Research ArticleHIV

Protection against a mixed SHIV challenge by a broadly neutralizing antibody cocktail

See allHide authors and affiliations

Science Translational Medicine  20 Sep 2017:
Vol. 9, Issue 408, eaao4235
DOI: 10.1126/scitranslmed.aao4235

Swatting away a swarm

Monoclonal antibodies are being tested as therapy for many infections, including HIV. However, increasing evidence suggests that therapy with a single agent may encourage resistant viral strains to emerge. Julg et al. tested how potent HIV neutralizing antibodies could combat challenge with a mix of SHIV strains in nonhuman primates. They found that monotherapy with anti-V3 PGT121 or anti-V2 PGDM1400 allowed for selection of resistant viruses and subsequent replication in the animals. Dual antibody administration conferred complete protection, confirming that combination therapy may be necessary for protective efficacy against HIV in people.


HIV-1 sequence diversity presents a major challenge for the clinical development of broadly neutralizing antibodies (bNAbs) for both therapy and prevention. Sequence variation in critical bNAb epitopes has been observed in most HIV-1–infected individuals and can lead to viral escape after bNAb monotherapy in humans. We show that viral sequence diversity can limit both the therapeutic and prophylactic efficacy of bNAbs in rhesus monkeys. We first demonstrate that monotherapy with the V3 glycan-dependent antibody 10-1074, but not PGT121, results in rapid selection of preexisting viral variants containing N332/S334 escape mutations and loss of therapeutic efficacy in simian-HIV (SHIV)–SF162P3–infected rhesus monkeys. We then show that the V3 glycan-dependent antibody PGT121 alone and the V2 glycan-dependent antibody PGDM1400 alone both fail to protect against a mixed challenge with SHIV-SF162P3 and SHIV-325c. In contrast, the combination of both bNAbs provides 100% protection against this mixed SHIV challenge. These data reveal that single bNAbs efficiently select resistant viruses from a diverse challenge swarm to establish infection, demonstrating the importance of bNAb cocktails for HIV-1 prevention.


Broadly neutralizing antibodies (bNAbs) against HIV-1 Env are currently being developed for both HIV-1 prevention and therapy (1, 2). However, HIV-1 diversity presents a major challenge for the clinical development of bNAbs. Recent studies have demonstrated that bNAb monotherapy in chronically HIV-1–infected individuals rapidly selects for resistant viral variants (37), suggesting that cocktails of bNAbs will be required for effective therapeutic strategies. However, the question of whether multiple bNAbs will be required to protect against acquisition of infection has not previously been studied in detail.

Previous studies in nonhuman primates that have demonstrated bNAb-mediated protection against simian-HIV (SHIV) challenge have typically used single challenge viruses (810). During sexual transmission of HIV-1 across mucosal barriers, however, exposure to a swarm of viral variants is common, although only a small number of founder viruses typically establish the infection (11, 12). These data suggest that bNAb combinations may be required for HIV-1 prevention strategies.

Here, we explored whether single bNAbs would select for resistant SHIV variants in both therapeutic and prophylactic experiments in rhesus monkeys. We also evaluated whether a bNAb cocktail would be required to protect against a mixed SHIV challenge. We selected the V3 glycan-specific antibody PGT121 and the V2 glycan-specific antibody PGDM1400 (13, 14) for this study because these bNAbs target different epitopes and are both currently being evaluated in clinical trials (NCT02960581 and NCT03205917).


Rapid viral escape from single bNAb therapy

It has been hypothesized that certain bNAbs, such as the V3 glycan-dependent antibody PGT121 (14), might have an intrinsically higher bar to escape than other bNAbs targeting this epitope (15) as a result of making multiple glycan contacts on the Env surface, a phenomenon termed “glycan promiscuity” (16). We therefore first compared the therapeutic efficacy of two related V3 glycan-dependent antibodies, PGT121 and 10-1074 (14, 17), in chronically SHIV-SF162P3–infected rhesus monkeys. Animals were infected with SHIV-SF162P3 about 7 months before a single infusion with PGT121 [10 mg/kg intravenously (iv)] (n = 4) or 10-1074 (10 mg/kg iv) (n = 2). Both antibodies had comparable neutralizing potency against SHIV-SF162P3 in vitro [IC50 (median inhibitory concentration) of 0.1 μg/ml for 10-1074 and IC50 of 0.1 μg/ml for PGT121] but resulted in markedly different therapeutic efficacy in vivo (Fig. 1). Consistent with our previous findings (15), PGT121 infusion resulted in up to a 3-log reduction of plasma viral loads to undetectable levels, which was followed by viral rebound when PGT121 titers declined to subtherapeutic levels. In contrast, 10-1074 infusion resulted in only a transient 1.5-log decline of plasma viral loads followed by rapid viral rebound to baseline levels by day 14, which is comparable to the reported efficacy of 10-1074 in SHIV-AD8–infected monkeys (18) and in HIV-1–infected humans (6).

Fig. 1. SHIV-SF162P3 escape from 10-1074 in rhesus monkeys.

Log plasma viral RNA copies per mililliter in chronically SHIV-SF162P3–infected rhesus monkeys after infusion with a single dose of 10-1074 (10 mg/kg) or PGT121 (10 mg/kg) on day 0. Detection limit is 50 copies/ml. mAb, monoclonal antibody.

Viral sequencing by single-genome amplification (SGA) after 10-1074 infusion demonstrated the emergence of Env mutations at critical contact sites (N332D/T/S/I/K and S334N/R), which eliminate the N-linked glycan at position 332 (Fig. 2). These mutations were present at detectable but low levels in the challenge stock and baseline plasma samples (Fig. 3A), suggesting that 10-1074 selected rare preexisting viral escape variants. In contrast, no mutations were observed in the PGT121-treated animals after rebound (fig. S1), consistent with our prior observations (15). Moreover, an N332A point mutation in SHIV-SF162P3 completely abrogated 10-1074 neutralization activity in vitro but only partially reduced PGT121 neutralization activity (Fig. 3B). Together, these data suggest that PGT121 has a higher bar to escape than 10-1074 for SHIV-SF162P3, presumably as a result of the ability of PGT121 to make multiple glycan contacts on HIV-1 Env (16), and thus a single N332/S334 point mutation in the context of SHIV-SF162P3 may be insufficient to escape from PGT121.

Fig. 2. Analysis of SHIV env sequences in 10-1074–treated animals.

Viral sequences by SGA of plasma viruses before and after 10-1074 administration. Amino acid residues that were detected at days 14 and 42 are highlighted in red.

Fig. 3. Frequencies of amino acids at critical PGT121 and 10-1074 contact sites in the SHIV-SF162P3 challenge stock.

(A) “O” indicates an Asn that is part of a PNGS (positions 295, 301, and 332). The critical N332 and S334 residues are circled in red. aa, amino acid. (B) PGT121 and 10-1074 neutralization potency of SHIV-SF162P3 (wild-type) and SHIV-SF162P3 containing an N332A mutation.

Complementary HIV-1 coverage by PGT121 and PGDM1400

PGT121 covers about 60 to 70% of global viruses (14), and thus we explored which bNAbs would complement PGT121 to improve global virus coverage (19). PGDM1400 is a V2 glycan-dependent bNAb with exceptional neutralizing potency and breadth (13). We evaluated the neutralizing activity of PGT121, PGDM1400, and the combination of PGT121 + PGDM1400 against a panel of 118 multiclade pseudoviruses in TZM-bl assays. PGDM1400 covered 77% of viruses in this panel with an IC80 of 0.57 μg/ml, and PGT121 covered 60% of viruses in this panel with an IC80 of 3.26 μg/ml (Fig. 4A). The combination of both bNAbs neutralized 97% of pseudoviruses with an IC80 of 0.03 μg/ml, indicating that the combination was substantially more potent than either bNAb alone (P = 1.0 × 10−6 and P = 4.2 × 10−11 comparing IC80 titers for the combination versus PGDM1400 alone and PGT121 alone, respectively, Wilcoxon rank sum tests) and also had greater breadth than either bNAb alone (P = 2.2 × 10−6 and P = 2.2 × 10−13 comparing breadth for the combination versus PGDM1400 alone and PGT121 alone, respectively, Fisher’s exact tests) (Fig. 4B). These data suggest a remarkable complementarity between PGT121 and PGDM1400, as viruses resistant to one of these antibodies are largely susceptible to the other antibody. The combination of PGT121 + PGDM1400 covered 97 to 100% of viruses from clades A, B, C, and CRF01 in this global panel (Fig. 4C). Moreover, the combination led to complete (>95%) neutralization of 86% of viruses in this panel, as compared with 59% for PGDM1400 and 53% for PGT121 (P = 9.3 × 10−6 and P = 5.1 × 10−8, respectively, Fisher’s exact tests) (fig. S2).

Fig. 4. Complementarity of neutralization profiles of PGDM1400 and PGT121 against HIV-1.

(A) IC80 titers for PGDM1400, PGT121, and the combination of PGT121 + PGDM1400 against a panel of 118 multiclade viruses. IC80 titers are shown as a heatmap with viruses represented in rows. Dark red indicates more potent neutralization, and light yellow indicates less potent neutralization. Blue indicates IC80 neutralization titers >50 μg/ml. The IC80 titers in the combination reflect the concentration of each bNAb. (B) IC80 breadth-potency plots for PGDM1400, PGT121, and the combination of PGT121 + PGDM1400. (C) IC80 titers for PGDM1400, PGT121, and the combination of PGT121 + PGDM1400 for pseudoviruses from clades A, B, C, and CRF01. Bold horizontal lines represent medians, and thin horizontal lines are 25th and 75th percentiles. The percentage of viruses with IC80 titers >50 μg/ml is shown on the top of each panel. P values reflect Fisher’s exact tests. thr, threshold; CRF01, circulating recombinant form 01.

In vivo protection against mixed SHIV challenge

The complementarity of PGT121 and PGDM1400 in vitro suggests the combination of these two bNAbs as an HIV-1 prevention strategy. We therefore evaluated the protective efficacy of PGDM1400 alone, PGT121 alone, or the combination of PGT121 + PGDM1400 in rhesus monkeys against a mixed SHIV challenge with the clade B SHIV-SF162P3 (20) and the clade C SHIV-325c (21). SHIV-SF162P3 was sensitive in vitro to PGT121 but resistant to PGDM1400, whereas SHIV-325c was sensitive in vitro to PGDM1400 but resistant to PGT121 (Table 1). Twenty rhesus monkeys (n = 5 per group) were infused with PGDM1400 (10 mg/kg iv), PGT121 (10 mg/kg iv), the combination of PGT121 (5 mg/kg iv) + PGDM1400 (5 mg/kg iv), or phosphate-buffered saline as control. Animals were challenged the following day by the intrarectal route with 500 TCID50 (median tissue culture infectious dose) of SHIV-SF162P3 + 500 TCID50 of SHIV-325c. Pharmacokinetics of the two bNAbs was similar by both idiotype-specific enzyme-linked immunosorbent assays (ELISAs) and virus-specific neutralizing antibody (NAb) assays (figs. S3 to S5).

Table 1. IC50 and IC80 neutralization titers (in μg/ml) of PGT121 and PGDM1400 against SHIV-SF162P3 and SHIV-325c.
View this table:

After challenge, all sham control animals became infected with peak plasma viral loads of 6.60 to 7.73 log RNA copies/ml on day 14 (Fig. 5). Similarly, 100% (five of five) of animals that received PGDM1400 alone became infected with peak plasma viral loads of 5.46 to 7.74 log RNA copies/ml, and 100% (five of five) of animals that received PGT121 alone also became infected but with lower peak plasma viral loads of 3.97 to 5.26 log RNA copies/ml. In contrast, 0% (zero of five) of animals that received the combination of PGT121 + PGDM1400 exhibited detectable plasma viral loads (>50 RNA copies/ml) at any point in time (Fig. 5), indicating that only the PGT121 + PGDM1400 combination protected against this mixed SHIV challenge (P = 0.008 comparing the combination group with either the PGDM1400 alone group or the PGT121 alone group, Fisher’s exact test).

Fig. 5. Protective efficacy of the combination of PGT121 + PGDM1400 against a mixed SHIV challenge in rhesus monkeys.

Five animals per group received an intravenous single dose of PGDM1400, PGT121, the combination of PGT121 + PGDM1400, or saline (Sham) before being rectally challenged with a high dose of both SHIV-SF162P3 and SHIV-325c. Log plasma viral RNA copies per milliliter after mixed challenge with SHIV-SF162P3 and SHIV-325c. Red line indicates median values. Detection limit is 50 copies/ml.

Viral sequencing by SGA at weeks 2 to 6 after challenge indicated that 100% of clones in the PGT121-treated animals were SHIV-325c, whereas 100% of clones in the PGDM1400-treated animals were SHIV-SF162P3 (Fig. 6 and Table 2), consistent with the resistance patterns of these viruses (Table 1). The sham controls were detectably infected with only SHIV-SF162P3 (Fig. 6 and Table 2), which has greater replication capacity than SHIV-325c. These findings were confirmed by strain-specific reverse transcription polymerase chain reaction (RT-PCR) assays (figs. S6 and S7), demonstrating that each bNAb alone efficiently selected resistant variants from the challenge swarm for the establishment of primary infection.

Fig. 6. Analysis of SHIV env sequences in breakthrough infections.

Maximum likelihood tree depicting SHIV env sequences by SGA of plasma viruses at weeks 2 to 6 after challenge. The sequences of the challenge stock SHIVs are highlighted in green.

Table 2. Number of SHIV strain–specific single-genome sequences in the animals that experienced breakthrough infection.
View this table:

Diversity of HIV-1 sequences in target bNAb epitopes

We next evaluated the variation in key PGDM1400 and PGT121 contact and signature positions in HIV-1–infected humans. Signature residues were resolved by a phylogenetically corrected strategy using large panels of M group or C clade viruses, and contact sites were found by structural modeling using structures of PGT145 and PGT122 complexed with HIV-1 Env and of PGDM1400 and PGT121 Fabs (fig. S8) (13, 2224). The variation of key signature residues at contact sites for PGDM1400 and PGT121 binding to HIV-1 Env from multiple clades was then defined (figs. S9 and S10). In these logo plots, O represents an Asn that occurs in a potential N-linked glycosylation site (PNGS).

We assessed sequence variability in these key PGDM1400 and PGT121 signature positions in 10 HIV-1–infected individuals who had been sampled longitudinally and for whom there were >150 full-length published env sequences in the Los Alamos Sequence Database. The key N332 PNGS for PGT121 exhibited variability in 8 of 10 individuals, including those with primarily sensitive virus, and the D325 site showed variability in 7 of 10 individuals (Fig. 7 and table S1). D325 is a key contact residue, and a D325A substitution can reduce neutralization sensitivity by ninefold and can completely abrogate sensitivity in conjunction with the loss of N332 glycosylation site (25). The key N160 PNGS for PGDM1400 was variable in only 3 of 10 individuals, but a critical R166 signature was either variable or lost in 5 of 10 individuals, and the key K/R169 positive charge was either variable or lost in 10 of 10 individuals (Fig. 7 and table S1) (22). These findings in a random sample of HIV-1–infected individuals suggest that diverse resistance profiles evolve over time in HIV-1–infected humans.

Fig. 7. Intrapatient variation in key contact signatures for PGDM1400 and PGT121.

Logo plots of viral sequence diversity in the key PGDM1400 and PGT121 contact sites (figs. S8 to S10) in 10 HIV-1–infected individuals for whom >150 full-length env sequences were available in the Los Alamos Sequence Database. The subject ID is indicated above each plot. The x axis indicates the amino acid position based on HXB2 numbering. The y axis indicates the probability of an amino acid at this location. O indicates an Asn that is part of a PNGS. Blue reflects sensitivity signatures, red reflects resistance signatures, and black reflects no statistically significant associated with antibody sensitivity.


Our data demonstrate that bNAbs can rapidly and specifically select for resistant viral variants within a diverse challenge swarm, resulting in efficient breakthrough infection with resistant SHIV strains. Both PGDM1400 alone and PGT121 alone failed to protect against challenge with a mixture of SHIV-SF162P3 and SHIV-325c in rhesus monkeys, whereas the combination of both bNAbs protected with 100% efficacy. These findings suggest that a combination of bNAbs will be required to provide optimal protection against acquisition of HIV-1 infection.

All the sham controls exhibited only SHIV-SF162P3 replication by both SGA and subtype-specific RT-PCR assays. The absence of detectable SHIV-325c replication in the sham controls suggests that SHIV-SF162P3 has higher replicative capacity than SHIV-325c in vivo, although we cannot exclude the possibility that a low level of SHIV-325c replication occurred in these animals below the level of detection (<50 copies/ml). In the presence of PGT121 selection pressure, however, robust infection with SHIV-325c occurred in 100% of animals. These findings are consistent with the observation that during primary HIV-1 infection, transmitted founder viruses with higher replication rates are selected from a mixed swarm (26). However, whether bNAb-mediated selection of resistant viral variants occurs entirely at the mucosal site of inoculation or at distal sites of early infection (27, 28) remains to be determined.

Multiple previous studies have shown that single bNAbs and bNAb combinations robustly protect against SHIV challenges in rhesus monkeys (810), presumably as a result of relatively limited viral sequence diversity in most SHIV challenge stocks. Our findings suggest that the increased viral diversity observed in HIV-1 swarms may compromise the efficacy of single bNAbs for prevention of acquisition of HIV-1 infection, as a result of a greater frequency of resistant viral variants and the remarkable efficiency of bNAbs to select resistant viruses within a swarm. In the plasma viruses of the 10 randomly selected individuals that we analyzed, variations in key PGDM1400 and PGT121 contact and signature positions were frequent. Moreover, bNAbs rapidly select minor resistant variants in chronically infected hosts, as shown here for 10-1074, as well as in prior studies in SHIV-infected rhesus monkeys (18) and HIV-1–infected humans (37), although certain bNAbs such as PGT121 may have a higher bar to escape than 10-1074 against a subset of viruses. PGT121 and PGDM1400, both alone and in combination, are currently being explored in clinical trials (NCT02960581 and NCT03205917). A recent study has similarly shown that a trispecific Env antibody can also protect against a mixed SHIV challenge (29).

A limitation of this study is that two SHIVs might not be representative of a diverse HIV-1 swarm. In addition, this study assessed only one bNAb combination, and thus, testing additional bNAb combinations against mixed SHIV challenges will be required to further define the generalizability of our findings. Together, these data demonstrate the importance of developing bNAb cocktails for both prevention and therapy of HIV-1 infection to cover viral diversity and resistant variants within diverse viral populations.


Animals and study design

The overall objective of this study was to investigate the protective efficacy of PGDM1400 and PGT121, alone or in combination, against a challenge with two SHIVs in rhesus monkeys. One additional study was performed to test the therapeutic activity of the antibodies PGT121 and 10-1074 in chronically SHIV-SF162P3–infected macaques. Animals were randomly allocated to groups, and virologic data were generated blinded. Placebo controls (saline) were used for the protection study. Five animals per group were used, similar to previous studies (10, 30, 31). The number of animals analyzed is stated in the figure legends. Primary data can be found in the figures. All animal studies were approved by the appropriate Institutional Animal Care and Use Committee. Monkeys were housed at Bioqual and Alpha Genesis.

Single SHIV treatment study. Six outbred, Indian-origin male and female rhesus monkeys (Macaca mulatta) were genotyped and selected as negative for the protective major histocompatibility complex (MHC) class I alleles Mamu-A*01, Mamu-B*08, and Mamu-B*17. Tripartite motif-containing protein 5 (TRIM5) polymorphisms were balanced equally among groups. Animals were chronically infected with SHIV-SF162P3 and received PGT121 (10 mg/kg iv) or 10-1074 (10 mg/kg iv).

Double SHIV challenge study. Twenty outbred, Indian-origin male and female rhesus monkeys (M. mulatta) were genotyped and selected as negative for the protective MHC class I alleles Mamu-A*01, Mamu-B*08, and Mamu-B*17. TRIM5 polymorphisms were balanced equally among groups. Animals received PGT121 (5 to 10 mg/kg iv), PGDM1400 (5 to 10 mg/kg iv), or both and were challenged intrarectally with 500 TCID50 of SHIV-SF162P3 + 500 TCID50 of SHIV-325c. Generation of these SHIV stocks has been previously described (20, 21, 32).


10-1074, PGT121, and PGDM1400 were produced, as previously described (14), and purified by Protein A affinity matrix (GE Healthcare). All the monoclonal antibody preparations were endotoxin-free.

Enzyme-linked immunosorbent assays

PGDM1400 and PGT121 antibody titers were measured by a quantitative ELISA, as previously described (14), using anti-idiotypic antibody-coated microtiter plates to capture the monoclonal antibodies followed by detection using a horseradish peroxidase–conjugated anti-human immunoglobulin G antibody.

TZM-bl neutralization assays

Neutralization of the two SHIV challenge stocks was evaluated in vitro by using TZM-bl target cells and a luciferase reporter assay as described (33).

Viral RNA assays

Plasma SHIV RNA levels were measured using a gag-targeted quantitative real-time RT-PCR assay, and tissue levels of SHIV RNA and DNA were measured using gag-targeted, nested quantitative hybrid real-time/digital RT-PCR and PCR assays, as previously described (15).

Strain-specific RT-PCR assays

Viral RNA was reverse-transcribed using the universal SHIV primer vl-EnvR (5′-CAATAATTGTCTGGCCTGTACCGTC-3′). Probes and primers were designed to target the specific regions of either SHIV-SF162P3 or SHIV-325c. Primers including SF162P3 vl-EnvF (5′-TTGGAGAATGCTACTAATACCAC-3′), vl-EnvR (5′-CCTATGCTTGTGGTGACGTT-3′), and probe (5′-ATGAACAGAGGAGAAATA-3′) linked to FAM and BHQ (Biosearch Technologies) were used for SHIV-SF162P3–specific viral load assay; primers including 325C vl-EnvF (5′-CGGGAAACATAACATGTAGATCAAA-3′), vl-EnvR (5′-TTCTCCCTTTTCTCCTCCATCA-3′), and probe (5′-ACAGGACTACTGGTGACAC-3′) linked to Cy5 (Integrated DNA Technologies) were used for SHIV-325c–specific viral load assay.

SHIV-SF162P3 deep sequencing

Env amplicon deep sequencing was performed using pairs of primers designed to target V3 and V2 regions of SHIV-SF162P3 Env. SHIV-SF162P3 RNA was isolated and reverse-transcribed by SF162P3 NGS-V3R (5′-TTCTGGGTCCCCTCCTGAGG-3′). Primers including NGS-V3F (5′-GTACAGCTGAAGGAATCTGTAG-3′) and NGS-V3R were used for V3 amplification; primers including NGS V2F (5′-GTAATTGGAAAGAGATGAACAGAGGAG-3′) and NGS V2R (5′-GGGCACAATAATGTATGGGAATTGG-3′) were used for V2 amplification. NEBNext DNA Library Prep Kit for Illumina (New England Biolabs) was used for V3 and V2 amplicon library preparation on the basis of the manufacturer’s protocols. Illumina MiSeq Sequencing 150–base pair paired-end (PE150) was performed by the Dana-Farber Cancer Institute Molecular Biology Core Facility.

Sequence analysis

Adaptor sequences and low-quality bases (quality scores < 30) were removed from Illumina sequences using BBDuk from the BBTools suite ( The following adaptor sequences were used: 5′-AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGATCT-3′ and its complement (the NEBnext “universal” primer), 5′-CAAGCAGAAGACGGCATACGAGAT-CACTGT-GTGACTGGAGTTCAGACGTGTGCTCTTCCGATC-3′ (for the V2 sequences) and 5′-CAAGCAGAAGACGGCATACGAGAT-TCAAGT-GTGACTGGAGTTCAGACGTGTGCTCTTCCGATC-3′ (for the V3 sequences) (two index primers), and their reverse complements (dash characters delimit the complements of the sequencing bar codes). Adaptor trimming used the parameters “ktrim=r k=27 mink=4 hdist=2 tpe tbo”; quality trimming used “qtrim=rl trimq=30.” Paired-end sequences were aligned against SF162P3 Env using Bowtie 2 (; version 2.1.0, parameters “--very-sensitive-local--reorder”), sorted and indexed using SAMtools, processed into paired-end FASTA files, and translated into amino acids using custom Perl code (wfischer{at} PCR primer–derived sequences were removed by visual inspection of translated sequences (V3-F, VQLKESV; V3-R, SGGDPE; V2-F, NWKEMNRG; V2-R, PIPIHYCA) using AliView (34). Sequences were removed from the data set if they were not complete throughout the regions of interest or if they contained stop codons. PNGSs were computed by matches to the recognition sequence N-x-[ST], where x represents any amino acid other than proline, and represented in alignments by the letter O. Finally, columns matching signature positions were extracted from the alignment on the basis of HXB2 numbering ( and the Protein Data Bank entry 5FYJ, chain G).

Single-genome amplification

SGA assays were performed essentially as described (11) except that the primers used here were designed to target conserved region of SHIV-SF162P3 and SHIV-325c. Briefly, viral RNA was isolated and reverse-transcribed to viral complementary DNA (cDNA) using SHIV SGA EnvR1 (5′-CAATAATTGTCTGGCCTGTACCGTC-3′). First-round PCR was carried out with Q5 High-Fidelity 2X Master Mix (New England Biolabs) together with primers SHIV SGA EnvF1 (5′-TATGGGGTACCTGTGTGGAA-3′) and SHIV SGA EnvR1. PCR conditions were programmed as follows: 1 cycle of 98°C for 30 s, 35 cycles of 98°C for 15 s, 55°C for 15 s, and 72°C for 55 s, followed by a final extension of 72°C for 2 min. One microliter of first-round PCR product was added to Q5 Master Mix with primers SHIV SGA EnvF2 (5′-GCCTGTGTACCCACAGAC-3′) and EnvR2 (5′-ATAGTGCTTCCTGCTGC-3′). PCR conditions were programmed as above but increased to 45 cycles for the second step. Amplicons from cDNA dilutions resulting in less than 30% positive were considered to result from amplification of a single cDNA amplification and were processed for sequencing. For each sample, 15 to 30 sequences were analyzed.

Env contact sites for PGDM1400

Because no structures exist for PGDM1400 in complex with Env trimers, a cryogenic electron microscopy (cryo-EM) structure of the related antibody PGT145 in complex with Env trimer (22) and structural modeling was used to find Env contact sites for PGDM1400. The heavy-chain complementarity-determining region 3 (HCDR3) from the crystal structure of the PGDM1400 Fab (13) to the PGT145 Fab was aligned using positions 88 to 108 and the “align” function from PyMOL. All Env amino acids and glycans that had any heavy atoms within 8.5 Å of PGDM1400 heavy atoms were isolated. The list below shows PGDM1400 HXB2 positions of contact sites (including glycans). Most of the contact sites are shared between PGDM1400 and PGT145, with 119, 129, 130, and 158 in contact with PGT145 alone and 200 and 314 in contact with PGDM1400 alone. When the PGDM1400 contact sites contained amino acids that were significantly associated with sensitivity or resistance with a q value of <0.2, they are shown below in bold, and these are the positions that were tracked globally and within subjects using published longitudinal sequencing data (Fig. 4). Positions 160 to 162 comprise a glycosylation site with the motif Nx[ST] that is essential for an Env to be sensitive to PGDM1400 neutralization. A positive charge at position 169 was also particularly highly significantly associated with PGDM1400 sensitivity. The other sites (161, 165, 166, and 167) had significant but more modest associations. PGDM1400 contact sites: 120, 121, 122, 123, 124, 125, 126, 127, 128, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 200, 309, 312, 313, 314, and 315.

Env contact sites for PGT121

Because no structures exist for PGT121 in complex with Env trimers, a cryo-EM structure of the related antibody PGT122 in complex with Env trimer (23) was used, and the contact surface of PGT121 on the SOSIP trimer was modeled. Specifically, HCDR3 from the crystal structure of the PGT121 Fab (24) to the PGT122 Fab was aligned using positions 88 to 101 and the align function from PyMOL. Then, all Env amino acids and glycans that had any heavy atoms within 8.5 Å of PGT121/PGT122 Fab heavy atoms were isolated. These contact sites are listed below. Env sequences in hypervariable region of V1 cannot be aligned, so standard signature analysis on this region was not performed, and these positions are shown in italics. The contact sites with significant PGT121 resistance signatures (q < 0.2) are shown below in bold. PGT121 contact sites: 133, 134, 135, 136, 137, 138, 138A, 138B, 138C, 138D, 138E, 151, 152, 154, 175, 296, 297, 298, 299, 300, 301, 302, 322, 322A, 323, 324, 325, 326, 327, 328, 329, 330, 331, 332, 385, 414, 415, 416, 417, 418, and 419. PGT121 glycan sites: 156, 301, 332, 392, 413, and 442.

Several strong signatures that were not directly in the contact surface were also included. The PNGS at 295 is a favored signature for PGT121, but it is not present in this particular structure. The PNGS at 334, which is often gained when the key signature site of the PNGS at N332 is lost, was also included. The signature sites were then resolved using a phylogenetically corrected strategy (35, 36) to identify amino acid in key positions that are associated with either resistance or sensitivity to PGT121 to large panels of 207 M group or 200 C clade viruses.

Statistical analyses

Analyses of independent data were performed by Wilcoxon rank sum tests, and comparisons of categorical variables were performed using Fisher’s exact tests. P values less than 0.05 were considered significant. Combination IC80 titers were calculated using the Bliss-Hill model, which has been shown to provide accurate predictions of combination neutralization properties using those of individual bNAbs (19). Statistical analyses were performed using GraphPad Prism or the Stats module in SciPy ( MEGA6 was used to do the single-genome sequence alignments and create the maximum likelihood tree.


Fig. S1. Analysis of SHIV env sequences in PGT121-treated animals.

Fig. S2. Complementarity of complete neutralization profiles of PGDM1400 and PGT121 against HIV-1.

Fig. S3. Serum concentration of PGDM1400 and PGT121 in antibody-treated animals by ELISA.

Fig. S4. Serum concentration of PGDM1400 and PGT121 in antibody-treated animals by TZM-bl NAb assays.

Fig. S5. Comparison of ELISA and TZM-bl neutralization assays for evaluating pharmacokinetics of PGDM1400 and PGT121.

Fig. S6. Protective efficacy as measured by SHIV-SF162P3–specific viral loads.

Fig. S7. Protective efficacy as measured by SHIV-325c–specific viral loads.

Fig. S8. Structural modeling of PGT121 and PGDM1400 contact sites.

Fig. S9. Clade diversity of key signature residues for PGDM1400.

Fig. S10. Clade diversity of key signature residues for PGT121.

Table S1. Intrapatient variation in key contact signatures for PGDM140 and PGT121.


  1. Acknowledgments: We thank S. Mojta, P. Gandhi, B. Keele, W. Rinaldi, M. Ferguson, M. Lewis, and J. Misamore for advice, assistance, and reagents. Author contributions: B.J., P.-T.L., B.T.K., and D.H.B. designed the studies and interpreted the data. P.-T.L., P.A., N.B.M., J.B.W., and D.R.B. led the virologic assays. K.W., W.M.F., and B.T.K. led the sequence analyses, model development, and statistical analyses. J.P.N., K.M., E.N.B., S.K., M.K., J.A.L., and M.S.S. led the immunologic assays. D.H.B., L.J.T., and J.R.M. led the analysis of the challenge stocks. B.J., P.-T.L., B.T.K., and D.H.B. wrote the paper with all coauthors. Funding: We acknowledge support from the American Foundation for AIDS Research (109219), the NIH (AI096040, AI100663, AI106408, AI124377, and AI126603), and the Ragon Institute of Massachusetts General Hospital, Massachusetts Institute of Technology, and Harvard University. Competing interests: D.R.B. is a co-inventor on U.S. Patent 9,464,131 held by co-owners Theraclone Sciences, International AIDS Vaccine Initiative (IAVI), and The Scripps Research Institute (TSRI) that covers PGT121. D.R.B. is a co-inventor on U.S. Patent application no. 14/731,621 held by co-owners Cornell University, IAVI, and TSRI that covers PGDM1400. The other authors declare that they have no competing interests. Data and materials availability: The data presented in this paper are tabulated in the main paper and in the Supplementary Materials. Antibodies and SHIVs are available with an appropriate material transfer agreement, and requests for materials should be addressed to D.H.B. (dbarouch{at}

Stay Connected to Science Translational Medicine

Navigate This Article