Hyperimmune Bovine Colostrum as a Low-Cost, Large-Scale ... · into 293T cells and pN1-AD8-140 (AD8 clone of ADA; provided by M. Martin) into HeLa cells as previously described (11).

Hyperimmune Bovine Colostrum as a Low-Cost, Large-Scale Source ofAntibodies with Broad Neutralizing Activity for HIV-1 Envelope withPotential Use in Microbicides

Marit Kramski,a Rob J. Center,a Adam K. Wheatley,a Jonathan C. Jacobson,a Marina R. Alexander,a Grant Rawlin,b andDamian F. J. Purcella

Department of Microbiology and Immunology, University of Melbourne, Parkville, Victoria, Australia,a and Immuron Ltd., North Melbourne, Victoria, Australiab

Bovine colostrum (first milk) contains very high concentrations of IgG, and on average 1 kg (500 g/liter) of IgG can be harvestedfrom each immunized cow immediately after calving. We used a modified vaccination strategy together with established produc-tion systems from the dairy food industry for the large-scale manufacture of broadly neutralizing HIV-1 IgG. This approach pro-vides a low-cost mucosal HIV preventive agent potentially suitable for a topical microbicide. Four cows were vaccinated pre-and/or postconception with recombinant HIV-1 gp140 envelope (Env) oligomers of clade B or A, B, and C. Colostrum and puri-fied colostrum IgG were assessed for cross-clade binding and neutralization against a panel of 27 Env-pseudotyped reporter vi-ruses. Vaccination elicited high anti-gp140 IgG titers in serum and colostrum with reciprocal endpoint titers of up to 1 � 105.While nonimmune colostrum showed some intrinsic neutralizing activity, colostrum from 2 cows receiving a longer-durationvaccination regimen demonstrated broad HIV-1-neutralizing activity. Colostrum-purified polyclonal IgG retained gp140 reac-tivity and neutralization activity and blocked the binding of the b12 monoclonal antibody to gp140, showing specificity for theCD4 binding site. Colostrum-derived anti-HIV antibodies offer a cost-effective option for preparing the substantial quantities ofbroadly neutralizing antibodies that would be needed in a low-cost topical combination HIV-1 microbicide.

In the absence of an effective prophylactic vaccine against humanimmunodeficiency virus type 1 (HIV-1), there is an urgent need

for female-controlled, safe, effective, and inexpensive biomedicalpreventions, such as topical microbicides for the prevention ofsexually transmitted HIV-1 infections (16, 34, 37). Despite thefailure of previous microbicide trials (14, 51, 58, 59), the antiret-roviral drug tenofovir demonstrated significant reduction in HIVacquisitions by 39% if used topically in a 1% gel (CAPRISA 004trial) (24) but did not provide any protection if used orally(VOICE trial) (36). Irrespective of differences in outcome result-ing from delivery modality and adherence, the use of antiretrovi-ral drugs as microbicides is controversial in developing countries,where affordability and accessibility to antiretroviral drugs areextremely limited. Further, the use of current therapeutic drugsfor prophylaxis may increase the selection pressure for drug-resis-tant HIV escape mutants. Maximum microbicide potency is likelyto require combination microbicides incorporating differentcomponents, such as antibodies (Abs) capable of blocking HIVinfection.

Broad and potent neutralizing Abs (NAbs), primarily raisedagainst the envelope protein (Env), have been isolated from theserum of HIV-1-infected individuals. These monoclonal NAbs(mNAbs) bind to conserved functional epitopes on the gp140Env: b12 and VRC01 targeting the CD4 binding site, 2G12 target-ing glycan, 2F5 and 4E10 targeting the membrane proximal re-gion, 447-52D targeting a CD4-induced face, and PG9/16, target-ing oligomeric V3 structures (5, 7, 13, 61, 63). mNAbs targetingthe membrane proximal regions are broad acting, with 2F5 and4E10 neutralizing 67 and 100%, respectively, of a panel of 90 di-vergent viruses (5), indicating strongly conserved epitopes. TheCD4 binding site targeting mNAb b12 neutralized 50% of theviruses in the same panel, and 2G12 neutralized 41%. 447-52Dwas less broad acting, neutralizing 19% of viruses. The VRC01

mNAb, which also targets the CD4 binding site, neutralized 91%of a 190-virus panel (63), indicating a high degree of conservationof the contact residues of the Env epitope for this antibody.

Intravenous and vaginal application of patient-derived anti-HIV-1 immunoglobulin and/or mNAbs 2G12, 2F5, and b12 canafford dose-dependent sterile protection to primates from intra-venous or vaginal challenge with chimeric simian-human immu-nodeficiency virus (SHIV) (32, 33, 45). Protection correlated withNAb concentration and in vitro neutralizing activity (21, 32, 33,40, 45, 47, 60). In contrast, a nonneutralizing variant of b12 didnot provide protection in primates (8).

High concentrations of mNAbs show promise for microbicideformulations, but they are currently prohibitively expensive toproduce in the large amounts required. An alternative source forlow-cost HIV-1-specific NAbs is bovine colostrum (BC). BC ishighly enriched with maternal immunoglobulin that is activelydrawn from the serum. The most abundant immunoglobulins inBC are IgG, with up to 50 mg/ml (primarily IgG1), but IgA andIgM are also present (up to 4 mg/ml) (55). BC also contains anti-microbial peptides and proteins, such as lactoferrin lactoperoxi-dase and lysozyme, that can stimulate innate antiviral pathwaysand adaptive immune responses (52, 55, 57). Vaccination of cowsagainst specific pathogens results in polyclonal pathogen-specific

Received 28 February 2012 Returned for modification 26 March 2012Accepted 23 May 2012

Published ahead of print 4 June 2012

Address correspondence to Damian F. J. Purcell, [email protected].

Supplemental material for this article may be found at http://aac.asm.org/.

Copyright © 2012, American Society for Microbiology. All Rights Reserved.

doi:10.1128/AAC.00453-12

4310 aac.asm.org Antimicrobial Agents and Chemotherapy p. 4310–4319 August 2012 Volume 56 Number 8

Abs in BC (hyperimmune BC [HBC]). Purified HBC Abs havesuccessfully been used for the oral treatment of dental caries (26)and a variety of gastrointestinal infections caused by pathogenicbacteria (20, 25, 46, 56) or virus (27, 35, 39).

Here, we developed a vaccination strategy in cows using oligo-meric HIV-1 gp140 antigens that induced high titers of anti-gp140Abs in HBC. Both HBC and purified HBC IgG showed broadneutralizing activity against a diverse array of Env-pseudotypedreporter viruses. This approach provides low-cost production ofHIV-1 NAbs for a potentially efficacious and affordable microbi-cide.

MATERIALS AND METHODSProduction and purification of oligomeric gp140. Oligomeric gp140 ofEnv truncated at the membrane proximal external region was created bystably transfecting pN1-UG8-140 and pN1-MW-140 (the 92UG037.8 and93MW965.26 strains, respectively; provided by F. Gao and B.H. Hahn)into 293T cells and pN1-AD8-140 (AD8 clone of ADA; provided by M.Martin) into HeLa cells as previously described (11). Secreted gp140 waspurified from cell culture supernatants by lentil-lectin affinity chromatog-raphy and size exclusion chromatography using a 16/60 Superdex 200column as previously described (10). Fractions enriched for gp140 trimerswere pooled and concentrated for use in vaccinations.

Vaccination of cows. Vaccination of four female cows was performedwith Ethics Committee approval [POCTAA (3)003 A04]. Each vaccina-tion consisted of 100 �g of purified gp140 oligomers in adjuvant (Mon-tanide ISO 206; Seppic, France) injected intramuscularly into the flank(vaccination regime) (Fig. 1A). Primary vaccination took place either be-fore conception (group NP) or during the second trimester of pregnancy(group P) followed by three or two boost vaccinations after conception,respectively. Final vaccination was given approximately 4 weeks beforegiving birth. Within each group, one animal was vaccinated either withclade B (AD8) only (B) or with equal amounts of clade A (UG8), B (AD8),and C (MW) gp140 (33.33 �g each) (trimix). Blood samples were col-lected into serum tubes (Venoject glass tubes; Venoject) at regular inter-vals, avoiding the period 4 weeks before birth to reduce the risk of mis-carriage. Blood was centrifuged at 1,300 � g for 12 min at roomtemperature. Serum was removed and stored in aliquots at �80°C. HBCwas collected within 6 h postpartum by milking and was stored at �20°Cprior to use.

Preparation of colostrum and purified colostrum IgG. BC was defat-ted and pasteurized by centrifugation at 10,000 � g for 30 min at 4°C,incubation at 63°C for 30 min, and further centrifugation at 10,000 � g for10 min. For IgG purification, casein-depleted colostral whey was preparedby adjusting the pH to 4.6 with an equal volume of 0.2 M sodium acetatesolution (pH 4.0) while mixing at 37°C for 2 h followed by cooling andcentrifugation at 10,000 � g for 30 min. Colostral whey was adjusted topH 6.6 and then dialyzed against phosphate-buffered saline (PBS) using a30-kDa cutoff ultrafiltration membrane (Amicon Ultra, 15 ml; Millipore).IgG was purified by protein G Sepharose chromatography (GE Health-care). After elution with 50 mM citrate, pH 2.6, IgG-containing fractionswere neutralized to pH 7.0 by the addition of 1 M Tris, pH 8.0, and thendialyzed against PBS. Purified IgG was filter sterilized, and concentrationswere measured by reading absorbance at 280 nm.

Anti-bovine IgG ELISA. gp140-specific IgG binding titers in sera andwhole HBC and purified HBC IgGs were measured by enzyme-linkedimmunosorbent assay (ELISA). Purified gp140 in 100 �l/well coating buf-fer (pH 9.8; 2 mM Tris, 10 mM NaCl) was coated onto 96-well polyvinylflat-bottom plates (Pathtech, Australia) at 100 ng/well overnight at 4°C.All subsequent steps were performed at room temperature. Wells werewashed sequentially in PBST (PBS-0.1% Tween 20) and PBS and thenblocked with 5% casein in PBS for 2 h. Wells were washed as describedabove, followed by the addition of half-log dilutions of sera or HBC or2-fold dilutions of purified HBC IgG in 5% casein. After 4 h of incubation,

wells were washed and horseradish peroxidase (HRP)-conjugated rabbitanti-bovine IgG antibody (Sigma) was added and incubated for 1 h. Colorreactions were developed using 3,3=-5,5=-tetramethylbenzidine (TMB),and absorbance was measured at 405 nm against a reference of 492 nm. Apositive IgG signal was defined as an optical density (OD) of �2-foldhigher than that obtained with either serum samples taken prior to vacci-nation, whole nonimmune BC, or IgG purified from nonimmune BC, asappropriate. IgG titers are presented as averages from 3 repeats.

Competition ELISA. To determine purified HBC IgG-specific bind-ing sites upon the gp140 Env protein, a competition ELISA was carried outwith human MAbs obtained through the AIDS Research and ReferenceReagent Program, Division of AIDS, NIAID, NIH: 2G12 and 2F5 (bothprovided by H. Katinger), 447-52D (provided by S. Zolla-Pazner), VRC01(provided by John Mascola), and b12 (provided by D. Burton and C.Barbas). Briefly, 96-well polystyrene plates were coated with gp140 (AD8),washed as described above, and then blocked for 1 h with 5% skim milkpowder in PBST. A total of 1 �g/�l of purified HBC IgG or a 1:100 dilutionof whole HBC diluted in 5% skim milk powder in PBST was added andincubated for 2 h and washed as described above before serial dilutions ofthe MAbs were added. After 2 h of incubation, wells were washed andbinding of MAbs to gp140 was detected using HRP-conjugated goat anti-human IgG antibody (Dako). Color reactions were developed using TMBsubstrate and stopped with 1 M H2SO4. Absorbance was measured at 450nm against a reference of 690 nm.

SDS-PAGE and Western blotting. Protein samples were resolved by8% reducing SDS-PAGE and transferred to a polyvinylidene difluoride(PVDF) membrane, blocked with 5% casein in PBS overnight at 4°C, andprobed with sera or HBC for 4 h at room temperature. Bound IgG wasdetected using HRP-conjugated rabbit anti-bovine IgG antibody (Sigma)and Supersignal West Pico chemiluminescent substrate (Thermo Scien-tific). Bands were visualized using Kodak 4000MM Image Station and IMsoftware (Carestream Health). As a loading control, all blots were strippedusing RestoreTM Plus Western blot stripping buffer (Thermo Scientific)according to the manufacturer’s protocol, washed with PBS and PBST,and blocked in 5% skim milk powder in PBST before being reprobed withpooled human HIV-positive serum (clade B) for 2 h. Bound human IgGwas detected using goat anti-human IgG HRP-conjugated antibody (In-vitrogen).

Production of EGFP Env-pseudotyped HIV-1 viruses. Env-pseu-dotyped HIV-1 particles were produced by cotransfecting HeLa cells withan EGFP-expressing proviral reporter plasmid that terminates Env after26 amino acids (pNL-4.3�envNef-EGFP) (1) and a vector expressing thefull-length Env protein of the strain of interest. The indicated HIV-1 Env-encoding plasmids were used to produce EGFP Env-pseudotyped viruses:AD8 (pCMV-AD8; prepared from pAD8 provided by M. Martin); MN(pSVIII-MN; provided by J. Sodroski); SF162 (pCAGGS-SF162; providedby L. Stamatatos and C. Cheng-Mayer); NL4.3 (pCMV-NL4.3; preparedfrom pNL4.3 provided by M. Martin); 89.6 (pSVIII-89.6; J. Sodroski); 966(pCRII-93TH966-89; provided by B. H. Hahn); plasmids from theclade B and C NIH reference panel (AIDS Research and ReferenceReagent Program); Du156.12 and Du172.17 (D. Montefiori, F. Gao, S.Abdool Karim, and G. Ramjee.); ZM197M.PB7, ZM214M.PL15,ZM233M.PB6, and ZM249M.PL1 (B. H. Hahn, Y. Li, and J. F. Salazar-Gonzalez); ZM53M.PB12, ZM109F.PB4, and ZM135M.PL10a (E.Hunter and C. Derdeyn); CAP45.2.00.G3 and CAP210.2.00E8 (L.Morris, K. Mlisana, and D. Montefiori); 6535, clone 3, PVO, clone 4,QH0692, clone 42, TRO, clone 11, AC10.0, clone 29, and SC 422661.8(B. H. Hahn and J. F. Salazar-Gonzalez); and pWITO4160 clone 33,pTRJO4551 clone 58, pREJO4541 clone 67, and pRHPA4259 clone 7(B. H. Hahn, X. Wei, and G. M. Shaw).

EGFP neutralization assay. The neutralization assay was performedas previously described (11). Briefly, EGFP Env-pseudotyped viruses wereincubated in a total volume of 30 �l for 1 h at 37°C with various concen-trations of HBC (defatted and pasteurized) or purified HBC IgG. CF2th/CD4/CCR5/CXCR4 canine cells (17, 18) that express only the stable-

Bovine Colostrum HIV-Neutralizing Antibodies

August 2012 Volume 56 Number 8 aac.asm.org 4311

Kramski et al.

4312 aac.asm.org Antimicrobial Agents and Chemotherapy

transfected human receptors CD4, CCR5, and CXCR4 on the cell surfacewere used as target cells to exclude other human receptors that mightcomplicate the analysis from any antibody reactive to any possible humancontaminants in the gp140 vaccine. After incubation of virus and Ab,CF2th/CD4/CCR5/CXCR4 cells (2 � 104/well) were added followed byspinoculation (42) at 1,200 � g for 2 h at room temperature. Residualvirus and antibody were removed, and 200 �l of fresh medium was addedto the cells, which were cultured for an additional 2 days. Target cells wereanalyzed for infection (positive for EGFP expression) by flow cytometry(FACSort; Becton, Dickinson). The percent neutralization was calculatedas [1 � (virus � HBC or purified HBC IgG/virus � media)] � 100 andpresented as the mean from duplicate samples. Our assay was previouslyvalidated with broadly NAbs 2F5, 2G12, and 4E10, giving a maximum of75% neutralization at a concentration of 0.5 �g/�l (data not shown).

RESULTSVaccination elicits high IgG binding titers to gp140 in serumand colostrum. All vaccinated cows seroconverted 2 weeks afterthe first boost vaccination (week nine), with mean anti-AD8gp140 (clade B) IgG reciprocal serum endpoint titers of 1 � 102.75

(P-B), 1 � 101.5 (P-trimix), 1 � 104.5 (NP-B), and 1 � 104.75

(NP-trimix) (Fig. 1B). A second boost vaccination (week 12) didnot lead to an increase in serum reciprocal IgG endpoint titers.Overall, serum IgG specific to clade B (AD8) was detectable in allcows but was highest for NP-B and NP-trimix. Serum IgG specificto clade A (UG8) and clade C (MW) was detected only in the NPgroup of cows and did not require vaccination with the trimix-clade vaccine (Fig. 1C), suggesting reactivities to conservedepitopes.

Three out of four cows had readily detectable anti-AD8 gp140IgG titers of 1 � 103.5 (P-trimix), 1 � 105 (NP-B), and 1 � 104.5

(NP-trimix) in HBC (Fig. 1D). In agreement with serum titers andbreadth of reactivity, NP-B and NP-trimix HBC IgG endpointtiters against clade A and clade C gp140 were up to 1 � 104.75 and1 � 103.5, respectively. The gp140-specific IgG titer of HBC fromP-B (1 � 102) was very low and therefore was not included infurther characterizations.

Total IgG was purified from HBC to an overall purity of �90%(data not shown) and tested for binding to gp140. Purified HBCIgG from P-trimix showed weak binding to clade A, B, and Cgp140 (Fig. 1E). In contrast, purified HBC IgG from NP-B andNP-trimix showed strong and equivalent reactivities to gp140 ofall three clades, resulting in low endpoint concentrations between0.78 �g/�l and 6.26 �g/�l. No binding was detected for a nonim-mune BC IgG control.

Serum and colostrum IgG are specific to gp140 of clade A, B,and C. Serum and HBC from both NP cows demonstrated strongand specific binding to secreted clade A, clade B, and clade Cgp140 by immunoblotting (Fig. 2). No specific IgG responses wereraised to proteins of the producer cell line (mock-transfected cellsand empty vector pN1 control). No gp140 reactivity was observedfor serum and HBC at the concentrations tested from animal P-

trimix or prevaccination sera from all cows (data not shown).Nonimmune BC also had no gp140 reactivity (see Fig. S1 in thesupplemental material).

HBC mediates broad neutralization of HIV-1-pseudotypedviruses. We assessed HIV-1 neutralization breadth of one nonim-mune BC and three HBCs using a panel of 27 Env-pseudotypedHIV-1 viruses with a broad range of neutralization sensitivitiesfrom clades B (n � 15), C (n � 11), and the circulating recombi-nant form (CRF) A/E (n � 1). In the neutralization assay used,typically 30% to 80% neutralization is seen using sensitive Envstrains and patient-derived sera at a similar dilution range (datanot shown). Therefore, neutralization was divided in 3 categories:absent/weak (0 to 33%), moderate (33 to 66%), and strong(�66%) neutralization. A summary of the neutralization profilesof HBC is given in Table 1. We found that nonimmune BC showedintrinsic neutralization with a baseline activity between 13.7% and64%. This neutralizing activity may result from high concentra-tions of low-affinity antibodies against Env that could dissociate inother assays (e.g., ELISA) that use lower concentrations of anti-bodies. Surprisingly HBC from NP-trimix, with the highest gp140IgG binding titer, was less potent than HBC from animal P-trimixor NP-B which showed strong to moderate neutralization for al-most all tested viruses ranging from 33.7 to 89.7% (P-trimix) and42.8 to 94.6% (NP-B) neutralization. HBC from NP-B not onlyhad the most potent neutralization activity but also exhibitedstrong breadth, neutralizing 1/1 CRF AE and 15/15 clade B and9/11 clade C strains. Thus, 2 of 3 HBC samples showed broadneutralizing activity against sensitive to moderately resistantHIV-1 Env clones from different clades. Total IgG concentrationsin HBC, however, varied from 57.7 �g/�l (P-trimix), 60 �g/�l(NP-B), 26.6 �g/�l (NP-trimix), and 20 �g/�l (nonimmune),and this may in part explain the different neutralization profiles.

Purified HBC IgG retains broad neutralizing activity. As BCcontains different concentrations of IgG and intrinsic antibacte-rial and antiviral proteins and peptides, we investigated if broadHIV-1-neutralizing activity was also mediated by IgG. Neutraliz-ing activity of purified HBC IgG (0.01 to 1 �g/�l) was testedagainst a representative panel of Env-pseudotyped HIV-1 strains,including 1 clade CRF AE strain, 5 common lab strains of clade B,and 2 clade B and 2 clade C NIH reference Env strains describedabove.

Purified HBC IgG from NP-B showed the highest activity withstrong to moderate neutralization observed for all viruses tested(n � 10) at IgG concentrations of 1 to 0.1 �g/�l (Fig. 3A). IgGfrom NP-trimix also displayed strong to moderate neutralizationfor all viruses at concentrations of 1 �g/�l and 0.5 �g/�l and 5/10viruses at concentrations of 0.1 �g/�l. IgG from P-trimix was lesspotent and showed moderate neutralization for 10/10 and 8/10viruses at a concentration of 1 �g/�l and 0.5 �g/�l, respectively.Neutralization was weak or absent at IgG concentrations of less

FIG 1 Vaccination regime and gp140-specific IgG titers. (A) Primary vaccination was initiated during the second trimester of pregnancy (P cows, n � 2) or priorto conception (NP cows, n � 2). Cows received 100 �g of gp140 Env oligomers consisting of either clade B alone or a mixture of clade A, B, and C (trimix). (B)Reciprocal Env-specific IgG endpoint titers in serum against gp140 AD8 over the time of vaccination. Arrows indicate additional boost vaccinations; results showthe means from 3 repeats. (C) Serum IgG titers 9 weeks after primary vaccination. (D) Colostrum IgG titers specific against clade A (UG8), B (AD8), and C (MW)gp140 Env; results show the means from 3 repeats; error bars represent the standard deviation (SD). (E) gp140-specific binding of purified colostrum IgG againstclade A, B, and C gp140 Env. Endpoint concentrations are listed in the table below the graphs. Reciprocal IgG endpoint titers were determined by ELISA againstthe respective gp140 Env proteins, and a positive signal was defined as an OD of �2-fold of preimmunization samples, nonimmune colostrum, or nonimmunepurified IgG.

Bovine Colostrum HIV-Neutralizing Antibodies

August 2012 Volume 56 Number 8 aac.asm.org 4313

than 0.1 �g/�l except for NP-B, which displayed moderate neu-tralization at this concentration. No viral neutralization was ob-served using equivalent concentrations of control IgG purifiedfrom human or bovine sera (Sigma). The breadth of neutraliza-tion for each IgG sample at a concentration of 0.5 �g/�l is shownin Fig. 3B.

We tested BC-purified IgG from an increased number (n � 6)of HIV-1 Env naïve animals and detected a wide range of intrinsicneutralization potency. Figure 3B shows the highest- and lowest-neutralizing samples of nonimmune BC IgG at concentrations of0.5 �g/�l. Measured neutralizing activity of these BC-purifiedIgGs were similar to the neutralizing activity of correspondingwhole BC samples (see Fig. S2 in the supplemental material).

Purified HBC IgG shows activity against the CD4 bindingsite of Env. To define potential neutralizing sites of Env targetedby colostrum IgG from NP-B and NP-trimix, we performed com-petition ELISAs with human MAbs 2G12, 2F5, 44752D, VRC01,and b12. Purified HBC IgG from NP-B and NP-trimix specificallyinhibited the binding of b12 and VRC01 to gp140 with similarpotencies but did not inhibit binding of MAb 2G12 (Fig. 4) or 2F5or 44752D (data not shown). This b12/VRC01-blocking activitywas not observed using high concentrations of IgG from the non-

immune control, suggesting that vaccination of cows with gp140can induce Abs which target the CD4 binding site of Env. Com-petition with MAb b12 and VRC01 binding was not observed forthe same concentrations of purified HBC IgG from P-trimix (datanot shown).

DISCUSSION

A vaccination or immunotherapy strategy inducing broadly neu-tralizing Abs capable of preventing transmission of HIV-1 is con-sidered the best method of curtailing the increasing costs of HIV-1treatment; however, such a vaccine has not yet been produced. Asan alternative, we aimed to produce broadly neutralizing anti-HIV Abs at a sufficiently low cost that they could be coformulatedinto a topical microbicide as an HIV prevention strategy. Here, wesuccessfully induced high titers of gp140-specific polyclonal Abswith broad HIV-1-neutralizing activity in HBC by vaccinatingpregnant cows with purified gp140 oligomers.

Vaccination with gp140 from a single clade (AD8) was suffi-cient to induce IgG with robust cross-clade binding and neutral-ization. We did not observe any increased neutralization breadthby vaccination with multiclade gp140 antigens. In contrast withother studies (2, 11, 19, 53), we found broad neutralizing activity

FIG 2 Binding specificity of antibodies determined by Western blotting. Specificity of serum (A and B) and colostrum (C and D) IgG of NP cows against secretedgp140 clade A (UG8), B (AD8), and C (MW) Env (black arrows). Cell culture supernatant from mock-transfected cells and cells transfected with vector backbonepN1 was used as a control. Proteins were separated on an 8% SDS gel under reducing conditions, and transferred proteins were blotted against a serum dilutionof 1:100 or a colostrum dilution of 1:10,000 and rabbit anti-bovine IgG HRP-conjugated antibody. Efficient loading of Env on the membranes was assessed byreprobing with pooled HIV-positive sera (clade B) (1:5,000) and goat anti-human IgG HRP-conjugated antibody (1:1,000).

Kramski et al.

4314 aac.asm.org Antimicrobial Agents and Chemotherapy

TABLE 1 HIV-1 neutralization profile of whole HBC

a Reduction of EGFP-expressing CF23 canine thymocyte cells after infection with Env-pseudotyped viruses preincubated with defatted and pasteurized colostrum was detected induplicate wells. Numbers represent percent neutralization for a 1:16 dilution against the indicated Env-pseudotyped viruses, including common lab strains and the NIH clade B andC reference panels. The percentage of neutralization was calculated as follows: [1 � (virus � colostrum or antibody/virus � media)] � 100. Neutralization is presented as themeans SD from duplicate samples. Results are shown for one representative experiment of 3 repeats.b EGFP Env-pseudotyped virus with respective envelope protein.c Neutralization sensitivity for NIH reference Env clones are given according to Li et al. (28, 29) and include genetic and antigenic diversity while avoiding strains that are unusually sensitiveor resistant to neutralization. Pseudovirus sensitivities are defined as follows: sensitive, readily neutralized by the majority of patient serum and monoclonal antibodies; moderately sensitive/resistant, neutralized by a small proportion of patient serum and monoclonal antibodies; resistant, neutralized by very few patient serum and monoclonal antibodies.

August 2012 Volume 56 Number 8 aac.asm.org 4315

from HBC and more importantly from purified HBC IgG in apseudovirus-based assay that included tier 1 and tier 2 Env andprimary isolates from heterologous clade viral isolates. Thebreadth of neutralization seen with our purified HBC IgG, includ-ing primary Env clones derived from sexually acquired acute/earlyinfections, demonstrates clear potential for inclusion of HBC-de-rived polyclonal antibodies in candidate microbicides aiming toprevent sexually acquired HIV.

The highest serum and HBC titers of gp140-specific Abs withbroad neutralizing activity were induced in cows vaccinated be-fore conception (NP). The extended duration and one extra boostof the NP vaccination protocol may favor IgG affinity maturationthat has been associated with effective HIV neutralization (64).Further improvements in the magnitude of vaccine-elicited im-mune responses might be achieved by optimization of the doseand adjuvant used with the gp140 immunogens. Previous immu-nogenicity studies of gp140 trimers in mice, rabbits, guinea pigs,and primates have used protein concentrations ranging from 5 �gand 500 �g (2, 6, 11, 23, 41). Considering body weight, vaccina-tion in the current study with only 100 �g of gp140 may be sub-

optimal but still produced high levels of neutralizing antibodies.Although the number of animals used in this study is limited,results suggest that for future vaccination studies, vaccinationshould be carried out prior to conception followed by multipleboost vaccinations during pregnancy with gp140 oligomer. Wefound that AD8 gp140 alone was sufficient. AD8 expresses well incell culture, allowing for production of sufficient amounts neces-sary for vaccination. Abundant antigen is important, as vaccinat-ing with an increased amount of gp140 in conjunction with at least10 cows may be a strategy to reduce variability and improve con-sistency. Another effective way to reduce biological variability is toscreen the colostrum for each animal prior to pooling or in large-scale manufacture to have at least 100 cows per production run.This is a standard method of Immuron Ltd. to achieve consistencyfor the production of their oral colostrum IgG product to preventtraveler’s diarrhea (Travelan). In that case, non- or low respondersto vaccination would have a limited impact on the final product.

Polyclonal IgG, including the HBC IgG from this study, havespecificity for many different epitopes and generally contain onlya proportion of IgG specific to the vaccine antigen, such as gp140in our case. Therefore, a relatively small proportion of HBC IgG islikely to be directly involved in neutralization. This explains whythe HBC IgG here typically had lower potency than the previouslycharacterized human broadly mNAbs, despite its broad neutraliz-ing activity. Protocols that further refine the vaccination for poly-clonal HIV-neutralizing IgG may improve this yield. Neverthe-less, substantial amounts of antibody could be prepared with thedescribed protocols by harnessing commercial dairies that rou-tinely process milk products for human consumption. This vastprotein production system could produce unprecedented quanti-ties of NAb for mucosal applications at a relatively low cost.

We observed a relationship between IgG concentration andHIV neutralization from whole HBC and purified HBC IgG. Forneutralization measurements of purified HBC IgG, IgG concen-trations were equalized to better reflect the IgG-mediated neutral-izing activity. The presence and concentrations of antimicrobialcomponents within BC, including lactoferrin, lysozyme, or lac-toperoxidase, vary considerably among healthy cows, are highlydependent on the exact timing of BC collection, and could there-fore significantly influence the neutralization profile of HBC (12,15, 38, 55). We mapped the Ab-binding sites within gp140 con-ferring the neutralizing activities in purified HBC IgG from bothNP cows using competition with MAbs with defined epitopes.Purified HBC IgG was able to block the binding of the MAb b12and VRC01, which target the CD4 binding site of Env and pre-vents CD4 attachment (65). The functionally conserved CD4binding region of gp120 is a crucial Ab epitope and has previouslybeen linked with exceptionally broad neutralization (63). More-over, defined neutralizing activity within the plasma of HIV-in-fected patients is most commonly directed toward to the CD4binding site (4, 30). Although binding to the CD4 binding sitedoes not necessarily guarantee broad neutralization, our b12blocking and VRC01 data demonstrates binding of purified HBCIgG to the CD4 binding site, which may explain, at least in part, thebroad neutralizing activity of purified HBC IgG. Molecular char-acterization of b12 and several other patient-derived broadlymNAbs has revealed that these Abs have an unusually long com-plementarity-determining region (CDR) of the antibody heavychain 3 (H3) (18 residues and longer) that is vital for Ab-neutral-izing activity (7, 9, 22, 43, 48, 49, 54, 61, 66, 67). Extended CDR H3

FIG 3 Neutralization profile of purified colostrum IgG against a representa-tive panel of Env-pseudotyped viruses with different neutralization sensitivi-ties. Colostrum purified IgG was used at 0.01 to 1 �g/�l. (A) Average neutral-ization of all viruses tested plotted against the respective purified colostrumIgG concentration. IgGs purified from human and bovine HIV-negative se-rum (Sigma) were included as controls for nonspecific neutralization. (B)Representative neutralizing profile for all IgG samples against the 10 viruses foran IgG concentration of 0.5 �g/�l. NImax and NImin represent a strong neu-tralizing and nonneutralizing nonimmune colostrum sample, respectively(out of 6 viruses tested). Dotted lines indicate the moderate and strong neu-tralization cutoffs and 0% neutralization. Results are expressed as mean per-centages of neutralization from two independent experiments; percentage ofneutralization was calculated as [1 � (virus � colostrum or antibody/virus �media)] � 100. Neutralization is shown as means from duplicate samples forone representative experiment of 3 repeats.

Kramski et al.

4316 aac.asm.org Antimicrobial Agents and Chemotherapy

regions represent an important alteration of Ab structure thatfacilitates the insertion of binding domains into recessed cavitieswithin gp120. The bovine heavy chain repertoire utilizes Ig VHgenes from only a single family, and significant Ab diversity isgenerated by extensive somatic mutation (3, 31). The bovine CDRH3 is particularly variable and ranges in length from 13 to 28amino acids, with the majority of Abs containing a CDR H3 inexcess of 21 amino acids (50). This is longer than has been re-ported for humans (62) and may provide one potential explana-tion for the exceptional neutralization breadth of HBC IgG.

Our results demonstrate that vaccination of cows with purified

gp140 oligomers is an effective strategy for the production of highconcentrations of HIV-1-specific Abs with gp140 binding andbroad neutralizing activity. Although Abs can be produced usingrecombinant biological manufacturing techniques, anti-HIV IgGfrom HBC represents an inexpensive source of very large quanti-ties of Abs using established dairy manufacturing protocols thatare potentially transferrable to developing countries.

Further clinical development of the HIV-1-specific HBC IgGwill require thorough confirmation of the neutralizing activity byan external reference laboratory with standardized virus produc-tion and assay conditions and different cells, as it was recently

FIG 4 Mapping of Env binding epitopes. Colostrum IgG competes with human neutralizing MAb b12 and VRC01 for binding at the gp140 CD4 binding site.(A) Cartoon of an HIV gp140 envelope protein (V2/3 loop in red, V3 loop in gray, and gp41 in orange) with the respected binding sites of the monoclonalantibodies used. Competition ELISAs were performed by titrating the human neutralizing MAb b12, VRC01, and 2G12 against a constant background of 100 �gpurified colostrum IgG (B to E) or a 1:100 dilution of whole hyperimmune colostrum (F). The ability of b12, VRC01, and 2G12 to bind to gp140 Env (AD8) inthe presence or absence of colostrum IgG was detected by anti-human IgG HRP-conjugated antibody.

Bovine Colostrum HIV-Neutralizing Antibodies

August 2012 Volume 56 Number 8 aac.asm.org 4317

shown that neutralizing activity of polyclonal Ab preparations ifused to prevent HIV infection is essential (44). In addition, in vivoexperiments confirming the safety and antiviral potential of HIV-1-specific HBC IgG are warranted. The ultimate goal would be tocombine HIV-1-specific IgG with other suitable components,such as antiretroviral drugs, or compounds that are currently un-der evaluation in topical microbicide formulations in human clin-ical trials.

ACKNOWLEDGMENTS

Funding for these studies came from a grant to DFP from the AustralianCenters for HIV and Hepatitis Virology (ACH2) and from NHMRC Aus-tralia Program grant number 510488.

We thank Victor Wong, Shahan Campbell, Carly Siebentritt, and JaneHoward for technical assistance and Paul Gorry, Luk Rombauts, Roy Rob-ins-Browne, and Gottfried Lichti for helpful discussions.

REFERENCES1. Alexander MR, Wheatley AK, Center RJ, Purcell DF. 2010. Efficient

transcription through an intron requires the binding of an Sm-type U1snRNP with intact stem loop II to the splice donor. Nucleic Acids Res.38:3041–3053.

2. Beddows S, et al. 2007. A comparative immunogenicity study in rabbits ofdisulfide-stabilized, proteolytically cleaved, soluble trimeric human im-munodeficiency virus type 1 gp140, trimeric cleavage-defective gp140 andmonomeric gp120. Virology 360:329 –340.

3. Berens SJ, Wylie DE, Lopez OJ. 1997. Use of a single VH family and longCDR3s in the variable region of cattle Ig heavy chains. Int. Immunol.9:189 –199.

4. Binley JM, et al. 2008. Profiling the specificity of neutralizing antibodiesin a large panel of plasmas from patients chronically infected with humanimmunodeficiency virus type 1 subtypes B and C. J. Virol. 82:11651–11668.

5. Binley JM, et al. 2004. Comprehensive cross-clade neutralization analysisof a panel of anti-human immunodeficiency virus type 1 monoclonalantibodies. J. Virol. 78:13232–13252.

6. Blish CA, et al. 2010. Comparative immunogenicity of subtype A humanimmunodeficiency virus type 1 envelope exhibiting differential exposureof conserved neutralization epitopes. J. Virol. 84:2573–2584.

7. Burton DR, et al. 2004. HIV vaccine design and the neutralizing antibodyproblem. Nat. Immunol. 5:233–236.

8. Burton DR, et al. 2011. Limited or no protection by weakly or nonneu-tralizing antibodies against vaginal SHIV challenge of macaques com-pared with a strongly neutralizing antibody. Proc. Natl. Acad. Sci. U. S. A.108:11181–11186.

9. Cardoso RM, et al. 2005. Broadly neutralizing anti-HIV antibody 4E10recognizes a helical conformation of a highly conserved fusion-associatedmotif in gp41. Immunity 22:163–173.

10. Center RJ, Lebowitz J, Leapman RD, Moss B. 2004. Promoting tri-merization of soluble human immunodeficiency virus type 1 (HIV-1) Envthrough the use of HIV-1/simian immunodeficiency virus chimeras. J.Virol. 78:2265–2276.

11. Center RJ, et al. 2009. Induction of HIV-1 subtype B and AE-specificneutralizing antibodies in mice and macaques with DNA prime and re-combinant gp140 protein boost regimens. Vaccine 27:6605– 6612.

12. Cheng JB, et al. 2008. Factors affecting the lactoferrin concentration inbovine milk. J. Dairy Sci. 91:970 –976.

13. Corti D, et al. 2010. Analysis of memory B cell responses and isolation ofnovel monoclonal antibodies with neutralizing breadth from HIV-1-infected individuals. PLoS One 5:e8805. doi:10.1371/journal.pone.0008805.

14. Feldblum PJ, et al. 2008. SAVVY vaginal gel (C31G) for prevention ofHIV infection: a randomized controlled trial in Nigeria. PLoS One3:e1474. doi:10.1371/journal.pone.0001474.

15. Fonteh FA, Grandison AS, Lewis MJ. 2002. Variations of lactoperoxidaseactivity and thiocyanate content in cows’ and goats’ milk throughout lac-tation. J. Dairy Res. 69:401– 409.

16. Garg AB, Nuttall J, Romano J. 2009. The future of HIV microbicides:challenges and opportunities. Antivir. Chem. Chemother. 19:143–150.

17. Gray L, et al. 2006. Genetic and functional analysis of R5X4 human

immunodeficiency virus type 1 envelope glycoproteins derived from twoindividuals homozygous for the CCR5delta32 allele. J. Virol. 80:3684 –3691.

18. Gray L, et al. 2005. Uncoupling coreceptor usage of human immunode-ficiency virus type 1 (HIV-1) from macrophage tropism reveals biologicalproperties of CCR5-restricted HIV-1 isolates from patients with acquiredimmunodeficiency syndrome. Virology 337:384 –398.

19. Grundner C, et al. 2005. Analysis of the neutralizing antibody responseelicited in rabbits by repeated inoculation with trimeric HIV-1 envelopeglycoproteins. Virology 331:33– 46.

20. Hammarström L, Weiner CK. 2008. Targeted antibodies in dairy-basedproducts. Adv. Exp. Med. Biol. 606:321–343.

21. Hessell A, et al. 2009. Broadly neutralizing human Anti-HIV antibody2G12 is effective in protection against mucosal SHIV challenge even at lowserum neutralizing titers. PLoS Pathog. 5:e1000433. doi:10.1371/journal.ppat.1000433.

22. Julien JP, et al. 2010. Ablation of the complementarity-determining re-gion H3 apex of the anti-HIV-1 broadly neutralizing antibody 2F5 abro-gates neutralizing capacity without affecting core epitope binding. J. Virol.84:4136 – 4147.

23. Kang YK, et al. 2009. Structural and immunogenicity studies of a cleaved,stabilized envelope trimer derived from subtype A HIV-1. Vaccine 27:5120 –5132.

24. Karim QA, et al. 2010. Effectiveness and safety of tenofovir gel, an anti-retroviral microbicide, for the prevention of HIV infection in women.Science 329:1168 –1174.

25. Kelly GS. 2003. Bovine colostrums: a review of clinical uses. Alt. Med. Rev.8:378 –394.

26. Koga T, Oho T, Shimazaki Y, Nakano Y. 2002. Immunization againstdental caries. Vaccine 20:2027–2044.

27. Korhonen H, Marnila P, Gill HS. 2000. Bovine milk antibodies forhealth. Br. J. Nutr. 84(Suppl 1):S135–S146.

28. Li M, et al. 2005. Human immunodeficiency virus type 1 env clones fromacute and early subtype B infections for standardized assessments of vac-cine-elicited neutralizing antibodies. J. Virol. 79:10108 –10125.

29. Li M, et al. 2006. Genetic and neutralization properties of subtype Chuman immunodeficiency virus type 1 molecular env clones from acuteand early heterosexually acquired infections in Southern Africa. J. Virol.80:11776 –11790.

30. Li Y, et al. 2007. Broad HIV-1 neutralization mediated by CD4-bindingsite antibodies. Nat. Med. 13:1032–1034.

31. Lopez O, Perez C, Wylie D. 1998. A single VH family and long CDR3s arethe targets for hypermutation in bovine immunoglobulin heavy chains.Immunol. Rev. 162:55– 66.

32. Mascola JR. 2002. Passive transfer studies to elucidate the role of anti-body-mediated protection against HIV-1. Vaccine 20:1922–1925.

33. Mascola JR, et al. 1999. Protection of macaques against pathogenic sim-ian/human immunodeficiency virus 89.6PD by passive transfer of neutral-izing antibodies. J. Virol. 73:4009 – 4018.

34. McGowan I. 2010. Microbicides for HIV prevention: reality or hope?Curr. Opin. Infect. Dis. 23:26 –31.

35. Mehra R. 2006. Milk Ig for health promotion. Int. Dairy J. 16:1262–1271.36. Microbicide Trials Network. 2011. MTN statement on decision to

discontinue use of tenofovir gel in VOICE, a major HIV preventionstudy in women. Microbicide Trials Network, Pittsburgh, PA. http://www.mtnstopshiv.org/node/3909.

37. Minces LR, McGowan I. 2010. Advances in the development of mi-crobicides for the prevention of HIV infection. Curr. Infect. Dis. Rep.12:56 – 62.

38. Moore M, Tyler JW, Chigerwe M, Dawes ME, Middleton JR. 2005.Effect of delayed colostrum collection on colostral IgG concentration indairy cows. J. Am. Vet. Med. Assoc. 226:1375–1377.

39. Ng WC, Wong V, Muller B, Rawlin G, Brown LE. 2010. Prevention andtreatment of influenza with hyperimmune bovine colostrum antibody.PLoS One 5:e13622. doi:10.1371/journal.pone.0013622.

40. Nishimura Y, et al. 2002. Determination of a statistically valid neutral-ization titer in plasma that confers protection against simian-human im-munodeficiency virus challenge following passive transfer of high-titeredneutralizing antibodies. J. Virol. 76:2123–2130.

41. Nkolola JP, et al. 2010. Breadth of neutralizing antibodies elicited bystable, homogeneous clade A and clade C HIV-1 gp140 envelope trimersin guinea pigs. J. Virol. 84:3270 –3279.

42. O’Doherty U, Swiggard WJ, Malim MH. 2000. Human immunodefi-

Kramski et al.

4318 aac.asm.org Antimicrobial Agents and Chemotherapy

ciency virus type 1 spinoculation enhances infection through virus bind-ing. J. Virol. 74:10074 –10080.

43. Ofek G, et al. 2004. Structure and mechanistic analysis of the anti-humanimmunodeficiency virus type 1 antibody 2F5 in complex with its gp41epitope. J. Virol. 78:10724 –10737.

44. Onyango-Makumbi C, et al. 2011. Safety and efficacy of HIV hyperim-mune globulin for prevention of mother-to-child HIV transmission inHIV-1-infected pregnant women and their infants in Kampala, Uganda(HIVIGLOB/NVP STUDY). J. Acquir. Immune Defic. Syndr. 58:399 –407.

45. Parren PW, et al. 2001. Antibody protects macaques against vaginalchallenge with a pathogenic R5 SIV HIV at serum levels giving completeneutralization in vitro. J. Virol. 75:8340 – 8347.

46. Playford RJ, Macdonald CE, Johnson WS. 2000. Colostrum and milk-derived peptide growth factors for the treatment of gastrointestinal disor-ders. Am. J. Clin. Nutr. 72:5–14.

47. Prince AM, et al. 1991. Prevention of HIV infection by passive immuni-zation with HIV immunoglobulin. AIDS Res. Hum. Retroviruses 7:971–973.

48. Saphire EO, et al. 2001. Crystal structure of a neutralizing human IGGagainst HIV-1: a template for vaccine design. Science 293:1155–1159.

49. Scherer EM, Leaman DP, Zwick MB, McMichael AJ, Burton DR. 2010.Aromatic residues at the edge of the antibody combining site facilitate viralglycoprotein recognition through membrane interactions. Proc. Natl.Acad. Sci. U. S. A. 107:1529 –1534.

50. Sinclair MC, Gilchrist J, Aitken R. 1997. Bovine IgG repertoire is dom-inated by a single diversified VH gene family. J. Immunol. 159:3883–3889.

51. Skoler-Karpoff S, et al. 2008. Efficacy of Carraguard for prevention ofHIV infection in women in South Africa: a randomised, double-blind,placebo-controlled trial. Lancet 372:1977–1987.

52. Smolenski G, et al. 2007. Characterisation of host defence proteins inmilk using a proteomic approach. J. Proteome Res. 6:207–215.

53. Spearman P, et al. 2011. A trimeric, V2-deleted HIV-1 envelope glyco-protein vaccine elicits potent neutralizing antibodies but limited breadthof neutralization in human volunteers. J. Infect. Dis. 203:1165–1173.

54. Stanfield RL, Gorny MK, Williams C, Zolla-Pazner S, Wilson IA. 2004.Structural rationale for the broad neutralization of HIV-1 by humanmonoclonal antibody 447-52D. Structure 12:193–204.

55. Stelwagen K, Carpenter E, Haigh B, Hodgkinson A, Wheeler TT. 2009.Immune components of bovine colostrum and milk. J. Anim. Sci. 87:3–9.

56. Struff WG, Sprotte G. 2008. Bovine colostrum as a biologic in clinicalmedicine: a review—part II: clinical studies. Int. J. Clin. Pharmacol. Ther.46:211–225.

57. Tharpa B. 2005. Health factors in colostrum. Ind. J. Pediatr. 72:579.58. Van Damme L, et al. 2008. Lack of effectiveness of cellulose sulfate gel for

the prevention of vaginal HIV transmission. N. Engl. J. Med. 359:463–472.

59. Van Damme L, et al. 2002. Effectiveness of COL-1492, a nonoxynol-9vaginal gel, on HIV-1 transmission in female sex workers: a randomisedcontrolled trial. Lancet 360:971–977.

60. Veazey RS, et al. 2003. Prevention of virus transmission to macaquemonkeys by a vaginally applied monoclonal antibody to HIV-1 gp120.Nat. Med. 9:343–346.

61. Walker LM, et al. 2009. Broad and potent neutralizing antibodies from anAfrican donor reveal a new HIV-1 vaccine target. Science 326:285–289.

62. Wu TT, Johnson G, Kabat EA. 1993. Length distribution of CDRH3 inantibodies. Proteins 16:1–7.

63. Wu X, et al. 2010. Rational design of envelope identifies broadly neutral-izing human monoclonal antibodies to HIV-1. Science 329:856 – 861.

64. Wu X, et al. 2011. Focused evolution of HIV-1 neutralizing antibodiesrevealed by structures and deep sequencing. Science 333:1593–1602.

65. Wyatt R, et al. 1998. The antigenic structure of the HIV gp120 envelopeglycoprotein. Nature 393:705–711.

66. Zwick MB, et al. 2004. The long third complementarity-determiningregion of the heavy chain is important in the activity of the broadly neu-tralizing anti-human immunodeficiency virus type 1 antibody 2F5. J. Vi-rol. 78:3155–3161.

67. Zwick MB, et al. 2003. Molecular features of the broadly neutralizingimmunoglobulin G1 b12 required for recognition of human immunode-ficiency virus type 1 gp120. J. Virol. 77:5863–5876.

Bovine Colostrum HIV-Neutralizing Antibodies

August 2012 Volume 56 Number 8 aac.asm.org 4319