-
Kazuhiro Hayashida,
Akito Daiba,
Akito Sakai,
Takeshi Tanaka,
Kyosuke Kaji,
Niro Inaba,
Satoshi Ando,
Naoki Kajiyama, Hiroshi Terasaki,
Aki Abe,
Masanori Ogasawara,
Michinori Kohara,
Mine Harada,
Takeshi Okanoue,
Satoru Ito,
Shuichi Kaneko
[show abstract]
[hide abstract]
ABSTRACT: Interferon has been used widely to treat patients with chronic hepatitis C infections. Prediction of interferon efficacy before treatment has been performed mainly by using viral information, such as viral load and genotype. This information has allowed the successful prediction of sustained responders (SR) and non-SRs, which includes transient responders (TR) and nonresponders (NR). In the current study we examined whether liver messenger RNA expression profiles also can be used to predict interferon efficacy.
RNA was isolated from 69 liver biopsy samples from patients receiving interferon monotherapy and was analyzed on a complementary DNA microarray. Of these 69 samples, 31 were used to develop an algorithm for predicting interferon efficacy, and 38 were used to validate the precision of the algorithm. We also applied our methodology to the prediction of the efficacy of interferon/ribavirin combination therapy using an additional 56 biopsy samples.
Our microarray analysis combined with the algorithm was 94% successful at predicting SR/TR and NR patients. A validation study confirmed that this algorithm can predict interferon efficacy with 95% accuracy and a P value of less than .00001. Similarly, we obtained a 93% prediction efficacy and a P value of less than .0001 for patients receiving combination therapy.
By using only host data from the complementary DNA microarray we are able to successfully predict SR/TR and NR patients for interferon therapy. Therefore, this technique can help determine the appropriate treatment for hepatitis C patients.
Clinical Gastroenterology and Hepatology 01/2006; 3(12):1253-9. · 5.63 Impact Factor
-
[show abstract]
[hide abstract]
ABSTRACT: In this study, we used the rat liver as a model system to optimize the conditions for extracting RNA from liver biopsies for use in cDNA microarrays. We found that a 5-mm biopsy with a 16-gauge needle and storage in RNA later at 4 degrees C were optimal conditions for RNA extraction. The most important factor for the quantity and quality of RNA extraction was the sample diameter. Using the optimized sampling conditions and a cDNA microarray, we compared the expression of genes in the normal and the fibrotic tissues of the LEC rat liver, a model of liver tumorigenesis, with SD rat liver RNA as a reference. We found 29 genes that were up-regulated and 33 genes that were down-regulated in the fibrotic part of the liver. Furthermore, with the help of the reference RNA, we were able to classify the expression profiles into five groups without complex mathematical analyses; without the reference RNA, the genes could be classified into only two groups. Finally, we found that osteopontin was expressed at a very high level in the fibrotic portion of the LEC rat liver. This cDNA microarray result was validated by immunohistochemistry, which showed an elevated expression of osteopontin in the region of cholangiocarcinoma and a lack of expression in normal tissues. With optimized conditions, we should be able to apply the microarray system for routine practice.
Analytical Biochemistry 03/2005; 337(2):224-34. · 3.00 Impact Factor
-
Akito Daiba,
Niro Inaba,
Satoshi Ando,
Naoki Kajiyama,
Hiroshi Yatsuhashi, Hiroshi Terasaki,
Atsushi Ito,
Masanori Ogasawara,
Aki Abe,
Junichi Yoshioka,
Kazuhiro Hayashida,
Shuichi Kaneko,
Michinori Kohara,
Satoru Ito
[show abstract]
[hide abstract]
ABSTRACT: We have designed and established a low-density (295 genes) cDNA microarray for the prediction of IFN efficacy in hepatitis C patients. To obtain a precise and consistent microarray data, we collected a data set from three spots for each gene (mRNA) and using three different scanning conditions. We also established an artificial reference RNA representing pseudo-inflammatory conditions from established hepatocyte cell lines supplemented with synthetic RNAs to 48 inflammatory genes. We also developed a novel algorithm that replaces the standard hierarchical-clustering method and allows handling of the large data set with ease. This algorithm utilizes a standard space database (SSDB) as a key scale to calculate the Mahalanobis distance (MD) from the center of gravity in the SSDB. We further utilized sMD (divided by parameter k: MD/k) to reduce MD number as a predictive value. The efficacy prediction of conventional IFN mono-therapy was 100% for non-responder (NR) vs. transient responder (TR)/sustained responder (SR) (P < 0.0005). Finally, we show that this method is acceptable for clinical application.
Biochemical and Biophysical Research Communications 03/2004; 315(4):1088-96. · 2.48 Impact Factor
