| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Physiology and
* Division of Pediatric Surgery, The Johns Hopkins School of Medicine, Baltimore, Maryland 21205, USA
1Correspondence: Biophysics 203, The Johns Hopkins University School of Medicine, 725 N. Wolfe St., Baltimore, MD 21205, USA. E-mail: rreeves{at}welch.jhu.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: mapping • quantitative trait loci • inbred strains (A/J and C57BL/6J) • sepsis • multiple organ dysfunction syndrome
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and is characterized
clinically by a hypermetabolic state that arises as a consequence of an
exaggerated host inflammatory response to infection or to severe injury
(2
, 3)
. One detrimental consequence of sepsis is the
development of multiple organ dysfunction syndrome (MODS), the leading
cause of morbidity and mortality in intensive care units
(4)
. Despite improvements in the care of critically ill
patients during the past two decades, mortality rates from sepsis and
MODS remain unchanged. Development of better therapeutic strategies
requires a comprehensive understanding of the pathophysiological
mechanisms underlying these disorders, particularly that of the
inflammatory response.
The inflammatory response is characterized by a complex cascade of
cellular changes including the activation of phagocytic cells,
secretion of cytokines such as tumor necrosis factor (TNF) and
interleukins (IL) 1, 6, and 8, and the expression of cell surface
receptors like ICAM-1 on activated endothelial cells that guide the
adhesion and extravasation of polymorphonuclear leukocytes (PMN)
(5)
. In addition, different organ systems respond to these
changes by expressing various groups of genes such as the IL-1- and
IL-6-induced production of acute-phase proteins in the liver. Although
these events are beneficial to the individual in the early stages of
injury, an exaggerated or uncontrolled response can be detrimental.
Under these circumstances, the infiltration of PMN in major organ
systems like the lung and liver may result in secondary organ damage
due to the release of lytic enzymes and oxygen radicals
(6)
.
The genetic background of an affected individual may be an important
component in the regulation of inflammation. Epidemiological studies
have demonstrated that a familial clustering of susceptibility to
infectious agents to chronic inflammatory diseases, and to autoimmune
disease is due, in part, to genetic factors (7
, 8)
.
Polymorphic variants in the human TNF 
and ß loci that
produce higher amounts of circulating TNF have been associated with
higher mortality rates after severe sepsis (9)
. A
polymorphism in the IL1RA gene that results in high levels
of IL-1 receptor antagonist protein is a marker of susceptibility to
sepsis (10)
.
Studies performed in animal models support the assertion that there is
a genetic contribution to the inflammatory response. Differences in the
systemic inflammatory response to lipopolysaccharide (LPS)
administration among inbred strains of mice are well documented.
Extensive analyses have focused on LPS hyporesponsive C3H/HeJ and
C57BL/10SCr mice. It was recently discovered that these mice have a
mutation in the toll-like receptor 4 (Tlr4) gene, a
signal-transducing component of the cellular response to LPS (11
, 12)
. Other inbred strains also show defects in certain aspects
of the LPS-response, but do not have mutations in Tlr4. For
example, A/J mice were found to be resistant to LPS-induced mortality
at dosages where C57BL/6J (B6) mice were susceptible, and these
observations were correlated with differences in the inflammatory
response including the level of circulating TNF-, IL-1ß, IL-6, and
the expression of acute-phase proteins in the liver (13)
.
The degree of infiltrating leukocytes in the liver and lung also
differs between these strains after administration of LPS
(14)
. Several loci that control differential LPS-induced
proliferation of cultured splenocytes derived from A/J or B6 mice were
mapped in recombinant inbred (RI) and congenic strains
(15)
.
Intraperitoneal (i.p.) administration of LPS results in greater hepatic
PMN infiltration and ensuing necrosis in B6 than in A/J mice
(Fig. 1
), and this difference is highly significant (14)
.
F1 offspring from a cross between a B6 female and
an A/J male (B6AF1 mice) or from a cross between
an A/J female and a B6 male (AB6F1 mice) show a
PMN infiltration profile similar to B6 mice, indicating that genes
substantially contributing to this phenotype are neither sex linked nor
imprinted. To map genes that predispose to high or low levels of
endotoxin-induced hepatic PMN infiltration, linkage analysis was
performed using the AXB/BXA RI strain set, and a low-resolution genome
scan using 76 polymorphic markers evenly spaced throughout the genome
were used on a panel of 122 intercross animals.
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Phenotyping
Male mice, 8–9 wk of age, were injected i.p. with 15 mg/kg of
LPS derived from Escherichia coli O26:B6 (Difco, Detroit,
Mich.; manufacturer’s LD50=12.53 mg/kg)
resuspended in 0.9% NaCl (Abbott, Abbott Park, Ill.). Animals were
killed by cervical dislocation 24 h postinjection. Livers were
removed and fixed in 10% neutral buffered formalin. After being
mechanically rinsed, dehydrated, and embedded in paraffin, the tissue
was sectioned (5 µm) and stained with hematoxylin and eosin.
Neutrophil infiltration was quantified by counting the number of PMN in
10 high-power fields (h.p.f.; 400x magnification). Only fields that
were 90–95% liver parenchyma with nonfocal necrosis were selected.
The average number of PMN from the 10 fields was the infiltration score
for that animal. When slides from 20 animals were blinded and recounted
by the same investigator, the average difference in infiltration score
was found to be 3.35 where the mean score was 34 ± 17.5. All
experimental procedures were carried out in accordance with protocols
approved by the Johns Hopkins School of Medicine Animal Care and Use
Committee and in conformance with guidelines established in the
National Research Council’s Guide to the Care and Use of
Laboratory Animals.
Genotyping
DNA was extracted from tails of intercross progeny by
salting-out as described (16)
. PCR was carried out using
100 ng of template DNA. Primers for simple sequence length
polymorphisms (SSLP) that differ between A/J and B6 progenitor strains
were purchased from Research Genetics (Huntsville, Ala.) and used as
suggested by the manufacturer.
Linkage analysis
Linkage analysis was performed using Map Manager QT, v28
(17)
. For RI strains, infiltration scores from three
animals were averaged and entered as quantitative trait loci (QTL).
These were compared against the AXB/BXA set for QTL mapping
(18)
. Linkage and % variance accounted for by a QTL was
detected for the intercross set by comparing the infiltration scores of
122 progeny to the genotypes of those animals at 76 loci throughout the
genome. Suggestive, significant, and highly significant threshold
levels were determined by the permutation test function of Map Manager,
which is based in the statistical methods developed by Churchill and
Doerge (19)
. A likelihood ratio statistic (LRS) value of
4.6 is equivalent to one LOD. A two-LOD support interval was determined
for significant loci as the region at which the LRS value was within
9.2 units (two LOD units) of the peak value.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). The average infiltration score (PMN per h.p.f.) from each RI strain
was entered into Map Manager QT (17)
as a quantitative
trait, and linkage was evaluated against the edited AXB/BXA data set
(18)
. Suggestive linkage was detected at four loci. SSLP
marker D5Mit13 had the highest LRS, a measure of linkage.
The LRS score for this locus was 9.6, where 7.6 is suggestive and 12.8
is significant. The other markers that showed suggestive linkage in
this data set were D6Mit50, D9Mit16, and a
restriction fragment length polymorphism in the Tnf gene
that served as a marker on mouse chromosome (Chr) 17.
|
The relatively low statistical confidence in linkage yielded by the RI
analysis reflects, in part, the limited number of meioses represented
in RI sets. Although these results do not prove linkage, they do
suggest areas of the genome that should be investigated in greater
detail. Further examination of genes that contribute to the
differential PMN infiltration in B6 and A/J mice were pursued by
mapping in 122 intercross animals (244 meioses). The infiltration
scores for these animals were distributed continuously from 7 to 100,
consistent with a quantitative trait (Fig. 3
). Genotypes at 76 loci were determined in the intercross progeny
(Fig. 4
). Assuming a mouse genome length of 1500 cM, these 76 markers occur, on
average, every 20 cM. Therefore, any contributing QTL is no more
than 10 cM from the closest typed marker. Only one marker was typed on
the X chromosome because F1 data suggested that
major contributors are not sex linked (14)
. Analysis of
linkage using Map Manager QT (17)
detected significant
linkage at marker D13Mit17 on Chr 13 and suggestive linkage
with markers D1Mit318, D5Mit13,
D5Mit197, D8Mit80, D9Mit352, and
D16Mit181 (Fig. 4)
.
|
|
Refinement of QTL map positions
Six additional Chr 13 SSLP markers that map near
D13Mit17 were typed to confirm this linkage. Also, five Chr
5 markers that map between D5Mit13 and D5Mit197
were typed since D5Mit13 exhibited a suggestive linkage in
both the intercross and RI analyses. These genotypes helped refine the
localization of two contributing Hpi (hepatic
PMN infiltration) QTL (Fig. 5
). The linkage peak on Chr 13 was detected between markers
D13Mit236 and D13Mit248. The threshold value for
‘highly significant’ linkage corresponds to P<0.001 for
the genome scan (20)
. The two-LOD support interval around
Hpi1 is defined as the region where the LRS score is within
9.2 units of the peak value (see Materials and Methods) and spans 15
cM. The Hpi1 interval accounts for 16% of the total trait
variance. Using a constrained recessive regression model, highly
significant linkage (P<0.001) was detected at one of the
added markers, D13Mit88 (LRS=22.1).
|
The addition of markers on Chr 5 raised the suggestive LRS for this locus to a significant level (LRS=12.7) in the intercross analysis using either an additive or recessive model for regression. The two LOD support interval extends 34 cM from the proximal end of Chr 5. The peak LRS score was observed at D5Mit233. Hpi2 accounts for 8% of the total trait variance. This data set also defines the best marker map order for four SSLP markers to be D5Mit81-D5Mit233-D5Mit391-D5Mit394 rather than D5Mit391-D5Mit81-D5Mit233-D5Mit394, as reported in the composite Chr 5 map by Mouse Genome Informatics (http://www.jax.informatics.org).
Epistasis in predisposition to high PMN infiltration
The average infiltration scores seen in mice with each of the nine
possible combinations of genotypes at Hpi1 and
Hpi2 were compared to examine possible interactions between
these loci. A homozygous B6 genotype (B/B) at Hpi1 on Chr 13
was seen to contribute to a significantly higher overall infiltration
score (Table 1
, shaded cells), consistent with the recessive model that provided the
strongest evidence for linkage at this locus. The high infiltration
scores in Hpi1 B/B animals were unaffected by either an A/A
or A/B genotype at Hpi2 on Chr 5, but the scores increased
substantially (to an average of 69) when the Hpi2 genotype
was B/B. That is, the Hpi2 genotype had no apparent effect
on infiltration score in mice with any genotype at Hpi1
other than B/B, and only a B/B genotype at Hpi2 affected the
phenotype in mice that were B/B at Hpi1. This epistatic
interaction was also seen using the composite interval mapping function
of Map Manager QT. When the analysis was run to control for the effect
of Hpi1, linkage was no longer detected at Hpi2.
In contrast, significant linkage was still detected at Hpi1
when effects of a Chr 5 locus were controlled.
|
Simple inspection of genotypes showed that among the 122 intercross
progeny, 9 of the 14 with the highest infiltration scores were
homozygous for the B allele at both Hpi1 and
Hpi2. The lowest infiltration score seen in a double
homozygous B6-type mouse was 33 PMN/h.p.f., more than three standard
deviations higher than the A/J mean (13.8±5.3) and substantially
higher than any A/J mouse in the study (Fig. 3)
. Intercross progeny
with the lowest infiltration scores had a paucity of B6 alleles at
Hpi1 and showed the 1:2:1 distribution of genotypes at
Hpi2 expected for a locus with no involvement in the
phenotype. There was no obvious ‘protective’ combination of genotypes
at these two loci corresponding to the predisposition toward high PMN
infiltration in B6-type double homozygous progeny (data not shown).
Candidate gene analysis
D5Mit13, the marker detected by both RI and intercross
analysis, maps just distal to the gene encoding IL-6 (Il6),
an important cytokine mediator of the inflammatory response. The
contribution of Il6 to the LPS-induced PMN infiltration
response was evaluated in C57BL/6J-Il6tm1Kopf mice, which
have a homozygous deletion of the Il6 gene on a B6
background (21)
. Absence of Il6 had no effect
on LPS-induced PMN infiltration. Mean infiltration scores for five B6
and five C57BL/6J-Il6tm1Kopf were 43.5 ± 15.1 and
46.7 ± 13.9, respectively.
The region of suggestive linkage on Chr 17 determined by RI analysis
includes the major histocompatibility complex, H2, and the
proinflammatory cytokine Tnf. A/J and B6 mice are known to
have allelic variants at the Tnf locus, which have been
associated with different levels of TNF- in several strains of mice
(22
, 23)
. The possible contribution to the infiltration
phenotype of H2, Tnf, or another gene in this
region of Chr 17 was assessed using congenic mice. The region of Chr 17
in B10.A-H2a-H2-T18a/SgSnJ congenic mice includes
the H2 complex and Tnf (22)
, and is
derived from the inbred strain A/WySnJ, a low responder in the hepatic
PMN infiltration assay like A/J. The remainder of the genome of this
congenic strain is derived from C57BL/10SnJ, which demonstrated a high
level of LPS-induced PMN infiltration like B6. These congenic mice
showed high PMN infiltration after LPS, suggesting that H2
and Tnf are not pivotal to this response (Table 2
). The PMN phenotype of the reciprocal congenics,
A.BY-H2b-H2-T18b/SnJ, supports this conclusion.
These mice have a B-derived MHC on an A background and show the low PMN
infiltration response to LPS. Thus, allelic differences of genes that
map within the MHC do not correlate with PMN infiltration phenotypes.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The endotoxin-induced
infiltration often results in liver damage, and this hepatotoxicity can
be significantly reduced in experimental animal models if neutrophils
are depleted before endotoxin is administered (25)
.
Antibody therapies have been devised to reduce liver damage during a
systemic inflammatory response. Most of these strategies try to
eliminate the interaction between infiltrating PMN and endothelial
cells (26
27
28)
or the downstream effectors of the LPS
signaling pathway (29)
. These preventative therapies
against endotoxin-induced damage have been tested in animal models with
varying success. The mouse system provides a powerful tool for unraveling the complex interaction of genes and environment in the inflammatory response. Whereas an inflammatory response arises as a consequence of stochastic events in humans, the mouse system allows for controlled initiation of this response in animals of the same age that have been housed under identical pathogen-free conditions since birth. The use of inbred strains defines and limits the sources of genetic variation, providing a greatly simplified set of polymorphic alleles whose interactions are accordingly observed much more readily than in an outbred population. Genetic approaches using mouse are further facilitated by unique genetic tools such as RI strains, congenic strains, and mice with targeted mutations.
Two loci on Chr 5 and 13 that contribute to the differential PMN infiltration response in A/J and B6 mice were identified through RI mapping and a low-resolution genome scan of 122 intercross animals. The support intervals around QTL identified in this first level of mapping are broad and include too many candidate genes for a comprehensive mutation screen. On the other hand, the Chr 13 and Chr 5 2-LOD support intervals each exclude more than 97% of mouse genes as candidates. These significant linkages and the several suggestive linkages identified in the first stage analysis focus future mapping efforts using a high-resolution intercross panel (in progress). Candidate genes from the Hpi1 and Hpi2 loci with particular biological relevance to the PMN infiltration phenotype were assessed by examination of congenic and gene-targeted mice. These tools rapidly and definitively eliminated these candidates, focusing interest on other genes in the target regions.
The demonstration in congenic strains that IL-6 is not involved in
hepatic PMN recruitment after endotoxin exposure is interesting because
it had been reported previously that IL-6 was associated with
neutrophil recruitment and lung injury in hemorrhagic shock in mice
(30)
, with PMN infiltration and ocular injury in
endotoxin-induced uveitis in the rat (31)
and with
activation of leukocytes and severity of organ dysfunction in the
systemic inflammatory response syndrome in humans (32)
.
Our results show identical degrees of hepatic PMN infiltration in the
presence or absence of the Il6 gene, indicating that IL-6 is
dispensable in this aspect of the inflammatory response. Others have
observed that LPS-induced lethality was unaffected by the presence of
functional IL-6 (33
, 34)
. Considered together, these data
suggest that the role, if any, played by IL-6 in this hepatic PMN
infiltration phenotype is subtle or redundant.
Endothelin-1 (Edn1) maps close to D13Mit88, the
marker that showed highly significant linkage at Hpi1 on Chr
13. Edn1 is a vasoconstrictor produced by endothelial cells
that can act as an autocrine factor to stimulate the production of
nitric oxide and prostacyclin, important downstream effectors of the
LPS response (35)
. Further, endotoxemia has been shown to
increase hepatic endothelin-1 production (36)
. This could
result in vasoconstriction and ensuing ischemia/reperfusion in the
liver, which would exacerbate PMN infiltration and liver damage in B6
mice. Thus, functional allelic differences of Edn1 in A/J
and B6 mice could contribute to the phenotype described here.
Clues to other plausible candidates come from the Lpsd
allele, which encodes a missense mutation in Tlr4 (11
, 12)
. This component of the LPS signaling pathway is a member of
the same gene family as the IL-1 receptor and is also involved in the
activation of proinflammatory genes (37
, 38)
. The key role
of Tlr4 in the LPS response illustrates the possibility that
mutations affecting elements of the signal transduction cascade
initiated by this or other LPS receptors may play important roles in
variable response to endotoxemia and, hence, are candidate genes for
the phenotype examined here.
In this study, the contribution of the Hpi1 locus was shown
to be epistatic to Hpi2. Hpi2 had no apparent affect on the
LPS-induced hepatic PMN infiltration phenotype unless it was homozygous
for the B6-type allele in the presence of a homozygous B6 phenotype at
Hpi1. All intercross mice with B/B genotypes at both
Hpi1 and Hpi2 had high infiltration scores in
response to endotoxin, including most of the highest infiltration
scores among the 122 intercross progeny. This prediction held when RI
genotypes at Hpi1 and Hpi2 were compared to PMN
infiltration scores in AXB/BXA RI strains (Fig. 1)
. The four strains
with B/B, B/B genotypes showed 63 ± 14.8 PMN/h.p.f. as compared
to 44 ± 12 (A/A, A/A), 37.2 ± (B/B, A/A), and 37.1 ±
9.8 (A/A, B/B). Thus, ascertainment of these two genotypes can identify
a subset of mice that will exhibit a strong inflammatory response.
The endotoxin-induced PMN infiltration phenotype seen in mice is directly analogous to the pathology of a number of human diseases characterized by an exaggerated inflammatory response, and therefore the findings of this study have important implications for the management of human patients. A high level of PMN infiltration in response to endotoxin can be predicted based only on a knowledge of genotypes at two loci. Our observations validate such a prediction only in the context of the limited genetic variability between these specific inbred strains. However, these results clearly demonstrate the basis for identifying a larger subset of the repertoire of polymorphic genes that affect predisposition to exaggeration or attenuation of specific aspects of the inflammatory response manifested in mice. Comparative mapping provides a means of extending these observations to human beings, with the possibility that individuals predisposed to an exaggerated inflammatory response could be identified by genetic testing at the time of admission to a surgical intensive care unit, prior to the onset of disease.
|
ACKNOWLEDGMENTS |
|---|
Received for publication January 3, 2000.
Revision received April 28, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. Laragione, N. C. Yarlett, M. Brenner, A. Mello, B. Sherry, E. J. Miller, C. N. Metz, and P. S. Gulko The Arthritis Severity Quantitative Trait Loci Cia4 and Cia6 Regulate Neutrophil Migration into Inflammatory Sites and Levels of TNF-{alpha} and Nitric Oxide J. Immunol., February 15, 2007; 178(4): 2344 - 2351. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. B. Fulton, R. H. Reeves, M. Takeya, and A. De Maio A Quantitative Trait Loci Analysis to Map Genes Involved in Lipopolysaccharide-Induced Inflammatory Response: Identification of Macrophage Scavenger Receptor 1 as a Candidate Gene J. Immunol., March 15, 2006; 176(6): 3767 - 3773. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. I. Rutitzky, H. J. Hernandez, Y.-S. Yim, D. E. Ricklan, E. Finger, C. Mohan, I. Peter, E. K. Wakeland, and M. J. Stadecker Enhanced Egg-Induced Immunopathology Correlates With High IFN-{gamma} in Murine Schistosomiasis: Identification of Two Epistatic Genetic Intervals J. Immunol., January 1, 2005; 174(1): 435 - 440. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Del Pozzo, D. Mascolo, A. Prisco, P. Barba, A. Anzisi, and J. Guardiola Lack of patent liver autoimmunity after breakage of tolerance in a mouse model Int. Immunol., October 1, 2003; 15(10): 1173 - 1181. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
