Tumour Necrosis Factor Microsatellite

Derek A O’Reilly
1,2
, Simon Dunlop
3
, Kevin Sargen
1,2
, Andrew Demaine
2
, Stephen Wilkinson
3
,
Andrew N Kingsnorth
1
1
Department of Surgery Derriford Hospital;
2
Department of Molecular Medicine, University of
Plymouth;
3
Department ofGastroenterology, Derriford Hospital. Plymouth, United Kingdom
ABSTRACT
Context Alcohol is the major aetiological
agent for both chronic pancreatitis and
alcoholic liver disease. However, as only a
minority of alcoholics develop either chronic
pancreatitis or alcoholic liver disease, there
are clearly genetic or environmental cofactors
that determine individual susceptibility to
these diseases.
Objective
To
determine
whether
polymorphisms of the TNF gene may account
for individual susceptibility to develop
chronic pancreatitis or alcoholic liver disease.
Design A controlled study.
Patients We analyzed 73 patients with
chronic pancreatitis, 103 healthy controls, 39
patients with alcoholic liver disease and 29
alcoholics without liver or pancreatic disease.
Results The intermediate/low TNF secreting
haplotype a6b5c1d3e3 was over-represented
in chronic pancreatitis compared to healthy
controls (OR=2.08; 95% CI: 1.07-4.06);
P=0.019) and in alcoholic chronic pancreatitis
compared to healthy controls (OR=2.08; 95%
CI: 1.01-4.29; P=0.029). The high TNF
secreting haplotypes, a2b3c1d1e3 and
a2b5c2d4e3 were under-represented in
chronic pancreatitis compared to healthy
controls (OR=0.48; 95% CI: 0.22-1.04; P=
0.043) and in alcoholic chronic pancreatitis
compared to alcoholic controls (OR=0.20;
95% CI: 0.05-0.77; P=0.014), respectively.
Conclusion A reduced capacity to produce
TNF may be responsible for the induction of
chronic pancreatitis.
INTRODUCTION
No single theory fully explains the clinical
and morphological evolution of chronic
pancreatitis. Theories to explain the
pathogenesis of chronic pancreatitis include
pancreatic duct obstruction by protein plugs
[1], the toxic effects of ethanol and its
metabolites [2], mounting oxidant stress [3]
and recurrent acute pancreatitis [4]. More
recent studies, which demonstrate the
presence of activated cytotoxic T-cells within
chronic pancreatitis resection specimens,
suggest a primary pathogenic role for the
immune system [5]. Similarly, a number of
hypotheses regarding the pathogenesis of
alcoholic liver disease exist but the role of
immune responses leading to tissue injury has
recently been emphasized. Studies have
demonstrated circulating antibodies to
acetaldehyde
modified
self-proteins,
increased expression of proinflammatory
cytokines and decreased anti-inflammatory

Page 2
JOP. J Pancreas (Online) 2006; 7(1):14-26.
JOP. Journal of the Pancreas – http://www.joplink.net – Vol. 7, No. 1 - January 2006. [ISSN 1590-8577]
15
cytokines. In particular, tumor necrosis factor
(TNF) has emerged as the main candidate
accounting for liver toxicity [6, 7, 8].
Alcohol is the major etiological agent for both
chronic pancreatitis and alcoholic liver
disease. However, as only approximately 10-
35% of all alcoholics develop alcoholic liver
disease, 5% chronic pancreatitis, and only 1%
develop both conditions [9, 10], there are
clearly genetic or environmental cofactors
that determine individual susceptibility to
these diseases. A role for heredity in the
pathogenesis of pancreatitis is suggested by
familial clustering of alcoholic pancreatitis
[11, 12] and by studies of genetic markers.
HLA antigens [13], mutations of the cystic
fibrosis transmembrane conductance regulator
gene [14, 15] and SPINK1 gene mutations
[16] have been associated with susceptibility
to alcoholic chronic pancreatitis. For
alcoholic liver disease, twin studies [17] and
associations with polymorphisms of the genes
encoding the cytochrome P-450-2E1 [18, 19],
alcohol dehydrogenase [20], TNF [21] and
interleukin-10 [22] also suggest a genetic
component. It is likely that the expression of
both these disease phenotypes, in common
with other complex human diseases, is
influenced by strong environmental risk
factors plus the possession of a range of
susceptibility genes.
TNF is produced predominantly by
macrophage/monocytes but also by T and B-
lymphocytes, neutrophils and endothelial
cells. Its functions include a pivotal role in the
initiation of the release of other
proinflammatory mediators, activation of
endothelial cells, upregulation of the
expression of intercellular adhesion
molecules, induction of T and B-lymphocyte
proliferation, enhancement of T-cell mediated
cytotoxicity and increased MHC (class I and
class II) expression [23, 24]. Stable
interindividual variations exist in levels of
production of TNF, suggesting inherited
individual differences [25]. Family studies
show that up to 60% of the variability in TNF
production between individuals may be
genetically determined [26]. Association
studies have been widely used to investigate
the putative association of TNF
polymorphisms and susceptibility to various
diseases. TNF polymorphisms have been
associated with susceptibility and outcome of
various infectious [27, 28, 29], inflammatory
[30, 31] and neoplastic [32, 33] diseases.
In acute pancreatitis, the haplotype
TNFa2b1c2 has a reduced frequency among
those of alcoholic aetiology [34]. The finding
of this genetic difference among alcoholics
who develop acute pancreatitis led us to
hypothesize that this locus may play an even
more important role in the development of
chronic pancreatitis and liver disease, where
alcohol predominates as an aetiological
factor.
Table 1. Age and sex characteristics of patients and controls.
Age (years)
Sex
Median
Range
Males
Females
Chronic pancreatitis (n=73)
- Alcoholic (n=47)
- Idiopathic (n=23)
- Hereditary (n=3)
50.5
50.5
58.0
46.0
19-87
26-68
31-87
19-50
48 (65.8%)
36 (76.6%)
11 (47.8%)
1 (33.3%)
25 (34.2%)
11 (23.4%)
12 (52.2%)
2 (66.7%)
Alcoholic liver disease (n=39)
53.5
35-70
27 (69.2%)
12 (30.8%)
Mann-Whitney P=0.580a
Chi-squared=0.14; P=0.834 a
Alcoholic controls (n=29)
53.0
36-73
26 (89.7%)
3 (10.3%)
Mann-Whitney P=0.370a
Fisher’s P=0.015 a
Healthy controls (n=103)
N/A
N/A
44 (42.7%)
59 (57.3%)
-
Chi-squared=9.09; P=0.003 a
P values vs. chronic pancreatitis
N/A: not applicable

Page 3
JOP. J Pancreas (Online) 2006; 7(1):14-26.
JOP. Journal of the Pancreas – http://www.joplink.net – Vol. 7, No. 1 - January 2006. [ISSN 1590-8577]
16
MATERIALS AND METHODS
Study Subjects and Diagnostic Criteria
DNA was extracted from 73 patients with
chronic pancreatitis (47 alcoholic, 23
idiopathic and 3 hereditary), 103 healthy
controls, 39 patients with alcoholic liver
disease and 29 alcoholics with no clinical
evidence of either liver or pancreatic disease.
Demographic data are provided in Table 1.
Analysis of the demographic data showed no
significant difference with respect to age but
significant differences were seen for
male/female sex ratios between chronic
pancreatitis and alcoholic controls (Fisher’s
P=0.015) and between chronic pancreatitis
and healthy controls (chi-squared=9.09;
P=0.003). All patients and controls were
Caucasoid.
The diagnostic criteria for chronic pancreatitis
was based upon the classification system
adopted at the Zurich Workshop of 1996 [35].
For diagnosis, a typical history of recurrent
clinical acute pancreatitis was required in
addition to one or more of the following:
pancreatic calcification, ductal lesions [36],
steatorrhoea reduced by enzyme supplement-
ation, and histology. The definition of
“definite alcoholic chronic pancreatitis” was
based upon the above plus an excessive
alcohol intake (greater than 80 g/day for more
than 2 years) [35].
Alcoholic liver disease patients were patients
with acute alcoholic hepatitis or alcohol-
induced cirrhosis who were without clinical
evidence of pancreatitis, attending the
Gastroenterology Department, Derriford
Hospital, Plymouth, UK. All had drunk at
least 8 units of alcohol per day for at least two
years.
Alcoholic controls were patients undergoing
treatment for alcoholism at the Broadreach
Rehabilitation Unit, Plymouth, UK. All had
drunk at least 56 units of alcohol per week for
two or more years. They were without clinical
or biochemical evidence of either chronic
pancreatitis or alcoholic liver disease.
Healthy controls were Caucasoid cord-blood
samples following a normal, healthy obstetric
delivery at Derriford Hospital, Plymouth, UK.
DNA Preparation
High molecular weight DNA was prepared
from peripheral venous blood using the
Nucleon II DNA extraction kit (Scotlab,
Lanarkshire, UK).
TNF Microsatellite Typing
The five microsatellite loci: TNFa, TNFb,
TNFc, TNFd, TNFe were amplified by the
polymerase chain reaction as a one-step
procedure using oligonucleotide amplimers
(MWG-Biotech, Buckinghamshire, UK)
synthesized from previously described
sequences [37] (Table 2). One-hundred to 500
ng of genomic DNA was added to a final
reaction volume of 20 µL with 0.5 µM of
Table 2. Amplimer pairs used in PCR reactions for TNF microsatellites a-e, with specific annealing temperatures (Ta)
used during PCR.
Microsatellite
Amplimer pairs
Ta
TNFa
5’-GCCTCTAGATTTCATCCAGCCACA-3’
5’-CCTCTCTCCCCTGCAACACACA-3’
65°C
TNFb
5’-GCACTCCAGCCTAGGCCACAGA-3’
5’-GTGTGTGTTGCAGGGGAGAGAG-3’
60°C
TNFc
5’-GGTTTCTCTGACTGCATCTTGTCC-3’
5’-TCATGGGGAGAACCTGCAGAGAA-3’
57°C
TNFd
5’-AGATCCTTCCCTGTGAGTTCTGCT-3’
5’-CATAGTGGGACTCTGTCTCCAAAG-3’
64°C
TNFe
5’-GTGCCTGGTTCTGGAGCCTCTC-3’
5’TGAGACAGAGGATAGGAGAGACAG-3’
61°C

Page 4
JOP. J Pancreas (Online) 2006; 7(1):14-26.
JOP. Journal of the Pancreas – http://www.joplink.net – Vol. 7, No. 1 - January 2006. [ISSN 1590-8577]
17
each amplimer pair (reverse amplimer 5’; end
labeled with 50,000 cpm of
32
P), 10 mM Tris-
HCl (pH 9.0), 50 mM KCl, 2.5 mM MgCl2,
0.1% Triton X-100, 300 µM of each
deoxyribonucleoside 5'-triphosphates (dNTPs),
and 0.8 units of Taq DNA polymerase (HT
Biotechnology, Cambridge, UK). PCR was
performed in a cyclogene thermocycler
(Techne, Cambridge, UK) under the
following conditions: 95°C for 4 minutes,
then 30 cycles of 95°C for 2 minutes, 2
minutes at the specific annealing temperature
(Table 2), 62°C for 2 minutes, followed by an
extension period of 10 minutes at 68°C. Non-
radioactive PCR products underwent
electrophoresis on 1% agarose gel, to confirm
the presence of product of anticipated size.
Amplification products (6 µL) were then
mixed with 3 µL of formamide containing
Stop solution (Amersham Life Science,
Buckinghamshire, UK) and then separated on
a 6% polyacrylamide gel containing 8 M urea
at 1,700 V for 2.5 hours (TNFa, TNFc, TNFd,
TNFe) or 3 hours (TNFb). After drying, the
gels were exposed to Kodak XLS5 X-ray film
(Scientific Imaging Systems, Cambridge, UK)
with intensifying screens at -80°C for 18
hours.
Haplotype Designation
Alleles that combine in non-random
association across the TNF five locus
microsatellite system are termed haplotypes.
Haplotype studies are more powerful than
gene frequency studies [38]; therefore TNF
haplotype data have been presented rather
than a comparison of allele frequencies. These
TNF haplotypes have previously been
extensively characterized, by maximum
likelihood estimates in population and family
studies and their existence confirmed in
homozygous cell lines [39, 40]. Associations
between TNF-alpha and TNF-beta secretion
levels and the frequent Caucasian TNF five
locus microsatellite haplotypes have been
determined in macrophages and lymphoid
cells in response to a variety of stimuli,
Table 3. Definitions of TNF 5-locus haplotypes, with accompanying secretor status for TNF-alpha and TNF-beta,
whether stimulated by lipopolysaccharide (TNF-alpha), T-cell mitogens (TNF-beta) or constitutive B-cell secretion
(TNF-alpha). Adapted from Weissensteiner and Lanchbury [41].
TNF 5-locus haplotypes
Secretor status
TNFa TNFb TNFc TNFd TNFe
Lipopolysaccharide
T-cell mitogens
Constitutive B-cell
secretion
A
1
5
1
3
3
B
1
5
2
4
3
High
Low
C
2
1
2
4
1
High
Low
Low
D
2
3
1
1
3
High
High
High
E
2
5
2
4
3
High
Low
F
3
3
1
1
3
G
4
5
1
3
3
Low
H
4
5
1
4
3
I
4
7
2
5
3
High
J
5
5
2
5
3
High
K
6
5
1
3
3
Inter
Low
Low
L
6
5
1
4
3
M
6
5
1
5
3
Inter
N
6
5
1
7
4
O
7
1
2
2
3
P
7
4
1
3
3
Inter
Inter
Q
7
5
1
3
3
R
8
4
1
3
3
S
10
4
1
3
3
Low
Low
T
10
4
1
4
3
U
11
4
1
3
3
Low
Inter
Low

Page 5
JOP. J Pancreas (Online) 2006; 7(1):14-26.
JOP. Journal of the Pancreas – http://www.joplink.net – Vol. 7, No. 1 - January 2006. [ISSN 1590-8577]
18
including lipopolysaccharide and T-cell
mitogens [41]. TNF alleles at each of the five-
microsatellite loci were ascribed to a
haplotype, as previously defined [39, 40, 41],
for each sample. Table 3 provides the
definition of each haplotype and their
secretory status, under specified conditions.
ETHICS
Local Research Ethics Committee approval
was obtained. In accordance with the
American Society of Human Genetics
(ASHG) guidelines on genetic research [42],
informed consent was obtained from all
subjects in whom specimens were obtained
prospectively, i.e. those with pancreatitis and
those with alcoholic liver disease. Additional
clinical, pathological and demographic
information was obtained before subject
identifiers were removed and the samples
“anonymised”. The healthy control study
utilized retrospective samples collected
anonymously, in whom there was no
possibility, or need, to obtain consent.
STATISTICS
Comparison of haplotype frequencies
between patients and controls were made in
2x2 contingency tables using the chi-squared
method. Corrections were made, if necessary,
for small sample numbers using Fisher’s
exact test (i.e. where a cell in a 2x2
contingency table contained a number less
than 5). The odds ratio (OR) and the 95%
confidence interval (CI) were also evaluated.
Because a single hypothesis was being tested,
i.e. that our previously reported TNF
haplotype association with acute alcoholic
pancreatitis [34], would show a similar
association with chronic pancreatitis, a two
tailed P value of less than 0.05 was
considered statistically significant, without
further correction [43]. A recent review of
immunogenetics has confirmed the legitimacy
of this approach and states: “the alternative to
correction is to verify (replicate) in
subsequent studies” [44]. Ages were
compared by means of the Mann-Whitney U
test. The SPSS 8.0 for Windows was used to
analyse the data.
RESULTS
The TNF a6b5c1d3e3 Haplotype Is
Positively Associated with Chronic
Pancreatitis; the TNF a2b3c1d1e3 and
a2b5c2d4e3 Haplotypes are Negatively
Associated (Table 4)
The intermediate/low TNF secreting
haplotype a6b5c1d3e3 (K) was over-
represented in patients with chronic
pancreatitis compared to healthy controls. It
was present in 25 out of 104 (24.0%)
identified haplotypes in patients with chronic
pancreatitis compared with 24 out of 182
(13.2%) identified haplotypes among healthy
controls. The OR was 2.08 (95% CI: 1.12-
3.88; chi-squared=5.49, P=0.019). This
haplotype was specifically over-represented
in patients with alcoholic chronic pancreatitis
compared to healthy controls. It was present
in 19 out of 79 (24.1%) identified haplotypes
in patients with alcoholic chronic pancreatitis
compared with 24 out of 182 (13.2%)
haplotypes among healthy controls. The OR
was 2.08 (95% CI: 1.06-4.08; chi-
squared=4.72, P=0.030).
The high TNF secreting haplotype
a2b3c1d1e3 (D) was under-represented in
patients with chronic pancreatitis compared to
healthy controls. It was present in 11 out of
104 (10.6%) identified haplotypes in patients
with chronic pancreatitis compared with 36
out of 182 (19.8%) haplotypes among healthy
controls. The OR was 0.48 (95% CI: 0.23-
0.99; chi-squared=4.08, P=0.043). This
haplotype was specifically under-represented
in patients with alcoholic chronic pancreatitis
compared to healthy controls. It was present
in 8 out of 79 (10.1%) identified haplotypes in
patients with alcoholic chronic pancreatitis
compared with 36 out of 182 (19.8%)
haplotypes among healthy controls. The OR
was 0.46 (95% CI: 0.20-1.03; chi-
squared=3.66, P=0.056). Furthermore, a
similar high TNF secreting haplotype
a2b5c2d4e3 (E) was underrepresented in

Page 6
JOP. J Pancreas (Online) 2006; 7(1):14-26.
JOP. Journal of the Pancreas – http://www.joplink.net – Vol. 7, No. 1 - January 2006. [ISSN 1590-8577]
19
patients with alcoholic chronic pancreatitis
compared to alcoholic controls. It was present
in 3 out of 79 (3.8%) identified haplotypes in
patients with alcoholic chronic pancreatitis
compared with 9 out of 54 (16.7%)
haplotypes among alcoholic controls. The OR
was 0.20 (95% CI: 0.05-0.77; Fisher’s
P=0.014).
Low-Intermediate
TNF
Secreting
Haplotypes Are Positively Associated with
Chronic Pancreatitis; High Secreting
Haplotypes Are Negatively Associated
(Table 4)
The low-intermediate secreting TNF
haplotypes (K, M, P) were over-represented
in patients with chronic pancreatitis compared
to healthy controls. They were present in 34
out of 96 (35.4%) identified haplotypes in
patients with chronic pancreatitis compared to
38 out of 167 (22.8%) haplotypes in healthy
controls. The OR was 1.86 (95% CI: 1.07-
3.24; chi-squared=4.92, P=0.027). These
haplotypes were specifically over-represented
in patients with alcoholic chronic pancreatitis
compared to healthy controls. They were
present in 26 out of 73 (35.6%) identified
haplotypes in patients with alcoholic chronic
pancreatitis compared to 38 out of 167
(22.8%) haplotypes in healthy controls. The
OR was 1.88 (95% CI: 1.03-3.42; chi-
squared=4.30, P=0.038).
All high secreting TNF haplotypes (B, C, D,
E, I, J) were under-represented in patients
Table 4. TNF microsatellite haplotype frequencies.
Chronic pancreatitis
Overall
(n=73)
Alcoholic
(n=47)
Idiopathic
(n=23)
Hereditary
(n=3)
Alcoholic
liver disease
(n=39)
Alcoholic
controls
(n=29)
Healthy
controls
(n=103)
A
0
0
0
0
0
0
0
B
0
0
0
0
0
0
0
C
12.5%
13.9%
9.5%
0
13.1%
13.2%
13.7%
D
10.6% a
10.1% b
9.5%
20.0%
19.6%
13.2%
19.8% ab
E
3.8%
3.8% c
4.7%
0
9.8%
16.7% c
6.6%
F
0
0
0
0
0
0
0
G
5.8%
6.3%
4.7%
0
4.9%
0
4.9%
H
1.9%
1.3%
4.7%
0
3.2%
1.8%
1.6%
I
0
0
0
0
1.6%
1.8%
1.6%
J
1.9%
1.3%
4.7%
0
0
0
0.5%
K
24.0% d
24.1% e
14.3%
60.0%
13.1%
24.5%
13.2% de
L
1.9%
2.5%
0
0
1.6%
0
0.5%
M
1.9%
1.3%
4.7%
0
0
0
1.6%
N
0
0
0
0
0
0
0
O
0.9%
0
0
20.0%
0
0
0
P
6.7%
7.6%
4.7%
0
8.1%
5.6%
6.0%
Q
0
0
0
0
0
1.8%
1.0%
R
0.9%
1.3%
0
0
0
0
1.6%
S
6.7%
5.0%
19.0%
0
4.9%
5.6%
5.4%
T
1.9%
2.5%
0
0
0
0
3.3%
U
18.2%
18.9%
19.0%
0
19.6%
15.0%
18.1%
K, M, P
35.4% f
35.6% g
26.3%
75.0%
22.4%
32.7%
22.8% fg
B, C, D, E, I, J
31.3% h
31.5% i
31.5%
25.0%
46.5%
46.1%
46.1% hi
aHaplotype D: chronic pancreatitis vs. healthy controls; OR=0.48 (95% CI: 0.23-0.99, chi-squared=4.08, P=0.043)
bHaplotype D: alcoholic chronic pancreatitis vs. healthy controls; OR=0.46 (95% CI: 0.20-1.03, chi-squared=3.66, P=0.056)
cHaplotype E: alcoholic chronic pancreatitis vs. alcoholic controls; OR=0.20 (95% CI: 0.05-0.77; Fisher’s P=0.014)
dHaplotype K: chronic pancreatitis vs. healthy controls: OR=2.08 (95% CI: 1.12-3.88); chi-squared=5.49, P=0.019)
eHaplotype K: alcoholic chronic pancreatitis vs. healthy controls; OR=2.08 (95% CI: 1.06-4.08; chi-squared=4.72, P=0.030)
fHaplotypes K,M,P: chronic pancreatitis vs. healthy controls; OR=1.86 (95% CI: 1.07-3.24, chi-squared=4.92, P=0.027)
gHaplotypes K,M,P: alcoholic chronic pancreatitis vs. healthy controls; OR=1.88 (95% CI: 1.03-3.42, chi-squared=4.30, P=0.038)
hHaplotypes B,C,D,E,I,J: chronic pancreatitis vs. healthy controls; OR=0.53 (95% CI: 0.31-0.93, chi-squared=5.58, P=0.018)
iHaplotypes B,C,D,E,I,J: alcoholic chronic pancreatitis vs. healthy controls; OR=0.54 (95% CI: 0.30-0.96, chi-squared=4.46, P=0.035)
All other comparisons among groups were not significant.

Page 7
JOP. J Pancreas (Online) 2006; 7(1):14-26.
JOP. Journal of the Pancreas – http://www.joplink.net – Vol. 7, No. 1 - January 2006. [ISSN 1590-8577]
20
with chronic pancreatitis compared to healthy
controls. They were present in 30 out of 96
(31.3%) identified haplotypes in patients with
chronic pancreatitis compared to 77 out of
167 (46.1%) haplotypes in healthy controls.
The OR was 0.53 (95% CI: 0.31-0.93; chi-
squared=5.58, P=0.018). These haplotypes
were specifically under-represented in
patients with alcoholic chronic pancreatitis
compared to healthy controls. They were
present in 23 out of 73 (31.5%) identified
haplotypes in patients with alcoholic chronic
pancreatitis compared to 77 out of 167
(46.1%) haplotypes in healthy controls. The
OR was 0.54 (95% CI: 0.30-0.96; chi-
squared=4.46, P=0.035).
No Association between Idiopathic Chronic
Pancreatitis or Alcoholic Liver Disease and
TNF Haplotypes (Table 4)
Trends towards association of the alcoholic
liver disease patients with high secreting
microsatellite haplotypes were found, e.g. for
haplotype D (a2b3c1d1e3) when compared to
alcoholic pancreatitis, but these did not reach
statistically significant levels (chi-
squared=2.56, P=0.100).
No significant association between idiopathic
chronic pancreatitis and TNF haplotypes was
observed.
DISCUSSION
The TNF gene cluster is located within the
central MHC (class III region) on the short
arm of chromosome six. It spans 12 kilobases
(kb) and maps 250 kb centromeric to the
HLA-B locus and 850 kb to the class II region
[45]. The locus contains genes coding for
TNF, lymphotoxin alpha (previously TNF-
beta), lymphotoxin beta and leukocyte
specific transcript 1 (LST1) [46]. Two
clusters of single nucleotide polymorphisms
and five microsatellite markers are also found
within this locus [47]. Stable interindividual
variations exist in levels of production of
TNF, suggesting inherited individual
differences [25]. Family studies show that up
to 60% of the variability in TNF production
between individuals may be genetically
determined [26] and this has been correlated
with specific MHC alleles and ancestral
haplotypes [48]. An important feature of the
MHC is the strong linkage disequilibrium
between particular alleles across this region.
This has prompted speculation that the
association between high TNF production and
HLA-DR alleles may be related to
polymorphism within the TNF region itself.
Allele 2 of the TNF promoter polymorphism
at position -308 has been correlated with high
TNF production by peripheral blood
mononuclear cells [49] and a seven-fold
increase of TNF transcription in a CAT
reporter gene transiently transfected into a
human B cell line [50]. The highly
polymorphic TNF microsatellite markers may
be more informative, with high levels of TNF
production associated with TNFa2, TNFc2
and TNFd3 alleles and low levels with TNFa6
[51, 52]. Thus, the TNF response correlates
with a number of genetic markers all mapped
within a region with strong linkage
disequilibrium.
Linkage analysis has been used successfully
to find major gene effects but has limited
power to detect more modest effects.
Association studies, which utilize candidate
genes, have far greater power to delineate the
genetics of complex human diseases [53].
This approach has been widely used in
relation to the putative association of TNF
polymorphisms and susceptibility to various
diseases. In Crohn’s disease, the five
microsatellite haplotype TNFa2b1c2d4e1 is
overrepresented compared to patients with
ulcerative colitis and to healthy controls [30].
This haplotype is in linkage disequilibrium
with
HLA-DR1/DQ5.
A
different
microsatellite haplotype, TNFa6b5c1d3e, has
been described in rheumatoid arthritis [31]. In
contrast to Crohn’s disease, this association is
independent of HLA linkage. TNF
polymorphisms have also been associated
with susceptibility and outcome of various
infectious [27, 28, 29] and neoplastic [32, 33]
diseases.
In a previous study, we evaluated the
association of the TNF and IL-10 genes with

Page 8
JOP. J Pancreas (Online) 2006; 7(1):14-26.
JOP. Journal of the Pancreas – http://www.joplink.net – Vol. 7, No. 1 - January 2006. [ISSN 1590-8577]
21
acute pancreatitis and resultant disease
severity [34]. When patients were stratified
according to disease aetiology, the haplotype
TNFa2b1c2 showed a trend toward reduced
frequency in the group with alcohol induced
acute pancreatitis. TNFa2b1c2 was present in
4.8% of individuals with alcohol induced
acute pancreatitis compared with 24.3% of the
controls, 31.9% of patients with non-alcoholic
pancreatitis. However, this was not significant
after correction for multiple comparisons. The
finding of a genetic difference among
alcoholics who develop acute pancreatitis in
our original study suggested that this locus
may play an even more important role in the
development of chronic pancreatitis, where
alcohol plays an even more predominant role
as an etiological factor. Therefore, we
expanded our study to determine the
frequencies of TNF gene markers in a larger
group of patients with alcohol induced
chronic pancreatitis. The entry criteria for the
current study differs from our previous study.
For the original study of patients with acute
pancreatitis a consistent clinical history,
appropriate radiological evidence, and a
serum amylase level greater than three times
the upper limit of normal, were required.
Patients were then classified as having mild
disease or severe local or systemic disease
according to criteria defined by the Atlanta
convention. The entry criteria for the current
study of chronic pancreatitis was based upon
the classification system adopted at the Zurich
Workshop, 1996 [35] i.e. a typical history of
recurrent clinical acute pancreatitis was
required in addition to one or more of the
following: pancreatic calcification, ductal
lesions [36], steatorrhoea reduced by enzyme
supplementation and histology. The finding
that the a6b5c1d3e3 is positively associated
with chronic pancreatitis and the haplotypes
a2b5c2d1e3 and a2b5c2d1e3 are negatively
associated, confirms the association between
alcoholic pancreatitis and genetic markers of
reduced TNF production.
Other studies of the TNF gene and its putative
association with chronic pancreatitis include a
recent investigation that demonstrated that the
frequencies of the TNF promoter
polymorphisms at position -238 and -308 in
patients with alcoholic chronic pancreatitis,
idiopathic pancreatitis and those with the
N34S mutation of the SPINK 1 gene, did not
differ significantly from the control group.
However, the variant TNF-238A was a risk
factor for disease manifestation in families
with hereditary pancreatitis [54].
Histological examination of the normal
pancreas reveals that inflammatory cells are a
rare finding. In addition, MHC class I
antigens, which are expressed by most
nucleated somatic cells, are not present on
exocrine pancreatic cells. Normal acinar and
ductal pancreatic epithelial cells do not
express MHC class II determinants [55, 56].
In chronic pancreatitis, this is radically
altered. Lymphocytic infiltrates are common,
with the cytotoxic CD8+ T-cell subset
predominating [56, 57, 58]. Abnormal
expression of class I or class II MHC
determinants, or both, by exocrine epithelial
cells was demonstrated in 89% of ninety-three
cases of chronic pancreatitis [56]. Until
recently, the cellular infiltrate was not
considered
to
be
of
primary
pathophysiological importance but merely an
epiphenomenon consequent upon other
primary pathological changes in the pancreas.
Recent findings however, provide increasing
evidence that immunologic effector
mechanisms might be instrumental in the
induction of chronic pancreatitis. Hunger has
demonstrated that CD8+ T-cells are activated
in cellular infiltrates of alcoholic chronic
pancreatitis, using perforin mRNA as a
specific in vivo activation marker [59]. In
MHC class II deficient mice, an animal model
originally designed to study ulcerative colitis,
Vallance et al. unexpectedly observed the
development of an immune-based pancreatitis
with selective loss of exocrine cells and
function. Furthermore, CD8+ cells alone were
able to adoptively transfer the disease to
athymic mice and, in their absence, the
remaining cells were unable to induce any
pancreatic pathology [60]. Autoreactive T-
cells normally exist within the immune
system but are kept quiescent, or anergic, by
other regulatory T lymphocytes. In cases in

There are no products listed under this category.