Full Text

Turn on search term navigation

ARTICLE

Received 23 Oct 2014 | Accepted 9 Feb 2015 | Published 10 Apr 2015

DOI: 10.1038/ncomms7565 OPEN

Analysis of immunoglobulin transcripts and hypermutation following SHIVAD8 infection and protein-plus-adjuvant immunization

Joseph R. Francica1,*, Zizhang Sheng2,*, Zhenhai Zhang2,3, Yoshiaki Nishimura4, Masashi Shingai4,Akshaya Ramesh5, Brandon F. Keele6, Stephen D. Schmidt1, Barbara J. Flynn1, Sam Darko1, Rebecca M. Lynch1, Takuya Yamamoto1, Rodrigo Matus-Nicodemos1, David Wolinsky1, NISC Comparative Sequencing Programw, Martha Nason7, Nicholas M. Valiante8, Padma Malyala8, Ennio De Gregorio8, Susan W. Barnett8, Manmohan Singh8, Derek T. OHagan8, Richard A. Koup1, John R. Mascola1, Malcolm A. Martin4,Thomas B. Kepler5,**, Daniel C. Douek1,**, Lawrence Shapiro2,** & Robert A. Seder1,**

Developing predictive animal models to assess how candidate vaccines and infection inuence the ontogenies of Envelope (Env)-specic antibodies is critical for the development of an HIV vaccine. Here we use two nonhuman primate models to compare the roles of antigen persistence, diversity and innate immunity. We perform longitudinal analyses of HIV Env-specic B-cell receptor responses to SHIVAD8 infection and Env protein vaccination with eight different adjuvants. A subset of the SHIVAD8-infected animals with higher viral loads and greater Env diversity show increased neutralization associated with increasing somatic hypermutation (SHM) levels over time. The use of adjuvants results in increased ELISA titres but does not affect the mean SHM levels or CDR H3 lengths. Our study shows how the ontogeny of Env-specic B cells can be tracked, and provides insights into the requirements for developing neutralizing antibodies that should facilitate translation to human vaccine studies.

1 Vaccine Research Center, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 20892, USA. 2 Department of Biochemistry, Columbia University, New York, New York 10032, USA. 3 State Key Laboratory of Organ Failure Research and National Clinical Research Center for Kidney Disease, Nanfang Hospital, Southern Medical University, Guangzhou, Guangdong 510515, China. 4 Laboratory of Molecular Microbiology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 20892, USA. 5 Department of Microbiology and Immunology, Boston University, Boston, Massachusetts 02118, USA. 6 AIDS and Cancer Virus Program, Leidos Biomedical Research Inc., Frederick National Laboratory, Frederick, Maryland 21702, USA. 7 Biostatistics Research Branch, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland 20892, USA. 8 Novartis Vaccines and Diagnostics, Cambridge, Massachusetts 02139, USA. * These authors contributed equally to this work. ** These authors jointly supervised this work. w A full list of consortium members appears at the end of the paper. Correspondence and requests for materials should be addressed to R.A.S. (email: mailto:[email protected]

Web End [email protected] ).

NATURE COMMUNICATIONS | 6:6565 | DOI: 10.1038/ncomms7565 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms7565

Ahighly successful preventative HIV vaccine will likely require antibodies that neutralize HIV and block virus acquisition. One of the greatest challenges to HIV vaccine

development is the elicitation of antibodies with sufcient breadth and potency to counter the genetic diversity of strains that may establish an infection1,2. Over the past several years, the cloning and characterization of a number of broadly neutralizing monoclonal antibodies (bNAbs) from HIV-infected humans has identied distinct sites on the HIV Envelope (Env) that are vulnerable to neutralization, and dened several characteristics critical for their protective function3,4. Many of these bNAbs have high levels of somatic hypermutation (SHM)57 or long third complementarity-determining regions of the heavy chain (CDR H3)3,8.

To improve HIV vaccine development, an experimental preclinical animal model is needed to assess how B-cell lineages are elicited and antibodies mature in response to vaccination or infection. Preclinical models using rodents and rabbits may be limited in their ability to elicit responses similar to those characteristic of human bNAbs because of their evolutionarily divergent immunoglobulin (Ig) gene repertoires9 or lack of natural CD4 expression. In contrast, nonhuman primates (NHPs) are capable of eliciting cross-reactive neutralizing responses to the V3/glycan site following SHIVAD8 infection10, tier 1 neutralizing antibodies to the CD4-binding site11 and binding antibodies to the membrane proximal external region12 following Env protein or peptide vaccination. The high similarity between the antibody genes of humans and NHPs may underlie the ability to elicit such similar responses9,11. Moreover, because NHPs are most similar to humans in other immunologic aspects, such as tissue-specic Toll-like receptor (TLR) expression13,14, they provide greater predictability of human vaccine responses than rodent models15,16.

Recent studies of neutralizing antibody responses in HIV-infected individuals have used next-generation sequencing (NGS) to study the genetic record of antibody development encoded in peripheral memory B cells1719. The germline-encoded antibody segments (V, D and J) provide critical elements for interpreting these data. The currently available heavy chain (HC) gene repertoire for rhesus macaque11,20 was obtained by whole-genome sequencing of a single animal with 5 depth of

coverage. Here we report a new draft database of VH gene sequences from 10 Indian-origin rhesus macaques acquired using Illumina deep sequencing to B50100 coverage. Using this

new draft database, we applied B-cell Ig transcript analysis methods similar to those used previously to interrogate human repertoires18,19. In this study, longitudinal analyses of two cohorts were performed to address the critical questions of how antigen load, diversity, persistence and innate immunity alter antibody responses: (1) NHP infected with SHIVAD8 provide a model with persistent and diverse Env antigens that has been shown to induce potent cross-clade serum-neutralization responses10,21,22; and (2) NHPs vaccinated with gp140 Env protein and eight different adjuvants (alum or MF59 with or without TLR4 or TLR7 ligands, pIC:LC or immune stimulator complexes (ISCOMs)), which were chosen because they are clinically approved or in advanced development, and because they mediate their effects through distinct innate mechanisms that could inuence B-cell immunity. From both of these studies, peripheral antibody transcripts isolated from Env-specic B cells were sequenced to assess SHM, CDR H3 length and variable heavy (VH) gene repertoire.

The data presented here using the NHP vaccine model conrm that a number of clinically based adjuvants are effective for enhancing the magnitude of antibody response but, remarkably, are not able to increase SHM. Alternatively, during chronic

SHIVAD8 infection, antigen diversity and persistence appear critical for enhancing the potency and breadth of neutralizing Env antibody responses and SHM.

ResultsDevelopment of an NGS platform for Env-specic B cells. To study B-cell ontogeny and characterize Ig maturation features, an NGS platform was developed that could be used to study a large number of NHPs (Supplementary Fig. 1). Briey, Env probe-specic B cells were bulk-sorted from individual animals at various time points after SHIVAD8 infection or Env and adjuvant vaccination (Supplementary Fig. 2), IgG HC transcripts were amplied by multiplexed primer PCR with unique barcodes, and sequenced by 454 pyrosequencing. Raw reads were then ltered for quality and redundancy and mapped to a newly generated rhesus macaque Ig draft reference database (new accession codes KP710506 to KP710583 and NW_001121239 to NW_001121240 from previous IMGT database). The draft database comprises 58 VH genes, with 98 alleles in total (Supplementary Table 1). There were 9, 3, 28, 12, 3, 1 and 2 members of the VH1, VH2, VH3, VH4, VH5, VH6 and VH7 families, respectively. Temporary names for these genes were assigned; the closest mapping between NHP and human VH genes23, as well as to a previously published

NHP database20, is shown in Supplementary Fig. 3.

To obtain a detailed understanding of B-cell responses to infection and vaccination, three major analyses were performed on the unique sequences obtained from NGS: (1) SHM, calculated as percent divergence from germline at the nucleotide (nt) level;(2) CDR H3 length; and (3) the VH gene origin. Unique sequences were used for each of these analyses because, for small sequence differences, true somatic variants cannot be distinguished from PCR-induced replicates and redundant reads could affect gene-usage distributions. The inclusion of redundant sequences yielded very similar results for the mean divergence analyses compared with using only unique reads (Supplementary Fig. 4a). To further ensure that gene-usage distributions obtained with multiplex primer PCR were not biased, rapid amplication of cDNA end (RACE) PCR, which uses a common 50 primer to avoid primer-associated amplication bias, was used for comparative purposes. Similar results were obtained by these different methods (Supplementary Fig. 5) demonstrating that the multiplex primer PCR approach did not signicantly change the VH gene distribution.

Viral load and Env sequence diversity during SHIVAD8 infection. To study the development of neutralizing antibodies, NHPs were infected with SHIVAD8, a tier 2B virus that has been shown to

induce potent cross-reactive serum neutralization in NHP10,21,22. This model shows features of HIV infection in humans such as continuous CD4 T-cell loss, sustained plasma viraemia and clinical immunodeciency. Plasma from infected NHPs was screened over the course of infection for neutralization against a diverse panel of tier 1 and 2 viruses. Animals were segregated based the basis of neutralization breadth and potency: four were categorized as good neutralizers and four as poor neutralizers (Fig. 1a). Poor neutralizers were categorized as demonstrating limited neutralization of tier 1A and 1B strains, while good neutralizers were those that showed much stronger tier 1 neutralization as well as moderate activity against several tier 2 strains, including the autologous Env, CK15 3-3. Generally, neutralization developed after 50 weeks post infection10,21 (Supplementary Fig. 6a).

To determine what factors may inuence the development of neutralization in the SHIVAD8 model, viral load and diversity were assessed during infection. Longitudinal analysis showed that

2 NATURE COMMUNICATIONS | 6:6565 | DOI: 10.1038/ncomms7565 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms7565 ARTICLE

ID50

Clade / type Tier

Other

2 2

1B 1B 1 2 2 1B 2 2 1 1B 2 1B 2 2 1A

Animal ID

Weeks post infection

KER2018.11

Q23.17

UG037.8

DJ263.8

BaL.01

JR-FL

JR-CSF.JB

6535.3

7165.18

AC10.0.29

ADA.DG

HXB2

TRO.11

SS1196.1

CAP45.2.00.G3

DU156.12

MW965.26

SHIVDH12 clone 7 V3 AD8

SHIVAD8 CK15 3-3

Poor neutralizers

DBJI DCV9 DBVC DBZ3

Good neutralizers

DC6W DC8T DCF1 DCC7

<20 <20 <20 <20

825 277 285

271

32 <20

<2029 <20

<2020 <20

<2067 <20

<20

1,486 2,247

29200 163

<20600 416

<20 <20

<20<20 <20

<20

382 <20

2960 <20

<20223 <20

<20137 <20

<20

100 102 100 100

<20 <20

<20<20 <20

<20

47 <20 <20 <20

104 96

<20786 108

<20<20 33

<20609 211

<20

<20 <20 <20 <20

<20 <20 <20

654 229 156

454

414 1,234

194

975

<20 <20 <20 <20

4,206 2,123 2,697

747

7,375 2,554

53782 643

<206,450 1,672

<205,633 175

<20

102 102

<20 <20

<20<20 <20

155

<20 <20 <20 <20

5,086 11,235

2,555 862

148

77 25 59

<20

<20 <20

<20<20 <20

<20

606161 <20

248

<20

16,005 19,831

7,671 5,292

102 100

Poor neutralizers Good neutralizers

na n.a.

n.a.

DBJI

wk6 wk26 wk54 wk99

DC6W

wk6 wk26 wk54 wk99

DCF1

wk6 wk26 wk54 wk99

DCV9

wk6 wk26 wk54 wk99

DC8T

wk6 wk26 wk54 wk99

0.0030

DBZ3

wk6 wk26 wk54 wk99

DBVC

wk6 wk26 wk54 wk99

DCC7

wk5,7 wk26 wk54 wk99

Mean divergence *

Mean diversity

2.5

1.0

0.5

2.0

Percent divergence

Percent diversity

1.0

0.5

2.0

1.5

Poor neutralizers

Good neutralizers

wk6

wk26

wk54

wk99

0 wk6

wk26

wk54

wk99

V2 loop

167 167 309 309

Poor neutralizers Good neutralizers Poor neutralizers Good neutralizers

V3 loop

wk 6

wk 54

wk 26

wk 99

Figure 1 | Serum neutralization breadth and viral diversity during SHIVAD8 infection. (a) Plasma neutralization of selected viruses 100102 weeks post infection with SHIVAD8. Eight animals were segregated into good and poor neutralizers on the basis of the potency and breadth of their responses.

ID50 values are shown; colours indicate potency: 4099, green; 100999, yellow; Z1,000, red. (bd) Viral sequencing was performed on plasma from SHIV-infected NHP at week 6, 26, 54 and 99, except where indicated. (b) Phylogenic trees of Env sequences rooted to the inoculum sequence for good and poor neutralizers. Colours indicate time points; scale bar indicates 0.3% divergence. The mean divergence (c) or diversity (d) was calculated for each animal at the indicated time points. Horizontal bars indicate group medians; *, signicant discovery by multiple t-test comparison. Sequence data were not available for DBVC wk 54, DCV9 wk 26 and DCF1 wk99 because of low viral titres. (e) Consensus Env sequences at the indicated time points highlighting the V2 and V3 loops. Conserved mutations at amino-acid positions 167 and 309 are shown (orange) compared with the original residue (green) in the good and poor neutralizer animals.

NATURE COMMUNICATIONS | 6:6565 | DOI: 10.1038/ncomms7565 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms7565

animals, DCF1, had a large population of HC sequences with CDR H3s of 31 aa in length (Fig. 2d), which gradually increased in proportion over time (Fig. 2e). The ontogeny of these sequences was further analysed to assess potential mechanisms of how long CDR H3 regions develop. Most of the long CDR H3 regions from this animal, such as in sequence no. 107, attained their length solely through nucleotide (N-) addition; however some of the CDR H3s, such as in sequence no. 4594, were generated through V(DD)J recombination (Fig. 2f). All of these CDR H3 regions are anionic and are all predicted to have one or more sulfated tyrosines, similar to human V1V2-targeted bNAbs25,26. These long CDR H3 antibodies comprised 14 lineages with CDR H3 lengths ranging from 28 to 31 aa, the largest of which includes sequence no. 4594 (Fig. 2g). In all, nine of the fourteen long CDR H3 lineages were rst observed at week 28 and three of these lineages grew in proportion and diversity over time (Fig. 2h), although the average CDR H3 length remained constant (Fig. 2i). These data demonstrate that CDR H3 length is generated primarily during recombination and show convergent evolution from separately generated lineages over the course of infection.

Inuence of adjuvants on antibody titre and breadth. Chronic SHIVAD8 infection results in persistent innate activation and uctuations in the type and number of CD4 T cells, which may be critical for the induction of cross-reactive serum neutralization. Therefore, we assessed how adjuvants that mediate their effects through different innate immune pathways would inuence B-cell maturation in the context of Env protein vaccination.MF59 has been shown to recruit neutrophils, increase antigen uptake and increase antibody afnity27,28. TLR4 agonists are the rst TLR ligands approved for use in humans and activate certain dendritic cell subsets and macrophages leading to potent antibody responses29. Agonists to TLR3 (pIC:LC) and TLR7 were used because they induce robust production of interleukin (IL)-12 and Type I interferon (IFN), leading to potent CD4 TH1 immunity and antibody production14; TLR7 is also expressed in B cells, allowing for direct adjuvant activation. ISCOMs were also tested, as they have also been shown to be potent stimulators of both T- and B-cell immunity30. Alum was used a benchmark adjuvant based on its longstanding clinical use for enhancing antibody responses.

Rhesus macaques (n 6 per group) were immunized intra

muscularly at 0, 4, 12 and 24 weeks with gp140 TV1DV2 Env protein alone, or formulated with alum or MF59 with or without agonists of TLR 4 or 7, or with pIC:LC or ISCOMs (Fig. 3a).It should be noted that unlike the clade B AD8 strain used in the SHIVAD8 infection, the clade C Envelope TV1 was used for vaccination. TV1 was used because Novartis had clinical formulations of the protein and adjuvants, thus facilitating the

Figure 2 | Next-generation sequencing of antigen-specic B cells after SHIVAD8 infection. Data are shown from four good neutralizers, red; and four poor neutralizers, blue. (a,b) SHM for each animal over time; each symbol represents the average percent divergence from germline for Env-specic (gp120 )

(a) or nonspecic (gp120 ) (b) sequences from a given animal; error bars indicate s.e.m. (c) Histogram representation of SHM distribution from all

sequences from 90110 weeks post infection; binning averaged in 2% increments. (d) CDR H3 length distribution of Env-specic sequences from individual animals for all time points combined. Colour/symbol scheme as in a; binning averaged in 3-aa increments; arrow indicates population of sequences with long CDR H3 regions. (e) Proportion of unique sequences with long CDR H3 regions from DCF1 at the indicated time points. w2-test with Yates correction was used to calculate P values between proportions from Env or samples at each time point; n.s., not signicant; *Po0.0001. (f) Long CDR H3

regions from DCF1 may be derived in at least two unique ways. Example 1 depicts the CDR H3 region of sequence 4594, arising from V(DD)J recombination. Example 2 depicts the CDR H3 region of sequence 107, arising from N-addition. Bold text indicates mature antibody sequence; red stars indicate predicted tyrosine sulfation; CDR H3 charge is shown. (g) Divergence over time of long CDR H3 antibodies from the parent lineage that includes sequence 4594. A phylogenic tree was constructed by maximum likelihood and rooted to the IGHV4D*01 allele and is colour-coded by time point. (h,i) Two-dimensional plots depicting Env-specic sequences from animal DCF1 at the indicated time points. Sequences are plotted by their VH divergence from germline and their identity to the 4594 HC (h); or by their CDR H3 length (i). Red triangle indicates sequence 4594; magenta triangles indicate sequences related to 4594; shaded boxes indicate reads with long CDR H3 regions.

good neutralizers maintained viral loads 12 orders of magnitude higher, and had a more rapid decline in peripheral CD4 counts, than did poor neutralizers (Supplementary Fig. 6b,c). To track viral evolution and diversity, single-genome amplication (SGA) of blood plasma virus was performed over the course of the infection (Fig. 1b). In total, 441 single genomes were amplied with a mean of 55 per animal. Notably, Env sequences in the blood from good neutralizers were signicantly more divergent from the infecting inoculum (Po0.0001, multiple t-tests, Fig. 1c)

and showed greater sequence diversity over time than virus from poor neutralizers (P 0.002, multiple t-tests, Fig. 1d). Virus in

good neutralizer animals evolved about four times faster than in poor neutralizers (0.02 versus 0.005 substitutions per site per year) over a period of 100 weeks. Of note, E to K mutations at position 167 in the V2 loop epitope and R to S mutations at position 309 in the V3 loop epitope were consistently found in virus from all of the good neutralizers but not the poor neutralizers (Fig. 1e). These conserved mutations were rst observed at week 26 and became predominant in the population over time, suggesting immunological pressure at these sites. Taken together, these data highlight the dynamic interactions between virus and neutralization in the macaque SHIVAD8 model

and provide an opportunity to study how these virologic characteristics relate to B-cell Ig sequences.

SHIVAD8 infection analysis of SHM and neutralization. To determine how SHIVAD8 infection inuences Ig maturation, AD8 gp120 Env-specic and nonspecic memory B cells from infected animals were isolated at various time points post infection, sequenced and analysed (Supplementary Table 2). A total of 5,526 unique HC sequences were derived from 28,000 Env-specic cells, and 311,274 HC sequences were derived from B1,313,000 nonspecic cells. Env-specic B cells accumulated mutations over the course of infection and two of the four good neutralizers had higher divergence levels than the poor neutralizers (Fig. 2a). As a control, sequences from nonspecic memory B cells did not signicantly accumulate mutations over the course of the infection (Fig. 2b). Indeed, by 110 weeks post infection Env-specic B cells had accumulated more mutations from germline than had nonspecic cells (Fig. 2c). These data show that SHIVAD8 infec

tion results in the accumulation of SHM that is HIV-specic and correlates with serum neutralization.

Tracking CDR H3 length during SHIVAD8 infection. Long CDR H3s are another dening characteristic of several human HIV bNAbs. Similar to what has been observed in humans24, nonspecic memory B cells from SHIVAD8-infected animals had a median CDR H3 length of 14 aa, and Env-specic memory B cells had a median length of 16 aa. One of the good neutralizer

4 NATURE COMMUNICATIONS | 6:6565 | DOI: 10.1038/ncomms7565 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms7565 ARTICLE

translation of preclinical study ndings to humans. Clade C strains are most common in Africa and Asia; therefore, this strain is relevant to these regions where several vaccine trials are currently being planned. Formulations derived from MF59 to include TLR agonists are referred to as adjuvant nano-emulsions (ANE).

To rst show how the adjuvants affected humoral immunity, Env-binding titres were assessed from the peak of the response after the nal immunization, week 26. Animals receiving alum/TLR7, MF59 and ANE/TLR4 produced the highest IgG titres, while alum/TLR4, pIC:LC and ISCOMs showed a trend towards enhancement over

Env-specific memory B cells

Weeks post infection

Nonspecific memory B cells

Weeks post infection

SHM distribution (90110 w.p.i.)

Relative frequency (percentage)

Good neutralizers Env + Good neutralizers Env Poor neutralizers Env + Poor neutralizers Env

% Divergence from germline

DCF1 DC6W DC8T DCC7 DCV9 DBVC DBJI DBZ3

Good neutralizers

Poor neutralizers

% Divergence from germline

80 120

0 80

120

0 0 2 4 6 8

10 12 14 16 18 20 22 24 26 28 30

% Divergence from germline

CDRH3 length distribution, Env-specific B cells

Relative frequency (percentage)

100 Proportion of DCF1 sequences with

long CDRH3 regions

Env + Env

* *

Percent

n.s.

0.1

10 0 3 6 9 12

wk 6

wk 28

wk 52

wk 102

wk 108

CDRH3 length (aa)

D segment no.1

D segment no.2

IGVH4-D*01

Example 1 (nt)

IGHD2-2*01-----AGGATATTGTAGTGGTGGTGTCTGCTACGCC---------------------------------------------------------IGHD3-1*01--------------------------------------GTATTACGAGGATGATTACGGTTACTATTACACCCACAGCGT-------------4594 GCGAG...G.GG....A.AA.........T...C.ATT.G.............A.T..........A..T..G.TCACAATCGGTTCGATGTC

D segment

N- addition N- addition

Week 28 52 102 108

IGHD3-1*01---------------------TATTACGAGGATGATTACGGTTACTATTACACCCACAGCGT----------------------------

Example 2 (nt) 107 TGTGCGAGAAGGGGGTCTTCG...........A....T....G........T.GGG.CCGACGACCCACGGAATAAAGTCATTGGATGTC

0.05

IGHD2-2*01--GYCSGGVCYA-------------------IGHD3-1*01-------------YYEDDYGYYYTHS-----

Example 1 (aa) 4594 AR.W.NN....PLD...EF...NSDHNRFDV

IGHD3-1*01------YYEDDYGYYYTHS----------

Example 2 (aa) 107 ARRGSS...E.F.D..SGPTTHGIKSLDV

Q = 3.9

Q = 1.8

4594_wk108

* *

Week 28 Week 52

Week 6 Week 108

100

4594

4,594 related lineage

Identity to 4594 (%)

104

103

102

101

100

Reads

Long CDRH3

CDRH3 length (aa)

0 30

Germline VH divergence (%)

NATURE COMMUNICATIONS | 6:6565 | DOI: 10.1038/ncomms7565 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms7565

Group

Number

Vaccine description

Immunization

1 6

Env

Env + alum

Env + alum / TLR4 agonist

Env + alum / TLR7 agonist

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 36

34 (Week)

Env + MF59

Env + ANE / TLR4 agonist

Peripheral sampling

Env + ANE / TLR7 agonist

Env + polyIC:LC

Env + ISCOM

Binding IgG

Tier 1A neutralization

* *

* * * *

End point titre (Log 10)

ID 50 (Log 10)

Env

Alum

Alum / TLR4

Alum / TLR7

MF59

ANE / TLR4

ANE / TLR7

pIC:LC

ISCOM

Pre-vax

Wk 6

Wk 26

Wk 36

SSC

0.695

11.8

13.1

2.45

gp140

Env-specific, memory B cells

Env-specific, nave B cells

EnvEnv + AlumEnv + Alum / TLR4 Env + Alum / TLR7 Env + MF59Env + ANE / TLR4 Env + ANE / TLR7 Env + pIC:LCEnv + ISCOM

* *

gp140 specific cells (%)

* *

Pre-vax

wk 6

wk 26

wk 36

Pre-vax

wk 6

wk 26

wk 36

Figure 3 | Vaccination with adjuvants results differential effects on humoral and B cellular responses. (a) Vaccination project overview. Nine vaccines were given to 53 NHP in a homologous prime-boost manner at 0, 4, 12 and 24 weeks (black arrows). PBMC sampling was performed before vaccination (pre-vax) and 6, 26 or 36 weeks after the prime (red arrows). (b) Env-specic IgG-binding titres at week 26. (c) Plasma neutralization of tier 1A MW965.26 Env-bearing pseudovirus at week 26. Horizontal bars indicate medians. *Po0.05; **Po0.01 compared with Env alum group by the

KruskalWallis test. (df) Antigen-specic B cells were identied from PBMCs at the indicated time points by binding to a gp140 protein probe followed by ow cytometry. (d) Representative ow cytometry plots of antigen-specic memory B cells at the indicated time points after prime. Antigen-specic memory (e) or naive (f) cells were enumerated at the indicated time points. Data points represent vaccine group means and are depicted as a percent of all IgG cells. *Po0.05 compared with pre-time point by two-way analyses of variance.

the benchmark adjuvant, alum (Fig. 3b). Plasma neutralization was observed against a panel of seven tier 1 viruses, and the potency among adjuvant groups was similar to the binding titres; activity against tier 2 viruses was largely undetected

(Fig. 3c). These results provide an extensive comparison of existing and novel clinically based adjuvant formulations, conrming their ability to increase the magnitude of antibody responses.

6 NATURE COMMUNICATIONS | 6:6565 | DOI: 10.1038/ncomms7565 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms7565 ARTICLE

Average SHM per animal

SHM per sequence

% Divergence from germline

Group mean

Env

Env + Alum

4 Env + Alum/TLR4

Env + Alum/TLR7

Env

Env + MF59

Env + ANE/TLR4

Env + ANE/TLR7

Env + pIC:LC

Env + ISCOM

Env + Alum

Env + Alum/TLR4

Env + Alum/TLR7

Env + MF59

Env + ANE/TLR4

Env + ANE/TLR7

Env + pIC:LC

Env + ISCOM

Distribution of somatic hypermutation by vaccine

Relative frequency (percentage)

Animal Adjuvant

CK8W No adjuvant

O4E107 MF59

CK8V Alum / TLR7

CL47 Alum / TLR7

CL24 pIC:LC A4E001 pIC:LC

LWK Alum / TLR7

EnvEnv + AlumEnv + Alum / TLR4 Env + Alum / TLR7 Env + MF59Env + ANE / TLR4 Env + ANE / TLR7 Env + pIC:LC

Env + ISCOM

% Divergence from germline

****

** ****

****

5 30

SHM in select animals over time

% Divergence from germline

Distribution of CDRH3 length by vaccine

Weeks post prime

Relative frequency (percentage)

Number of reads

10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44

0 2 4 6 8

CDRH3 length (aa)

Figure 4 | Next-generation sequencing of antigen-specic B cells after protein and adjuvant vaccination. (a) SHM for each animal (n 51) at week 26;

each symbol represents the average percent divergence from germline for sequences from a given animal. Horizontal bars indicate the medians. (b) SHM for each vaccine at week 26; each symbol represents the percent divergence from germline for a unique sequence. Horizontal yellow bars indicate the geometric means. (c) Histogram representation of SHM at week 26. The distribution is shown as a composite for all unique sequences from a given vaccine; binning averaged in 2% increments. (d) Effect of boosting on SHM. Percent divergence of antigen-specic cells sorted at week 6 or 26 is depicted for seven animals. Data points represent the mean SHM for each animal; error bars show s.e.s; vertical dashed lines represent immunization points; n.s., not signicant; **Po0.01; ****Po0.0001 compared with week 6 time point by two-way ANOVA. (e) CDR H3 length distribution as a composite from all sequences from a given vaccine; binning averaged in 2-aa increments. Inset highlights the low frequency of reads with long (Z28 aa) CDR H3 regions. Colours denote vaccines as in c.

The frequency of Env-specic B cells is modulated by adjuvants. To assess how adjuvants affect Env-specic B cells, these were rst quantied from peripheral blood mononuclear cells (PBMCs) by their binding of a labelled gp140 TV1DV2 Env probe throughout the course of vaccination (Fig. 3d; Supplementary Fig. 2b). Animals receiving Env alone or Env with alum showed a low frequency of Env-specic B cells, and were increased by addition of the TLR 4 or 7 agonists (Fig. 3e). The MF59, pIC:LC and ISCOM formulations all induced a striking increase in Env-

specic cells after two or four immunizations, although these populations signicantly contracted by week 36 (Fig. 3e). In contrast, Env-specic naive B cells were found at a much lower frequency throughout the course of vaccination (Fig. 3f).

Env-specic memory B cells were then sorted from 51 of the vaccinated animals after the fourth immunization and processed through the NGS pipeline, yielding 53,563 Ig HC unique sequences from B201,000 sorted Env-specic B cells (Supplementary Table 3). These data were subsequently analysed

NATURE COMMUNICATIONS | 6:6565 | DOI: 10.1038/ncomms7565 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms7565

on a per-animal or a per-vaccine basis; each vaccine group has at least four animals and a minimum of 2,000 unique sequences, allowing for robust comparisons at both levels (Supplementary Table 4).

Adjuvants do not alter overall SHM levels. For each VH sequence, CDR and framework regions (FRs) were delineated following IMGT denitions31 and the level of SHM within each

region was assessed (Supplementary Fig. 7a). Mutations were signicantly higher in the CDR1 and CDR2 regions (12%, 15% divergence from germline), compared with the FRs (46% divergence). However, the adjuvants did not affect the accumulation of mutations in these regions, nor did they affect the proportion of VH gene families (Supplementary Fig. 7b). The inuence of adjuvants on accumulated SHM levels of Env-specic B cells was next determined. The mean divergence from germline after four immunizations (week 26) was analysed for each animal (Fig. 4a) or by grouping all the individual sequences from each vaccine (Fig. 4b,c). Notably, animals immunized with the Env protein alone induced the highest average divergence, while the addition of MF59 (P 0.007, mixed effect linear regression

model) and ANE/TLR7 (P 0.007) had statistically lower levels.

The nding that these adjuvants did not increase SHM was subsequently conrmed in an independent follow-up experiment in which SHM levels were assessed after two or three homologous immunizations with several of the same adjuvants (Supplementary Fig. 8a,b). To extend the analysis, we determined whether germline divergence increased after successive immunizations. Env-specic B cells were sequenced from seven animals of different vaccine groups after two immunizations (week 6). Six of these animals showed an increase from week 6 to week 26 (Fig. 4d), showing that successive boosting did increase accumulated SHM for antigen-specic transcripts (Supplementary Fig. 8c).

Although the overall mean divergence levels was not increased with these adjuvants, the germline divergence of particular VH genes was highly variable among the vaccine groups. Using data from the two NHP protein-plus-adjuvant vaccination studies, we found that after mapping sequences to each of the 58 VH genes, some such as IGHV4L were largely unaffected by the adjuvant used. In contrast, others such as IGHV3Q showed large variations among the adjuvants (Supplementary Fig. 8e). These data suggest the ability of adjuvants to differentially promote the development of B cells with specic VH genes, although further studies should be performed to investigate this phenomenon.

Adjuvants do not alter the development of CDR H3 length. To conclude the analysis of how adjuvants inuence B-cell development, the CDR H3 length distribution for each vaccine group were assessed. The median CDR H3 length was 14 or 15 aa (Fig. 4e; Supplementary Table 4) and none of the vaccine adjuvants elicited a population of Env-specic B cells with long CDR H3 regions; only 11 sequences were found to have a CDR H3 Z28 aa in length (Fig. 4e, Supplementary Fig. 8d).

This nding is consistent with the frequency of long CDR H3 sequences from the naive repertoire (B0.03%), which was

Figure 5 | Characteristics of animals and sequences with high levels of SHM at week 26 after vaccination. (a) Histogram representation of SHM distribution from animals with an average 410% divergence from germline, compared with all vaccinated animals. Binning averaged in 2% increments.

(b) VH gene composition for sequences from animals with high SHM compared with the whole data set, all animals. The frequency of reads mapping to each VH gene are represented as a fraction of the total sequences. (c) VH gene composition for sequences 420% divergent from germline from a subset of samples taken pre-vaccination (bulk IgG reads)

or from Env-specic B cells sorted from all animals post-vaccination. (d) VH gene composition for sequences 420% divergent from germline from sorted Env-specic (gp120 ) or nonspecic (gp120 ) B cells from

SHIV-infected animals. P values are derived from the Fishers exact test for the proportion of IGHV3Q. The number of sequences in each population subset is indicated under the corresponding pie chart. Sequences mapping to IGHV4L (blue), IGHV3Q (yellow) and IGHV3J (red) are highlighted.

25 Distribution of SHM

% Divergence from germline

Relative frequency (percentage)

A4E001 CK8W

DBC8

O4E016

CL24

DBE6

All animals

0 2 4 6 8 10

All animals

A4E001

CK8W

CL24

IGHV4E

DBC8

DBE6 O4E016

n=235

IGHV3Q

IGHV4L

IGHV3J

Pre-vaccination, nonspecific reads >20% divergent

Post-vaccination, Env-specific reads >20% divergent

n=758

IGHV4L

IGHV3Q

P<0.0001

SHIV, nonspecific reads >20% divergent

SHIV, Env-specific reads >20% divergent

n=9

n=249

P<0.0001

8 NATURE COMMUNICATIONS | 6:6565 | DOI: 10.1038/ncomms7565 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms7565 ARTICLE

Env versus non-Env VH genes

Non-Env(% of total sequences)

SHIV versus vaccine VH genes

SHIV (% of total sequences)

Good vs. poor SHIV neutralizer VH genes

IGHV1E

IGHV4L

Env (% of total sequences)

Good neutralizers

(% of total sequences)

IGHV1E

IGHV4D IGHV4A

IGHV4D

IGHV4A

IGHV3J

IGHV4E

5 0 20

5 0 25

Vaccine(% of total sequences)

Poor neutralizers (% of total sequences)

Figure 6 | VH gene correlations between Env/adjuvant vaccination and SHIVAD8 infection models. The composition of individual VH genes is graphed as the percentage of sequences mapping to each VH gene within each data set. Each dot represents an individual VH gene; diagonal lines indicate the position of genes lacking a preference between data sets. (a) HIV Env-specic compared with nonspecic sequences combined from SHIVAD8 infection and Env/adjuvant vaccination data sets. (b) Env-specic sequences from the SHIV data set compared with the Env/adjuvant vaccination data set. VH genes with enriched composition in the SHIV data set are shown in green. (c) Env-specic sequences from SHIVAD8 good neutralizers compared with poor neutralizers. VH genes with enriched composition in the good neutralizer data set are shown in red; those enriched in the poor neutralizer data setare shown in blue.

measured by sorting naive B cells from two NHPs before vaccination (Supplementary Fig. 4b).

Highly mutated Env antibodies arise from specic VH genes. Because some HIV bNAbs accumulate high levels of SHM, the data set was examined for highly mutated antibodies. While the mean divergences from germline genes were similar between the vaccine groups, eight animals from various vaccine groups had high mean divergence levels (Fig. 4a). Furthermore, in all vaccine groups there were animals containing highly divergent sequences (Fig. 4b). Therefore, animals with 410% mean SHM (n 6,

excluding O4E064 and T5321, which had very low numbers of unique reads) were further assessed for SHM distribution and VH gene composition. These animals had a relatively uniformly higher distribution of mutated sequences compared with all vaccinated animals (Fig. 5a). Of note, they showed an increase in the proportion of two VH genes, IGHV3Q and IGHV3J, and a decrease in other VH genes, especially IGHV4L, which was the most prominent VH gene for all other animals (Fig. 5b). Overall, these data show that there were signicant alterations in VH gene usage in animals with the highest levels of SHM.

We next analysed all sequences that were 420% divergent in the Env vaccination study. IGHV3Q was again the predominant VH gene, comprising 37% of such reads, with a relatively small frequency of IGHV4L (Fig. 5c). As a control, we compared this distribution to reads that were 420% divergent from nonspecic memory B cells (from pre-vaccination samples, n 584,634).

Here IGHV4L was the predominant VH gene used, and IGHV3Q comprised only 5% (Fig. 5c). To see whether the preference of high SHM reads for IGHV3Q was limited to vaccinated NHP, we analysed highly divergent reads from the SHIVAD8-infected animals. Only nine unique highly divergent reads were found, yet six of these mapped to IGHV3Q (Fig. 5d). Among highly mutated reads from nonspecic B cells, IGHV4L was the most prominent VH gene, as was observed with nonspecic reads from the Env vaccination study. Taken together, these data suggest that high levels of SHM in Env-specic antibodies are associated with a preference for the IGHV3Q gene. This nding was conrmed in a follow-up Env protein-plus-adjuvant vaccination study (Supplementary Fig. 8f).

Preferential VH gene usage following SHIVAD8 infection. To address whether certain VH genes are preferentially used for Env

reactivity or neutralization, a composite analysis of data obtained from both the Env vaccination and SHIVAD8 infection studies was performed. Pre-vaccination IgG sequences and nonspecic memory B cells from the SHIVAD8 infection study provided a non-Env data set (n 890,460). Env-specic reads from B cells

sorted after vaccination and SHIVAD8 infection were combined for an Env-specic data set (n 58,942). The proportions of VH

genes between these data sets were compared and an enrichment of specic VH genes was assessed. However, there was no preferential VH gene usage between the sequences from the Env- and non-Env-specic data sets (Fig. 6a).

The VH gene composition of Env-specic sequences between the Env vaccination and the SHIVAD8 infection studies were then compared. In SHIVAD8-infected animals, sequences preferentially mapped to IGHV1E (Po0.0001, w2-test), 4A (Po0.0001) and 4D (Po0.0001), which could represent a VH gene signature of SHIVAD8 infection (Fig. 6b). Furthermore, in comparing VH gene composition between good and poor neutralizers from SHIVAD8-

infected animals, IGHV4A (Po0.0001, w2-test), 4D (Po0.0001) and 3J (Po0.0001) were more prevalent in the good neutralizers at a high level, while IGHV4E (Po0.0001) and 4L (Po0.0001) were more prevalent in the poor neutralizers (Fig. 6c). These ndings were substantiated by performing the same analysis using Sanger sequenced, single-cell sorted Env-specic B cells from a separate study of SHIVAD8-infected NHP (Supplementary

Fig. 9). This analysis shows how the NHP model can be used to evaluate which immunoglobulin VH genes play a role in specic immune responses.

DiscussionUsing a high-throughput NGS platform with a newly generated draft Rhesus macaque Ig HC reference database, Env-specic antibody responses were assessed following SHIVAD8 infection and Env protein-plus-adjuvant vaccination to gain insight into factors that inuence SHM and other characteristics associated with neutralizing antibodies. The SHIVAD8 infection model provides a striking contrast in the amount, duration, diversity and conformation of Env antigen compared with homologous Env protein immunization. Nevertheless, it was notable that antigen-specic B cells in both SHIVAD8-infected and Envvaccinated animals had similar mean germline divergence levels at both week 6 and week 26 (Figs 2a and 4d). As the SHIVAD8

infection progressed past week 50, animals that maintained their

NATURE COMMUNICATIONS | 6:6565 | DOI: 10.1038/ncomms7565 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 9

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms7565

viral loads had increased viral diversity and increased antibody germline divergence, which was associated with the development of cross-reactive neutralization. These data are consistent with several studies of HIV-infected individuals in whom increased viral loads and viral Env diversity were correlated with serum neutralization capacity3234. By contrast, and consistent with other gp140 vaccination studies35,36, Env-vaccinated NHP did not develop signicant cross-reactive neutralization responses. Of note, sequences greater than 20% divergent from germline (which is in the range of many characterized bNAbs5) were detected from some animals in all vaccine groups after four immunizations. These data show that levels of SHM accumulated following infection or vaccination may be necessary but are not sufcient to generate potent cross-reactive neutralization.

A major question addressed in this study was whether various adjuvant formulations differed in their capacity to increase SHM levels. Here clinically approved adjuvants (alum, MF59) were evaluated individually as benchmarks or combined with TLR ligands such as TLR 4 and 7 to enhance their potency. These ligands activate innate immune cells such as macrophages and specic DC subsets to secrete IL-6 and Type I IFN, respectively, mediators that can induce B-cell activation and enhance antibody production3739. TLR7 is also expressed in B cells and can directly induce B-cell activation. Moreover, TLR stimulation has been shown to activate AID40 and SHM41 in vitro, and other reports have showed afnity maturation using MF59 in an inuenza vaccine28 or CpGs in a hepatitis B vaccine42, although direct links between afnity and SHM were not explored. Remarkably, we found limited differences in the mean levels of SHM levels induced by the different adjuvants compared with the Env protein alone after four homologous immunizations. We note that for Env-specic B-cell isolation, trimeric gp140 was used in the vaccine adjuvant study and gp120 Env protein used in the SHIVAD8 infection study. Both of these proteins are uncleaved molecules, and thus are unable to bind quaternary-specic epitopes and potentially react with non-native epitopes. For future studies, cleaved, stable Env trimers will make better probes for isolating Env-specic B cells43.

The correlation over time between viral diversity and the development of neutralization breadth and potency in the SHIVAD8 model raises the question of how vaccines might be better designed to induce such a response. The SHIVAD8 data

(Supplementary Fig. 6a) indicate that some tier 1A reactivity develops before tier 2 neutralization in the good neutralizers. This suggests not that tier 1 antibodies are precursors to tier 2 antibodies but rather that tier 2 antibodies may take longer to develop. The virological differences between good and poor neutralizers do not simply result in improved tier 1 neutralization, but also give rise to greater breadth of neutralization, including tier 2 isolates. Therefore, vaccine regimens mimicking the antigen load and diversity seen during SHIVAD8 infection might induce better quality neutralization responses than what is currently observed after homologous immunization. To this end, strategies using sequential or concurrent immunization of multiple diverse Env proteins could be used. Indeed, such strategies have begun to be explored, using transmitter/founder virus envelopes44 and those derived from bNAb lineage studies45. We further suggest that immunization regimens with heterologous, cleaved Env trimers, given repeatedly over a prolonged period with adjuvants that improve innate stimulation, may enhance SHM and neutralization.

As long CDR H3 regions are a dening feature of bNAbs targeting the V1V2 epitope5,8, we explored how they were inuenced by vaccination and infection. The adjuvants studied here had no inuence on CDR H3 length. By contrast, and consistent with certain chronically infected HIV patients25, one of

the SHIVAD8-infected animals (DCF1, a good neutralizer) developed antibodies with long CDR H3 regions that are anionic and predicted to have tyrosine sulfation as do human V1V2 bNAbs25,26. Although we cannot address the neutralization capacity of these antibodies without matched light chains, we explored the mechanism underlying the generation of CDR H3 length to understand how the NHP model compares to what is observed in humans. The nding that V(DD)J recombination contributed to such antibodies in DCF1 was notable as this mechanism is thought to be relatively rare24,46. The data presented here best support the generation of long CDR H3 regions by N-addition during V(D)J recombination, with only small insertions added rarely through SHM. Importantly, naive B cells with long CDR H3 regions were readily detectable in NHP as has been reported in the human naive repertoire47. Thus, in designing vaccines to specically induce V1V2 antibodies, it should be possible to track the lineages of such long CDR H3 antibodies using the NHP model.

Finally, on the basis of alterations of the human B-cell repertoire after HIV and inuenza infection48,49, we assessed the effects of Env vaccination and SHIVAD8 infection on VH gene usage. Sundling et al.50 reported relative similarities in VH gene usage between Env-specic and total memory B cells after vaccination, a nding conrmed here. However, studies have demonstrated that certain epitope specicities are preferentially derived from certain VH genes (for example, VRC01-like class antibodies from VH12 (refs 5,7,9,19)). In addition, HIV infection has been associated with increased VH1 family usage, most notably VH169 (refs 51,52), whose NHP orthologue (IGHV1E) was found prominently in sequences from SHIVAD8-infected animals. In this study, the VH genes IGHV4A, 4D and 3J were specically associated with neutralizing responses following SHIVAD8 infection. Analysis of the VH repertoire after Env protein/adjuvant immunization in two independent studies revealed that highly mutated sequences preferentially mapped to the IGHV3Q gene (whose human homologue gives rise to the CAP256-VRC26 family of bNAbs25). Furthermore, we note that the adjuvants induced differences in specic VH genes, including

IGHV3Q. Understanding the mechanisms by which adjuvants inuence this process may be critical for strategies to rationally induce bNAbs.

In conclusion, on the basis of the similarities in innate and adaptive immune responses between macaques and humans, combined with the new draft Ig reference database and NGS platform, the NHP system allows for the evaluation of Env-specic B-cell development using future generations of HIV immunogens and adjuvant formulations.

Methods

SHIVAD8 infection and quantication of viral nucleic acids. Eight rhesus macaques of Indian origin ranging 28 years of age, either male or female, were maintained in accordance with the Association for Assessment and Accreditation of Laboratory Animal Care International standards, and were housed in a biosafety level 2 National Institute of Allergy and Infectious Diseases (NIAID) facility. Phlebotomies and sample collection were performed as previously described53. All animals were negative for the major histocompatibility complex class I Mamu-A*01 allele. Levels of CD4 T-cell subsets were measured using ow cytometry as previously reported22. The origin and preparation of the tissue-culture-derived

SHIVAD8 swarm and molecular cloned stocks have been previously described21,54. Viral RNA levels in the plasma were determined by real-time reverse transcription PCR (ABI Prism 7900HT sequence detection system; Applied Biosystems) as previously reported53.

Env vaccination study animals and immunizations. Fifty-three rhesus macaques of Indian origin ranging 38 years of age were divided into eight study groups of six and one group of ve (Fig. 3a). This number of animals was chosen to give B90% power to detect a 1.0 log difference in antibody titres between two vaccine groups and 65% power to detect 0.5 log differences, given reasonable estimates of group s.d.s from previous studies. Animals from previous vaccine studies were

10 NATURE COMMUNICATIONS | 6:6565 | DOI: 10.1038/ncomms7565 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms7565 ARTICLE

employed; these had not been exposed to any HIV or other retroviral proteins or nucleic acids and thus were naive for HIV Envelope; these weredistributed evenly among the vaccine groups. For vaccination, 100 mg trimeric gp140 TV1DV2 protein (Env)55 was adsorbed to alum and administered as such, or co-adsorbed with an agonist of TLR 4 (an MPL derivative, E6020) or an agonist of TLR 7 (a proprietary benzonaphthyridine of Novartis, Cambridge, MA). Env was also mixed with MF59 or with a similar ANE formulated with the TLR 4 or 7 agonists (Novartis56). Finally, Env was mixed with pIC:LC (Hiltonol; Oncovir, Washington, DC) or Abisco100 ISCOMs (Isconova AB, Stockholm, Sweden). Animals were immunized intramuscularly in the quadriceps in a homologous prime-boost manner with Env alone, or with the adjuvant formulations at 0, 4, 12 and 24 weeks. Animals were housed and cared for at the animal facility of BIOQUAL Inc. (Rockville, MD) in accordance with the American Association for Accreditation of Laboratory Animal Care standards in accredited facilities, and all animal procedures were performed according to protocols approved by the Institutional Animal Care and Use Committees of the National Institute of Allergy and Infectious Diseases, National Institutes of Health.

PBMC and plasma separation. Animals were bled into vacutainer tubes containing the anticoagulant and preservative, acidcitratedextrose. Blood was processed by Ficoll density centrifugation using Leucosep tubes (Greiner Bio-One, Monroe, NC). Plasma was then collected by aspiration above the resulting PBMC layer. PBMCs were transferred and washed with PBS before being cryopreserved in FBS DMSO. Serum was collected by bleeding animals into serum-separating

tubes that were then centrifuged.

Titre ELISA. Immulon 4HBX plates (Thermo Scientic) were coated overnight at 4 C with gp140 TV1DV2 protein diluted in carbonate buffer. Plates were then blocked with PBS FBS, and plasma was applied in serial 10-fold dilutions and

incubated at 37 C. Detection was performed at room temperature with anti-monkey IgG-horseradish peroxidase (HRP; 1:30,000 Bethyl Labs) followed by TMB substrate-chromogen (Dako) and a 2N sulfuric acid stop solution. Washing was performed between steps with PBS 0.05% Tween 20 on an ELx405 automated

plate washer (Bio-Tek Instruments). Plates were read on a SpectraMax Plus spectrophotometer (Molecular Devices), and data were analysed with the SoftMax Pro software (Molecular Devices).

HIV-1 neutralization assay. Plasma samples were heat-inactivated at 56 C for1 h and centrifuged at 16,000g for 10 min to pellet precipitated lipoproteins. Neutralizing activity of plasma supernatants was assessed against Env-pseudotyped isolates using the TZM-bl luciferase reporter gene assay as described5759. Plasma supernatants were diluted in a fourfold, 8-point dilution series, with a starting dilution of 1:20 after the supernatant was mixed with virus. Neutralization curves were t using a ve-parameter equation as described59. The amount of plasma required to inhibit infection by 50% is reported as the reciprocal 50% inhibitory dilution (ID50).

Envelope protein probes. For bulk-sorting and NGS-sequencing of Env-specic B cells from SHIVAD8-infected NHP, CK15 3-3 gp120 Env protein21 was used to make a B-cell probe. For Env-specic B cells from Env-vaccinated NHP, TV1DV2 gp140 Env protein55 Env protein was used. These proteins were biotinylated at free primary amines using sulfo-NHS-LC-biotin (Pierce). The biotinylated Env was then complexed with ExtrAvidin-phycoerythrine (PE) or ExrAvidin-FITC (Sigma) in a 1:3 molar ratio of Env:ExtrAvidin. The ExtrAvidin was divided into ve equal volumes and mixed sequentially for 20 min at 4 C with the total Env portion. For single-cell sorting and Sanger sequencing of B cells, an Avi-tagged YU2 gp140 protein60 was expressed, puried and biotinylated using the biotin ligase Bir A (Avidity, Denver, CO). Biotinylation of the YU2 gp140 protein was conrmed using enzyme-linked immunosorbent assay (ELISA).

B-cell culture. Memory B-cell culture protocol was adapted from previous reports61,62. Briey, 1 day before culture, NIH 3T3 feeder cells expressing mouse CD40 ligand were seeded into 384-well plates at B6,500 cells per well with100 U ml 1 IL-2 and 50 ng ml 1 IL-21 in IMDM with glutamax, 10% FBS and

MycoZap (Lonza). The following day, PBMCs were stained with a uorochromeconjugated gp140 probe and positive events in the IgG gate were automatically sorted at one cell per well into the 384-well plates containing feeder cells. Plates were incubated for 2 weeks at 37 C, 5% CO2. Supernatants were then applied at a 1:10 dilution to titre ELISA plates coated with anti-IgG (Rockland) or TV1 gp140. Bound IgG was detected with anti-IgG-HRP (Bethyl Laboratories) and read on a spectrophotometer.

Flow cytometry phenotyping. PBMCs were stained for viability in PBS with Aqua Dead Cell Stain (Invitrogen) followed by staining for surface markers. For B-cell enumeration from PBMCs: IgD FITC (1:33, Southern Biotech, cat. no. 2030), CD20 Ax700 PE (1:100, clone 2H7), IgM PE Cy5 (1:50, BD Pharmingen, clone G20-127), gp140-biotin/ExtrAvidin PE (Sigma), CD3 APC Cy7 (1:100, BD Pharm., clone

SP34-2), IgG APC (1:20, BD Pharm., clone G18-145), CD14 QD605 (1:300, clone M5E2), CD8 Pacic Blue (1:400, BD Pharm., clone RPA-T8); for B-cell sorting from vaccinated NHP: CD20 Ax700 PE (1:100, clone 2H7), IgM PE Cy5 (1:50, BD Pharm., clone G20-127), SA QD705, gp140-biotin/ExtrAvidin PE (Sigma), CD3 APC Cy7 (1:100, BD Pharm., clone SP34-2), IgG APC (1:20, BD Pharm., clone G18-145), CD14 QD605 (1:300, clone M5E2), CD8 Pacic Blue (1:400, BD Pharm., clone RPA-T8); for B-cell sorting from SHIV-infected NHP: CD20 Ax700 PE (1:100, clone 2H7), IgM PE Cy5 (1:50, BD Pharm.), gp120-biotin/ ExtrAvidin FITC (Sigma), gp120-biotin/ExtrAvidin PE (Sigma), CD3 APC Cy7 (1:100, BD Pharm., clone SP34-2), IgG APC (1:20, BD Pharm., clone G18-145), CD14 QD800 (1:200, Life Technologies, clone TK4) and CD8 Pacic Blue (1:400, BD Pharm., clone RPA-T8).

Events were acquired on a BD LSR II ow cytometer (BD Biosciences) and FACS data were analysed using the FlowJo software (Tree Star). Gating trees for all populations studied are described in Supplementary Fig. 2.

B-cell PCR. Antigen-specic B cells were sorted into OL-1 lysis buffer (Qiagen) and mRNA was extracted using an Oligotex Direct mRNA mini kit (Qiagen). RNA was then concentrated in a centrifugal concentrator (Millipore) and reverse transcription was performed using oligo-dT priming and Superscript II polymerase (Invitrogen). Multiplex 50 PCR was performed to amplify the IgG HC; reactions proceeded for 35 cycles using an annealing temperature of 48 C. Primers were derived from ref. 7, and include a ve- to nine-nt barcode (all barcodes have a minimum three nt difference), and XLR sequences for 454 sequencing (Supplementary Table 5). PCR products were resolved on a 1.5% agarose gel, and the appropriate bands were excised and extracted. Antigen-specic samples were combined so that each sorted B cell would be sequenced at least 10-fold, given estimates of sequencer read outputs.

RACE PCR technique. The majority of our data were generated using a standard multiplex primer PCR method employing 21 50 primers (Supplementary Table 5), which have a variety of TMs, thus potentially skewing our data set by preferential amplication with certain primers. To evaluate this, we compared our multiplex approach with the RACE technique, which uses a common 50 primer to avoid primer bias for certain VH genes. 50 RACE ready cDNA was prepared using a

SMARTer RACE cDNA Amplication Kit (Clontech) according to the manufacturers directions and as previously described63; however, the 50 CDS primer was added 30 min into the reverse transcription reaction. PCR was performed in two reactions (Supplementary Table 5); the latter used to add the XLR sequences. PCR products were resolved on a 1.5% agarose gel, and the appropriate bands were excised, extracted and sequenced by 454 pyrosequencing.

New NHP Ig reference database. Genomic DNA was extracted from skin punches from 10 Rhesus macaques (of Indian origin) using the Agilent SureSelect gDNA Extraction Kit. One to three micrograms of the gDNA was then sheared using the Covaris E-Series Sample Preparation system into 150- to 200-base-pair fragments. gDNA libraries were prepared according to the protocol provided in the SureSelect XT Target Enrichment System Kit for Illumina Multiplexed Sequencing manual. The prepped DNA libraries were then mixed with the SureSelect Capture Library, which consists of biotinylated RNA library baits that are complimentary to the target region on the prepped library. Streptavidin-coated magnetic beads were attached to the baits that had hybridized to the target region and were then magnetically separated from the unbound DNA fraction. The baits were then digested, leaving only single-stranded target DNA. A nal PCR was performed to amplify the captured DNA content and to add indices to the target DNA. The samples were then pooled in preparation for 100 base-pair, paired-end multiplexed sequencing on the Illumina HiSeq platform.

The resulting data were analysed, in part, with direct inference from 454 Ig transcripts, and annotated to form the reference database. Briey, candidate V genes were identied on the basis of identication and evaluation of these essential features of functional V genes: recombination signals, in-frame coding sequence with invariant cysteine codons, leader regions and splice sites. A gene was classied as functional if it contained all of these essential features. If a putative V gene had no stop codons in the coding region but was missing one or more of the essential features of a V gene, it was classied as an open reading frame (ORF). Genes that contained stop codons in the coding region were classied as nonfunctional. The functional genes from all monkeys were divided into families on the basis of sequence homology to human V genes. Phylogenetic analysis was used to estimate whether sequence differences should be attributed to allelism or distinct genetic loci. Because the organization of these genes on the locus could not be determined, numbers were not used in the gene names, since numbers are intended to indicate the genes position on the chromosome. Instead, letters were used. Alleles are indicated as usual by an integer following an asterisk. The database used in the analysis contained all and only the genes that were annotated as functional or an ORF. GenBank accession codes for these VH gene sequences are given in

Supplementary Table 1.

To compare this reference database with the previous rhesus database and that for humans, two phylogenetic trees were reconstructed using PhyML (Supplementary Fig. 3a,b). The human-NHP gene dendrogram was reconstructed

NATURE COMMUNICATIONS | 6:6565 | DOI: 10.1038/ncomms7565 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 11

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms7565

with TN93 substitution model; the NHP previous-new database dendrogram was built with the GTR substitution model. The optimal substitution models were estimated using MEGA5 (ref. 64); sequences were aligned using Muscle65. All available sequences and alleles were used for analysis.

454 deep sequencing. Sequencing was performed by the NIH Intramural Sequencing Center (NISC). For each sample, the amplicon library concentrations were determined with quantitative PCR (qPCR) using Library Quantication Kit

454 Titanium (Lib-L)/Universal (Kapa Biosystems) on a CFX96 Real-Time System (Bio-Rad). The libraries were then amplied on Capture Beads by emulsion PCR using a ratio of DNA molecules: DNA Capture Bead between 0.5 and 2.0 and GS Titanium emPCR Kit (Lib-L) and emPCR Breaking Kits (Roche) following the manufacturers instructions. Approximately 0.71.0 106 enriched Capture Beads

were applied to each region of a two-region PicoTiterPlate (70 75) and sequenced

on a GS FLX Sequencer (Roche-454 Life Sciences, Bradford, CT) using GS FLX Titanium Sequencing Kit for 200 cycles. Post-run signal processing, quality ltering (modied according to recommendations in Roche-454 Life Sciences Application Brief No. 001-2010) and conversion of owgram intensities to nucleotide sequence and Phred-based quality scores were performed on an off-instrument linux cluster running 454 application software version 2.6.

Bioinformatic analysis of sequencing data. A data set of high-quality unique HC sequences was collected using the following pipeline. First, raw reads were sorted according to the barcodes used for different time points or different animals, with one nucleotide mismatch allowed in the barcode. Next, reads with length o300 or 4700 bp were removed. Germline V gene and J gene were assigned for each read using BLASTN with E value cutoff 1E-10 and 1E-3, respectively66. 50 primers and 30 primers were removed on the basis of the assigned V and J gene boundary. Because single-nucleotide insertions and deletions are common in 454 pyrosequencing, we partially rectied high condence single-nucleotide insertions and deletions in the V segment using the following method: each read and its assigned germline V gene were aligned using Muscle65; single-nucleotide insertions and deletions were identied as a single gap in the aligned germline gene and read sequence, respectively. If the 50 two nucleotides and 30 two nucleotides of an insertion position mapped exactly to the germline V gene, the insertion was removed; if the 50 two nucleotides and 30 two nucleotides of a deletion position were a perfect match between read and germline V gene, the aligned germline nucleotide at the deletion position was used to rectify this deletion. However, reads were excluded if they contained more than three single-nucleotide deletions and/or insertions. Next, the correct ORF of a read was identied as the one that showed the highest amino-acid sequence identity to the germline V gene (only the three forward ORFs were compared). Furthermore, if the ratio of amino-acid identity between the read and germline V gene divided by nucleotide identity was lower than 0.7, the read was marked as having a frameshift. High-quality reads were then selected using the following criteria: no more than a 15-nucleotide truncation at the 50 terminus of the V region, no stop codon and no frameshift. Finally, a data set of unique HC sequences was collected by removing PCR and sequencing redundancy among the high-quality reads using CD-HIT with a sequence identity cutoff of0.975 and a length coverage cutoff of 0.99 (ref. 18).

To calculate divergence in the CDR and FR of each V segment, we rst aligned

amino-acid sequences of each antibody V segment and its assigned germline V gene using Muscle. The boundaries of the CDRs and FRs from germline genes were used to determine the corresponding boundaries of the sequence. The CDR and FR boundaries of germline V genes were assigned using the IMGT database67 and were manually inspected. The nucleotide-level divergence was then calculated. The amino-acid sequence alignment in each region was further used as a guide to align the sequence to its assigned germline gene at the DNA level so that divergence at the nucleotide level could be calculated. The CDR3 H3 sequence was extracted starting from the residue next to the second conserved cysteine of the V segment to the residue before the conserved HC W-G-X-G motif in the J gene where X represents any amino acid.

Long CDR H3 (Z28 aa) antibodies found in animal DCF1 were grouped into 14 lineages using the following parameters: all antibodies within a lineage have the same V and J assignments, CDR H3 identity Z80% and r40% variation in CDR

H3 length. All lineages were manually checked to exclude PCR crossover and sequencing error. The phylogenetic tree (Fig. 2g) was built using PhyML with GTR substitution model, which is the best model estimated by MEGA5. Tyrosine sulfation was predicted using the GPS-TSP programme (http://tsp.biocuckoo.org

Web End =http://tsp.biocuckoo.org). ProteinCaculator was used for charge calculations at pH 7 (http://protcalc.sourceforge.net

Web End =http:// http://protcalc.sourceforge.net

Web End =protcalc.sourceforge.net ).

SHIVAD8 viral sequencing. SGA was performed on blood plasma from eight SHIVAD8-infected animals sampled at weeks 6, 26, 54 and 99 except where indicated. For viral RNA extraction and cDNA synthesis, from each plasma specimen at least 20,000 viral RNA copies were extracted using the QIAamp Viral RNA Mini kit (Qiagen). Reverse transcription of RNA to single-stranded cDNA was performed using SuperScript III reverse transcription according to the manufacturers recommendations (Invitrogen). In brief, a cDNA reaction of 1 RT buffer,

0.5 mM of each deoxynucleoside triphosphate, 5 mM dithiothreitol, 2 U ml 1

RNaseOUT (RNase inhibitor), 10 U ml 1 of SuperScript III reverse transcription and 0.25 mM antisense primer SIVEnvR1 50-TGTAATAAATCCCTTCCAGTCCC CCC-30 was incubated at 50 C for 60 min, 55 C for 60 min and then heat-inactivated at 70 C for 15 min followed by treatment with 2 U of RNase H at 37 C for 20 min. The newly synthesized cDNA was used immediately or frozen at

80 C. For SGA of SHIV Env, a 3.3-kb fragment that includes the entire env gene was sequenced from each animal at various time points following infection using a limiting dilution PCR; therefore, only one ampliable molecule is present in each reaction. SGA was performed by serially diluting cDNA distributed among independent PCR reactions to identify a dilution where amplication occurred in o30% of the total number of reactions. PCR amplication was performed with1 PCR buffer, 2 mM MgSO4, 0.2 mM of each deoxynucleoside triphosphate,

0.2 mM of each primer and 0.025 U ml 1 Platinum Taq High Fidelity polymerase (Invitrogen) in a 20-ml reaction. First-round PCR was performed with primer SIVEnvF1 50-CCTCCCCCTCCAGGACTAGC-30 and antisense primer SIVEnvR1 under the following conditions: one cycle of 94 C for 2 min, 35 cycles at 94 C for 15 s, 55 C for 30 s and 68 C for 5 min, followed by a nal extension of 68 C for 10 min. Next, 1 ml from the rst-round PCR product was added to a second-round

PCR reaction that included the sense primer SHIVEnvF2 50-GACCTCCAGAAA ATGAAGGACCAC-30 and antisense primer SIVEnvR2 50-ATGAGACATRTCT ATTGCCAATTTGTA-30 performed under the same conditions used for rst-round PCR, but with a total of 45 cycles. Correct sized amplicons were identied using agarose gel electrophoresis and directly sequenced with second-round PCR primers and nine HIV-specic primers using Big Dye Terminator Technology (Applied Biosystems). To conrm PCR amplication from a single template, chromatograms were manually examined for multiple peaks, indicative of the presence of amplicons resulting from PCR-generated recombination events, Taq polymerase errors or multiple variant templates.

All 441 sequences were aligned using ClustalW and phylogenetic trees were constructed using the neighbour-joining method. Diversity measurements were determined at each time point for each animal using pair-wise comparisons. Similarly, divergence measurements were determined at each time point for each animal using the centre of the tree as the root. Viral sequences were aligned using Muscle with manual correction. Sequences logos for the V2 and V3 regions were generated by Weblogo68. All sequences were deposited in GenBank with accession codes KM082525 to KM082965.

To estimate the SHIV evolutionary rate within each host, we rst conrmed that no recombination took place in the ENV gene using RPD3 (ref. 69). The HKY gamma I substitution model was used, having been evaluated by MEGA5

to t the sequence substitution pattern. The evolutionary rate was then estimated using Beast with 108 Markov chain simulations70. During simulation, a relaxed log-normal molecular clock was selected for evaluating longitudinal rate variation and Bayesian Skyline plot method was used to model virus population dynamics. The rst 107 runs were used as burn in and discarded before rate estimation.

Statistical analysis and sample blinding. Data were graphed and analysed with the Prism software (GraphPad). For nonsequence data, a KruskalWallis test with a Dunns correction for multiple comparisons was used to compare means among vaccine groups, using the Env alum group as a control. A two-way analysis of

variance test with a Bonferroni correction for multiple comparisons was employed to analyse vaccine groups across multiple time points, using the pre time point as a control. For sequencing data, a mixed effect linear regression model, allowing for random effects for each animal, was employed to generate P values. Comparisons of VH gene proportions and correlations of VH genes to the neutralization status were performed using the w2-test with Yates correction or Fishers exact test. SHIV viral sequencing data were compared with multiple t-tests using a false discovery rate (Q) 1%. Investigators assembling data collected were not blinded. Plasma

and PBMC samples were analysed for viral loads, viral sequences, antibody responses and B-cell sequences in four geographically separate laboratories. Each laboratory was blinded with regard to the results from the other laboratories.

References

1. Excler, J. L., Robb, M. L. & Kim, J. H. HIV-1 vaccines: challenges and new perspectives. Hum. Vaccin. Immunother. 10, 17341746 (2014).

2. Klein, F. et al. Antibodies in HIV-1 vaccine development and therapy. Science

341, 11991204 (2013).

3. Walker, L. M. et al. Broad neutralization coverage of HIV by multiple highly potent antibodies. Nature 477, 466470 (2011).

4. Wu, X. et al. Rational design of envelope identies broadly neutralizing human monoclonal antibodies to HIV-1. Science 329, 856861 (2010).

5. Mascola, J. R. & Haynes, B. F. HIV-1 neutralizing antibodies: understanding natures pathways. Immunol. Rev. 254, 225244 (2013).

6. Klein, F. et al. Somatic mutations of the immunoglobulin framework are generally required for broad and potent HIV-1 neutralization. Cell 153, 126138 (2013).

7. Scheid, J. F. et al. Sequence and structural convergence of broad and potent HIV antibodies that mimic CD4 binding. Science 333, 16331637 (2011).8. Walker, L. M. et al. Broad and potent neutralizing antibodies from an African donor reveal a new HIV-1 vaccine target. Science 326, 285289 (2009).

12 NATURE COMMUNICATIONS | 6:6565 | DOI: 10.1038/ncomms7565 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms7565 ARTICLE

9. West, Jr. A. P., Diskin, R., Nussenzweig, M. C. & Bjorkman, P. J. Structural basis for germ-line gene usage of a potent class of antibodies targeting the CD4-binding site of HIV-1 gp120. Proc. Natl Acad. Sci. USA 109, E2083E2090 (2012).

10. Walker, L. M. et al. Rapid development of glycan-specic, broad, and potent anti-HIV-1 gp120 neutralizing antibodies in an R5 SIV/HIV chimeric virus infected macaque. Proc. Natl Acad. Sci. USA 108, 2012520129 (2011).

11. Sundling, C. et al. High-resolution denition of vaccine-elicited B cell responses against the HIV primary receptor binding site. Sci. Transl. Med. 4, 142ra196 (2012).

12. Dennison, S. M. et al. Induction of antibodies in rhesus macaques that recognize a fusion-intermediate conformation of HIV-1 gp41. PLoS ONE 6, e27824 (2011).

13. Ketloy, C. et al. Expression and function of Toll-like receptors on dendritic cells and other antigen presenting cells from non-human primates. Vet. Immunol. Immunopathol. 125, 1830 (2008).

14. Gujer, C., Sundling, C., Seder, R. A., Karlsson Hedestam, G. B. & Lore, K. Human and rhesus plasmacytoid dendritic cell and B-cell responses to Toll-like receptor stimulation. Immunology 134, 257269 (2011).

15. Morgan, C. et al. The use of nonhuman primate models in HIV vaccine development. PLoS Med. 5, e173 (2008).

16. Letvin, N. L. Nonhuman primate models for HIV vaccine development. Immunodec. Rev. 3, 247260 (1992).

17. Zhou, T. et al. Multidonor analysis reveals structural elements, genetic determinants, and maturation pathway for HIV-1 neutralization by VRC01-class antibodies. Immunity 39, 245258 (2013).

18. Zhu, J. et al. Somatic populations of PGT135-137 HIV-1-neutralizing antibodies identied by 454 pyrosequencing and bioinformatics. Front. Microbiol. 3, 315 (2012).

19. Zhu, J. et al. De novo identication of VRC01 class HIV-1-neutralizing antibodies by next-generation sequencing of B-cell transcripts. Proc. Natl Acad. Sci. USA 110, E4088E4097 (2013).

20. Rhesus Macaque Genome, S. et al. Evolutionary and biomedical insights from the rhesus macaque genome. Science 316, 222234 (2007).

21. Shingai, M. et al. Most rhesus macaques infected with the CCR5-tropic SHIV(AD8) generate cross-reactive antibodies that neutralize multiple HIV-1 strains. Proc. Natl Acad. Sci. USA 109, 1976919774 (2012).

22. Nishimura, Y. et al. Generation of the pathogenic R5-tropic simian/human immunodeciency virus SHIVAD8 by serial passaging in rhesus macaques.J. Virol. 84, 47694781 (2010).23. Giudicelli, V. et al. IMGT/LIGM-DB, the IMGT comprehensive database of immunoglobulin and T cell receptor nucleotide sequences. Nucleic Acids Res. 34, D781D784 (2006).

24. Briney, B. S., Willis, J. R., Hicar, M. D., Thomas, 2nd J. W. & Crowe, Jr. J. E. Frequency and genetic characterization of V(DD)J recombinants in the human peripheral blood antibody repertoire. Immunology 137, 5664 (2012).

25. Doria-Rose, N. A. et al. Developmental pathway for potent V1V2-directed HIV-neutralizing antibodies. Nature 509, 5562 (2014).

26. Pancera, M. et al. Structural basis for diverse N-glycan recognition by HIV-1-neutralizing V1-V2-directed antibody PG16. Nat. Struct. Mol. Biol. 20, 804813 (2013).

27. OHagan, D. T., Ott, G. S., De Gregorio, E. & Seubert, A. The mechanism of action of MF59an innately attractive adjuvant formulation. Vaccine 30, 43414348 (2012).

28. Khurana, S. et al. MF59 adjuvant enhances diversity and afnity of antibody-mediated immune response to pandemic inuenza vaccines. Sci. Transl. Med. 3, 85ra48 (2011).

29. Giannini, S. L. et al. Enhanced humoral and memory B cellular immunity using HPV16/18 L1 VLP vaccine formulated with the MPL/aluminium salt combination (AS04) compared to aluminium salt only. Vaccine 24, 59375949 (2006).

30. Paillot, R., Grimmett, H., Elton, D. & Daly, J. M. Protection, systemic IFNgamma, and antibody responses induced by an ISCOM-based vaccine against a recent equine inuenza virus in its natural host. Vet. Res. 39, 21 (2008).

31. Lefranc, M. P. et al. IMGT unique numbering for immunoglobulin and T cell receptor variable domains and Ig superfamily V-like domains. Dev. Comp. Immunol. 27, 5577 (2003).

32. Doria-Rose, N. A. et al. Breadth of human immunodeciency virus-specic neutralizing activity in sera: clustering analysis and association with clinical variables. J. Virol. 84, 16311636 (2010).

33. Piantadosi, A. et al. Breadth of neutralizing antibody response to human immunodeciency virus type 1 is affected by factors early in infection but does not inuence disease progression. J. Virol. 83, 1026910274 (2009).

34. Sather, D. N. et al. Factors associated with the development of cross-reactive neutralizing antibodies during human immunodeciency virus type 1 infection.J. Virol. 83, 757769 (2009).35. Earl, P. L. et al. Immunogenicity and protective efcacy of oligomeric human immunodeciency virus type 1 gp140. J. Virol. 75, 645653 (2001).

36. Sundling, C. et al. Soluble HIV-1 Env trimers in adjuvant elicit potent and diverse functional B cell responses in primates. J. Exp. Med. 207, 20032017 (2010).

37. Kwissa, M., Nakaya, H. I., Oluoch, H. & Pulendran, B. Distinct TLR adjuvants differentially stimulate systemic and local innate immune responses in nonhuman primates. Blood 119, 20442055 (2012).

38. Coffman, R. L., Sher, A. & Seder, R. A. Vaccine adjuvants: putting innate immunity to work. Immunity 33, 492503 (2010).

39. Boonstra, A. et al. Macrophages and myeloid dendritic cells, but not plasmacytoid dendritic cells, produce IL-10 in response to MyD88- and TRIF-dependent TLR signals, and TLR-independent signals. J. Immunol. 177, 75517558 (2006).

40. Pone, E. J. et al. BCR-signalling synergizes with TLR-signalling for induction of AID and immunoglobulin class-switching through the non-canonical NF-kappaB pathway. Nat. Commun. 3, 767 (2012).

41. Aranburu, A. et al. TLR ligation triggers somatic hypermutation in transitional B cells inducing the generation of IgM memory B cells. J. Immunol. 185, 72937301 (2010).

42. Siegrist, C. A. et al. Co-administration of CpG oligonucleotides enhances the late afnity maturation process of human anti-hepatitis B vaccine response. Vaccine 23, 615622 (2004).

43. Sanders, R. W. et al. A next-generation cleaved, soluble HIV-1 Env Trimer, BG505 SOSIP.664 gp140, expresses multiple epitopes for broadly neutralizing but not non-neutralizing antibodies. PLoS Pathog. 9, e1003618 (2013).

44. Malherbe, D. C. et al. Envelope variants circulating as initial neutralization breadth developed in two HIV-infected subjects stimulate multiclade neutralizing antibodies in rabbits. J. Virol. 88, 1294912967 (2014).

45. Liao, H. X. et al. Co-evolution of a broadly neutralizing HIV-1 antibody and founder virus. Nature 496, 469476 (2013).

46. Sanz, I. Multiple mechanisms participate in the generation of diversity of human H chain CDR3 regions. J. Immunol. 147, 17201729 (1991).

47. Briney, B. S., Willis, J. R. & Crowe, Jr. J. E. Human peripheral blood antibodies with long HCDR3s are established primarily at original recombination using a limited subset of germline genes. PLoS ONE 7, e36750 (2012).

48. Yin, L. et al. IgM repertoire biodiversity is reduced in HIV-1 infection and systemic lupus erythematosus. Front. Immunol. 4, 373 (2013).

49. Lingwood, D. et al. Structural and genetic basis for development of broadly neutralizing inuenza antibodies. Nature 489, 566570 (2012).

50. Sundling, C. et al. Single-cell and deep sequencing of IgG-switched macaque B cells reveal a diverse Ig repertoire following immunization. J. Immunol. 192, 36373644 (2014).

51. Xiao, M., Prabakaran, P., Chen, W., Kessing, B. & Dimitrov, D. S. Deep sequencing and Circos analyses of antibody libraries reveal antigen-driven selection of Ig VH genes during HIV-1 infection. Exp. Mol. Pathol. 95, 357363 (2013).

52. Breden, F. et al. Comparison of antibody repertoires produced by HIV-1 infection, other chronic and acute infections, and systemic autoimmune disease. PLoS ONE 6, e16857 (2011).

53. Endo, Y. et al. Short- and long-term clinical outcomes in rhesus monkeys inoculated with a highly pathogenic chimeric simian/human immunodeciency virus. J. Virol. 74, 69356945 (2000).

54. Gautam, R. et al. Pathogenicity and mucosal transmissibility of the R5-tropic simian/human immunodeciency virus SHIV(AD8) in rhesus macaques: implications for use in vaccine studies. J. Virol. 86, 85168526 (2012).

55. Lian, Y. et al. Evaluation of envelope vaccines derived from the South African subtype C human immunodeciency virus type 1 TV1 strain. J. Virol. 79, 1333813349 (2005).

56. Baudner, B. C. et al. MF59 emulsion is an effective delivery system for a synthetic TLR4 agonist (E6020). Pharm. Res. 26, 14771485 (2009).

57. Li, M. et al. Human immunodeciency virus type 1 env clones from acute and early subtype B infections for standardized assessments of vaccine-elicited neutralizing antibodies. J. Virol. 79, 1010810125 (2005).

58. Shu, Y. et al. Efcient protein boosting after plasmid DNA or recombinant adenovirus immunization with HIV-1 vaccine constructs. Vaccine 25, 13981408 (2007).

59. Seaman, M. S. et al. Tiered categorization of a diverse panel of HIV-1 Env pseudoviruses for assessment of neutralizing antibodies. J. Virol. 84, 14391452 (2010).

60. Scheid, J. F. et al. A method for identication of HIV gp140 binding memory B cells in human blood. J. Immunol. Methods 343, 6567 (2009).

61. Huang, J. et al. Isolation of human monoclonal antibodies from peripheral blood B cells. Nat. Protoc. 8, 19071915 (2013).

62. Huang, J. et al. Broad and potent neutralization of HIV-1 by a gp41-specic human antibody. Nature 491, 406412 (2012).

63. Douek, D. C. et al. A novel approach to the analysis of specicity, clonality, and frequency of HIV-specic T cell responses reveals a potential mechanism for control of viral escape. J. Immunol. 168, 30993104 (2002).

NATURE COMMUNICATIONS | 6:6565 | DOI: 10.1038/ncomms7565 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 13

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms7565

64. Tamura, K. et al. MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol. Biol. Evol. 28, 27312739 (2011).

65. Edgar, R. C. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32, 17921797 (2004).

66. Altschul, S. F. et al. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25, 33893402 (1997).

67. Lefranc, M. P. IMGT, the International ImMunoGeneTics Information System for Immunoinformatics: methods for querying IMGT databases, tools, and web resources in the context of immunoinformatics. Mol. Biotechnol. 40, 101111 (2008).

68. Crooks, G. E., Hon, G., Chandonia, J. M. & Brenner, S. E. WebLogo: a sequence logo generator. Genome Res. 14, 11881190 (2004).

69. Martin, D. P. et al. RDP3: a exible and fast computer program for analyzing recombination. Bioinformatics 26, 24622463 (2010).

70. Drummond, A. J. & Rambaut, A. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol. Biol. 7, 214 (2007).

Acknowledgements

We thank K. McKee, D. Ambrozak, R. Kruthers, R. Plishka, A. Buckler-White, B. Skopets and R. Petros for expert technical help. We also thank N. Longo, B. Hill, N. Doria-Rose,R. Kong, B. Haynes and M. Roederer for helpful discussions. This work was supported in part with federal funds from the National Cancer Institute; National Institutes of Health under contract HHSN261200800001E.

Author contributions

J.R.F., Z.S., L.S. and R.A.S. conceived the study and wrote the paper. N.M.V., P.M., E.D.G., S.W.B., M.S. and D.T.O. developed the vaccine formulations. J.R.F., Y.N., S.D.S., B.J.F., T.Y., R.M.-N. and R.M.L. designed and carried out experiments. B.F.K. performed

the viral sequencing. NISC performed the 454 sequencing. Z.S., Z.Z. and L.S. performed the 454 bioinformatic analyses. S.D., D.W., B.F.H. and D.D. resequenced the macaque Ig locus. A.R. and T.B.K. analysed the macaque DNA sequences. R.A.K., J.R.M., M.A.M. and R.A.S. gave conceptual advice and project oversight.

Additional information

Accession codes: The SHIV viral sequences have been deposited in the GenBank database with accession codes KM082525 to KM082965. Novel NHP Ig heavy chain sequences have been deposited in the GenBank database with accession codes KP710506 to KP710583.

Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Web End =http://www.nature.com/ http://www.nature.com/naturecommunications

Web End =naturecommunications

Competing nancial interests: N.M.V., P.M., E.D.G., S.W.B., M.S., and D.T.O.are employees of Novartis Vaccines. The other authors do not have any conicts of interest.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions

Web End =reprintsandpermissions/

How to cite this article: Francica, J.R. et al. Analysis of immunoglobulin transcripts and hypermutation following SHIVAD8 infection and protein-plus-adjuvant immunization.

Nat. Commun. 6:6565 doi: 10.1038/ncomms7565 (2015).

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/

NISC Comparative Sequencing Program

Betty Barnabas,9 Robert Blakesley,9 Gerry Bouffard,9 Shelise Brooks,9 Holly Coleman,9 Mila Dekhtyar,9 Michael Gregory,9 Xiaobin Guan,9 Jyoti Gupta,9 Joel Han,9 Shi-ling Ho,9 Richelle Legaspi,9 Quino Maduro,9 Cathy Masiello,9 Baishali Maskeri,9 Jenny McDowell,9 Casandra Montemayor,9 James Mullikin,9 Morgan Park,9 Nancy Riebow,9 Karen Schandler,9 Brian Schmidt,9 Christina Sison,9 Mal Stantripop,9 James Thomas,9 Pamela Thomas,9 Meg Vemulapalli9 & Alice Young9

9 NIH Intramural Sequencing Center, National Human Genome Research Institute, Bethesda, Maryland 20852, USA.

14 NATURE COMMUNICATIONS | 6:6565 | DOI: 10.1038/ncomms7565 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

Word count: 13471

Show less

Abstract

Translate

Developing predictive animal models to assess how candidate vaccines and infection influence the ontogenies of Envelope (Env)-specific antibodies is critical for the development of an HIV vaccine. Here we use two nonhuman primate models to compare the roles of antigen persistence, diversity and innate immunity. We perform longitudinal analyses of HIV Env-specific B-cell receptor responses to SHIV_AD8 infection and Env protein vaccination with eight different adjuvants. A subset of the SHIV_AD8 -infected animals with higher viral loads and greater Env diversity show increased neutralization associated with increasing somatic hypermutation (SHM) levels over time. The use of adjuvants results in increased ELISA titres but does not affect the mean SHM levels or CDR H3 lengths. Our study shows how the ontogeny of Env-specific B cells can be tracked, and provides insights into the requirements for developing neutralizing antibodies that should facilitate translation to human vaccine studies.

Details

Title

Analysis of immunoglobulin transcripts and hypermutation following SHIVAD8 infection and protein-plus-adjuvant immunization

Author

Francica, Joseph R; Sheng, Zizhang; Zhang, Zhenhai; Nishimura, Yoshiaki; Shingai, Masashi; Ramesh, Akshaya; Keele, Brandon F; Schmidt, Stephen D; Flynn, Barbara J; Darko, Sam; Lynch, Rebecca M; Yamamoto, Takuya; Matus-nicodemos, Rodrigo; Wolinsky, David; Nason, Martha; Valiante, Nicholas M; Malyala, Padma; De Gregorio, Ennio; Barnett, Susan W; Singh, Manmohan; O'hagan, Derek T; Koup, Richard A; Mascola, John R; Martin, Malcolm A; Kepler, Thomas B; Douek, Daniel C; Shapiro, Lawrence; Seder, Robert A

Pages

6565

Publication year

2015

Publication date

Apr 2015

Publisher

Nature Publishing Group

e-ISSN

20411723

Source type

Scholarly Journal

Language of publication

English

DOI

https://doi.org/10.1038/ncomms7565

ProQuest document ID

1672119424

Analysis of immunoglobulin transcripts and hypermutation following SHIVAD8 infection and protein-plus-adjuvant immunization

Jump to:

Full Text

Abstract

Details

Suggested sources