(Chest. 2002;121:21S-25S.)
© 2002
American College of Chest Physicians
Uses of Expression Microarrays in Studies of Pulmonary Fibrosis, Asthma, Acute Lung Injury, and Emphysema*
Roger S. Mitchell Lecture
Dean Sheppard, MD
*
From the Lung Biology Center, Center for Occupational and Environmental Health, Cardiovascular Research Institute, Department of Medicine, University of California, San Francisco, San Francisco, CA.
Correspondence to: Dean Sheppard, MD, Lung Biology Center, University of California, San Francisco, Box 0854, San Francisco, CA 94143; e-mail: deans{at}itsa.ucsf.edu
 |
Abstract
|
|---|
Expression microarrays are a powerful tool that could provide new information about
the molecular pathways regulating common lung diseases. To exemplify
how this tool can be useful, selected examples of informative
experiments are reviewed. In studies relevant to asthma, the cytokine
interleukin-13 has been shown to produce many of the phenotypic
features of this disease, but the cellular targets in the airways and
the molecular pathways activated are largely unknown. We have used
microarrays to begin to dissect the different transcriptional responses
of primary lung cells to this cytokine. In experiments designed to
identify global transcriptional programs responsible for regulating
lung inflammation and pulmonary fibrosis, we performed microarray
experiments on lung tissue from wild-type mice and mice lacking a
member of the integrin family know to be involved in activation of
latent transforming growth factor (TGF)-ß. In addition to identifying
distinct cluster of genes involved in each of these processes, these
studies led to the identification of novel pathways by which TGF-ß
can regulate acute lung injury and emphysema. Together, these examples
demonstrate how careful application and thorough analysis of expression
microarrays can facilitate the discovery of novel molecular targets
for intervening in common lung diseases.
Key Words: acute lung injury • integrins • microarrays • pulmonary fibrosis
Expression microarrays are a recently
developed method that should soon allow pulmonary investigators to
simultaneously evaluate messenger RNA concentrations for every
transcript from a given genome simultaneously in a single
experiment.1
Although this approach has all of the
limitations of classical methods for measuring RNA abundance
(eg, reverse transcriptase-polymerase chain reaction,
Northern blotting, or ribonuclease protection assays), it has
the major advantage of eliminating the need to design specific reagents
and conditions for each target of interest, one at a time. Because
expression arrays are currently expensive, and because the results of a
single experiment can generate enormous amounts of data, thoughtful
experimental design and improved methods for data analysis are likely
to be critical determinants of how successful investigators are in
applying this new tool. In this review are discussed a few examples of
how we have utilized expression microarrays to develop a few new
insights into the molecular pathways underlying common lung diseases.
|
Effects of Interleukin-13 ON
PRIMARY LUNG CELLS
|
|---|
T lymphocytes of the T-helper 2 (Th2) subtype have been
shown to be prominent in the airways of patients with
asthma.2
In murine models, T-cell transfer
studies3
have demonstrated that cells differentiated into
Th2 cells in vitro are sufficient to induce most of the
phenotypic features of allergic asthma (ie, airway
responsiveness, mucus metaplasia, and eosinophilic inflammation). The
principal product of Th2 cells responsible for inducing all of these
features during the effector phase of the immune response appears to be
interleukin (IL)-13.4
5
IL-13 modifies cell behavior by
activating the signal transducer and activator of transcription
(STAT)-6, which translocates to the nucleus and regulates transcription
of target genes.6
However, the cellular targets of IL-13
responsible for each of the phenotypic features of asthma and the
specific genes whose expression is regulated in these target cells
remain largely unknown. Since IL-13 works in large measure by
regulating gene expression, these questions seem ideally suited to
expression array experiments.
As a first step to address these questions, Lee et al7
utilized Affymetrix hu6500 GeneChips (Affymetrix; Santa Clara,
CA) to evaluate expression of approximately 6,500 genes in
primary cultures of airway epithelial cells, airway smooth-muscle
cells, and lung fibroblasts 6 h after addition of either IL-13 or
an equivalent volume of phosphate-buffered saline solution to
subconfluent cultures. First, to determine whether each cell type was
capable of activating the same STAT-6 signaling pathway, they performed
STAT-6 immunoprecipitations followed by antiphosphotyrosine Western
blot tests and demonstrated that STAT-6 was phosphorylated in response
to IL-13 in each case. Surprisingly, however, despite activation of the
same canonical signaling pathway, the genes induced and inhibited by
IL-13 were virtually nonoverlapping in these cells. In fact, although
several hundred genes were induced in at least one cell type, there was
not a single gene that met our criteria for induction in all three cell
types (Fig 1
).
One clue to why the pattern of gene expression was so different came
from evaluation of the transcription factors induced in each cell type.
Although transcription factors were among the most prominently induced
genes, completely distinct groups of transcription factors were induced
in each cell type.7
The most prominent genes induced in
airway smooth-muscle cells were signaling effectors and receptors and
contractile proteins. Different signaling receptors and effectors were
prominently induced in fibroblasts, suggesting that IL-13 might be
priming each of these cell types to respond to other signals. In
epithelial cells, the most prominent genes induced were components of
the extracellular matrix.
>In summary, these studies demonstrated a striking effect of
cellular differentiation in determining the transcriptional response to
IL-13 in primary lung cells. The patterns of gene expression of each
cell type provide some clues to the likely in vivo cellular
and molecular targets of IL-13 that could contribute to allergic
asthma. Of course, an important caveat is that these responses may not
fully reflect the in vivo transcriptional responses because
of the well-known effects of in vitro culture conditions on
cellular differentiation.
|
Combined Use of Gene Knockouts and Expression Arrays to Identify
in Vivo Pathways Involved in Acute Lung Injury,
Pulmonary Fibrosis, and Emphysema
|
|---|
The development of lines of mice expressing homozygous null
mutations of individual genes provides an opportunity to utilize
expression microarrays to identify molecular pathways downstream of
these inactivated genes. This approach is especially attractive when
applied to knockout lines with disease-related phenotypes because it
has the potential to identify previously unexpected features of
molecular pathogenesis. For example, several years ago we generated a
line of mice lacking expression of the epithelial integrin,
vß6.8
These mice are dramatically protected from
pulmonary fibrosis despite developing enhanced pulmonary
inflammation.8
9
Based on in vitro
observations, we suspected that both the protection from fibrosis and
enhancement of inflammation were a consequence of loss of a normal
pathway for activation of latent transforming growth factor
(TGF)-ß.9
Expression array experiments on lung tissue
from these mice, at baseline and after treatment with the
fibrosis-inducing drug, bleomycin, have provided both expected and
surprising insights into the molecular mechanisms underlying pulmonary
fibrosis, emphysema, and acute lung injury.10
With regard to pulmonary fibrosis, cluster analysis based on pairwise
comparisons of gene expression after saline solution or at two time
points after bleomycin in wild-type and ß6 knockout mice identified
two distinct clusters of genes whose expression was increased in
response to bleomycin.10
One cluster consisted of 63 genes
that were expressed at higher levels at baseline in the knockout mice
but were expressed at similar levels in both strains after bleomycin.
We reasoned that these genes might be regulators of lung inflammation,
since the knockout mice had substantial inflammation, but there was a
robust inflammatory response to bleomycin in both strains. This
hypothesis turned out to be substantially correct, since 38 of these
genes were known regulators of inflammation. The other 25 genes in this
cluster are thus candidate regulators of inflammation. The second
cluster consisted of 66 genes that were expressed at the same level at
baseline, but were preferentially induced by bleomycin in wild-type
mice. We reasoned that these genes might encode regulators of the
fibrotic response, since wild-type mice were substantially more
susceptible to pulmonary fibrosis. Again, the genes in this group
encoded 42 proteins that were known matrix components, matrix response
genes, or known mediators of tissue remodeling, suggesting that our
hypothesis was substantially correct. Interestingly, most of the known
TGF-ß–inducible genes on the arrays we used were included in this
cluster, providing support for the hypothesis that ß6 knockout mice
are protected from pulmonary fibrosis as a consequence of failure to
activate TGF-ß.
Further insights into a previously unexpected in vivo
pathway came from examination of the differences in baseline gene
expression between ß6 knockout and wild-type mice. From scattergram
analysis, a small group of only four genes stood out as most
dramatically induced in the lungs of ß6 knockout mice (Fig 2
). Of these, the most highly induced was the gene encoding matrix
metalloelastase (matrix metalloproteinase [MMP]-12).
MMP-12 was of interest since it is a macrophage-restricted protease
that has been strongly implicated in the induction of emphysema in
mice.11
Quantitative polymerase chain reaction of RNA from
alveolar macrophages obtained from wild-type or ß6 knockout mice
confirmed that MMP-12 messenger RNA concentration was increased > 200
fold in macrophages from ß6 knockout animals. We therefore wondered
whether loss of vß6 might contribute to the development of
emphysema through induction of this metalloprotease. Indeed, when we
performed quantitative morphometry of aging ß6 knockout mice, it
became clear that these animals had a progressive increase in alveolar
diameter characteristic of emphysema. This effect was clearly a
consequence of MMP-12 induction, because when we crossed ß6 knockout
mice onto an MMP-12 knockout background the induction of emphysema was
completely abolished. This effect, like the role of this integrin in
the development of pulmonary fibrosis, is likely to depend on
vß6-mediated activation of TGF-ß, since TGF-ß is an extremely
potent inhibitor of MMP-12 induction in
vitro.12
13
View larger version (40K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2. Mean values for expression of approximately 6,000
murine genes (arbitrary units) in lungs from ß6 knockout (-/-) or
wild-type (+/+) mice. Genes whose values were within twofold of each
other in lungs from each strain of mice are omitted.
|
|
Additional insights into the in vivo functions of the
integrin vß6 and TGF-ß itself came from analysis of the time
course of induction of these TGF-ß–inducible genes. Utilizing a
method called self-organizing maps, we identified a small subcluster of
11 genes that were all preferentially induced by bleomycin in wild-type
mice and were all induced with a similar time course.10
All 11 of these genes were known to be TGF-ß inducible.
Interestingly, all of these genes were already substantially induced by
2 to 5 days after treatment with bleomycin, a time point well in
advance of the first detectable fibrosis. In addition to causing
fibrosis, bleomycin is a potent cause of acute lung injury in mice (and
humans), and this effect is maximal approximately 5 days after
treatment.14
We therefore wondered whether TGF-ß itself
might participate in induction or modulation of the pulmonary edema
that characterizes acute lung injury. To examine this possibility, we
first compared bleomycin-induced pulmonary edema in wild-type and ß6
knockout mice and found that ß6 knockout mice were completely
protected from bleomycin-induced pulmonary edema, despite the expected
enhancement of the acute inflammatory response to
bleomycin.14
Although the finding that ß6 knockout mice were protected from
pulmonary edema suggested a possible role for TGF-ß as an effector in
this process, we could not exclude an unrelated protective effect of
loss of this integrin, for example as a compensatory response to
chronic inflammation. To more directly examine the role of TGF-ß
itself, we examined the effects of a potent TGF-ß inhibitor (a
chimeric molecule composed of the extracellular domain of the
high-affinity TGF-ßII receptor fused to the Ig Fc
domain15
16
) on wild-type mice treated with bleomycin.
This chimeric inhibitor also completely prevented bleomycin-induced
pulmonary edema, thus directly implicating TGF-ß in this
process.14
Furthermore, this effect was not limited to
pulmonary edema induced by bleomycin, since the TGF-ß–receptor
chimera also prevented endotoxin-induced pulmonary edema in wild-type
mice. Thus, based on an unexpected pattern of gene expression in
response to bleomycin, we were able to identify a novel effector of
pulmonary edema in acute lung injury and have identified the integrin
vß6 and TGF-ß itself as potential therapeutic targets for
improving the treatment of this largely untreatable group of disorders.
|
Pilot Study of Human Pulmonary Fibrosis
|
|---|
>In contrast to the lung injury and pulmonary fibrosis
induced by bleomycin in genetically identical mice, human pulmonary
fibrosis is etiologically and temporally heterogeneous and obviously
occurs in people who are genetically heterogeneous. Interpretation of
microarray analysis of tissue samples obtained from such patients is
thus clearly more challenging and likely to pose greater problems in
distinguishing true signals from noise. Nonetheless, we have begun
pilot studies to assess the utility of this approach. In our first
effort, lung biopsy samples from five patients with pathologic features
consistent with usual interstitial pneumonitis were analyzed and
compared to samples from three resected lungs with normal histologic
findings and RNA from a pool of five normal lungs. Because of the
heterogeneity of these samples, a critical first step was deciding
which genes were differentially expressed in a meaningful fashion. For
this determination, we utilized methods specially designed for this
purpose to generate an "information score" for each of the 8,400
genes being analyzed. The first method consisted of first
computationally determining, for each gene, the optimal value that
would separate control from experimental (patient) values, and then
counting the number of values that were "misclassified" into the
wrong group. A second method involved mathematically determining a
Gaussian distribution curve for patient and control values for each
gene and then calculating the overlap between these distributions. With
these approaches we identified 164 genes that were likely to be
informative in this data set. Encouragingly, many of these overlapped
with the genes present in the "fibrosis cluster" from our studies
of murine pulmonary fibrosis. Of these, the individual gene with the
highest information score was the metalloprotease martilysin (MMP-7).
>MMP-7 is of interest because it has been reported to be involved
in a number of processes that go beyond its role in degrading
components of the extracellular matrix, including activation of tumor
necrosis factor-17
and generation of soluble Fas
ligand,18
two effects that could be predicted to enhance
pulmonary fibrosis. We therefore examined the role this protein played
in fibrosis in more detail. Immunostaining demonstrated dramatic
induction of MMP-7 in epithelial cells overlying fibroblastic foci in
other patients with usual interstitial pneumonitis. Furthermore, MMP-7
knockout mice, on two different genetic backgrounds, were substantially
protected from bleomycin-induced pulmonary fibrosis. Thus, microarray
analysis, even on a small number of samples from patients with
pulmonary fibrosis, successfully identified at least one unexpected
protein that appears to contribute to pathogenesis of this disease.
|
Future Directions
|
|---|
>Despite the encouraging results of the initial studies
described above, it is clear that optimal application of microarray
technology to the study of diseases of complex organs like the lung
will be limited by spatial and temporal heterogeneity of disease, and
by dramatic differences in cellular composition of affected and
unaffected tissue. It will therefore be critical to develop improved
methods for unbiased amplification of small RNA samples so that
meaningful information can be obtained by utilizing microarrays on
small tissue samples and pure cell populations (eg, samples
obtained by microdissection of tissue sections). The combination of
these approaches with proteomic analysis and follow-up functional
evaluation of identified candidates is likely to greatly accelerate our
understanding of molecular pathogenesis over the next several years.
|
Footnotes
|
|---|
Abbreviations: IL = interleukin; MMP = matrix
metalloproteinase; STAT = signal transducer and activator of
transcription; TGF = transforming growth factor; Th2 = T-helper 2
|
References
|
|---|
-
Albelda, SM, Sheppard, D (2000) Functional genomics and expression profiling: be there or be square. Am J Respir Cell Mol Biol 23,265-269[Free Full Text]
-
Robinson, DS, Hamid, QW, Ying, S, et al (1992) Predominant Th2-like bronchoalveolar T-lymphocyte population in atopic asthma. N Engl J Med 326,298-304[Abstract]
-
Cohn, L, Tepper, JS, Bottomly, K (1998) IL-4-independent induction of airway hyperresponsiveness by Th2, but not Th1, cells. J Immunol 161,3813-3816[Abstract/Free Full Text]
-
Grunig, G, Warnock, M, Wakil, AE, et al (1998) Interleukin 13 and interleukin 4 mediate the asthmatic phenotype through interleukin 4 receptor . Science 282,226-228
-
Wills-Karp, M, Luyimbazi, J, Xu, X, et al (1998) Interleukin-13: central mediator of allergic asthma. Science 282,2258-2261[Abstract/Free Full Text]
-
Lin, J-X, Migone, T-S, Friedman, M, et al (1995) The role of shared receptor motifs and common Stat proteins in the generation of cytokine pleiotrophy and redundancy by IL-2, IL-4, IL-7, IL-13, and IL-15. Immunity 2,331-339[CrossRef][ISI][Medline]
-
Lee, JH, Kaminski, N, Dolganov, G, et al (2001) Interleukin-13 induces dramatically different transcriptional programs in three human airway cell types. Am J Respir Cell Mol Biol 25,474-485[Abstract/Free Full Text]
-
Huang, XZ, Wu, JF, Cass, D, et al (1996) Inactivation of the integrin ß-6 subunit gene reveals a role of epithelial integrins in regulating inflammation in the lung and skin. J Cell Biol 133,921-928[Abstract/Free Full Text]
-
Munger, JS, Huang, XZ, Kawakatsu, H, et al (1999) The integrin vß6 binds and activates latent TGFß1: a mechanism for regulating pulmonary inflammation and fibrosis. Cell 96,319-328[CrossRef][ISI][Medline]
-
Kaminski, N, Allard, J, Pittet, J-F, et al (2000) Global analysis of gene expression in pulmonary fibrosis reveals distinct programs regulating lung inflammation and remodeling. Proc Natl Acad Sci U S A 97,1778-1783[Abstract/Free Full Text]
-
Hautamaki, RD, Kobayashi, DK, Senior, RM, et al (1997) Requirement for macrophage elastase for cigarette smoke-induced emphysema in mice. Science 277,2002-2004[Abstract/Free Full Text]
-
Werner, F, Jain, MK, Feinberg, MW, et al (2000) Transforming growth factor-ß1 inhibition of macrophage activation is mediated via Smad3. J Biol Chem 275,36653-36658[Abstract/Free Full Text]
-
Feinberg, MW, Jain, MK, Werner, F, et al (2000) Transforming growth factor-ß 1 inhibits cytokine-mediated induction of human metalloelastase in macrophages. J Biol Chem 275,25766-25773[Abstract/Free Full Text]
-
Pittet, J-F, Griffiths, MJD, Geiser, T, et al (2001) TGFß is a critical mediator of acute lung injury. J Clin Invest 107,1529-1536[ISI][Medline]
-
Wang, Q, Wang, Y, Hyde, DM, et al (1999) Reduction of bleomycin induced lung fibrosis by transforming growth factor beta soluble receptor in hamsters. Thorax 54,805-812[Abstract/Free Full Text]
-
Smith, JD, Bryant, SR, Couper, LL, et al (1999) Soluble transforming growth factor-ß type II receptor inhibits negative remodeling, fibroblast transdifferentiation, and intimal lesion formation but not endothelial growth. Circ Res 84,1212-1222[Abstract/Free Full Text]
-
Haro, H, Crawford, HC, Fingleton, B, et al (2000) Matrix metalloproteinase-7–dependent release of tumor necrosis factor- in a model of herniated disc resorption. J Clin Invest 105,143-150[ISI][Medline]
-
Mitsiades, N, Yu, WH, Poulaki, V, et al (2001) Matrix metalloproteinase-7-mediated cleavage of Fas ligand protects tumor cells from chemotherapeutic drug cytotoxicity. Cancer Res 61,577-581[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
E. K. Chu and J. M. Drazen
Asthma: One Hundred Years of Treatment and Onward
Am. J. Respir. Crit. Care Med.,
June 1, 2005;
171(11):
1202 - 1208.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
Abstract
Effects of Interleukin-13 ON...
