Nano Today 36 (2021) 101056 Contents lists available at ScienceDirect Nano Today journal homepage: www.elsevier.com/locate/nanotoday Antibody cooperative adsorption onto AuNPs and its exploitation to force natural killer cells to kill HIV-infected T cells ]] ]]]]]] ]] Antonio Astorga-Gamaza a, Michele Vitali a, Mireya L. Borrajo a, Rosa Suárez-López b, Carlos Jaime b, Neus Bastus c,d, Carla Serra-Peinado a, Laura Luque-Ballesteros a, Oscar Blanch-Lombarte e,f, Julia G. Prado e,f, Juan Lorente g, Felix Pumarola g, Marc Pellicer g, ⁎ ⁎⁎ Vicenç Falcó a, Meritxell Genescà a, Víctor Puntes a,c,d, , Maria J. Buzon a, a Infectious Disease Department, Hospital Universitario Vall d'Hebron, Institut de Recerca (VHIR), Universitat Autònoma de Barcelona, Barcelona, Spain b Departament de Química, Universitat Autònoma de Barcelona, Bellaterra, 08193 Barcelona, Spain c Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC and BIST, Campus UAB, Bellaterra, 08193 Barcelona, Spain d Institució Catalana de Recerca i Estudis Avançats (ICREA), 08010 Barcelona, Spain e IrsiCaixa AIDS Research Institute, Badalona, Spain f Germans Trias i Pujol Research Institute (IGTP), Universitat Autònoma de Barcelona, Badalona, Spain g Otorhinolaryngology Department, Vall d'Hebron University Hospital, Passeig Vall d'Hebron 119-129, 08035 Barcelona, Spain a r t i cl e i nfo a bstr ac t Article history: HIV represents a persistent infection which negatively alters the immune system. New tools to reinvigorate Received 17 July 2020 different immune cell populations to impact HIV are needed. Herein, a novel nanotool for the specific Received in revised form 26 October 2020 enhancement of the natural killer (NK) immune response towards HIV-infected T-cells has been developed. Accepted 3 December 2020 Bispecific Au nanoparticles (BiAb-AuNPs), dually conjugated with IgG anti-HIVgp120 and IgG anti-human Available online 20 December 2020 CD16 antibodies, were generated by a new controlled, linker-free and cooperative conjugation method promoting the ordered distribution and segregation of antibodies in domains. The cooperatively-adsorbed Keywords: HIV antibodies fully retained the capabilities to recognize their cognate antigen and were able to significantly Bispecific nanoparticles enhance cell-to-cell contact between HIV-expressing cells and NK cells. As a consequence, the BiAb-AuNPs NK cells triggered a potent cytotoxic response against HIV-infected cells in blood and human tonsil explants. Polarization Remarkably, the BiAb-AuNPs were able to significantly reduce latent HIV infection after viral reactivation in a primary cell model of HIV latency. This novel molecularly-targeted strategy using a bispecific nanotool to enhance the immune system represents a new approximation with potential applications beyond HIV. © 2020 The Author(s). Published by Elsevier Ltd. CC_BY_NC_ND_4.0 Introduction several sophisticated biological mechanisms, including antibody- dependent cellular cytotoxicity (ADCC), to accomplish their mission The augmentation of the effector immune responses towards [2–4]. Unfortunately, in the context of a persistent viral infection, specific targets, such as virally-infected or cancerous cells, is posi­ like the one established by the human immunodeficiency virus tioning itself as a new promising therapeutic intervention [1]. In­ (HIV), the virus integrates into the genome of host cells and is able to deed, an efficient immunological synapse between immune effectors escape from the immune response, therefore perpetuating the in­ and aberrant cells is a key factor to promote an effective immune- fection [5]. In turn, chronic infection leads to a persistent immune mediated killing of such undesired cells. In this context, cell-to-cell activation that promotes the appearance of dysfunctional NK cells contact is required for natural killer (NK) cells, lymphocytes of the with diminished capacity to kill infected cells [6], further impairing innate immune system that eliminate abnormal cells through immune defences. Importantly, this functionality is not completely restored with the current antiretroviral therapy, and discontinuation of the treatment leads to a rapid viral rebound [7]. Thus, new tar­ ⁎ Corresponding author at: Catalan Institute of Nanoscience and Nanotechnology geted interventions are needed to help the immune system of HIV- (ICN2), CSIC and BIST, Campus UAB, Bellaterra, 08193 Barcelona, Spain infected patients to control the virus. In this regard, novel ther­ ⁎⁎ Corresponding author. apeutic strategies focused on the re-direction, potentiation and/or E-mail addresses: victor.puntes@icn2.cat (V. Puntes), promotion of an efficient immunological synapse between NK and mariajose.buzon@vhir.org (M.J. Buzon). https://doi.org/10.1016/j.nantod.2020.101056 1748-0132/© 2020 The Author(s). Published by Elsevier Ltd. CC_BY_NC_ND_4.0 A. Astorga-Gamaza, M. Vitali, M.L. Borrajo et al. Nano Today 36 (2021) 101056 Fig. 1. Characterization of the bispecific gold nanoparticles (BiAb-AuNPs). (a) Representative TEM image of the ~40 nm AuNPs. (b) A32-AuNPs saturation curve. Zeta potential (mV) versus the antibody concentration is plotted. The red arrow indicates the saturating concentration for the A32 antibody. (c) 3G8-AuNPs saturation curve. The blue arrow indicates the saturating concentration for the 3G8 antibody. (d) Representative UV-Vis spectra of the BiAb-AuNPs compared to naked AuNPs. (e) Z-potential measurements of BiAb-AuNPs, naked AuNPs, and the single antibodies A32 and 3G8 (n = 3, mean with SD). (f) Representative Dynamic light scattering (DLS) characterization of the BiAb-AuNPs compared to naked nanoparticles. (g) Representative UV-Vis spectra of the BiAb-AuNPs at 48 h and 1 month after antibody conjugation. infected T-cells are highly desirable to efficiently combat such per­ biological identity to the NP [15]. In these conditions, similar pro­ sistent viral infections. teins pack better than those which are different, favoring thus the Nano-based approaches with a dual capacity of recognition are formation of homogeneous coating domains from a mix of coating emerging as promising bispecific formats to engage cell-to-cell molecules [16]. This evolution towards a more stable state, an ex­ contacts and potentiate antiviral and antitumor immunity [8]. Such pression of the Vroman effect on the surface of NPs [17], is what nanotools based on nanoparticles (NPs) displaying mixtures of an­ allows to explore enthalpy-driven, rather than entropy-driven states, tibodies showed significant advantages over the more conventional which implies macromolecular ordering and domain segregation bispecific antibodies, such as the possibility to regulate biological [18]. Thus, by controlling the principal parameters affecting the NP- processes in a size-dependent manner, their capacity to penetrate protein interactions, such as relative concentrations and affinity, into some tissues and cells and their enhanced potency mediated by incubation time, pH, temperature, media ionic composition and their multivalent properties [9–11]. solubility, it is possible to find the optimal conditions to promote the The interaction of proteins, including antibodies, onto inorganic cooperative adsorption of proteins onto the NP [14] and consequently surfaces, such as gold, has been a subject of intense studies in the the formation of highly polarized NP conjugates [12,19,20]. field of biochemistry and medicine since implants were developed. Here, with the purpose to enhance the NK immune response These studies on the adsorption of proteins at surfaces, a phenom­ towards HIV-expressing cells we have constructed bispecific nano­ enon ubiquitous in nature, have shown two models: Random particles following a novel cooperative adsorption conjugation pro­ Sequential (disordered) vs Clustering (cooperative) adsorption [12]. tocol of a mixture of antibodies, inducing their domain segregation The first indications on the mode in which IgGs adsorb onto AuNPs is at the NP surface and resulting in a certain degree of NP functionality the kinetics of the adsorption process that already points towards a polarization. This novel molecularly-targeted approach uses the cooperative mode: proteins at the NPs surface evolve from forming a antibodies A32 (IgG anti-HIVgp120), which recognizes the protein transient corona to become permanent with time, known as soft and gp120 expressed on the membrane of HIV-infected cells, and the hard NP-Protein Corona (PC) respectively [13]. As it has been well 3G8 (IgG anti-human CD16, clone 3G8) that recognizes the CD16 described, when NPs are dispersed in protein solutions, these rapidly receptor expressed on NK cells, conjugated to a 40 nm AuNP. This form a soft protein corona around the NP, formed of loosely bound novel nanotool appears very promising for the targeting and im­ proteins, in equilibrium with free proteins in solution. With time, mune-mediated elimination of HIV-infected cells. this PC evolves, due to rearrangements and crowding effects, to­ wards a permanent corona since protein-dense clusters are more Results stable at the NP surface than single proteins. Indeed, cooperative adsorption implies that a protein adsorbed to a surface binds more Generation of functionalized bispecific nanoparticles strongly to it if it is surrounded by similar proteins thanks to the possibility of build compact arrangements with identical Our goal was to develop multivalent bispecific NPs after a co­ building blocks, increasing coating stability (cooperative effect) as operative adsorption of the antibodies at their surface. For that, we the protein domain grows. This evolves towards the formation of a carefully studied the conjugation medium and the conditions to complete hard corona of proteins permanently attached to the NP achieve a cooperative adsorption protocol and we validated the surface that remains associated to the NP even after NP precipitation approach. First, we generated monoconjugates with the antibodies and resuspension into protein-free media [13,14], thus conferring A32 and 3G8 separately and studied the impact of the conjugation 2 A. Astorga-Gamaza, M. Vitali, M.L. Borrajo et al. Nano Today 36 (2021) 101056 process on the capacity of these antibodies to still recognizing their of antibody coatings, UV–Vis, DLS and Z-potential peaks would be­ cognate antigens. In order to assess the affinity of the different an­ come broader than in the case of non-conjugated NPs, integrating tibodies for the AuNP surface, we performed Langmuir-type iso­ the dispersion of both, the NP core and the NP coating. thermal adsorption curves to identify the concentration needed to Using a theoretical approximation of 150 nm2 per antibody on obtain surface-saturated monoconjugates. AuNPs of 42 ± 6 nm dense antibody coatings [26], we estimated that 32 antibodies were (Fig. 1a) with a surface charge of −44 mV, colloidally stable in a displayed on the surface of our pBiAb-AuNPs (Supplementary Note 2.2 mM sodium citrate (SC) at 5 × 1010 NPs/ml concentration and 1). Moreover, thermogravimetric analysis (TGA) of the conjugates, with an UV-Vis Spectroscopy absorption peak at 532 nm, were ex­ considering a density of 1.35 and 19.30 g/ml for an antibody and the posed to the different antibodies dispersed in a 2.2–4 mM SC-Borate Au, respectively, led to similar results. In this case, the evaporation at selected conjugating medium at different concentrations. From the 450 °C of a 10 nm corona of antibodies in a ~40 nm diameter of solid adsorption curves, 5 µg/ml and 20 µg/ml were identified as the an­ AuNP accounted for about 10% of the total weight loss tibody saturation concentrations needed for A32 and 3G8, respec­ (Supplementary Fig. 2). tively (Fig. 1b and c). The specific cell-surface binding abilities of both monoconjugates were evaluated by flow cytometry. We found Polarization assessment after following the cooperative adsorption that A32-AuNPs retained the same binding capability than the free conjugation protocol antibody and recognized the entire population of HIV+ cells, in­ dicating that the conjugation process did not alter its targeting ca­ Molecular mixtures on a NP surface are difficult to characterize pacity (median 98.60% for the conjugate and 99.60% for free A32) [27], especially in their working environment. In the early 1990s, (Supplementary Fig. 1a). Similarly, 3G8-AuNPs were able to re­ studies suggested that two alkanethiols of different chain length cognize the CD16 molecules expressed on NK cells in a similar segregate on a gold surface, forming microscopic islands, despite the manner as the non-conjugated free antibody (median 58.60% for the strong bond between thiols groups and Au [16]. Since then, other conjugate vs. 51.30% for free 3G8) (Supplementary Fig. 1b), indicating mixtures of alkanethiols [19], organic molecules as MUA and PEG that antibody adsorption did not result in any decrease in its ability [28] and also peptides did show a tendency to segregate onto the NP to recognize the cognate antigen. surface, and this effect was discussed not without controversy Next, we generated the polarized bispecific nanoparticles (pBiAb- [19,29–31]. This tendency to grow domains was described to occur AuNPs) by mixing equal volumes of the AuNPs solution with the mix during the self-assembly process onto the NP surface, and not when of antibodies in solution for 48 h at 4 °C carefully controlling the the molecules were free in solution [32]. Simulations performed conjugating medium. By mildly destabilizing ~40 nm AuNPs in a mix mimicking our experimental conditions showed the tendency of of A32 and 3G8 antibodies, we observed how the system slowly cooperative adsorption of antibodies onto the NP surface when en­ evolved towards a new equilibrium state where NPs got fully coated ough time was allowed to the antibodies to organize themselves by the two antibodies (vide infra). In a close-up, A32 and 3G8 pre­ onto the AuNP surface. Similarly, once few clusters had nucleated sent differences in their fine structure and physicochemical prop­ onto the NP surface, these grew until the whole surface was densely erties, that enabled its segregation when adsorbed at the NP surface. covered (Fig. 2a–d). For instance, A32 has a reported isoelectric point (pI) of 8.8 [21] and Describing the organization of molecules at the NP surface is a surface charge of ≃−15 mV, meanwhile the 3G8 antibody has an challenging. Despite the literature is full with examples of functio­ estimated pI of 8.2 (according to its sequence) and a more negative nalized nanoparticles generally illustrated with inspiring schemes, measured surface charge. The aim was to work under conditions the actual distribution and conformation of molecules at the NP where antibodies and NPs were rather stable and therefore slowly surface is still unknown and often remain poorly characterized. We conjugated, allowing them to explore their minimum energy posi­ obtained the first evidence for the formation of antibody domains on tions which correspond with clustering and crystallization. top of the AuNPs from the antibody loading curves and geometrical The ratio of each antibody in the solution was chosen according considerations. The progressive coating of the antibodies on the NPs to the saturation curves and their different relative affinities for the can be observed in the UV–Vis, DLS and Z-potential measurements AuNP surface. In order to obtain an approximate 50:50 proportion of after purification as a function of incubation time. If the adsorbed each antibody at the NP surface, a ratio of 100:60 for 3G8 and A32 proteins were distributed purely at random, those NPs with just respectively was finally chosen. After conjugation, polarized bispe­ above the 50% of coverage should be resistant to aggregation during cific nanoparticles (pBiAb-AuNPs) presented a redshift in the UV–Vis the purification process (that forces NPs together). In a perfect spectra of 4 nm (Fig. 1d), consistent with a full coating of the NP random distribution, after loading more than 50%, the remaining surface with proteins [22]. The surface charge of the pBiAb-AuNPs grooves would not be large enough to efficiently fit another protein was also characterized (Fig. 1e), reporting a lower negative value (carrying an attached NP), nor the two AuNP surfaces could be in (−31.4 mV), corresponding to half way between the values of the two contact (see schema in Supplementary Fig. 3). In our experiments, homoconjugates (Supplementary Table 1) indicating the presence of NPs aggregated during purification up to surface coverages close to a 50% mixture of antibodies in the fully coated surface of AuNPs [23]. the 80%, well away from the theoretical 50% suggesting cooperative The DLS (Dynamic Light Scattering) hydrodynamic radius signal also growth of the antibody coating (Supplementary Fig. 4a and b). This increased upon conjugation (about 10 nm), correlating well with the time-dependent hardening of the protein corona was also confirmed UV-Vis spectroscopy measurements and attributed to the formation by the analysis of the evolution of the surface charges. A minimum of of a dense antibody monolayer on top of the AuNP (Fig. 1f). No 24 h of antibody exposure was needed for the formation of a stable further shifts of the UV–Vis post conjugation were observed after a antibody layer (Supplementary Fig. 4c). This time point for con­ month (Fig. 1g), indicating the high stability of the formed con­ jugation (24 h) was associated to an 80% of surface coverage jugates. Note that if antibodies were denaturalized as a consequence (Supplementary Fig. 4d). Note that in a Langmuir-like isothermal of the interaction with the AuNP surface, they would inevitably lead adsorption behavior, the observed shift of the characteristic plas­ to antibody aggregation and detectable NP cross-linking [24]. Im­ monic peak position between zero and saturation is directly pro­ portantly, the characteristic width of the measured peaks after portional to the degree of coating [28]. We used this information to conjugation remained similar to the measurements before con­ prepare partially and fully antibody-coated AuNPs, and gain further jugation, further confirming the formation of homogeneous anti­ insight into the growth regimen of the antibody layer under our body coronas [25]. Otherwise, if a dispersion in the conjugation protocol conditions by visualizing the prepared conjugates through process would occur leading to NPs with different degrees or types transmission electron microscopy (TEM) and scanning electron 3 A. Astorga-Gamaza, M. Vitali, M.L. Borrajo et al. Nano Today 36 (2021) 101056 a b c d e 60% COVERAGE 100% COVERAGE anti-BSA IgG anti-HSA IgG f g HSA BSA anti-BSA IgG anti-HSA IgG anti-HSA IgG anti-BSA IgG BSA HSA BSA HSA anti-HSA IgG anti-BSA IgG Fig. 2. Simulations and Transmission electron microscopy (TEM) and Scanning electron microscopy (SEM) analysis of conjugates following the cooperative adsorption protocol. Four snapshots of the resultant simulations are shown (a–d). From the left to the right, the water-antibodies bead interacting parameters are increased from 24 to 50. (a) The adsorption of these proteins starts to be observed but without any preferences. (b–d) All the antibodies are adsorbed onto AuNP for minimizing their interaction with water. Note that, the higher is the repulsion with water, the closer will be the antibodies. Polarized AuNPs are observed when water-antibodies repulsion is bigger than its own interaction. (e) In the upper panel TEM images of partially-coated (60% surface cover) and fully-coated Au conjugates can be seen. In the lower panel SEM images are shown. (f) AuNPs functionalized following cooperative adsorption with anti-BSA IgG and anti-HSA IgG, and exposed to smaller conjugates as labels (AuNPs of 25 nm functionalized with BSA and of 15 nm with HSA), are shown. It can be observed a hemispherical distribution of the labels, with three BSA-AuNPs distributed in one side of the bispecific AuNP and one HSA-AuNP in the opposite side. (g) The same can be observed in this panel but with an inverted proportion of the labels, further confirming polarity. microscopy (SEM). We cannot distinguish by TEM different anti­ shown. In the TEM images (Fig. 2e upper panel) we observe how in bodies on top of the NP, but we can observe if before saturation, the case of saturation, AuNPs do not touch each other, while for the distribution of proteins is random or if the coating grows co­ unsaturated samples the AuNP-AuNP direct contact is more evident. operatively onto the AuNP. In Fig. 2e, representative TEM and SEM In the SEM images (Fig. 2e lower panel), the charging effects of the images of partially (at the 60%) and fully antibody coated AuNPs are organic layer on the edge of the NPs can be clearly seen, and it can be 4 A. Astorga-Gamaza, M. Vitali, M.L. Borrajo et al. Nano Today 36 (2021) 101056 observed how the coating is complete in the case of antibody satu­ significantly increased the mean of cell doublets by over 7-fold rated samples while only one side of the partially coated AuNPs (p = 0.01) (Fig. 3a and b). Of note, the rBiAb-AuNPs with a random appears to wear it. (non-polarized) disposition of the A32 and 3G8 antibodies at their In addition, to directly analyze the distribution of antibodies per surface, were not efficient at performing cell-to-cell engagement NP following our conjugation protocol, we used doubly functiona­ (Fig. 3b). These results position our pBiAb-AuNPs as better candi­ lized AuNPs with anti-HSA and anti-BSA IgGs. We then used 25 nm dates to achieve such biological effects. In addition, we performed AuNPs functionalized with BSA and 15 nm AuNPs functionalized a different analysis and assessed the composition of the cell with HSA, as labels to directly observe the presence and distribution doublets in the absence and presence of pBiAb-AuNPs and rBiAb- of the two antibodies at the AuNP surface. This experiment was not AuNPs. First, we observed that our pBiAb-NPs were better nano­ performed with the A32 and 3G8 antibodies because their cognate conjugates at inducing the formation of cell-doublets (Fig. 3c antigens are embedded at the cell surface, and therefore, it will re­ upper panel), and within these cell doublets, we observed an in­ quire the use of recombinant gp120 and CD16 to functionalize the crease in the specific CD4-NK doublets and formed less homo­ label nanoparticles. Moreover, anti-human secondary antibodies doublets CD4-CD4 compared to the medium control (Fig. 3c lower could not be used because it does recognize not only the A32 but panel and 3d), further supporting their polarized nature. Besides, also the 3G8 antibody, providing a high unspecific background signal rBiAb-AuNPs generated significantly less cell doublets, with in­ and preventing the real discrimination between the A32 and 3G8 creased percentages of NK-NK doublets (Fig. 3d). Note that in the domains. After conjugation, the three functionalized AuNPs were absence of AuNP polarization, one would expect also an increase of incubated together and analyzed using TEM. The cooperative ad­ homodoublets (NK-NK or CD4-CD4) due to the random distribu­ sorption approach was confirmed by the observation of NPs with a tion of both antibodies on the surface of the NPs. This is consistent hemispherical distribution of the labels (Fig. 2f and g). After fully with results observed for rBiAb-AuNPs which increased NK-NK characterization, 50% of the sample presented non- or mono- con­ doublets (Fig. 3d). After the precise quantification by flow cyto­ jugated NPs (identified by the absence of labels or by only one type metry, we imaged the preparation by confocal microscopy, since of the secondary labels), 26% of the multiconjugated NPs showed the very low dielectric constant of AuNPs scatters light very effi­ some polarity, while 24% of them seemed to present stochastic or­ ciently [34] and therefore the ~40 nm AuNPs can be observed in ientation of the antibodies. All in all showing the tendency of pro­ the confocal microscope, especially if they form any kind of ag­ teins to grow forming domains in these conditions. A gallery of TEM gregates. Direct contact of NK and HIV+ cells zipped by the pBiAb- images representing different structures indicating the cooperative AuNPs, which appear in red, can be observed (Fig. 3e). adsorption of the antibodies is shown in Supplementary Information 2. pBiAb-AuNPs induce a potent cytotoxic response against primary CD4 T cells from HIV-infected patients and against ex vivo human infected BiAb-AuNPs promote specific cell-to-cell contact between CD4- tonsillar explants expressing HIV and NK cells Next, in order to evaluate the ADCC-like activity of NK cells to­ Once the pBiAb-AuNPs presenting A32 and 3G8 antibodies wards HIV-expressing cells in the presence of the pBiAb-AuNPs, we were characterized, we investigated their potential to bring closer performed the corresponding functional assays. CEM NKR CCR5 cells cytotoxic NK cells (characterized by dim expression of the marker coated with gp120 were co-cultured with NK cells during 4 h in the CD56 and high levels of the CD16 molecule) and primary presence of different nanoconjugates. Cells non-coated with gp120 CD4 T cells coated with the viral protein gp120. We performed were included as a control. Cell killing was measured by the loss of cell-to-cell contact experiments by culturing both cell types in the the marker eFluor670 by flow cytometry. A representative example presence of different bioconjugates. As control, irrelevant bispe­ and gating strategy are shown in Fig. 4a and Supplementary Fig. 5, cific AuNPs (Irre-AuNPs) coated with polyclonal IgG mouse and respectively. As a positive control we used plasma from a viremic polyclonal IgG rat antibodies, were similarly prepared. In addition, HIV+ patient containing a pool of antibodies targeting the gp120 viral we prepared 10:100 and 40:100 A32:3G8 ratio conjugates. The protein. Values were normalized to this positive control which is conjugation ratio 100:10 A32:3G8 was also prepared but it rapidly considered the maximum ADCC response that could be achieved for led to the irreversible aggregation of the AuNPs precluding its the individual samples. First, we tested NK cells from healthy donors further use. Besides, we also compared our pBiAb-AuNPs gener­ and we observed that pBiAb-AuNPs were able to promote a highly ated using the cooperative approach, with an already well de­ potent and specific ADCC-like response (Fig. 4b). The effect of the scribed conjugation protocol expected to yield randomly antibody- pBiAb-AuNPs was significantly better than the cytotoxicity reached coated NPs (rBiAb-AuNPs) [33]. Results for the employed con­ by the addition of IrreAuNPs or the free A32 antibody, which is de­ jugates are shown in Supplementary Table 1. No direct toxicity in scribed in the literature as one of the most potent non-neutralizing ex-vivo HIV-infected CD4+ T cells was associated with the presence antibodies [35]. Value for the pBiAb-AuNPs was close to 100% of the of any pBiAb-AuNPs or Irre-AuNPs in culture after 4 h incubation maximum NK immune effector response (median 97.04% vs. 44.09%, (mean percentage 95.83 ± 6.56 and 90.10 ± 7.78 for pBiAb-AuNPs 37.37% and 18.63% for A32, Irre-AuNPs and the combination of A32 and Irre-AuNPs, respectively). Longer incubation times (48 h) did and 3G8 unconjugated antibodies, respectively) (Fig. 4b). Note that not significantly affect cell viability either (91.2% and 86.67% cell the addition of both 3G8 and A32 antibodies to cell culture de­ viability, for pBiAb-AuNPs and Irre-AuNPs, respectively). We ob­ creased the ADCC levels compared to A32 alone. 3G8 is commonly served that 60% and 100% of the saturating concentration for each used to block Fc receptors (through the interaction with the variable antibody (A32 and 3G8, respectively) produced the highest fraction) at concentrations near 5 µg/ml, thus when presented in a number of NK-HIV+ CD4 T doublets, with a 2.5-fold increase non-conjugated form it may block part of the ADCC response pro­ compared with controls (median % doublets of 5.66 compared to moted by A32. Next, we assessed the capacity of the pBiAb-AuNPs to 2.23%, 3.30% and 4.65% for control medium, 10:100 and 40:100 restore the functional activity of NK cells from HIV+ patients, since it ratios, respectively) (Fig. 3a and b). Thus, 60% and 100% of the is well known that HIV infection impairs their effector immune re­ saturating concentration for A32 and 3G8 were selected as the sponses [6]. These patients presented a median CD4 T cell count in optimized concentrations for the generation of the pBiAb-AuNPs. blood of 590 cells/µl and a median time under suppressive anti­ Importantly, a dose-dependent response was observed, and the retroviral treatment of 69 months (plasma viral load < 50 copies/ml). addition of the pBiAb-AuNPs coated with 10 µg/ml of antibodies Using the A32 antibody we observed, as expected, that NK cells from 5 A. Astorga-Gamaza, M. Vitali, M.L. Borrajo et al. Nano Today 36 (2021) 101056 Fig. 3. The pBiAb-AuNPs induce cell-to-cell contact between HIV-expressing cells and NK cells. (a) Representative flow cytometry plots showing cell doublets corresponding to primary CD4+ T cells coated with gp120 (CD3+) and primary NK cells (CD56+) after 20 min co-culture in the presence of the different nanoconjugates. Double positive cells for the markers CD3 and CD56 denote pairs of cells formed by CD4-HIV+ and NK cells. Controls corresponding to medium alone and non-conjugated (naked) nanoparticles are shown. Different ratios for the generation of BiAb-AuNPs and different total antibody concentrations on the nanoparticles (2.5 µg/ml for all conjugates and an additional dose of 10 µg/ml for 60:100 ratio) are shown. (b) Summary graph for the induction of cell-to-cell contacts in the presence of different nanoconjugates. Mean with SEM for the experimental replicates are shown. Statistical analysis consisted on a paired t-test. *p < 0.05. (c) Representative example of the gating strategy to identify cell doublets using the FSC-H and FSH- A signals, and the subsequent analysis of the frequency of different types of cell pairs using the cell markers CD3 and CD56. (d) Distribution of the cell doublets. Cell doublets were identified by flow cytometry using the FSH-H and FSH-A plot. The total cell doublets were distributed in T-T cells (CD3+CD3+), NK-NK (CD56+CD56+) or T-NK (CD3+CD56+) doublets and the percentages are plotted. Data shows mean with SEM of n = 6 independent experiments. Statistical comparisons were performed using Wilcoxon matched-pairs signed rank test. (e) Representative confocal micrographs of HIV-expressing cells (CEM.NKR CCR5+ coated with HIV gp120 recombinant protein and stained with PKH67 in green, and DAPI nuclei staining in blue) and primary NK cells (DAPI nuclei staining). Both cell types were co-cultured at 1:1 ratio for 20 min in the presence of the pBiAb-AuNPs. pBiAb-AuNPs are visualized in red. Scale bar represents 5 µm. HIV-infected patients had a diminished capacity to kill by ADCC in cell immune response, since we found that non-polarized rBiAb- comparison to healthy donors (Fig. 4c). Importantly, after the addi­ AuNPs could not strongly potentiate the NK cytotoxic response tion of the polarized pBiAb-AuNPs, the functionality of NK cells was (Fig. 4d). significantly restored (Fig. 4d). Of note, polarization of the antibodies As the triggering through the CD16 receptor can be a potent ac­ at the AuNPs surface was a requisite for the enhancement of the NK tivating signal, we addressed the possibility of direct NK activation 6 A. Astorga-Gamaza, M. Vitali, M.L. Borrajo et al. Nano Today 36 (2021) 101056 Fig. 4. NK-mediated cytotoxic response against HIV-expressing cells promoted by the pBiAb-AuNPs. (a) Representative flow cytometry plots for NK-mediated cytotoxicity assays. CEM-NKR.CCR5+ cells were double-stained with PKH67 and eF670, and coated with the HIVBal gp120 recombinant protein. Cells were co-cultured with primary NK for 4 h at 1:10 target/effector ratio and in the presence of different nanoconjugates at the following antibody doses: 10 µg/ml for pIrre-AuNPs, 10 µg/ml for pBiAb-AuNPs and 5 µg/ml for each free antibody. Loss of the eF670 marker was used to determine the percentage of dead cells in an eF670 versus PKH67 plot. (b) Summary graph (n = 6) of the ADCC activity mediated by NK cells from healthy donors in the presence of different nanoconjugates at the following antibody doses: 10 µg/ml for irrelevant AuNPs, 10 µg/ml for pBiAb-AuNPs and 5 µg/ml for each free antibody. Represented data is normalized to the positive control condition (plasma from a viremic HIV+ patient). Statistical comparisons were performed using Wilcoxon matched-pairs signed rank test. Median with range (min-max) are shown. *p < 0.05. (c) Graph comparing the ADCC NK-mediated response between healthy donors and treated HIV-infected patients in the presence of the anti-gp120 antibody A32. Statistical comparisons were performed using Mann-Whitney test. Median with range (min-max) are shown. *p < 0.05; **p < 0.01. (d) Summary graph (n = 6) of the ADCC activity mediated by NK cells from treated HIV-infected patients in the presence of different nanoconjugates at the following antibody doses: 10 µg/ml for irrelevant AuNPs, 10 µg/ml for pBiAb-AuNPs and non-polarized rBiAb-AuNPs, and 5 µg/ml for each free antibody. Represented data is normalized to the positive control condition (plasma from a viremic HIV+ patient). Statistical comparisons were performed using Wilcoxon matched-pairs signed rank test. 7 A. Astorga-Gamaza, M. Vitali, M.L. Borrajo et al. Nano Today 36 (2021) 101056 Median with range (min-max) are shown. *p < 0.05. (e) Representative flow cytometry plots showing CD16-expressing cells in digested cells after 24 h of tonsil histoculture in the presence of the pBiAb-AuNPs or Irre-AuNPs (5 µg/ml of antibody burden dose). (f) Summary graph showing CD16-expressing cells after different tissue culture conditions (basal condition, 24 h incubation with 5 µl of pBiAb-AuNPs at 5 µg/ml dose or 5 µl of Irre-AuNPs at 5 µg/ml dose per block) in n = 8 histocultures. Median and ranges (min-max) are shown. Statistical comparisons were performed using Wilcoxon matched-pairs one tailed signed rank test. *p < 0.05; **p < 0.01. (g) Representative flow cytometry plots showing reduction in HIV infection (by detection of the viral protein p24) after 6 days of infection of the tonsil histocultures, in the presence of the pBiAb-AuNPs or Irre-AuNPs (5 µl of the 5 µg/ml dose per tonsil block). (h) Graph summarizing the results of the HIV inhibition assays (measured at day 6 of HIV infection by detecting the p24 antigen expression by flow cytometry) in tonsil histocultures in the presence of the pBiAb-AuNPs or Irre-AuNPs (5 µl of the 5 µg/ml dose per tonsil block) (n = 8). Median with ranges are represented. Statistical comparisons were performed using Wilcoxon matched-pairs signed rank test. *p < 0.05. by our pBiAb-AuNPs. Human peripheral blood mononuclear cells up to 50% reduction in some cases (Fig. 4g and h). This is noteworthy, (PBMCs) were cultured with the pBiAb-AuNPs for 4.5 h and the since the majority of the persistent HIV infection in antiretroviral different activation and cytotoxicity markers were measured by flow treated individuals reside in secondary lymphoid tissue such as cytometry. We observed that after incubation with the pBiAb-AuNPs, lymph nodes and the gut-associated lymphoid tissue (GALT), being primary NK cells (CD3- CD56+) produced IFN-γ (Supplementary the main barrier to reach a cure [38]. Interestingly, NPs are being Fig. 6a). Accordingly, 3G8 free antibody and IrreAuNPs also increased extensively studied as drug carriers and targeting vehicles for their IFN-γ production in NK cells, showing that in general, targeting Fc ability to target some of these tissue compartments [39]. receptors could also stimulate NK cells to produce IFN-γ, as pre­ viously reported [36]. By contrast, the pBiAb-AuNPs not always in­ pBiAb-AuNPs strongly enhance the NK-mediated killing of latent HIV- duced the expression of CD107a (a degranulation marker) infected cells after viral reactivation (Supplementary Fig. 6b), nor significantly affected cell activation, measured by the expression of HLA-DR (Supplementary Fig. 6c). Reducing or eliminating the persistent reservoir is the main We also used a human tissue histoculture model [37] to assess challenge in HIV research field. Recently, therapeutic strategies the penetrability of the pBiAb-AuNPs in blocks of lymphoid tissue named “shock and kill” are being designed and tested to clear the and the NK immune effector function during ex-vivo HIV infection HIV reservoir. This approach tries to induce the expression of dor­ (Supplementary Fig. 7). Note that the only valuable animal model to mant HIV viruses with latency reversal agents (LRAs), relying on study the NK biology in tissues during HIV disease are primarily achieving also a potent immune response able to eliminate the non-human primates. Tissue blocks were infected with HIV, and HIV-infected cells. For assessing the suitability of the pBiAb-AuNPs after 15 min pBiAb-AuNPs were added to the culture. After 24 h in­ to be used in the context of a shock and kill strategy, we tested their cubation, we checked the penetrability of the pBiAb-AuNPs by de­ capacity to induce a cytotoxic response on viral reactivated HIV- tection of the CD16 marker in NK cells. 6 days later, the infection was infected cells with LRAs, first on the latently infected cell line ACH- measured by intracellular staining of the HIV p24 viral protein. We 2 and then on a primary cell model of HIV latency. When pBiAb- observed that pBiAb-AuNPs were able to penetrate these cultured AuNPs were added to ACH-2 cells and subjected to viral reactiva­ tonsil blocks, as demonstrated by the specific loss in the detection of tion, we consistently observed a highly potent response against the CD16 marker by flow cytometry in the presence of the pBiAb- HIV+ cells, inducing a median of 80.0% of cell death in 4 h (Fig. 5a). AuNPs (Fig. 4e and f). This reduction of CD16 expression is attributed This potent response was also reproduced with NK cells when to the blockage of CD16 by the 3G8 antibody present in the pBiAb- pBiAb-AuNPs were added to the primary cell model of HIV latency AuNPs after tissue penetration. In addition, the pBiAb-AuNPs sig­ after viral reactivation; in the absence of nanoparticles, NK cells nificantly decreased HIV infection in this lymphoid tissue, reaching were able to reduce the pool of reactivated cells by a median of Fig. 5. The pBiAb-AuNPs enhance the killing of the latently HIV-infected cells after viral reactivation. (a) Cytotoxic response against latently-infected ACH-2 cells after viral reactivation. PHA and PMA-reactivated ACH-2 cells were cultured with primary NK cells at ratio 1:10 target/effector for 4 h in the presence or absence of the pBiAb-AuNPs at 10 µg/ml. Cytotoxicity was calculated as the disappearance of ACH-2 target cells using flow cytometry count beads. Data is normalized to the control condition in absence of the pBiAb-AuNPs. Mean with SD of 4 independent experiments is shown. (b) Graph representing the cytotoxic response mediated by NK cells in a primary cell model of HIV latency and after viral reactivation with Ingenol (ING) (100 nM) (ratio reactivated cells:NK cells 1:1) in the absence of pBiAb-AuNPs. Median with range of n = 5 experiments is re­ presented. (c) Summary graph of the enhancement of the NK-mediated cytotoxic response against viral reactivated CD4+ T cells induced by the presence of the pBiAb-AuNPs or Irre-AuNPs (2.5 µg/ml total antibody burden) (n = 5). Median with range is represented. Statistical comparisons were performed using the one sample t-test. *p < 0.05. 8 A. Astorga-Gamaza, M. Vitali, M.L. Borrajo et al. Nano Today 36 (2021) 101056 23.08% (Fig. 5b). However, this killing was highly enhanced when CD16, and provide unspecific cell activation. Of note, it is uncertain cells were cultured in the presence of the pBiAb-AuNPs, further the exact orientation of the antibodies on top of the AuNPs. The cell reducing the pool of viral reactivated CD4 T cells by a median of targeting results indicate that the variable chain of the antibodies 58.0% (in comparison to median of 23.20% achieved by the control remains available for antigen recognition. However, it is also possible Irre-AuNPs) (Fig. 5c). Therefore, the properties of our pBiAb-AuNPs that the Fc portion stills accesible and might stimulate NK cells make them also highly valuable in the context of emerging shock through Fc receptors. Both scenarios are likely possible if antibodies and kill clinical strategies. are adsorbed laterally on AuNPs. This would explain the results obtained in the cytotoxic assays, both against cultured gp120-coated Discussion cells and HIV-infected tonsillar tissue cells, in which control Irre- AuNPs showed certain effect. In the mentioned cytotoxicity assays, The development of novel therapeutic strategies to target per­ pBiAb-AuNPs efficiently eliminated gp120-coated cells, providing sistent HIV in patients treated with antiretroviral therapy represents better response than Irre-AuNPs and the A32 free antibody, a non- a current priority in the search for an HIV cure [40,41]. Here, with neutralizing antibody highly potent at performing ADCC [35]. Be­ the aim to generate a potent immune-based approach to specifically sides, the combination of unconjugated A32 and 3G8 resulted in a target and deplete productively HIV-infected and reactivated la­ weak cytotoxic response, further supporting that their combination tently-infected cells, we have constructed novel targeted and po­ and confinement in a polarized and multivalent AuNP platform is larized bispecific nanoparticles. For that, we have optimized and essential to mediate potent anti-HIV effect. developed a linker-free method for promoting cooperative antibody Furthermore, pBiAb-AuNPs were able to rescue NK cells from conjugation, generating AuNPs with 2 specificities; on the one hand HIV-infected patients, triggering a cytotoxic response similar to the the A32 antibody recognizes the HIV protein gp120 displayed in the one provided by the positive control. This is noteworthy, since HIV cell membrane of productively or viral-reactivated HIV-infected persistent infection leads to the appearance of dysfunctional NK cells cells, and on the other hand, the anti-CD16 monoclonal antibody with diminished capacity to kill cell targets [6], and this NK cell (3G8) recognizes the molecule CD16 mainly expressed on the cyto­ exhaustion persists despite the antiretroviral treatment [47–49]. toxic subset of NK cells (CD56dim NK cells). Using this novel con­ Thus, all these properties place pBiAb-AuNPs as an attractive mole­ jugation strategy, antibodies were attached to the AuNPs surface cularly-targeted nanotool to redirect and potentiate the NK cell re­ forming ordered domains and fully retaining their antigen specifi­ sponse to combat persistent HIV infection. Other nano-based city. Thus, pBiAb-AuNPs were able to significantly enhance the approaches trying to boost anti-viral immunity have been previously generation of desired cell doublets, and activate NK cells which in­ developed [50,51]. duced a potent cytotoxic response against both, cultured HIV-ex­ To our knowledge, the pBiAb-AuNPs are the first nanotool de­ pressing cells and infected lymphoid tissue cells. Importantly, and scribed engaging NK cells with HIV-infected cells. Bispecific AuNPs supporting its application in “shock-and-kill” approaches, the pBiAb- have been previously developed to target malignant B cells [52]. AuNPs efficiently reduced the pool of latent HIV-infected cells after However, such dual-affinity AuNPs were generated attaching both pharmacological reactivation. antibodies via a linker in a randomly-distributed manner, which The results presented here support not only the beneficial effects consequently also attached cells of a similar type. In a similar of the pBiAb-AuNPs bispecific platform to eliminate HIV-infected manner, several bispecific antibody derivatives targeting NK cells cells, but also the importance of developing polarizing antibody through CD16 have been explored for cancer therapy [53,54], conjugation methods to construct bispecific nanotools in order to showing CD16 engagement as a powerful way for tilting the balance achieve better biological results. Herein, we carefully studied the in favor of NK activation. conjugation medium and the process to obtain bispecific hard pro­ Importantly, the pBiAb-AuNPs were able to penetrate and greatly tein coronas, driven by a cooperative adsorption of the mix of anti­ reduce the HIV infection in tonsillar lymphoid human tissue. This is bodies, leading to their organization in domains at the AuNP surface. of relevance since the majority of the persistent HIV infection in In the case of non-conjugated non-neutralizing anti-HIV antibodies, antiretroviral treated individuals, which hinders the eradication of such as free A32, indirectly recruit NK cells through their Fc portion the HIV [55], reside in secondary lymphoid tissue such as the lymph and promote ADCC activity. The potency of the ADCC immune re­ nodes and the GALT [38]. The histocultures assays performed here sponse depends on many factors, such as the valence of the anti­ are highly valuable as evidence for in vivo performance, since the body-epitope binding, the antibody orientation on the bounded only animal models that fully recapitulate NK cell function are non- antigen and the resulting size of the immune complex [42–44]. In­ human primates [56]. terestingly, pBiAb-AuNPs multivalency (they held about 32 anti­ Finally, we tested the potential of the pBiAb-AuNPs to promote a bodies at their surface) favored all these conditions to take place and NK-mediated clearance of the HIV-latently infected cells after viral triggered a potent anti-HIV NK-mediated cytotoxic response, which reactivation, mimicking “shock and kill” assays. Latently HIV-in­ was also facilitated by their specific cell engaging capacity. fected cells are supposed to be efficiently killed by the immune NK cells express a broad variety of inhibitory and activating cell system after the pharmacologically reactivation of dormant viruses. receptors, and only a specific balance between the expressions of However, reactivation of viral expression alone has appeared not these receptors triggers a cytotoxic immune response [45]. However, sufficient to eradicate cell reservoirs [57–59]. Thus, the appropriate CD16, the cognate antigen of 3G8 antibody, has been reported as the potentiation of the immune system is essential to fully eliminate the only receptor able to activate cytotoxicity and cytokine secretion latent reservoir. NK cells could help to completely clear viral-re­ without the need of additional signals [46]. We proved that direct activated cells [4,60,61]. The A32 antibody is a good candidate for CD16 engagement by pBiAb-AuNPs helped to achieve a potent anti- targeting latently infected cells in the setting of viral reactivation for HIV cytotoxic response mediated by the stimulation of NK cells and several reasons; it broadly reacts against envelopes from all HIV-1 the enhanced contact with cell targets. Free 3G8 antibody and clades [62], A32-like epitope residues are highly conserved [63], IrreAuNPs slightly increased the percentage of NK cells producing there is no evidence of epitope escape mutations [64], and it targets IFN-γ, showing that in general, targeting Fc receptors could stimulate the earliest known epitope expressed post-infection [35]. Im­ NK cells to produce this cytokine, as previously reported [36]. These portantly, pBiAb-AuNPs extensively reduced the pool of reactivated irrelevant nanoparticles (IrreAuNPs), used as an unspecific control, latently infected cells by triggering a potent NK immune response. are composed of polyclonal IgG mouse and polyclonal IgG rat anti­ This is very relevant since viral-reactivated cells have been reported bodies, which can bind Fc receptors expressed on NK cells, such as to be partially resistant to the well-studied CD8 cell-mediated 9 A. Astorga-Gamaza, M. Vitali, M.L. Borrajo et al. Nano Today 36 (2021) 101056 immune responses, and therefore novel strategies are urgently Nanoparticle conjugation needed [65]. Following our protocol, antibodies were dispersed in the con­ jugating medium, consisting in a 4 mM borate 2.2 mM SC solution. The initial antibody concentrations were 3 µg/ml of the anti- Materials and methods HIVgp120 (A32), and 20 µg/ml of the anti-CD16 (clone 3G8). Then, the solution of AuNPs 2.2 mM was added drop-wise (volume ratio Ethics statement antibodies solution:NP solution of 1:1). The mixture was incubated for 48 h at 4 °C under mild agitation conditions (without magnet). PBMCs from healthy donors were obtained from the Blood and Finally, we removed the conjugates from the unbound antibodies by Tissue Bank, Barcelona, Spain. The study protocol was approved by centrifugation at 5000 g× for 40 min. The difference in the antibody the Comitè d’Ètica d’Investigació Clínica (Institutional Review Board concentration used for each antibody is directly related to their number PR(AG)350/2017) of the Hospital Universitari Vall d’Hebron, different affinity for the AuNP surface, as demonstrated in the iso­ Barcelona, Spain. Samples were obtained from adults, who all pro­ thermal adsorption curves (Fig. 1b and c). We optimized our protocol vided written informed consent, and were prospectively collected of conjugation by exploring different mixtures of ions to increase the and cryopreserved in the Biobanc (register number C.0003590). All stability of the process and the conjugates. We found that 3–10 mM samples received were totally anonymous and untraceable. PBMCs range of borate and 2.2 mM SC solution conferred high stability to from HIV-1-infected patients were also obtained from the HIV unit both antibodies and the nanoparticles (data not shown), allowing a of the Hospital Universitari Vall d’Hebron (Institutional Review soft interaction between them. In addition, it is well known that a Board number PR(AG)350/2017). In addition, tonsil tissue was ob­ slow and controlled addition of nanoparticles to an ion-optimized tained from eight individuals who were undergoing routine tonsil­ solution of diluted antibody (protein), and long-term incubation of lectomy from the Otolaryngology Department of the Hospital the mixture at 4 °C under mild agitation conditions, lead to an or­ Universitari Vall d´Hebron, Barcelona, Spain (Institutional Review dered disposition of the different antibodies into segregated do­ Board number PR(AG)116/2018). All participants gave written in­ mains, forming polarized nanoparticles [13]. This cooperative formed consent for their participation in these studies. adsorption phenomenon, facilitated by a weak perturbation of the conjugation system, translates the mixture to a new equilibrium, leading to polarized NPs [12,19]. Cells, antibodies and reagents In addition, we reproduced a previously reported antibody-ad­ sorption conjugation protocol to generate bispecific gold nano­ We used the cell lines CEM.NKR CCR5+ and ACH-2 (obtained particles [33]. This protocol leads to a random disposition of the through the NIH AIDS Reagent Program, Division of AIDS, NIAID, NIH, antibodies on the surface of gold nanoparticles (non-polarized). from Dr. Alexandra Trkola [66,67] and Dr. Thomas Folks [68,69], Briefly, 133 µl of 15 mM borate buffer pH 8.7 was added to 1 ml of respectively). PBMCs were isolated by Ficoll-Paque density gradient AuNPs. 3 µg/ml of A32 antibody and 20 µg/ml of 3G8 were added and centrifugation and cryopreserved in liquid nitrogen. Cells were cul­ allowed to react under agitation for 30 min. Afterwards, sucrose was tured in RPMI 1640 medium supplemented with 10% fetal bovine incorporated to a final concentration of 5% and incubated for 30 min. serum (FBS; Gibco, Life Technologies, Inc.), 100 U/ml penicillin, and Then, 160 µl of 3% bovine serum albumin (BSA) were added and 100 µg/ml streptomycin (Life Technologies, Inc.) (R10; control shaken for 10 min. Thereafter, we centrifuged the sample (7500 g× medium), and maintained at 37 °C in a 5% CO2 incubator. The A32 for 30 min) to remove unbound antibodies and AuNPs were re-sus­ antibody was obtained through the AIDS Research and Reference pended in 1 ml of 2 mM borate buffer pH 8.7 containing 5% sucrose, Program, NIAID, NIH (Cat#11438) from Dr. James E. Robinson [70,71]. 2% glycerol, 0.5% BSA, and 0.01% Tween. The washing step was re­ The anti-CD16 clone 3G8 was purchased from Stemcell. BaL gp120 peated and the conjugate was re-suspended in 100 µl of the men­ recombinant protein was also obtained through the NIH AIDS Re­ tioned borate buffer. agent Program. Polyclonal IgG mouse and IgG rat were purchased from Sigma-Aldrich. IL-2 was obtained from the Vall d´Hebron Physicochemical characterization of BiAb-AuNPs Hospital pharmacy. Physicochemical characteristics of the synthesized nanoparticles Synthesis of gold nanoparticles and conjugation with antibodies were assessed by different methodologies, such as electron micro­ scopy. To characterize the different conjugates, we evaluated sam­ Chemicals ples by UV-Vis spectrophotometry (Cary 60 UV-Vis Sodium citrate tribasic dihydrate (≥99%), gold (III) chloride tri­ Spectrophotometer, Agilent), Dynamic Light Scattering (DLS) and hydrate HAuCl4•3H2O (99.9% purity) and sodium tetraborate dec­ Zeta potential experiments. DLS and Zeta potential experiments ahydrate were purchased from Sigma-Aldrich. All reagents were were conducted in a Zetasizer Nano series instrument (Malvern). used as received without further purification and all glass material The particles size and Zeta potential were measured simultaneously was sterilized and dehydrogenated in an oven prior to use. Mili-Q three times. Previously to the characterization experiments, samples water was used in the preparation of all solutions. were centrifuged to remove the unbound antibodies. TEM - SEM analysis Nanoparticle synthesis Transmission Electron Microscopy. Gold nanoparticles and Gold Aqueous solutions of sterile endotoxin free ~40 nm citrate-sta­ nanoparticles – antibody complexes were visualized using FEI bilized AuNPs were synthetized according to previously developed MAGELLAN 400 L XHR SEM. This microscope allows for images ac­ seeded-growth method following a citrate reduction protocol as quisition in the high energy (15–30 kV) SEM and STEM mode. The previously described [72]. AuNPs were obtained after different se­ ultrathin Formvar-coated 200-mesh copper grid (Ted-pella, Inc.) quential steps of growing, yielding 5 × 1010 NPs/ml (33 µg/ml). After were directly dipped in the sample solution and left to dry in air purification by centrifugation, to discard by-products and con­ overnight. TEM images of the prepared colloidal AuNP were used for taminants, NPs were resuspended in a solution of 2.2 mM sodium the size distribution measurements. For each sample, the size of at citrate (SC) and stored at 4 °C in the dark. All the particles were used least 100 particles was measured and the average size and the within 20 days after their synthesis. standard distribution were obtained. TEM and SEM images of the 10 A. Astorga-Gamaza, M. Vitali, M.L. Borrajo et al. Nano Today 36 (2021) 101056 prepared AuNP – antibody complexes were used for the analysis of HSA for 24 h. Then, the sample containing the three types of AuNPs surface contact between AuNPs in the case of total antibody surface was purified from the excess of smaller AuNPs by decantation and passivation or partial antibody surface passivation. The charging after that, the TEM grids were prepared by dipping, by directly im­ effect observable on the edge of the NPs visualized by SEM is asso­ mersing the grids in the AuNPs solution and letting them dry in an ciated to the presence of the organic layer. This effect provides in­ open atmosphere overnight. Finally, the sample was analyzed by TEM, formation on the degree of NP surface passivation by the antibody. using an 80-keV TEM JEOL 1010 equipped with an Orius (Gatan) CCD Camera. TEM grids consisted on an ultrathin-formvar coated 200 mesh UV-Visible absorption spectroscopy copper grid covered with a layer of carbon (Ted-Pella). The TEM UV-Visible absorption (UV-Vis) spectra were acquired on an images of 232 AuNPs of 60 nm were randomly obtained and classified Agilent Cary 60 UV-Vis Spectrophotometer. A total volume of 1 ml of depending on the apparent polarization formed by the specific re­ the analysed sample was placed in a plastic cell and the spectral cognition between 25 nm and 15 nm label AuNPs. analysis was performed between 300 nm and 800 nm at room temperature. Cary WinUV software by Agilent has been used to Thermogravimetric assay collect the spectrophotometer output. Thermogravimetric analyses (TGA) were carried out on a Pyris DLS and Zeta potential measurements TGA 8000 under N2 flow with a temperature range from 30 °C to The hydrodynamic diameter and the surface charge of the col­ 600 °C at a heating rate of 5 °C min−1. Considering that the limit of loids were measured by Dynamic Light Scattering (DLS) and Zeta detection is a loss in mass of 5 μg of antibody, a total volume of 417 µl Potential on a Malvern Zetasizer Nano ZS90 which incorporates a Z- of AuNPs was conjugated with mouse IgG producing 0.269 mg of potential analyser (Malvern Instruments Ltd, Worcestershire, UK). bioconjugates. The purified bioconjugates were deposited in a ceramic crucible, allowing the excess of water to evaporate during Bioconjugates analysis by TEM-SEM 24 h in a laminar flow cabinet. For conjugation, 50 nm AuNPs were used at final concentration Simulation model 4 × 1010 NP/ml (total surface area approximately 2 ×1014 nm2/ml). AuNPs were exposed to IgG molecules at two different final protein Coarse-grained (CG) models were employed to extract the re­ concentration, corresponding to (i) a situation in which the number levant characteristics for the studied processes [74]. In our model, of IgG molecules theoretically cover the 60% of the total NPs surface, one bead comprises three heavy atoms. The bead interaction para­ and (ii) a situation in which IgG molecules theoretically cover the meters, topology and composition are chosen to mimic the inherent total NP surface with a 100x excess of antibody molecules. The total chemistry for the atoms. The employed model is Dissipative Particle NP surface was calculated knowing the surface area of the 50 nm Dynamics (DPD) [75], a stochastic CG simulation model that has been AuNPs (7.85 × 103 nm2 per particle). The IgG surface area used to used to study soft matter [76], complex fluids and biomolecular estimate the theoretical number of IgG molecules in solutions, and systems, such as self-assembling amphiphilic micelles and bilayers, consequently, the total NP surface coverage was calculated according polymers, proteins, nanoparticles [18,77] and colloidal systems [78]. to the relation with molecular weight specific for globular proteins The cooperative adsorption of immunoglobulins onto AuNPs is [73] and approximating the IgG molecules to a sphere. IgG was ad­ studied by DPD-MC hybrid model. All the simulations are performed sorbed onto AuNPs by incubating the AuNPs with IgG in 10 mM in a NPT ensemble (constant number of particles N, pressure P and Borate Buffer pH 8.5 for a total time of 48 h at 4 °C under mild agi­ temperature, T) using a dimensionless units system. For our para­ tation (650 RPM). After 48 h of incubation time, IgG-conjugated NPs meterization, ρ = 3, m = 3 and V w = 30 Å and the friction constant (λ) were analyzed without further purification. necessary for evaluating the dissipative force is set equal to 4.5. For each system, the length of the simulation is normally 300,000 cycles, Bioconjugation to Nivolumab equivalent to 1350 ns if using 0.03τ as time step. The working tem­ perature use throughout the simulations is Kb T = 0.42 [79,80]. 50 nm AuNPs were conjugated to Nivolumab antibody. To per­ All experiments are performed in a cubic box water of form bioconjugation, we calculated the theoretical number of bio­ L x × L y × L z = 30 × 30 × 30d03. Simulations contain 80,000 beads of molecules required to cover the whole NP surface knowing the water (equivalent to 240,000 molecules of water), 3000 beads surface area of the 50 nm AuNPs (7.85 × 103 nm2 per particle) and the of gold (equivalent to 9000 gold atoms) that form a unique AuNP and surface coverage of Nivolumab molecule. For Nivolumab, we esti­ 100 molecules of each biological ligand. All compounds are ran­ mated the number of antibodies molecules needed to cover the NP domly distributed along the box. surface according to the relation with molecular weight specific for globular proteins [73]. Incubation was performed in 10 mM Phos­ Cell targeting assays phate Buffer, pH 7.4 at Room Temperature, by directly diluting the AuNP in the protein solution in the presence of a 100x Nivolumab We assessed the binding capabilities of conjugated A32 (anti- excess, in order to ensure complete surface coverage. HIVgp120) and 3G8 (anti-CD16) antibodies to their targets by flow cytometry. Briefly, for evaluation of A32-monoconjugates we used Distribution of antibodies on cooperatively-conjugated AuNPs by TEM the CEM.NKR CCR5+ cell line. Cells were coated with 1 µg of the HIV- 1 BaL gp120 recombinant protein during 1 h at room temperature With the aim to directly observe if the optimized conjugation (RT). After incubation, cells were extensively washed in ice-cold R10 method leads to the segregation of a mix of antibodies in two different medium and staining buffer (PBS containing 3% FBS). Thereafter, domains (polarization), we performed a straight forward experiment. cells were incubated for 20 min at RT in R10 medium containing On the one hand, AuNPs of ~60 nm were doubly conjugated with naked AuNPs, AuNPs conjugated with A32 or free A32, and stained rabbit anti-HSA IgG (18 µg/ml) and rabbit anti-BSA IgG (18 µg/ml). On with an anti-human FITC-labelled secondary antibody (dil 1:100) the other hand, we conjugated smaller AuNPs as labels: AuNPs of (Thermo Fisher) for A32 detection (20 min at RT). To assess the 25 nm were functionalized with BSA (5 µM) and of 15 nm functiona­ targeting capacity of 3G8 monoconjugates, NK cells were obtained lized with HSA (5 µM). The bispecific conjugates were incubated with from PBMCs of healthy donors by negative isolation using magnetic the AuNPs functionalized with BSA and the AuNPs functionalized with beads (MagniSort™ Human NK cell Enrichment Kit, eBioscience). 11 A. Astorga-Gamaza, M. Vitali, M.L. Borrajo et al. Nano Today 36 (2021) 101056 Then, NK cells were incubated in medium containing naked AuNPs, and anti-HLA-DR-SB600 (LN3; eBioscience) antibodies for 20 min at AuNPs conjugated with 3G8 or free 3G8. For detection of 3G8 by flow RT. Cells were then fixed and permeabilized with Fixation/ cytometry, we used an anti-mouse AF647-labelled secondary anti­ Permeabilization Solution (Becton Dickinson) for 20 min at 4 °C, body (dil 1:200) (Invitrogen). R10 media controls were included for washed with BD Perm/Wash buffer and stained with anti-IFN-ɤ both conjugates. Cells were then fixed with paraformaldehyde so­ AF700 (Life technologies) for 30 min at 4 °C. After washing, cells lution (PFA; Affymetrix) (2%) and acquired in an LSR Fortessa flow were fixed with PFA (2%) and acquired on an LSR Fortessa flow cytometer (Becton Dickinson). Results were analyzed with FlowJo cytometer (Becton Dickinson). Data were analyzed using FlowJo V10 v10 software (Tree Star). software. Cell-to-cell contact assays Antibody-dependent cell-mediated cytotoxicity (ADCC) assay CD4+ T cells and NK cells were isolated from cryopreserved The ADCC assay was performed as described by others with some PBMCs of healthy donors by commercial kits (MagniSort Human modifications [81]. Briefly, CEM-NKR. CCR5 cells were used as target CD4+ T Cell Enrichment; Affymetrix, and MagniSort™ Human NK cell cells after being double-stained with PKH67 (Sigma-Aldrich) and Enrichment; eBioscience). CD4+ T cells were coated with 1 µg of re­ eF670 (Labclinics) dyes following manufacturer’s instructions. Then, combinant gp120 protein during 1 h at RT. After extensive washes, cells were coated with 1 µg of the HIV-1 BaL gp120 recombinant CD4+ T cells were mixed with NK cells in a 1:1 ratio and stained with protein for 1 h at RT and extensively washed in ice-cold R10 medium. anti-CD56 (PE; Becton Dickinson) and anti-CD3 (PE-Cy7; Becton Target cells were dispensed in U-bottom 96-well plates (5000 Dickinson) antibodies for detection of NK cells and CD4 T cells, re­ cells/well) and incubated for 15 min with 5 µg/ml of free A32, 5 µg/ spectively. After washing, we added naked AuNPs or BiAb-AuNPs at ml of free 3G8, Irre-AuNPs or BiAb-AuNPs (10 µg/ml antibody two different doses (2.5 µg/ml and 10 µg/ml of total antibody). After burden), and plasma (1:1000 dilution) from a viremic (high viral 20 min incubation at RT, cells were washed and fixed with PFA (2%). load in blood) HIV+ patient as a positive control. After incubation, NK Samples were then acquired on a LSR Fortessa flow cytometer effector cells isolated from PBMCs of healthy donors or antiretroviral (Becton Dickinson) and data were analyzed using FlowJo V10 therapy (ART) suppressed HIV+ patients were added at 1:10 target/ software. effector ratio. Plates were centrifuged and incubated for 4 h at 37 °C and 5% CO2. Finally, cells were washed, fixed with PFA (2%), acquired Confocal microscopy on a LSR Fortessa flow cytometer (Becton Dickinson) and analyzed using FlowJo software. Flow cytometry count beads (AccuCount Attachment of BiAb-AuNPs to their targets and subsequently Blank Particles, Cytognos) are used to normalize cell collection to a promotion of dual cell conjugates were assessed by confocal mi­ constant number of particles (1000 events). Target cells were iden­ croscopy. First, CEM.NKR CCR5+ cells were stained with 2 µM of the tified in a PKH67-versus side scatter (SSC) plot. Loss of the eF670 membrane lipid marker PKH67 (Sigma-Aldrich) according to the marker was used to determine the percentage of killing in an eF670 manufacturer instructions. These cells were also coated with 1 µg of versus PKH67 plot. ADCC on virally-reactivated cells was performed HIV-1 BaL gp120 recombinant protein for 1 h at RT. Then, NK cells as follows; the latently-infected cell line ACH-2 was stimulated to isolated from PBMCs of healthy donors were obtained as described produce HIV by the addition of 10 µg/ml PHA (Phytohemagglutinin; above. NKs and PKH67-labelled CEM.NKR CCR5+ cells were mixed at Fisher Scientific) and 10 nM PMA (phorbol myristate acetate; 1:1 ratio and incubated for 20 min at RT with pBiAb-AuNPs Abcam) during 17 h at 37 °C. Then, cells were subjected to the ADCC (2.5 µg/ml of total antibody burden), then attached to coverslips (by assay as described above. ADCC of viral-reactivated cells was cal­ centrifugation at 800 ×g for 5 min) previously coated with 0.1 mg/ml culated as the fraction of cells that disappeared within the target of poly-L-lysine (Sigma-Aldrich) and fixed with PFA (4%) for 15 min population after the addition of pBiAb-AuNPs in comparison to the at RT. Cells were finally stained with DAPI (1:5000 dilution) (Thermo control condition with targets and NK cells, but lacking the nano­ Fisher) and mounted with Fluoromount G (eBioscience). particles. To assess the absolute number of cells killed by ADCC, we Preparations were imaged with an Olympus Spectral Confocal added to each well flow cytometry particles for absolute cell Microscope FV1000. ImageJ software was used for image composi­ counting (5 × 104/ml) (AccuCount Blank 5.0–5.9 µm, Cytognos). tions. A total of 20 images were counted, containing 1650 cells. Note that AuNPs of 40 nm do not reflect or diffract light because they are Tonsillar tissue processing and infection below the visible light resolution limit, but they are good at dis­ persing light, especially if aggregated. Thus, the light scattering/ Tonsils from healthy children were obtained and processed as dispersion can be observed with a confocal microscope [25]. previously described [82]. Briefly, the tissue was processed within the first 2 h following surgery. We dissected the tissue into uniform NK cell activation and functional activity assays blocks of approximately 2 mm × 2 mm x 1 mm, after discarding all cauterized, bloody or necrotic areas. Ten blocks per condition were The activation and cytotoxicity phenotype of NK cells after sti­ placed on top of a piece of sponge (Absorbable gelatin Surgispon; mulation with pBiAb-AuNPs was evaluated. Cytotoxicity was as­ AEGIS LIFESCIENCES) suspended in RPMI1640 medium supple­ sessed by the measurement of CD107a and IFN-ɤ. Briefly, PBMCs mented with 20% FBS (R20) and 100 U/ml penicillin and 100 µg/ml were cultured in R10 medium with 10 µg/ml pBiAb-AuNPs, 10 µg/ml streptomycin in a 6-well plate. Then, for viral infection experiments, Irre-AuNPs, 5 µg/ml of A32, 5 µg/ml of 3G8 or 10 ng/ml PMA plus 5 µl of HIVBal virus (457 TCID50) were added on top of each tissue 1 µM ionomycin (positive control) for 4.5 h in a 96-well plate at 37°C block. Some pieces were not infected as a control. After 15 min in­ and 5% CO2. CD107a-PE-Cy5 (H4A3; Becton Dickinson), BD GolgiPlug cubation, we added 5 µl of pBiAb-AuNPs or Irre-AuNPs at 5 µg/ml of Protein Transport Inhibitor (Beckton Dickinson) and BD GolgiStop total antibody burden on each block and placed the plate in the Protein Transport Inhibitor containing monensin (Becton Dickinson) incubator (37 °C 5% CO2). We performed two different assays with were also added to each well at the recommended concentrations. the tonsillar blocks. First, after one day incubation with the nano­ Cells were then washed and stained with a viability dye (LIVE/DEAD particles, we digested the blocks as previously described [82] and Fixable Violet dead cell stain; Thermo Fisher). To measure the acti­ performed the staining with LIVE/DEAD Far Red viability (Thermo vation marker HLA-DR, cells were stained with anti-CD56-FITC Fisher) to identify dead cells for 30 min at RT. After washing with (B159; Becton Dickinson), anti-CD3-PE-Cy7 (SK7; Becton Dickinson) PBS, we stained the digested cells with the following labeled 12 A. Astorga-Gamaza, M. Vitali, M.L. Borrajo et al. Nano Today 36 (2021) 101056 antibodies: anti-CD45-FITC (HI30; Biolegend), anti-CD3-PerCP (SK7; Spanish “Ministerio de Economía y Competitividad, Instituto de Becton Dickinson), anti-CD8-APC (RPA-T8; Becton Dickinson) and Salud Carlos III” (ISCIII, PI14/01058) to J.G.P, a research grant from anti-CD16-PE (3G8; Biolegend), for 20 min at RT. Then, cells were Gilead Sciences (GLD17-00204 and GLD19-00084) to M.B, GeSIDA washed and fixed with PFA (2%). Second, we assessed HIV infection and the Spanish AIDS network “Red Temática Cooperativa de at day 6 of tonsil culture by tissue digestion and staining with LIVE/ Investigación en SIDA” (RD16/0025/0007). The Miguel Servet pro­ DEAD Far Red viability, anti-CD45-FITC, anti-CD3-PerCP and anti- gram funded by the Spanish Health Institute Carlos III (CP17/00179) CD8-APC as described. After the surface staining, cells were fixed to M.B and J.G.P (CPII15/00014). The “Pla estratègic de recerca i and permeabilized with Fixation/Permeabilization Solution (Becton innovació en salut” (PERIS), from the Catalan Government to M.G. Dickinson) for 20 min at 4 °C, washed with BD Perm/Wash buffer and The Spanish Secretariat of Science and Innovation Ph.D. fellowship to stained with anti-p24-PE (Beckman Coulter) for 20 min on ice and A.A-G (BES-2016-076382), AGAUR-FI-B-00582 Ph.D. fellowship from 20 min at RT. Finally, cells were washed and fixed with PFA (2%). the Catalan Government to O.BL, and PIF-UAB Ph.D. fellowship from Samples were acquired in a FACSCalibur flow cytometer (Becton Universitat Autònoma de Barcelona to R.SL. The funders had no role Dickinson) and analyzed with FlowJo V10 software. in study design, data collection and analysis, the decision to publish or preparation of the manuscript. Generation of latency model, viral reactivation and cytotoxicity assays CRediT authorship contribution statement PBMCs from healthy donors or HIV-infected patients were thawed. Then, CD4 T cells were isolated using a commercial kit (MagniSort Antonio Astorga: Conceptualization, Methodology, Validation, Human CD4+ T Cell Enrichment; Affymetrix) and infected with HIV Formal analysis, Data curation, Writing - original draft, Visualization. (350,000 TDCI50/million cells) with NL4.3 viral strain for 4 h at 37 °C Michele Vitali: Methodology, Formal analysis, Visualization. Mireya 5% CO2 or by spinoculation at 1200 ×g and 37 °C for 2 h. Cells were L. Borrajo: Methodology, Visualization. Rosa Suárez-Lopez: washed twice with PBS and cultured in R10 with IL-2 (10 U/ml) and Methodology, Software, Formal analysis. Carlos Jaime: Methodology, IL-7 (1 nM) (Bio-Techne R&D Systems). After 2 days of infection, Software, Formal analysis. Neus Bastus: Methodology. Carla Serra- Raltegravir (1 µM), Darunavir (1 µM) and Nevirapin (1 µM) were Peinado: Methodology. Laura Luque-Ballesteros: Methodology. added to the cell culture to prevent new rounds of viral infection. Next Oscar Blanch-Lombarte: Methodology. Julia G. Prado: day, cells were incubated with Q-VD (10 µM) for 2 h at 37 °C and 5% Methodology, Writing - review & editing. Juan Lorente: Resources. CO2, in R10 medium containing IL-2, IL-7, Raltegravir, Darunavir and Felix Pumarola: Resources. Marc Pellicer: Resources. Vicenç Falcó: Nevirapin at the indicated concentrations, and then latent viral in­ Resources, Writing - review & editing. Meritxell Genescà: Resources, fection was reactivated during 22 h with Ingenol (100 nM) and sub­ Writing - review & editing. Victor Puntes: Conceptualization, jected to the cytotoxic NK assay. NK cells were added at ratio 1:1 in Methodology, Validation, Investigation, Writing - original draft, the presence or absence of pBiAb-AuNPs or irrelevant AuNPs (2.5 ug/ Supervision, Project administration. Maria J. Buzon: ml total antibody burden), for 22 h at 37 °C and 5% CO2. In some cases, Conceptualization, Methodology, Validation, Investigation, Writing - we included as a negative control AuNPs conjugated with the anti­ original draft, Supervision, Project administration, Funding acqui­ body Palivizumab. The cytotoxic assays were performed in round sition. bottom 96 wells plates, at a cellular concentration of 2.5 M cells/ml. Finally, cells were stained with LIVE/DEAD violet viability (Thermo Declaration of competing interest Fisher) for 20 min at RT. After washing once with staining buffer, cells were stained with anti-CD3-AF700 (SK7; Biolegend) and anti-CD56- The authors declare that they have no known competing fi­ FITC (B159; Becton Dickinson) antibodies for 20 min at RT. Cells were nancial interests or personal relationships that could have appeared then fixed and permeabilized with Fixation/Permeabilization Solution to influence the work reported in this paper. (Becton Dickinson) for 20 min at 4 °C, washed with BD Perm/Wash buffer and stained with anti-p24-PE (Beckman Coulter) for 20 min on Acknowledgements ice and 20 min at RT. After washing, cells were fixed with PFA (2%). Samples were acquired on an LSR Fortessa flow cytometer (Becton The authors thank Jana Oliveras and Javy Patarroyo for helping in Dickinson) and data analyzed using FlowJo V10 software. the acquisition of TEM and SEM images of NP antibody complexes. Cell toxicity of pBiAb-AuNPs and Irre-AuNPs was measured using the primary cell line CEM-NKR and ex-vivo infected primary CD4+ T Appendix A. Supporting information cells. Cells were incubated with the different nanoconjugates for 4 and 48 h. Then, cell viability was determined by staining with a LIVE/ Supplementary data associated with this article can be found in DEAD viability dye and measured by flow cytometry. the online version at doi:10.1016/j.nantod.2020.101056. Statistical analysis References Analyses were performed with GraphPad Prism v6 and statistical [1] E.A. Zhukovsky, R.J. Morse, M.V. Maus, Bispecific antibodies and CARs: gen­ eralized immunotherapeutics harnessing T cell redirection, Curr. Opin. Immunol. tests are reported within each figure legend. p values < 0.05 were 40 (2016) 24–35. considered statistically significant. [2] N. Huot, B. Jacquelin, T. Garcia-Tellez, P. Rascle, M.J. Ploquin, Y. Madec, R.K. Reeves, N. Derreudre-Bosquet, M. Müller-Trutwin, Natural killer cells mi­ grate into and control simian immunodeficiency virus replication in lymph node Funding follicles in African green monkeys, Nat. Med. 23 (11) (2017) 1277–1286. [3] G. Alter, M.P. Martin, N. Teigen, W.H. Carr, T.J. Suscovich, A. Schneidewind, This study was supported by the Spanish Secretariat of Science H. Streeck, M. Waring, A. Meier, C. Brander, J.D. Lifson, T.M. Allen, M. Carrington, M. Altfeld, Differential natural killer cell-mediated inhibition of HIV-1 replication and Innovation and FEDER funds (grants SAF2015-67334-R and based on distinct KIR/HLA subtypes, J. Exp. Med. 204 (12) (2007) 3027–3036. RTI2018-101082-B-I00 [MINECO/FEDER]), American National [4] F. Marras, A. Casabianca, F. Bozzano, M.L. Ascierto, C. Orlandi, A. Di Biagio, Institutes of Health (grant R21AI118411 to M.B), an unrestricted re­ E. Pontali, C. Dentone, G. Orofino, L. Nicolini, L. Taramasso, M. Magnani, F.M. Marincola, E. Wang, L. Moretta, A. De Maria, Control of the HIV-1 DNA re­ search grant from Bristol-Myers Squibb S.A.U (PfC-2015-AI424-564) servoir is associated in vivo and in vitro with NKp46/NKp30 (CD335 CD337) to M.B, the Spanish “Ministerio de Economía y Competitividad, inducibility and interferon gamma production by transcriptionally unique NK Instituto de Salud Carlos III” (ISCIII, PI17/01470) to M.G and the cells, J. Virol. 91 (23) (2017). 13 A. Astorga-Gamaza, M. Vitali, M.L. Borrajo et al. Nano Today 36 (2021) 101056 [5] T.W. Chun, L. Stuyver, S.B. Mizell, L.A. Ehler, J.A.M. Mican, M. Baseler, A.L. Lloyd, [34] P.K. Jain, K.S. Lee, I.H. El-Sayed, M.A. El-Sayed, Calculated absorption and scat­ M.A. Nowak, A.S. Fauci, Presence of an inducible HIV-1 latent reservoir during tering properties of gold nanoparticles of different size, shape, and composition: highly active antiretroviral therapy, Proc. Natl. Acad. Sci. U. S. A. 94 (24) (1997) applications in biological imaging and biomedicine, J. Phys. Chem. B 110 (14) 13193–13197. (2006) 7238–7248. [6] J. Bi, Z. Tian, NK cell exhaustion, Front. Immunol. 8 (2017) 760. [35] G. Ferrari, J. Pollara, D. Kozink, T. Harms, M. Drinker, S. Freel, M.A. Moody, [7] L. Zhang, C. Chung, B.S. Hu, T. He, Y. Guo, A.J. Kim, E. Skulsky, X. Jin, A. Hurley, S.M. Alam, G.D. Tomaras, C. Ochsenbauer, J.C. Kappes, G.M. Shaw, J.A. Hoxie, B. Ramratnam, M. Markowitz, D.D. Ho, Genetic characterization of rebounding J.E. Robinson, B.F. Haynes, An HIV-1 gp120 envelope human monoclonal anti­ HIV-1 after cessation of highly active antiretroviral therapy, J. Clin. Investig. 106 body that recognizes a C1 conformational epitope mediates potent antibody- (7) (2000) 839–845. dependent cellular cytotoxicity (ADCC) activity and defines a common ADCC [8] H. Yuan, W. Jiang, C.A. von Roemeling, Y. Qie, X. Liu, Y. Chen, Y. Wang, epitope in human HIV-1 serum, J. Virol. 85 (14) (2011) 7029–7036. R.E. Wharen, K. Yun, G. Bu, K.L. Knutson, B.Y.S. Kim, Multivalent bi-specific na­ [36] C. Jacobi, M. Claus, B. Wildemann, S. Wingert, M. Korporal, J. Römisch, S. Meuer, nobioconjugate engager for targeted cancer immunotherapy, Nat. Nanotechnol. C. Watzl, T. Giese, Exposure of NK cells to intravenous immunoglobulin induces 12 (8) (2017) 763–769. IFN gamma release and degranulation but inhibits their cytotoxic activity, Clin. [9] W. Jiang, B.Y.S. Kim, J.T. Rutka, W.C.W. Chan, Nanoparticle-mediated cellular Immunol. 133 (3) (2009) 393–401. response is size-dependent, Nat. Nanotechnol. 3 (3) (2008) 145–150. [37] J. Cantero, M. Genesca, Maximizing the immunological output of the cervi­ [10] S. Bhattacharyya, R. Bhattacharya, S. Curley, M.A. McNiven, P. Mukherjee, covaginal explant model, J. Immunol. Methods 460 (2018) 26–35. Nanoconjugation modulates the trafficking and mechanism of antibody in­ [38] J.D. Estes, C. Kityo, F. Ssali, L. Swainson, K.N. Makamdop, G.Q. Del Prete, duced receptor endocytosis, Proc. Natl. Acad. Sci. U. S. A. 107 (33) (2010) S.G. Deeks, P.A. Luciw, J.G. Chipman, G.J. Beilman, T. Hoskuldsson, A. Khoruts, 14541–14546. J. Anderson, C. Deleage, J. Jasurda, T.E. Schmidt, M. Hafertepe, S.P. Callisto, [11] C. Garrido, C.A. Simpson, N.P. Dahl, J. Bresee, D.C. Whitehead, E.A. Lindsey, H. Pearson, T. Reimann, J. Schuster, J. Schoephoerster, P. Southern, K. Perkey, T.L. Harris, C.A. Smith, C.J. Carter, D.L. Feldheim, C. Melander, D.M. Margolis, Gold L. Shang, S.W. Wietgrefe, C.V. Fletcher, J.D. Lifson, D.C. Douek, J.M. McCune, nanoparticles to improve HIV drug delivery, Future Med. Chem. 7 (9) (2015) A.T. Haase, T.W. Schacker, Defining total-body AIDS-virus burden with implica­ 1097–1107. tions for curative strategies, Nat. Med. 23 (11) (2017) 1271–1276. [12] J.J. Ramsden, G.I. Bachmanova, A.I. Archakov, Kinetic evidence for protein clus­ [39] J. Kennedy, E. Larrañeta, M.T.C. McCrudden, C.M. McCrudden, A.J. Brady, tering at a surface, Phys. Rev. E Stat. Phys. Plasmas Fluids Relat. Interdiscip. Top. S.J. Fallows, H.O. McCarthy, A. Kissenpfennig, R.F. Donnelly, In vivo studies in­ 50 (6) (1994) 5072–5076. vestigating biodistribution of nanoparticle-encapsulated rhodamine B delivered [13] E. Casals, T. Pfaller, A. Duschl, G.J. Oostingh, V. Puntes, Time evolution of the via dissolving microneedles, J. Control. Release 265 (2017) 57–65. nanoparticle protein corona, ACS Nano 4 (7) (2010) 3623–3632. [40] S. Cao, K.A. Woodrow, Nanotechnology approaches to eradicating HIV reservoirs, [14] V.V. Hlady, J. Buijs, Protein adsorption on solid surfaces, Curr. Opin. Biotechnol. 7 Eur. J. Pharm. Biopharm. 138 (2019) 48–63. (1) (1996) 72–77. [41] A. Bowen, E.E. Sweeney, R. Fernandes, Nanoparticle-based immunoengineered [15] M.P. Monopoli, C. Åberg, A. Salvati, K.A. Dawson, Biomolecular coronas provide approaches for combating HIV, Front. Immunol. 11 (2020) 789. the biological identity of nanosized materials, Nat. Nanotechnol. 7 (12) (2012) [42] A.W. Boesch, E.P. Brown, M.E. Ackerman, The role of Fc receptors in HIV pre­ 779–786. vention and therapy, Immunol. Rev. 268 (1) (2015) 296–310. [16] E. Colangelo, J. Comenge, D. Paramelle, M. Volk, Q. Chen, R. Lévy, Characterizing [43] A. Lux, X. Yu, C.N. Scanlan, F. Nimmerjahn, Impact of immune complex size and self-assembled monolayers on gold nanoparticles, Bioconjugate Chem. 28 (1) glycosylation on IgG binding to human FcgammaRs, J. Immunol. 190 (8) (2013) (2017) 11–22. 4315–4323. [17] L. Vroman, Effect of absorbed proteins on the wettability of hydrophilic and [44] P. Acharya, W.D. Tolbert, N. Gohain, X. Wu, L. Yu, T. Liu, W. Huang, C. Huang, hydrophobic solids, Nature 196 (1962) 476–477. Y.D. Kwon, R.K. Louder, T.S. Luongo, J.S. McLellan, M. Pancera, Y. Yang, B. Zhang, [18] A. Raman, C. Jaime, V.F. Puntes, Domain formation and conformational changes R. Flinko, J.S. Foulke, M.M. Sajadi, R. Kamin-Lewis, J.E. Robinson, L. Martin, in gold nanoparticle conjugates studied using DPD simulations, Langmuir 33 P.D. Kwong, Y. Guan, A.L. DeVico, G.K. Lewis, M. Pazgier, Structural definition of (50) (2017) 14502–14512. an antibody-dependent cellular cytotoxicity response implicated in reduced risk [19] A.M. Jackson, J.W. Myerson, F. Stellacci, Spontaneous assembly of subnanometre- for HIV-1 infection, J. Virol. 88 (21) (2014) 12895–12906. ordered domains in the ligand shell of monolayer-protected nanoparticles, Nat. [45] L.L. Lanier, Up on the tightrope: natural killer cell activation and inhibition, Nat. Mater. 3 (5) (2004) 330–336. Immunol. 9 (5) (2008) 495–502. [20] C. Singh, P.K. Ghorai, M.A. Horsch, A.M. Jackson, R.G. Larson, F. Stellacci, [46] Y.T. Bryceson, M.E. March, H.G. Ljunggren, E.O. Long, Synergy among receptors S.C. Glotzer, Entropy-mediated patterning of surfactant-coated nanoparticles on resting NK cells for the activation of natural cytotoxicity and cytokine se­ and surfaces, Phys. Rev. Lett. 99 (22) (2007) 226106. cretion, Blood 107 (1) (2006) 159–166. [21] M.M. Sajadi, G.K. Lewis, M.S. Seaman, Y. Guan, R.R. Redfield, A.L. DeVico, [47] L. Azzoni, E. Papasavvas, J. Chehimi, J.R. Kostman, K. Mounzer, J. Ondercin, Signature biochemical properties of broadly cross-reactive HIV-1 neutralizing B. Perussia, L.J. Montaner, Sustained impairment of IFN-gamma secretion in antibodies in human plasma, J. Virol. 86 (9) (2012) 5014–5025. suppressed HIV-infected patients despite mature NK cell recovery: evidence for [22] J. Piella, N.G. Bastus, V. Puntes, Size-dependent protein-nanoparticle interactions a defective reconstitution of innate immunity, J. Immunol. 168 (11) (2002) in citrate-stabilized gold nanoparticles: the emergence of the protein corona, 5764–5770. Bioconjugate Chem. 28 (1) (2017) 88–97. [48] J. Chehimi, L. Azzoni, M. Farabaugh, S.A. Creer, C. Tomescu, A. Hancock, [23] P.P. Pillai, B. Kowalczyk, W.J. Pudlo, B.A. Grzybowski, Electrostatic titrations re­ A. Mackiewicz, L. D’Alessandro, S. Ghanekar, A.S. Foulkes, K. Mounzer, J. Kostman, veal surface compositions of mixed, on-nanoparticle monolayers comprising L.J. Montaner, Baseline viral load and immune activation determine the extent of positively and negatively charged ligands, J. Phys. Chem. C 120 (2016) reconstitution of innate immune effectors in HIV-1-infected subjects undergoing 4139–4144. antiretroviral treatment, J. Immunol. 179 (4) (2007) 2642–2650. [24] D. Zhang, O. Neumann, H. Wang, V.M. Yuwono, A. Barhoumi, M. Perham, [49] G.F. Lichtfuss, W.J. Cheng, Y. Farsakoglu, G. Paukovics, R. Rajasuriar, J.D. Hartgerink, P. Wittung-Stafshede, N.J. Halas, Gold nanoparticles can induce P. Velayudham, M. Kramski, A.C. Hearps, P.U. Cameron, S.R. Lewin, S.M. Crowe, the formation of protein-based aggregates at physiological pH, Nano Lett. 9 (2) A. Jaworowski, Virologically suppressed HIV patients show activation of NK cells (2009) 666–671. and persistent innate immune activation, J. Immunol. 189 (3) (2012) 1491–1499. [25] L. García-Fernández, J. Garcia-Pardo, O. Tort, I. Prior, M. Brust, E. Casals, [50] E.E. Sweeney, P.B. Balakrishnan, A.B. Powell, A. Bowen, I. Sarabia, R.A. Burga, J. Lorenzo, V.F. Puntes, Conserved effects and altered trafficking of Cetuximab R.B. Jones, A. Bosque, C.R.Y. Cruz, R. Fernandes, PLGA nanodepots co-en­ antibodies conjugated to gold nanoparticles with precise control of their capsulating prostratin and anti-CD25 enhance primary natural killer cell anti­ number and orientation, Nanoscale 9 (18) (2017) 6111–6121. viral and antitumor function, Nano Res. 13 (3) (2020) 736–744. [26] B. Saha, T.H. Evers, M.W. Prins, How antibody surface coverage on nanoparticles [51] R.B. Jones, S. Mueller, S. Kumari, V. Vrbanac, S. Genel, A.M. Tager, T.M. Allen, determines the activity and kinetics of antigen capturing for biosensing, Anal. B.D. Walker, D.J. Irvine, Antigen recognition-triggered drug delivery mediated by Chem. 86 (16) (2014) 8158–8166. nanocapsule-functionalized cytotoxic T-cells, Biomaterials 117 (2017) 44–53. [27] Q. Ong, Z. Luo, F. Stellacci, Characterization of ligand shell for mixed-ligand [52] D. Yeheskely-Hayon, L. Minai, L. Golan, E.J. Dann, D. Yelin, Optically induced cell coated gold nanoparticles, Acc. Chem. Res. 50 (8) (2017) 1911–1919. fusion using bispecific nanoparticles, Small 9 (22) (2013) 3771–3777. [28] J. Comenge, The role of PEG conformation in mixed layers: from protein corona [53] A. Wiernik, B. Foley, B. Zhang, M.R. Verneris, E. Warlick, M.K. Gleason, J.A. Ross, substrate to steric stabilization avoiding protein adsorption, Sci. Open Res. X. Luo, D.J. Weisdorf, B. Walcheck, D.A. Vallera, J.S. Miller, Targeting natural killer (2015). cells to acute myeloid leukemia in vitro with a CD16 x 33 bispecific killer cell [29] Y. Cesbron, C.P. Shaw, J.P. Birchall, P. Free, R. Lévy, Stripy nanoparticles revisited, engager and ADAM17 inhibition, Clin. Cancer Res. 19 (14) (2013) 3844–3855. Small 8 (24) (2012) 3714–3719 author reply 3720-6. [54] M.K. Gleason, M.R. Verneris, D.A. Todhunter, B. Zhang, V. McCullar, S.X. Zhou, [30] Q.K. Ong, F. Stellacci, Response to “Critical assessment of the evidence for striped A. Panoskaltsis-Mortari, L.M. Weiner, D.A. Vallera, J.S. Miller, Bispecific and nanoparticles”, PLoS One 10 (11) (2015) e0135594. trispecific killer cell engagers directly activate human NK cells through CD16 [31] L. Duchesne, G. Wells, D.G. Fernig, S.A. Harris, R. Lévy, Supramolecular domains signaling and induce cytotoxicity and cytokine production, Mol. Cancer Ther. 11 in mixed peptide self-assembled monolayers on gold nanoparticles, (12) (2012) 2674–2684. Chembiochem 9 (13) (2008) 2127–2134. [55] B. Joos, M. Fischer, H. Kuster, S.K. Pillai, J.K. Wong, J. Boni, B. Hirschel, R. Weber, [32] R. Bonomi, A. Cazzolaro, L.J. Prins, Assessment of the morphology of mixed SAMs A. Trkola, H.F. Gunthard, HIV rebounds from latently infected cells, rather than on Au nanoparticles using a fluorescent probe, Chem. Commun. 47 (1) (2011) from continuing low-level replication, Proc. Natl. Acad. Sci. U. S. A. 105 (43) 445–447. (2008) 16725–16730. [33] P. Ciaurriz, F. Fernández, E. Tellechea, J.F. Moran, A.C. Asensio, Comparison of four [56] C. Manickam, S.V. Shah, J. Nohara, G. Ferrari, R.K. Reeves, Monkeying around: functionalization methods of gold nanoparticles for enhancing the enzyme- using non-human primate models to study NK cell biology in HIV infections, linked immunosorbent assay (ELISA), Beilstein J. Nanotechnol. 8 (2017) 244–253. Front. Immunol. 10 (2019) 1124. 14 A. Astorga-Gamaza, M. Vitali, M.L. Borrajo et al. Nano Today 36 (2021) 101056 [57] T.A. Rasmussen, M. Tolstrup, C.R. Brinkmann, R. Olesen, C. Erikstrup, [67] A. Trkola, J. Matthews, C. Gordon, T. Ketas, J.P. Moore, A cell line-based neu­ A. Solomon, A. Winckelmann, S. Palmer, C. Dinarello, M. Buzon, M. Lichterfeld, tralization assay for primary human immunodeficiency virus type 1 isolates S.R. Lewin, L. Østergaard, O.S. Søgaard, Panobinostat, a histone deacetylase that use either the CCR5 or the CXCR4 coreceptor, J. Virol. 73 (11) (1999) inhibitor, for latent-virus reactivation in HIV-infected patients on suppressive 8966–8974. antiretroviral therapy: a phase 1/2, single group, clinical trial, Lancet HIV 1 (1) [68] K.A. Clouse, et al., Monokine regulation of human immunodeficiency virus-1 (2014) e13–e21. expression in a chronically infected human T cell clone, J. Immunol. 142 (2) [58] N.M. Archin, A.L. Liberty, A.D. Kashuba, S.K. Choudhary, J.D. Kuruc, A.M. Crooks, (1989) 431–438. D.C. Parker, E.M. Anderson, M.F. Kearney, M.C. Strain, D.D. Richman, [69] T.M. Folks, K.A. Clouse, J. Justement, A. Rabson, E. Duh, J.H. Kehrl, A.S. Fauci, M.G. Hudgens, R.J. Bosch, J.M. Coffin, J.J. Eron, D.J. Hazuda, D.M. Margolis, Tumor necrosis factor alpha induces expression of human immunodeficiency Administration of vorinostat disrupts HIV-1 latency in patients on antiretroviral virus in a chronically infected T-cell clone, Proc. Natl. Acad. Sci. U. S. A. 86 (7) therapy, Nature 487 (7408) (2012) 482–485. (1989) 2365–2368. [59] J.H. Elliott, F. Wightman, A. Solomon, K. Ghneim, J. Ahlers, M.J. Cameron, [70] J.P. Moore, M. Thali, B.A. Jameson, F. Vignaux, G.K. Lewis, S.W. Poon, M. Charles, M.Z. Smith, T. Spelman, J. McMahon, P. Velayudham, G. Brown, J. Roney, M.S. Fung, B. Sun, P.J. Durda, Immunochemical analysis of the gp120 surface J. Watson, M.H. Prince, J.F. Hoy, N. Chomont, R. Fromentin, F.A. Procopio, glycoprotein of human immunodeficiency virus type 1: probing the structure of J. Zeidan, S. Palmer, L. Odevall, R.W. Johnstone, B.P. Martin, E. Sinclair, S.G. Deeks, the C4 and V4 domains and the interaction of the C4 domain with the V3 loop, J. D.J. Hazuda, P.U. Cameron, R.P. Sékaly, S.R. Lewin, Activation of HIV transcription Virol. 67 (8) (1993) 4785–4796. with short-course vorinostat in HIV-infected patients on suppressive anti­ [71] R. Wyatt, J. Moore, M. Accola, E. Desjardin, J. Robinson, J. Sodroski, Involvement retroviral therapy, PLoS Pathog. 10 (10) (2014) e1004473. of the V1/V2 variable loop structure in the exposure of human im­ [60] B.F. Haynes, P.B. Gilbert, M.J. McElrath, S. Zolla-Pazner, G.D. Tomaras, S.M. Alam, munodeficiency virus type 1 gp120 epitopes induced by receptor binding, J. D.T. Evans, D.C. Montefiori, C. Karnasuta, R. Sutthent, H.X. Liao, A.L. DeVico, Virol. 69 (9) (1995) 5723–5733. G.K. Lewis, C. Williams, A. Pinter, Y. Fong, H. Janes, A. DeCamp, Y. Huang, M. Rao, [72] N.G. Bastus, J. Comenge, V. Puntes, Kinetically controlled seeded growth synth­ E. Billings, N. Karasavvas, M.L. Robb, V. Ngauy, M.S. de Souza, R. Paris, G. Ferrari, esis of citrate-stabilized gold nanoparticles of up to 200 nm: size focusing versus R.T. Bailer, K.A. Soderberg, C. Andrews, P.W. Berman, N. Frahm, S.C. De Rosa, Ostwald ripening, Langmuir 27 (17) (2011) 11098–11105. M.D. Alpert, N.L. Yates, X. Shen, R.A. Koup, P. Pitisuttithum, J. Kaewkungwal, [73] H.P. Erickson, Size and shape of protein molecules at the nanometer level de­ S. Nitayaphan, S. Rerks-Ngarm, N.L. Michael, J.H. Kim, Immune-correlates ana­ termined by sedimentation, gel filtration, and electron microscopy, Biol. Proced. lysis of an HIV-1 vaccine efficacy trial, N. Engl. J. Med. 366 (14) (2012) Online 11 (2009) 32–51. 1275–1286. [74] L. Monticelli, S.K. Kandasamy, X. Periole, R.G. Larson, D.P. Tieleman, S.J. Marrink, [61] R. Olesen, S. Vigano, T.A. Rasmussen, O.S. Søgaard, Z. Ouyang, M. Buzon, The MARTINI coarse-grained force field: extension to proteins, J Chem Theory A. Bashirova, M. Carrington, S. Palmer, C.R. Brinkmann, X.G. Yu, L. Østergaard, Comput 4 (5) (2008) 819–834. M. Tolstrup, M. Lichterfeld, Innate immune activity correlates with CD4 T cell- [75] S.F. Afshar Y, A. Pishevar, S. Worley, Exploiting seeding of random number associated HIV-1 DNA decline during latency-reversing treatment with pano­ generators for efficient domain decomposition parallelization of dissipative binostat, J. Virol. 89 (20) (2015) 10176–10189. particle dynamics, Comput. Phys. Commun. 184 (4) (2013) 1119–1128. [62] J.P. Moore, F.E. McCutchan, S.W. Poon, J. Mascola, J. Liu, Y. Cao, D.D. Ho, [76] M.Aa.J. A, Dissipative particle dynamics: effects of parameterization and ther­ Exploration of antigenic variation in gp120 from clades A through F of human mostating schemes on rheology, Soft Mater. 13 (2) (2015) 106–117. immunodeficiency virus type 1 by using monoclonal antibodies, J. Virol. 68 (12) [77] A. Franconetti, J.M. Carnerero, R. Prado-Gotor, F. Cabrera-Escribano, C. Jaime, (1994) 8350–8364. Chitosan as a capping agent: Insights on the stabilization of gold nanoparticles, [63] W.D. Tolbert, N. Gohain, M. Veillette, J.P. Chapleau, C. Orlandi, M.L. Visciano, Carbohydr. Polym. 207 (2019) 806–814. M. Ebadi, A.L. DeVico, T.R. Fouts, A. Finzi, G.K. Lewis, M. Pazgier, Paring down HIV [78] A. Moreno, J.A, R. Grillo, L.F. Fraceto, C. Jaime, A study on the molecular existing Env: design and crystal structure of a stabilized inner domain of HIV-1 gp120 interactions in nanoherbicides: a chitooligosaccharide/tripolyphosphate kloaded displaying a major ADCC target of the A32 region, Structure 24 (5) (2016) with paraquat case, Colloids Surf. A 562 (2018) 220–228. 697–709. [79] F.J.M. De Meyer, R.J.M, T.F. Willems, B. Smit, Molecular simulation of the effect of [64] G.K. Lewis, M. Pazgier, A.L. DeVico, Survivors remorse: antibody-mediated pro­ cholesterol on lipid-mediated protein-protein interactions, Biophys. J. 99 (11) tection against HIV-1, Immunol. Rev. 275 (1) (2017) 271–284. (2010) 3629–3638. [65] K. Deng, M. Pertea, A. Rongvaux, L. Wang, C.M. Durand, G. Ghiaur, J. Lai, [80] M. Venturoli, S.B.a.S, M. M, Simulation studies of protein-induced bilayer de­ H.L. McHugh, H. Hao, H. Zhang, J.B. Margolick, C. Gurer, A.J. Murphy, formations, and lipid-induced protein tilting, on a mesoscopic model for lipid D.M. Valenzuela, G.D. Yancopoulos, S.G. Deeks, T. Strowig, P. Kumar, J.D. Siliciano, bilayers with embedded proteins, Biophys. J. 88 (3) (2005) 1778–1798. S.L. Salzberg, R.A. Flavell, L. Shan, R.F. Siliciano, Broad CTL response is required to [81] V.R. Gómez-Román, R.H. Florese, L.J. Patterson, B. Peng, D. Venzon, K. Aldrich, clear latent HIV-1 due to dominance of escape mutations, Nature 517 (7534) M. Robert-Guroff, A simplified method for the rapid fluorometric assessment of (2015) 381–385. antibody-dependent cell-mediated cytotoxicity, J. Immunol. Methods 308 (1–2) [66] D.N. Howell, et al., Natural killing target antigens as inducers of interferon: (2006) 53–67. studies with an immunoselected, natural killing-resistant human T lympho­ [82] J.C. Grivel, L. Margolis, Use of human tissue explants to study human infectious blastoid cell line, J. Immunol. 134 (2) (1985) 971–976. agents, Nat. Protoc. 4 (2) (2009) 256–269. 15