British Journal of Rheumatology 1998;37:525–531
CHARACTERIZATION OF THE HUMORAL IMMUNE RESPONSE TO KLEBSIELLA
SPECIES IN INFLAMMATORY BOWEL DISEASE AND ANKYLOSING
SPONDYLITIS
H. TIWANA,* R. S. WALMSLEY,† C. WILSON,* J. Y. YIANNAKOU,‡ P. J. CICLITIRA,‡
A. J. WAKEFIELD† and A. EBRINGER*§
*Division of Life Sciences, Infection and Immunity Group, King’s College, †Inflammatory Bowel Disease Study Group, Royal
Free Hospital and School of Medicine, ‡Gastroenterology Unit, UMDS St Thomas’ Hospital and §Department of
Rheumatology, UCL School of Medicine, Middlesex Hospital, London
SUMMARY
This study was carried out to characterize the antibody class response by ELISA to seven Klebsiella pneumoniae serotypes ( K2,
K3, K17, K21, K26, K36, K50) in five different groups, 40 HLA-B27-positive ankylosing spondylitis (AS) patients, 46 patients
with Crohn’s disease (CD), 38 patients with ulcerative colitis ( UC ), 50 patients with active anti-endomysial antibody-positive
coeliac disease and 40 healthy controls, using whole bacteria and capsular polysaccharide. IgG antibody levels were significantly
elevated in AS patients to K17, K36, K50; IgA to K2, K3, K21, K26, K36 and K50; and IgM to serotype K21 when compared
to normal controls. Furthermore, IgG antibody levels were significantly elevated in CD patients to K2, K17, K21, K26, K36
and K50; IgA to K2, K3, K21, K26, K36 and K50; and IgM to K2, K3, K17, K21 and K50. Increased IgG antibody levels
in the UC group were limited only to K17, K36 and K50. No antibody class was increased to any of the K. pneumoniae
serotypes in the coeliac disease group. The immune responses in AS patients also involve Klebsiella bacteria having capsular
serotypes other than K26, K36 and K50. The similarity in the immune responses between CD and AS groups suggests that
many AS patients may have occult bowel inflammation.
K : Ankylosing spondylitis, Crohn’s disease, Ulcerative colitis, Klebsiella pneumoniae, Capsular serotypes.
A spondylitis (AS) can be considered as a
reactive arthritis following Klebsiella pneumoniae infection in HLA-B27-positive patients, based on consistent
findings of raised anti-Klebsiella antibodies [1–3] and
the presence of molecular mimicry between the HLAB27 molecule and bacterial antigens [4–6 ]. Ileocolonoscopic studies show that bowel inflammation
occurs in up to two-thirds of spondyloarthropathy
patients during active phases of the disease [7–10].
Antibodies to various microbes of the gut bacterial
flora have been found in inflammatory bowel disease
(IBD), including K. pneumoniae spp. [11–13], and
bacterial antigens and complexes of cell wall components have been detected in areas of inflammation
beneath ulcers and fissures [14, 15]. Exposure of the
mucosal immune system to bacterial antigens is likely
to promote an inflammatory response [16–18] with the
production of antibodies against K. pneumoniae which
may lead to AS in the genetically predisposed HLAB27-positive individual. We have previously shown
that our population of AS and IBD patients have
elevated total antibody titres to K. pneumoniae, but
not to Eubacterium, Bacteroides, Peptostreptococcus,
Escherichia coli or Proteus mirabilis [19]. It has been
suggested that immune responses to K. pneumoniae
serotypes K26, K36 and K50 are of importance in
HLA-B27-positive AS patients [20] with IgA antibod-
ies in particular having a pathogenic role, as their titre
correlates with disease activity [13, 21].
In order to clarify this further, antibody responses
to seven K. pneumoniae serotypes have been studied in
patients with chronic bowel inflammation and AS, and
compared to healthy controls.
MATERIALS AND METHODS
Patients and controls
Sera were collected from five patient groups: AS,
Crohn’s disease (CD), ulcerative colitis ( UC ), coeliac
disease and normal blood donors. Forty unrelated
HLA-B27-positive AS patients (New York criteria)
[22], with active disease and an erythrocyte sedimentation rate ( ESR) >15 mm/h, were obtained from the
AS Research Clinic at the Middlesex Hospital,
London. The IBD groups were obtained from the
Department of Gastroenterology, Royal Free Hospital,
London. Forty-six unrelated patients had CD. Active
disease was defined by physician’s assessment with
raised C-reactive protein (CRP) and/or a Harvey–
Bradshaw score >4 [23] (19 active). Twenty-three had
terminal ileal and colonic disease, 10 had colonic
disease alone and 13 had disease limited to the small
bowel. Thirty-eight unrelated patients had UC. Active
disease was based on physician’s assessment with
accompanying raised ESR, or a simple colitis activity
score >3 [24], or an integrated activity score >150
[25] (21 active). Eight had total/extensive disease, 20
had left-sided disease and 10 had proctitis alone. Three
had clinically diagnosed AS; two were HLA-B27 positive and one HLA-B27 negative. Fifty unrelated patients
with active coeliac disease [26 ] were obtained from the
Gastroenterology Unit, UMDS St Thomas’ Hospital,
Submitted 12 August 1997; revised version accepted 5 November
1997.
Correspondence to: A. Ebringer, Infection and Immunity Group,
Division of Life Sciences, King’s College, Campden Hill Road,
London W8 7AH.
© 1998 British Society for Rheumatology
525
526
BRITISH JOURNAL OF RHEUMATOLOGY VOL. 37 NO. 5
London. Nineteen patients had newly diagnosed disease, seven patients were undergoing gluten challenge
and 24 patients had active disease while on treatment.
Forty-nine patients had positive anti-endomysial antibodies, demonstrated by immunohistochemistry on
human umbilical cord [27]. Forty healthy control sera
were supplied by the Blood Transfusion Service,
London ( Table I ).
Klebsiella pneumoniae capsular serotypes
Klebsiella pneumoniae isolates of capsular types K2
(NTCC-5055), K3 (5056), K21 (9141), K26 (9146),
K36 (9156) and K50 (9170) were kindly provided by
Dr T. Pitt (Division of Hospital Infection, Central
Public Health Laboratory, Colindale). The K17 serotype used was isolated from the faecal sample of a CD
patient. We selected the three most frequent serotypes
found in the UK, K2, K3 and K21, which are commonly isolated from blood/throat, urine and faeces,
respectively [28], and K26, K36 and K50 serotypes
which had been reported to evoke higher antibody
levels in HLA-B27-positive AS patients compared with
other K. pneumoniae capsular serotypes [20].
Isolation of K. pneumoniae
Faecal samples were collected from six CD and three
UC patients, attending the Royal Free Hospital,
London. A loopful of test material was streaked across
plates of nutrient agar ( Unipath Ltd, Hampshire) to
which had been added 1% w/v inositol, 0.3% v/v of
1% bromothymol blue, 1% w/v Lab lemco and 1%
bacteriological peptone. The plates were incubated for
18 h at 37°C.
Inositol-fermenting colonies were identified using the
biochemical API20E kit (Biomérieux, Hampshire) as
K. pneumoniae spp. Three K. pneumoniae isolates from
individual CD patients were kindly serotyped by Dr
T. Pitt (Central Public Health Laboratory, Colindale).
Two positive inositol-fermenting colonies could not be
identified further. The third was identified as K. pneumoniae serotype K17 and was isolated from a 24-yrold patient with CD having inactive disease, who
previously had a resection of both sigmoid colon and
terminal ileum, and was taking 5-ASA.
Preparation of whole bacteria
The individual K. pneumoniae serotypes were grown
in batch culture as described previously [29]. Starter
cultures containing 20 ml of nutrient broth ( Unipath
13 g/l ), were set up by inoculating with two loopfuls
of bacteria from the nutrient agar slopes. The cultures
were incubated at 37°C for 7 h on an orbital shaker,
after which 200 ml of nutrient broth in 2 l flasks were
inoculated with 10 ml of the corresponding cultures
and left shaking at 37°C for 16 h. The bacterial cultures
were harvested by centrifugation at 6000 r.p.m. (MSE
High Speed18, 6 × 250 ml rotor) for 20 min, washed
and suspended in 0.15  phosphate-buffered saline
(PBS; pH 7.4); stock solution was prepared to give an
OD reading at 540 nm of 0.25 on the Corning
Spectrophotometer (Model 258) which gives a suspension of 6 × 108 bacteria/ml.
Preparation of capsular extract
The capsular polysaccharide of K. pneumoniae was
extracted as described by Sahly et al. [20]. Klebsiella
pneumoniae isolates of capsular types K2, K3, K17,
K21, K26, K36 and K50 were grown on Worfel–
Ferguson agar [30] at 37°C for 24 h, followed by
another 24 h at room temperature to increase capsule
production. Using a sterile, disposable cell scraper
(Bibby Sterilin Ltd, Staffordshire), the bacterial lawns
were suspended in 10 ml of 0.9% NaCl in a conical
flask containing 0.2 mm glass beads. The bacterial
suspensions were shaken vigorously for 7 min to detach
the capsular polysaccharides, on an orbital shaker,
glass beads removed and further centrifuged at 4°C
for 20 min at 6500 r.p.m. (Beckman JA-20).
Supernatants were filtered using a membrane with a
porosity of 0.22 mm (Millipore, Hertfordshire). The
filtrate was lyophilized for each of the capsular serotypes and the resulting lyophilizates were dissolved in
1 ml of distilled water. Total carbohydrate content was
determined by the phenol–sulphuric acid method [31].
The capsular polysaccharide preparations were diluted
in carbonate buffer (0.05 , pH 9.6), giving a final
concentration of 2 mg of carbohydrate/ml.
Enzyme-linked immunosorbent assay (ELISA)
Ninety-six-well, flat-bottomed, rigid polystyrene
microtitre plates (Dynatech, Billinghurst) were coated
with 200 ml of bacteria or capsular extract in carbonate
buffer (0.05 , pH 9.6), overnight at 4°C. The plates
were then washed three times with PBS (pH 7.4),
300 ml of blocking solution were added to each well
[0.1% (w/v) bovine serum albumin (BSA), 0.05% (v/v)
Tween 20] in PBS and incubated for 1 h at 37°C.
Excess blocking solution was removed and 200 ml
TABLE I
General characteristics of the patient and control groups
Number
Male:female
Mean age, yr (range)
Mean ESR (range)
Mean CRP ± ..
Healthy
controls
Coeliac
disease
Ulcerative
colitis
Ankylosing
spondylitis
Crohn’s
disease
40
20:20
41 (30–59)
ND
1.2 ± 0.5
50
19:31
35 (11–75)
ND
5.2 ± 0.9
38
19:19
46 (24–79)
ND
19.2 ± 2.0
40
31:9
50 (19–72)
41 (17–133)
14.3 ± 2.0
46
19:27
41 (20–82)
ND
19.7 ± 2.1
ESR, erythrocyte sedimentation rate (mm/h); CRP, C-reactive protein (mg/l ); ND, not done.
TIWANA ET AL.: IMMUNE RESPONSE TO KLEBSIELLA IN IBD AND AS
527
serum samples, test or control, diluted 1/200 in PBS–
Tween were added in duplicate. Plates were incubated
for 2 h at 37°C, followed by the addition of 200 ml of
peroxidase-conjugated rabbit anti-human class-specific
IgG, IgA or IgM (Dako Ltd, Buckinghamshire) diluted
1/500 in PBS–Tween, and incubated for a further 2 h
at 37°C. Development of the colorimetric assay took
place at room temperature for 20 min after the addition
of 200 ml/well 0.5 g/ml 2,2∞-azinobis(3-ethylbenzthiazoline-6-sulphonic acid) (Sigma, Dorset) in citrate/
phosphate buffer (pH 4.1), containing 0.98 mM H O
2 2
(Sigma). The reaction was stopped with 100 ml of
2 mg/ml sodium fluoride (Sigma) and ODs measured
at a wavelength of 630 nm with a micro-ELISA plate
reader (Dynatech MR600). All assays were carried out
under code, in that the tester did not know whether a
sample was a test or control serum.
C-Reactive protein determination
CRP levels were determined by the single radial
immunodiffusion method of Mancini et al. [32] and
the results expressed in mg/l of serum.
Statistical analysis
Differences in immunoglobulin class levels against
the various K. pneumoniae isolates in the serum samples
of patients and those classified according to disease
site were assessed by comparing the proportion in each
group having OD units greater than the 95% confidence
limits for the population of controls (one-tailed test)
using the x2 test, with Yates correction (significance
was taken at P < 0.05). The reproducibility of the
assay was tested by calculating the coefficient of
variation.
RESULTS
Antibodies to whole bacteria
Levels of IgG, IgA and IgM antibodies to each of
the seven different K. pneumoniae serotypes were measured by ELISA, and the percentage of subjects (disease
and control ) with immunoglobulin levels above the
95% confidence limit of the control population (onetailed test) were determined.
A significant percentage of AS patients showed
raised IgG levels to whole K. pneumoniae spp. K17
(x2 = 5.60, P < 0.05), K36 (x2 = 13.20, P < 0.001)
and K50 (x2 = 4.24, P < 0.05) compared with the
healthy control population (Fig. 1A). Furthermore,
IgA antibody levels to K21 (x2 = 7.58, P < 0.01), K26
(x2 = 4.78, P < 0.05), K36 (x2 = 6.33, P < 0.05), K50
(x2 = 12.11, P < 0.001) (Fig. 1B) and of IgM to K21
(x2 = 2.85, P < 0.05) were also found to be elevated
in a greater number of AS patients when compared
with controls (Fig. 1C ).
In a significant percentage of CD patients, IgG
antibody levels were raised to K17 (x2 = 23.44,
P < 0.001), K26 (x2 = 8.16, P < 0.01), K36
(x2 = 21.92, P < 0.001) and K50 (x2 = 13.93,
P < 0.001) when compared with controls ( Fig. 1A). A
similar IgA response was detected in both CD and AS
patients. The IgA antibody level was found to be raised
F. 1. —Percentage of ankylosing spondylitis, Crohn’s disease,
ulcerative colitis and coeliac disease patients with antibodies ( ELISA
OD 630 nm) IgG (A), IgA (B) and IgM (C ) elevated above the 95%
confidence limit of controls to whole bacterial preparations of K.
pneumoniae serotypes K2, K3, K17, K21, K26, K36 and K50. An
asterisk indicates statistical significance compared to the healthy
control group; *P < 0.05, **P < 0.01, ***P < 0.001 using x2 test
with Yates correction.
in a greater proportion of patients above the 95%
limit to K2 (x2 = 4.23, P < 0.05), K21 (x2 = 12.24,
P < 0.001), K26 (x2 = 13.01, P < 0.001), K36
(x2 = 15.23, P < 0.001) and K50 (x2 = 22.10,
P < 0.001) (Fig. 1B). When measuring IgM antibody
levels, a significant response was seen using K21
(x2 = 7.01, P < 0.01) and K50 (x2 = 11.40, P < 0.001)
(Fig. 1C ).
No significant differences in mean antibody levels to
any of the K. pneumoniae spp. tested were found when
either CD or UC patients were divided into clinically
active and inactive groups. Furthermore, there was no
significant difference in the number of both CD and
528
BRITISH JOURNAL OF RHEUMATOLOGY VOL. 37 NO. 5
UC patients with antibody levels above the confidence
limits when patients were classified according to disease
site ( Table II ).
A proportion of patients with UC exhibited elevated
IgG antibody levels to K. pneumoniae spp. K17
(x2 = 4.51, P < 0.05), K36 (x2 = 9.19, P < 0.01) and
K50 (x2 = 5.25, P < 0.05) (Fig. 1A). Two UC patients
with AS and who were also HLA-B27 positive had
IgG, IgA and IgM antibody levels above the confidence
limit to the whole bacteria K2, K3, K17 and K50, and
antibodies to capsular extracts of K2, K3, K21, K26
and K50. Both were male; one had disease limited to
the left colon, and the second had total colitis. Both
had active disease requiring steroid therapy at the time
of sampling. One male UC patient with AS who was
HLA-B27 negative with total extensive active disease
had increased IgG and IgA immunoglobulin levels
above the 95% limit to K36 and K50 only. Patients
with coeliac disease did not have any antibody class
levels elevated above the confidence limits to any of
the K. pneumoniae spp. tested.
The mean coefficient of variation for IgG bacterial
serotype assays was 3.9% (range 2.7–6.9%), for IgA
TABLE II
Numbers of patients with immunoglobulin class reactivity to K.
pneumoniae serotypes (whole bacteria) above the 95% confidence
limit of controls in Crohn’s disease and ulcerative colitis patients
classified according to disease state
Immunoglobulin class
Klebsiella
serotype
Study group
IgG
IgA
7/23a
2/10b
2/13c
IgM
K2
CD
–
–
–
–
–
–
K17
CD
10/23a
8/10b
7/13c
–
–
–
–
–
–
UC
1/8e
4/10f
3/20g
–
–
–
–
–
–
K21
CD
–
–
–
12/23a
6/10b
6/13c
6/23a
2/10b
6/13c
K26
CD
10/23a
2/10b
5/13c
11/23a
5/10b
7/13c
–
–
–
K36
CD
14/23a
3/10b
6/13c
12/23a
5/10b
6/13c
–
–
–
UC
3/8e
4/10f
5/20g
–
–
–
–
–
–
CD
11/23a
7/10b
4/13c
14/23a
5/10b
7/13c
8/23a
3/10b
5/13c
UC
3/7e
4/10f
5/20g
–
–
–
–
–
–
K50
a, terminal ileal and colonic disease; b, colonic disease alone;
c, small bowel disease only; e, total/extensive disease; f, proctitis
alone; g, left-sided disease.
F. 2. —Percentage of ankylosing spondylitis, Crohn’s disease,
ulcerative colitis and coeliac disease patients with antibodies ( ELISA
OD 630 nm) IgG (A), IgA (B) and IgM (C ) elevated above the 95%
confidence limit of controls to capsular extracts of K. pneumoniae
serotypes K2, K3, K17, K21, K26, K36 and K50. An asterisk
indicates statistical significance compared to the healthy control
group; *P < 0.05, **P < 0.01, ***P < 0.001 using x2 with Yates
correction.
assays 7.8% (range 4.7–9.6%) and for IgM assays 4.1%
(range 2.8–6.6%).
Antibodies to capsular extract
IgG antibody levels to capsular extract of K36
(x2 = 3.80, P < 0.05) were found to be raised in a
significant percentage of AS patients (Fig. 2A); IgA
antibody levels were elevated to K2 (x2 = 11.96,
P < 0.001), K3 (x2 = 17.20, P < 0.001), K21
(x2 = 3.79, P < 0.05), K26 (x2 = 8.06, P < 0.01) and
K36 (x2 = 6.33, P < 0.05) extracts (Fig. 2B), but no
increased IgM antibody elevation to any capsular
TIWANA ET AL.: IMMUNE RESPONSE TO KLEBSIELLA IN IBD AND AS
extract was detected in AS patients when compared to
healthy controls (Fig. 2C ).
A significant number of CD patients showed
increased IgG antibody levels to capsular extracts
of K2 (x2 = 12.69, P < 0.001), K21 (x2 = 7.14,
P < 0.01), K26 (x2 = 10.33, P < 0.01), K36
(x2 = 8.04, P < 0.01) and K50 (x2 = 7.39, P < 0.01)
in comparison with the controls (Fig. 2A), whilst IgA
antibodies were detected to the same K. pneumoniae
extracts as those found to be elevated in AS
patients: K2 (x2 = 18.48, P < 0.001), K3 (x2 = 25.86,
P < 0.001), K21 (x2 = 6.04, P < 0.01), K26
(x2 = 10.61, P < 0.001) and K36 (x2 = 11.48,
P < 0.001) ( Fig. 2B). When testing for IgM antibodies,
a significant percentage of CD patients had increased
levels to K2 (x2 = 3.56, P < 0.05), K3 (x2 = 3.57,
P < 0.05), K17 (x2 = 3.82, P < 0.05) and K50
(x2 = 6.04, P < 0.05) extracts compared to controls
( Fig. 2C ).
There was no significant difference between mean
antibody levels to the K. pneumoniae capsular extracts
in CD and UC patients when divided according to
disease activity. Furthermore, there was no significant
difference in the number of CD patients having antibody class levels above the confidence limit to
K. pneumoniae extracts when classified according to
disease site ( Table III ). In addition, no antibody class
levels were elevated in the UC or coeliac disease groups
to any of the capsular extracts tested when compared
TABLE III
Numbers of patients with immunoglobulin class reactivity to
K. pneumoniae serotypes (capsular extract) above the 95% confidence
limit of controls in Crohn’s disease patients classified according to
disease state
Immunoglobulin class
Klebsiella
serotype
Study group
IgG
IgA
IgM
K2
CD
11/23a
5/10b
5/13c
15/23a
6/10b
6/13c
6/23a
1/10b
5/13c
K3
CD
–
–
–
15/23a
5/10b
10/13c
5/23a
2/10b
5/13c
K17
CD
–
–
–
–
–
–
5/23a
3/10b
6/13c
K21
CD
9/23a
2/10b
5/13c
8/23a
2/10b
3/13c
–
–
–
K26
CD
11/23a
2/10b
6/13c
9/23a
4/10b
8/13c
–
–
–
K36
CD
9/23a
2/10b
4/13c
10/23a
5/10b
5/13c
–
–
–
K50
CD
9/23a
2/10b
7/13c
–
–
–
6/23a
2/10b
6/13c
a, terminal ileal and colonic disease; b, colonic disease alone;
c, small bowel disease only.
529
to healthy controls, and there was no statistical difference in the number of UC patients who showed raised
IgG antibodies above the confidence limits when classified according to disease site.
The mean coefficient of variation value for the
capsular IgG assays was 4.8% (range 4.2–5.4%), for
IgA assays 6.6% (range 5.4–8.0%) and for IgM assays
3.9% (range 2.3–6.9%).
DISCUSSION
This study confirms that both AS and CD patients
have elevated levels of IgG, IgA and IgM antibodies
to K. pneumoniae, whilst UC patients had elevated IgG
only. The immune responses also involve Klebsiella
bacteria having capsular serotypes other than K26,
K36 and K50. The capsular serotypes of K. pneumoniae
vary in their pathogenicity. Sahly et al. [20] undertook
an extensive investigation of the role of all 77 capsular
serotypes in the immune responses of AS patients,
comparing them with reactive arthritis following
Yersinia infection, rheumatoid arthritis, psoriatic arthritis, systemic lupus erythematosus patients and normal
controls. They found that the sera from HLAB27-positive AS patients had significantly higher levels
of IgG antibodies to the capsular serotypes K26, K36
and K50 compared to all other groups, suggesting a
pre-eminent role for these microbes.
We have shown that a certain percentage of AS
patients in our populations have significant IgG antibody responses to the capsular extract of K36 and to
the whole bacteria of K17, K36 and K50. IgA antibodies to capsular extracts of K2, K3, K21, K26, and to
K36 and to whole bacterial preparations of K21, K26,
K36 and K50 were also found. In contrast, apart from
K21, IgM levels to any of the serotypes could not be
demonstrated. To some degree the difference between
our results and those of Sahly et al. [20] may be due
to different frequencies of K. pneumoniae spp. in the
two countries. For instance, in Japan, K. pneumoniae
serotypes frequently isolated from the general population are K1, K2, K7, K10, K33 and K43 [33].
In CD patients, we have shown serum IgG elevations
to capsular extracts of K2, K21, K26, K36 and K50,
and IgA to the same isolates as those detected in AS
patients, and IgM antibodies to capsular extract of
K2, K3, K17 and K50 with additional raised IgM
levels to whole bacteria K21 and K50. The presence
of an IgM response to both the whole bacteria and
capsular extract may suggest a recent/ongoing exposure
to K. pneumoniae, implying that the disease is not
quiescent. However, in both CD and UC groups, there
was no relationship between increased antibody levels
and disease activity or disease site.
In UC patients, we found predominantly an
increased IgG response to whole K. pneumoniae which
was limited to K17, K36 and K50. Two UC patients
who had AS and were HLA-B27 positive demonstrated
increased humoral immune responses to a greater
number of both whole and capsular extracts of
Klebsiella serotypes tested compared to the UC patient
with AS who was HLA-B27 negative. The CD patients
530
BRITISH JOURNAL OF RHEUMATOLOGY VOL. 37 NO. 5
used were not tissue typed for HLA-B27, but none of
the patients studied had AS.
An early study reported an antibody elevation in
both CD and UC patients to Yersinia enterocolitica
0:3, 0:9 and to K. pneumoniae. However, the serotype
of the Klebsiella used was unknown [11]. In related
studies, antibody levels to K43 have been shown to be
elevated in CD, AS and RA patients [12, 13], but this
particular serotype was not used for this study.
The association of AS with overt bowel disease is
well established with an increased prevalence of IBD
in families of AS patients of up to 5%. Oligoarthritis
or classical AS occurs in up to 20% of IBD patients
[34] and there is good evidence for occult bowel
inflammation in many of those with spondyloarthropathies. In a study of ileocolonoscopic findings in
232 patients with seronegative spondyloarthropathy,
Mielants et al. [7] observed inflammatory gut lesions
in 57% of AS patients. In the follow-up study, 25% of
those who had chronic inflammatory changes on initial
biopsy had developed CD [35]. In a similar study in
Finland, a diagnosis of CD was made on follow-up in
26% of patients with chronic spondyloarthropathy [9].
However, there was no association of gut lesions with
the use of non-steroidal anti-inflammatory drugs and
similar observations have been made by other groups
[8, 10].
This study provides supportive evidence for the
possible role of the capsular serotypes K26, K36 and
K50 in AS, but also suggests that a broader range of
serotypes may be involved. It confirms the presence of
both IgG and IgA anti-K. pneumoniae antibodies in
AS. In IBD, we found IgG, IgA and IgM antibodies
to the majority of the serotypes tested in the CD
patients, with a restricted but significant immune
response in the UC group. It is possible that both
CD/UC and a considerable proportion of AS patients
have bowel inflammation, and that the increased humoral immune response towards K. pneumoniae may
be a secondary phenomenon, following intestinal
inflammation. Furthermore, the similarity in the
immune responses between CD and AS patients,
together with the mounting evidence from endoscopic
studies, suggest that AS patients represent a section of
the population with occult IBD who develop spondyloarthropathy as a result of being HLA-B27 positive and
being exposed to K. pneumoniae bacteria.
A
The authors gratefully acknowledge the support of
the Arthritis and Rheumatism Council and the Trustees
of the Middlesex Hospital.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
R
1. Maki-Ikola O, Lehtinen K, Granfors K, Vainionpaa R,
Toivanen P. Bacterial antibodies in ankylosing spondylitis. Clin Exp Immunol 1991;84:472–5.
2. Shodjai-Moradi F, Ebringer A, Abuljadayel I. IgA antibody response to Klebsiella in ankylosing spondylitis
measured by immunoblotting. Ann Rheum Dis 1992;
51:233–7.
3. Collado A, Gratacos J, Ebringer A, Rashid T, Sanmarti
18.
19.
R, Munoz-Gomez J. Serum IgA anti-Klebsiella antibodies in ankylosing spondylitis patients from Catalonia.
Scand J Rheumatol 1994;23:119–23.
Schwimmbeck PL, Yu DTY, Oldstone MBA.
Autoantibodies to HLA-B27 in the sera of HLA-B27
patients with ankylosing spondylitis and Reiter’s syndrome: molecular mimicry with Klebsiella pneumoniae as
potential mechanism of autoimmune disease. J Exp Med
1987;166:173–81.
Scofield RH, Kurien B, Gross T, Warren WL, Harley
JB. HLA-B27 binding of peptide from its own sequence
and similar peptides from bacteria: implications for
spondyloarthropathies. Lancet 1995;345:1542–4.
Fielder M, Pirt SJ, Tarpey I et al. Molecular mimicry
and ankylosing spondylitis: possible role of a novel
sequence in pullulanase of Klebsiella pneumoniae. FEBS
Lett 1995;369:243–8.
Mielants H, Veys EM, Cuvelier C, de Vos M.
Ileocolonoscopic findings in seronegative spondylarthropathies. Br J Rheumatol 1988;27(suppl. II ):95–105.
Simenon G, van Gossum A, Adler M, Rickaert F,
Applebloom T. Macroscopic and microscopic gut lesions
in seronegative spondyloarthropathies. J Rheumatol
1990;17:1491–4.
Leirisalo-Repo M, Turunen U, Stenman S, Helenius P,
Seppala K. High frequency of silent inflammatory bowel
disease in spondyloarthropathy. Arthritis Rheum 1994;
37:23–31.
Grillet B, de Clerck L, Dequeker J, Rutgeerts P,
Geboes K. Systematic ileocolonoscopy and bowel biopsy
study in spondylarthropathy. Br J Rheumatol 1987;
17:1491–4.
Ibbotson JP, Pease PE, Allan RN. Serological studies in
Crohn’s disease. Eur J Clin Microbiol 1987;6:286–90.
Cooper R, Fraser SM, Sturrock RD, Gemmell CG.
Raised titres of anti-Klebsiella IgA in ankylosing spondylitis, rheumatoid arthritis and inflammatory bowel
disease. Br Med J 1988;296:1432–4.
O’Mahony S, Anderson N, Nuki G, Ferguson A.
Systemic and mucosal antibodies to Klebsiella in patients
with ankylosing spondylitis and Crohn’s disease. Ann
Rheum Dis 1992;51:1296–300.
Liu Y, van Kruningen HJ, West AB, Cartun RW,
Cortot A, Colombel J. Immunocytochemical evidence of
Listeria, Escherichia coli and Streptococcus antigens in
Crohn’s disease. Gastroenterology 1995;108:1396–404.
Klasen IS, Melief MJ, van Halteren AGS et al. The
presence of peptidoglycan-polysaccharide complexes in
the bowel wall and the cellular responses to these complexes in Crohn’s disease. Clin Immunol Immunopathol
1994;71:303–8.
Blaser MJ, Miller RA, Lacher J, Singleton JW. Patients
with active Crohn’s disease have elevated serum antibodies to antigens of seven enteric bacterial pathogens.
Gastroenterology 1984;87:888–94.
Duchmann R, Kaiser I, Hwermann E et al. Tolerance
exists towards resident intestinal flora but is broken in
active inflammatory bowel disease. Clin Exp Immunol
1995;102:448–55.
Macpherson A, Khoo UY, Forgacs I, Philpott-Howard
J, Bjarnason I. Mucosal antibodies in inflammatory
bowel disease are directed against intestinal bacteria.
Gut 1996;38:365–75.
Tiwana H, Wilson C, Walmsley RS et al. Antibody
response to gut bacteria in ankylosing spondylitis,
rheumatoid arthritis, Crohn’s disease and ulcerative colitis. Rheumatol Int 1997;17:11–6.
TIWANA ET AL.: IMMUNE RESPONSE TO KLEBSIELLA IN IBD AND AS
20. Sahly H, Kekow J, Podschun R, Schaff M, Gross WL,
Ullmann U. Comparison of the antibody responses to
the 77 Klebsiella capsular types in ankylosing spondylitis
and various rheumatic diseases. Infect Immun
1994;62:4838–43.
21. Trull A, Ebringer R, Panayi GS, Colthorpe D, James
DCO, Ebringer A. IgA antibodies in ankylosing spondylitis. Scand J Rheumatol 1983;12:249–53.
22. Bennet PH, Wood PHN. Population studies of the
rheumatic diseases. In: Bennet PH, Wood PHN, eds.
Proceedings of the Third International Symposium. New
York: Excerpta Medica Foundation, 1966.
23. Harvey RF, Bradshaw JM. A simple index of Crohn’s
disease activity. Lancet 1980;1:514.
24. Walmsley RS, Ayres RCS, Allan RN. A really simple
colitis activity index. Gut 1994;35(suppl. 5):S29.
25. Seo M, Okada M, Yao T, Ueki M, Arim S,
Okumara M. An index of disease activity in ulcerative
colitis. Am J Gastroenterol 1992;87:971–6.
26. Walker-Smith JA, Guandalini S, Schmitz J, Shmerling
DH, Visakorpi JK. Revised criteria for the diagnosis of
coeliac disease. Arch Dis Child 1990;65:909–11.
27. Ladinser B, Rossipal E, Pittschieler K. Endomysium
antibodies in coeliac disease: an improved method. Gut
1994;35:776–8.
28. Gaston MA, Ayling-Smith BA, Pitt TL. New bacteriophage typing scheme for subdivision of the frequent
capsular serotypes of Klebsiella spp. J Clin Microbiol
1987;25:1228–32.
531
29. Khalafpour S, Ebringer A, Abuljadayel I, Corbett M.
Antibodies to Klebsiella and Proteus microorganisms in
ankylosing spondylitis and rheumatoid arthritis patients
measured by ELISA. Br J Rheumatol 1988;27(suppl.
II ):86–9.
30. Worfel MT, Ferguson WW. Media and reagents. In:
Edwards PR, Ewing WH, eds. Identification of
Enterobacteriaceae, 3rd edn. Atlanta, GA: Burgess
Publishing Co., 1972, p. 354.
31. Chaplin MF. Monosaccharides. In: Chaplin MF,
Kennedy JF, eds. Carbohydrate analysis: a practical
approach, 2nd edn. New York: Oxford University Press,
1994, p. 2.
32. Mancini G, Carbonara HO, Heremans JF.
Immunological quantification of antigens by single
radial immunodiffusion. Immunochemistry 1965;2:
235–54.
33. Mori M, Ohta M, Agata N et al. Identification of species
and capsular types of Klebsiella clinical isolates, with
special reference to Klebsiella planticola. Microbiol
Immunol 1989;33:887–95.
34. Macrae I, Wright V. A family study of ulcerative colitis
with a particular reference to ankylosing spondylitis.
Ann Rheum Dis 1973;32:16–20.
35. De Vos M, Mielants H, Cuvelier C, Elewaut A, Veys E.
Long-term evolution of gut inflammation in patients
with spondyloarthropathy. Gastroenterology 1996;
110:1696–703.