Susceptibility to Neutralization by Broadly Neutralizing Antibodies Correlates with Infected Cell Binding for a Panel of Clade B HIV Reactivated from Latent Reservoirs

Efforts to HIV cure are obstructed by reservoirs of latently infected CD4+ T-cells that can re-establish viremia. Broadly neutralizing HIV-specific antibodies (bNAbs), defined by unusually high neutralization breadths against globally diverse viruses, may contribute to the elimination of these reservoirs by binding to reactivated cells, targeting them for immune clearance. However, the relationship between neutralization of reservoir isolates and binding to corresponding infected primary CD4+ T-cells has not been determined. Thus, the extent to which neutralization breadths and potencies can be used to infer the corresponding parameters of infected-cell binding is currently unknown. We assessed the breadths and potencies of bNAbs against 36 viruses reactivated from peripheral blood CD4+ T-cells of ARV-treated HIV-infected individuals, using paired neutralization and infected-cell binding assays. Single antibody breadths ranged from 0–64% for neutralization (IC80≤10μg/ml) and 0–89% for binding, with two-antibody combinations reaching 0-83% and 50-100%, respectively. Infected-cell binding correlated with virus neutralization for 10 out of 14 antibodies (e.g. 3BNC117, r=0.87, p<0.0001). Heterogeneity was observed, however, with a lack of significant correlations for 2G12, CAP256.VRC26.25, 2F5, and 4E10. Our results provide guidance on the selection of bNAbs for interventional cure studies; both by providing a direct assessment of intra-and inter-individual variability in neutralization and infected cell binding in a novel cohort, and by defining the relationships between these parameters for a panel of bNAbs. Importance Although anti-retroviral therapies have improved the lives of people who are living with HIV, they do not cure infection. Efforts are being directed towards harnessing the immune system to eliminate the virus that persists, potentially resulting in virus-free remission without medication. HIV-specific antibodies hold promise for such therapies owing to their abilities to both prevent the infection of new cells (neutralization), and also to direct the killing of infected cells. We isolated 36 HIV strains from individuals whose virus was suppressed by medication, and tested 14 different antibodies for neutralization of these viruses and for binding to cells infected with the same viruses (critical for engaging natural killer cells). For both neutralization and infected-cell binding, we observed variation both between individuals, and amongst different viruses within an individual. For most antibodies, neutralization activity correlated with infected cell binding. These data provide guidance on the selection of antibodies for clinical trials.

assessed the breadths and potencies of bNAbs against 36 viruses reactivated 48 from peripheral blood CD4 + T-cells of ARV-treated HIV-infected individuals, using 49 paired neutralization and infected-cell binding assays. Single antibody breadths 50 ranged from 0-64% for neutralization (IC80≤10μg/ml) and 0-89% for binding, with 51 two-antibody combinations reaching 0-83% and 50-100%, respectively. Infected-52 cell binding correlated with virus neutralization for 10 out of 14 antibodies (e.g. 53 3BNC117, r=0.87, p<0.0001). Heterogeneity was observed, however, with a lack 54 of significant correlations for 2G12, CAP256.VRC26.25, 2F5, and 4E10. Our 55 results provide guidance on the selection of bNAbs for interventional cure 56 studies; both by providing a direct assessment of intra-and inter-individual 57 variability in neutralization and infected cell binding in a novel cohort, and by 58 defining the relationships between these parameters for a panel of bNAbs. in long-lived resting memory CD4 + T-cells (1-3). Evidence from in vitro latency 83 models supports that these reservoirs can be eliminated by combining latency 84 reversal agents (LRAs), which induce the expression of viral antigens, with 85 enhanced immune effectors; a paradigm referred to as "kick and kill" or, 86 alternatively, as "shock and kill" (4-6). One strategy to harness immune effectors 87 for these strategies is to target reactivated infected cells with HIV-specific 88 antibodies, resulting in the engagement of natural killer (NK) cells, monocytes, 89 and granulocytes which eliminate infected cells through antibody-dependent cell-90 mediated cytotoxicity (ADCC) and/or antibody-dependent cell-mediated 91 phagocytosis (ADCP) (7-9). For this purpose, it will be crucial for the HIV-specific 92 antibodies to bind to Env protein expressed on the surface of the reactivated 93 latent infected cells. The current study focuses on correlating the susceptibility of 94 neutralization against viral isolates reactivated from patient CD4 + T-cells by a 95 panel of HIV-specific broadly neutralizing antibodies (bNAbs) with their capacity 96 to bind to Env expressed by the reactivated latent infected cells, thereby 97 providing guidance on the selection of bNAbs to optimally support the clinical 98 translation of kick and kill strategies. 99 bNAbs tested in study all share the same IgG1 Fc domain, differing only in their 146 Fab fragments. The current study thus focuses on providing guidance with 147 respect to the selection of the antigen binding Fab fragments of Abs for use in 148 cure strategies. To maximize potency, these Fab fragments may ultimately need 149 to be combined with Fc domains that are designed to maximally engage ADCC 150 effectors (39). reported a lack of infected-cell binding and ADCC activity by 3BNC117 against a 160 multi-clade panel of HIV (40), suggesting a lack of correspondence with its 161 breadth of neutralizing activity (15). Although this relationship has been explored 162 indirectly, to our knowledge, only one study has directly compared infected cell 163 binding or ADCC of bNAbs versus neutralizing activity across different viral 164 isolates. This study showed a correlation between these functions, but was 165 limited to the use of two viral isolates of HIV (NL4-3 and JR-FL) and SHIV . We therefore perceived a need to define the relationship between 167 neutralization and infected cell binding of clinically relevant bNAbs to HIV 168 produced by reactivated latent infected CD4 + T cells. To test the ability of bNAbs to neutralize reservoir virus, we obtained a 189 panel of 14 bNAbs that are currently being developed for clinical use in humans 190 and categorized these by their targeted epitope (See Methods). We measured 191 the neutralizing activities of these bNAbs against 36 viral isolates that had been 192 reactivated from the latent reservoirs of 8 individuals from limiting dilution 193 quantitative viral outgrowth assays (QVOA) (Fig 1 and 2A).  specific bNAbs PGT121 and 10-1074 and the V1V2-specific bNAb PG9 exhibited 195 potent but relatively narrow activity, exhibiting detectable neutralization (IC50 < 50 196 g/ml) of 53 -69% of viruses, with geometric mean IC50 values ranging from 0.3 197 -0.6 g/ml (Fig 2B). In contrast, the CD4 binding site (CD4bs)-specific 198 antibodies VRC01, VRC07-523, N6 and 3BNC117, as well as the  targeting antibody 10E8 exhibited broad activity, with a detectable neutralization 200 77 -100% of viruses (IC50 < 50 g/ml), but with substantially higher IC50 values 201 (geometric mean IC50 between 2.1 -8.9 g/ml) (Fig 2B). These trends parallel 202 previous reports using pseudovirus assays, which also observed that CD4bs 203 antibodies and 10E8 were generally much broader but less potent than V3-204 glycan and V1V2 apex antibodies (42,43). In the current experiment, 205 CAP256.VRC26.25 only neutralized 9 of 36 reactivated reservoir viruses (26%) 206 with a detectable IC50 (IC50 < 50 g/ml) (Fig 2B). Because CAP256.VRC26.25 207 has been reported to preferentially neutralize subtype C, and the QVOA viral 208 isolates tested here are all subtype B (Table 1), the low neutralization breadth we 209 observed is compatible with published data (22). 4E10 and 2F5 are known to be 210 less broad and potent than more recently published antibodies, so their lack of 211 breadth against these viruses is expected. One exception to the general 212 agreement between our data and those from published pseudovirus panels was 213 for 2G12 which, although not broadly neutralizing against genetically diverse 214 viruses, has been shown to potently neutralizes subtype B viruses in published 215 pseudovirus panels (19,44), but we observed only weak neutralization in our 216 assays, with only two viruses reaching 80% neutralization (Fig 2). 217 218 We frequently observed high degrees of similarity in neutralization 219 sensitivities within an individual's viral quasispecies, consistent with genetic 220 relatedness. For example, the five viral isolates from CIRC1096 were all sensitive 221 to neutralization by CD4bs and MPER antibodies, but resistant to V3-glycan and 222 V1V2 antibodies (Fig 2B & C). Exceptions to this, however, were not uncommon. 223 For example, for the four QVOA viruses from OM5346, two of these viruses (#2 224 and #4) were highly sensitive to V1V2 antibodies (PG9, CAP256-VRC26.25, 225 PGDM1400) and resistant to V3-glycan antibodies (PGT121, 10-1074 and 2G12) 226 whereas virus #3 exhibited the opposite sensitivity profile (Fig 2B & C) Fig 2C). 266

267
In order to establish breadth, we defined binding as a MFI ratio > 2. In 268 general, with the exception of VRC01, CD4bs Abs exhibited superior breadths of 269 infected-cell binding, covering 83 -89% of reservoir isolates when tested at the 270 neutralization IC80 concentrations (Fig 3C & D). The binding potencies of CD4bs 271 were relatively modest, however, with most exhibiting MFI ratios of between 2 -4 272 (Fig 3B & C). The V3-Glycan antibodies PGT121, 2G12, and 10-074 exhibited 273 more limited breadths as compared to CD4bs antibodies, but showed 274 substantially higher levels of specific binding to cells infected with susceptible 275 viruses, with many MFI ratios exceeding 5. Sensitivity/resistance profiles were 276 generally related for different viral isolates from the same individual, e.g. 10-1074 277 bound strongly to all isolates from 5/8 participants (Fig 3B), but exhibited a lack 278 of binding to all viruses from CIRC0196 (at both concentrations). Intra-patient 279 variability was observed, however, for example with 1 out of 5 viruses from 280 OM5162 exhibiting high sensitivity to 10-1074 and the remaining 4 exhibiting 281 resistance. With the exception of CAP256.VRC26.25 [which is predominately 282 clade C specific (22)], the V1/V2 bNAbs showed potent binding activity, 283 particularly in the case of PG9 which, at IC80 concentration, showed high levels of 284 specific binding to 16 of 36 reservoir viruses with an MFI ratio greater than 4 (Fig  285   3C). Infected cell binding of MPER-specific antibodies varied: 10E8v4-V5R-286 100cF (a version of 10E8 optimized for increased solubility and potency(45)) at 5 287 g/ml, bound to 30 of 36 isolates, with high-level binding observed for 13 of these 288 (MFI ratios > 4). However, 10E8 and 10E8v4-V5R-100cF also showed 289 substantial binding to uninfected bystanders (Gagpopulation) (see Fig 3A,  in binding to V3-glycan-specific bNAbs too, as shown PGDM1400 and PG9 298 bound robustly to viruses #1 and #3 (MFI ratio > 6), while no binding was 299 observed for viruses #2 and #4 (Fig 3B & C). Our data indicate both intra-and 300 inter-individual variability in binding to cells infected with reservoir viral isolates, 301 highlighting the limitations of using any single antibody in a therapeutic. 302 303 Achieving broad coverage of viral reservoir isolates in a population is likely 304 to require combinations of at least two bNAbs. To assess this in the current 305 population, we calculated the binding coverage of all possible two antibody 306 combinations using the binding data obtained with the neutralization IC80 307 antibody concentration (MFI ratio > 2) (Fig 3D). All CD4bs (excluding VRC01) 308 antibodies, when combined with 2G12 or V1/V2 antibodies or MPER antibodies 309 (except for 4E10), reached ≥ 92% coverage. Notably, the combinations of 2G12 310 with VRC07-523 or N6, or 10E8 or 10E8v4-V5R-100cF reached 100% coverage, 311 however, as previously mentioned, 10E8v4-V5R-100cF showed a high level of 312 bystander binding in our in vitro assays. 3BNC117 + 2G12 and VRC07-523 + 313 PG9 reached 97% coverage, thus representing promising combinations for 314 targeting reactivated clade B reservoir viruses (Fig 3D). 315

316
With respect to the effects of the different concentrations of antibodies 317 tested on binding, 10E8v4-V5R-100cF exhibited generally more favorable 318 binding profiles (MFI ratios) at 5 g/ml, due to a reduction in the background 319 binding that was observed at its IC80 concentration of 9.3 g/ml. In contrast, 10-320 1074 showed a lack of background binding even at 5 g/ml, and thus displayed 321 favorable binding profiles at this higher concentration, compared to its IC80 322 concentration at 0.7g/ml (Fig 3B & D). 323 324

Infected Cell Binding Correlates with Elimination by ADCC 325 326
Our primary interest in assessing infected-cell binding is to predict the ability of a 327 bNAb to direct ADCC against these cells. Infected cell binding is a prerequisite 328 for ADCC, and multiple studies have indicated that, where antibody Fc domains 329 are matched (as all bNAbs tested here share the same IgG1), levels of binding 330 correlate with ADCC activity (8,(32)(33)(34). To confirm this relationship under our 331 experimental conditions, we performed paired infected cell binding and ADCC 332 assays using two reservoir isolates (OM5334#7 and OM5162#1) in combination 333 with 9 bNAbs. Two types of NK cells were tested in parallel as effectors: i) haNK 334 cells (NantKwest)a derivative of the NK-92 cell line (46) that has been 335 enhanced for ADCC by expressing high affinity (ha) huCD16 V158 FcγRIIIa 336 receptor, as well as engineered to express IL-2 (47) ii) Freshly isolated NK cells 337 from the peripheral blood of an HIV-uninfected donor. Binding assays were 338 performed in parallel with ADCC assays using the same conditions -10g/ml 339 over a total of 7 hours at 37C. For both haNK cells and primary NK cells, we 340 observed moderate levels of NK-cell mediated elimination of HIV-infected cells in 341 the absence of bNAbs, likely due in part to HIV-mediated downregulation of HLA 342 molecules "missing self" (Fig 3E,F) (48,49). As expected, we observed 343 additional elimination of infected cells with the addition of bNAbs, and significant 344 direct correlations between total levels of elimination of HIV-infected cells (haNK, 345 r = 0.69, p < 0.001; primary NK cells, r = 0.65, p < 0.001), as well as ADCC-346 specific elimination of infected cells (% killing in +bNAb conditions -% killing in -347 bNAb conditions) (haNK, r = 0.73, p < 0.0001; primary NK cells, r = 0.65, p < 348 0.001) (Fig 3E, F) The infection of a cell by HIV results in the progressive, and almost 355 complete, loss of surface CD4 expression, through the concerted actions of Nef,356 Vpu,. Thus, in short-term in vitro infections of activated CD4 + T- We assessed whether differential binding of bNAbs to early (Gag + CD4 + ) 368 versus late (Gag + CD4 -) infected cells was present in our assays. Upon gating on 369 viable HIV-infected cells (lymphocytes, live cells, CD3 + , HIV-Gag + ), we observed 370 that some bNAbs, such as 3BNC117, 10-1074, and PG9 showing preferential 371 binding to late infected cells (CD4 -) (Fig 4A), while others, such as 10E8v4-V5R-372 100cF, showing similar, or slightly higher binding to early versus late populations 373 (Fig 4A). To test this systematically, we selected virus/bNAb combinations that 374 showed specific binding (Gag + / Gag -bNAb MFI ratio > 2 when tested at 375 neutralization IC80 concentrations) and compared levels of bNAb binding in the 376 Gag + CD4 + versus Gag + CD4populations. We calculated fold differences 377 between these early and late infected populations = (Geometric Mean MFI ratio 378 of Gag + CD4 -) / (Geometric Mean MFI ratio of Gag + CD4 + ). We observed that all 379 gp120-specific bNAbs exhibited higher levels of binding to late-infected 380 populations than to matched early-infected populations (Fold differences: 1.7 -381 3.9, Fig 4B). In contrast, each of the gp41-specific bNAbs exhibited similar or 382 slightly higher levels of binding to early-versus late-infected populations (Fold 383 differences: 0.90 -0.97, Fig 4B). A mechanistic explanation for this discrepancy 384 is beyond the scope of the current manuscript. However, we raise the possibility 385 that it may be related to the in cis interactions that have been shown to occur on 386 early infected cells between gp120 and CD4 on the same cell surface (54). 387 Binding of CD4 to functional trimers can induce gp120 shedding from Env 388 trimers, and enhance exposure of the gp41 membrane proximal external region 389 to antibody binding (55). Our data are consistent with such conformational 390 differences in Env favoring gp41-specific antibody binding to early-infected cells. 391

392
One implication of these results is that the binding data presented in Fig 3  393 -which was generated based on total Gag + cellsover-represents binding to 394 early-infected cells for gp120-specific antibodies, and under-represents binding 395 to late-infected cells. Data calculated based only on the late-infected populations 396 show a substantially intensified binding profile for most of the bNAbs used in this 397 studymost notably for the CD4bs bNAbs and PG9 (binding @ geographic 398 mean IC80 concentration, Supplementary Fig 1). A second implication is that 399 cellular infection dynamics may impact the ability to detect relationships between 400 infected-cell binding and virus neutralization. For example, if virus 1 replicated 401 with faster kinetics than virus 2, and thus had a greater proportion of Gag + CD4 -402 versus Gag + CD4 + cells, then this would skew bNAb binding profiles in a way that 403 was not intrinsic to the Env itself. To account for this, we have assessed these 404 relationships based on both total Gag + cells and on only the Gag + CD4late 405 infected populations (below). is an important pre-requisite for ADCC (8,(32)(33)(34). Efforts to harness bNAbs to 413 direct ADCC against infected cells would therefore benefit from an understanding 414 of the degree to which infected cell binding can be inferred from neutralizing 415 activity against a given virus. Our paired binding and neutralization data sets 416 allowed us to assess this using a number of analytic approaches in regards to 417 both concentrations of bNAbs used for binding assays and to stage of infection of 418 target cells. With respect to bNAb concentrations, binding to infected cells was 419 assessed for each bNAb at 5 μg/ml, and at the geometric mean IC80 420 neutralization concentration of that antibody against the same panel of reservoir 421 viruses. For the latter, this meant that some antibodies were tested at > 5 μg/ml 422 (ex. 4E10 at 49.2 μg/ml), while other antibodies were tested at substantially lower 423 concentrations (e.g. PGT121 at 0.6 μg/ml) (Geo Mean IC80 concentrations are 424 given below the heat-map in Fig 2C). This approach thus seeks to normalize for 425 intrinsic differences in avidity between different bNAbs. With respect to stage of 426 infection of target cells, we separately tested for correlations between 427 neutralization IC80 and binding to either total infected cells (Gag + ) or to late-428 infected cells (Gag + CD4 -), based on the differential binding patterns described 429 above. Of these, the most appropriate method for assessing the relationship 430 between binding and neutralization likely depends on the question being asked. 431 Importantly, however, the relationships that we observed, as described below, 432 turned out to be conserved across these different approaches. 433

434
We first tested for correlations between neutralization IC80 and the level of 435 binding (MFI ratio) at 5 μg/ml bNAb concentrations. As is described above, since 436 cells in early versus late stages of HIV infection exhibit differential bNAb binding 437 profiles, replication dynamics have the potential to impact overall assessments of 438 binding. In order to increase our ability to discern Env-intrinsic relationships 439 between binding and neutralization we therefore limited this initial analysis to the 440 late-infected (Gag + CD4 -) population. When all antibodies were considered 441 together, we observed a significant, direct correlation between virus 442 neutralization and infected cell binding (p < 0.0001, Spearman's r = 0.63) (Fig  443  5A). For each of the bNAbs that showed appreciable neutralizing activity 444 (VRC01, VRC07, 3BNC117, N6, PGT121, 10-1074, PGDM1400, PG9, 10E8, 445 and 10E8v4-V5R-100cF) we observed significant direct correlations between 446 neutralizing activity and infected-cell binding (Fig 5B). The antibodies 2F5 and 447 CAP256.VRC26.25 showed little in the way of either neutralization or binding, 448 precluding the possibility of detecting a relationship between these factors. 2G12 449 and, to lesser extent, 4E10 were notable outliers as they showed appreciable 450 binding capacity to many of the viruses in this panel, but very little corresponding 451 neutralizing activity. This lack of potent neutralization activity is inconsistent with 452 data from pseudovirus assays, but in agreement with previous data using virus 453 produced from T-cells, suggesting that 2G12 sensitivity is particularly tied to the 454 source of virus (56)(57)(58). bNAbs, e.g. for 3BNC117: Spearman's r = 0.82 for 5 μg/ml total Gag + 465 ( Supplementary Fig 2) vs Spearman's r = 0.60 for IC80 concentration total Gag + 466 ( Supplementary Fig 4); for PGT121: Spearman's r = 0.47 for 5 μg/ml total Gag + 467 ( Supplementary Fig 2) vs Spearman's r = 0.71 for IC80 concentration total 468 Gag + (Supplementary Fig 4). Overall, however, each of the antibodies that 469 exhibited a significant correlation by one analytic approach also exhibited 470 significant correlations by the other three approaches, and vice versa for those 471 lacking significant correlations. Thus, for 10 out of 14 bNAbs tested in this study, 472 the ability of a bNAb to neutralize a given virus is strongly correlated with its 473 ability to bind to a corresponding infected cell. In these in vitro assays, this 474 correlation was robust enough to be observed with or without controlling for 475 avidity of a given bNAb or for infection dynamics. The primary conclusion of the current study is that the ability of a given 480 bNAb to neutralize clinical viral isolates is a strong correlate of its ability to bind to 481 cell-surface Env on primary CD4 + T-cells infected with the same virus. 482 Furthermore, in comparing across a large panel of bNAbs, relative levels of 483 infected-cell binding and virus neutralization continued to correlatefor example, 484 10-1074 showed both high-level infected-cell binding and potent neutralization 485 compared to VRC01. Thus, we conclude thatwith respect to the Fab 486 component of Abs, when sharing the same Fcthe selection of Abs based on 487 broad and potent neutralizing activity is very likely to also select for those that are 488 suitable for infected-cell clearance. Of note, the reciprocal was not always true; 489 with 2G12 exhibiting reasonably potent and broad infected-cell binding, 490 contrasted by a general lack of neutralization of these reservoir-derived primary 491 isolate viruses. Though less strikingly, the MPER-specific bNAbs 2F5 and 4E10 492 also exhibited appreciable infected-cell binding (similar in breadths and 493 magnitudes to VRC01), but with minimal neutralizing activity. We propose that 494 the differences based on the directionality of this relationship may be related to 495 the differential antigen conformational requirements for these two functions. For a 496 bNAb to neutralize virus, it must bind functional Env trimers present on the 497 surface of cells producing infectious virus. In contrast, an antibody that also binds 498 to nonfunctional envelope proteins, such as gp41 stumps (59), may bind to 499 infected cells to a greater degree than they mediate neutralization (if they 500 neutralize at all). Thus, virus neutralization is a predictor of infected-cell binding, 501 but the reciprocal relationship does not hold. 502

503
While it may be intuitive that virus neutralization would correlate with 504 infected-cell binding, we do not feel that this could have been assumed to be the 505 case without experimental evidence. The conformation of Envs may be affected 506 by differences between the cell-surface vs virion environments, and this 507 variability could impact different viral isolates. For example, in cis interactions 508 between CD4 and Env on the surfaces of infected cells have been shown to 509 induce gp120 shedding, and expose gp41 stumps. This has been reported to 510 enhance infected-cell binding by gp41-specific Abs, while diminishing binding by 511 gp120-specific Abs (35). Such an effect might differentially impact different 512 virusesfor example, Horwitz et al. reported that the R456K mutation on YU2 513 gp120 decreased gp120 shedding, which led to less bystander (Gag -CD4 + ) 514 binding (60). Our data are consistent with these observations, and provide further 515 evidence of cis binding of CD4 modulating the binding of bNAbs to infected-cells. 516 We find gp120-specific bNAbs bind preferentially to cells in a late stage of 517 infection (CD4 low ) while gp41-specific bNAbs bind similarly or slightly better to 518 cells in an early stage of infection (CD4 high ). To address more mechanisms of 519 these findings, future studies may benefit from including gp120/gp41 interface 520 bNAbs, such as 8ANC195 (61), PGT151, PGT158 (62). However, despite any 521 such differences between the virion and cell-surface environments, the ability to 522 neutralize virus was significantly correlated with infected-cell binding, and these 523 relationships held whether we considered all infected cells (Gag + ) or only late 524 infected cells (Gag + CD4 -). 525

526
To investigate factors that may predict the efficacy of bNAb treatment to 527 contribute to HIV cure we felt it important to study the properties of bNAbs 528 against viruses derived from reactivated latent reservoirs. By combining a QVOA 529 approach with isolation of virus from dilutions of CD4 + T cells from different ART-530 suppressed patients where <50% of wells were p24 + , we were able to isolate 531 viruses that were likely clonal to test bNAb binding and neutralization profiles 532 (Fig 1) and assess both intra-and inter-patient variability. We observed a 533 considerable level of heterogeneity, even within a given individual, such that in 534 the majority of cases any single bNAb failed to provide universal coverage of an 535 individual's reservoir isolates. However, combinations of two antibodies provided 536 broad coverage both within and across individuals, reaching up to 100% 537 coverage as assessed by binding. Note that as our study population was derived 538 from a single site (Toronto, Canada), from a clinical perspective this assessment 539 of breadth is representative of what might be expected in a single-site study in a 540 North American clade B infected cohort. We propose that the method presented 541 here could be applied to different populations as a means of prioritizing antibody 542 combinations for a given regional population of patients and personalizing 543 individual HIV cure strategies as ART drug resistance is used to guide ART 544 therapy. Clinical use of the QVOA assay will likely be limited by its expense, cell 545 number requirements, and protracted timeline (14 days) for results. However, a 546 notable opportunity is present in the fact that infectious clonal autologous 547 reservoir viruses are generated as a byproduct of the primary measurement. The 548 pairing of quantitative and qualitative assessments of the HIV reservoir in this 549 way has been previously termed the Q 2 VOA (63). 550

551
The potencies of neutralization observed in the current study are overall 552 weaker than those that have been previously reported using pseudovirus assays 553 most notably for 2G12, which failed to achieve 80% neutralization for all but 554 two viruses. While this is likely due in part to our use of clinical viral isolates, 555 which are generally less sensitive to bNAbs than laboratory-adapted viruses (64, 556 65), we also note the role of virus producing cells in modulating sensitivity to 557 neutralization. Studies addressing this issue have reported that T-cell derived 558 virus is more resistant to neutralization than pseudovirus generated by 559 transfected 293T cells and, in particular, that replication competent virus 560 produced by PBMCs are more neutralization resistant than Env matched 561 pseudoviruses (56-58). However, there appear to be antibody-specific 562 differences in the level of influence that a producer cell has on sensitivity to 563 neutralization. For example, one study reported that PG9 is not very sensitive to 564 differences in producer cell (66), while large differences in IC50 have been 565 reported between T cell and pseudovirus for antibody 2G12 (56, 57). These data 566 suggest that producer cells differentially influence the conformations of Env on 567 resulting virions, as well as their densities and glycosylation, or numbers of 568 gp120 molecules in the viral membrane. As PG9 preferentially targets well-569 ordered, closed, trimeric viral spikes, it indicates that an equal number of well-570 folded spikes exists on virions produced by either cell type, whereas perhaps 571 bNAbs such as 2G12 can bind equally well to mis-folded trimers and are 572 therefore more sensitive to increases in the latter. Furthermore, the epitopes of 573 certain antibodies, such as 2G12, include glycans, and producer cells can affect 574 glycosylation patterns of gp120 (66). Thus, in addition to the comparison 575 between neutralization and infected-cell binding, the current study contributes a 576 reassessment of bNAb neutralization potency that may be more clinically 577 applicable than data from pseudovirus assays. 578 579 In conclusion, our study provides novel insights into the relationship 580 between infected-cell binding and virus neutralization that may help to guide 581 immunotherapeutic strategies aimed at either curing infection, or enabling 582 durable immune control of viral replication. The degree of intra-and inter-583 individual variation in bNAb sensitivity within even this geographically discrete 584 clade B population reinforces the importance of utilizing combinations of at least 585 two bNAbs in such therapies. Screening reactivated reservoir viruses for 586 sensitivity to bNAbs, either at an individual or population level, can help select 587 antibody combinations for optimal coveragefor example, with combinations of 588 PG9 and either 3BNC117 or N6 providing potent infected-cell binding coverage 589 of 94% and 72-78% coverage of neutralization (IC80 ≤ 10μg/ml) of viruses in the 590 current study population. For the bNAbs that exhibited correlations between 591 infected-cell binding and neutralization, our study indicates that screening for 592 either one of these factors is sufficient to infer that both functions will be present 593 against reactivated reservoir viruses. Consistent with previous studies, we also 594 confirmed that this infected cell bindingas measured by our assaycorrelated 595 well with NK cell mediated ADCC, suggesting that it is a reasonable surrogate. It 596 will be of interest, however, for future studies to build upon these results with 597 more extensive functional assays (potentially using varying Fc domains and/or 598 effector cells). Such future directions could potentially uncover more subtle 599 aspects of the relationship between virus neutralization and the targeting of cell-600 Review Boards. All subjects were adults, and gave written informed consent. 612 Clinical data for these participants are given in Table 1. Human CD4 T cells were enriched from the peripheral blood mononuclear cells 631 (PBMCs) (Stemcell Technologies), processed from leukapheresis, which were 632 drawn from long-term ARV-treated HIV-infected participants ( Table 1) SpectraMax Glo Steady-Luc reporter assay (Molecular Devices, LLC., CA) 672 reagent was added to the cells. After a 10-min incubation at room temperature to 673 allow cell lysis, the luminescence intensity was measured using a SpectraMax 674 i3x multi-mode detection platform per the manufacturers' instructions. 675 Neutralization curves were calculated comparing luciferase units to virus-only 676 control after background subtraction and fit by nonlinear regression using the 677 assymetric five-parameter logistic equation in GraphPad Prism (Fig 2A). The 678 50% and 80% inhibitory concentrations (IC50 and IC80, respectively) were defined 679 as the antibody dilution that caused a 50% and 80% reduction in neutralization. 680 681 bNAb binding assay 682 All binding assays were tested with the unconjugated bNAbs. CD4 + T cells 683 (which were all CD3 + ) were isolated with the Human CD4 T cell enrichment kit 684 (Stemcell Technologies) and activated with CD3/28 antibodies (Biolegend) for 48 685 hours. Supernatants collected from QVOA wells (p24 + , the same viruses with 686 neutralization assay) were used for infection by adding into the activated CD4 + T 687 cells, followed by spinnoculation for 1 hour and 6 days in culture with media 688 change every 3 days. Infection rate was checked on days 3 and 5 post infection.   Outgrowth Assays (QVOA) were performed using CD4 + T-cells from ARV-1080 suppressed study participants. Virus was isolated from HIV-p24 + wells at a 1081 dilution where < 50% of wells were positive. A portion of the supernatants from 1082 each of these wells was used directly to assess virus neutralization using a TZM-1083 bl assay. Another portion was used to infect activated primary CD4 + T-cells. 1084 Binding of bNAbs to these infected cells was assessed by flow cytometry, co-1085 staining with CD3, CD4 and HIV-Gag to identify infected cells. showing paired comparisons of MFI ratios between CD4 + and CD4populations. 1144 MFI ratio is defined as (MFI of bNAbs in Gag + CD4 + )/ (MFI of bNAbs in Gag -) 1145