(Chest. 2002;121:42S-44S.)
© 2002
American College of Chest Physicians
Application of Expression Profiling to the Developing Lung*
Thomas A. Neff Lecture
Thomas J. Mariani, PhD and
Steven D. Shapiro, MD, FCCP
*
From the Departments of Pediatrics, Medicine, Cell Biology and Physiology, and the Program in Lung Development, Washington University School of Medicine and St. Louis Children’s Hospital, St. Louis, MO.
Correspondence to: Stephen D. Shapiro, MD, FCCP, Pulmonary and Critical Care Division, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, 75 Francis St, Boston MA; e-mail: sshapiro{at}rics.bwh.harvard.edu
 |
Abstract
|
|---|
If
we hope to repair damaged lung tissue associated with a variety of
acquired and developmental diseases, we must first gain a full
appreciation of normal lung development. As an approach, we have
utilized Affymetrix (Santa Clara, CA) high-density,
oligonucleotide-based microarrays to generate an expression profile of
the entire process of rodent lung development, which will be made
publicly available. Our initial results were internally
consistent and correlated closely with those generated with standard
expression techniques such as Northern hybridization. We have verified
known expression of genes, found other genes with previously
unsuspected expression during lung development, as well as uncovered
many expressed sequence tags whose role in lung development awaits
further study. Data mining reveals close relationships of expression
profiles between specific genes, suggesting novel regulatory
relationships. In the future, application of these methods to the study
of gene-targeted mice with abnormal lung development should uncover
pathways of airway and alveolar development. Ultimately, expression
profiling of diseased lungs might allow us to understand why the lung
fails to repair, and strategies to influence repair might become
apparent.
Genome-wide expression analysis allows the
interrogation of many if not all messenger RNAs expressed in a given
cell or tissue. Improvements in technology and widespread use have made
this technique affordable for individual laboratories. These types of
experiments challenge the hypothesis-driven method that investigators
traditionally use to approach biological problems. In contrast, this
process-driven approach should lead to a variety of testable hypotheses
without suffering from difficulties in predicting candidate genes. This
technology has largely been applied as a discriminate tool to determine
differences between two conditions. For example, one can determine
differences in gene expression in a cell type with and without
treatment, or one can compare normal and diseased tissue.1
Examination of tissues as opposed to cells has additional complexity in
accounting for different cell types associated with different samples.
Another application of expression profiling is to compare gene
expression over time.2
Here, the technology can be used as
a correlative tool in order to identify genes with similar expression
patterns across the time series. Genes sharing similarities in
expression often are functionally related.
Publishing databases of large microarray studies in traditional journal
format is difficult. Indeed, as more expression databases are
constructed, communication of results will become a major issue.
Traditional journals struggle with the nonhypothesis-driven approaches
and incomplete description of massive amounts of data. In-depth
investigation into specific aspects of the data will conform to usual
types of publication. In lieu of or in addition to traditional
publication, young investigators that immerse themselves in these
altruistic tasks must be rewarded in academic circles in other ways.
For example, in addition to publications, Web sites (and number of hits
or major advances based on the data) should be incorporated into
curriculum vitas.
|
Methods
|
|---|
Two general formats are widely used. Complementary DNA (cDNA)
microarrays contain long nucleic-acid probes immobilized on membranes
or glass slides. The distinguishing features of cDNA microarrays are
that the expression values are based on the competitive hybridization
of two samples being directly compared, and a single hybridization
event for each gene/probe. Conversely, oligonucleotide-based
microarrays utilize a noncompetitive strategy, where each sample is
hybridized independently. This technology is dominated by Affymetrix
Inc. (Santa Clara, CA) and their GeneChip arrays. This is the
technology that will be the focus of this report.
Technically, total (> 5 µg) or polyA RNA (> 0.2
µg) isolated from each sample is used to generate a biotinylated
"target" complementary RNA (cRNA). This is performed in a two-step
process, beginning with linear (or amplified) generation of a cDNA
library and ending with in vitro transcription of this cDNA
into cRNA. Following fragmentation of the cRNA, to improve
hybridization to the short oligonucleotide "probes" immobilized on
the chip, hybridization of each individual sample is performed. Each
chip consists of hundreds of thousands of short oligonucleotides,
representing between approximately 7,000 to 13,000 genes. Each gene is
represented as a "probe set" for which individual expression data
are generated, and multiple probe sets can interrogate the same gene.
Each probe set consists of 16 to 20 pairs of oligonucleotides
complementary to overlapping segments of an expressed sequence (cloned
gene or expressed sequence tag [EST]). One of each pair of
oligonucleotides has the correct sequence, while the other contains a
single mismatched nucleotide. This strategy is utilized in order to
control for random hybridization, with the mismatch oligonucleotide
serving as a baseline. For each probe set (gene), evaluation for
expression is based on the hybridization characteristics of all 16 to
20 sets of probe pairs (32 to 40 hybridization events).
For comparison purposes, the output must be scaled, usually by one of
two strategies: (1) using a small group of "house-keeping" genes
that show invariant expression, or (2) using a transcriptome-equivalent
strategy, with the assumption that the total sum of all transcripts are
similar between samples. Obviously, each strategy has its limitations,
but the transcriptome-equivalent strategy is currently more commonly
used. After scaling, hard data are generated for each probe set (and
each probe, though these data are typically not evaluated
individually). Multiple values are generated in an effort to fully
describe the expression characteristics of each probe set. No single
value gives a complete picture of the data set, yet for intuitive
simplicity the relative expression value (average difference value) is
most often used. Additional metrics can be used to further describe the
data. For instance, the absent/present call may indicate if a given
gene was or was not expressed in a sample.
|
Approach
|
|---|
We have applied gene expression profiling to whole-lung
tissue throughout lung development. Our intent was to generate a
profile of normal mouse development that will serve as a resource of
gene expression information and a baseline for future analysis of
murine models of abnormal development. Lung development was assessed
using Mu11Kchipset subA and subB oligonucleotide microarrays
(Affymetrix) as described.3
This high-density
oligonucleotide array encompasses > 11,000 cloned genes and ESTs.
Initial analysis was performed on Swiss Webster mice due to their large
lung size throughout early lung development. Future applications will
include other strains, both to observe similarities and differences
between strains, and to have a firm grasp of lung development in
strains used for gene targeting (such as C57BL/6-J). Our approach was
to combine multiple (three or more) lungs and perform a single
microarray analysis. Tissue was obtained every 2 to 3 days from
E12-P14. Adult lung tissue was also obtained. Total cellular RNA
was isolated, and 10 µg was used to generate target cRNA for
hybridization to the chip.
The initial strategy of pooling multiple lungs for a single array
per time point was based on our experience with the low technical
variability of this system as opposed to large biological variability.
Data evaluation using a variety of internal controls, such as multiple
probe sets for some individual genes (fibronectin,
1[I] procollagen) which gave similar
expression values, supported the accuracy of our data set. More
importantly, application of Northern hybridization to a small number of
cloned genes demonstrated a high degree of concordance between the
microarray data and traditional expression techniques. Additionally,
data mining revealed genes that clustered together most closely in
their developmental expression profile are genes that are of the same
family or closely related functions. In the future, these experiments
will be repeated to test biological and technical variability.
|
Extracellular Matrix Gene Expression During Lung
Development
|
|---|
Our initial focus has been on the expression of genes
encoding proteins that comprise the extracellular matrix (ECM). ECM
formation is an essential component to lung development and repair. The
ECM can serve as a structural support for organ morphogenesis, regulate
cellular activity with structural cues, and modulate growth factor
availability or activity.4
Examination of groups of ECM
genes shows similar profiles for molecules sharing functional
classification.3
Large-scale mathematical
clustering5
6
of the entire data set also reveals
expression profile similarities among genes sharing functional roles.
For instance, groups of genes encoding interstitial collagens clustered
together, using both hierarchical and agglomerative methods. Basement
membrane collagens clustered together with a distinct profile. Included
in the collagen nodes were other genes with similar expression patterns
including ESTs of unknown function and known transcription factors
(Table 1
). Coordinate regulation of transcription factors with ECM proteins
leads to hypotheses regarding regulation of ECM gene expression during
lung development
|
Conclusions
|
|---|
Expression profiling of lung development is a useful tool capable
of simultaneously identifying the expression patterns of large numbers
of genes and ESTs. These data should serve as a clearinghouse of
information for investigators interested in knowing the expression
pattern of any specific gene or group of genes in the lung. Even in the
context of the dynamic changes in cell populations of whole lung
tissue, this technique consistently reported similar expression
profiles for genes with previously known functional similarities. Data
mining of this massive data set can generate novel, testable hypothesis
related to the regulation of lung development. Further experimentation
will be essential to validate the specific hypotheses generated by
these approaches, as well as the utility of this method to identify
regulatory networks essential to the process of lung development.
|
Footnotes
|
|---|
This work was supported by the National Heart, Lung, and Blood
Institute; Francis Families Foundation; and the Washington University
School of Medicine Program in Lung Development.
Abbreviations: cDNA = complementary DNA;
cRNA = complementary RNA; ECM = extracellular matrix;
EST = expressed sequence tag
|
References
|
|---|
-
Golub, TR, Slonim, DK, Tamayo, P, et al (1999) Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science 286,531-537[Abstract/Free Full Text]
-
Spellman, PT, Sherlock, G, Zhang, MQ, et al (1998) Comprehensive identification of cell cycle-regulated genes of the yeast Saccharomyces cerevisiae by microarray hybridization. Mol Biol Cell 9,3273-3297[Abstract/Free Full Text]
-
Mariani T, Reed J, Shapiro S. Expression profiling of the developing mouse lung: insights into the establishment of the extracellular matrix. Am J Respir Cell Mol Biol. (in press)
-
Lukashev, M, Werb, Z (1998) ECM signalling: orchestrating cell behaviour and misbehaviour. Trends Cell Biol 8,437-441[CrossRef][ISI][Medline]
-
Eisen, MB, Spellman, PT, Brown, PO, et al (1998) Cluster analysis and display of genome-wide expression patterns Proc Natl Acad Sci U S A 95,14863-14868[Abstract/Free Full Text]
-
Alon, U, Barkai, N, Notterman, DA, et al (1999) Broad patterns of gene expression revealed by clustering analysis of tumor and normal colon tissues probed by oligonucleotide arrays. Proc Natl Acad Sci U S A 96,6745-6750[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
J. A. Whitsett, C. J. Bachurski, K. C. Barnes, P. A. Bunn Jr., L. M. Case, D. N. Cook, D. Crooks, M. W. Duncan, L. Dwyer-Nield, R. C. Elston, et al.
Functional Genomics of Lung Disease
Am. J. Respir. Cell Mol. Biol.,
August 1, 2004;
31(2/S1):
S1 - S81.
[Full Text]
[PDF]
|
|
|
TOP
Abstract
Methods
