The transcriptional program in the response of human fibroblasts to serum

Science

Volumn 283, Issue 5398, 1999, Pages 83-87

  Iyer, Vishwanath R ^a   Eisen, Michael B ^a   Ross, Douglas T ^a   Schuler, Greg ^b   Moore, Troy ^c   Lee, Jeffrey C F ^d   Trent, Jeffrey M ^e   Staudt, Louis M ^f   Hudson Jr , James ^c   Boguski, Mark S ^b   Lashkari, Deval ^d   Shalon, Dari ^d   Botstein, David ^a   Brown, Patrick O ^a  

^a STANFORD UNIVERSITY SCHOOL OF MEDICINE   (United States)

^b NATIONAL INSTITUTES OF HEALTH   (United States)

^c Research Genetics   (United States)

^d Experimental Station   (United States)

^e NATIONAL HUMAN GENOME RESEARCH INSTITUTE   (United States)

^f NATIONAL CANCER INSTITUTE   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

ACTIN; CYCLOOXYGENASE 2; INTERLEUKIN 8; MESSENGER RNA;

ARTICLE; CELL ACTIVATION; CELL KINETICS; CONTROLLED STUDY; FIBROBLAST; GENE EXPRESSION REGULATION; HUMAN; HUMAN CELL; PRIORITY JOURNAL; PROTEIN EXPRESSION; SERUM; TRANSCRIPTION REGULATION;

BLOOD; CA(2+)-CALMODULIN DEPENDENT PROTEIN KINASE; CELL CYCLE; CELL LINE; CHOLESTEROL; CULTURE MEDIA; CULTURE MEDIA, SERUM-FREE; EXPRESSED SEQUENCE TAGS; FIBROBLASTS; FLUORESCENT DYES; GENE EXPRESSION REGULATION; GENES, IMMEDIATE-EARLY; HUMANS; OLIGONUCLEOTIDE ARRAY SEQUENCE ANALYSIS; POLYMERASE CHAIN REACTION; SOFTWARE; TIME FACTORS; TRANSCRIPTION FACTORS; TRANSCRIPTION, GENETIC; WOUND HEALING;

EID: 0032903908     PISSN: 00368075     EISSN: None     Source Type: Journal    
DOI: 10.1126/science.283.5398.83     Document Type: Article

References (25)

1. J. A. Winkles, Prog. Nucleic Acid Res. Mol. Biol. 58, 41 (1998).
2. A normal human diploid fibroblast cell line derived from foreskin (ATCC CRL 2091) in passage 8 was used in these experiments. The protocol followed for growth arrest and stimulation was essentially that of (16) and (17). Cells were grown to about 60% confluence in 15-cm petri dishes in Dulbecco's minimum essential medium containing glucose (1 g/liter), the antibiotics penicillin and streptomycin, and 10% (by vol) FBS (Hyclone) that had been previously heat inactivated at 56°C for 30 min. The cells were then washed three times with the same medium lacking FBS, and low-serum medium (0.1% FBS) was added to the plates. After a 48-hour incubation, the medium was replaced with fresh medium containing 10% FBS. mRNA was isolated from several plates of cells harvested before serum stimulation; this mRNA served as the serum-starved or time-zero reference sample. Cells were harvested from batches of plates at 11 subsequent intervals (15 min, 30 min, 1, 2, 4, 6, 8, 12, 16, 20, and 24 hours) after the addition of serum. mRNA was also isolated from exponentially growing fibroblasts (not subjected to serum starvation). mRNA was isolated with the FastTrack mRNA isolation kit (Invitrogen), which involves lysis of the cells on the plate. The growth medium was removed, and the cells were quickly washed with phosphate-buffered saline at room temperature. The lysis buffer was added to the plate, transferred to tubes, and frozen in liquid nitrogen. Subsequent steps were performed according to the kit manufacturer's protocols.
3. The National Center for Biotechnology Information maintains the UniGene database as a resource for partitioning human sequences contained in GenBank into clusters representing distinct transcripts 01 genes (78, 19). At the time this work began, this database contained about 40,000 such clusters. We selected a subset of 10,000 of these UniGene clusters for inclusion on gene expression microarrays. UniGene clusters were included only if they contained at least one done from the I.M.A.G.E. human cDNA collection (20), so that a physical clone could easily be obtained (all I.M.A.G.E. clones are available commercially from a number of vendors). We attempted to include as complete as possible a set of the "named" human genes (about 4000) and genes that appeared to be closely related to named genes in other organisms (about an additional 2000). The remaining 4000 clones were chosen from among the "anonymous" UniGene clusters on the basis of inclusion on the human transcript map (www.ncbi. nlm.nih.gov/SCIENCE96/) and the lade of apparent homology to any other genes in the selected set. A physical done representing each of the selected genes was obtained from Research Genetics. This "10K set" is included in a more recent "15K set" described at www. nhgri.nih.gov/DIR/LCG/15K/HTML/p15Ktop.html. Of these clones, 472 are absent from the current edition of UniGene and were presumed to be distinct genes. The remainders represent 8141 distinct clusters, or human genes, in UniGene. These clones, thus presumed to represent 8613 different genes, were used to print microarrays according to methods described previously (21, 22).
4. One microgram of mRNA was used for making fluorescently labeled cDNA probes for hybridizing to the microarrays, with the protocol described previously (23). mRNA from the large batch of serum-starved cells was used to make cDNA labeled with Cy3. The Cy3-labeled cDNA from this batch of serum-starved cells served as the common reference probe in all hybridizations. mRNA samples from cells harvested immediately before serum stimulation, at intervals after serum stimulation, and from exponentially growing cells were used to make cDNA labeled with Cy5. Ten micrograms of yeast tRNA, 10 μg of polydeoxyadenylic acid, and 20 μg of human CoT1 DNA (Gibco-BRL) were added to the mixture of labeled probes in a solution containing 3X standard saline citrate (SSC) and 0.3% SDS and allowed to prehybridize at room temperature for 30 min before the probe was added to the surface of the microarray. Hybridizations, washes, and fluorescent scans were performed as described previously (23, 24). All measurements, totaling more than 180,000 differential expression measurements, were stored in a computer database for analysis and interpretation.
5. The nominal identities of a number of cDNAs (currently about 3750) on the microarray were verified by sequencing. The clones that were sequenced included many of the genes whose expression changed substantially upon serum stimulation, as well as a large number of genes whose expression did not change substantially in the course of this experiment. About 85% of the clones on the current version of this microarray that were checked by resequencing were correctly identified. In all the figures, gene names or EST numbers are given only for those genes on the microarrays whose identities were reconfirmed by resequencing. In the cases where a human gene has more than one name in the literature, we have tried to use the name that is most evocative of its presumed role in this context. The remainder of the clones have been assigned a temporary identification number (format: SID######) and a putative identity pending sequence verification. The correct identities of these genes will be posted at our Web site (genome-www.stanford.edu/serum) as they are confirmed by resequencing.
6. M. B. Eisen, P. T. Spellman, P. O. Brown, D. Botstein, Proc. Natl. Acad. Sci. U.S.A. 95, 14863 (1998).
7. 2(expression ratio) measured for the gene in this time course exceeded 0.7. In addition, observations in which the pixel-by-pixel correlation coefficients for the Cy3 and Cy5 fluorescence signals measured in a given array element were less than 0.6 were excluded. This selection criteria yielded a computationally manageable number of genes while minimizing the number of genes that were included because of noise in the data.
8. A more complete analysis and interpretation of the results of this experiment, as well as a searchable database, can be found at genome-www.stanford.edu/ serum
9. K. J. Livak, S. J. Flood, J. Marmaro, W. Giusti, K. Deetz, PCR Methods. Appl. 4, 357 (1995).
10. The apparent dip in the profile of COX2 at the 2-hour time point in the microarray data appears to result from a localized area of low intensity on the corresponding array scan resulting in an underestimation of the expression ratio. The expression ratios measured for mast/stem cell growth factor receptor are somewhat lower in the microarray data. This discrepancy is probably a consequence of the conservative background subtraction method used for quantitating the signal intensities on the array scans (23). The sequences of the PCR primer pairs (5′ to 3′) that were used are as follows: COX2, CCGTGGCTCTCTT-GGCAG and CTAAGTTCTTTAGCACTCCTTGGCA; IL-8, CGATGCTGTGGAGCTGTATC and CCATGGTTTC-ACCAAAGATG; mast/stem cell factor receptor, ACA-GAAGCCCGTGGTAGACC and GAGGCTGGGAGGAG-GAAG; B4-2, AAACCCCCCTCAGGAAAGAG and CC-ATGAACAAGCTGGCCAT; and actin, AGTACTCCGTGT-GGATCGGC and GCTGATCCACATCTGCTGGA
11. V. R. Iyer et al., unpublished data. The gene expression data for the early time points in the presence of cycloheximide will be available at our Web site (genome-www.stanford.edu/serum)
12. T. Hunter, Cell 80, 225 (1995).
13. J. Leppaluoto and H. Ruskoaho, Ann. Med. 24, 153 (1992).
14. A. C. Chang et al., Mol. Cell. Endocrinol. 112, 241 (1995).
15. K. L. Madsen et al., Am. J. Physiol. 274, G96 (1998).
16. W. Krek and J. A. DeCaprio, Methods Enzymol. 254, 114 (1995).
17. R. A. Tobey, J. G. Valdez, H. A. Crissman, Exp. Cell Res. 179, 400 (1988).
18. M. S. Boguski and G. D. Schuler, Nature Genet. 10, 369 (1995).
19. G. D. Schuler, J. Mol. Med. 75, 694 (1997).
20. G. Lennon, C. Auffray, M. Polymeropoulos, M. B. Soares, Genomics 33, 151 (1996).
21. I.M.A.G.E. clones were amplified by PCR in 96-well format with amino-linked primers at the 5′ end. Purified PCR products were suspended at a concentration of ∼0.5 mg/ml in 3X SSC, and ∼5 ng of each product was arrayed onto coated glass by means of procedures similar to those described previously (22). A total of 9996 elements were arrayed onto an area of 1.8 cm by 1.8 cm with the elements spaced 175 μm apart. The microarrays were then postprocessed to fix the DNA to the glass surface before hybridization with a procedure similar to previously described methods (22).
22. M. Schena, D. Shalon, R. W. Davis, P. O. Brown, Science 270, 467 (1995).
23. J. L DeRisi, V. R. Iyer, P. O. Brown, ibid. 278, 680 (1997).
24. J. DeRisi et al., Nature Genet. 14, 457 (1996).
25. We thank E. Chung for help with sequencing, A. Alizadeh for help with sequence verification, K. Ranade for advice on the TaqMan assay, and J. DeRisi and other members of the P.O.B. and D.B. labs for discussions. Supported by a grant from the National Human Genome Research Institute (NHCRI) (HG00450) and the National Cancer Institute (NIH CA 77097). V.R.I. was supported in part by an Institutional Training Grant in Genome Sciences (T32 HG00044) from the NHGRI. M.B.E. is an Alfred E. Sloan Foundation Postdoctoral Fellow in Computational Molecular Biology, and D.T.R is a Walter and Idun Berry Fellow. P.O.B. is an Associate Investigator of the Howard Hughes Medical Institute.