INFECTION AND IMMUNITY, Jan. 1976, p. 195-203 Copyright C) 1976 American Society for Microbiology

Vol. 13, No. 1 Printed in U-SA.

Distribution of Cholera Organisms in Experimental Vibrio cholerae Infections: Proposed Mechanisms of Pathogenesis and Antibacterial Immunity GORDON D. SCHRANKl* AND W. F. VERWEY Department of Microbiology, University of Texas Medical Branch, Galveston, Texas 77550 Received for publication 7 July 1975

This study was undertaken to determine the sequence of events in the microenvironment of the intestinal tract that culminate in the symptoms of cholera and to attempt to define more clearly the mechanisms involved in antibacterial immunity. The extent to which mucus occurs in the normal intestine of rabbits and the appearance of the intestinal villi in unfixed frozen sections was demonstrated. The villi and intervillous spaces were found to be normally covered by a layer of mucoid material that formed a mucous zone between the intestinal contents and the tips of the villi. The distribution of cholera organisms in normal and immunized animals was demonstrated by the staining of frozen-tissue sections with specific fluorescent antibody. Study of tissue sections from normal animals showed that the onset of fluid accumulation was concomitant with the establishment of large masses of organisms in the intervillous spaces and crypts of the intestine after the successful penetration of this mucous zone. Tissue sections from animals actively or passively immunized against a cell wall antigen of Vibrio cholerae showed clumping of vibrios in the lumen and restricted distribution in the lumen and luminal border of the mucous zone. Antibody was not lytic in vivo.

In recent years, considerable attention has been given to production of more effective vaccines against cholera and the study of mechanisms of action of cholera toxin on mucosal cells (5). The effectiveness of antibacterial immunity has been quite thoroughly proved, though the mechanisms involved are still subjects of speculation. Earlier studies in this laboratory (25) provided an in vitro demonstration ofthe inhibition of mobility of motile vibrios in soft agar containing complement-free antisomatic antibody. These findings provided a basis for a hypothesis of antibacterial immunity in cholera: in the intestine immunity prevents the migration of cholera organisms into the intervillous spaces (24). The objective of this work was to determine the pattern of distribution of vibrios within ligated ileal loops of normal rabbits and rabbits actively and passively immunized with purified somatic antigen (18, 23) and its corresponding antibody. Special tissue handling techniques (20) were employed to maintain normal intestinal relationships from the time when loops were removed from animals until the tissues were frozen and sectioned. 1 Present address: Department of Microbiology, The University of Tennessee Center for the Health Sciences, Memphis, Tenn. 38163.

195

(These experiments represented a portion of a dissertation submitted by G. D. Schrank to the University of Texas Medical Branch in partial fulfillment ofthe requirements for the Ph. D degree.) MATERIALS AND METHODS Animals. White, female New Zealand outbred rabbits weighing 3 to 5 kg were used throughout these experiments. Animals with no evidence ofdiarrheal disease were used after they had been allowed to acclimate for 2 weeks at the laboratory animal care facility of the University of Texas Medical Branch. Bacteria. Vibrio cholerae (eltor) Inaba strain V86, a stock culture of this laboratory, was used throughout these experiments. Organisms were maintained at -20 C in the lyophilized state. Each week, the organisms were suspended in heart infusion broth (Difco), and smooth colonies were subcultured twice on heart infusion agar plates at 37 C. The second of these plates was retained as a source of organisms throughout the week. Isolated "typical" colonies were streaked for confluent growth on heart infusion agar and incubated at 37 C for 4 h. Vibrios were harvested with 5 ml of sterile 0.005 M phosphate-buffered normal saline with gelatin (PBSG, 0.1% gelatin [wt/vol], pH 7.4) and adjusted to 170 Klett units (green filter, Klett-Summerson photometer). This reading represents an optical density of 0.340 and approximately 1.3 x 109 viable

196

SCHRANK AND VERWEY

INFECT. IMMUN.

y44 I

'

me

'

-1w

4,

FIG. 1. Frozen-tissue section from an unwashed and uninoculated ileal loop ofa normal adult rabbit. Tips of villi are seen at the upper left of the photograph, whereas gut contents are seen on the right half with the intermediate mucous zone (x125).

C

S At~~~AW

'~va

kr

FIG. 2. Tissue section from an ileal loop ofa normal adult rabbit inoculated with a5% suspension ofcarbon particles. Tips of villi are seen in the upper portion of the photograph, whereas dark areas of accumulated carbon are seen at the bottom of the photograph with the intermediate mucous zone (x125).

vibrios per ml for organisms grown under these conditions. Immunization. A purified Inaba antigen (IN3) prepared in this laboratory (23) was employed. Hyperimmune antiserum preparation and active immunization were accomplished by injecting adult rabbits subcutaneously in the nuchal region and intra-

venously in the marginal ear vein with 0.5 ml ofIN3 according to the following schedule: days 1 and 13, 20 ,ug/ml subcutaneously; days 5, 9, 17, and 21, 40 /g/ml intravenously. Ten days after the last injection, animals were bled for hyperimmune serum or were challenged by the ileal loop technique. A single pool of hyperimmune rabbit antiserum was used for

VOL. 13, 1976

CHOLERA ORGANISMS IN V. CHOLERAE INFECTIONS

TABLE 1. Relationship between organisms and the intestine in nonimmune animals

197

nate. Labeling with fluorescein isothiocyanate was accomplished by the method of Goldman (11) and was followed by dialysis to remove unreacted dye Location of organisms and fractionation on diethylaminoethyl-cellulose to Clump- Gross minimize nonspecificity of the staining reaction. All Time Mu- Intervil- Lu- ing of fluid acprotein determinations were made by the Lowry mens organ- cumula(h) Lu- cous OUS of ismsa tionb IOUs method (17). mencous zone spaces crypts Vibriocidal antibody. Vibriocidal antibody determinations were made according to the procedure of _ _ _ _ Verwey et al. (22). +c + 0 Operative procedure. After they had been fasted + + 1 _ _ _ for 72 h (water ad libitum), animals were anesthe+ + 2 + (1/2)d sized with sodium pentobarbital (Abbott Laborato+ (3/4) + + 3 _ _ _ ries) at a dose of 35 mg/kg diluted to a final volume + + + + 4 + _ of 10 ml in sterile normal saline. Ileal loops were + + + + + 5 prepared with slight modifications of earlier meth+ + + + + 6 ods (4). Eight ileal loops approximately 8 cm in + + + + + 7 length were prepared in each animal. The loops + + + + 8 + were inoculated by means of a 22-gauge needle at+ + + + + 9 tached to a 1-ml disposable syringe. Starting in the a See Fig. 5 and 6. distal ileum, loops 1, 3, 4, 6, 7, and 8 were infected b Two milliliters of fluid or more recovered. with viable vibrios, whereas loops 2 and 5 were c +, At least 10 organisms seen in area indicated treated as controls and injected with PBSG. Inocula on any section. Five sections were studied from each were chosen to provide adequate visualization of the organism with fluorescent antibody in tissue secileal loop. d Refers to the relative amount of the total length tions. All infected loops examined at a given time of the intervillous spaces that contained organisms. period received the same inoculum in normal and immunized animals. For loops studied at zero time, 1.3 x 108 vibrios were used in the infecting dose; for three types of passive immunization: type 1, intra- those studied at 1, 2, or 3 h, 1.3 x 107 vibrios; and for vascular administration of 6 ml of hyperimmune study at 4, 5, 6, 7, 8, or 9 h, 1.3 x 106 vibrios were serum 24 h before surgery; type 2, mixing of 0.3 ml of used. Treatment of intestinal loops. Ileal loops reantiserum with 0.7 ml of a dilution of inoculum to give a final inoculum of 1 ml containing 1.3 x 106, moved from animals were assigned reference num1.3 x 107, or 1.3 x 108 viable vibrios; or type 3, bers and treated by the tissue handling procedures exposure of each 1-ml aliquot of inoculum to 0.6 ml developed by Savage et al. (20). Intact ileal loops of antiserum, followed by three washings. The wash- were suspended in aluminum-foil cylinders containings were made with sterile PBSG after centrifuga- ing 2% methylcellulose, 15 CP (Fisher) in 0.15 M tion at 5,900 x g at 4 C. This pool of antiserum had a saline. The tissue and supporting medium were frovibriocidal titer (22) of 1:204,800. All sera were inac- zen in liquid nitrogen. Tissue sections were cut at 6 tivated at 56 C for 30 min immediately before use. ,am on a microtome-cryostat (International EquipPurification and labeling of antibodies. Hyper- ment Co.) and stained with fluorescent antibody by immune serum was fractionated by the method of the direct technique, after fixing in absolute methaCherry (3). Specific antibodies were absorbed and nol for 90 s. All tissue sections were counterstained eluted from heat-killed cells. Approximately 1.95 x with Evan blue (0.01% solution, Eastman Kodak 1010 heat-killed V. cholerae strain V86 cells were Co.). Other loops were used for viable-count determiwashed three times with cold 0.2 M glycine-hydro- nations. chloride buffer (pH 3.25) and three times with cold Determinations of viable V. cholerae recovered 0.05 M tris(hydroxymethyl)aminomethane-hydro- from ileal loops were made by a microdrop method chloride buffer (pH 7.5). These cells were suspended on dextrin heart infusion agar plates, a noninhibiin 30-ml aliquots of the globulin preparations and tory differential medium for V. cholerae (21). Serial kept for 2 h at 4 C. The cells were washed with cold 10-fold dilutions of fluids were made in PBSG. An Tris-ethylenediaminetetraacetic acid buffer (pH 7.5) 18-gauge blunt needle (slightly bent) on a 1-ml tuber(6) until the supernatant showed an optical density culin syringe was used to deliver 6 drops of fluid to of 0 at 280 nm (Beckman spectrophotometer). The each petri plate, starting with the highest dilution. specific antibody was eluted by washing with 1-ml The drops were permitted to fall from the end of the aliquots of cold 0.2 M glycine-hydrochloride buffer needle with the orifice parallel to the culture me(pH 3.25). When the optical density at 280 nm of the dium surface. Each drop delivered in this manner supernatant reached 0, the pH of the solution was approximated 0.02 ml. The plates were allowed to adjusted to 7.0 by titration with 4 N NaOH. The stand at room temperature for 30 min and then neutralized solution was dialyzed for 12 h against incubated at 37 C for 18 h without inversion. The Tris-ethylenediaminetetraacetic acid buffer (pH 7.0) numbers of drops on countable plates (5 to 40 coloand concentrated by the use of 20 M Carbowax (Un- nies/drop) were averaged and then multiplied by 50 ion Carbide). The solution was then dialyzed aginst to give colony-forming units per milliliter of original PBSG before labeling with fluorescein isothiocya- fluid. Counts could be determined in this manner

198

SCHRANK AND VERWEY

INFECT. IMMUN.

FIG. 3. Frozen-tissue sections stained with fluorescent antibody and counterstained with Evan blue. All sections shown were from loops infected with V. cholerae that were taken from normal animals. (A) Tissue section showing the central lumen of ileum at zero time with fluorescing organisms. Fluorescing organisms appear as isolated bacterial cells (x550). (B) Tissue section showing two villi and an intervillous space (arrow) at zero time. No organisms are present in this region at this time period (x550).

since V. cholerae organisms occur as single cells. Comparison counts of organisms obtained from the intestinal lumen and those of organisms adjacent to the intestinal mucosa were made by a method similar to that of Freter (10). Sterile 30-ml homogenizing flasks (VirTis) containing 30 ml of brain heart infusion broth (Difco) were prepared, and after removal of the ligated loop 10 ml of broth from the flask was injected into the lumen of the loop, which was gently washed by inverting the closed segment four to five times.

Loops distended with accumulated fluid were first emptied and then washed with an amount of broth such that secreted fluid plus the wash was equivalent to 10 ml. The loops were opened at one end and the wash fluids were emptied into a homogenizing flask. Determination of the number of organisms remaining in the mucous zone was made by homogenizing the washed tissue loop in brain heart infusion. This broth is used in the vibriocidal assay (22) because of its anticomplementary nature. It was likewise used in the system described here to pre-

VOL. 13, 1976

CHOLERA ORGANISMS IN V. CHOLERAE INFECTIONS

vent low counts that might result if complement were released into the system by homogenization of ileal loops. Demonstration of an intestinal mucous zone. Ileal loops were prepared as before in normal animals. These loops were either removed from animals without washing, injection, or manipulation and frozen, or they were injected with a 5% carbon suspension (Darco G-60, Atlas Powder Co.). Injected loops

0

199

were removed from animals after 1 h and frozen. The untreated loops were taken from animals that had not been fasted. RESULTS

Critical to these studies was the demonstration of a gel-like mucous zone separating the villi of the rabbits from particulate intestinal

;I '~,%X }s';.t.2

IL

_

:.

FIG. 4. Frozen-tissue sections stained with fluorescent antibody and counterstained with Evan blue. All sections shown were from loops infected with V. cholerae that were taken from normal animals. (A) Tissue section showing the central lumen of ileum at 4 h with fluorescing organisms in association with intestinal contents (x550). (B) Tissue section showing two villi and an intervillous space (arrow) at 4 h. Organisms can be seen in the intervillous spaces at this time. With increasing time, tissue sections show increasing numbers of organisms in the intervillous spaces (x550).

200

SCHRANK AND VERWEY

INFECT. IMMUN.

FIG. 5. Frozen-time section stained with fluorescent antibody and counterstained with Evan blue. The section shown was from a loop infected with V. cholerae that was taken from a normal animal. The organisms appear as isolated fluorescing rods at the tips of the villi in the mucous zone (section taken 3 h postinfection) (x550).

FIG. 6. Frozen-tissue section stained with fluorescent antibody and counterstained with Evan blue. The section shown was from a loop infected with V. cholerae that was taken from an actively immunized animal. The organisms appear as clumps offluorescing rods at the tips of the villi in the mucous zone (section taken 3 h postinfection). Similar results were obtained in passively immunized animals (x550).

contents. As can be seen in Fig. 1, such an area existed within the ileum of the rabbit. This cross section demonstrated an area like the intervillous spaces that was free of intestinal

contents. This area extended about 100 ,um above the tips of the villi in an unstained and unfixed section of tissue as it appeared immediately after sectioning. Toluidine blue- or thio-

201

CHOLERA ORGANISMS IN V. CHOLERAE INFECTIONS

VOL. 13, 1976

TABLE 2. Relationships between organisms and intestine in actively and passively immunized animals Location of organisms

Time (h)

0

1

2 3 4 S 6 7 8 9

Lumen

Mucous zone

1

2

3

A 1

2

Clumping of organismsa

Lumens of crypts

Intervillous spaces

3

A 1

2

3 A

1

A 1

2 3c

A

+ + + + ++

+ +d + + ++

+ + + + + + + + + +

+ +.++ ++.+ + - ++.+ - - + + ++ - - + + ++ - - +

+ + + + ++

+ + - + + +.+ - - + + ++

++++ ++++

+ +

+ +

++ ± + + + + +++ + +++

+ + + +

2

3

Gross fluid accumulationb

A

1

2

3

-4-+

-

+ + +

+ + +

+

+

+ + + ++ + + -+ ++ + + ± -

a See Fig. 5 and 6. 'Two milliliters of fluid or more recovered. c A, Actively immunized animals; 1, passively immunized animals (type 1, intravenously injected); 2, passively immunized animals (type 2, hyperimmune serum mixed with infecting inoculum); 3, passively immunized animals (type 3, sensitized vibrios injected into ileal loops). d +, At least 10 organisms seen in area indicated on any section. Five sections were studied for each ileal loop.

TABLE 3. Comparison of geometric means (log values) of viable cells recovered from the lumen of ileal loops in normal and immune animals imani- Actively immu- Passively Time (h) Normal ml aninized animals umunized mals mals

0 1 2 3 4 5 6 7 8 9

6.50a 6.12 6.94 7.56 7.71 8.36 8.35 8.64 8.75 8.76

6.18 6.47 7.20 7.27 7.27 7.03 7.27 7.49 7.70 8.01

6.72b 6.93 7.25 7.45 7.65 7.24 7.37 7.59 7.98 8.37

a Eight ileal loops used for each geometric mean given for normal and actively immunized animals. b Twelve loops used for the values given for passively immunized animals.

nine-stained fixed-tissue sections gave a characteristic mucin staining reaction in this region. The ability of the mucous gel to restrict carbon particles to the central lumen is shown in Fig. 2. These sections further demonstrate that in reality the intervillous spaces were small and the villi were quite plump, in contrast to the impression gained from the usual histological preparations. Similar findings could be demonstrated throughout the full length of rabbit ileum and jejunum. The mucous zone did not increase or diminish in fasted or unfasted animals.

The distribution of vibrios in the small intestine was first studied in normal animals. Sequential groups of animals were studied from zero time to 9 h postinfection. The change in distribution of organisms is shown in Table 1 and also demonstrated in Fig. 3 and 4. The effectiveness of antisomatic antibody in preventing fluid accumulation has been previously described (10, 13). This protection can be demonstrated with either actively or passively immunized animals (9, 10). The primary influence of antibody upon vibrio distribution is shown in Fig. 5 and 6. In Fig. 5, which shows tissue from a normal animal, vibrios are seen to be distributed individually in the lumen of the ileum. In marked contrast, note the pronounced agglutination of vibrios in the lumen of an immune animal in Fig. 6. This clumping of organisms occurred in both actively and passively immunized animals. Table 2 shows the restriction of vibrios to the lumen of the intestine in these immunized animals. The occurrence of vibrios at the tips of villi in some sections likely represented variability in the thickness and consistency of the mucous zone, as well as the movement of organisms into portions of the mucous zone where sufficient quantities of antibodies resulted in clumping of organisms. Loops were protected regardless of the route by which antibody gained access to the system. In vivo killing of organisms was not necessary for protection to be apparent. These findings, as shown in Table 3, are in agreement

202

INFECT. IMMUN.

SCHRANK AND VERWEY

TABLE 4. Comparison of geometric means (log values) of viable cells recovered from the intestinal lumen and mucous zone in normal and immune animals Passively immunized animals (type 2) Time (h) Lumen count

0 1 2 3 4 5 6 7 8 9

Mucous zone count

6.50a

3.84b

6.12 6.94 7.56 7.71 8.36 8.35 8.64 8.75 8.76

5.03 5.07 6.81 7.01 6.81 7.25 7.74 7.64 7.74

Lumen reaction count + +c

+ + + +

6.42a 6.82 7.15 7.20 7.49 7.75 7.79 7.88 8.41 8.88

Mucous zone count

Loop

reaction

2.11b

-

3.71 3.95 4.79 4.82 4.44 4.54 4.62 5.03 6.14

-

a Eight ileal loops used for each geometric mean given. bFour ileal loops for each geometric mean given. c Two milliliters of fluid or more recovered.

with earlier investigations (2, 10) concerning independence of fuctional immunity and bactericidal activity. No significant difference existed between the numbers of viable cells recovered at any time, as determined by analysis of variance. As seen in Table 4, luminal viablecell concentration was not significant; rather, cell counts obtained from the mucous zone and intervillous spaces were critical to determining whether or not infected loops became positive. For all time periods, a significant difference existed between counts from the lumen and mucous zone in immunized animals. However, the difference between the two counts was significant only for 0, 1, and 2 h in normal animals. Fluid accumulation became evident in these animals at 4 h.

16). These sections were made from intestinal tissue that had been split longitudinally, rolled, frozen, and sectioned. The change in distribution of vibrios from the intestinal lumen to the intervillous spaces was not clearly demonstrated because of this manipulation of the tissue. The tissue handling methods we used in our studies permitted less manipulation of the tissue. Our studies indicated that clumping of organisms in the mucous zone by somatic antibody resulted in protection of infected loops (Table 2). The close association between vibrios and the intestinal mucosa would allow for the most efficient release of toxin, since the toxin specifically and selectively adsorbs to the intestinal mucosa (19). The restriction of cholera organisms to the intestinal lumen would greatly decrease the efficiency of toxin delivery to susceptible cells. Some authors (12, 25) have indicated that motile toxigenic cholera vibrios are more virulent than nonmotile toxigenic strains. From the data presented here, we conclude that the increased virulence exhibited by the motile strains could be a result of the enhanced ability of these organisms to transverse the mucous zone. To reach susceptible host cells, organisms must first penetrate the mucous zone. Immunity with antisomatic antibody prevented this free, random distribution of bacteria that were consequently seen as clumps of organisms in the intestinal lumen and mucous zone. It is evident that antibacterial immunity did not function by killing of organisms (Table 3), but rather by affecting distribution or mobility of vibrios (Tables 2 and 4). Although antisomatic antibody appeared to inhibit mobility of cholera organisms (1) by reacting with the entire organism (25), including the flagellum (Follett and Gordon [7] hypothesized that the exterior portion of the cholera flagellum was of cell wall origin), vibrios grown in the presence of antisomatic antibody in broth did produce toxin (G. Schrank and C. Stager, unpublished observations). These studies were carried out in ileal loops; but when the intestinal tract is patent, organisms existing only in the more central mucous zone and lumen of the intestine may be either removed from the area by normal intestinal flow and eventually pass to the outside or exposed to the influences of the colonic bacterial flora. The above discussion of the role of antisomatic antibody does not exclude a possible additive role for antitoxic immunity.

DISCUSSION The demonstration in these experiments of the presence and extent of the mucous zone (Fig. 1 and 2) provided a basis for hypothesizing a mechanical barrier that prevented or at least slowed the penetration of vibrios into the intervillous spaces. Earlier workers (9, 10, 14-16) have suggested that an association between vibrios and the intervillous spaces and crypts is necessary for disease production. The results of our experiments support this assumption (Table 1) and provide a view of the establishment ACKNOWLEDGMENTS of this host-parasite relationship (Table 1; Fig. 3 and 4). Previous workers have used fluoresThis study was supported in part by the James W. Mccent antibody-stained frozen-tissue sections (10, Laughlin Fellowship Fund and by Public Health Service

VOL. 13, 1976

CHOLERA ORGANISMS IN V. CHOLERAE INFECTIONS

Grant AI-08518 from the National Institute of Allergy and Infectious Diseases.

1. 2.

3.

4.

5. 6.

7.

8.

9. 10.

11.

12. 13. 14.

15.

LITERATURE CITED Beneson, A. S., M. R. Islam, and W. B. Greenough III. 1964. Rapid identification of Vibrio cholerae by darkfield microscopy. Bull. W. H. 0. 30:827-831. Burrows, W., M. E. Elliott, and I. Havens. 1947. Studies on immunity to Asiatic cholera. IV. The excretion of coproantibody in experimental cholera in the guinea pig. J. Infect. Dis. 81:261-281. Cherry, W. B. 1970. Fluorescent antibody techniques, p. 693-704. In J. E. Blair, E. H. Lennette, and J. P. Turant (ed.), Manual of clinical microbiology. The Williams & Wilkins Co., Baltimore. De, S. N., and D. N. Chatterjee. 1953. An experimental study ofthe mechanism of action of Vibrio cholerae on the intestinal mucous membrane. J. Pathol. Bacteriol. 66:559-562. Finkelstein, R. A. 1973. Cholera. Crit. Rev. Microbiol. 2:553-623. Finkelstein, R. A., and J. J. LoSpalluto. 1969. Pathogenesis of experimental cholera: preparation and isolation of choleragen and choleragenoid. J. Exp. Med. 130:185-202. Follett, E. A. C., and J. Gordon. 1963. An electron microscope study of vibrio flagella. J. Gen. Microbiol. 32:235-239. Freter, R. 1964. Comparison of immune mechanisms in various experimental models of cholera. Bull. W. H. 0. 31:825-834. Freter, R. 1969. Studies on the mechanism of action of intestinal antibody in experimental cholera. Tex. Rep. Biol. Med. 27(Suppl. 1):299-316. Freter, R. 1970. Mechanism of action of intestinal antibody in experimental cholera. II. Antibody-mediated antibacterial reaction at the mucosal surface. Infect. Immun. 2:556-562. Goldman, M. 1968. Fluorescent antibody methods. Academic Press Inc., New York. Guentzel, M. N., and L. J. Berry. 1975. Motility as a virulence factor for Vibrio cholerae. Infect. Immun. 11:890-897. Kaur, J., W. Burrows, and M. A. Furlong. 1971. Immunity to cholera: antibody response in the lower ileum of the rabbit. J. Infect. Dis. 124:359-366. Lankford, C. E. 1960. Factors of virulence of Vibrio cholerae. Ann. N. Y. Acad. Sci. 88:1203-1212. Lankford, C. E., and U. Legsomburana. 1965. Viru-

16.

17.

18.

19. 20.

21. 22.

23.

24.

25.

203

lence factors of choleragenic vibrios, p. 109-121. In Proceedings of the cholera research symposium. Government Printing Office, Washington, D.C. LaBrec, E. H., H. Sprintz, H. Schneider, and S. B. Formal. 1965. Localization of vibrios in experimental cholera: a fluorescent antibody study in guinea pigs, p. 272-276. In Proceedings of the cholera research symposium. Government Printing Office, Washington, D.C. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193:265-275. Mosley, W. H., K. M. S. Ariz, A. S. M. M. Rahman, A. K. M. A. Chowdbury, A. Ahmed, and J. C. Feeley. 1970. The 1968-69 cholera vaccine field trial in rural East Pakistan: effectiveness of monovalent Ogawa and Inaba vaccines and a purified Inaba antigen, with comparative results of serological and animal protection tests. J. Infect. Dis. 121(Suppl. 1):51-59. Peterson, J. W., J. J. LoSpalluto, and R. A. Finkelstein. 1972. Localization of cholera toxin in vivo. J. Infect. Dis. 126:617-628. Savage, D. C., R. Dubos, and R. W. Schaedler. 1968. The gastro-intestinal epithelium and its autochthonous bacterial flora. J. Exp. Med. 127:67-81. Schrank, G. D., C. E. Stager, and W. F. Verwey. 1973. Differential medium for Vibrio cholerae. Infect. Immun. 7:827-829. Verwey, W. F., Y. Watanabe, J. C. Guckian, H. R. Williams, Jr., P. E. Phillips, S. S. Rocha, Jr., and E. B. Bridgeforth. 1969. Serological responses of human volunteers to cholera vaccine. Tex. Rep. Biol. Med. 27(Suppl. 1):243-274. Verwey, W. F., Y. Watanabe, P. E. Phillips, and H. R. Williams, Jr. 1965. The partial purifilcation and some properties of the mouse protection antigen of the Inaba subtype of Vibrio cholerae El Tor), p. 259-263. In Proceedings of the cholera research symposium. Government Printing Office, Washington, D.C. Verwey, W. F. 1972. An hypothesis of antibacterial immunity in cholera, p. 186-188. In Proceedings of the 8th joint conference, U.S.-Japan cooperative medical science program, cholera panel. Fuji Printed Co., Ltd., Tokyo. Williams, H. R., Jr., W. F. Verwey, G. D. Schrank, and E. K. Hurry. 1973. An in vitro antigen-antibody reaction in relation to an hypothesis of intestinal immunity to cholera, p. 161-173. In Proceedings of the 9th joint cholera research conference. U.S. Department of State, Washington, D.C.