaldehyde, and subsequently acquired on a FACSCalibur (BD Biosciences, San Jose, CA). Data analyses were performed on gated CD14 + cells, using FlowJo 8.8.6 software (TreeStar Inc., Ashland, OR).

Quantitation of cell surface receptors. The number of receptors/cell was determined using Quantum
Simply Cellular Beads (Bangs Laboratories Inc., IN, USA) according to the manufacturer's instructions, and as previously described 25 . Briefly, MDM (0.5 × 10 6 /tube) were preincubated with 10% normal goat sera, followed by the addition of a mAb cocktail for 30 min on ice. Cells were then fixed with 2% formalin. Beads from the kit were stained with anti-human HLA-A, B, C-FITC, anti-human HLA-DR, DP, DQ-FITC, anti-human CD80-PE, anti-human CD86-APC, anti-human CD16-FITC, anti-human CD32-PE, anti-human CD169-APC. Individual standard curves were established using the stained beads. Samples were acquired on a FACSCalibur and analyzed with FlowJo 8.8.6 software. Data were placed into QuickCal version 2.3 software (Bangs Laboratories), and the number of receptors/cell was extrapolated from the standard curves generated with the stained beads.
HIV-1 infection of MDM. MDM (3-4 × 10 5 cells/well) were pre-treated with ALF55, ALFQ or ALFQsuv for 24 h at 37 °C/5%CO 2 . MDM not exposed to any adjuvant served as controls. Cells were gently rinsed, and 300 μl infection media containing 3 ng HIV-1 p24 (US-1) was added to each well. www.nature.com/scientificreports/ and then incubated in 1 × PBS with goat anti-rabbit HRP secondary antibody (System Biosciences, CA) at RT for 1 h. The membranes were then washed for 5 min in PBS-T (twice) and with PBS (once) prior to developing using Clarity (Bio-Rad, CA) and/or Clarity Max ECL Western Blotting Substrates (Bio-Rad, CA). Chemiluminescence was imaged using a ChemiDoc Touch Imaging System (BIO-RAD) and images processed using Image Lab (Software version 6.0.1.0; Bio-Rad, CA). Densitometry analysis was used to measure the signal from each protein band and analyzed using ImageJ software (National Institutes of Health; Bethesda, MD). The signal density counts were obtained for each band, subtracting background signal, using the same capture area for all bands. Each protein band signal was normalized to its corresponding GAPDH band to account for potential variations between samples.
Statistical analysis. Statistical methods used are detailed in the figure legends. Cytokine responses were compared by two-tailed Student t-test. Cellular anti-viral genes were compared by two-tailed Mann-Whitney test. Group means of experimental data for number of cell receptors, percentage of HIV-1 infected MDM, apoptotic MDM and HIV-1 RNA and DNA copies were compared by One-way ANOVA. Data are displayed as scatter dot plots or bar graphs with mean ± standard deviation. No data point was excluded from the analyses. p-values less than or equal to 0.05 were considered statistically significant. Analyses were carried out with GraphPad Prism 7 (version 5.0c; GraphPad Software (La Jolla, CA, USA).
Ethics statement. RV229B (WRAIR Protocol #1386): This protocol "Apheresis of blood components from healthy volunteers for in vitro research" and all related documents were approved by the following independent Institutional Review Boards (IRBs): Division of Human Subject Protection, Walter Reed Army Institute of Research; Ethical Review Committee for Research in Human Subjects. All volunteers provided written informed consent following discussion and counseling by the clinical study team prior to enrollment and before the blood draw.