An analysis of gene expression profiles obtained from cervical cancers was performed to find those genes most aberrantly expressed. Total RNA was prepared from 29 samples of cervical squamous cell carcinoma and 18 control samples, and hybridized to Affymetrix oligonucleotide microarrays with probe sets complementary to over 20,000 transcripts. Unsupervised hierarchical clustering of the expression data readily distinguished normal cervix from cancer. Supervised analysis of gene expression data identified 98 and 139 genes that exhibited >2-fold upregulation and >2-fold downregulation, respectively, in cervical cancer compared to normal cervix. Several of the genes that were differentially regulated included SPP1 (Osteopontin), CDKN2A (p16), RPL39L, Clorf1, MAL, p11, ARS and NICE-1. These were validated by quantitative RT-PCR on an independent set of cancer and control specimens. Gene Ontology analysis showed that the list of differentially expressed genes included ones that were involved in multiple biological processes, including cell proliferation, cell cycle and protein catabolism. Immunohistochemical staining of cancer specimens further confirmed differential expression of SPP1 in cervical cancer cells vs. nontumor cells. In addition, 2 genes, CTGF and RGS1 were found to be upregulated in late stage cancer compared to early stage cancer, suggesting that they might be involved in cancer progression. The pathway analysis of expression data showed that the SPP1, VEGF, CDC2 and CKS2 genes were coordinately differentially regulated between cancer and normal. The present study is promising and provides potential new insights into the extent of expression differences underlying the development and progression of cervical squamous cell cancer. This study has also revealed several genes that may be highly attractive candidate molecular markers/targets for cervical cancer diagnosis, prognosis and therapy. ' 2005 Wiley-Liss, Inc.