Downloaded from https://academic.oup.

com/jid/article-abstract/184/6/803/847412 by Universidad de La Salle user on 31 July 2020

Enteropathogenic Escherichia coli and Ulcerative Colitis
in Cotton-Top Tamarins (Saguinus oedipus)
Keith G. Mansfield, Kuei-Chin Lin, Dongling Xia, Joseph V. Newman, David B.
Schauer, John MacKey, Andrew A. Lackner, Angela Carville
 803

CONCISE COMMUNICATION

Enteropathogenic Escherichia coli and Ulcerative Colitis in Cotton-Top

Tamarins (Saguinus oedipus)
Keith G. Mansfield,1 Kuei-Chin Lin,1 Dongling Xia,1

Downloaded from https://academic.oup.com/jid/article-abstract/184/6/803/847412 by Universidad de La Salle user on 31 July 2020

1
New England Regional Primate Research Center, Harvard Medical
Joseph V. Newman,2 David B. Schauer,2 John MacKey,1 School, and 2Division of Comparative Medicine, Massachusetts
Institute of Technology, Cambridge
Andrew A. Lackner,1 and Angela Carville1

The cotton-top tamarin (CTT; Saguinus oedipus) is an endangered New World primate that
develops a highly prevalent idiopathic colitis resembling human ulcerative colitis. This study
found that enteropathogenic Escherichia coli (EPEC) caused acute colitis in CTTs, which was
associated with ulcerative colitis. EPEC clinical isolates revealed localized adherence patterns
by HEp-2 assay and were devoid of Shiga-toxin production. Sequencing of the eae gene
(GenBank accession no. AF319597) revealed 99.2% identity to sequences of human isolates
(GenBank AF116899) and corresponded to the e intimin gene subtype. Detection of intimin
sequences by polymerase chain reaction on primary fecal cultures indicated widespread EPEC
infection in the CTT colony. Prospective analysis revealed that animals with fecal cultures
positive for intimin sequences had a higher frequency of active colitis (75.0% vs. 27.2%; P !
.005, x2 test) and higher histological scores of colonic inflammation (0.875 vs. 0.455, respec-
tively; P ! .05, Mann-Whitney rank sum test).

The cotton-top tamarin (CTT; Saguinus oedipus) is an en-
dangered primate native to the forests of northern Colombia.
A highly prevalent idiopathic colitis resembling human ulcer-
ative colitis has been recognized in CTTs housed at the New
England Regional Primate Research Center (NERPRC) and
elsewhere since the 1970s. This idiopathic colitis shares similar
clinical, endoscopic, and histopathological features with ulcer-
ative colitis in humans and has been widely used as a nonhuman
primate model to investigate immunopathogenesis, risk factors,
and novel therapeutic interventions for the human condition
[1]. CTTs with chronic colitis have a high incidence of colonic
carcinoma. Although the etiology of CTT colitis and colonic
carcinoma is unknown, several lines of evidence suggest that
environmental factors play a role: A prospective study of the
epizootiology of colitis and adenocarcinoma showed that
chronic mucosal changes were modified by diet [2]. Further-
more, the marked decreased incidence of both colitis and car-
cinoma of animals raised in isolation from the rest of the colony
suggested the existence of an infectious agent as a cofactor in
Received 2 April 2001; revised 24 May 2001; electronically published 8
August 2001.
Animals were housed at the New England Regional Primate Research
Center in accordance with standards of the American Association for Ac-
creditation of Laboratory Animal Care and of the Harvard Medical School’s
Animal Care and Use Committee.
Financial support: National Institutes of Health (RR-00168).
Reprints or correspondence: Dr. Keith G. Mansfield, Harvard Medical
School, New England Regional Primate Research Center, PO Box 9102,
Southborough, MA 01772-9012 (Keith_Mansfield@HMS.Harvard.edu).
The Journal of Infectious Diseases 2001; 184:803–7
䉷 2001 by the Infectious Diseases Society of America. All rights reserved.
0022-1899/2001/18406-0021$02.00
disease [2]. Circumstantial evidence has implicated Campylo-
bacter jejuni, corona-like viruses, and Helicobacter species as
agents that may play a role in producing this disease syndrome
[2, 3].
An infectious etiology has also been suggested as a cofactor
of ulcerative colitis in humans, and enteroadherent Escherichia
coli has been implicated in the pathogenesis of this disease. The
adhesive properties of E. coli isolates from patients with ulcer-
ative colitis differ from those of isolates from healthy control
patients, and intimin-positive strains have been identified in
some affected persons [4, 5]. Furthermore, treatment of patients
with ulcerative colitis by means of probiotic nonpathogenic E.
coli has an effect similar to that of mesalazine in maintaining
remission [6]. Here we describe enteropathogenic E. coli (EPEC)
infection in CTTs and its association with ulcerative colitis in
this species.
Methods
Animals and housing. CTTs were housed at the NERPRC. All
CTTs were housed in pairs or grouped into small extended families
consisting of 1 adult breeding male CTT, 1 adult breeding female
CTT, and 2–6 siblings !18 months old, as described elsewhere [7].
Average colony census during the study period was ∼200 CTTs.
Bacteriology. Rectal swabs of clinically affected animals were
cultured by standard techniques. Samples were plated on Mac-
Conkey, Hektoen, and blood agar at 37⬚C in 5% CO2 and on
Campylobacter agar at 42⬚C with 5% O2, 85% N, and 10% CO2.
Plates were evaluated at 24, 48, and 72 h, and individual bacterial
colonies were identified by standard biochemical techniques (API
Rapid 20E; BioMerieux Vitek). A HEp-2 adherence assay was done
as described elsewhere [8]. Bacterial isolates were serotyped at the
804 Mansfield et al.
Pennsylvania State University E. coli Reference Laboratory (Uni-
versity Park, PA). Shiga-toxin production was assessed by Vero cell
cytotoxicity assay, as described elsewhere [9].
Polymerase chain reaction (PCR) and sequencing. For char-
acterization of individual bacterial clones, DNA was isolated from
single colonies and grown overnight in broth. PCR utilizing primers
directed at the intimin (eae), Shiga-like toxin (stx1 and stx2), and
hemolysin (hlyA) genes was done as described elsewhere [10]. For
detection of bacteria-specific virulence sequences in primary fecal
cultures, rectal swabs or colonic tissue were placed in Luria broth
(Difco Laboratories) and grown overnight. DNA was then isolated
from cell pellets. The intimin gene was amplified from isolates ob-
tained from 2 animals by use of primers SK1 and LP5 [11]. The
PCR products were cloned (Invitrogen) and sequenced utilizing an
ABI automated sequencer (Perkin-Elmer Cetus) and forward/re-
verse primers. The bfpA gene was amplified from EPEC isolates,
using primers Donne-28 and Donne-29, and sequenced as described
elsewhere [12].
Colonic biopsy specimens and histopathology. Colonic biopsy
specimens were collected, stained with hematoxylin-eosin, and ex-
amined in a blinded fashion for acute colitis and evidence of struc-
tural alterations associated with chronicity, as described elsewhere
[2]. For statistical work, we used a commercial software package
(Jandel Scientific). Groups were compared by the x2 test of con-
tingency or the Fisher’s exact and Mann-Whitney rank sum tests,
as appropriate.
Results
Five CTTs presented with acute onset of profuse diarrhea.
Affected animals became anorexic and inactive and developed
clinically recognized dehydration. Laboratory results revealed
marked neutrophilic leukocytosis, metabolic acidosis, and hy-
pokalemia. Hyperglycemia and hypoalbuminemia were less fre-
quent. Despite the apparent blood loss in feces, animals did
not develop significant anemia. Animals received supportive
therapy and were treated with enrofloxacin (2 mg/kg once a
day intramuscularly) and lactated Ringer’s solution. Despite
therapy, 4 of 5 animals died or were euthanized after a disease
course of 2–10 days.
Biopsy specimens and/or tissue obtained at necropsy were
evaluated from all clinically affected animals, to assess mor-
phological alterations associated with colitis. Findings revealed
features diagnostic of an attaching and effacing lesion. Colonic
epithelium appeared rounded or irregular, giving the surface a
scalloped or cobblestone appearance (figure 1). Bacilli were
intimately associated with the apical cytoplasmic membrane,
and individual epithelial cells often exfoliated with variable cy-
toplasmic vacuolization. Surface changes were accompanied by
marked crypt cell hyperplasia characterized by loss of goblet
cells, cytoplasmic basophilia, increased mitotic rate, and loss
of nuclear polarity. Crypt abscesses were visible in most sec-
tions, and neutrophilic infiltrates were often present within the
lamina propria. Ultrastructurally, there was effacement of nor-
mal microvillus architecture, and adherent bacilli were attached
JID 2001;184 (15 September)
to the apical cytoplasmic membrane with pedestal formation
and rearrangement of the underlying cytoskeleton. Morpho-
logical alterations were limited to the large intestine, which was
diffusely involved in all cases. In contrast, the small intestine
was spared and was normal in all animals.
Rectal swabs and fecal samples were obtained from all clin-
Downloaded from https://academic.oup.com/jid/article-abstract/184/6/803/847412 by Universidad de La Salle user on 31 July 2020
ical cases, for bacterial isolation and identification, and were
negative for common enteric pathogens of nonhuman primates,
including Salmonella, Shigella, and Campylobacter species. Rec-
tal swabs plated on MacConkey agar produced almost pure
growth of E. coli from all animals. Multiple isolates tested by
HEp-2 adherence assay indicated a localized adherence pattern
characteristic of EPEC. DNA was isolated from clones from
each animal, and the presence of the eae gene was confirmed
by PCR amplification of a 384-bp product. A 2.6-kb fragment
was amplified from 2 animals and sequenced (GenBank acces-
sion no. AF319597). A BLAST similarity search revealed 99.2%
identity to sequences obtained from human and rhesus ma-
caque isolates (GenBank AF116899 and AF301015, respec-
tively) that correspond to the e intimin gene subtype [11, 13].
Similarly, the bfpA gene was detected, by PCR, in EPEC iso-
lates, and sequencing revealed 98.4% identity to bfpA sequences
obtained from human EPEC isolates (GenBank AF382948 and
AF304484, respectively). PCR performed on isolates was neg-
ative for the detection of stx1, stx2, and hlyA sequences. Fur-
thermore, no Shiga-toxin production was demonstrated by the
Hela cell cytotoxicity test. Bacterial isolates positive for the eae
gene and demonstrating the localized adherence pattern were
serotyped and identified as O26:HNM.
To investigate the relationship between EPEC infection and
colitis in the CTT colony, we did a prospective analysis to
determine the incidence of EPEC infection and its relationship
to ulcerative colitis. Paired fecal and colonic biopsy specimens
were obtained from colony animals (n p 38 ) 2 weeks after reso-
lution of the last clinical case and were evaluated for the pres-
ence of intimin sequences by PCR and of inflammatory infil-
trates by histopathology. Of 38 animals, 16 (42%) were PCR
positive for intimin sequences in primary stool cultures. PCR
performed on fecal cultures was negative for the detection of
stx1, stx2, and hlyA sequences.
Twelve of the 16 animals positive, by PCR, for intimin se-
quences had histological evidence of active colitis, compared
with 6 (27.2%) of 22 that were PCR negative (P ! .005, x2 test).
Mean scores for active colitis were significantly higher for ani-
mals with fecal cultures positive for intimin sequences, com-
pared with those for animals with negative cultures (mean,
0.875 vs. 0.455, respectively; P ! .05 , Mann-Whitney rank sum
test). No significant differences between animals with positive
and those with negative fecal cultures were noted for chronicity
scores (mean, 0.875 vs. 0.864, respectively) or age (8.5 and 9.5
years, respectively). An attaching and effacing lesion was iden-
tified morphologically in only 1 of the 12 animals with active
colitis that had fecal cultures positive for intimin sequences. In
JID 2001;184 (15 September) Enteropathogenic E. coli in Tamarins 805

Downloaded from https://academic.oup.com/jid/article-abstract/184/6/803/847412 by Universidad de La Salle user on 31 July 2020

Figure 1. Enteropathogenic Escherichia coli (EPEC) infection of colonic epithelium in cotton-top tamarins. A, Irregular colonic surface epi-
thelium accompanied by marked epithelial crypt cell hyperplasia (hematoxylin-eosin stain; magnification, ⫻85). Inset, Bacilli intimately associated
with apical cytoplasmic membrane (toludene blue stain; magnification, ⫻428). B, Ultrastructural features of EPEC infection. Adherent bacilli
with effacement of normal microvillus border (magnification, ⫻5000). Inset, Pedestal formation with rearrangement of underlying cytoskeleton
(magnification, ⫻13,000).

this animal, the lesion was present multifocally in 1 of 3 colonic
sections examined. In the remaining intimin-positive animals,
inflammatory changes did not differ from those typically at-
tributed to idiopathic ulcerative colitis in CTTs.
The colony was retested for the presence of EPEC 7 months
after the first survey. Of 35 animals tested, intimin sequences
were detected by PCR on DNA isolated from primary fecal
cultures from 12 (34.3%) animals. Despite the absence of severe
acute colitis, the incidence of infection was not statistically dif-
ferent from that 7 months earlier.
806 Mansfield et al.
Discussion
Here we describe EPEC infection as a cause of severe acute
colitis in CTTs. The clinical disease was characterized by pro-
fuse and sometimes hemorrhagic diarrhea and was often ac-
companied by neutrophilic leukocytosis, metabolic acidosis,
and hypokalemia. Animals with acute disease had morpholog-
ical findings characteristic of EPEC infection in other species,
including the diagnostic attaching and effacing lesion. Defini-
tive diagnosis was obtained by histological evaluation and was
supported through bacteriologic culture and by use of PCR
and the HEp-2 adhesion assay. Sequencing of a 2.6-kb fragment
of the intimin gene revealed 99.2% identity to sequences ob-
tained from human EPEC isolates and corresponded to the e
subtype. This subtype was previously associated with entero-
hemorrhagic E. coli (EHEC) isolates obtained from cattle and
humans and from rhesus macaques with AIDS [11, 13]. There
was no evidence that isolates from CTTs produced Shiga toxin,
as determined by cell cytotoxicity assays or PCR, nor were
histomorphological findings at death compatible with hemo-
lytic uremic syndrome. PCR also identified bfpA sequences con-
sistent with the observed localized adherence phenotype.
Molecular techniques showed widespread EPEC infection in
the rest of the CTT colony. Routine cultures for common bac-
terial pathogens in nonhuman primates will likely underesti-
mate the true incidence of EPEC infection, and specialized tech-
niques must be pursued for definitive diagnosis. E. coli is the
predominant facultative anaerobe found within the colon, and
most clinical laboratories do not distinguish between species of
lactose fermentors or diarrheagenic from nondiarrheagenic E.
coli. Furthermore, differentiation requires specialized tech-
niques such as PCR and adhesion assays that may not be read-
ily available in most clinical laboratories. Although colonic
biopsy was reliable for the identification of animals with clinical
disease, PCR performed on DNA isolated from primary fecal
cultures was required to identify enzootic infection in the CTT
colony.
As in humans with ulcerative colitis, the etiology of colitis
in CTTs is unknown, and a variety of factors have been im-
plicated in producing the full clinical syndrome, including diet,
environmental stress, and genetic influences [7, 14]. Wild CTTs
in their natural environment have a markedly decreased inci-
dence of colitis, epithelial dysplasia, and overt carcinoma, com-
pared with captive animals, indicating that environmental fac-
tors initiate or promote disease development [14]. It has also
been suggested that infectious agents may play a role in the
clinical manifestation of disease. Corona-like viruses and Cam-
pylobacter and Helicobacter species have been suggested as po-
tential pathogens that may initiate or influence disease course
[2, 3].
EPEC infection in the CTT colony was a common enzootic
infection, and the incidence was essentially unchanged over a
7-month observation period. Although animals with clinically
unrecognized EPEC infection had a higher incidence of colitis
and higher histological scores of colitic activity, whether EPEC
JID 2001;184 (15 September)
infection plays an initiating or an etiologic role in producing
colitis remains undetermined. For a variety of reasons, animals
with preexisting colitis have altered mucosal defenses and may
be prone to colonization by potentially pathogenic organisms.
Such changes may include alterations in lipid metabolism, mu-
cosal architecture, microbial flora, epithelial maturity, and cell
Downloaded from https://academic.oup.com/jid/article-abstract/184/6/803/847412 by Universidad de La Salle user on 31 July 2020
surface receptor expression. Of particular interest may be
changes in phospholipid turnover found in inflamed mucosa,
leading to increased phosphatidylethanolamine levels. Expres-
sion of phosphatidylethanolamine was previously correlated
with binding of EHEC and EPEC strains [15].
Although animals with preexisting colitis may be more read-
ily colonized by potential pathogens, EPEC may potentiate
chronic inflammatory changes through alterations in mucosal
permeability and activation of proinflammatory signaling path-
ways. EPEC infection of CTTs may provide an experimental
system with which to investigate the role of adherent bacteria
in the pathogenesis of ulcerative colitis in humans and an op-
portunity to examine whether infection potentiates mucosal
injury.
References
1. Watkins PE, Warren BF, Stephens S, Ward P, Foulkes R. Treatment of ul-
cerative colitis in the cottontop tamarin using antibody to tumour necrosis
factor alpha. Gut 1997; 40:628–33.
2. Johnson LD, Ausman LM, Sehgal PK, King NW. A prospective study of
the epidemiology of colitis and colon cancer in cotton-topped tamarins
(Saguinus oedipus). Gastroenterology 1996; 110:102–15.
3. Saunders KE, Shen Z, Dewhirst FE, Paster BJ, Dangler CA, Fox JG. Novel
intestinal Helicobacter species isolated from cotton-top tamarins (Saguinus
oedipus) with chronic colitis. J Clin Microbiol 1999; 37:146–51.
4. Giaffer MH, Holdsworth CD, Duerden BI. Virulence properties of Escher-
ichia coli strains isolated from patients with inflammatory bowel disease.
Gut 1992; 33:646–50.
5. Kallinowski F, Wassmer A, Hofmann MA, et al. Prevalence of enteropatho-
genic bacteria in surgically treated chronic inflammatory bowel disease.
Hepatogastroenterology 1998; 45:1552–8.
6. Rembacken BJ, Snelling AM, Hawkey PM, Chalmers DM, Axon AT. Non-
pathogenic Escherichia coli versus mesalazine for the treatment of ulcer-
ative colitis: a randomised trial. Lancet 1999; 354:635–9.
7. Bertone ER, Giovannucci EL, King NW Jr, Petto AJ, Johnson LD. Family
history as a risk factor for ulcerative colitis–associated colon cancer in
cotton-top tamarin. Gastroenterology 1998; 114:669–74.
8. Yamamoto T, Endo S, Yokota T, Echeverria P. Characteristics of adherence
of enteroaggregative E. coli to human and animal mucosa. Infect Immun
1991; 59:3722–39.
9. Beutin L, Zimmermann S, Gleier K. Rapid detection and isolation of Shiga-
like toxin (verocytotoxin)–producing Escherichia coli by direct testing of
individual enterohemolytic colonies from washed sheep blood agar plates
in the VTEC-RPLA assay. J Clin Microbiol 1996; 34:2812–4.
10. Paton AW, Paton JC. Detection and characterization of Shiga toxigenic Es-
cherichia coli by using multiplex PCR assays for stx1, stx2, eaeA, ente-
rohemorrhagic E. coli hlyA, rfbO111, and rfbO157. J Clin Microbiol
1998; 36:598–602.
11. Oswald E, Schmidt H, Morabito S, Karch H, Marches O, Caprioli A. Typing
of intimin genes in human and animal enterohemorrhagic and entero-
pathogenic Escherichia coli: characterization of a new intimin variant.
Infect Immun 2000; 68:64–71.
12. Blank TE, Zhong H, Bell AL, Whittam TS, Donnenberg MS. Molecular
JID 2001;184 (15 September) Enteropathogenic E. coli in Tamarins 807

variation among type IV pilin (bfpA) genes from diverse enteropathogenic
Escherichia coli strains. Infect Immun 2000; 68:7028–38.
13. Mansfield KG, Lin KC, Newman J, et al. Identification of enteropathogenic
Escherichia coli in simian immunodeficiency virus–infected infant and
adult Rhesus macaques. J Clin Microbiol 2001; 39:971–6.
14. Wood JD, Peck OC, Tefend KS, et al. Colitis and colon cancer in cotton-
top tamarins (Saguinus oedipus oedipus) living wild in their natural hab-
itat. Dig Dis Sci 1998; 43:1443–53.
15. Barnett FD, Philpott D, Abul-Milh M, Huesca M, Sherman PM, Lingwood
CA. Phosphatidylethanolamine recognition promotes enteropathogenic E.
coli and enterohemorrhagic E. coli host cell attachment. Microb Pathog
1999; 27:289–301.

Downloaded from https://academic.oup.com/jid/article-abstract/184/6/803/847412 by Universidad de La Salle user on 31 July 2020