INFECIION AND IMMUNITY, Oct. 1972, p. 600-609Copyright © 1972 American Society for Microbiology

Vol. 6, No. 4Printed in U.S.A.

Isolation of Porcine Immunoglobulins andDetermination of the ImmunoglobulinClasses of Transmissible Gastroenteritis

Viral Antibodies1LINDA J. SAIF, EDWARD H. BOHL, AND R. K. PAUL GUPTA2

Department of Veterinary Science, Ohio Agricultural Research and Development Center, Wooster, Ohio 44691

Received for publication 23 May 1972

The porcine immunoglobulins M (IgM), A (IgA), and G (IgG) were isolatedand purified and some of the properties of the porcine milk IgA were examined.Monospecific antisera which were prepared against these immunoglobulins in rab-bits were then used to absorb a particular class of immunoglobulin from sowserum, colostrum, and milk in an attempt to identify the immunoglobulin classesof neutralizing antibodies to the porcine enteric virus, transmissible gastroenteritis(TGE). The results of these absorption studies suggest that in colostrum and milkfrom sows experimentally (orally) or naturally infected with live virulent TGEvirus, IgA is the predominant immunoglobulin class of TGE antibodies. Both IgAand IgG TGE antibodies appeared to be present in the serum from these sows, butwith IgG TGE antibodies predominating. In contrast, in the serum, colostrum andmilk from sows vaccinated intramuscularly or intramammarily with live at-tenuated TGE virus, the TGE antibody activity was associated mainly with theIgG class of immunoglobulins. These results provide additional data indicatingthat the route of infection or vaccination markedly influences the immunoglobulinclass of antibodies in colostrum and milk. Secondly, IgA antibodies in mammarysecretions are probably essential for providing optimal passive immunity of nursingpigs against infection with TGE virus.

The absence of the transplacental passage ofmaternal antibodies in the pig has been reportedin studies by Sterzl et al. (23) and Kim, Bradley,and Watson (10). Sow mammary secretions aretherefore the sole source of passively acquiredantibodies to the newborn piglets. Because it isduring this period that piglets are most sus-ceptible to infection with the transmissiblegastroenteritis (TGE) virus, neutralizing anti-bodies transmitted in the milk and colostrum areof primary significance for passive immunity soas to provide protection to the epithelial cells ofthe small intestine (8, 9).The results reported in our previous papers

(la, 2) based on G-200 gel filtration studies haveindicated that TGE antibodies in milk from vac-cinated sows were associated mainly with second

1 Journal article no. 49-72 from the Ohio Agricultural Researchand Development Center, Wooster, Ohio 44691.

2 Present address: Department of Veterinary Pathology,College of Veterinary Medicine, Haryana Agricultural University,Hissar, India.

peak fractions which contained predominantlyimmunoglobulin G (IgG), whereas in milk frominfected sows they appeared mainly in fiist peakfractions where immunoglobulin A (IgA) wasmost concentrated. Moreover, piglets from in-fected sows were better protected against chal-lenge than piglets from vaccinated sows. The im-portance of the IgA secretory antibody system inhumans in the protection of mucous membraneshas been demonstrated in a number of studies(5, 22, 27) and secretory IgA has been shown topossess neutralizing-antibody activity to anumber of respiratory and enteric viruses (5, 14,22, 27). The protective role of secretory IgA inalimentary immunity against poliovirus has beenreported in studies by Ogra (14, 15).The isolation of IgA from porcine serum and

secretions has only recently been reported (3, 17,20, 24, 25). Its possible role in local infections inthe pig has been suggested also in studies byPorter, Noakes, and Allen (19) who found thatcolostral and milk antibodies to somatic antigens

600

on June 30, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

PORCINE IMMUNOGLOBULINS

of Escherichia coli were predominantly asso-ciated with IgA. This study was undertaken toisolate the porcine immunoglobulins IgM, IgA,and IgG and to identify the immunoglobulinclasses of TGE antibodies in the serum, colos-trum, and milk from vaccinated and infectedsows. Some of the properties of the isolatedporcine milk IgA are described, including itssignificance as neutralizing antibody in passiveimmunity to TGE.

MATERIALS AND METHODS

Experimental animals. The TGE viral preparationsand the methods employed in the infection, vaccina-tion, and challenge. of swine were described in the pre-ceding report (la).

Infection with live TGE virus. The pregnant sows,255 and 83-1, were naturally infected with TGE virus.An outbreak of TGE occurred in the herd of sow 255approximately 157 days prefarrowing and in the herdof sow 83-1 about 80 days prefarrowing. Sows 107and 76-10 were experimentally infected (orally andintranasally with 5 ml of Miller no. 3, virulent strainof TGE virus) at 32 and 33 days prefarrowing, respec-tively.

Vaccination with attenuated TGE virus. Sows 5-1and 179 were vaccinated twice intramammarily (imm)with the high cell-culture-passaged Purdue (HPP)strain ofTGE virus, each time using 5 ml of inoculumdivided among 3 glands on the left side of the udder.Sow 5-1 was injected at 43 and 14 days prefarrowingand sow 179 received imm injections at 34 and 12 daysprefarrowing. Sow 74-1 was injected intramuscularly(im) in the ham on days 41 and 13 prefarrowing, usinga 5-mi dose of the HPP strain. Piglets from both theinfected and vaccinated sows (except the piglets of sow107) were challenged orally with virulent TGE virusat about 3 days postfarrowing (DPF) as described pre-viously (1).

Collection of specimens. Methods used for the col-lection and processing of serum, colostrum, and milkwere the same as those reported in the preceding paper(la). Colostrum and milk samples from imm vac-cinated sows were collected from the injected glandsonly.

Virus neutralization test. TGE virus-neutralizingantibodies were detected using the plaque reductiontest as described previously (la). Antibody titers were

expressed as the reciprocal of the sample dilutionresulting in an 80% reduction in plaques.

Chromatographic methods. Gel filtration chroma-tography was performed on Sephadex G-200 in 0.1 Mtris(hydroxymethyl)aminomethane (Tris)-0.2 M NaClbuffer adjusted to pH 8 with HCI. Two columns (2.5by 45 cm and 2.5 by 100 cm) were connected in seriesand a flow rate of about 10 ml/hr was maintained withthe use of a peristaltic pump (Buchler Instruments,Fort Lee, N.J.). The eluate was collected in fractionsof 3 ml and the optical density (OD) at 280 nm was de-termined. Selected fractions were pooled and concen-

trated by ultrafiltration through an Amicon DiafloXM-100 membrane.

Ion-exchange chromatography was performed ondiethylaminoethyl (DEAE)-cellulose (Whatman DE52) using a column 1.5 by 30 cm. Protein was elutedusing a stepwise buffer elution procedure at a flow rateof 15 ml/hr. Samples used for the isolation of IgA wereeluted using the following stepwise changes in the mol-larity of Tris-hydrochloride buffer at pH 7.4: (i) 0.02 M,(ii) 0.1 M, (iii) 0.15 M, (iv) 0.2 M, and (v) 0.4 M plus2 M NaCl. Following a procedure similar to that out-lined by Mach and Pahud (11), samples used for theisolation ofIgM or IgG were eluted using the followingphosphate buffers: (i) 0.01 M, pH 7.4 (ii) 0.02 M, pH7.2; (iii) 0.06 M, pH 7; (iv) 0.1 M, pH 6; (v) 0.2 M, pH6; (vi) 0.3 M, pH 6; and (vii) 0.4 M, pH 6 plus 2 M NaCl.Fractions of 5 ml were collected and their OD valuesat 280 nm were determined.

Isolation of porcine immunoglobulins. Milk IgA.Three times concentrated sow's milk whey, late inlactation (10-11 weeks postfarrowing), was the sourceof porcine IgA. A typical elution profile obtained withsuch a sample of milk whey on Sephadex G-200 isshown in Fig. 1. On the basis of immunodiffusion andimmunoelectrophoresis using specific antisera, theprotein later identified as IgA appeared to be mostconcentrated in fractions 2 and 3 from the apex anddescending portions of the first peak. These fractionswere pooled, concentrated, and applied to a DEAE-cellulose column after thorough dialysis at 4 C againstthe first Tris-hydrochloride buffer. Porcine IgA waseluted with the second, third, and fourth Tris-hydro-chloride buffers. Fractions eluted with the third andfourth buffers contained both IgM and IgA. However,the fractions eluted with the second buffer containedno other immunoglobulins except IgA when tested byimmunodiffusion and immunoelectrophoresis usingspecific antisera prepared against porcine IgM, IgA,and IgG and anti-porcine colostral whey. Therefore,these fractions were pooled, concentrated and refil-tered on Sephadex G-200. A single symmetrical peakwas eluted after the void volume. Fractions from theapex and the descending portions of this peak werepooled, concentrated and used for the preparation ofanti-IgA serum.

IgG. Porcine colostral whey was the source of IgG,which is the predominant immunoglobulin in this se-cretion (17). Ten milliliters of colostral whey weredialyzed against the first phosphate buffer and thenfractionated on DEAE-cellulose. IgG was the onlyimmunoglobulin detectable in fractions 1, 2, and 3which were eluted with the first and second phosphatebuffers (Fig. 2). The fall-through peak, fraction 1, con-tained mainly slow IgG (IgG2), but with small amountsoffastIgG (IgGj) also present. IgG1 was present mainlyin fractions 2 and 3 which contained little IgG2 How-ever, a complete separation of these two subclasseswas not attempted. This first fraction, when concen-trated and filtered on Sephadex G-200, resulted in asingle highly symmetrical peak which eluted in the 7Sregion of the chromatogram. Peak fractions were con-centrated and used for the production of anti-porcineIgG serum.

IgM. Lipoproteins were precipitated from porcine

601VOL. 6, 1972

on June 30, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

SAIF, BOHL, AND GUPTA INFECr. IMMUNITY

I-

._

c

0

cc

Fractions

FIG. 1. Gel filtration on Sephadex G-200 of 70-day post-farrowing (DPF) milk whey from a naturally infectedsow. Vertical bars represent the TGE-neutralizing antibody titers in unconcentraled eluates from individual tubes.Immunoglobulinis were identified in the concentrated fractions (four to five times) by means of specific antisera.

serum as described by Bourne (4). A euglobulin pre-cipitate was then obtained by dialysis of the lipopro-tein-free serum against distilled water for 48 hr at 4 C.This euglobulin precipitate was dissolved in the thirdphosphate buffer and applied to a DEAE-cellulosecolumn equilibrated with this same buffer. IgM wasthe only immunoglobulin detected in fractions 10 and11 which were eluted with the sixth and seventh phos-phate buffers (Fig. 2). These fractions were concen-trated and recycled twice on Sephadex G-200'to give asingle sharp peak which eluted with the exclusion vol-ume. Peak fractions were concentrated and used toprepare anti-IgM serum.

Antisera. Rabbit antisera were prepared to solu-tions of purified porcine IgA, IgM, and IgG and toporcine serum, colostral, and milk whey. Rabbits re-ceived a primary im injection of about 2 mg of protein(0.5 ml) emulsified in an equal volume of CompleteFreund's Adjuvant. Thereafter, rabbits were given imor intravenous (iv) injections, or both, of 2 to 5 mg ofprotein every 2 weeks for a period of approximately 2months. Rabbits were exsanguinated by cardiac punc-ture. A small quantity of rabbit anti-human myelomaIgA was kindly supplied by Karen Costa (UniversityHospital, Columbus, Ohio). Rabbit anti-human IgMserum was obtained from Pentex (Miles Laboratories,Kankakee, Ill.; lot 4194).

Rabbit antisera to porcine IgA, IgM, and IgGwere rendered monospecific for a, ,u, and heavychains by absorption with soluble antigens. Anti-IgAand anti-IgM were absorbed with small samples ofporcine IgG until all anti-light chain activity was re-moved (200 ,ug of IgG/ml of anti-IgA and 400 g ofIgG/ml of anti-IgM). Similarly, anti-IgG was ab-

sorbed with small samples of a mixture containingporcine IgM and IgA until no more anti-light chainactivity was observed.The anti-IgM antiserum contained an a-macro-

globulin contaminant which was removed by absorp-tion with precolostral piglet serum (17).

Immunoelectrophoresis and immunodiffusion. Im-munoelectrophoresis was performed by a slight modi-fication of the micromethod of Scheidegger (21) in aGelman electrophoresis apparatus. A 1% solution ofpurified Noble agar in 0.05 M sodium barbital buffer,pH 8.6, was used.

Double immunodiffusion using 1% Noble agar(Difco) in 1% NaCl with 1:10,000 merthiolate wasperformed by the micromodification of Ouchter-lony's method described by Wadsworth (26).

Protein concentrations. The concentrations of solu-tions of purified porcine immunolobulins were de-termined using the extinction coefficients for porcineIgA, IgM, and IgG reported by Curtis and Bourne (7)at 280 nm.

Sucrose density gradient ultracentrifugation. Amodification of the method described by Martin andAmes (12) was used. Samples of 0.2 ml of purifiedporcine IgG and IgA containing 5 to 7 mg of protein/ml were applied to the top of linear 5 to 20% su-crose density gradients prepared in Tris-NaCl buffer,pH 8. Centrifugation was in a Beckman model L3-50ultracentrifuge in an SW41 ti rotor at an average speedof 40,000 rev/min for 18 hr at 4 C. A purified sampleof bovine lactoferrin with a sedimentation coefficientof 5.3 S was used as a sedimentation marker in someof the gradients. The gradients were fractionated bypuncturing the bottom of the tubes and eluting with a

602

E0GoCN4at0ci

on June 30, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

PORCINE IMMUNOGLOBULINS

DEAE Fractionation of Colostrum

4 5 6 7+ + + +

FractionsIgM Hi-i H H

IgA e- HI 1_-IgG E-E H H i _ H -4

FIG. 2. DEAE-cellulose chromatography of porcine colostral whey. Arrows indicate buffer changes and the

numbers above the arrows represent the following phosphate buffers: (2) 0.02 M, pH 7.2; (3) 0.06 M, pH 7; (4)

0.1 m, pH 6; (5) 0.2 m, pH 6; (6) 0.3 M, pH 6; (7) 0.4 m, pH 6 plus 2 m NaCI. The fall-through peak (fraction 1)

was eluted with 0.01 M phosphate; pH 7.4. Total elution volume was 1,100 ml. Immunoglobulins were detected in

concentrated fractions (8 to 12 times) using specific antisera.

50% sucrose solution in an upward direction. Peakswere monitored by absorbance at 280 nm.

Absorption of colostral and milk whey samples withheavy chain-specific antisera. Representative porcinecolostral and milk whey and serum samples were col-lected from vaccinated and infected sows. These milkand colostral samples were diluted (1:4 to 1:20, re-spectively) and then absorbed with monospecific rab-bit anti-porcine IgA, anti-IgM, or anti-IgG to removethe corresponding class of immunoglobulin. One-milliliter samples of diluted whey secretions weremixed with the specific antiserum and then incubated,first at 37 C for 30 min and then at 4 C for 24 to 48 hr.Controls incubated in the same manner contained 1 mlof the diluted whey sample and saline instead of anti-serum. Precipitates which formed in the tubes con-

taining the whey and antiserum were removed bycentrifugation. The same procedures were repeated onthe supernatants until all of the particular class of im-munoglobulin was precipitated by the specific anti-serum. Removal of an immunoglobulin class was con-

sidered to be complete when the sample no longershowed a line of precipitation on immunodiffusionagainst the specific antiserum. However, as shown byimmunodiffusion using specific antisera, the nonho-mologous classes of immunoglobulins were still pres-ent in the samples following absorption. As a furthercontrol, a sample of purified porcine milk IgA, whichhad a TGE-neutralizing antibody titer of 74 was ab-sorbed with monospecific anti-IgG to determine thepossible significance of nonspecific precipitation ofimmunoglobulins from the samples. However, no sig-nificant reduction in antibody titer was observed afterabsorption with anti-IgG.

After absorption, samples were filtered through0.45 lMm membrane filters and subsequently tested in

the virus neutralization test. Virus neutralization titerswere expressed in terms of the final dilutions of the ab-sorbed samples and saline controls. Differences intiters between the absorbed samples and saline con-

trols of less than twofold were not considered sig-nificant. All immunoglobulin antisera used for ab-sorption were first tested for TGE virus-neutralizingactivity and were found to be negative.

RESULTS

Gel filtration of porcine whey and serum.When porcine colostral and milk whey werefiltered on Sephadex G-200, four peaks were ob-served as was shown in Fig. 1 for milk wheycollected late in lactation.Immunoglobulins were identified in the various

pooled and concentrated fractions by immuno-diffusion and immunoelectrophoresis using mono-specific antisera. IgM was generally detected infractions throughout the first peak, but was mostconcentrated in fractions from the ascendingportion of the first peak. IgA was eluted in frac-tions throughout the first peak, often as a shoulderon the first peak, and sometimes was also detectedin the second peak fractions. It appeared to bemost concentrated in fractions from the apex anddescending portions of the first peak. IgG was

detected in second peak fractions, in colostralwhey, and often in first peak fractions as well.The third peak fractions contained mainlyalbumin. The milk-specific proteins (they did not

cross-react with anti-porcine serum) which were

5-E0co

04

VOL. 6, 1972 603

on June 30, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

SAIF, BOHL, AND GUPTA

eluted in fourth peak fractions were possiblyalpha lactalbumin and beta lactoglobulin.As reported by others, IgG, which is the pre-

dominant immunoglobulin in colostrum, de-creases approximately 30-fold during the firstweek of lactation whereas the concentration ofIgA decreases only two- to threefold, becomingthe predominant immunoglobulin in porcinemilk (7, 17). These findings were suggested in thisstudy by the changes observed in the G-200elution profiles of colostrum and milk. Thesecond peak which contains predominantly IgGwas the major peak eluted from colostral wheyand the minor peak eluted from late milk whey(Fig. 1). In all the 4- to 5-DPF milk whey samplesstudied, the first peak (which contains IgA andIgM) was the major protein peak eluted fromG-200.

In this paper, only the first two peaks areshown in most of the G-200 chromatograms ofporcine whey since the immunoglobulins andTGE antibody activity were detected only inthese fractions.Gel filtration of porcine serum on Sephadex

G-200 resulted in the appearance of three proteinpeaks. Using immunodiffusion and immuno-electrophoresis, IgM was detected throughoutthe first peak fractions. IgA was generally presentin fractions from the descending portion of thefirst peak and the ascending and top portionof the second peak. IgG was detected throughoutthe second peak and also sometimes in first peakfractions. Third peak fractions contained al-bumin and other serum proteins.

Purity and identification of isolated porcineimmunoglobulins. When porcine IgM, IgA, andIgG (concentration of 5 to 12 mg of protein/ml)were tested by immunodiffusion and immuno-electrophoresis against rabbit anti-porcine serumand anti-colostral whey, single precipitin arcswere observed with each immunoglobulinpreparation. On immunoelectrophoresis theporcine IgA and IgM migrated in the B2 positionand the IgG in the gamma position. Rabbitantisera prepared against each purified porcineimmunoglobulin were adsorbed to remove anti-light chain reactivity as described, and thenexamined for heavy chain specificity by immuno-diffusion and immunoelectrophoresis. As shownin Fig. 3, each antiserum developed only one

precipitin arc which was characteristic for eachclass of immunoglobulin when tested againstporcine serum or colostral whey. All three anti-sera failed to react with the nonhomologousclasses of immunoglobulins (at concentrations of10 mg/ml) when tested by immunodiffusion.The identity of porcine IgM and IgA was

further confirmed by their cross-reactions with

anti-human IgA or IgM antisera. When analyzedby immunoelectrophoresis, porcine IgM, but notIgA or IgG, showed a precipitin arc with anti-human IgM. Similarly, porcine IgA, but notIgM or IgG, revealed an arc of precipitationagainst anti-human IgA.The porcine milk IgA isolated in this study was

further characterized by Ouchterlony analysis todetermine its relationship to porcine serum IgA.As shown in Fig. 4, the purified porcine milkIgA, colostral, and milk whey lines spurred overthe porcine serum line when tested against anti-milk IgA, indicating the presence of additionalantigenic determinants on the colostral and milkIgA. When examined by sucrose density ultra-centrifugation, one preparation of milk IgAshowed a single symmetrical peak with a sedi-mentation coefficient of 11.0 S. A second IgApreparation showed a peak sedimentation coeffi-cient of 11.0 S with a shoulder which had a sedi-mentation coefficient of 9.55 S. Porcine IgG sedi-mented as a single peak with a sedimentationcoefficient of 6.5 S.

Changes in TGE antibody titers after absorption

rmw

nr ei

P .)

,i. I

t

! t: #A

S)£.

.._nt

X \0 X̀ ..,%.?

(t~~~~~~I5

FIG. 3. Immunoelectrophoresis of porcine colostralwhey (CW) and serum (PS) developed with porcine,heavy chain-specific anti-IgM, anti-IgA, and anti-IgG, and also anti-porcine serum (anti-PS). Anode tothe left.

604 INFECT. IMMUNITY

.. ... .....

p c.

on June 30, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

PORCINE IMMUNOGLOBULINS

of samples with monospecific antisera. (i) Experi-mental or natural infection. In G-200 fractionatedcolostrum and milk whey from naturally orexperimentally infected sows, the TGE antibodytiters were highest in fractions from the apex anddescending portions of the first peak, where IgAis most concentrated. Antibody activity wasabsent or low in second peak fractions whereIgG is most concentrated. These results are illus-trated by the G-200 chromatograms shown forcolostral and 5-DPF milk whey from a naturallyinfected sow, 255 (Fig. 5). Similar results wereseen after gel filtration of milk or colostral wheyfrom experimentally infected sows (la). Whenserum from experimentally or naturally infectedsows was filtered on G-200, TGE antibodyactivity was highest in fractions from the ascend-ing portion and top of the second peak, whereboth IgG and IgA were detected by immuno-diffusion. This finding is illustrated by the G-200chromatogram of serum from sow 76-10 shown inFig. 6.To confirm the immunoglobulin classes of

TGE antibodies suggested by these observations,samples were absorbed with heavy chain-specificantisera to remove a particular class of immuno-globulin. The virus neutralization titers of theseabsorbed samples and saline controls are sum-marized in Table 1. When colostral whey fromthe infected sow 255 was absorbed with anti-IgA,the TGE antibody titer was decreased at least70-fold (to the limit of the dilution factor) com-pared to the control. However, absorption of this

FIG. 4. Immunodiffusion of porcine serum (PS),colostral whey (CW), milk whey (MW), and purifiedporcine milk IgA with anti-porcine milk IgA (centerwell) showing spurring of milk and colostral IgA linesover serum IgA lines.

sample with anti-IgG or anti-IgM did not sig-nificantly reduce the antibody titer. Absorptionof colostral whey samples from infected sows83-1, 107, and 76-10 with anti-IgA reduced theantibody titers from about seven- to twofold.When these samples were absorbed with anti-IgG, two- to threefold reductions in antibodytiters were observed. This suggests the associationof antibody activity with both IgG and IgA,but predominantly with the latter in these colos-tral samples. Absorption of 4- to 5-DPF milkwhey samples from these sows (255, 107, and76-10) with anti-IgA resulted in 72- to 6-folddecreases in antibody titers, all of which werereduced to the level of the final dilution factors.This significant reduction in antibody activityafter absorption with anti-IgA is in good agree-ment with the observation that absorption withanti-IgG or anti-IgM produced little or no changein the antibody titers of these samples. As ex-pected, similar results were seen when a G-200milk whey fraction (sow 255, G-200 Fr. 4, Fig. 5),which contained TGE antibody activity and bothIgG and IgA, was absorbed with anti-IgA andanti-IgG (Table 1). Following absorption of aserum sample from sow 76-10 with anti-IgA,only a slight decrease in titer was observed, whileabsorption with anti-IgG resulted in about athreefold reduction. Since both absorptionsresulted in reductions in antibody titers, butneither anti-IgA nor anti-IgG reduced the anti-body titer completely, this suggests that bothIgG and IgA TGE antibodies were present.

(ii) im and imm vaccination. Results of gelfiltration studies of colostral and 4- to 7-DPFmilk whey and serum from both im and immvaccinated sows suggested that TGE antibodyactivity was mainly associated with second peakfractions which contained predominantly IgG.A typical G-200 elution profile of colostral andmilk whey from the imm vaccinated sow 179is shown in Fig. 7. Also shown in this figure isthe G-200 elution curve of 29-DPF milk wheywhich was collected 26 days after the exposure ofthis sow's piglets to TGE virus. Whereas TGEantibody activity was highest in second peakfractions from colostral whey and 4-DPF milkwhey, in the 29-DPF milk whey, the antibodyactivity was highest in fractions from the descend-ing portion of the first peak. Similar results, withthe primary localization of the TGE antibodyactivity in second peak fractions, were observedafter gel chromatography of serum from im or

imm vaccinated sows. This is exemplified by theG-200 elution curve of serum from the imm vac-

cinated sow 5-1 (Fig. 6). Gel filtration on G-200of colostral and milk whey and serum from imvaccinated sows resulted in similar curves with

VOL. 6) 1972 605

on June 30, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

606~~~~~SAIF,BOHL, AND GUPTA

14-b

-2561.2

-64~o0>0E- a 0 0.

INFECT. IMMUNITY

A,

n

0

IL

Tube NumberlgM -IgA __X_a__IgG

FractionsIgM l-4IgA -IgG

FIG. 5. Gelfiltration on Sephadex G-200 of colostral whey (a) and 5-day postfarrowing (DPF) milk whey (b)from a naturally infected sow (255). Vertical bars represent the TGE-neutralizing antibody titers ofunconcentratedeluate from individualfractions (a) or concentrated (five times) pooledfractions (b). Immunoglobulins were identi-fied in concentratedpooledfractions (five times) using monospecific antisera.

0

0

0

Tube NumberIgM - + +

IgA - + +

IgG- + + + +

Tube NumberIgM - + + +IgA - ++ + + + _IgG - + + + + + +

FIG. 6. Gelfiltration on Sephadex G-200 ofserum (2 ml) from an inttramammarily vaccinated sow, 5-1 (a), andant experimentally infected sow, 76-10 (b). TGE-neutralizing antibody titers in individual unconcentratedfractionsare indicated by vertical bars. Immunoglobulins were identified in these fractions by means of immunodiffusionusing monospecific antisera.

the highest antibody titers in second peak frac-tions (la).The results of absorption studies of samples

from these sows are also summarized in Table 1.In all cases, absorption of colostral whey, 4- to7-DPF milk and serum samples from vaccinatedsows (74-1, 5-1, and 179) with anti-IgG resultedin significant reductions in antibody titers rangingfrom 15- to 12-fold decreases compared to thecontrols. In all the samples from vaccinated

sows, except colostral whey from sow 179 and 4-DPF milk whey from the imm vaccinated sows5-1 and 179, absorption with anti-IgG reducedthe titer to the limit of the final dilution factor.When the samples from sows 74-1, 5-1, and 179were absorbed with anti-IgA or anti-IgM (sow5-1), no significant reductions in antibody titersoccurred compared with the controls. Similarly,when a G-200 colostral whey fraction (Fig. 7a,sow 179, G-200 Fr. 1) from the descending por-tion of the first peak which contained both IgA

0cocod

-00

0v

p.2.w0ce

606

on June 30, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

PORCINE IMMUNOGLOBULINS 607

TABLE 1. Comparisonis of the TGE-nieutralizinzg anitibody titers of paired samples after absorptioni withrabbit aniti-porcinie heavy chain-specific sera or saline

I I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Sow Route of exposure to TGE

Natural infectionNatural infectionNatural infection

Natural infectionExptl infectionExptl infectionExptl infectionExptl infectionExptl infectionim Vaccinationbim Vaccinationimm Vaccinationbimm Vaccinationimm Vaccinationimm Vaccinationimm Vaccination

imm Vaccinationimm Vaccinationimm Vaccination

Sample

Colostral whey5-DPFb milk whey5-DPF milk wheyG-200 Fr. 4c

Colostral wheyColostral whey5-DPF milk wheyColostral whey4-DPF milk whey4-DPF serumColostral whey7-DPF milk wheyColostral whey4-DPF milk whey6-hr PF serumdColostral wheyColostral wheyG-200 Fr. le

4-DPF milk whey29-DPF milk whey29-DPF serum

Antibody titera after absorption with

Anti-IgM

4,080820

15,6001,184

Salcnt Anti-IgA

3,840 <41840 <13.2

<2.5

14,8201,152

17087

<12400<1676

10,32086

15,4001,0883,200

41,40017

84634

3,500

Salinecontrol

2,80095050

950690444960109116

8,370100

12,2501,2002,00060,750

15

6841,0243,750

Anti-IgG Saline;control

1,80052831

3373852794788039

<100<5.8<97.592

<52.55104,

39480

<105a Reciprocal of the dilution resulting in 80%o plaque reduction.I Abbreviations: DPF, day postfarrowing; im, intramuscular; imm, intramammary.c See Fig. 5.d See Fig. 6.e See Fig. 7.

a [ C MilCeIestrum So0 179 0..Sw 179 Milk 4 DPF 6

Iii I

E 3 Sow107 0O5.0-II 10024;'T Iv\ 04

3,37068034

1,080880315

1,32085133

14,20093

14,6251,1002,83552,500

27

704688

3,570

19M - + + _IgA - + + + _

+ IgG - -- - + +

FIG. 7. Gel filtrationt on Sephadex G-200 of colostral whey (a), 4-day postfarrowing (DPF) milk whey (b),and 29-DPF milk whey (c) from an intramammarily vaccinated sow. Vertical bars represent TGE-neutralizingantibody titers in individual unconicentrated fractions. Immunoglobulins were detected in these fractionls by means

of monospecific antisera.

and IgG was absorbed with anti-IgG, a seven-fold reduction in titer occurred. Absorption withanti-IgA, however, did not significantly reducethe titer. An exception to these findings was the

29-DPF milk whey from sow 179, which hadbeen collected 26 days after exposure of thissow's piglets to virulent TGE virus. Both thepiglets and sow developed clinical signs of TGE.

255255255

83-1107107

76-1076-1076-1074-174-15-15-15-1179179

179179179

0

0

._lu1

Tube Numb.rIgM - + - - -

IgA - + + + -

IgG - - - + +

VOL. 6, 1972

on June 30, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

SAIF, BOHL, AND GUPTA

When this sample was absorbed with anti-IgGthere was only a 1.4-fold decrease in titer,whereas absorption with anti-IgA resulted in a30-fold reduction in titer.

DISCUSSION

The porcine immunoglobulins, IgG, IgA, andIgM isolated in this study had properties similarto those described in the literature. Porcine IgGhad a sedimentation coefficient of 6.5 S which isconsistent with the S20, value of 6.7 S reportedby Porter and Allen (18). As noted by others (7,13), at least two subclasses of IgG were observedin this study, IgG, and IgG2.

Identification of porcine IgM and porcine milkIgA was based, in part, on their cross-reactivitywith anti-human IgM or IgA serum. The isolatedporcine milk IgA had properties similar to thosedescribed by other investigators (3, 17, 19, 20).The heterogeneity in the molecular size of IgA inmilk and colostrum was indicated by its elutionthroughout the first peak and often in secondpeak fractions on Sephadex G-200 (Fig. 1, 5).The isolated porcine milk IgA had a sedimenta-tion coefficient of 11.0 S which agrees well withthe S20, value of 11.2 S reported for porcinecolostral IgA by Richardson and Kelleher (20).On sucrose density gradient ultracentrifugation,one preparation of milk IgA had a shoulder onthe 11 S peak which had a sedimentation coeffi-cient of 9.55 S. This is similar to the sedimenta-tion coefficient of 9.3 S reported by Porter andAllen (18) for dimeric serum IgA which lacks thesecretory component. Therefore, the 11.0 S IgApeak might represent dimeric IgA with secretorycomponent and the 9.55 S shoulder couldrepresent the presence of dimeric IgA lackingsecretory component. While there are presentlyno reports in the literature describing the isolationof porcine secretory component, the possibilityof its presence is further suggested by the obser-vation that porcine milk and colostral IgApossessed additional antigenic determinants notpresent on serum IgA (Fig. 4). Similar findingswere reported by Bourne (5), but Richardsonand Kelleher (20) found no unique antigenicdeterminants present on colostral IgA comparedto serum IgA.The results of the absorption studies (Table 1)

using monospecific antisera indicate that in themilk and colostrum from sows infected eithernaturally or experimentally with live TGEvirus, the TGE antibody is associated predomi-nantly with the IgA class of immunoglobulin,although IgG TGE antibodies also were evidentin colostrum. In the serum from such animals,the TGE antibody appears to be associated with

both the IgA and IgG classes, but with more ofthe antibody activity in the IgG class.

In contrast, in the serum, colostrum and milkfrom sows vaccinated im or imm with live at-tenuated TGE virus, the TGE antibody is asso-ciated primarily, if not entirely, with the IgGclass of immunoglobulin.Such findings suggest that following im or

imm vaccination, the attenuated virus primarilystimulates the peripheral lymphoid tissues re-sulting in the production of antibodies predomi-nantly if not solely of the IgG class in serum,colostrum, and milk. This suggests that the IgGTGE antibodies found in colostrum and milkmay originate primarily from the serum. Thehigher antibody titers observed in colostrum fromthe injected glands than from the noninjectedglands, as previously reported (la), may be theresult of an inflammation in the injected glandswhich might permit an increased passage ofcirculating antibodies (which were of the IgGclass) into the colostrum of the injected glands.An alternative explanation would be that somelocal synthesis of IgG occurred in the injectedmammary glands.The mechanism accounting for the predomi-

nance of IgA TGE antibodies in the milk andcolostrum of naturally and experimentally in-fected swine is not known. Studies by Ogra (14)have demonstrated the highly localized nature ofthe secretory IgA antibody response in the in-testinal tract. Local immunization of the colonwith attenuated or inactivated poliovaccine pro-duced a secretory IgA antibody response only inthe immunized segment of the gut and not in thenasopharynx. However, in this study, the pre-dominance of IgA TGE antibodies in colostrumand milk seems to be associated with an infectionof the intestinal tract with- the virulent TGE virus.In other studies (1, 6), to account for the extra-intestinal IgA observed after oral administrationof certain antigens, it was proposed that in-testinal IgA immunocompetent cells, after beingstimulated by the antigen, emigrate from the gutto colonize distant lymphoid tissues. Migrationof sensitized IgA immunocompetent cells fromthe intestine to the mammary gland mightbe one explanation for the predominance of IgATGE antibodies observed in the milk and colos-trum of infected sows as was suggested in ourprevious studies (la, 2). An alternate explanationwould be that orally administered live virus wascarried by the circulation to the mammary glandwhere it stimulated local production of IgATGE antibodies which were secreted into thecolostrum and milk. However, the low antibodytiters of the infected sow's serum (la) suggeststhat this is a less-plausible explanation.

608 INFECT. IMMUNITY

on June 30, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

PORCINE IMMUNOGLOBULINS

Another explanation which has been proposedfor the origin of antibodies in colostrum andmilk is their transport from the serum. In thebovine, IgGj, the predominant immunoglobulinin colostrum and milk, appears to be selectivelytransported to the mammary gland from theserum (16). However, there is no evidence for a

selective transport of IgA or IgG in the pig, andIgG2 seems to be the predominant subclass ofIgG in porcine serum, colostrum, and milk (7).

After exposure of an imm vaccinated sow

(179) to live TGE virus through contact withher challenged piglets, the immunoglobulin classof antibodies in milk changed from the pre-

dominance of IgG TGE antibodies (4-DPF milkwhey) to mainly IgA TGE antibodies (29-DPFmilk whey). However, in the 29-DPF serum fromthis sow, the results of the absorption studiesindicated that IgG TGE antibodies were presentwith little or no antibody activity in the IgAclass. This suggests that synthesis of IgA TGEantibodies by plasma cells in the mammary

gland, either of local origin or relocated from theintestinal tract, may be a more plausible ex-

planation for the origin of IgA TGE antibodiesin milk than transport from the serum

ACKNOWLEDGMENTS

We are grateful for the advice and assistance of Floyd Schan-bacher in the performance of the sucrose density gradient studies.We thank Kathy L. Miller and Dennis Huebner for their technicalassistance.

This investigation was supported by contract no. 12-14-140-2085-94 from the Biologics Program, Veterinary Services, Ani-mal and Plant Health Inspection Service, U. S. Department ofAgriculture and by a grant from the National Pork ProducersCouncil.

LITERATURE CITED

1. Bazin, H., G. Levi, and G. Doria. 1970. Predominant contri-bution of IgA antibody-forming cells to an immune response

detected in extraintestinal Iymphoid tissues of germ-free miceexposed to antigen by the oral route. J. Immunol. 105:1049-1051.

la. Bohl, E. H., R. K. Paul Gupta, M. V. F. Olquin, and L. J.Saif. 1972. Antibody responses in serum, colostrum, andmilk of swine after infection or vaccination with transmis-sible gastroenteritis virus. Infect. Immunity 6:289-301.

2. Bohl, E. H., R. K. P. Gupta, L. W. McCloskey, and L. Saif.1972. Immunology of transmissible gastroenteritis. J. Amer.Vet., Med. Ass. 160:543-549.

3. Bourne, F. J. 1969. IgA immunoglobulin from porcine milk.Biochim. Biophys. Acta 181:485-487.

4. Bourne, F. J. 1969. IgA immunoglobulin from porcine serum.

Biochem. Biophys. Res. Commun. 36:138-145.5. Cate, T. R., R. D. Rosen, R. G. Douglas, W. T. Butler, and

R. B. Couch. 1966. The role of nasal secretion and serum

antibody in the rhinovirus common cold. Amer. J.Epidemiol. 84:352-363.

6. Crabbe, P. A., D. R. Nash, H. Bazin, H. Eyssen, and J. F.Heremans. 1969. Antibodies of the IgA type in intestinalplasma cells of germ-free mice after oral or parenteral im-munization with ferritin. J. Exp. Med. 130:723-738.

7. Curtis, J., and F. J. Bourne. 1971. Immunoglobulin quantita-tion in sow serum, colostrum and milk and the serum ofyoung pigs. Biochim. Biophys. Acta 236:319-332.

8. Haelterman, E. 0. 1963. Transmissible gastroenteritis of

swine. Proc. 17th World Vet. Cong. 1:615-618.9. Hooper, B. E., and E. 0. Haelterman. 1966. Concepts of

pathogenesis and passive immunity in transmissible gastro-

enteritis of swine. J. Amer. Vet. Med. Ass. 149:1580-1586.10. Kim, Y. B., S. G. Bradley, and D. W. Watson. 1966. Ontogeny

of the immune response. I. Development of immuno-

globulins in germfree and conventional colostrum-deprivedpiglets. J. Immunol. 97:52-63.

11. Mach, J. P., and J. J. Pahud. 1971. Secretory IgA, a major

immunoglobulin in most bovine external secretions. J.Immunol. 106:552-563.

12. Martin, R. G., and B. N. Ames. 1961. A method for determin-ing the sedimentation behavior of enzymes: application to

protein mixtures. J. Biol. Chem. 236:1372-1379.13. Metzger, J. J., and M. Fougereau. 1968. Caracterisation

biochemique des immunoglobulines yG et yM porcines.Rech. Vet. 1:37-61.

14. Ogra, P. L. 1971. The secretory immunoglobulin system

of the gastrointestinal tract, p. 259-279. In The secretory

immunologic system. U. S. Government Printing Office,Washington, D.C.

15. Ogra, P. L., and D. T. Karzon. 1969. Poliovirus antibodyresponse in serum and nasal secretions following intranasalinnoculation. J. Immunol. 102:15-18.

16. Pierce, A. E., and A. Feinstein. 1965. Biophysical and im-munological studies on bovine immune globulins with evi-dence for selective transport within the mammary glandfrom maternal plasma to colostrum. Immunology 8:106.

17. Porter, P. 1969. Transfer of immunoglobulins IgG, IgA andIgM to lacteal secretions in the parturient sow and theirabsorption by the neonatal piglet. Biochim. Biophys.Acta 181:391-392.

18. Porter, P., and W. D. Allen. 1972. Classes of immuno-globulins related to immunity in the pig. J. Amer. Vet. Med.Ass. 160:511-517.

19. Porter, P., D. E. Noakes, and W. D. Allen. 1970. SecretoryIgA and antibodies to Escherichia coli in porcine colostrumand milk and their significance in the alimentary tract of theyoung pig. Immunology 18:245-257.

20. Richardson, A. K., and P. C. Kelleher. 1970. The 1iS-sowcolostral immunoglobulin isolation and physicochemicalproperties. Biochim. Biophys. Acta 214:117-124.

21. Scheidegger, J. J. 1955. Une micromethode d'immunoelec-trophorese. Int. Arch. Allergy Appl. Immunol. 7:103-110.

22. Smith, C. B., R. H. Purcell, J. A. Bellanti, and R. M. Chanock.1966. Protective effect of antibody to parainfluenza type Ivirus. N. Engl. J. Med. 275:1145-1152.

23. Stertzl, J., J. Kostka, I. Riha, and L. Mandel. 1960. Attemptsto determine the formation and character of -y-globulin andof natural and immune antibodies in young pigs rearedwithout colostrum. Folia. Microbiol. 5:29-40.

24. Vaerman, J. P., J. P. Arbuckle, and J. F. Heremans. 1970.Immunoglobulin A in the pig. II. Sow's colostral and milkIgA-quantitative studies and molecular size estimation. Int.Arch. Allergy 39:323-333.

25. Vaerman, J. P., and J. F. Heremans. 1970. ImmunoglobulinA in the pig. I. Preliminary characterization of normal pigserum IgA. Int. Arch. Allergy 38:561-572.

26. Wadsworth, C. A. 1957. A slide microtechnique for the anal-ysis of immune precipitates in gel. Int. Arch. Allergy Appl.Immunol. 10:355-360.

27. Waldman, R. H., J. J. Mann, and P. A. Small. 1969. Immuni-zation against influenza: prevention of illness in man byaerosolized inactivated vaccine. J. Amer. Med. Ass. 207:520-541.

VOL. 6, 1972 609

on June 30, 2020 by guesthttp://iai.asm

.org/D

ownloaded from