EXPERIMENTAL COLI‘BACILLOSIS IN GNOTOBIOTIC PIGS .,. Thesis for the Degree of M. S. M!CHIGAN STATE UNWERSITY BRUCE ROBERT CHRiSTIE E967 LIB R A R Y Michigan State University IHESXS ABSTRACT EXPERIMENTAL COLIBACILLOSIS IN GNOTOBIOTIC PIGS by Bruce Robert Christie Research was conducted using a total of 62 gnotobiotic pigs in 3 experiments to determine the clinical effects, pathogenesis and lesions produced by 2 closely related sero- types of Escherichia coli--Ol38:K81:NM (Michigan origin) and 0138:K81 (Minnesota origin). It was determined that both serotypes readily colonized in the intestinal tract from experimental infection simula- ting neonatal contamination of the umbilical stump. Oral exposure was precluded. A bacteremia resulted within 1 1/2 hours after injec- tion, and by 24 hours the infection was generally established in the mucosa of the gastrointestinal tract. By 48 hours after injection the bacteremia had subsided so that only an occasional isolation from organs other than the gastroin- testinal tract was made. The clinical signs of profuse watery diarrhea, dehy- dration and elevated haircoat characteristic of colibacillosis in the neonatal pig were not apparent until the organism be- came well established in the gastrointestinal tract. The Bruce Robert Christie clinical signs of infection with E. ggli 0138:K81 (Minn.) were significantly milder than those resulting from infec- tion with E. ggli 0138:K81:NM (Mich.), but edema of the sub- cutis was a constant feature of infection with the former. Oral exposure of l germfree litter to E. ggli 0138: K81:NM (Mich.) produced both heavy colonization of the en— tire intestinal tract and diarrhea within 4 hours. Inter- mittent bacteremia was evident in this litter. No definite trends in body temperature or hematologic findings were observed, although in some animals there was a marked reduction in leukocyte counts. The necropsy findings, in general, agreed with those reported in the literature from field cases of colibacillosis in neonatal pigs. Histologically the lesions were predominantly in the gastrointestinal tract and ranged from an acute hemorrhago- necrotic enteritis to a histological picture, in persistently scouring pigs, microscopically indistinguishable from that seen in clinically normal germfree animals. The most common lesion observed in the epithelium of the villi of infected pigs was hydrOpic degeneration. Basic changes in apparent function of the villi were also noted. It is hypothesized that the changes observed result in malabsorption. EXPERIMENTAL COLIBACILLOSIS IN GNOTOBIOTIC PIGS BY Bruce Robert Christie A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Pathology 1967 ACKNOWLEDGMENTS I wish to express my gratitude to my major professor, Dr. Glenn L. Waxler, for his guidance, consideration, en— couragement, understanding and practical help throughout the course of these studies. His energy and skill in help- ing to secure and raise the experimental animals was invalu- able, and his ready assistance at any hour during the course of the experiments was very much appreciated. My sincere thanks to my academic committee: Dr. C. K. Whitehair, who assisted so much in getting me started along the way in this research program; Dr. D. A. Schmidt, for his friendly help and advice from the outset of the work; and Dr. C. C. Morrill, Chairman of the Department of Pathology, Michigan State University, for providing facilities and sup- plies for this research. I am extremely appreciative of the technical assistance given by Mr. Robert A. Brooks, former Senior Pathologist of this department, Mesdames Sunderlin, Miller and Whipple of the histopathology laboratory, and Messrs. Lynn Guenther and James Southern, animal caretakers at the Veterinary Research Farm, Hagadorn Road, East Lansing. ii All serotyping of the reclaimed organisms from these experiments was done through the courtesy of Dr. Paul Glantz of The Pennsylvania State University. The author is very appreciative of the kindness of Dr. Glantz in providing this invaluable assistance. iii 'fiv ”J:-(< " /7€5' — I757. iv TABLE OF CONTENTS INTRODUCTION . O O O O O O O O O I O O O LITEMTURE REVIEW 0 O O O O O O O . . . Colibacillosis in Human Medicine . . Colibacillosis in Veterinary Medicine Colibacillosis in Swine . . . . . . The Pathogenesis of Colibacillosis . Colibacillosis in "Specific Pathogen and Gnotobiotic Pigs . . . . . . MATERIALS AND METHODS O O O O O O O O 0 General Plan . . . . . . . . . . . . Animals 0 O O O O O O O O O O I O 0 Determination of Initial Sterility of Test Animals . . . . . . . . . Control Animals . . . . . . . . . . Infective Agents . . . . . . . . . . Injection Technique within the Isolator Hematology . . . . . . . . . . . . . Necropsy and Microbiologic Procedures Microbiologic Technique . . . . Histopathologic Technique . . . Reclamation of Test Serotype . . . . ll 14 18 18 19 20 22 22 24 26 27 28 30 32 Page RESULTS . . . . . . . . . . . . . . . . . . . . . . . 33 Experiment 1 . . . . . . . . . . . . . . . . . . 33 Experiment 2 . . . . . . . . . . . . . . . . . . 48 Experiment 3 . . . . . . . . . . . . . . . . . . 53 Recovery and Serotyping of the Organisms . . . . 59 Clinical and Pathologic Findings . . . . . . . . 60 Clinical Findings . . . . . . . . . . . . . 60 Hematology . . . . . . . . . . . . . . . . 65 Histopathologic Findings . . . . . . . . . 68 DISCUSSION . . . . . . . . . . . . . . . . . . . . . 105 Microbiologic Aspects of Experiment 1 . . . . . . 105 Microbiologic Aspects of Experiment 2 . . . . . . 107 Microbiologic Aspects Common to Experiments 1 and 2 . . . . . . . . . . . . . . 107 Microbiologic Aspects of Experiment 3 . . . . . . 108 Clinical and Pathologic Findings (Experiments 1, 2 and 3) . . . . . . . . . . . 110 SUMMARY . . . . . . . . . . . . . . . . . . . . . . . 121 REFERENCES . . . . . . . . . . . . . . . . . . . . . 123 VITA . . . . . . . . . . . . . . . . . . . . . . . . 128 vi LIST OF TABLES Table Page 1 Details of experimental animals, control animals, serotype of infective organisms, and number of infective organisms per inoculum . . . . . . . . . . . . . . . . . 21 2 Bacteriologic findings for Litter 1. . . . . 34 3 Bacteriologic findings for Litter 2. . . . . 39 4 Bacteriologic findings for Litter 3. . . . . 44 5 Bacteriologic findings for Litter 4. . . . . 49 6 Bacteriologic findings for Litter 5. . . . . 56 vii Figure 10 11 12 13 LIST OF FIGURES Coiled wire loop and spatula used for tissue sampling 0 O O O O O O O O O O O A typical culture plate . . . . . . . . . Pig J4669. Hemorrhagic contents of the jejunum and ileum. E. coli was injected 32 hours before necropsy . . . . . . . Pig J4668. E. coli was injected 40 hours before necropsy. Edema of the mesocolon and hyperemia of the umbilical entrance to the peritoneal cavity . . . . . . . External intestinal surface of an infected pig 0 O O O I O O O O O O O O O O O O O Pig J4662. Injection site 4 hours after injection of E. coli . . . . . . . . . Pig J4662. Same field as Figure 6. Peri- vascular neutrophils .. . . . . . . . . Pig J5722. The injection site. E, coli was injected 168 hours before necropsy. Pig J6205. Prescapular lymph node. E. coli was injected 24 hours before " necropsy O O O O O O O O I O O I O O I Pig J4673. Prescapular lymph node. E. coli was injected 101 hours before necropsy O O I O I O O O I I O O O O O Pig J5716. Myocardium . . . . . . . . . Pig J5712. Aorta. E. coli was injected 5 hours before necropsy . . . . . . . . Pig J5716. Esophagus. Parakeratosis of the eSOphageal epithelium . . . . . . . viii Page 29 29 37 37 54 69 69 71 71 73 75 75 78 Figure Page 14 Pig J6207. Esophagus. E. coli was injec- ted 72 hours before necrOpsy . . . . . . 78 15 Pig K63. Mucosa of the cardiac region of the stomach from a germfree control . . 79 16 Pig K63. Mucosa of the fundic region of the stomach from a germfree control . . 79 17 Pig J5718. Fundic region of the stomach E. coli was injected 40 hours before necropsy . . . . . . . . . . . . . . . . 80 18 Pig J4669. Pyloric region of the gastric mucosa. E. coli was injected 32 hours before necropsy . . . . . . . . . . . . 82 19 Pig J4669. Pyloric region of the stomach. E. coli was injected 32 hours before necropsy O O O O O O O O O O O O O O O O 82 20 Pig J6207. Fundic region of the stomach. E. coli was injected 72 hours before necropsy . . . . . . . . . . . . . . . . 83 21 Pig J5723. Terminal ileum. Germfree contrOl O O O O O O I O O O O O O O O O 85 22 Pig K57. Terminal jejunum. E. coli was introduced per cs 4 hours before necropsy O O O O O O O O O O O O O O O O 85 23 Pig J5721. Midjejunum. E. coli was in- jected 121 hours before necropsy . . . . 87 24 Pig J4672. Midjejunum. E. coli was in- jected 77 hours before necrOpsy . . . . 87 25 Pig J5721. Terminal ileum. E. coli was injected 121 hours before necropsy . . . 88 26 Pig J5721. Terminal ileum. From the same field as Figure 25 . . . . . . . . 88 27 Pig J6206. Jejunum. E. coli was injected 48 hours before necropsy . . . . . . . . 89 28 Pig J5721. Duodenum. E. coli was inject- ed 48 hours before necropsy. . . . . . . 89 ix Figure 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 Page Pig K60. Midjejunum. E. coli was intro- duced per os 40 hours before necropsy . . 90 Pig K62. Midjejunum. E. coli was intro- duced per cs 8 hours before necropsy . . 90 Pig K57. Terminal jejunum. E. coli was introduced per cs 4 hours before necropsy 91 Pig K54. Terminal ileum. E. coli was introduced per os 16 hours before necropsy . . . . . . . . . . . . . . . . 91 Pig J4672. Midjejunum. E. coli was in- jected 77 hours before necropsy . . . . . 93 Pig J4672. Midjejunum. From the same field as Figure 33 . . . . . . . . . . . 93 Pig K55. Terminal jejunum. E. coli was introduced per os 22 hours before necropsy O O O 0 O O O O O O O O O O O O 94 Pig K55. Midjejunum. E. coli was intro- duced per os 22 hours before necropsy . . 94 Pig J4669. Mesocolon. E. coli was in- jected 32 hours before necropsy . . . . . 96 Pig J6206. Colon. E. coli was injected 48 hours before necropsy . . . . . . . . 98 Pig J6206. Colon. From the same field as Figure 38 O O 0 O O O O O O O O O O O 98 Pig J4669. Duodenum. E. coli was injected 32 hours before necrOpsy . . . . . . . . 99 Pig J4669. Terminal jejunum. E. coli was injected 32 hours before necropsy . . . . 99 Pig J4669. Terminal jejunum. From the same field as Figure 41 . . . . . . . . . 100 Pig J4669. Terminal ileum. E. coli was injected 32 hours before necropsy . . . . 100 Pig J4669. Colon. E. coli was injected 32 hours before necropsy . . . . . . . . lOl Figure Page 45 Pig J6206. Liver. E. coli was injected 48 hours before necropsy . . . . . . . . 101 46 Pig J57l4. Liver. E. coli was injected 17 hours before necropsy . . . . . . . . 102 47 Pig K62. Liver and cholecyst. E. coli was introduced per cs 8 hours Before necropsy . C C C O C O . . O C O O C C C 102 48 Pig J6211. Urocyst epithelium from a germfree control . . . . . . . . . . . . 104 49 Pig J6791. Urocyst epithelium. E. coli was injected 72 hours before necropsy. . 104 xi INT RODUCTI ON The distribution of various bacterial species in the intestine of the normal pig is influenced by the age, diet, level of intestine and the bacterial species themselves. In recent years there has been a renewed interest in the association of Escherichia coli with neonatal diarrhea in human infants, calves and pigs. There have been periods in the study of this association when an appreciation of the commensal nature of the organism in the normal gut flora has led investigators away from serious consideration of the role of E. 22$; as a pathogen. The ability to differentiate strains of E. ggli_by serological typing, initiated by Kauffman (1943) and de- veloPed by others, has enabled relationships between strains, host animals and pathogenicity to be established. Pigs derived by hysterotomy or hysterectomy and main- tained free of all detectable microorganisms, or specifically contaminated with known organisms, are called "gnotobiotic" pigs. The recent development of techniques that make such animals available has been of value to pathology and micro— biology. There have been a number of attempts to character- ize the etiology and pathogenesis of colibacillosis in neo- natal pigs. Some of the most valuable work has been done with hysterectomy-derived, colostrum-deprived and artifi- cially reared pigs. There are, however, many concepts con- cerning pathogenesis of colibacillosis that have little basis in fact. It is likely that the basic problems of the elucidation of the syndrome of colibacillosis can most suc- cessfully be investigated in "germfree" animals. The objectives of this study were: 1. To observe the route by which E. 22$; 0138:K81:NM (Michigan origin) established an infection in gnotobiotic pigs following introduction of the organism into the subcutis of the umbilical stump without contamination of the external environment or oral contamination of the experimental animals. 2. To observe the route by which E. 22$; Ol38:K81 (Minnesota origin)--the same serotype referred to above, but from a different source and with a known different clinical expression--established itself in the gnotobiotic pigs, using the same experimental conditions as used with the above sero- type in order to provide a basis for comparison of these organisms. 3. To observe the route by which E. Egli Ol38:K81:NM established an infection in gnotobiotic pigs following oral inoculation of the organism. 4. To compare the clinical syndrome produced by oral inoculation of gnotobiotic pigs with E. Egll Ol38:K81:NM, and the syndrome produced by injection of the same organism into the subcutis of the umbilical stump. 5. To observe and record the gross and microscopic lesions resulting from infection of germfree pigs with these two serotypes of E. coli. LITERATURE REVIEW Escherichia coli is an organism of many races, of great variety serologically and continually varying in viru- lence. It was Theobald Escherich, in 1885, who first found and described "Bacterium coli commune" from the feces of breast-fed babies. His detailed description leaves no doubt that this was the organism we now know as E. ggEE_(Sojka, 1965). Escherich did, however, consider this organism.to be a harmless saprophyte, and his published opinions might have had considerable bearing on the fact that it has been only in recent years that a renewed interest has been taken in the potential of E. 32;; as an important pathogen of both man and animals. Reliable methods of classifying E. ggii_became avail- able in the 1940s with the discovery of the "O" agglutina- tion masking effect of the "L" antigen and subsequent pub- lications concerning the antigenic structure of E. 22;; (Kauffman, 1944; Kauffman, 1947; Kauffman and Dupont, 1950). Since then classification of the organism according to sero- type has been widely accepted, and the literature pertaining to serotyping of E. 29;; antigens has been reviewed and tabulated by Edwards and Ewing (1962). Some serotypes of E. coli are pathogenic in their own right. Most serotypes, 4 however, are of low virulence and cause disease because of either their overwhelming numbers or because the host has been weakened by other factors. Certain strains of E. 22$; produce powerful endotoxins capable of evoking tissue reaction when the organism invades tissues outside those of its normal habitat (Ribi EE_El°' 1964), but one assumes that the power to invade tissues is not necessarily associated with the pathogenicity of the endotoxin of the organism. Colibacillosis in Human Medicine It is of concern to note in important medical publi- cations in this field statements such as that of Robbins (1962) that strains of E. Egli are separable by serologic methods, but that these immunologic differences are "ap- parently not important in clinical medicine." In human medicine certain serotypes of this organism are important etiologic agents of focal pyogenic skin in- fections, urinary tract infections, peritonitis, acute ap— pendicitis, cholecystitis, cholangitis, and infectious biliary cirrhosis. In all of these sites E. 22;; evokes a nonspecific suppurative reaction with a degree of abscess formation that is often indistinguishable from reaction of tissues to invasion with Staphylococcus EEE’ (Robbins, 1962). In addition, there are specific strains of E. coli, identi— fiable by serotyping procedures, that produce severe forms epidemic diarrhea of infants. These infections are often associated with pneumonia and other bacteremic complica- tions. The importance of the immunologic differences between strains of these organisms is that strains that are impor- tant in clinical medicine can now be identified and separ- ated from those of less clinical significance. Infections with E. 22;; can occur at any age but are particularly severe in infancy and in advanced life. The age distribution is probably due to lack of immunity in the young or general debility and increased susceptibility of the aged (Jubb and Kennedy, 1963). Escherichgg coli sets up a symbiotic residence in the intestinal tract almost at birth, and it is thought that the organisms can enter the tissues of the host from this source via the lymphatic route or by contamination of the body surfaces or through breaks in these surfaces (Hopps, 1966). Colibacillosis in Veterinary Medicine In the veterinary field, Jubb and Kennedy (1963) sum- marized the prevailing concepts when they listed the three basic situations in which this organism behaves as a pathogen. Certain strains act as primary pathogens; others of lower virulence act as pathogens only when they are present in numbers sufficient to overwhelm the host tissue defenses; and other strains act as secondary or Opportunist invaders which enter and produce, or combine with other organisms to produce, disease in a host tissue weakened by a previous invader or stress situation. In the latter case it..is con- sidered that, once the strain of E. 32$; has become estab- lished, the predisposing factors may become unnecessary. As an example, young animals contract the disease and may die without the predisposing factors' operating. These authors (Jubb and Kennedy 1963) list the predisposing stress factors in young nonprimates: (1) inadequate housing with exposure to extremes of weather; (2) mistakes in dietary management (often overfeeding); (3) insufficiency of colos- trum; (4) congenital weakness; and (5) specific vitamin de- ficiencies. In a study of "calf scours" Smith and Crabb (1956) could not find evidence of a specific strain of E. 22;; associated with an individual outbreak of the disease. They found, in the healthy control group, 22 of the 23 phage types of E. 39;; that had been isolated from the diseased group. They also observed a change in the pre- dominant phage type during the course of the disease. It is not yet known what factors are associated with virulence in a particular strain, and there is no in vitro test for this prOperty available at this time. Gay (1965) has recorded that, in a particular species of host animal, there is variation and overlapping in the lethal dose be- tween serotypes considered to be pathogens and nonpathogenic strains from healthy animals. Studies on the pathogenicity of these strains, using laboratory animals or chick embryos, are of little value. In the experience of some workers (Saunders EE_El°' 1960) a specific serotype may or may not be hemolytic, the conditions of incubation (particularly the carbon dioxide potential) being a determinant in this phenomenon. In general there is no relationship between virulence and O and K antigens. With serotypes isolated from "gut edema" and coliform gastroenteritis of swine, however, Gay (1965) recorded a relationship between these two antigens and pathogenicity. The most recent review of the identification and classi- fication of the serotypes of the E. 22;; group of importance in veterinary medicine is well documented, brief and quite comprehensive (Gay, 1965). Colibacillosis in Swine Much confusion has resulted from the consideration by many workers that diarrhea of baby pigs is an entity rather than a clinical sign. The relationship between E. 92;; and porcine neonatal diarrhea has been accepted for a number of years. It was Jensen (1899) who first recorded that an infec- tion by E. coli could produce diarrhea in baby pigs. In 1934, there was described an outbreak of diarrhea in 3-day-old pigs that was considered to be a manifestation of E. 22$; infection (McBryde, 1934). In his account McBryde described a duodenitis and a jejunitis with no obvious le- sions in other organs. He cultured E. Egli from internal organs and mesenteric lymph nodes of these pigs. Glasser, Hupka and Wetzel (1950) described an acute outbreak of swine enteritis associated with E. 22$; infec- tion as being characterized by diarrhea, anorexia and, at necropsy, catarrhal gastritis and enteritis with redding and swelling of the intestinal mucous membrane. In their investigation of an acute diarrhea affecting pigs within 6 hours and up to a few days after farrowing, Gordon and Luke (1958) found that the pathogen involved was E. EQEE. The disease was characterized by high mortality in a litter that was apparently normal at the time of far- rowing. They observed that shortly after birth the pigs became dull and were disinclined to suckle. A profuse diar- rhea (with yellow to orange feces) resulted in dehydration, followed at the end of the first and second day by coma and death. Less acute cases were seen to have "scalded" hind- quarters and, in some pigs, necrosis and sloughing of the tail. Gastroenteritis was a consistent sign at necropsy. Almost pure cultures of E. ggli were grown from large in- testine and stomach samples, and occasional isolations of the same organism were made from liver and spleen. 10 Saunders EE_El' (1960), while investigating 58 out- breaks of disease in pigs in which E. ggli_was considered significant, came to recognize a clinically well defined syndrome affecting piglets within an age range of 48 hours to 2 weeks. The syndrome that these workers attributed to infection with E. 39;; was characterized by purity, abun- dance and consistency of the serotype; absence of these features from other cases examined routinely during the same period was noteworthy. The finding of Saunders 2E_El° (1960) that E. 22;; was not often recovered from viscera other than the intestines, even when there was a delay of several days between death and post-mortem examination, con- curs with the finding of Stevens (1961). In a recent review, Sojka (1965) listed 37 papers pub- lished since 1928 in which E. 22;; was considered to be involved in gastroenteritis and septicemia of young pigs. In a recent study of 100 pigs submitted to a diagnos- tic laboratory, Gossling and Rhoades (1966) obtained isolates of E. EQEE_0138:K81:NM from 6 pigs with "enteritis." These authors stated that "two serotypes, 08:K85:Hl9 and 0138: K81:NM,,were identified as possible specific etiologic agents of enteritis in baby pigs." In a comprehensive analysis of outbreaks of enteritis and diarrhea over a 5-year period, Stevens (1963) was able to discern 3 groups in which a characteristic syndrome was associated with a specific age group. These syndromes ll seemed to be associated with certain serotypes of E. EQEE. These syndromes and the age groups were: (1) piglet enter- itis, l to 4 days of age; (2) enteritis of unweaned piglets, 3 weeks of age; and (3) post-weaning enteritis of pigs, 10 to 12 weeks of age or shortly after weaning, manifested chiefly as a "catarrhal enteritis." Stevens included hemor- rhagic enteritis, edema disease and necrotic enteritis in his "post-weaning enteritis" group. The experimental reproduction of piglet enteritis with 2 serotypes frequently associated with field cases is re- corded by Saunders et a1. (1963). The disease resembled natural infections associated with these serotypes, diarrhea following the experimental exposure within 12 to 24 hours and some deaths occurring within 3 days. At necropsy the same serotypes were obtained in profuse pure culture from the intestines and, at times, from other viscera. Varying degrees of ileitis and colitis, commonly associated with slight edema of the colonic mesentery, were found in the dead pigs, and mild gastritis was seen in approximately 1/3 of these animals. The Pathogenesis of Colibacillosis In studies of experimental colibacillosis using a particular serotype, Saunders et al. (1960) found the or— ganism to be present in all the viscera of some piglets but only in the intestines of others in the same group. 12 This phenomenon was observed when 2 other serotypes were used, and these workers concluded that bacteremia was a stage in the development of the infection. Jubb and Kennedy (1963) listed the favored sites of localization of E. 2211 as the intestinal tract and de- scribed the infection as septicemic since it is possible to isolate the organism from all tissues. Localization of E. 22;; in the meninges, joints and lungs was described by these authors, who pointed out that, although these infec- tions accompany the enteric phase, their manifestations appear later than does diarrhea. Moreover, umbilical infec- tions might never show themselves as enteric infections but remain as tissue infections. The favored Sites of locali— zation of E. 99;; in the tissues are the meninges, joints, the 1ungs--especially if virus porcine pneumonia lesions are present--and probably the kidney, in which site it pro- duces the interstitial nephritis known as "white spotted kidneys." The renal infection may produce a descending pyelonephritis from which the organism is recoverable. It is evident from bacteriological examination of cases of colisepticemia that systemic invasion can occur from the umbilical stump, nasal or pharyngeal mucosae, or from the intestinal tract (Fey and Margadant, 1961; Fey, 1962; Jubb and Kennedy, 1963). Fey g£_gE. (1962) produced an E. 22;; septicemia by oral or intranasal infection of 3 colostrum-deprived neonatal 13 calves despite ligation and section of the esophagus. A fourth calf, with esophagus left intact and exposed intranasally to the same serotype, had an E. gel; bacteremia 20 hours after challenge, but the organism was not obtained from the contents of the small inteStine, collected by laparotomy, nor from bile. These workers concluded that, in calves, in- fection of the intestine was by the circulatory system rather than by the digestive system. A number of workers have re- ported that strains of E. ESL; from cases of infant epidemic diarrhea have the ability to cause diarrhea in adults as well as in infants when the organism is fed in sufficient numbers (Furguson, 1956; Neter and Shumway, 1950). The so-called "three-week enteritis" of Stevens' classi- fication (Stevens, 1963) is considered to be due to an in- crease in the number of E. 22$; organisms already existing in the alimentary tract and not to the introduction of another serotype. Of the third of Stevens'clinical groups (Stevens, 1963), the only specific manifestation that need be mentioned here is of hemorrhagic enteritis. This acute disease is character- ized by sudden death and is often associated with dietetic or environmental change. At necropsy the intestinal tract is partly filled with ingesta and blood, and one finds gross and microscopic lesions of severe hemorrhagic enteritis (Jennings, 1959; Philip and Shone, 1960; Thomlinson, 1963; Kinnaird, 1964; Bennett, 1964). Buxton and Thomlinson (1961) l4 and Thomlinson and Buxton (1962) prOposed that the hemore rhagic enteritis syndrome was an anaphylactic reaction. Stevens (1963) supported this View. Colibacillosis in "Specific Pathogen Free" and Gnotobiotic Pigs Experimental reproduction of colibacillosis in "pathogen-free," "germfree" or gnotobiotic pigs is generally not well documented. Saunders, Stevens, Spence and Betts (1963) used hysterectomy-derived, colostrum-deprived and artifiCially reared pigs in a series of experiments. In one experiment, using a known pathogenic serotype of E. 22;; (E68 type I) and oral doses ranging from 1000 to 100 x 106 organisms, they were able to produce diarrhea within 15 to 36 hours and recovery or, in some instances, death within 30 to 33 hours. In those pigs which died, the test organism was isolated from some or all of the tissues--intestines, liver, brain and lung. Those pigs which recovered were killed 6 days after infection, and the organism was cultured at that time from some or all of the sites--intestines, brain, heart, blood and liver. In a second experiment, using a "non-pathogenic" sero- type (P50) in 3 pigs, some diarrhea was observed within 20 hours of oral inoculation with 1 x 106, 100 x 106 and 100 x 106 organisms, respectively. These pigs recovered quickly and were soon seen to be healthy with soft to normal feces. 15 Two of these pigs were subsequently inoculated orally with serotype E68, type 1, and died within 2 and 5 days, respec- tively. The test organism was recovered in profuse pure culture from the intestinal tract and in sparse but pure culture from the brain, heart blood and liver. In the final experiment of the series Saunders §E_El° (1963) found that the pathogenic serotype used in the pre- vious experiments in neonatal pigs was equally pathogenic when used in 6-day-old pigs. They considered that there was no correlation between the severity of effects and the number of organisms administered and that the use of "patho- gen-free" pigs, which were shown to be free of cytopatho- genic viruses, indicated that their pathogenic serotype of E. 22;; was a primary pathogen. Using E. gel; 08:K?:H21 in gnotobiotic pigs infected at 4 to 6 days of age, Kohler and Bohl (1966) detected a bacteremia as a constant finding for 3 days after infection. They attributed this bacteremia to the fact that the experi- mental animals were colostrum-deprived, since bacteremia was not detected in the clinically ill, conventionally reared pigs from which the strain of E. ggli was originally isolated. These workers did not observe enteritis in gross examination of pigs which died in the acute phase of the disease. In a comprehensive study of the pathogenicity of E. coli, serotype 083, Britt and Waxler (1964) used 7 litters 16 of gnotobiotic pigs and recorded characteristic lesions in 78% of the 64 piglets exposed at ages ranging from 2 to 12 days. The outstanding lesion reported by these workers was a serofibrinous or fibrinopurulent polyserositis, including polyarthritis. Kenworthy and Allen (1966), in studies on germfree and gnotobiotic pigs, demonstrated a variation in intestinal villus and crypt structure associated with the degree of bacterial contamination of the intestional tract. The villi of pig intestine monocontaminated with E. 39;; Ol4l:K85a, c,(B):H4 were generally similar to those in the germfree control animal. Pigs contaminated with E. Egli serotypes Ol4l:K85a,c,(B):H4 and O8:H had intestinal villi with con- siderable variation in shape, including some branching forms. Slight edema in the lamina propria of the intes- tines of these duocontaminates was also observed. In pigs maintained in a "normal" environment, leaf-shaped villi were observed, with evidence of stunting, clubbing and fusion of villi. Those villi most severely affected were at the crests of the plicae. Cellular infiltration into the lamina propria was heavy, and reticular tissue appeared coarser and more abundant than in the monocontaminated pigs. These workers concluded that a biochemical interaction be- tween the intestinal flora and the diet might result in the appearance of an altered metabolic pathway which, in turn, depended on the symbiosis or antagonism of the ,17 bacterial species present and the various substrates present in the intestinal tract. The morphology of the mucosa of the small intestine is dependent, in some measure, on these factors (Kenworthy and Allen, 1966). MATERIALS AND METHODS General Plan In a study that involves following the progress of a specific organism through a sequence of body tissues or systems and the collection of tissues from the test animals so that the tissue changes can be correlated with the move- ment of the organism, it is necessary to make some assump- tions regarding the behavior of organisms in host tissues. The validity of these assumptions can be increased by stan- dardizing genetic and environmental factors in the experiment. It was expected that litter mates held under the same environmental conditions and submitted to identical treat- ments, provided they were of similar physiological activity, would generally respond or react in a similar way. It was not possible to carry out the extensive site sampling called for in this experimental method without killing the individual pig; hence, it was assumed that observations made on each in- dividual animal at a particular time interval following ex- posure to pathogens would constitute a general picture of the progress of infection through any particular pig of that litter when held under the same circumstances and when the disease was permitted to run its course. The gnotobiotic pigs were obtained according to the modified hysterotomy l8 19 method described by Waxler EE_El° (1966). This technique has been further modified by Dr. A. L. Britt* and Dr. G. L. Waxler,* in that a standard surgical skin preparation of detergent, alcohol, then chloroform, was used prior to the sealing of the skin of the sow to the plastic sheet that formed the floor of the fiberglass ring set in the base of the surgical isolator (Waxler, Schmidt and Whitehair, 1966). Other modifications have included the use of Allis tissue forceps to clamp the edge of the incised plastic sheet to the apposite edge of the incised skin of the sow at approxi- mately 2-cm. intervals. This modification has helped to maintain the adhesion between these 2 surfaces and to pre- vent movement of fluid from the surgical field out between the skin of the sow and the wall of the isolator. The use of umbilical cord clamps** to occlude the umbilical cord vessels prior to separation of the pig from the placenta has decreased the operation time and consider- ably facilitated manipulations within the surgical isolator. Animals An experimental unit consisted of l or more litters of gnotobiotic pigs, using some as control animals and some as test animals in an accepted experimental design. Between- litter variation was minimized so far as possible. The pigs *Department of Pathology, Michigan State University. **Disposab1e Cord-Clamp, Hollister "Double Grip," Hol- lister, Inc., Chicago, Ill. 20 used in this series of experiments were from related sows and were the progeny of 1 boar; they were fed a standard volume of the same diet;* they were maintained in sterile plastic-film isolators at an environmental temperature of 85 to 90 F. Details of litter size, sex distribution, birth weights, bacterial serotype, numbers of organisms administered, route of exposure, and age of pigs at expos- ure are shown in Table 1. Determination of Initial Sterility of Test Animals Immediately prior to the introduction of the infec- tive material into the sterile isolators, composite fecal, oral, nasal, and fecal tray samples were taken with sterile swabs and inoculated into 3 tubes of thioglycollate medium.** These tubes were incubated at 37 C, 50 C, and at room tem- perature. (In litters 3 and 5 the tubes were incubated only at 37 C.) These tubes were examined daily for 7 days. If there was no growth after 7 days, the isolators were con- sidered to have been free of bacteria. Since these were very short-term experiments, the sterility of the isolators was determined immediately prior to the exposure of the pigs to the test organism, and generally this was done within 48 hours of the delivery of the pigs into the isolators. *SPF Lac, Borden Co., New York, N.Y. **Bacto Fluid Thioglycollate Medium (Dehydrated), Difco Laboratories, Detroit 1, Michigan, U.S.A. 21 «N mo 2 A.£oflzvzz m m m mm m we no mom om.e "meummao E h o.H m moa ona 2 sea ema m macaw l.acflzv m R mm we m m HMOAHHQED oo.mmH ammummao 2 m m.o e em we 2 osspm A.n0Hsczz m m m mm 2 om HMUHHHQED om.o "ammummao S m H.H m we HOH z ossum A.noflzvzz m m mm mm m mm HMUHHHQED oo.mo "meummao z m o.H N mmH mma z madam A.30flzvzz m n poouoomu em on m we HMOflHflQED Ne.o “ammuwmao E o won a .o.uncr.H ),V S quv and N S o_N M V 1 IuJuo H6 3 x6 xo m a 3m in ,& 5.4 onu Tie x .d.e sun 1 . e 3 T.1.31 .. O 01. q 0 LG: I 1. MTans (9 so sa 8 1 as )e a e s I 1 n J n 1 A s.l HR” 1 sloqwe.& 1 1 o d .4 are 0 4+1 N e.a a 3 X a e . N XudTo e T Vu .18 m I;+I.Omw o ‘16 I unv I. I745 u 1 q 0 T. I q TLC": O The 9 ma: 3 a e.L d .14 9.9 H 1 P S (. le A s moan Houucou mamflcmmuo .ESHsoocfi Mom mamflcmmuo mo HmQEDG pom .mEmH Icmmuo m>fluommcfl mo mmmuonmm .mHmEHcm Houucoo .mamfiflcm HMUGOEHHmmxm mo mHHmqu .H manms 22 Control Animals Although it is essential that a sample of pigs from each litter should be kept as control animals in order to check the sterility of the intrauterine environment during gestation and the sterility of the hysterotomy and delivery procedures, only 2 or 3 pigs from each litter were retained for this purpose. It was planned that control animals from each litter be killed on a time schedule so that, together, they would adequately cover the actual-age scale of the test animals in the experiments constituting the whole of this series. Since between-experiment variation was negligible in terms of diet and environment, and genetic and congenital vari- ation was minimal (see birth weights recorded in Table 1), it was considered that control animals from any 1 litter in the series would provide adequate tissue controls for test animals of equal age from other litters in the series. Infective Agents The organism used in litter 2 was Escherichia coli 0138:K81 (Minn.) and was obtained from Dr. D. K. Sorenson of the University of Minnesota. This organism was reported to be one of the strains associated with edema disease and gastroenteritis and which had been shown to cause an enteric mucoserous exudate. This organism did not produce marked clinical signs when inoculated into gnotobiotic pigs. In 23 all other litters in this series of experiments the same serotype, from a different source, was used. This organism was E. 22;; 0138:K81:NM (Mich.) and was isolated in August, 1959, by Dr. G. L. Waxler from experimental pigs at Michigan State University. These organisms had been cultured, lyOphi— lized in glass ampules, and stored at -20 C until required. Twenty-four hours before the pigs were due to be in- fected, an ampule of the culture was broken open and its contents inoculated into liquid brain-heart infusion medium* and streaked onto a bovine blood agar plate and a MacConkey Agar plate.** These cultures were incubated at 37 C. After 24 hours' incubation, immediately before the experimental animals were due to be inoculated and when it was apparent that a pure culture was present, the tube culture was di- luted 1 in 1000 in sterile saline (0.85% NaCl). This saline suspension of the organisms was the material used in the isolators to inoculate the test animals. Serial tenfold dilutions of this inoculum.were made in sterile saline and 0.1 ml. aliquots of these dilutions were pipetted into standard Petri dishes, to which was added melted Eugonagar.+ Immediately, the contents were gently *Bacto Brain Heart Infusion (B 37); Difco Laboratories, Detroit 1, Michigan, U.S.A. **Bacto MacKonkey Agar (B 75); Difco Laboratories, Detroit 1, Michigan, U.S.A. TEugonagar Vera, Baltimore Biological Laboratory, Bali- more, Maryland, U.S.A. 24 mixed, allowed to cool and solidify, and were then incubated overnight. The viable count of the inoculum was calculated from the colony count on plates growing clearly separable colonies after 24 hours' incubation at 37 C. The inoculum was loaded into a sterile disposable tuberculin syringe* using strict aseptic technique. Once the syringe was loaded, the loading needle was removed, the tip of the syringe was flamed almost to the point of melting the plastic tip, and the tip of the syringe was capped with a covered sterile disposable 23 G needle.** The capped sy- ringe was then thoroughly sprayed with 2% peracetic acid (with approximately 0.1% wetting agent added).+ The syringe was then introduced into the port of the sterile isolator and the port resprayed and closed. After 30 minutes the sy- ringe was introduced into the isolator through the internal cap, and the inoculum was then ready for use. A standard dose volume of 0.251mL per pig was used throughout the ex- periments. Injection Technique within the Isolator An assistant restrained the pig in lateral recumbency a few inches above and separated from a sterile towel. i on the floor of the isolator. Meanwhile the operator, having *Tomac Disposable Tuberculin Syringe, American Hospital Supply Co., Evanston, Ill., U.S.A. **Discardit Sterile Disposable Yale 23G l, Becton-Dick- inson, N.Yq.U.S.A. _ +Nacconal, N.R.S.F., National Aniline Division, Allied Chemical Corporation, New York, N.Y., U.S.A. 25 donned sterile plastic gloves over the normal isolator gloves, removed the outer cap of the syringe, exposing the needle and thereby theoretically exposing the interior of the isolator to the organisms within the syringe. Holding the umbilical clamp in the left hand, it was possible for the Operator to insert the tip of the needle into the um- bilical stump distal to the clamp. The needle was then passed through the umbilical cord tissue held between the jaws of the clamp and then, with care, it was possible to place the tip of the needle in the subcutis of the umbili- cal stump immediately proximal to the clamp. The injection of 0.25 ml. of a suspension of the organisms was made at this stage. As the needle was withdrawn from the stump, the clamp pressure prevented the flow of inoculum back along the track of the needle. This, combined with the fine gauge needle that was used, meant that an absolute minimum of the sus- pension of the organism remained at the site of withdrawal of the needle. This site was immediately swabbed with COpious amounts of a 2% solution of peracetic acid. At no stage of the operation did the assistant or the pig touch the operator's hand which held the syringe. The only con- tact with the other hand of the Operator was with the end of the umbilical clamp, and this in turn did not contact the needle, the inoculum or the hand working the syringe. After all pigs in the isolator had been inoculated in similar 26 fashion, materials used in the injection technique, includ— ing the outer plastic gloves, were immersed in a flask of 2% peracetic acid solution, and the flask was immediately removed from the isolator. Although fraught with a number of hazards from the point of aseptic technique, especially since the operation was performed in an otherwise sterile environment, this technique was apparently successful in that, at no time during these experiments, was there detected contamination of the oral, skin, or isolator environment until the in4 jected organisms had become established in the intestinal tract and these were being excreted in the feces. In litter 5, in which the oral route was used, the syringes of inoculum were prepared as described (see above) and sprayed into the isolator. After each pig had been fed 4 fluid ounces of the standard milk formula diet, the 0.25 ml. dose was squirted onto the back of the tongue. Hematology Five-milliliter amounts of blood were collected from the anterior vena cava according to the method of Carle and Dewhirst (1942). An initial blood sample was taken within 24 hours of each pig having been delivered to the rearing isolator. A terminal sample was taken from each animal im- mediately prior to death. Total and differential leukocyte counts were carried out as described by Benjamin (1964). 27 Hemoglobin estimation was by the cyanmethemoglobin method and packed-cell volume by the micro-hematocrit method (Ben- jamin, 1964). Necropsy and Microbiologic Procedures Although Stevens (1961) has recorded that coliform bacteria do not usually invade the general viscera within 24 hours of death, even at the environmental temperature of the isolators that were used in these experiments, Dunne (1964) is emphatic in his opinion that E. 22$; is capable of invading the blood stream from the intestinal tract within an hour of death and states that ". . . isolation of the organisms from tissues under such conditions would be of little immediate value in establishing the presence of a septicemia." In order to minimize post-mortem migration and/or multiplication of test organisms within tissues, necropsies were completed within 90 minutes of euthanasia. Immediately after the pigs were removed from the isolaters, they were killed by concussion and exsanguination (by severing the axillary vessels). At necropsy the pig was placed in lateral recumbency on its right side and a midline skin incision was made from the symphysis pubis to the body of the mandible. The skin fold, including the foreleg, was reflected, the medial mass of the hind leg muscles was separated, and the hind leg was 28 laid back. Both the prescapular and prefemoral lymph nodes were exposed in this fashion. Microbiologic technique. With a hot spatula it was possible to dry and singe the surface of the lymph nodes and with a flamed, tightly wound, sterile wire loop* (Figure l) to pene- trate the singed surface, sample the contents of the node, and streak the loop onto the marked portion of a MacConkey Agar plate (Figure 2). The thoracic cavity was opened with heated scissors and a sample of pleural fluid collected with the sterile wire loop and streaked onto culture media. The pleural serosa of the parietal pericardium was in turn singed with the spatula, the pericardium was broken with a touch of the hot 100p of wire, and a sample of pericardial fluid was collected and streaked onto the medium. In similar fashion, using the small heated spatula, the myocardial surface was dried and singed, and the sterile wire loop was forced through the myocardium into the right ventricle. A loop of heart blood was withdrawn and streaked onto the medium. Using the hot spatula it was possible to first dry, then sterilize, the surface of any of the organs or tissues prior to using the sterile wire loop to collect material for culture. In sampling the humeroradial joint and the *Nichrome wire, 25 gauge; Scientific Products, Evans- ton, Illinois, U.S.A. 29 i —: Figure l. A coiled nickel-chromium alloy wire was used to penetrate and sample the selected organ and tissue sites. The spatula was used to sear tissue sur- faces before sampling. Figure 2. MacConkey Agar plate. A typical culture plate on which E. coli colonies are growing. 3O midbrain and medulla of the central nervous system a hot scalpel was also needed to clear away superficial tissues prior to sampling. The MacConkey Agar plates were streaked centripetally, and an attempt was made to use a standard stroke while making the streak and to use a minimal amount of inoculum. This procedure resulted in adequate separation of the streaks from different organs; and, in lightly contaminated tissue, separation into relatively small groups or organisms, or even individualization of organisms was possibly achieved (Figure 2). The MacConkey Agar plates were seeded as follows: the first plate was used for prefemoral lymph node, pre- scapular lymph node, pleura, pericardium, heart blood and lung samples; the second plate was inoculated with material from the peritoneum, liver, spleen, kidney, urocyst, iliac lymph node, humeroradial joint fluid and brain stem; the third plate was used for gastric mucosa, jejunal mucosa, mucosa of the ileocecal valve, rectal mucosa, the injection site (or the oral cavity in the experiment in which the in- fecting dose was given per os) and the subcutaneous tissue 2 cm. lateral to the inoculation site. Histopathologic technique. Tissues taken for histologic examination were: prescapular lymph node, transverse sec- tion of the thyroid gland, thymus, transverse section of the trachea and eSOphagus at the mid-cervical level, cross 31 section of the right cardiac lobe of the lung, transverse section of the myocardium (including left and right ventri- cles with their serosal surfaces and portion of the interh ventricular septum), transverse section of the aorta in the mid-thoracic region, the skin and subcutis at the injection site, the skin and subcutis 2 cm. lateral to the injection site, cross section of the right adrenal gland, pancreas, midsection of the spleen, posterior section of the right central lobe of the liver, transverse section of the chole- cyst in situ, transverse section of the right kidney at the level of the pelvis, section from the anterior dorsal wall of the urocyst, sections of cardiac, fundic and pyloric re- grions of the stomach wall, duodenum, transverse sections of the jejunum at the 50 cm., lm., and 2m. levels, trans- verse section of the ileum immediately cranial to the ileo- cecal valve, and seetions of the cecum, spiral colon (in- cluding relationship with other coils and the mesocolon) and a transverse section of the rectum at the level of the pelvic brim. Tissue sections or slices approximately 0.5 cm. thick were identified and fixed in 10% formalin and processed for examination by procedures described in the Manual of HistOr logic and Special Staining Techniques of the Armed Forces Institute of Pathology, Washington, D.C. (1957). 32 Reclamation of the Test Serotype After the MacConkey Agar plate cultures had been ex- amined and read, a sterile loop was used to sample a typical colony from each of the positive streaks, and this loop was used in turn for sampling the positive isolations from each of the pigs which had inhabited a particular isolator during the experiment. This loop of organisms selected from different tissue. isolations and different animals, but all from the one iso- lator, was inoculated into 2 tubes of boiled thioglycollate medium and incubated for 16 to 24 hours, one tube at 37 C and the other at 55 C. A composite of the growth from either or both of these tubes was inoculated onto a nutrient agar slant and onto 2 bovine blood agar plates, one to be incubated aerobically and the other anaerobically for 18 to 24 hours at 37 C. The nutrient agar slant was incubated for 24 hours at 37 C, sealed with a foam plug, and stored at room temperature until all samples were ready for dis: patch to the serotyping service. In this way it was possible not only to check the validity of the monocontamination but also to insure that the serotype isolated from the test animals was, in fact, the serotype that was used in the initial infecting expo- sure and that a mutant of altered serotype had not appeared during the course of the experiment. RESULTS Experiment 1 The Experimental Plan In this experiment 3 litters of pigs were infected by injection of E. coli 0138:K81:NM ”fixfln) into the subcutis of the umbilical stump but without oral or envirnomental contamination. In this way it was hoped to simulate infec- tion of the umbilical stump of neonatal pigs and to study the course of infection subsequent to this in colostrum- deprived animals. Litter 1. These 13 pigs were 46 hours old when 0.42 x 106 E. coli 0138:K81:NM (Mich.) organisms were injected into the subcutis of the umbilical stump of each of 11 of them. The animal necropsy numbers, sex distribution, period of elapsed time between inoculation and necropsy, and sites of positive isolation from tissues are tabulated (Table 2). Four hours after injection it was recorded that there was heavy contamination of the injection site and moderate contamination of the subcutis 2 cm. lateral to the injection site. One organism each was cultured from the heart blood and the lung, and 8 organisms were detected in the loop of fluid taken from the spleen. 33 34 .nu3oum o: I .mchoHoo OOH swap mHoE u ++ .mchoHoo HOH coco mmmH pom mchoHoo om cmnu oHoE + .mchoHoo Hm Conn mmoH n “@985: one "mBOHHom mm oommmumxm mum mucsoo >c0Hoo HMHHopomm« I I ++ ++ ++ ++ ++ ++ ++ m I I I enuomm I I ++ ++ ++ ++ ++ ++ om m I H I m>Hm> HmomoomHH I I ++ ++ ++ + + ++ ++ om I I I .Eucsmmnon I I ++ + + ++ ++ ++ om OH I I I comsoum I I I I I I I I ++ I I I + mgHm coHuowncH 0:» Op HmnmuMH .80 N mesoQom I I I + ++ I I om ++ om ++ ++ ++ muHm coHuochH I I I I I I I I om + I m I ucHon HMHpmuoumEDm I I I I I I I I ++ + I m I :Hmum I I I I I I I on ++ m I I I moo: smasH unflHH I I I I I I I I + om I I I unhooub I I I H I I I I ++ + I H I mocon I I I I I I I m ++ + I mH m comHmm I I I w I I I I ++ + I m I um>Hq I I I I I I I I om om I I H EsmcouHuom I I I I I I I I ++ ++ I I I mood I I I H I I I I ++ ++ I + H oopo undo: I I I I I I I I I + I I I EdHonoHumm I I I I I I I I om om I I I musmHm I I m I I N I I ++ + I H I moo: nmemH Hmuoammonm I I I I I I I I ++ ++ I I I moon £QE>H HmHsmmommum «mmuHm quBOHHom map EOHM mSchmmHo Mo coHumHOmH mm. «N am. mmH HOH an we ow mm am HH m a coflnomflafl “when musom 2 m E m m 2 S m h 2 h m 2 wa P. P. P. f P. P. P. P. n... P. P. n... n... .V .V .V .7 .V .V .7 .V .V .V .7 .7 .7 9 9 9 9 9 9 9 9 9 9 9 9 9 I. I. 8 8 /. I. 9 9 9 9 9 9 9 HMQEZ mflm TI. 0 T. 0 PC 7... I. 8 6 9 .7 Cr. 7. . mHouucou .H Hmuqu CH mmflm EOHH meHUCHM OflwOHOHkuomm .N mHQMB 35 No gross lesions were observed in this pig. After 8 hours, heavy growth was obtained from the in- jection site, and light to moderate growth was obtained from the heart blood. One isolation was made from the prefemoral lymph node, and a few organisms were cultured from liver, spleen, kidney, brain and humeroradial joint cavity. One organism was obtained from the mucosa of the ileocecal valve. Gross abnormalities observed in this pig were ventral abdominal edema and micrognathia. In the pig subjected to necropsy 11 hours after injec- tion, organisms were isolated only from the injection site. Culture of this site indicated that a heavy contamination was present. The organisms were not present in the subcutis 2 cm. lateral to the injection site. There were no gross lesions in this animal. Twenty-four hours after injection very few organisms remained in the injection site, no organisms were detected in the subcutis 2 cm. lateral to that site, but there was a heavy systemic contamination of the pig. At this stage very light contamination of the mucosa of the intestinal tract at the test sites was recorded (Table 2). The.gross lesions in this animal were hemorrhagic: edema of the injec- tion site and an area of erythema, 1 cm. in diameter, of the parietal peritoneum at the point of entry of the umbili- cal vessels. 36 At 32 hours after injection there was moderate to heavy systemic and gastrointestinal tract contamination, positive isolations being made from all 20 sites sampled. Grossly, the liver was soft and very dark. Throughout its length the ileum was distended with thick, intensely red fluid consisting mostly of blood lost from the raw, denuded surface of the ileum (Figure 3). The gross appearance was that of hemorrhagic enteritis. After 40 hours a very few organisms were isolated from the injection site, iliac lymph node and spleen. All other systemic sites were nega- tive. Very heavy contamination was recorded from the 4 sites of the gastrointestinal tract. Clinically, this animal was drinking very well and had a profuse diarrhea. The feces were watery, yellow, and flecked with fine white curds. Gross lesions included edema of the mesocolon, slight hemorrhagic edema of the injection site, and petechiae of the visceral peritoneum at the umbilical entrance (Figure 4). The remaining pigs in this litter were killed at 48, 77, 101, 123, and 151 hours, respectively, post injection. In each case the gastrointestinal tract remained heavily contaminated, which suggests that this organ had become colonized by the test organism. Very few organisms were found in other sites (Table 2). Gross lesions consistently found in these animals were enlargement and edema of the superficial lymph nodes, slight to moderate hyperemia of the injection site, and 37 Figure 3. Pig J4669. Hemorrhagic con- tents of the jejunum and ileum. E. coli was injected 32 hours before necropsy. 5') ‘ ff 7" . Figure 4. Pig J4668. E. coli was in- jected 40 hours before necropsy. Edema of the mesocolon (supported by the spatula) petechiation and marked hyperemia of the um- bilical entrance to the peritoneal cavity (towel clamp). 38 slight to moderate localized edema of the ventral abdominal wall in the umbilical region. Ample food was present in the anterior gastrointestinal tract, and it was apparent from this and the feeding records that the appetites of the animals had not been impaired. Some hyperemia of the sero- sal surface of the stomach and the jejunum was recorded, and in some instances the jejunum appeared thicker-walled than normal. The contents of the cecum and colon were con- sistently watery, and moderate to marked edema of the meso- colon was observed in these pigs. The 2 control pigs from this litter were free of de- tectable microorganisms at necropsy, and all attempts to culture bacteria from the standard test sites on MacConkey Agar plates failed. These animals were killed at 70 and 198 hours of age, respectively, and were used for clinical and histological controls for this and.other litters in this experimental series. Litter 2. This litter was 29 hours old when 63 x 106 E. coli 0138:K81:NM (Mich.) organisms Were injected into the subcutis of the umbilical stump of each of IQ of the 12 pigs in the litter. The animal necropsy numbers, sex distribution, period of elapsed time between inoculation and necropsy, and sites of positive isolation from tissues are tabulated (Table 3). At 1 l/2 hours after injection only 1 organism was isolated from the injection site. The subcutis 2 cm. lateral 39 .cu3oum o: u I “mchoHoo 00H cusp whoa u ++ .mchoHoo HOH cmcu mmmH new mchoHoo om swap whoa n + .mchoHoo Hm cusp mmmH u Hogans one "m3oHH0m mm pmmmmumxm mum mucooo hooHoo HMHHmuomm« I I ++ ++ ++ I ++ ++ I I I I ,Esuomm I .I ++ ++ ++ ++ ++ ++ ++ I I H m>Hm> HmomoomHH I I ++ ++ ++ ++ ++ m H NH I I .85csnmflon I I ++ ++ ++ ++ + m we H I I someovm I I I I N I ++ I I I I I muHm coHuomncH may on HMHmHMH .80 N mHusonsm I I I ++ ++ v N I ++ I ++ H muHm coHuomncH I I I I I I I I I I I I ucHon HMHomuoumssm I I I I I I I I I I I I :Hmum I I I I OH I I I I I I I mp0: £QE>H OMHHH I I I I I I I H I I I I unmoouo I I I I I I I I I I I I mocon + I I I mm cmmHmm I I H I I I I N I I I mH um>HH I I I I I EsocouHumm I I I I mcsH H I I I H wooHn unmmm I I I I I I I I I I I I EprumoHumm H I I. I I m I I I I I I musmHm I I I I moo: cmexH HMHoEmmmum I I I I I I I I I I I I moo: nmfihH HmHsmmomonm ammuHm mCHBOHHom 03w Eoum mEchmmHo mo coHumHOmH mm Nb Nb mv vN mH NH m m.w w m.N m.H COH#OmflGH Hmpmm mHSOE .m .m 2 E .m E 2 2 .m .m S .m xmm P. P. P. f P. P. P. P. P. P. C P. 9 9 9 9 9 9 9 9 9 9 9 9 7.. z 7.. 7.. 7. 7.. z z z 7. z z m n w w w m w m. w w w m. .35 E mHouucoo .N HmpuHH CH mmHm EOHM mmchaHm UHmoHOHHmuomm .m OHQMB 40 to this site was sterile. Light contamination was recorded from the heart blood, liver and spleen cultures, and l or- ganism was isolated from the ileocecal valve mucosa. No gross abnormalities were observed in this pig. Culture of organisms was successful only from the in- jection site in the pig submitted to necropsy 2 1/2 hours after injection. Gross abnormalities observed were slight edema of the superficial lymph nodes and the mesocolon and marked hemorrhagic edema of the umbilical stump. At 4 hours after injection the organisms were not re- covered from the injection site, and isolations indicating very light contamination were obtained only from the stomach and midjejunal mucosae. During this necropsy a few fibrin strands were seen on the surface of the liver and spleen, and there was slight excess of faintly cloudy peritoneal fluid. After 6 1/2 hours, heavy growth was obtained from cul- ture of the injection site, but it was apparent that the organisms had not spread laterally in the subcutis, as cul— ture from the latter site was negative. Moderate growth from the gastric and midjejunal mucosae and very heavy growth from the mucosa of the ileocecal valve were recorded. The rectum was sterile at this stage. I At necropsy the iliac lymph nodes were enlarged and moderately edematous. The mesenteric lymph nodes were nor- mal, but there was marked hyperemia of the vessels located 41 between these lymph nodes and the mesenteric attachment of the jejunum. The aortic lymph nodes were enlarged, edema- tous and hyperemic. The injection site was markedly ede- matous, and the skin 2 cm. lateral to the umbilical stump was hyperemic and edematous. In the pig killed 8 hours after inoculation, light to moderate contamination was recorded from heart blood, pleura, liver, spleen and urocyst. There was light con- tamination of the gastric and midjejunal mucosae and very heavy contamination of the mucosae of the ileocecal valve and rectum. At gross examination of this pig the superfi- cial, iliac and aortic lymph nodes were slightly enlarged, and there were 2 subcapsular ecchymoses and areas of sub- capsular hyperemia of the liver. There was slight edema of the mesocolon, moderate edema of the injection site, and an associated ventral edema. The cecal contents were very c0pious and of watery consistency. The rectal contents were firm. Twelve hours after injection very few organisms re- mained in the injection site, although the subcutis lateral to the umbilical stump was heavily contaminated. All sites in the mucosa of the gastrointestinal tract were heavily colonized; all other test sites were sterile. At necropsy the injection site was moderately edematous and there was .an associated ventral abdominal edema. 42 After 16 hours, light contamination was recorded from culture of the injection site, and no growth was obtained from the subcutis lateral to that site. Very heavy growth was obtained from the mucosal samples of stomach, midjejunum and ileocecal valve. The rectal mucosa was sterile. All other sites were sterile. The only gross lesions in this animal were slight hyperemia of the injection site and slight edema of the prescapular lymph node. The remaining infected pigs in this litter were killed at 24, 48 and 72 hours after injection. At 24 hours after injection, moderate contamination of the lung, spleen and iliac lymph node was recorded. The organisms were found in profuse numbers in the injection site and had spread in the subcutis to a slight degree. The entire gastrointestinal tract was heavily colonized at this time. All other sites were sterile. Immediately prior to euthanasia a bright yellow, curdy and watery stool was observed. This pig had an optimum food intake, and there was no evidence of anorexia. At necropsy the stomach was filled with clotted milk and the intestinal tract was filled with watery yellow fluid in which flakes of milk curd were sus- pended. There was an ecchymosis at the injection site; there was also hyperemia of the peritoneum at the point of entry of the umbilical vessels. 43 In each of the remaining 2 pigs the gastrointestinal tract was heavily colonized and was the only site of any significant isolation of organisms. In each case the hair coat was erect, the animal was scouring profusely before death, and it was dehydrated. The appetite was normal, the stomach was filled with clotted milk, and the intesti- nal tract was uniformly flaccid and filled with the watery contents as described for the previous pig. The two control pigs from this litter were free from detectable microorganisms at necropsy, and all attempts at culture from the routine tissue sites on the MacConkey agar plates were negative. These animals were killed at 65 and 101 hours of age, respectively, and were used for clinical and histological controls for this and other litters in this experimental series. In the older pig, moderately enlarged and edematous prescapular lymph nodes were observed. The intestinal tracts of both control pigs were of uniform color throughout. Litter 3. This litter was 20 hours old when 0.8 x 106 E. coli 0138:K81:NM (Mich.) organisms were injected into the subcutis of the umbilical stump of each of 9 of the 11 pigs in the litter. The animal necropsy numbers, sex distribution, period of elapsed time between inoculation and necropsy, and sites Of positive isolation from tissues are tabulated (Table 4). «mchoHoo 00H cmnu «mmHGOHoo Hm cmnu mmmH mun .cuzoum o: 08 u ++ “mchoHoo HOH cmnu mmmH can mmHaoHoo om swap mHOE u + “m3oHHow mm Ummmmumxm mum mucsoo mcoHoo HMHMmpomm* Hogans was I I ++ I I ++ I I ++ I I ++ zHm> HmoooooHH asasnmmoHs somaoum muHm cowuomflaH on» on HMHmUMH .50 N mHusonsm wuHm GOHuooncH ucHon HMHpmuoumfism chum moo: smfihH UMHHH unknown ammon :mmHmm Hw>HH Esmcoanmm wand nooHn unmom EDHUHMUHHmm musmHm moo: smexH HMHOEmmmHm moo: amemH HMHsmmommnm «mmuHm mcHsoHHOM may Bonn mEchmmHo mo GOHuwHomH coHuomncH noumm musom xmm E6L9P Z6L9f T6L9D mHouucou 06L9f 68L9D €8L9f 38L9f V8L9C SBL9D HmnEsz mHm Hmuufln CH mmflm Scum mmGHUCHm OHmOHOHHmuomm .q magma 45 At 1 1/2 hours after injection only the injection site was positive for presence of the test organism. The only gross post-mortem lesion in this pig was a moderate hemor- rhagoserous exudate at the injection site. In the pig killed 2 1/2 hours after injection, the organisms could not be recovered from any of the routine sites, and it was considered that the antiseptic procedures used during the injection technique could have sterilized the inoculum in this instance. There were no gross abnormalities observed in this animal at necropsy. After 9 hours had elapsed between injection and eu- thanasia, moderate growth was obtained from the injection site and heavy growth was obtained from the mucosa of the ileocecal valve. All other sites were negative. At necropsy the superficial lymph nodes were slightly edematous, and there was apparent flaccidity of the ileum and large intes- tine. There was slight edema of the injection site and peri— umbilical edema. The pig subjected to necropsy at 24 hours was sterile in all culture sites. There was no edema of the injection Site. The midjejunum was distended and flatulent, and the rectum was partly filled with meconium. Thirty-two hours after injection there was heavy con- tamination of the injection site. One colony each was iso- lated from the urocyst and from the gastric mucosa. All Other sites were negative. 46 Gross lesions recorded for this pig were as follows. The terminal jejunum and ileum were moderately enlarged, with flaccid walls and very soft contents; there was slight edema of the mesocolon, and meconium was still present in the rectum. The injection site was slightly hemorrhagic and the periumbilical peritoneum was hyperemic in an area 1 cm. in radius. At the peritoneal attachment of the falci- form ligament there was a 3-mm.-diameter vesicle in the sub- serosa. This vesicle contained a slightly cloudy serous fluid. Forty-eight hours after injection, very light growth was obtained from culture of the injection site, and heavy growth was obtained from the 4 alimentary tract sites. All other sites were sterile. At necropsy there was moderate edema of the superficial lymph nodes and the mesenteric lymph nodes. The mediastinal lymph nodes were hyperemic, there was moderate edema of the mesocolon, the jejunum was uniformly flaccid, and the ileum and large intestine ap- peared to have normal intestinal tone. At the injection site there was a 5-mm.-diameter vesicle containing a sero- purulent fluid. This vesicle was surrounded by a l-cm.- wide hyperemic annulus. The subcutis lateral to this lesion was not noticeably edematous. Prior to euthanasia this pig had profuse watery stool and was the first animal in this litter to show this sign. 47 At 72 hours after injection, very light growth was ob- tained from the prescapular lymph node, and heavy growth was recorded from each of the alimentary tract sites. It was observed that the perineal epithelium of this female pig was inflamed and partly eroded and macerated as a result of the continual diarrhea. The culture of the mucosa of the urocyst was positive with a very heavy growth of organisms. Since all other systemic sites, including the kidney, were sterile in this pig it was presumed that this urocyst infection could have resulted from ascending contamination with organ- isms from the perineum. The intestinal tract was flaccid and filled with a watery fluid in which fine yellow curds were suspended. There was slight hemorrhagic edema of the injection site, and the superficial lymph nodes were slightly edematous. The 2 control animals from this litter were free from detectable microorganisms at necropsy, and all attempts at culture from the standard test sites on MacConkey plates were negative. These animals were killed at 23 and 44 hours of age, respectively, and were used for clinical and histo- logical controls for this and other litters in this experi- mental series. In the older pig the thymus was very small and was found only with difficulty. 48 Experiment 2 The Experimental Plan In this experiment a litter of pigs was infected by injection of E. coli Ol38:K81 (Minn.) into the subcutis of the umbilical stump but without oral or environmental con- tamination. This serotype of E. coli, although possessing the same 0 and K antigens as the organism used in Experi- ment 1, was from a different source and was milder in its clinical expression in infected pigs. It was proposed to observe the pathogenicity of this organism and to compare these observations with those recorded in Experiment 1 in order to determine microbiologic and pathologic differences between organisms of the same broad serotype but with dif- fering clinical manifestations in gnotobiotic pigs. Litter 4. This litter was 8 hours old when 135 x 106 E. 321i 0138:K8l (Minn.) organisms were injected into the subcutis of the umbilical stump of each of 12 of the 15 pigs in the litter. The animal numbers, sex distribution, period of elapsed time between inoculation and necropsy, and sites of positive isolation from tissues are tabulated (Table 5). At 4 hours after injection there was moderately heavy contamination of the injection site, and light to moderate LPOPMlations of the test organism were obtained from heart lilood, liver, spleen and kidney. Both the adrenals in this 49 «mmHCoHoo 00H “moHCoHoo Hm Cmnu Cmnu mHoE u mmmH ++ .Cuzoum 0C u I ammunCOHOU HOH CMSH wme UGM mmflGOHOU om Cwflu. whoa H + HoQECC one "m3oHH0m mm pmmmwnmxo mum muCCoo MCOHOO HMHCmuomm« I I I ++ I I I ++ I I I mH I I I HH Q‘ V In H ++ ++ ++ mH H ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ fi' ++ ++ ++ ++ N ++ ++ NH ++ ++ NH I—{ OH OH I _Esuomm I o>Hm> HmomoooHH I ascsHmHeHz I COMEOHm I muHm CoHuownCH may on HmanMH .80 N Cme + muHm COHuoonCH I UCHofl HMHUmuoumfism I CHmnm I mUOC CQENH UMHHH I ammoouo m emceHm NH :mmHmm ow Ho>HH I EsoCouHumm I mCCH m pooHQ pumom I ECHUHmoHumm I MHCmHm I m©0C CQENH Hmuoamwmum I oUOC CQEMH CMHCmmommum ImmuHm mCHonHOH on» Eoum mEchmmuo mo CoHumHOmH COHuommCH Coumm musom xom m 2 D S L Z S EZLSD RIC mHoupCou 6TLSD haw SILSL‘ [LI 0 LILSD zcv 9TLSP St“ STLSC hIo VILSI‘ In I\ ETLSD 2Hm> HmoooowHH ++ ++ eucsHmHon ++ ++ .ECCmoosa ++ ++ Comaoum ++ ++ XCNHMCQOHO I I uCHom HMHcmuoumECm I I CHmum I I mUOC smemH omHHH I I ammooua I I mmceHm I I CmmHmm I I Hm>HH I I ECOCOHHHmm I I OCCH I I cooHn pummm I I ECHUCMOHHom I I musmHm I I mUOC CQENH HanoEmmmum I I OUOC CQEMH HmHCmmommHm h.meeHm OCHBOHHOM mCu Eoum mEmHCmmno mo COHHMHomHIImCmmHB m.H m.H mudmomxm Hmuo Cmumm mnsom z m xmm CmnECz mHm umuuHH CH mmHm Eoum mmCHpCHm UHOOHOHHmuomm .m OHQMB 57 epithelium was recorded, and l bacterial colony was cultured from the material taken from the lymph node draining the urocyst. All other routine sites were negative on culture, and it is possible that in this female pig the colonization of the urocyst was a result of perineal contamination and an ascending infection. The early involvement of the iliac lymph node indicates that systemic infection could have de- veloped from contamination of the urocyst. At necropsy the stomach was seen to be filled with clotted milk, and there was a marked flaccidity of the in- testinal tract. The tract was filled throughout with watery yellow fluid, and this pig had just started to scour prior to euthanasia. Another pig was killed 4 hours after inoculation, and the organism was cultured from all levels of the gastroin- testinal tract. Some systemic contamination had occurred, with positive cultures of E. 391; being obtained from pre- scapular lymph node and heart blood. All other systemic sites were sterile. This pig did not have diarrhea, and the intestinal tract contents were not markedly fluid. At 8 hours after inoculation, l pig was killed, and heavy growth of organisms was obtained from the mucosae of the oropharynx, stomach, duodenum, and ileocecal value. Only 4 colonies were cultured from the midjejunal mucosa, and the rectal sample was sterile. The skin surface and all systemic sites in this animal were sterile. This pig 58 was clinically normal when killed and, at necropsy,-the in- testinal contents were very fluid within the small intestine, cecum and colon. There was still some meconium retained within the rectum. The remaining inoculated pigs in this litter were killed at 16, 22, 24, 32, 40 and 123 hours, respectively, after oral exposure with the bacterial suspension. Except for the last pig, these animals were uniformly contaminated to a moderate or marked degree in each of the routine sites sampled within the gastrointestinal tract. In 1 pig, 1 or- ganism was detected in a prefemoral lymph node. In a second pig, 1 organism was detected in an iliac lymph node. In the pig killed 40 hours after inoculation the stomach and mid- jejunum were found to be very lightly contaminated, and the duodenum was sterile. The remainder of the intestinal tract in this pig was heavily contaminated. Each of these pigs had profuse watery stool up to the time of euthanasia. Generally the intestinal tract was markedly flaccid and filled with watery fluid at necropsy. The pig killed 16 hours after inoculation had scoured for several hours, and the skin was markedly inflamed in the perineal region. However, the intestine was of normal tone in this animal, even though the intestinal contents were characteristically fluid. Edema of the mesocolon was not observed in any of the test or control animals in this litter. 59 In the pig killed 123 hours after inoculation, all cultures from all 21 routine sampling sites were negative. This animal had all the signs of baby pig enteritis; it had prolific diarrhea with watery stool and was markedly dehy- drated. It had shared a plastic isolator with 2 other in- fected pigs used in this experiment and from which positive cultures had been obtained. As this was the final pig in the series, the isolator had been dismantled, and tissues remaining after necrOpsy had been incinerated or preserved in formalin before the cultures were found to be negative. No explanation was found for this phenomenon. The 2 control pigs from this litter were free from detectable microorganisms at the time of necropsy. These animals were killed at 8 and 24 hours of age, respectively, and were used for clinical and histological controls for this and other litters in this experimental series. There were no gross abnormalities observed in either animal at necropsy. Recovery and Serotyping of the Organisms Seventeen cultures were grown on nutrient agar slants. Each individual culture represented a composite of organisms harvested from the positive culture plates of the experi- mental pigs of a particular litter that were housed in the same plastic isolator. They were forwarded to a serotyping 60 service* at the conclusion of the experiments. This service center reported that all cultures belonged to the E. coli 0 group 138 anui had the K8l antigen, but no check was made for the presence of H antigen nor for hemolytic activity. Clinical and Pathologic Findings Observations were made during the preceding experiments on the clinical effect of oral inoculation and subcutaneous injection of neonatal gnotobiotic colostrum-deprived pigs with E. gel; serotypes 0138:K81:NM (Mich.) and 0138:K81 (Minn.). A record and study of the lesions associated with colibacillosis due to these specific serotypes was made. Clinical Findings There was a distinct difference between the clinical expression of the infections with E. 32;; Ol38:K81:NM (Mich.) and E. 22;; Ol38:K81 (Minn.). Also, a difference was found in the time taken for clinical signs of disease to be ob- served when l serotype was used but introduced by a differ- ent route. Results were as follows: Experiment 1 [E. coli Ol38:K81:NM (Mich.) injected into the subcutis of the umbilical stump] Details of the time of development of systemic contami— nation and enteric infection and the time of disappearance *Dr. Paul Glantz, The National Escherichia Center for Animal Disease, The Pennsylvania State University, University Park, Pennsylvania 16802. 61 of the organisms from the specific testing sites are given in Tables 2, 3 and 4. Litter 1. Although pigs killed at 4, 8 and 24 hours after injection had varying degrees of systemic and enteric involvement with the organism (Table 2), the first clinical sign of disease was seen in a pig when it was killed 32 hours after injection. This pig had marked hemorrhagic panenter- itis (found at necrOpsy), and prior to euthanasia the soft pasty feces of this pig were seen to be very dark and blood- stained. After 32 hours other pigs in this litter began to pass copious amounts of soft or watery feces. By 40 hours after injection the remaining pigs in the litter were scouring. Pigs were killed at 40, 48, 77, 101, 123 and 151 hours after injection. The fecal material from those killed early in this series was of watery consistency, pale yellow, and with suspended fine whitish curds. As time passed the color of the watery material varied between pigs, and the suspended curds were less commonly seen, even though the pigs continued to drink well throughout the experiment. At about 100 hours after injection the pigs were arch- ing their backs as though discomforted by the diarrhea. The first signs of raised hair coat and "scalding" (erythema venenatum) of the perineum were seen in the pig killed 123 hours after injection. At 151 hours after injection the last pig in this experiment had a raised hair coat, was 62 apparently losing weight, had a "scalded" perineum, and had a very watery stool in which there were thin white curds. This animal did not have anorexia at any stage of the ex- periment. The germfree control animals continued to pass soft or typically pasty feces, gained weight, and maintained ap- parent good health until the time of euthanasia. Temperatures of control and test pigs were taken daily in the morning, and all readings tended to be on the low side of an accepted normal (102.0 F.) for pigs (Blood and Henderson, 1963). There were no significant changes in temperature recorded during the experiment. Litter 2. Although pigs killed at l l/2, 2 l/2, 4, 6 1/2, 8, 12 and 16 hours after injection had varying de- grees of systemic and enteric involvement with E. 32;; 0138:K81:NM (Mich.) (Table 3), the first clinical sign of disease was seen in a pig when it was killed 24 hours after injection. This pig had bright yellow watery feces in which find white curds were suspended. At 48 hours after inocu- lation the watery feces were light yellow and there was a loss of "bloom" in the hair coat. In the pig killed 72 hours after injection, diarrhea was recorded. The pig had an arched back, was thin, was apparently dehydrated, and had an erect hair coat. Anorexia was not seen in any pig in this litter dur- ing the experiment. 63 Temperatures were taken under the same conditions as for Litter l, and no significant differences were noted. The control pigs remained germfree and healthy through- out the experiment. Litter 3. Although pigs killed 1 l/2, 2 1/2, 9, 24, and 32 hours after injection had varying degrees of systemic and enteric involvement with E. gel; 0138:K81:NM (Mich.) (Table 4), the first clinical sign of disease in this litter was seen in a pig when it was killed 48 hours after injec- tion. Both this pig and the animal killed at 72 hours after injection had profuse watery diarrhea at the time of eutha- nasia. No instances of anorexia were recorded in this litter during the experiment. Temperature readings were not taken in this experi- mental group. The control animals remained germfree and in good health throughout the experiment. Experiment 2 [E. coli 0138:K81 (Minn.) injected into the subcutis of the umbilical stump] Litter 4. Although pigs killed at 4, 8, 12, 17, 20, 25, 32 and 40 hours after injection had varying degrees of systemic and enteric involvement with the test organism (Table 5), the first clinical sign of colibacillosis in this litter was seen in a pig when it was killed 48 hours 64 after injection. This animal had very soft feces and marked subcutaneous edema of the axilla and hind legs. In the pig killed 72 hours after injection, ventral abdominal edema was observed and the feces were very sloppy in consistency. At 121 hours, the feces were excessively soft, and there was moderate ventral edema. The pig killed at 168 hours after injection had excessively soft feces, but no ventral edema was recorded. Temperatures were taken daily in the morning from this litter, but there were no significant changes in the readings recorded during the course of the experiment. Control animals remained germfree and in good health throughout the experiment. Experiment 3 [E. coli Ol38:K81:NM (Mich.) inoculated per os] Litter 5. Details of organ and tissue isolations with respect to time after inoculation are recorded in Table 6. The first sign of diarrhea was observed in l of the pigs submitted to necropsy 4 hours after inoculation. The entire gastrointestinal tract of this pig was heavily con- taminated by that time. A second pig killed 4 hours post- inoculation did not have clinical evidence of disease and was found to have its gastrointestinal tract heavily con- taminated only so far as the ileocecal valve. The rectum of this pig was still bacteriologically sterile at the time of necropsy (Table 6). 65 One pig killed 8 hours after inoculation did not have a fully colonized gastrointestinal tract (Table 6) and did not have clinical signs of disease. The single pig killed 16 hours after inoculation had profuse, clear, watery, yellow feces in which fine yellow curds were suspended. This pig had a "scalded" perineum and had been scouring for i an unrecorded period of time prior to euthanasia. Profuse watery yellow feces and "scalding" of the per- ineum were characteristic of the pigs killed at 22, 24, 32, 40 or 123 hours pestinoculation.. The pig killed at 123 hours postinoculation was extremely thin and apparently de- hydrated. Anorexia was not reported regarding any pig in this litter during the experiment. Temperatures of the pigs in this litter were taken 3 times daily (8 a.m., noon, and 5 p.m.). During the first 2 postnatal days the temperatures were on the low side of normal for all pigs in the experiment, but by the 3rd day of life (first day postinoculation) the temperatures of the remaining pigs were within the range of 100.4 to 103.2 F and remained at approximately this level throughout the ex- periment. Control pigs remained germfree and in good health throughout the experiment. Hematology In only 3 litters was it possible to get satisfactory blood samples from the experimental animals both immediately 66 after delivery and again immediately prior to euthanasia. These were Litters 3, 4 and 5. Litter 3. In this litter both control pigs had a marked fall in total leukocyte counts to 1/2 and 1/3, re- spectively, of the numerical values recorded from samples taken on the day of delivery. The depletion of lymphocytes was relatively greater in both pigs than was the neutrophil depletion. At the time of necropsy the age of these 2 pigs was 23 hours and 44 hours, respectively. The fall in leukocyte counts was not uniform through- out the injected pigs in this litter, and no trend could be seen either in total leukocyte count or in the absolute dif- ferential leukocyte counts of these animals. Litter 4. In 2 of the 3 control pigs complete hemo- grams were recorded. In 1 pig, killed at an age of 44 hours, there was a slight drop in the leukocyte count to a value still within the normal range. In the second pig, killed at an age of 176 hours, there was a marked drOp in the leu- kocyte count to a value 2/3 of that accepted as a normal minimum (Calhoun and Smith, 1964). In this case, however, the decrease was due to a loss of neutrophils. The lympho- cytes were actually increased in total numbers in this animal. The general trend in the hemograms recorded for day- of-birth and terminal blood samples was a moderate to marked fall in the leukocyte count, usually from initial values at 67 the lower level of the accepted normal for conventional pigs of this age (as cited by Calhoun and Smith, 1964) to values far below the acceptable normal minima. This decrease was due either to an equal decrease in both neutrophils and lymphocytes or to a marked decrease in the lymphocyte count which in 2 instances was associated with a concomitant ab- solute increase in the neutrOphilic fraction. Litter 5. The hemograms from the germfree animals in this litter were insufficient in that the terminal hemogram from 1 of the animals was not recorded because the blood sample clotted while being removed from the isolator. The second control animal was killed 24 hours after inoculation. It had a ninefold increase in total leukocytes, due almost entirely to an increase in the neutrophilic fraction of the leukocytes. The function of these hemograms as control hemograms was discounted. In the hemograms of the 2 pigs killed 1 1/2 hours postinoculation, there was an approximate 2/3 reduction in the circulating leukocytes, the major reduction being in the circulating neutrOphils. A similar change was recorded in the hemograms of the 2 pigs killed 4 hours after inocu- lation. In the pig killed 16 hours postinoculation the leuko- cyte count was approximately half of that recorded at birth (60 hours before), and the fall in the count was apparently 68 due to approximately equal decreases in both the lymphocyte and neutrophil counts. This trend persisted in the hemograms of the pigs killed 22, 24, 32 or 40 hours after inoculation. At 40 hours postinoculation, the total hemoglobin content and leukocyte counts were still below the normally acceptable limits for conventional pigs of this age range (Calhoun and Smith, 1964). Histopathologic Findings» The histopathologic lesions found in this study were predominantly in the gastrointestinal tract. The lesions ranged from those of an extensive acute hemorrhagonecrotic enteritis to a histologic picture, in persistently scouring pigs, microscopically indistinguishable from that seen in clinically normal germfree control animals. Studies were made of a number of systems, and lesions were recorded as follows: Skin and umbilical stump. Within 1 1/2 hours of in- jection of E. 22;; Ol38:K81:NM (Mich.) there was mild hem— orrhagic edema and perivascular fibrinous exudate in the subcutis of the umbilical stump. By 2 1/2 hours neutrOphils had entered but were still concentrated in perivascular lo- cations (Figures 6 and 7). Neutrophils, eosinOphils and lymphocytes increased in numbers in the site of inflammatory edema for the next 12 to 24 hours. At 12 hours postinjec- tion, necrosis of inflammatory cells was evident and, at 69 Figure 6. Pig J4662. Injection site 4 hours after injection of E. coli. Connective tissue edema (A) and perivascular collection of neutrophils (N). Hematoxylin and eosin. x 190. o¢n'lpf- ’ I Figure 7. Pig J4662. Same field as Fig- ure 6. Perivascular neutrophils are spreading into the injection site from neighboring ves- sels. Hematoxylin and eosin. x 470. 70 14 hours, macrophages were observed in the area. At 72 hours, attempts to enclose a necrotic mass of neutrophils with fibroblasts were observed. By 123 and 168 hours there was abscess formation with necrotic cells walled off by foamy macrophages and granulation tissue (Figure 8). In the first 40 hours following injection of E. 39;; Ol38:K81 (Minn.) (as in Experiment 2) the tissue response was essentially the same as for E. 92;; Ol38:K81:NM (Mich.). A significant observation was that in each of the 4 pigs killed after longer periods of time (48, 72, 121 or 168 hours postinjection) there was abscess formation at the site with the increasing develOpment of a wall of macro- phages and fibroblasts. Peripheral to the developing abs- cess there was slight to moderate hemorrhage and infiltra- tion with inflammatory cellular (primarily neutrophilic) exudate. Superficial lymph nodes. Within 4 to 8 hours after injection of the test organisms, there was reduction in the number of mature lymphocytes present in the superficial lymph nodes (Figure 9). At time intervals beyond 8 hours there was an increase in the numbers of neutrophils and. eosinophils in the nodes and an associated peritrabecular and perinodal edema. Eosinophils were markedly increased in the lymph nodes by 100 hours postinjection. Neutrophilic lymphadenitis was observed in l pig killed 101 hours after 71 o , '9 _ ._ ’ ‘ 'l.‘ Figure 8. Pig J5722. The injection site. E. coli was injected 168 hours before necropsy. An abscess surrounded by macrophages , fibroblasts, other inflammatory cells and edema. Hematoxylin and eosin. x 75. Figure 9. Pig J6205. Prescapular lymph node. E. coli was injected 24 hours before necropsy. There is edema and an apparent lack of mature lymphocytes in the germinal centers. Hematoxylin and eosin. x 75. 72 injection (Figure 10). Superficial lymph nodes from the germfree controls were also edematous and depleted of mature lymphocytes, but these changes were not nearly so marked in degree as were those of the infected pigs. Trachea and lung;. Consistent findings in histopatho- logical examination of the lungs were atelectasis and emphy- sema. There were, in addition, many instances of alveoli being filled with homogeneous eosinophilic material. It is possible that these phenomena, equally present in control animals, are a result of the hysterotomy procedures at de- livery. In 2 instances, hydropic degeneration of the terminal bronchiolar epithelium was observed--both pigs were under— going early systemic involvement and early establishment of the test organism in the intestinal tract. In 2 pigs heavily contaminated in all routine sites (including lung) and in which gross intestinal lesions were observed, there were significant changes in the lungs. In the first pig, killed 24 hours postinjection, there were hyperemia of the alveolar capillaries and fibrinous exudate partly or completely filling some of the alveoli. In other alveoli an exudate of erythrocytes, fibrin, proteinaceous fluid, neutrophils, lymphocytes and macrophages with desqua- mated alveolar cells was observed. These were also vacuola- tion and desquamation of the mucous epithelium of the associated bronchioles. 73 Figure 10. Pig J4673. Prescapular lymph node. E. coli was injected 101 hours before necrOpsy. Acute lymphadenitis. The lymph node is infiltrated with neutrophils (N). Hematoxy- lin and eosin. x 190. 74 In the second pig, killed 32 hours after injection, there was marked hyperemia of the alveolar capillaries. No other changes were observed. There were no apparent changes of significance in the trachea. Heart and aorta. No consistent changes were observed in the heart or the aorta. Ianalmost all instances cells with the configuration of nuclear chromatin character— istic of the Anitschkow myocyte were extremely common in the sections of myocardium of both infected and control animals; in some instances they were so common as to suggest that the Anitschkow myocyte is not a myocardial histiocyte but an immature normal myocardial cell and of little or no signifi- cance when observed in sections of the myocardium of a germ- free pig (Figure 11). Myocardial edema and diffuse intramyocardial hemorrhages were observed in l pig with systemic bacterial contamination and established enteric infection. This pig was killed 20 hours after injection of the test organism. In another pig with systemic and early enteric contami— nation resulting from injection with the test organism 8 hours earlier, the aortic endothelium was distended, swollen and vacuolate, with the nuclei consistently displaced toward the lumen of the aorta (Figure 12). Almost every muscle cell in the section of the myocardium of this pig was an Anitsch- kow myocyte. 75 Figure 11. Pig J5716. Myocardium. A majority of the myocardial nuclei have the chromatin configuration characteristic of Anitschkow myocytes. Hematoxylin and eosin. x 750. . Figure 12. Pig J5712. Aorta. E. coli was injected 5 hours before necropsy. Vacuoli- zation of endothelial cells. Hematoxylin and eosin. x 75. 76 Spleen. It was observed that there was an apparent slight to moderate depletion of mature lymphocytes from the spleens of some of the infected animals when compared with those of the germfree controls. Thymus. No consistent changes were observed in the thymus. In 1 pig (J4663) killed 8 hours after injection, from which there was isolation of the organism at 7 sys- temic sites, the interlobular septa of the thymus were in- filtrated with edema fluid containing many mature lympho- cytes. In another pig with systemic and early enteric con- tamination with E. 92;; (J5712), a discrete mass of neutro- phils was observed in the medulla of l lobule. .In the same experimental group, a pig (J57l7) with slight systemic contamination and heavy colonization of the entire intestinal tract had a thymus markedly infiltrated with eosinophils. Thyroid and adrenals. There were no consistent changes observed in the thyroid. In 1 pig killed 20 hours after injection (J5715), in which there was systemic con- tamination and enteric colonization, edema of the inter- stices of the thyroid was observed. A generally consistent finding in the adrenals of both the monocontaminated and the germfree pigs was hyper- emia of the medulla and the zona reticularis. This hyper- emia was more pronounced in those pigs in which there was 77 only systemic bacterial involvement than in those killed 72 hours or more after injection. The hyperemic response in those pigs killed 72 hours or more after injection was more apparent than in the germfree control group. Foci of eosinophils were observed in the adrenal cor- tex of l pig killed 25 hours after injection with the test organism (J57l6). Alimentary tract Esophagus. No significant changes were observed. The test organism when given by mouth was seen to colonize in the debris characteristically attached to the esophageal epithelial surface in gnotobiotic pigs on liquid diets. In some instances there was parakeratosis of the epithelium (Figure 13) and infiltration of the epithelium with neutro- phils was observed (Figure 14). Stomach. Sections were cut from the cardiac, fundic and pyloric areas of the stomach of each pig. There was wide variation in the lesions observed in these tissues, and at the end of the scale of decreasing severity of lesions there were no apparent differences between the tissues of the monocontaminated pigs and the germfree pigs. In these pigs the simple columnar mucous epithelium was blocky and arranged in orderly fashion in a single layer of cells, the lamina pr0pria and submucosa being infiltrated with occa- sional lymphocytes and mononuclear cells (Figures 15, 16 and 17). 78 Figure 13. Pig J5716. Esophagus. Para- keratosis of the esophageal epithelium (P). Hematoxylin and eosin. x 75. Figure 14. Pig J6207. Esophagus. E. coli was injected 72 hours before necropsy. Neutrophils (N) infiltrating the squamous epi- thelium of the eSOphagus near the cardia. Hematoxylin and eosin. x 750. 79 Figure 15. Pig K63. Mucosa of the cardiac region of the stomach from a germfree control, illustrating minimal epithelial activity to be seen in control animals. Hematoxylin and eosin. x 470. Figure 16. Pig K63. Mucosa of the fundic region of the stomach from a germfree control illustrating maximal epithelial activity seen in control animals. Hematoxylin and eosin. x 470. 80 Figure 17. Pig J5718. Fundic region of the stomach. E. coli was injected 40 hours before necropsy. Moderate epithelial activity of the heavily colonized gastric mucosa of a monocontaminated pig. Hematoxylin and eosin. x 470. 81 In the mucosa of most of the pigs from which the test organism was cultured, the mucous activity of the epithe- lium was increased. The individual cells were larger, more densely packed, and elongated, and in some instances the cytoplasm was so pale staining and increased in amount as to give the mucous cells a vacuolate appearance. In each instance where distinct microscopic lesions were observed, the cultures were found to be positive, in- dicating heavy contamination of the stomach by the test organism. Changes observed in the mucosa were distention and vacuolation of individual epithelial cells. Instances of massive simultaneous discharge of the epithelial cell contents with outpouring of mucus onto the gastric surface were seen. Necrosis, dissolution of the epithelial surface and some epithelial sloughs were recorded from some infected pigs (Figures 18 and 19). In the fundic mucosa of l pig there were foci of neutrophils (Figure 20). Hyperemia of the gastric mucosa of infected pigs was common, as were edema and hypercellularity of the lamina propria and areolar coat of the submucosa. Major cell types present to varying degrees in these tissues were primarily lymphocytes, plasma cells and neutrophils, with some macrophages and an unex- pectedly high proportion of eosinophils. The small intestine. Sections were routinely made from duodenum, jejunum at the 50-cm., lOO-cm. and 200- cm. levels, and ileum close to the ileocecal valve. It is 82 Figure 18. Pig J4669. Pyloric region of the gastric mucosa. E. coli was injected 32 hours before necropsy. There is necrosis (N) and dissolution (D) of portions of the epithe- lial surface and the lamina propria. Seromucus (S) is seen to be streaming from some epithe- lial elements. Hematoxylin and eosin. x 190. Figure 19. Pig J4669. Pyloric region of the stomach. E. coli was injected 32 hours before necropsy. There is a focus of necrosis and early ulceration of the mucosa. Hematoxy- lin and eosin. x 190. 83 Figure 20. Pig J6207. Fundic region of the stomach. E. coli was injected 72 hours before necrOpsy. A few scattered neutrophils (F) are in the lamina propria. Hematoxylin and eosin. x 750. 84 apparent that marked variations from the normal can occur as a result of infection with these 2 serotypes of E. 2211' It is also apparent that much variation from the normal to abnormal can occur as one examines tissues taken from neigh- boring sites in a heavily colonized intestinal tract. (i) The epithelium and lamina propria. In some infected pigs and some germfree control animals the surface epithelium was ordered and regular with little or no mucous activity and with occasional active goblet cells filled with mucin and becoming more frequent in more caudal sites (Figure 21). According to the level of the section, lympho- cytes were present singly or in aggregates in the submucosa and in increasing numbers caudad. Some slight but distinct changes from the above were also seen in the control and monocontaminated pigs. The volume of the cytoplasm of the individual columnar cells of the mucous epithelium was uniformly greater and less in- tensely stained. In some instances the epithelial nuclei were uniformly centrifugally placed in relation to the cell cyt0plasm and the axis of the villus (Figure 22); in other sections the epithelial cell nuclei were uniformly axially placed, close to the basement membrane of the epithelium. The cytoplasm of these cells was negative or very faintly positive to Mayer's mucicarmine stain. A more advanced degenerative change very frequently observed was that in both control and infected animals columnar epithelial cells of the villi were markedly 85 Figure 21. Pig J5723. Terminal ileum. Germfree control. Moderate vacuolization of the epithelium of the villus has little or no apparent effect on the ability of the villus to absorb some fluid from the lumen of the intestine. Lymphatic vessels and lacteals are visible (P), indicating a degree of function. Hema- toxylin and eosin. x 190. Figure 22. Pig K57. Terminal jejunum. E. coli was introduced per cs 4 hours before necropsy. Trans- verse section of the villi. Some slight lymphatic and lacteal function is apparent (F); the epithelial cells are enlarged and vacuolated (E). Epithelial cell nuclei are abaxially situated. Hematoxylin and eosin. x 470. 86 distended and/or vacuolated either singly or en masse (Figures 23, 24, 25 and 26). Varying degrees of subepithelial edema of the villi were seen to be associated with l or more of the epithelial changes described above (Figure 27). Commonly observed was edema of the tips of the villi, forming subepithelial bullae (Figures 28 and 29). In some instances this edema was re- stricted to the tip of the villus, but it was also seen to extend in some cases the entire length of the villus. In other cases the subepithelial edema was observed at the base or midvillus region leaving the tip relatively un- changed (Figure 30). In a few instances distinct separation of the basement membrane from the lamina propria indicated that the subepi- thelial space was a fixation artefact. ‘Generally, the con- dition of subepithelial edema was indicated by subepithelial vacuolation bounded by fibrous stroma (Figures 30 and 31) and sometimes containing pale-staining homogeneous or granu- lar eosinOphilic material and/or fibrin strands (Figures 27, 28, 29 and 32). At times, edema of the axial tissues of the lamina propria of the villus was seen. Other changes commonly observed in the villi and not seen in the villi of the intestines of control pigs were hy- peremia with or without hemorrhage and/or necrosis. In some instances there was surface necrosis with sloughing of the tips of the villi and shortening of these structures to pro- duce a short, stumpy villus. 87 Figure 23. Pig J5721. Midjejunum. E. coli was injected 121 hours before necrospy. Epithelial cells are enlarged and vacuolated. Nuclei are abaxially situated. Hematoxylin and eosin. x 470. Figure 24. Pig J4672. Midjejunum. E. coli was injected 77 hours before necropsy. Marked vacuolization of the epithelial cells (V). Neighboring villi are very closely ap- posed and the laminae propriae of individual villi are compressed. Hematoxylin and eosin. x 470. 88 Figure 25. Pig J5721. Terminal ileum. E. coli was injected 121 hours before necropsy. very pronounced vacuolization of the epithelial cells. Compression and apparent inactivity of elements of the lamina propria. Hematoxylin and eosin. x 190. Figure 26. Pig J5721. Terminal ileum. From the same field as Figure 25. Marked vacu- olization of epithelial cells. Hematoxylin and eosin. x 470. 89 Figure 27. Pig J6206. Jejunum. E. coli was in- jected 48 hours before necropsy. There is early vacu- olization of the epithelium, but it is within the range of normality of jejunal epithelium of germfree control pigs. There is evidence of fluid transfer from the jejunal lumen into the subepithelial spaces and the vessels of the lamina propria. The lamina pr0pria is infiltrated with some neutrophils. The villus has undergone partial longitudinal contraction. Hematoxylin and eosin. x 190. L ' .' . . Figure 28. Pig J5721. Duodenum. E. coli was in- jected 121 hours before necropsy. Bullae at the tips of duodenal villi. There are some neutrOphils in the lami- na prOpria, and fluid absorbed from the lumen of the duo- denum has been transported into the subepithelial spaces (S). Hematoxylin and eosin. x 470. 90 Figure 29. Pig K60. Midjejunum. E. coli was introduced er os 40 hours before necropsy. A —villus undergoing an 1nadequate" longitudinal contraction. There is some vacuolization of the epithelium and the villi are distended with eosinophilic granular material (M). Hematoxylin and eosin. x 190. ‘.’ J - .¢.: ‘1 . . 4". " ‘. ‘r‘m. «Inga; NH 3. .. ‘ 4,‘ ~. $~\!‘\.'UA- 3 .4flb‘l'l‘"""“""' "< 9-2 § . - * .4; 28”» 3» .. . ‘3”;‘5; K ‘ L 3&¥3f§‘$,§. \ 1‘» (Wai‘ pH .8. ~. Os. Figure 30. Pig K62. Midjejunum. E. coli was introduced per cs 8 hours before necropsy. There is evidence that active longitudinal contraction has occurred. The epithelium is minimally vacuolated and apparently there is fluid transfer from the in- testinal lumen into the subepithelial spaces, lac- teils and lymphatic vessels. Hematoxylin and eosin x 90. Figure 31. Pig K57. Terminal jejunum. E. coli was introduced er cs 4 hours before necropsy. Trans- verse section oE the villi of the terminal jejunum. There are varying degrees of vacuolization of the villus epithelium. Where vacuolization is marked, villus function indicated by subepithelial collection of tranSmitted fluid, and activity of lymphatic vessels and lacteals, is minimal. Hematoxylin and eosin. x 470. Figure 32. Pig K54. Terminal ileum. E. coli \was introduced per os 16 hours before necropsy. There is accumulation of fluid in the epithelium and the sub- Iepithelial spaces of the villi (F). The laminae pro- Ipriae are infiltrated with neutrophils. Hematoxylin and.eosin. x 470. 92 A further lesion seen in the infected intestine, and generally not associated with many of the changes described above, was that in which the villi were markedly enlarged due to vacuolization of the epithelial surface or due to subepithelial edema, so that the villi were in very close apposition and appeared adherent to each other, the micro- villi forming a dense eosinophilic boundary common to both villi (Figures 33 and 34). The gross appearance of this intestine is of a thicker-walled intestine that is firm to the touch compared to the flaccid intestine that is commonly seen in colibacillosis. At times in the infected intestine the lamina propria of some of the villi were infiltrated with inflammatory cells, primarily neutrophils (Figures 32 and 35), and in some in- stances these cellular aggregates formed microabscesses that distended the shaft of the villus. Other inflammatory cells commonly associated with a general edema of the villi were lymphocytes, eosinophils, plasma cells and occasional macro- phages. In some instances, the jejunal villi were seen to be very long and slender. There was almost uniform, advanced hydropic change in the epithelium of the villi in these sec- tions, there was little or no subepithelial vacuolation, and Breucke fibers were not contracted (Figure 36). Fre- tquently the laminae propriae of these villi were infiltrated with foci of neutrophils. 93 Figure 33. Pig J4672. Midjejunum. E. coli was injected 77 hours before necropsy. The villi are mark- edly enlarged and closely apposed. The intervillous spaces are closed, effectively reducing the surface available for absorption of fluids from the intestinal lumen. Hematoxylin and eosin. x 190. Figure 34. Pig J4672. Midjejunum. From the same field as Figure 33. The mucous membrane is markedly vacuolated and there is minimal functional activity in the lamina propria. Hematoxylin and eosin. x 470. 94 Figure 35. Pig K55. Terminal jejunum. E. coli was introduced er os 22 hours before necropsy. There are few changes. The lamina propria is infiltrated with some neutrophils. Hematoxylin and eosin. x 470. Figure 36. Pig K55. Midjejunum. E. coli was introduced per os 22 hours before necropsy. The ‘Villi are elongated and the Breucke fibers are not contracted. There is little evidence of absorption <3f fluids from the lumen of the jejunum. Hematoxylin and eosin. x 75. 95 (ii) The submucosa and other structures of the small intestine. The changes observed in the submucosa and the muscular and serosal layers of the small intestine of the monocontaminated pigs were primarily inflammatory edema and infiltration with lymphocytes, neutrophils, eosinophils and macrophages. Hyperemia and distended lymphatic vessels of the submucosa and occasionally intramuscular edema were observed in the infected pigs. Mature lymphocytes were present in the submucosa of the terminal jejunum and ileum of both the infected and the germfree control pigs, but there were increased numbers of lymphocytes in these tissues in the pigs heavily contaminated with the test organisms. Mesenteric lymph nodes. Changes recorded in the mesenteric lymph nodes were inconsistent. In one in— stance there was a marked increase in the numbers of neu- 'trophils present in foci in the interstices of the lymph IIOde. Edema was common, and in many instances there was ian apparent depletion of mature lymphocytes. Large intestine. Sections of cecum and colon firom the infected group were characterized by submucosal eniema and hypercellularity of the lamina propria. In some illstances there were hyperemia and excessive goblet cell a£rtivity in the mucosa. Edema of the mesocolon was a fea- tuI‘e in both the infected and control pigs. The edema of the infected pigs, however, was inflammatory (Figure 37) , anti in one instance the edema of the mesocolon was hemor- rhag ic . .96 Figure 37. Pig J4669. Mesocolon. E. coli was injected 32 hours before necropsy. Edema and hyperemia of the mesocolon; lymphocytes predominate, with some neutrophils and histiocytes also present in the exudate. Hematoxylin and eosin. x 75. 97 Microabscesses and ulceration of the mucosa, and edema, fibrin and hemorrhage in the submucosa were observed in the colon of l pig killed 48 hours after injection with the test organism (Figures 38 and 39). In 1 pig (J4669) classical lesions of hemorrhagic enteritis were noted, the lesions involving stomach, small intestine and large intestine. These lesions were charac- terized by necrosis, hemorrhage and sloughing of extensive areas of the mucosa. There were also hyperemia and edema of the muscular coats of the intestine (Figures 40, 41, 42, 43 and 44). In some areas of the jejunum, total mucosal slough, including portions of the underlying muscle coats, was observed (Figures 41 and 42). In the large intestine there were hyperemia, hemorrhage, and necrosis of the mu- Icous epithelial surface (Figure 44) and catarrhal exudate vmith some pseudomembrane formation in the colom. The rec- ‘tal mucosa of this pig was relatively undamaged. Liver and pancreas. No consistent changes in the Jriver and pancreas were associated with infection. In 1 liver section from which the organism had been CLthured there were marked hyperemia and increased numbers 015 neutrophils scattered throughout the liver parenchyma and in the vicinity of larger blood vessels (Figure 45). In thfi! livers of some pigs with an established enteric infection there was mild centrolobular vacuolization of the hepatic Parenchymal cells (Figure 46) . Edema of the submucosa of thE3 cholecyst was observed in some infected pigs (Figure 47)- I! M Q . . . Q ’0‘; \ .' ’ Ifoya,‘ J' ’ Figure 38. Pig J6206. Colon. E. coli was in- jected 48 hours before necrOpsy. Focal ulceration and suppuration (S) of the mucosa of the colon. Hematoxylin and eosin. x 75. Figure 39. field as Figure 38. NeutrOphilic cellular exudate from the ulcerated mucosa. Pig J6206. Colon. From the.same Hematoxylin and eosin. x.470. 99 Figure 40. Pig J4669. Duodenum. E. coli was injected 32 hours before necropsy. Hyperemia (R), hemorrhage (S), necrosis and sloughing of the villi (T). Hematoxylin and eosin. x 75. Figure 41. Pig J4669. Terminal jejunum. E. <:oli was injected 32 hours before necropsy. Hemor- Jrhage and necrosis of the mucosa. In this instance 'the entire mucous membrane, lamina propria and mus- <:ularis mucosa are in the process of sloughing. liyperemia and hemorrhage of the submucosa (X). Hematoxylin and eosin. x 75. 100 I . Figure 42. Pig J4669. Terminal jejunum. From the same field as Figure 41. Hyperemia (K), hemor- rhage, necrosis and epithelial sloughs. Hematoxylin and eosin. x 190. Figure 43. Pig J4669. Terminal ileum. E. coli was injected 32 hours before necropsy. Hyperemia, hemorrhage and necrosis. Hematoxylin and eosin. x 75. 101 Figure 44. Pig J4669. Colon. E. coli was in- jected 32 hours before necropsy. Hyperemia and hemor- rhage of the lamina propria of the mucous membrane. There is some necrosis of the epithelium (M). Hema- toxylin and eosin. x 190. Pig J 6206. Liver. E. coli was in- Foci of neutrophils E. coli Hematoxylin and eosin. Figure 45. jected 48 hours before necropsy. infiltrating the parenchyma of the liver. was cultured from this liver. x 750. 102 . ~113‘5'“ h \. ~47" ".4 l» -. I, Figure 46. Pig J57l4. Liver. E. coli was in- jected 17 hours before necropsy. Centrolobular vacu- olization of hepatic cells was seen in a number of the pigs with enteric infections. Hematoxylin and eosin. x 190. Pig K62. Liver and cholecyst. E. er cs 8 hours before necropsy. The lymph Hematoxylin and eosin. x 75. Figure 47. coli was introduced Edema of the submucosa of the cholecyst. vessels are distended. 103 Urinary system. Positive isolations of a few organisms were common from the kidney during the early systemic phase of the infection. Lesions noted in these instances were edema and hemor— rhage of the medulla, hyperemia of the arcuate and interlo- bular vessels and the glomerular tuft, mild infiltration of the interstices of the cortex with mature lymphocytes, vacu- olization of the epithelium of the excretory tubules and in- filtration with mononuclear cells. In a few instances, isolations from the epithelium of the urocyst indicated heavy contamination and the possibility of establishment of an infection at that site. In only 4 pigs did the lesions observed suggest sig- nificance. These lesions were hyperemia, hemorrhage, edema .and.perivascular lymphocytosis of the lamina propria, mono- riuclear cell infiltration into the subepithelial tissues; and vacuolation, hydropic degeneration and necrosis of the txransitional epithelium (Figures 48 and 49). Joints. Bacteriological and gross examination of the l'11-‘lmeroradial joint of each pig was undertaken as a routine necropsy procedure. Positive isolations of the organism were obtained from 1:l'lis site from 3 pigs with systemic contamination in .the 8- tZCD 32-hour postinjection period. There were no clinical signs or gross lesions of ar- thritis in these pigs. Histopathologic examination of the joints was not done. 104 11". Figure 48. Pig J6211. Urocyst epithelium from a germfree control pig. Hematoxylin and eosin. x 470. .“t ‘ 1. - ' 311$; Figure 49. Pig J6791. Urocyst epithelium. E. coli was injected 72 hours before necropsy. Hydrop1c degeneration and distended lymphatics (L). Cultures from this epithelial surface indicated that a large population of E. coli 0138:K81:NM (Mich.) was present. Hematoxylin and eos1n. x 470. DISCUSSION Microbiologic Aspects of Experiment 1 The organisms did not actively spread from the injec- There was tion site along tissue planes in the subcutis. some degree of spread from the injection site but this was minimal and the result probably of mechanical transfer with The organism was tissue fluid and inflammatory exudate. non-motile, and this might explain the lack of active move- ment from the injection site. It was demonstrated in Experiment 1 that pathogenic serotypes of E. coli could systemically invade colostrum- uld be infected immediately after birth and develop bac- teremia and early enteric infection before receiving its firstcolostrum, even if lactation was delayed for only 4 Therefore, contamination 11cburs after parturition (Table 3). (’15 'the exposed tissues of the umbilical stump with coliforms c‘i’t-l-‘Ii‘ing the neonatal period is possibly of greater signifi- carIce than has previously been considered. 105 106 In the 3 litters used in Experiment 1 it took from 6 to 24 hours for the organism to pass through a systemic or bacteremic phase and to colonize the intestinal tract in significant numbers. The consistency with which this course of infection was followed indicates that the serotype under test will very readily colonize the intestinal tract from systemic or bacteremic contamination and that this may in fact be a common route of infection in slightly older pigs which have active gastric secretion and low pH of gastric contents. The reduction in numbers of organisms isolated from the stomach mucosa, as pigs approached an age of ap- proximately 150 hours, indicates that developing gastric acidity might have had an adverse effect on the colonization of the stomach with this serotype. It is of interest to note that, 24 to 48 hours after injection, systemic contamination was not common (Tables 3, 4 and 5). There was occasional localization in specific sites such as the kidney and urocyst, and isolations could be made from the lymph nodes draining these organs, indi- cating the possibility of transient or persistent bactere- mias derived from such localization of infection. It was also apparent that occasional reentry of the organism into the circulation could occur from heavily populated enteric sites. Local damage from microabscesses and trauma of the nnucosa by the ingesta could readily expose vessels to in- \rasion by organisms resident in the vicinity. 107 Microbiologic Aspects of Experiment 2 Escherichia coli 0138:K81 (Minn.) produced a milder clinical disease in gnotobiotic pigs than did the serotype used in Experiment 1. It did, however, behave in an essen- tially similar way--there was early bacteremia in the 4 to 12-hour postinjection period, and by 12 hours there was heavy colonization of the gastrointestinal tract. Systemic contamination was uncommon 20 hours after exposure, and only occasional systemic isolations were recorded after that time. Microbiologic Aspects Common to Both Experiments I—and 2 Contrary to the classically held concept that encap- sulated virulent bacteria require sensitization with speci- fic antibody before being susceptible to phagocytosis, the work of Wood eE_3E. (1946) on the phenomenon called "surface phagocytosis" is partly borne out by observations made inI these experiments. In independent experiments Brooks (1966) has shown that the pigs used in this series of experiments were agammaglobulinemic. It is apparent that bacteremia and localization of organisms in various systemic sites of these moncontaminated pigs was a transient phenomenon, and that a mechanism was operating to remove organisms from these tissue sites. Again it is clear from the histologic study of the injection sites and adjacent skin sections that 'there is a cellular response on the part of the baby pig in 'the infiltration of these sites with neutrophils, eosinophils 108 and macrophages. One can postulate, therefore, that despite the lack of gamma globulin, these animals were able to promptly phagocytose and/or inactivate and remove contami- nant microorganisms from certain tissue sites, provided that the organisms were not present in excessive numbers. The differences between specific and nonspecific opso- nization seem to be quantitative rather than qualitative (Suter, 1956). It is probable that the numbers of pathogens introduced at the injection site far exceeded the number that the mobilized phagocytes could remove by "spontaneous phagocytosis" (phagocytosis unassisted by specific serum components). As a consequence, the attempted repair pro- cess at the injection site involved necrosis, abcess forma- tion and fibrosis. Wood et al. (1946) observed, however, that phagocytosis in the absence of serum components did not occur upon "smooth" surfaces or substances, such as when bacteria and phagocytes were placed together on a thin layer of mucus on a glass slide. It may be that the reason for the persistence of the organisms in the intestinal tract is the inability of the phagocyte to trap coliforms in the mucous menstruum such as is found in the intervillous lumen of the intestine of the agammaglobulinemic monocontaminated pig. Surface phenomena in the tissues assist phagocytes in their activities in clearing systemic sites such as liver, kidney, spleen and lung of contaminating E. coli. 109 Microbiologic Aspects of Experiment 3 The site consistently infected by this organism was the gastrointestinal tract. Following oral dosing the in- fection was immediately established in this site and did not necessarily pass through a pre-enteric bacteremic phase as recorded in the pigs from Experiments 1 and 2. It was evident from the results of this experiment that a neonatal colostrum-deprived pig is very readily in- fected by oral contamination. It is apparent that even a brief period of agalactia in a parturient sow would permit the gastrointestinal tract of neonatal pigs exposed to pathogenic serotypes to become heavily contaminated, thereby constituting a serious threat to the survival of the litter. Stevens (1963) states that conventionally-reared pigs undergo an invasion of the viscera with E. 99;; in the termi- nal stages of colibacillosis and that the organism can be more frequently recovered from the brain than from "other sites." He considered a possible reason for this phenomenon to be that there are fewer substances inhibitory to growth of the organisms in the brain. There were occasional in- stances of transient bacteremia recorded in Litter 5, and in l pig with an apparent ascending infection of the urinary tract there was a positive isolation from the iliac lymph node draining the urocyst (Table 6). It is likely that the profuse diarrhea characterizing colibacillosis in gnotobiotic pigs due to E. coli 0138:K81:NM 110 (Mich.) is dependent on the presence either in the mucosa of the intestinal tract or the lumen of this organ, of suf— ficient numbers of the organisms, and a latent period of 4 to 18 hours. Presumably, the shorter period of time related to the contaminating dose having reached this site in the gastrointestinal tract by a direct route. It is apparent that the oral dose of approximately 5 x 106 organisms of this serotype is sufficient to produce diarrhea in a gnoto- biotic pig within 4 hours of the exposure. In gnotobiotic pigs systemically infected with this serotype of E. 2211' sufficient time must elapse for the organism to colonize the gastrointestinal tract, and in sufficient numbers, before the physiological activity of the intestine is so altered as to produce diarrhea. Clinical and Pathologic Findings from Experiments 1, 2 and 3 Clinical aspects There was a significant difference in the clinical expression of the disease associated with each of the 2 serotypes. The serotype, E. 92;; 0138:K81:NM (MichJ, produced very pronounced diarrhea and dehydration and, under field conditions, one would anticipate infection with this sero- type to result in a high mortality rate. Systemic infection with E. gel; 0138:K81 (Minn.) was associated with a higher incidence of localized ventral 111 edema. The diarrhea and dehydration produced in these animals, under the environmental conditions of the experi- ment, was relatively mild. With both serotypes, in those instances where the perineum was grossly and persistently contaminated with feces containing the organism, erythema venenatum was ob- served. In these circumstances it is not surprising that an occasional urocystitis, probably from an ascending in- fection of the female urethra, was recorded. The clinical signs of the disease produced by both oral exposure and subcutaneous injection of gnotobiotic neonatal pigs with E. gel; Ol38:K81:NM (Mich.) closely re- sembled the clinical descriptions given elsewhere, for other pathogenic serotypes of E. 92;; (Gordon and Luke, 1958; Saunders eE_EE., 1960; Stevens, 1963; Dunne, 1964). Anorexia may or may not be a part of the clinical syndrome as it is described in the literature. With more pathogenic serotypes of E. 2211' anorexia is more apparent as the pigs become dehydrated and lethargic, and it is likely that an- orexia is a sign of a deteriorating clinical state rather than a specific manifestation of colibacillosis. Anorexia was not a significant clinical finding in the series of ex- periments reported here. gross lesions. The gross pathological changes observed in the gnoto- laiotic pigs in this experimental series resemble in part 112 those descriptions given by previous workers and recorded in highly regarded veterinary texts (Jubb and Kennedy, 1963; Dunne, 1964); but it is apparent that textbook descriptions should now take into account that different serotypes, readily identifiable, produce basically different syndromes in neonatal pigs. The enteric syndrome produced in this series of experiments is clinically and pathologically separate from the polyserositis syndrome described by Britt and Waxler (1964). Because these syndromes are produced by serotypically different organisms and are clinically dis- tinct, it is misleading to combine the 2 clinicopathological syndromes in textbook descriptions of "colibacillosis." HistOpathologic lesions Smith and Jones (1963) have reported that there were no histologic indications of any inflammatory response in the intestinal tract of pigs with colibacillosis. From the histologic studies made with E. 32;; serotypes Ol38:K81 (Minn.) and Ol38:K81:NM (Mich.) and described above, it is clear that there are inflammatory changes associated with clinical colibacillosis but that these changes might be insufficiently marked to be readily apparent in tissues taken from field cases of the disease. There are apparently no lesions pathognomonic of the infection. Occasional foci of neutrophils were seen in the parenchymatous organs of these experimentally infected pigs, and inflammatory exu- dates were observed in sections from the lamina propria and 113 submucosa of the intestinal tract. Endotoxins have long been known to induce neutrophilia (Carpenter, 1965) and it may be that, as organisms multiply and "spill over" from various sites of phagocytosis, they or their endotoxins can produce transient microabscesses in these tissues. Bacterial exotoxins are potent, relatively unstable compounds with highly specific toxic effects (often suffi— ciently characteristic to identify the toxin) and are gen- erally highly antigenic. Endotoxins derived from the Enterobacteriaceae are weakly toxic, relatively stable prod- ucts of bacterial destruction. They are weakly antigenic and produce nonspecific toxic effects such as pyrexia, pros— tration, diarrhea, vascular malfunction (vasomotor shock, edema and hemorrhage) and at first hyperglycemia, then later, hypoglycemia. These so-called classical responses of host animals to endotoxins include altered response to epinephrine, and tissue necrosis at sites of endotoxin con- centration (Smith, Conant and Overman, 1964). (i) Histopathologic lesions in tissues other than the alimentary tract. The basic general histopathologic find- ings were specific epithelial degenerations in various organs and vascular changes indicated by edema. The most common epithelial change was hydrOpic de- generation, and this change was observed in bronchiolar epithelium, aortic endothelium, the gastrointestinal tract, and the transitional epithelium of the urocyst. These 114 changes were inconsistent and were possibly associated with incidental local concentrations of endotoxic, or possibly exotoxic activity of the test organism. There is no evidence that nonlethal levels of bacter- ial endotoxins have a generalized effect on the integrity of the capillary walls, but Zweifach ep_3l. (1956) and Zweifach (1958) have demonstrated vasoconstriction at the venular end of the capillary bed of laboratory mammals fol- lowing the application of endotoxin extracted from E. ggii. "Nontoxic" doses of this endotoxin produced a slowing of blood flow through the microcirculation, but there was no change in the systemic blood pressure during this reaction. With increased levels of endotoxin there was a brief hyper- reactive vascular phase as before, then a hyporeactive phase in which damage to the capillary and venular barrier per se was present. A continuing exposure to endotoxin led to pro- tracted derangement of vascular reactivity, then capillary hemorrhage and stasis (Zweifach, 1964). In this experiment, instances of these vascular changes were seen in some of the injection site sections, and the necrosis and hemorrhage observed in the intestine of l in- fected pig (J4669) might have been the result of either a massive release of endotoxin or high levels of endotoxin production from prolific growth of the coliforms in the in- testine. 115 It is possible that the general signs of edema so com- monly observed in tissues from these monocontaminants were a less severe manifestation of E. 22;; endotoxic activity on the vasculature. Dunne (1964) has associated the edema of the mesocolon seen in baby pigs with colibacillosis. It is apparent that edema of the colon may in fact be an indi- cation of the disease, but mild edema of this tissue is non- significant under the conditions of these experiments in that it was commonly found in the control animals. (ii) Histopathology of the alimentary tract. The changes in the epithelium and the lamina propria of the villi recorded as being common to both infected and germ- free pigs are described in detail in the results because the difference between the 2 groups is distinctly one of degree. It is apparent that in the process of absorption of fluids from the lumen of the intestine some of the histo- logic changes observed in the experimental pigs are likely to be found regardless of the presence or absence of E. 2211. (Figures 21 and 22). It may be that the presence of either endotoxin and/or exotoxin of E. gel; so modifies these changes that they become significant in a pathologic sense (Figure 27). It is suggested that the distinct hyperemic, hemor- rhagic, necrotic and inflammatory changes found only in the infected pigs are manifestations of E. coli endotoxins 116 whereas the less marked pathologic changes which probably account for the major clinical sign of diarrhea are pos- sibly due to the combined effects of exotoxin and-endotoxin and may even resemble to a degree some of the physiologic changes that one can see in the digestive tract of germfree pigs (Figure 27). In absorption from the intestinal tract the columnar epithelium actively transports fluid and metabolites from the lumen of the intestine into the lymphatics, lacteals and vessels of the lamina propria. A number of mechanisms can act to remove these intravillous fluids from the mucosa. One mechanism is the periodic rhythmic vertical contraction and horizontal swaying motion of the villi due to the con- tractions of the Breucke fibers. The vertical contraction has a pumping effect and occurs possibly at a level of vil- lous distention that stretches the Breucke fibers and stimu- lates them to contract. Evidence of this vertical contrac- tion activity of the villus can be readily seen in tissues from both the control pigs and those infected pigs which are not yet heavily colonized in the intestinal tract with E. ggli. These contracted villi are seen as being shortened with localized thickening of the lamina propria and an "oak- 1eaf" outline to the profile of the villus (Figures 27 and 29). With these possibilities in mind a number of hypothe- tical explanations for the intestinal tract changes observed .in these experiments present themselves. 117 The presence in a crypt or on the surface of a villus or within the lamina propria of even a small colony of E. 29;; elaborating both endotoxin and exotoxin, could produce sufficient toxin to inhibit or disrupt the cellular metabo- lism of the columnar epithelium of the villus, and perhaps neighboring villi. The expression of this toxic effect as hydrOpic degeneration or as marked distention and vacuola- tion of the columnar epithelium may and probably will result in malfunction or complete cessation of the active transport system of fluids from the lumen of the intestine into the villus. One could postulate this happening in a very re- stricted locale, or in certain segments of the small intes- tine, or even throughout the entire heavily colonized intestinal tract. This might serve to explain the presence of an obviously degenerating inactive villus together with an actively contracting villus in the same tissue section. It might also explain the presence of signs of both physi- ologic activity (expressed in active vertical villus con- tractions and active hyperemic vessels in the lamina propria) and pathologic changes such as distention and vacuolation of some or all of the epithelial cells in a single villus. Should the toxin be of sufficient concentration in the intestinal tract to produce necrosis of the epithelial surface, the sloughing of the necrotic epithelium, combined with hemorrhage and body fluid loss from exPosure of the lamina prOpria to the hypertonic lumen contents, will produce 118 a clinical syndrome in the pig that will vary according to the extent of the lesions. One could postulate first a stimulatory then an in- hibitory effect of E. 22;; endotoxin on the function of the Breucke fibers. According to the relative concentrations of the E. 99E; in certain regions of the lumen, it is pos- sible to visualize the presence in a tissue section of villi showing marked vertical contraction activity whereas else- where in the section one might see villi distended with fluids and unable to contract and eject this intravillous fluid because of toxic inactivation of the Breucke fibers. It is reasonable to assume that the lamina propria of such villi will absorb fluid to the point where the distention of the villus will prevent any further absorption of fluid. Indeed, it was observed in sections from the intes- tines of some of the infected pigs that distention of the villi reached a point at which the epithelial surfaces of neighboring villi were closely apposed thereby removing from function an extremely high.proportion of a potentially active absorptive surface (Figures 33 and 34). Occasionally, tissue sections were seen in which there was both widespread hydropic change in the epithelium, and in which the villi were elongated and the Breucke fibers not contracted (Figure 36). There was little or no evidence of absorption of fluids in these tissue sections, and pre- sumably the hydropic change in the epithelium had preceded 119 the Opportunity for the villi to absorb fluid from the in- testinal lumen. The lack of contraction of the Breucke fibers could be due either to the lack of stimulus for smooth muscle contraction because of a lack of distention of the villus with absorbed fluids, or the flaccid paraly- F”“x sis of the smooth muscle fibers due the the presence of high levels of bacterial endotoxin in the locale. Zweifach (1964), reviewing §°.2211 endotoxin activity, i records that it does not appear to have a direct cytotoxic :3 action, but he does present evidence of a direct effect on the smooth muscle of the vasculature. The pattern of the effect of increasing levels of endotoxin is first, an in- creased rate, then at higher levels of endotoxin, a decreased rate of sequential reaction of the vasculature due to endo- toxic sensitization of the muscular vessels to exogenous and endogenous stimuli. There is no overt evidence of damage to the endothelium of the capillaries and venules. There is increased capillary permeability, and E. gel; endotoxin--a lipopolysaccharide--does produce stasis of the capillary bed by a mechanism of narrowing of the lumina of the draining venules (Zweifach, 1964). This evidence can be applied to the hypothesis that the hyperemia, edema, hemorrhage, necrosis and desquamation observed in the in- testine of the infected pigs are direct and/or indirect results of vascular responses to the presence of E. coli endotoxin and its absorption either from the lumen or from the site of colonization. 120 It is possible, therefore, with a hypothesis such as this in mind, to reconcile some of the mass of confusing impressions gained from examination of the tissue sections from a number of sites of many pigs killed at various stages of the progress of colibacillosis. It is suggested that colibacillosis, as it was observed in these experiments, is an expression of a malabsorption syndrome. The intestinal mucosa has become ineffective as an absorbing organ because of a number of anatomic and func- tional changes--change in the epithelium rendering it in- capable of the passage of fluid materials through its walls; malfunction of the villus-draining mechanisms preventing movement of fluid from the intravillous site in the mucosa; reduction of effective absorptive surface area by swelling and compression of the villi; and vascular malfunction leading to hypofunction or anoxia and necrosis of the intes- tinal mucosa. The small intestine is a highly differentiated organ and has only a limited number of ways to react to stress. It is likely that malabsorption syndromes will result from a number of mechanisms, and we can safely assume that a com- plex organism like E. ggli_will affect at least a number of these mechanisms. S UMMARY Research was conducted using a total of 62 gnotobiotic pigs in 3 experiments to determine the clinical effects, pathogenesis and lesions produced by 2 closely related sero- types of Escherichia coli--Ol38:K81:NM (Mich.) and Ol38:K81‘ (Minn.). It was determined that both serotypes readily colonized the intestinal tract from experimental infection simulating neonatal contamination of the umbilical stump. Oral exposure was precluded. A bacteremia was produced within 1 1/2 hours of injec- tion, and by 24 hours the infection was generally established in the mucosa of the gastrointestinal tract. By 48 hours postinjection the bacteremia had subsided so that only an occasional isolation from organs other than the gastrointes- tinal tract was made. The clinical signs of profuse watery diarrhea, dehy- dration and erect hair coat characteristic of colibacillosis in the neonatal pig were not apparent until the organism be- came well established in the gastrointestinal tract. The clinical signs of infection with E. 22;; 0138:K81 (Minn.) were significantly milder but edema of the subcutis was a constant feature of infection with this serotype. 121 122 Oral exposure of l germfree litter to E. ggli 0138: K81:NM (Mich.) produced both heavy colonization of the en- tire intestinal tract and diarrhea within 4 hours. Inter- mittent bacteremia was evident in this litter. No definite trends in body temperature or hematologic findings were observed, although in some animals there was a marked re- duction in leukocyte counts. The gross necrOpsy findings, in general, agreed with those reported in the literature from field cases of coli- bacillosis in neonatal pigs. Histologically the lesions were predominantly in the gastrointestinal tract and ranged from a acute hemorrhago- necrotic enteritis to a histological picture, in persistently scouring pigs, microscopically indistinguishable from that seen in clinically normal germfree animals. An hypothesis of the pathogenesis of colibacillosis based on the histologic evidence is presented. REFERENCES Armed Forces Institute of Pathology: Manual of Histologic and Special Staining Techniques. Washington, D. C. Bennett, P. C.: Edema disease. In Diseases of Swine. 2nd ed., H. W. Dunne, ed., The Iowa State Univ. Press, Ames, Iowa (1964). Benjamin, M. E.: Outline of Veterinary Clinical Pathology. The Iowa State University Press, Ames, Iowa (1961). Blood, D. C. and Henderson, J. A.: Veterinary Medicine, 2nd ed., Williams and Wilkins Co., Baltimore, Md. (1963). Britt, A. L. and Waxler, G. L.: Proceedings, Gnotobiotic Symposium and Workshop, Michigan State University, East Lansing, Mich., June 7 (1964). Brooks, R. A.: Serum proteins and antibodies in the germ- free and gnotobiotic pig. M.S. Thesis, Michigan State University (1966). : Buxton, A. and Thomlinson, J. R.: The detection of tissue- sensitizing antibodies to Escherichia coli in oedema disease, haemorrhagic gastroenteritis and in normal pigs. Res. Vet. Sci., 2, (1961): 73-88. Calhoun, M. L. and Smith, E. M.: Hematology and hematopoie- tic organs. In Diseases of Swine. 2nd ed., H. W. Dunnenu1,The Iowa State Univ. Press, Ames, Iowa (1964). Carle, B. N. and Dewhirst, W. H.: A method for bleeding swine. J.A.V.M.A., 101, (1942): 495-496. Carpenter, P. L.: Immunology and Serology, 2nd ed. W. B. Saunders Co., Philadelphia, Pa. (1965). Dunne, H. W.: Streptococcosis, colibacillosis and bordetel- 10515. In Diseases of Swine, 2nd ed., H. W. Dunne, ed., The Iowa State Univ. Press, Ames, Iowa (1964). 123 124 Edwards, P. R. and Ewing, N. H.: Identification of Entero- bacteriaceae, 3rd ed., Burgess Pub. Co., Minneapolis, Minn. (1962). Ferguson, W. W.: Experimental diarrheal disease of human volunteers due to Escherichig coli. Ann. N. Y. Acad. Sci., 66, (1956): 71-77. Fey, H. and Margadant, A.: Zur Pathogenese der Kalber - Colisepsis. I. Verteilung des Sepsistyps in den Organen. Zbl. Bakt. I., 182, (1961): 71-79. Fey, H.: Neuere Untersuchungen fiber die Colisepsis des Kalbes. Schweiz. Arch. Tierheilk., 104, (1962): 1712. Fey, H., Lanz, E., Margadant, A. and Nicolet, J.: Zur Pathogenese der Kalbercolisepsis. Dtsch. tierartZl. Wschr., 69, (1962): 581-586. Gay, C. G.: Escherichia coli and neonatal disease of calves. Bacteriol. Rev., 29, (1965): 75-101. Glasser, K., Hupka, E. and Wetzel, R.: Die Krankheiten des Schweines, 5th ed. M. and H. Schaper, Hannover, (1950). Gordon, W. A. M. and Luke, D.: Gastroenteritis in young piglets associated with Escherichia coli. Vet. Rec., Gossling, J. and Rhoades, H. E.: Serologic types of Escherichia coli isolated from certain pigs with enteric disorders. Cornell Vet., 56, (1956): 344- 353. Hopps, C. H.: Bacterial diseases. In Pathology: 5th ed., Vol. 1, W. A. D. Anderson, ed. The C. V. Mosby Co., St. Louis, Mo. (1966). Jennings, A. R.: Gastroenteritis in the pig. Vet. Rec., 71 (1959): 766-771. Jensen, C. O.: (1889) C. O. Jensen's Selected Papers 1886- 1908. Ejnar Munksgaard, Copenhagen, 1, (1948): 290, cited by Sojka, W. J., (1956). Jubb, K. V. F. and Kennedy, P. C.: Pathology of Domestic Animals, Vol. 2. Academic Press, New York (1963). Kauffman, F.: Zur Serologie der Coli-Gruppe. Acta Pathol. Microbiol. Scand., 21, (1944): 20-45. 125 Kauffman, F.: The serology of the coli group. J. Immunol., 57, (1947): 71-100. Kauffman, F. and Dupont, A.: Escherichia strains from in- fantile epidemic gastroenteritis. Acta Pathol. Microbiol. Scand., 27, (1950): 552-564. Kenworthy, R. and Allen, D.: Influence of diet and bacteria on small intestinal morphology with special reference to early weaning and Escherichia coli. Studies with germfree and gnotobiotic pigs. J. Comp. Path., 76, (1966); 291-296. Kinnaird, P. J.: Progress of HyPAR pigs under commercial conditions at Dookie College. Vict. Vet. Proc., 22, (1964): 37-38. Kohler, E. M. and Bohl, E. H.: Studies of Escherichia coli in gnotobiotic pigs. I. Experimental reproduction of colibacillosis. Can. J. Comp. Med. Vet. Sci., 30, (1966): 199-203. McBryde, C. N.: Acute enteritis in young pigs due to infec- tion with the colon group. J.A.V.M.A., 84, (1934): 36-50. Neter, E. and Shumway, C. N.: Escherichia coli serotype D 433: Occurrence in intestinal and respiratory tracts, cultural characteristics, pathogenicity, sensitivity to antibiotics. Proc. Soc. Exp. Biol. Med., 75, (1950): 504-407. Philip, J. R. and Shone, D. K.: Some observations on oedema disease and a possibly related condition of pigs in Southern Rhodesia. J. So. Afr. Vet. Med. Assn., 31, (1960): 427-434. Ribi, E., Anacker, R., Fukushi, K., Haskins, W. T., Landy, M. and Milner, K. C.: Relationship of chemical com- position to biological activity. In Bacterial Endo- toxins. M. Landy and W. Brown, ed. Institute of M1croEiologYI Rutgers, The State University, New Brunswick, N. J. (1964). Robbins, S. L.: Textbook of Pathology with Clinical Appli- cation, 2nd ed. W. B. Saunders Co., Philadelphia, Pa. (1962). Saunders, C. N., Stevens, A. J., Spence, J. B. and Sojka, W. J.: Escherichia coli infection in piglets. Res. Vet. Sci., 1, (1960): 28-35. 126 Saunders, C. N., Stevens, A. J., Spence, J. B. and Sojka, W. J.: Escherichia coli infection: Reproduction of the disease in naturally reared piglets. Res. Vet. Sci., 4, (1964): 333-346. Saunders, C. N., Stevens, A. J., Spence, J. B. and Betts, A. O.: Escherichia coli infection: Reproduction of the disease in "Pathogen-Free" piglets. Res. Vet. Sci., 4, (1963): 347-357. Smith, D. T., Conant N. F. and Overman, J. R.: Microbiology, 13th ed., Appleton-Century-Crofts, New York (1964). Smith, H. W. and Crabb, W. E.: The typing of Escherichia coli by bacteriophage, its application to the study of the E. coli population of the intestinal tract of healthy calves and of calves suffering from white scours. J. Gen. Microbiol., 15, (1956): 556- 574. Smith, H. W. and Jones, J. E. T.: Observations on the ali- mentary tract and its bacterial flora in healthy and diseased pigs. J. Path. Bact., 86, (1963): 387-412. Sojka, W. J.: Escherichia coli in Domestic Animals and Poultry. Commonwealth Agricultural Bureaux, Farn- ham Royal, Bucks, England, (1965). Stevens, A. J.: A Laboratory Handbook of Veterinary Diag- nostic Procedures. Animal Health Branch Monograph No. 3., F.A.O., Rome (1961). Stevens, A. J.: Coliform infections in the young pig and a practical approach to the control of enteritis. Vet. Rec., 75, (1963): 1241-1246. Suter, E.: Interaction between phagocytes and pathogenic microorganisms. Bact. Rev., 20, (1956): 94-132. Thomlinson, J. R. and Buxton, A.: A comparison of experi- mental anaphylactic shock in guinea pigs with natur- ally occurring oedema disease and haemorrhagic gas- troenteritis in pigs. Res. Vet. Sci., 3, (1962): 186—202. Thomlinson, J. R. and Buxton, A.: Anaphylaxis in pigs and its relationship to the pathogenesis of oedema disease and gastroenteritis associated with Escheri- chia coli. Immunology, 6, (1963): 126-139. 127 Thomlinson, J. R.: Observations on the pathogenesis of gas- troenteritis associated with Escherichia coli. Vet. Rec., 75, (1963): 1246-1250. Waxler, G. L., Schmidt, D. A., and Whitehair, C. K.: Tech- nique for rearing gnotobiotic pigs. Am. J. Vet. Res., 27, (1966): 300-307. Wood, W. B., Smith, M. R. and Watson, R.: Studies on the mechanism of recovery in pneumococcal pneumonia. IV. The mechanism of phagocytosis in the absence of antibody. J. Exper. Med., 84, (1946): 387-402. Zweifach, B. W., Nagler, A. L., and Thomas, L: The role of epinephrine in the reactions produced by the endo- toxins of gram negative bacteria. J. Exptl. Med., 104, (1956): 881-896. Zweifach, B. W.: Microcirculatory derangements as a basis for the lethal manifestations of experimental shock. Brit. J. Anaesthesia, 30, (1958): 466-484. Zweifach, B. W.: Vascular effects of bacterial endotoxin. In Bacterial Endotoxins. M. Landy and W. Brown, ed., Institute of MiCrobiologYI Rutgers, The State University, New Brunswick, N. J. (1964). VI TA The author was born at Colac, Victoria, Australia, on September 22, 1932. His primary education was taken at Meeniyan and Hamilton, Victoria, and his secondary education completed at Ballarat College, Victoria, in 1950. The author attended the University of Melbourne from 1951 to 1952 and the University of Sydney from 1952 to 1956. He received the degree of Bachelor of Veterinary Science in 1957. From 1957 to 1961 he served as District Veterinary Officer in the Livestock Division of the Department of Ag- riculture, Victoria, and was Lecturer in Veterinary Science at Longerenong Agricultural College, Dooen, Victoria. In 1961 the author was appointed Veterinary Research Officer at the Cameron Laboratory, Animal Husbandry Research Centre, Werribee. He came to Michigan State University as an assistant instructor and graduate student in the Depart— ment of Pathology in 1964. 128 II I