(Circulation. 2000;101:1990.)
© 2000 American Heart Association, Inc.
Basic Science Reports |
From the Departments of Cardiovascular Research (P.D.S., J.T.N.T., L.A.B., R.Y., H.J., D.G.L.), Pathology (K.J.H), and Molecular Biology (A.G.), Genentech, Inc, South San Francisco, Calif.
Correspondence to David G. Lowe, PhD, Genentech, Inc, Cardiovascular Research, Mail Stop 42, 1 DNA Way, South San Francisco, CA 94080. E-mail lowe{at}gene.com
| Abstract |
|---|
|
|
|---|
Methods and ResultsMicroarrays with rat cDNAs for 86 known genes and 989 anonymous cDNAs obtained by molecular subtraction (representational difference analysis) of mRNA from sham-operated and 6-week post-MI samples were used in 2-color hybridization experiments. Twelve known genes previously associated with myocardial development were identified together with 10 uncharacterized expressed sequence tags and 36 genes not previously associated with cardiac development. After MI, genes associated with myocardial stress and wound healing exhibited differences in magnitude and expression kinetics, and 14 genes not previously associated with MI were identified. In situ hybridization revealed mRNA localization characteristic of wound healing and vascular and cardiomyocyte reactivity.
ConclusionsTissue analysis of gene expression with cDNA microarrays provides a measure of transcriptional or posttranscriptional regulation and cellular recruitment. Our results demonstrate the complexity of gene regulation in the developing myocardium and show that cDNA microarrays can be used to monitor the evolution of the cardiac stressinducible phenotype.
Key Words: genes molecular biology myocardial infarction
| Introduction |
|---|
|
|
|---|
The development of cDNA microarray technology to quantitatively monitor the expression of thousands of genes in parallel10 11 12 allows the detection of changes in the physiological status of tissues through changes in the molecular phenotype. These changes can reflect alterations in transcription, mRNA stability, or cellular composition in response to a changing environment.
In the present study, we used cDNA microarrays to analyze transcript levels in cardiac tissue samples. Our results provide a detailed view of the evolving phenotype of the developing heart and the myocardial response to stress.
| Methods |
|---|
|
|
|---|
100 embryonic or 10 neonatal samples. For the post-MI
study, coronary artery ligation was used to induce left
ventricle infarction13 in 12-week-old male Sprague-Dawley
rats. Sham surgeries were performed on the control groups for each time
point. ECGs were used to document the development of infarcts according
to the depth and persistence of pathological waves14 15
and to select animals with an infarct of at least
30%.16 17
cDNA Amplification and Microarray Preparation
Eighty-six known genes (Table 1![]()
) were polymerase
chain reaction (PCR)-amplified from rat heart cDNA by standard methods,
cloned into pCR2.1 (Invitrogen), and sequenced for verification.
Individual cDNA clones were PCR-amplified, purified, and prepared for
microarray printing as described previously18 19 (details
are available at http://cmgm.stanford/pbrown/index.html).
|
|
Generation of Differentially Expressed cDNAs
cDNA fragments enriched for sequences induced or repressed after
MI were prepared by representational difference
analysis20 with the use of modifications to the
PCR-Select cDNA Subtraction Kit (Clontech). Briefly, first-strand cDNA
was synthesized by using 2 µg of mRNA pooled from 3 sham-operated
ventricles or three 6-week post-MI ventricles and 1 µmol/L of an
oligo-dT primer with either RsaI or MslI
restriction sites at the 5' end. After second-strand synthesis, the
cDNA fragments were digested separately with RsaI or
MslI, and the protocol was followed according to the
manufacturers instructions. From the induced and repressed
subtractions, 3744 clones were picked and screened with
32P-labeled rat mitochondrial
DNA21 to identify mitochondrial genomic fragments;
1963 (52%) nonmitochondrial clones were identified; and 989 clones
were selected at random for microarray fabrication as described
above.
Fluorescent Probes
Total RNA was prepared from tissues by use of RNA-Stat 60
(Tel-Test), and poly(A)+ RNA was isolated by
using the PolyATract System (Promega). Reverse transcriptase (RT)
reactions were carried out by using Superscript II RNase
H- RT (GIBCO) as described by the manufacturer,
with the following modifications: 25 µL reactions contained 40
ng/µL poly(A)+ mRNA or 2 µg/µL total RNA,
80 ng/µL oligo dT primer, 1x Superscript first-strand buffer, 0.4
U/mL RNase Block (Stratagene), 500 µmol/L each of dATP, dGTP,
and dTTP, 280 µmol/L dCTP (Boehringer-Mannheim), 40
µmol/L Cy5-dCTP or Cy3-dCTP (Amersham), and 8 U/µL Superscript II
RT.
Data Analysis and Quantification
Hybridizations, washes, and array imaging were performed as
previously described10 18 22 at Synteni Inc. Raw scores
were normalized to the total fluorescence in either channel
(Cy3 or Cy5), which is the sum of all fluorescence intensity
values for the complete set of 1075 cDNAs. A ratio of differential gene
expression was calculated from the normalized fluorescence
intensity value obtained from each channel. Two different adult
ventricular RNA samples were labeled with Cy3- and Cy5-dCTP
and mixed in a reciprocal fashion to establish cutoff values for
assigning differential expression. Deviation from an ideal ratio of 1
was found to range from 1.4 to 0.65. Because of this deviation, we
chose a ratio of >1.5 for induction or <0.6 for repression.
Expression ratios are reported as mean±SEM.
In Situ Hybridization
Frozen sections were prepared from rat ventricular
tissue for in situ hybridization as previously
described.23 cDNA fragments were prepared for in vitro
transcription24 by PCR amplification with the use of
gene-specific primers incorporating T3 and T7 RNA polymerase
promoters.
| Results |
|---|
|
|
|---|
For the 989 anonymous cDNA clones on the array, we found differential
expression for 117 (11.8%). After they were sequenced, these clones
clustered to 48 genes modulated in either the embryonic or neonatal
myocardium (Table 2
). The
majority of these (38 genes) were known, although they were not
previously associated with cardiac development. In addition, we
identified 10 cDNA clones corresponding to uncharacterized expressed
sequence tags (ESTs) (Table 2
).
|
Gene Expression After MI
At 1, 3, 7, 14, 42, and 84 days after surgery,
ventricular tissue samples were harvested from groups
subjected to coronary artery ligation and from sham-operated
groups (n=4 to 7 animals per group per time point). No hybridization
signal was detected for cyclooxygenase 2, Bcl-2,
and parathyroid hormone-like protein (PTHLP) in the set of known genes,
whereas the remaining 83 genes (Table 1
) gave a signal at the 6
time points. Of these, only 5 were induced for at least 1 time point
after MI (Figure 1A
and 1B
). There is
early elevation of mRNA encoding atrial natriuretic peptide
(ANP) and brain natriuretic peptide (BNP) at 1 day after
MI, followed by further increases in ANP mRNA (Figure 1A
). In
contrast, BNP mRNA remains elevated at a constant level over the entire
12-week time course. The matrix proteins osteopontin, collagen III, and
fibronectin each exhibit different kinetics of expression, but all
returned to baseline within 6 weeks of surgery (Figure 1B
).
|
For the 989 unknown cDNAs, we found 63 clones exhibiting differential
expression; these clones clustered to 14 different genes with elevated
expression at 1 or more time points as long as 2 weeks after MI (Figure 1C
, Table 3
). No differential
expression was observed for these genes at 6 or 12 weeks after MI.
|
mRNA Localization by In Situ Hybridization
Five genes were studied by in situ hybridization to determine the
cellular source of differential expression. Hearts were harvested from
sham-operated and MI rats at 4 and 10 days after surgery for
analysis. Compared with normal myocardium (Figure 2A
), infarcted myocardium
revealed features typical of necrosis and hemorrhage (Figures 2B
, 3A
, and 3C
). In normal
heart, expression of vimentin (Figure 2C
), p41-Arp2/3
protein complex subunit (ARC) (Figure 2D
), and elongation
factor-1
(EF-1
, Figure 4A
and 4B
)
was confined to vascular structures. Vimentin and EF-1
were
expressed in blood vessel walls by smooth muscle cells and pericytes,
whereas p41-ARC was expressed by endothelial cells
lining small and large vessels as well as by vascular pericytes. The
post-MI expression of each increased in spindle-shaped fibroblast-like
cells within and adjacent to the region of damage (Figure 2E
, 2F
, and 2G
). Expression of EF-1
also increased in blood vessels
after MI (Figure 4
). In contrast, cathepsin B was expressed in
mononuclear phagocytes in infarcted hearts only (Figure 3A
and 3B
). Cathepsin B expression appeared higher at day 10 than at day 4,
and its temporal and distinctive regional localization would be
consistent with a role in organization and repair.
Phospholemman was expressed at low levels in normal cardiac myocytes,
with expression increasing after MI. The hybridization signal was
higher in the ventricles and was highest in myocytes adjacent to the
areas of damage (Figures 2H
, 3C
, and 3D
).
|
|
|
| Discussion |
|---|
|
|
|---|
,28 endothelin type A receptor,9
connexin40,29 osteopontin,30 manganese
superoxide dismutase,9 troponin I,31 and
heat-shock protein 22/
-B crystallin32 ) or during the
response to myocardial injury and stress (ANP and BNP,25
osteopontin,30 33 collagen III,34 and
fibronectin35 ) supports our identification of genes newly
associated with these cardiac phenotypes. However, these
results are not likely to be comprehensive even for the genes on our
microarray. Signal dilution due to cell-typespecific expression or
spatial regulation of mRNA levels could cause a signal below the
differential expression threshold of detection. These considerations
are particularly relevant to the heart, a complex arrangement of
cardiomyocytes and endothelial, smooth
muscle, endocardial, fibroblast, and neuroendocrine cells combined with
spatial and regional specialization.
Developmental Gene Expression
Cardiac gene expression changes markedly in the progression from
embryo to adult, reflecting the transition from a growing and
developing organ to a stable adult heart. The largest number of
differentially expressed mRNAs was detected when 13-day embryonic
ventricles were compared with adult ventricles. Our data are
consistent with results from EST cDNA sequencing, where clones
more abundant in fetal and neonatal cDNA libraries frequently encoded
regulatory and signal transduction proteins, reflecting a growing and
less differentiated phenotype.36
Reexpression of genes normally expressed in cardiac development has been observed in models of cardiac hypertrophy and failure.7 8 37 However, there is not a simple correlation between adaptive or pathological hypertrophy and developmental gene expression.4 9 14 Our data involving cardiac developmental gene expression provide a basis for more detailed testing of the relation between stress-inducible phenotypes in the adult myocardium and gene expression in myocardial development.
Expression Profiling in the Response to Injury
The rat model of surgical MI that we used is well characterized
and represents an initial wound-healing response, followed by
progression to cardiac dysfunction and failure.38 The
histological wound-healing response of the infarcted
myocardium is reflected in the expression pattern of
osteopontin,33 fibronectin,35 and collagen
III,34 which return to baseline expression levels between
2 and 6 weeks. In addition to gene expression reflecting the
wound-healing response, ANP and BNP represent a
cardiomyocyte response to stress,25 with
elevated expression as much as 12 weeks after MI. In situ localization
of mRNA for 5 genes newly associated with MI highlights different
cellular aspects of the tissue response to infarction, including wound
healing in the infarcted region (vimentin,39
EF-1
,28 and p41-ARC40 ), infiltrating
mononuclear phagocytes (cathepsin B41 ), vascular
reactivity (p41-ARC40 ), and myocyte reactivity
(phospholemman42 ).
Our data on changes in gene expression after MI illustrate how microarrays may be used to monitor the progression of the tissue molecular phenotype and, as a discovery tool, to identify new genes associated with myocardial stress. This technology should be a potent hypothesis generator in experiments designed to elucidate the mechanisms of disease pathogenesis and to determine the basis of drug action.
Study Limitations
The microarray that we designed was intended to evaluate the
expression of known and unknown genes by comparing developmental gene
expression with the response to MI. In fact, very few differentially
expressed genes were detected in the post-MI time course. This is
particularly striking given the extent of differential expression
identified in the embryo and neonate compared with the adult by use of
the same microarrays. We consider several possibilities. First, the
relation between the myocardial stress phenotype and
developmental gene expression7 8 37 may, in fact, be quite
limited, a conclusion supported by previous studies using fewer
candidate genes.4 5 9
Alternatively, there are still extensive changes in gene expression
after MI that are below the level of microarray sensitivity. This
limitation may be overcome by isolating selected regions of the
myocardium to enrich for expression differences, by
examining isolated cell populations, or by technically improving assay
precision and reproducibility. The latter case is illustrated by our
results with cathepsin B: multiple cDNAs were on the arrays, but
varying numbers of clones were hybridized at or above the differential
expression threshold (Tables 2
and 3
). Beyond 2 weeks
after MI, no cathepsin B differential expression was detected. However,
the cathepsin B cDNAs were obtained by molecular subtraction of mRNA of
sham-operated and 6-week post-MI myocardium. Enrichment for
cathepsin B cDNAs, as reflected in the large number of randomly picked
cathepsin B cDNAs on the microarray, clearly suggests that there is
elevated expression at later times. Although these results suggest that
molecular subtraction by representation difference
analysis does provide enrichment, improvements in sensitivity
will be required to evaluate the full advantage of this strategy.
However, when all expressed sequences are known as a result of the
genome projects, the molecular subtraction approach will not be a
necessary antecedent to array fabrication.
Distinguishing among closely related members of a gene family is also problematic with microarrays, inasmuch as we were not able to detect expected changes in actin and myosin gene expression.43 44 This is clearly an area requiring technical improvement with modified hybridization conditions or the use of shorter DNA fragments/oligonucleotides.
In summary, the present study establishes the utility of cDNA microarrays to monitor the expression profile of a tissue over time. In addition to function, anatomy, and histology, expression analysis by microarray can be expected to become an integral part of studying the response of the heart to a changing environment. Additional work on myocardial expression profiling can be expected to help unravel the relation between gene expression and cardiac function in the normal and diseased heart.
| Acknowledgments |
|---|
| Footnotes |
|---|
Received March 22, 1999; revision received October 25, 1999; accepted November 18, 1999.
| References |
|---|
|
|
|---|
-actin gene is
reactivated during cardiac hypertrophy provoked by
load. J Clin Invest. 1991;88:15811588.This article has been cited by other articles:
![]() |
M. Nahrendorf, D. E. Sosnovik, P. Waterman, F. K. Swirski, A. N. Pande, E. Aikawa, J.-L. Figueiredo, M. J. Pittet, and R. Weissleder Dual Channel Optical Tomographic Imaging of Leukocyte Recruitment and Protease Activity in the Healing Myocardial Infarct Circ. Res., April 27, 2007; 100(8): 1218 - 1225. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. Y. CHEUNG, L. I. ROTHBLUM, J. R. MOORMAN, A. L. TUCKER, J. SONG, B. A. AHLERS, L. L. CARL, J. WANG, and X.-Q. ZHANG Regulation of Cardiac Na+/Ca2+ Exchanger by Phospholemman Ann. N.Y. Acad. Sci., March 1, 2007; 1099(1): 119 - 134. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. Scholz Unraveling the basic principles Am J Physiol Regulatory Integrative Comp Physiol, June 1, 2006; 290(6): R1485 - R1487. [Full Text] [PDF] |
||||
![]() |
R. S. Vasan Biomarkers of Cardiovascular Disease: Molecular Basis and Practical Considerations Circulation, May 16, 2006; 113(19): 2335 - 2362. [Full Text] [PDF] |
||||
![]() |
X.-Q. Zhang, J. R. Moorman, B. A. Ahlers, L. L. Carl, D. E. Lake, J. Song, J. P. Mounsey, A. L. Tucker, Y.-m. Chan, L. I. Rothblum, et al. Phospholemman overexpression inhibits Na+-K+-ATPase in adult rat cardiac myocytes: relevance to decreased Na+ pump activity in postinfarction myocytes J Appl Physiol, January 1, 2006; 100(1): 212 - 220. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. V. Podgoreanu and D. A. Schwinn New Paradigms in Cardiovascular Medicine: Emerging Technologies and Practices: Perioperative Genomics J. Am. Coll. Cardiol., December 6, 2005; 46(11): 1965 - 1977. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. Bossuyt, X. Ai, J. R. Moorman, S. M. Pogwizd, and D. M. Bers Expression and Phosphorylation of the Na-Pump Regulatory Subunit Phospholemman in Heart Failure Circ. Res., September 16, 2005; 97(6): 558 - 565. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. V. Podgoreanu, G. A. Michelotti, Y. Sato, M. P. Smith, S. Lin, R. W. Morris, H. P. Grocott, J. P. Mathew, and D. A. Schwinn Differential cardiac gene expression during cardiopulmonary bypass: Ischemia-independent upregulation of proinflammatory genes J. Thorac. Cardiovasc. Surg., August 1, 2005; 130(2): 330 - 339. [Abstract] [Full Text] [PDF] |
||||
![]() |
B. A. Ahlers, X.-Q. Zhang, J. R. Moorman, L. I. Rothblum, L. L. Carl, J. Song, J. Wang, L. M. Geddis, A. L. Tucker, J. P. Mounsey, et al. Identification of an Endogenous Inhibitor of the Cardiac Na+/Ca2+ Exchanger, Phospholemman J. Biol. Chem., May 20, 2005; 280(20): 19875 - 19882. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. Song, X.-Q. Zhang, B. A. Ahlers, L. L. Carl, J. Wang, L. I. Rothblum, R. C. Stahl, J. P. Mounsey, A. L. Tucker, J. R. Moorman, et al. Serine 68 of phospholemman is critical in modulation of contractility, [Ca2+]i transients, and Na+/Ca2+ exchange in adult rat cardiac myocytes Am J Physiol Heart Circ Physiol, May 1, 2005; 288(5): H2342 - H2354. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. Rysa, H. Leskinen, M. Ilves, and H. Ruskoaho Distinct Upregulation of Extracellular Matrix Genes in Transition From Hypertrophy to Hypertensive Heart Failure Hypertension, May 1, 2005; 45(5): 927 - 933. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Freimann, M. Scheinowitz, D. Yekutieli, M. S. Feinberg, M. Eldar, and G. Kessler-Icekson Prior exercise training improves the outcome of acute myocardial infarction in the rat: Heart structure, function, and gene expression J. Am. Coll. Cardiol., March 15, 2005; 45(6): 931 - 938. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. Fransen Phospholemman, a chaperone of Na+,K+-ATPase? Cardiovasc Res, January 1, 2005; 65(1): 13 - 15. [Full Text] [PDF] |
||||
![]() |
H-W Chen, S-L Yu, W-J Chen, P-C Yang, C-T Chien, H-Y Chou, H-N Li, K Peck, C-H Huang, F-Y Lin, et al. Dynamic changes of gene expression profiles during postnatal development of the heart in mice Heart, August 1, 2004; 90(8): 927 - 934. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. V. Podgoreanu and D. A. Schwinn Genomics and the circulation Br. J. Anaesth., July 1, 2004; 93(1): 140 - 148. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. A. Mirza, X.-Q. Zhang, B. A. Ahlers, A. Qureshi, L. L. Carl, J. Song, A. L. Tucker, J. P. Mounsey, J. R. Moorman, L. I. Rothblum, et al. Effects of phospholemman downregulation on contractility and [Ca2+]i transients in adult rat cardiac myocytes Am J Physiol Heart Circ Physiol, April 1, 2004; 286(4): H1322 - H1330. [Abstract] [Full Text] [PDF] |
||||
![]() |
O. Tarnavski, J. R. McMullen, M. Schinke, Q. Nie, S. Kong, and S. Izumo Mouse cardiac surgery: comprehensive techniques for the generation of mouse models of human diseases and their application for genomic studies Physiol Genomics, February 13, 2004; 16(3): 349 - 360. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. J. Crampin, M. Halstead, P. Hunter, P. Nielsen, D. Noble, N. Smith, and M. Tawhai Computational physiology and the physiome project Exp Physiol, January 1, 2004; 89(1): 1 - 26. [Abstract] [Full Text] [PDF] |
||||
![]() |
B. Z. Simkhovich, P. Marjoram, C. Poizat, L. Kedes, and R. A. Kloner Brief episode of ischemia activates protective genetic program in rat heart: a gene chip study Cardiovasc Res, August 1, 2003; 59(2): 450 - 459. [Abstract] [Full Text] [PDF] |
||||
![]() |
C Napoli, L O Lerman, V Sica, A Lerman, G Tajana, and F de Nigris Microarray analysis: a novel research tool for cardiovascular scientists and physicians Heart, June 1, 2003; 89(6): 597 - 604. [Abstract] [Full Text] [PDF] |
||||
![]() |