Selective Induction of an Immune Response in Human

External Secretions by Ingestion of Bacterial Antigen

JIRI MESTECKY,JERRY R. MCGHEE,ROLANDR. ARNOLD,SUZANNEXI. MICHALEK, SHIRLEY J. PRINCE, and JAMES L. BABB,Department of Microbiology, the Comprehensive Cancer Center, and theInstitute of Dental Research, University of Alabama in Birmingham,University Station, Birmingham, Alabama 35294

A B S T RA C T Ingestion of capsules which containedkilled Streptococcus mlutans by four healthy humansubjects led to the appearance of specific antibodiesin external secretions. Salivary and lacrymal anti-bodies were detected within 1 wk of ingestion anidcontinued to increase throughout a 14-day immuniza-tion period, with a gradual decline during the 2ensuing months. A second period of immunizationresulted in a pronounced increase of specific anti-body levels which occurred earlier than in the pri-mary immunization period and reached peak levels byday 10. No change was detected in serum antibodylevels throughout either immunization period. Theantibody activity in all secretions was associatedwith the immunoglobulin A class, as determined byimmunochemical analyses. These data indicate thatingestion of bacterial antigens selectively stimulatesthe immune response in secretions.

INTRODUCTION

Secretory IgA (s-IgA)l is the prevalent class of immuno-globulin on mtucous membranes. Produced locallyby plasma cells that are distributed in the secretoryglands and in the lamina propria of intestinal andrespiratory tissues, s-IgA is selectively transportedinto external secretions (1). However, the mechanismsinvolved in the stimulation of the immune response inthese secretions remaini obscure. Parenteral immuniza-tion results in the production of serum antibodiesthat are principally associated with the IgG class,lbut is nevertheless iineffective in stimulating the secre-torv immune svstem (2-6). Onlv trace amounts of

Received for publication 1 August 1977 anid in revised formll24 October 1977.

'Abbreviations tused itn this paper: CFU, colony-formingunits; H chains, heavy chains; S-IgA, secretory IgA.

immunoglobulins derived from serum reach externalsecretions, and their lack of resistance to variousproteolytic enzymes makes them unsuitable to func-tion in the mucosal environment. Furthermore, evi-dence is available to suggest that the serum-derivedIgG antibodies may prevent the mucosal absorptionof an antigen used in systemic immunization, al-though the absorption of unrelated antigens is en-hanced (7). A secretory immune response, restrictedto the site of application of antigen, has been elicitedby antigens either locally applied to mucous surfaces(2, 4, 8, 9) or injected into secretory glands (10-15).The appearance of s-IgA-associated antibodies to lo-cally administered antigens (2, 4, 9) suggested that theinduction of a secretory immune response is the re-sult of local penetration of antigens with subsequentstimulation of immunocompetent cells in secretorytissues (2, 4, 8, 16). Although effective in stimulat-ing a secretory immune response, injections of antigensinto secretory glands are of questionable acceptancebecause of the anatomical inaccessibility of someglands, the possible requirement of adjuvants, the im-pairment of normal function of the gland, and theconcomitant stimulation of IgG-associated anti-bodies (15).

The demonstration that oral administration of anti-gens (in man and animals) induces the appearanceof s-IgA antibodies in mammary secretions (6, 17, 18)and the presence in saliva, milk, and colostrum ofnaturally occurring s-IgA-associated antibodies tobacterial antigens (Escherichia coli [19], Streptococcuspneunmoniae [17], and Streptococcus mutans [20])suggest that local stimulation of the gland may not berequired for this induction. We report that ingestionof a bacterial antigen induced the selective appear-ance of aintibodies of the IgA class in the externalsecretions of human salivary and lacrymal glands,without a serum antibody response.

The Journal of Clinical Investigation Voluime 61 March 1978 -731 -737 731

NIETHODS

iSubjects. Four individuials (one female and threemales, 30-35 yr old) voluniteered for this study. Unstimu-lated samples of parotid saliva were collected with the useof a plastic intraoral cup designed in this laboratory (21).This device also allowed for the simultanieouis collectioniof separate samples of suibmnlcnadibular-sublinigual salivaswithout the admiiixture of parotid secretion. This permittedthe coImlpariisoni of the titers of antibodies in secretionis of'anatomically distinct salivary glands. Whole saliva, whichcontaiins secretions of the miajor and minor salivary glandsas well as contributions fromii the crevicular fluid, was col-lected by having the subject dIrool into a sputunm cup packedin ice. Lacryymal secretionis were collected after irrita-tioll of the conijunctiva with a mist of oil expressed fromlemoni rind. Blood specimenis obtained by venipuniicture werecollected iinto heparinized tubes; leukocyte and differentialcounits were determined.

Prepa ration of antigen. The canididate immiiuniogeni wasselected through screeniing of saliva and serumii samples bymicrotitrationi (20) for agglutinin antibodies to over 40different strainis of the indigenous oral bacteriumil, Streptococ-cus mutatns. S. inutains OMZ-176 (sterotype d), to whichnone of the four subjects demiionistrated significant serumor salivary antibody activity, was selected for this study.

To prepare a large amount of bacterial antigeni for im-munization, 10 liters of dialyzed medium (22) was inlocU-late(d with a log phase cultuire of S. mutanls OMZ-176 andincubated at 37°C for 16 h. During the last 4 h of culture,the acidl formed was neutralized by sterile 1 N NaOH.Cells were harvested by cenitrifugation at 10,000 g for30 mill, washe(d three times with phosphate-buffered saline(pH 7.0), and resuspendedl in 0.5% formalin-saline at a con-centrationi of 5 x 109 colony-formiiing units (CFU)/ml. After3 days, sterility was tested by culturing in sodium thio-glvcolate broth and on blood agar. The cells were extensivelvwashed with sterile distilled water and then lyophilized.

Gelatini capsules (no. 000, Parke, Davis & Co., Detroit,Nlich.) were filled with 100 mg of desiccated cells (repre-senting approximately 1 x 10" eqjuivalent CFU). To avoid thecontact of extraneous antigeni with the oral mucosa, thecapsules were repeatedly rinsed with tap water and swal-lowed with a liberal amounlt of water; the oral cavity was thenthoroughly rinsed. Capsules were taken daily at specifiedtimes. In the initial immunizationi series, capsules were takenfor 14 conisecutive days; in the second series, for 7 consecu-tive days. Saliva, tears, and plasma samples were collectedat specific intervals.

Immllunoglobulin and antibody measurement. Serumimmunoglobulin levels were measured by single radialimmunodiffusion using commercially obtained immuno-plates and stand(lards (Behring Diagnostics, American HoechstCorp., Somerville, N. J.). IgG and IgM in external secretionswere measured with the use of single radial immuno-diffusion plates specifically prepared with commercialantisera (Behring Diagnostics) for determining low lev-els of these immunoglobulins. Dilutions of commercialpreparations of IgG and IgM were used as standards.Polymeric s-IgA purified from a pool of human colostrumby ammnonium sulfate precipitation, Sephadex G-200 andSepharose-6B gel filtration (Pharmacia Fine Chemicals,Inc., Piscataway, N. J.), and DEAE-cellulose chromatography(23) was used as the standard for single radial immuno-diffusion measurements of s-IgA. The concentration of poly-meric IgA was standardized spectrophotometrically usingan extinctioni coefficient of E""m 280 nm = 13.8. Since the col-lectioni of saliva and tears was not quantitative, s-IgA

levels in these secretionis were relate(d to the conicenltra-tion of total proteini (24) (in microgramiis of IgA per nmilli-gramii of total proteini) rather than to flow rates.

To dletermine the distribution of the antibody activitvin saliva andl tears, selected samples (first precipitatedwith amimniuoiilm stulfate, then re(lissolved and dialyzedagainist 1% amimiiloniiumii bicarbonate buffer pH 7.0) wereapplied to a columnlll of Sephadex G-200 (K9/70 cmll) e(ltqili-hrated in the samiie huffer. Protein concenitrationis, immintiio-globulin levels, andcl antibody activity vere determiniedl onpooledl fractions.

Bacterial aggluti nations were performedl hy microtitratiolns(20) in V-plates (Cooke Laboratory Prodcucts Div., Alexanl(lria,Va.). 50 ,ul of formaliniized S. mutacns OMZ-176 (2 x 108CFU/ml) was added to serial, twofold dilutions of the samiiplein barbital buffer (pH 7.4). The plates were incubated for2 h at 37°C andl allowed to stancd for 18 h at 4°C. Titerswere determinied by reading agglutination patterns. WNhenirelated to the absolute levels of IgA in selected samples ofsecretions, agglutinin titers displayed similar dynamics to un-adjusted samiples. In an effort to determinie the immunto-globulini class responisible for antibody activity, attemiptswere made to enhanice or inhibit the agglutinin titerswvith anitisera specific for heavy (H) chainis of immuntiiio-globulinls (IgA, IgM, and IgG). The agglutinated and sedi-menited bacteria were washed three times with buffer inthe mierotiter plates. 50 Al of an optimal concentration ofH-chain-specific antisera (Behring Diagnostics) was addedand the plates were incubated for 2 h at 37°C. After 18 hat 4°C, agglutinationi patterns were readl. Enhancemenit of'titers with H-chain-specific antisera suggested the classof Ig responsible for antibody activity. The inhibitioin of' theantibody activity was performed in aI manner similar tothat of titer enhancement, except that anti-H-chain or anti-secretory component (25) sera were added to the selecte(dsamples of saliva and tears and allowed to incubate for 3 hat 37°C before the antigen was added.

To f'urther confirnm the Ig class responsible for agglutininactivity, a suspension of S. tImutan1s OMZ-176 was heat-fixed on microscope slides and thoroughly washed withchilled 0.15 M carbonate-bicarbonate buffer, pH 8.5.The smears were incubated for 45 min at ambient tempera-ture with selected samples of saliva, tears, or serum andwashedl extensively with carbonate-bicarbonate buffer.The slides were incuibated for an additional 45 mmil withappropriate dilutions of fluorescein isothiocyanate-labeledantisera specific for H chains of human IgM, IgG, andlIgA, produced either in this laboratory (26) or obtainedfrom a commercial source. Then the slides were washed withbuffer, mounted in glycerol, and examined with a fluorescencemicroseope equipped with a vertical illuminator accordingto Ploem (Orthoplan, Leitz, Wetzlar, W. Germany). Thelight source was a mercury lamp (Osram HBO 100 W)e(juipped with a filter system for narrow-band excitationi.

Selected samples of plasma or sertum were tested for C-reactive protein, rheumatoid factor, and anti-streptolysin 0by the Department of Clinical Pathology at the Universityof Alabama in Birmingham. Since it has been suggested thatS. mutans may share antigenic determinants with the humanheart, selected serum and saliva samples were assayed forthe presence of heart-reactive antibodies by the indirectimmunofluorescence techniques of Zabriskie and Freimer(27). Fresh sections of human heart obtained at autopsywere imbedded in OCTcompound (Ames Co., Div. of MilesLab., Inc., Elkhart, Ind.), frozen in liquid nitrogen, andstored at -80°C until sectioning. Desiccated 4-,um sec-tions were incubated with saliva or serum for 1 h atambient temperature, washed in phosphate-buffered saline

732 Mestecky, McGhee, Arnold, Michalek, Prince, and Babb

(0.01 NI, pH 7.2), anid stained with fluorescein isothiocyanate-labeled, rabbit antihuman irnmunoglobulin (Behring Diag-nosties). All sectionis were examined with a Leitz Orthoplanifluoreseeniee n)icrroscope equipped with a vertical illunmina-tion and filter system appropriate for narrow-b)and excitation.

RESULTS

Measuirem)ient of the im munie respontse. After the 1stwk of immiluinization, significant agglutinin titers wereol)served in external secretions (Fig. 1 a-d). Thesetiters conitintued to rise throughout the immunizationperiod anid reached their highest levels 10-25 daysfromii the initial ingestion of antigen. During theenistuinig 50 days, the antibody titers slowly declinedbut did not decrease to preimmunization levels. 66dlavs after the termination of the primary immuniza-tion, a second (7-day) period of antigen ingestion wasbegun. This secondary challenge resulted in an earlier,more pronounced increase in the levels of specificaintil)odlies. Parallel dynamics of antibody appearanceand titers were observed in parotid, submandibular-sublingual, and whole salivas, as well as tears, afterboth primary and secondary stimulations. Augmenta-tion of agglutination with H-chain-specific antisera(Table I) revealed that agglutinin titers of parotidsaliva andl tears could be specifically enhanced whenthe washed and redispersed agglutinates were incu-bated with antisera specific for the a-chain; no effectwas evident with antisera to ,- and y-chains. Theagglutinin activity in selected samples of saliva andtears could be abolished or substantially reduced by theaddition of anti-a-chain-specific antiserum to thesesamples before titration by agglutination. Likewise,agglutinin activity could be inhibited by pretreat-ment of samples with anti-secretory component serum.Examination of bacterial smears that had been incu-bated with samples of secretions and then washedand reacted with fluorescein isothiocyanate-labeledantisera to H-chains corroborated the IgA nature ofthe induced antibodies. Also, when selected samplesof secretions were fractionated on Sephadex G-200,the antibody activity was detected in a fraction that

corresponided to the elution position of humiian s-IgA(Fig. 2). The s-IgA clharacter of the antibodies was coni-firmed by anti-H-chain enhancement and anti-secretorycomlpolnent inhibition of agglutiniationi aindl by in-direct i mmiuniofluoreseence o01 bacterial snmears. WN'iththe exceptioni of an apparently negligible increasein seruml agglutiniins in subject 1 (Table I), nO chaingesin serum titers were observed during either immuniza-tion period. Serum agglutininis to this strain of S.rmutitan.s that were evident before, during, and after theoral imimiilunlizatioin (Table I) belonged primarily to theIgM and IgG classes as ascertained by the enhance-menit of the agglutiniation with ,.t- and y-chaiin anti-sera, but inot a-chaini anitisera. The induced anti-bodies were specific for the immuniizing strain be-cautse there was 1)o compparable increase in agglutinintiters to other noni-cross-reacting serotypes of S.inutans (Table II). A slight increase in agglutininititers was observed with S. muttans 6715 (serotype g),which mav be explained by the presence of commonantigenis between serotypes d and g (28).

No alterations in body temperature, in the leuko-cvte couints and differentials, or in serum immunio-globulin levels were observed after ingestion oflarge doses of S. muutanis OMZ-176 (2.1 g). Likewise,C-reactive protein, anti-streptolvsin 0, and rheumiatoidfactor titers were uninfluenced by ingestion of thisbacterial antigen. An apparent rise was noted insalivary IgA levels and agglutinin titers (0.82 correla-tion). Humani heart-reactive antibodies could not bedetected by indirect immunofluorescence in salivaand serum samples collected before, during, andafter immunization.

DISCUSSION

The results of this study indicate that ingestion ofbacterial antigen by humans resulted in the selectiveinduction of a humoral imnmune response that appearedsimultaneously in secretions of salivary and lacrymalglands. The dynamics of antibody production ob-served after the second series of antigen ingestion

TABLE IAnti-S. mnutatns OMZ-176 Agglutinin Titers of Selected Samples of Parotid Saliva, Tears, and Serum

after Augmentation with H-Chain-Specific Antisera (Aniti-a, Anti-y, and Anti-,a)

Stubject 1 Subject 2 Stubject 3 Subject 4

Period Parotid Tears Serum Parotid Tears Serum Paroticl Tears Serumii Parotid Tears Sertum

Preimmilunization 1(2,1,1)* <1(1,<1,<1) 1(1,2,2) <1(1,< 1,<1) <1(1,<1,<1) 2(2,4,2) 1(2,1,1) 1(1,<1,< 1) 2(2,4,3) 1(2,<1,< 1) 1(2,1,1) 2(1,4,4)Peak primary 5(8,5,5) 5(9,5,5) 2(2,4,4) 4(7,4,4) 6(8,6,6) 1(1,3,2) 3(8,5,5) 6(8,6,6) 2(2,4,4) 6(9,6,6) 7(8,7,7) 2(2.4,4)Postprimiary 3(5,3,3) 2(6,2,2) 3(3,4,4) 3(5,3,3) 3(6,3,3) 2(2,4,3) 2(4,3,3) 3(5,3,3) 2(2,4,4) 3(5,3,3) 4(6.4,4) 2(3,4,4)Peak secondarv 10(12,10,10) 9(12,9,9) 2(2,4,5) 7(9,7,7) 11(12,11,11) 2(2,3,3) 8(11,8,8) 9(11,9,9) 3(3,5,4) 9(12,9,9) 10(>12,110,10) 2(2,4,4)Postsecondarv 5(8,5,5) 4(7,4,4) 1(1,3,2) 3(6,3,3) 6(8,6,6) 2(2,3,3) 5(7,5,5) 5(9,5,5) 2(2.4,4) 5(8,5,5) 8(12,8,8) 3(3,4,5)

*Log2 agglutnllin titer determined by microtitrations with 2 x 10' CFU/ml of fornmalin-killed S. "mutans ONMZ-176. Ntimbers in parentheses represent the log, agglti-tinin titers after wvashing ancl reacting with H-chain-specific antisera (anti-a, anti-y, and anti-A, respectisely).

Selective Induction of a Secretory Immune Response 733

0-

'a

* 4

* a* N

x

0-]U

'U

Nnw

l

I

0o

a

734 Mestecky, McGhee, Arnold, Michalek, Prince, and Babb

x

x

a

11

implied that an anamniestic responise could be in-duced in secretions. The presence of these anti-bodies raises a question as to the mechanisms in-volved in their inductioni.

Several possible explanations may be offered for theappearance of these antibodies in glandular secretionsremote from the apparent site of antigenic stimulationi.An ingested antigeni could be absorbed from thegastrointestinal tract, enter the circulation, and afterdeposition in secretorx tissties, stimulate immuno-competent cells. This explanation is unlikely becausecirculatinig antigens should also induce a pronouncedserum antibody response; this had not been ob-served in our experimenits or in investigations per-formed in other laboratories (6, 17, 19, 29). Directintroduction of the anitigeni into circulation by i.v.immunization induced an excellent response in serumbut stimulate(d no s-IgA antibodies in secretions (5, 6,17). Further, if an imnmlunogeni is carried to secretorytissues by the circulation,, serum antibodies shouldactually prevent the anitigen from reaching the secre-tory tissues and inhibit a secretory immune response.Montgomery anid co-workers (6) have recently demon-strated that high titers of antibodies in serum ofsystemically (intravenously) immunized animals didnot prevent the induction of a secretory responseby subsequent oral immunization with the identicalantigen. The (leposition of anitigens (with or withoutadjuvanits) into the parenchyma of secretory glandsresults in production of antibodies associated with IgAas well as IgG; the IgG class is usually m11ore prevalent(15).

A more probable explanationi for the occurreince ofantibodies in glandtular secretions, as a consecluenceof the ingestioni of anitigenis, involves the selectiveseedinig of commllitted precursors of IgA-producingcells friom a cenitral site of anitigeniic stiniulationi toremote secretory glainds (6, 17-20, 29-32). Thishypothesis is based UpoIn several experimenital ob-servationis. It has been demonstrated that the micro-pinocytotic activity of' the epithelial cells coveringgut (or bronchial-)-associated lymphoid tissues allowsan active tranisport of' both soluble anid particulateantigenis or their fragmiienits to the underlying lymphoidcells (31, 33). Antigenic stimulationi iniluces the differ-entiation anid proliferation of bonie marrow-derived (B)and thymus-derived (T) cells, followed bv an exodusof these cells through ductus thoracicus into thegeneral circulationi. The conmmitted cells subsequentlv

Ec-0

0.5-

C70.4 -

0

02

0.rinlllIzctO frmoeoth ujes; ;ldlilstne

10T2

the absorbancee profile repreesent the log, aniti-S. niutaiils OMIZ-176 agglutitiili] titers of the pooled fractioin. Brokeni Iines indi-ecate the titers of the saiie saimple.s after atignentatioil withaniti-a ani.tiseral, a.s deteriiiiiied by the micirotitration techniqtie.Levels of IgA, Ig.\I, aiicl IgG were deterimined by sinigle raclialiimiiiiiiiodifftisioii anid proteiin conieeiitratioiisi fv the Lowr,%miethod (24).

settle preferentially in the secretory tisstie and differ-eintiate into IgA-producing plasnma cells. Gut- andbronchial-associated lymphoid tissue.s and n)eseitericlvmlph nodes are iindeed enriched sources of IgA pre-cuirsor cells that exhibit a strong preferential hominiigfor the intestiical and respiratory tracts as well as f'orsecretory glainds (32, 34-40). A conviiiciiig denionstra-tion of the f'unctioin of Peyer's patclhes as sites forantigen stinmlilation of IgA prectirsor cells wvas provided

FIGURE 1 Dynamics of the aggltitninii antib)ody respoinse to S. m71tutan.s ONIZ-176 in glaiidtilarseeretiolns aiter primairv (10) and secondary (20) oral immiuinization: (a) parotid saliva; (b) subman-clil)ular-sul)lingual and iminior gland saliva; (C) whole saliva: ((l) teirs. Inidividtial aggltitiniin titersare represenitedl by the following smbl)ols: 0, subject 1; A, sutbject 2; +, subject 3; anidl x, suibject4. The curve represenits the mieani log.2 agglutin*in titers for all samiples collected per 3-clay period.

Selectire Iinductioni of ai Secretorij Ininiuncl7e Response 735

TABLE IIAnti-S. mutans (Serotype a-se and g) Agglutinitn Titers in Selected ParotidSaliva Samples after Oral Immunization with S. mutans OMZ-176

S. mutans strain

Stubject Period AHT (a) BHT (b) 10449 (c) OMZ-176 (d)* LM-7 (e) 6715 (g)

I Preimmunization 6 8 7 1 9 9Peak secondary 7 8 7 10 8 10

2 Preimmunizationi 5 7 7 <1 8 5Peak secondary 5 6 6 7 7 7

3 Preimnmunization 8 9 9 1 7 5Peak secondarv 8 8 8 8 6 9

4 Preimmunizationi 6 6 1 7Peak secondary 5 5 9 7

* Immunization strain.Log2 agglutinin titer determined by microtitrations with 2 x 108 CFU/ml of formalin-killed

S. mutatns a-e and g (20).

by Robertson and Cebra (41). When a single Peyer'spatch in an isolated rabbit intestinal loop was stimu-lated with a bacterial antigen, antibodies were pro-duced in other parts of the intestine. However, theintroduction of the same antigen into an ileal loop,which lacked Peyer's patches, failed to induce animmune response in other portions of the gut. Theappearance of s-IgA-associated antibodies in saliva andmilk, but not in serum, as a consequence of natural(20) or artificial oral immunization (6, 17-19, 29, 30),indicates that this mechanism of s-IgA induction isalso operational in secretory glands. The inability toinduce s-IgA antibodies in secretions remote fromthe site where antigens were applied (2, 4, 8, 9) maybe explained by the lack of involvement of lymphoidtissues such as Peyer's patches, which have the po-tential to seed distant secretory tissues with sensitizedprecursors of IgA-producing cells.

Although there is no direct evidence concerningthe origin of IgA-precursors in salivary and lacrymalglands, a recent report indicates that precursors of IgA-producing plasma cells found in mammary glandsoriginate from gut-associated lymphoid tissues andmesenteric lymph nodes (40). The suggestion thatprecursors of IgA-producing plasma cells of humansalivary glands may indeed originate at distant sitesis based on studies of patients undergoing extra-corporeal cesium irradiation of blood. A dramatic de-crease in salivary IgA levels was displayed, whereasthe serum levels remained unchanged (42).

The parallel appearance of antibodies at anatomi-cally remote secretory sites suich as salivary and lacry-mal glands (in the apparent absence of a serumresponise) as a con-sequlenice of the ingestion of bac-terial anitigen suggests the existence of a common

mechanism for the induction of an immune responseon mucosal surfaces. The reason for the absence ofa pronounced serum response after oral immuniza-tion is not clear; however, the induction of a selectivesystemic hyporesponsiveness could be considered (43).

ACKNOWLEDGMENTSWeare grateful to Drs. F. W. Kraus and H. W. Fullmer andMs. C. Sims for editorial advice, Ms. R. Kulhavy for excellenttechnical assistance, Dr. H. Dillon for the anti-streptolysinO titrations, and James Gerard for data analysis. Thanksis also expressed to Ms. A. Carlisle for typing this manu-script.

This research was supported by U. S. Public HealthService grants AI-10854, CA-13148, DE-04217, and DE-02670.

REFERENCES

1. Tomasi, T. B. 1976. The Immune system of Secretions.Prentice-Hall, Inc., Englewood Cliffs, N. J.

2. Ogra, P. L., R. B. Wallace, and G. Umana. 1974. Implica-tions of secretory immune system in viral infections.Adv. Exp. Med. Biol. 45: 277-282.

3. Ganguly, R., P. L. Ogra, S. Regas, and R. H. Waldman.1973. Rubella immunization of volunteers via the respira-tory tract. Infect. Immun. 8: 497-502.

4. Waldman, R. H., and R. Ganguly. 1974. The role of the se-cretory immune system in protection against agents whichinfect the respiratory tract. Adv. Exp. Med. Biol. 45: 283-294.

5. Eddie, D. S., M. L. Schulkind, and J. B. Robbins. 1971.The isolation and biologic activities of purified secretoryIgA and IgG anti-Salmonella typhimurium 'O' antibodiesfrom rabbit intestinal fluid and colostrum. J. Immunol.106: 181-190.

6. Montgomery, P. C., I. Lemaitre-Coelho, and E. T. Lally.1976. The effects of circulating antibodies on secretoryIgA antibody induction following oral immunization withdinitrophenylated Pneumococcus. Ric. Clin. Lab. 6(Stippl. 3): 93-99.

736 AMestecky, AMcGhee, Arniold, Michalek, Prince, acnd Babb

7. Brandtzaeg, P., and K. Tolo. 1977. Iticosal penetrabilityenhaniced l) seruim-derived antibodies. Natture (Louid.).266: 262-263.

8. Ogra, P. L., D. T. Karzon, F. Righthand, anid MI. MacGil-livrav. 1968. Immunoglobulin response in serumii anld se-cretionis alter iminiiiizatioin with live aud incactivated po-liovaccine and niatuiral inifection. N. Euigl. J. M1ed. 279:89:3-900.

9. Shlvartsimnan, Y. S., E. N. Agranovskaya, anid I. P. Zykov.1977. Forimiatioin of secretory and circtilating antibodiesafter imiminiltizationi with live and inactivated influenizavirtus vaccinies.J. Inifect. Dis. 135: 697-705.

10. Genico, R. J., and I. A. Taubman. 1969. Secretory yAantibodies induced by local imiiiniizationi. Nature(Louid.). 221: 679-681.

11. Taubman, NI. A., and R. J. Genco. 1971. Inductioni andproperties of ral)lit secretorx yA antibody directed togrouip A streptococcial carl)ohvdrate. Ii ininii uochem1mistry.8: 1137-1155.

12. Taubman, I. A., and D. J. Siith. 1974. Effects of localimiuiniitizationi with Streptococcus m11utauis on iniduictioni ofsalivary inmmunoglobulin A antibody and experimentaldlenital caries in rats. Infect. Inlnmun. 9: 1079-1091.

13. MlcGhee, J. R., S. NI. NMichalek, J. Webb, J. NI. Navia,A. F. R. Rahmiiain, and D. NV. Legler. 1975. Effective im-munitv to denital caries: protectioin of gnotobiotic rats bylocal immuniitiiiizatioin with Streptococcus tin utants. J. I in-111untlol. 114: 300-30.5.

14. Emmiings, F. G., R. T. Evans, aind R. J. Genco. 1975.Antibody responise in the parotid fluid anid serumil of irusmonkeys (MVlacaca fa.sciculari.s) after local immIunllizationlwith Streptococcus in1tutatns. In1fect. Im71t11unl. 12: 281-292.

15. Hurlinuiannl, J., and NI. Lichaa. 1976. Local immunizationiin the manmmary glands of the rabbit. J. Inin 7unlol. 116:1295-1301.

16. Crawford, J. NI., NI. A. Taubman, aI)d D. J. Smith. 1975.Mlinor salivary glandCls as a major source of secretory im-mulloglobtulin A in the humilan oral cavity. Scietnce (Wash.D. C.).190: 1206-1209.

17. Mlontgomerv, P. C., J. Cohn, and E. T. Lallv. 1974. TheilidItictionl and characterizationi of secretory IgA anti-bodies. Adv. Exp. Mled. Biol. 45: 453-462.

18. Bohl, E. H., L. J. Saif, R. K. P. Gupta, and G. T. Frederick.1974. Secretory antibodies in milk of swine against trans-missible gastroenteritis virus. Adv. Exp. Med. Biol. 45:337-342.

19. Goldblumn', R. NI., S. Ahlstedt, B. Carlsson, L. A. Hanson,U. Jodal, G. Lidin-janison, and A. Sohl-Ackerlunid. 1975.Antibody-forming cells in humain colostrumii after oral im-munization. Nature (Loud.). 257: 797-798.

20. Arnold, R. R., J. Nlestecky, and J. R. NMcGhee. 1976. Na-turally occuirring secretory imniunoglobulin A antibodiesto Streptococcus nmutatns in hutmiian colostrum anid saliva.Inifect. Iminiiiuni. 14: 355-362.

21. Schaefer, I. E., I. Rhodes, S. Prince, S. N. lichalek,and J. R. NcGhee. 1977. A plastic intraoral device for thecollection of human parotid saliva J. Dent. Res. 56:728-733.

22. Carlsson,J., E. Newbrun, and B. Krasse. 1969. Purificationand properties of dextran-sucrase from Streptococcus sani-guis. Arch. Oral Biol. 14: 469-478.

23. Nestecky, J., R. Kulhavy, and F. W. Kraus. 1972. StidiesOnl humani secretory immunoglobulini A. II. Subunit struc-ttire.. Iimuntol. 108: 738-747.

24. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J.Ranidall. 1951. Protein measurement with Folin phenolreagent.]. Biol. Chemi. 193: 265-275.

25. Mestecky, J., F. W. Kraus, and S. A. Voight. 1970. Propor-

tion of human colostral i mmunoglobulin-A mlolecuilescoIntaininiig the secretory deterimiinant. It7ntmun0ology. 18:237-243.

26. Kraus, F. WV., anld J. Mlestecky. 1971. Imimiuniohistochlelmli-cal localizationi of amvlase, lvsozvme aind immunoglobu-linis in the humilcani parotid gland. Arch. Oral Biol. 16: 781-790.

27. Zabriskie, J. B., anid E. H. Freiimier. 1966. Ain iimimunio-logical relatioinship between the group A streptococcuisacndl milammiiiialiani muscle. J. Exp). Aled. 124: 661-678.

28. Perch, B., E. Kjems, acnd T. Ravn. 1974. Biochemical andserological properties of Streptococcus ?n utanus f'romI vari-ous humlan and aninmal sources. Acta Pathol. Mllicrobiol.Scand. 82: 357-370.

29. Mlichalek, S. NM., J. R. NMcGhee, J. NMestecky, R. R. Arnold,and L. Bozzo. 1976. Ingestioni of Streptococcus mizutauisinduices secretory immunoglobulin A and caries immnu-nity. Scienice (Wzash. D. C.). 192: 1238-1240.

30. Ahlstedt, S., B. Carlsson, L. A. Hanson, and R. NI. Gold-blulm. 1975. Antibody productioni by human colostralcells. I. Immunoglobulin class, specificity, aind quantity.Scatd. J. Inini titnol. 4: 535-539.

31. Cooper, NI. D., P. WV. Kincade, D. E. Bockmiiani, and A. R.Lawvton. 1974. Origin, distribution and differenitiationi ofIgA-producing cells. Adv. Exp. Med. Biol. 45: 13-22.

32. Cebra, J. J., P. J. Gearhart, R. Kamat, S. NI. Robertson,and J. Tseng. 1977. Origin anid differentiation of lympho-cvtes involved in the secretory IgA response. Cold SpritngHarbor Sijmip. Quantt. Biol. 41: 201-215.

33. Bocknmani, D. E., and NI. D. Cooper. 1973. Pinocytosis byepitheliunm associated with lymphoid follicles in the bursaof Fabricius, appendix and Peyer's patches. An electronmicroscope study. Amtt. J. At at. 136: 455-477.

34. Hall, J. G., and NI. E. Smith. 1970. Homing of lxmph-bornie immunoblasts to the gut. Nature (Lond.). 226: 262-263.

35. Craig, S. XV., and J. J. Cebra. 1971. Peyer's patches: anieniriched source of precursors for IgA-producing inmmuno-cvtes in the rabbit. J. Exp. MAed. 134: 188-200.

36. Nloore, A. R., and J. G. Hall. 1972. Evidencee for a primaryassociation betweeni immuninoblasts and small gut. Nature(Lotud.). 239: 161-162.

37. Guy-Grand, D., E. Griscelli, and P. Vassalli. 1974. Thegut-associated lyimphoid system: nature and properties ofthe large dividing cells. Eur. J. Inimnititol. 4: 435-443.

38. Rudzik, R., R. L. Clancy, D. Y. E. Perey, R. P. Day, andJ. Bienenstock. 1975. Repopulationi with IgA-containingcells of bronchial and intestinal lamina propria after trans-fer of homologous Peyer's patch and bronchial lympho-cytes.J. I mmunol. 114: 1599-1604.

39. NicWilliams, NM., J. MI. Phillips-Quagliata, and NI. E.Lamm. 1977. Nlesenteric lymph node B lymphoblasts whichhome to the small intestine are precommitted to IgA syn-thesis. J. ExP. Med. 145: 866-875.

40. Lamm, NI. E., NM. E. Roux, P. Weisz-Carrington, NI.NMcWilliams, and J. NM. Phillips-Quagliata. 1978. Cellularorigin of the IgA in milk. Prot. Biol. Fluids Proc. Colloq.Bruiges. 25: 835-841.

41. Robertson, S. NI., and J. J. Cebra. 1976. A model for localimmunity. Ric. Cliii. Lab. 6(Suppl. 3): 105-119.

42. Friedman, B. K., and B. Greenberg. 1975. Effect of extra-corporeal cesium irradiation on secretory and serum im-munoglobulins.J. Dent. Res. (Spec. Issue A) 54: 68. (Abstr.96).

43. Andre, C., J. F. Heremans, J. P. Vaerman, and C. L.Cambiaso. 1975. A mechanism for the inductioni of im-munological tolerance by antigen feeding: antigein-anti-body complexes. J. Exp. Med. 142: 1509-1519.

Selective Induction of a Secretory Immune Respontse 737