MATERIALS AND METHODS

 

Patient Characteristics

Patients were included in this study based on the availability of freshly frozen lymph node biopsy material containing enough mRNA to allow cDNA microarray analysis.  Only patients with samples that had been obtained prior to any systemic therapy were included.  In all cases the pathological diagnosis was follicular NHL (follicular small cleaved, follicular mixed or follicular large cell histology).  Each patient received rituximab treatment with documentation of clinical outcome.  In all cases, biopsy and pathology review were performed at Stanford University Medical Center.

 

Microarray Procedures

 Freshly frozen lymph node samples were obtained from patients who underwent excisional biopsy at Stanford University Medical Center (SUMC) between 1984 and 1997, who subsequently received rituximab between 1994 and 2000 and whose clinical response to rituximab treatment had been recorded.  Tonsil and spleen samples were similarly obtained from patients treated at SUMC in 2000 or 2001.  Biopsy samples were stored frozen in optimal cutting temperature compound.  Poly-(A)+ mRNA was obtained from biopsy samples after homogenization of tissue using the FastTrack 2.0 kit (Invitrogen).  The quality of the resulting RNA was assessed by agarose gel electrophoresis prior to labeling and hybridization.  An experimental cDNA probe incorporating Cy5 dye was generated from mRNA from malignant and normal lymphoid tissues; a common reference cDNA probe incorporating Cy3 dye was from mRNA derived from a panel of cell lines and probes were hybridized to cDNA microarrays as previously described (1, 2).  Specific protocols used for mRNA isolation and for microarray post-processing and hybridization can be downloaded from the Brown Lab website here.  Two types of microarrays were used.  Some experiments in this study used Stanford Human (SH) arrays comprised of 38,431 DNA spots of 38,276 unique cDNA clones, representing approximately 31,139 unique Unigene clusters of which 16,152 correspond to unique named genes.  Some experiments were conducted with LC microarrays comprised of 37,632 DNA spots with 32,876 unique cDNA clones, representing approximately 17,622 Unigene clusters of which at least 10,250 are unique named genes.  The LC microarrays are enriched for lymphoid genes and are a later generation of the Lymphochip microarray (2).

 

Statistical Analysis

Prior to analysis, individual data points were median centered for each cDNA clone.  The dendrogram and gene clustering displayed in Figure 1 results from agglomerative hierarchical cluster analysis applied to the gene axis and to the sample axis, as previously described (3).  Prior to cluster analysis, the data was filtered for data quality and variance, as follows.  Data from 24 SH arrays representing malignant and normal tissue was selected for signal greater than 1.5 fold above local background in both the sample and reference channels.  Clones were identified by Stanford Unique Identifier (SUID), which averages data if multiple copies of the same clone are present on the array.  The resulting "precluster" file can be downloaded here.  Data were then filtered based on variance such that clones were included if expression was greater than two fold above or below the median on at least three arrays. Clones were required to have good data, by the previous criteria, on at least 80% of the arrays to be included in hierarchical cluster analysis.  Genes and arrays were then clustered by Pearson Correlation.  Hierarchical cluster analysis of LC data revealed a technical artifact that resulted is samples segregating by the date of the experiment.  Further investigation revealed that this artifact was likely due to differences in the calibration of the two scanners used to analyze the microarrays.  Singular value decomposition (SVD) was used to filter data for remove the pattern corresponding to this artifact prior to analysis (4) after missing data were estimated using a KNN impute algorithm with 10 missing values (5).  Prior to SVD and supervised analysis, clones from the LC data were selected based on data quality; clones were required to have signal greater than 2.5 fold above local background in either the sample or reference channel and good data for on at least 80% of the arrays.  The LC dataset prior to SVD analysis can be downloaded here.  The SVD filtered data can also be downloaded, either imputed or non-imputed.  Data from SH arrays was filtered prior to supervised analysis, using the same criteria as above; the resulting dataset is available here.  Supervised analysis taking into account known outcome to rituximab treatment was performed using Wilcoxon rank sum test to generate a rank list of genes whose corresponding mRNA levels differ significantly in rituximab responders versus non-responders (6).  For this analysis, patients were divided into two groups, rituximab responders (composed of  CR and PR) and non-responders (composed of NR and MR).

 

References

1.         Perou, C. M., Sorlie, T., Eisen, M. B., van de Rijn, M., Jeffrey, S. S., Rees, C. A., Pollack, J. R., Ross, D. T., Johnsen, H., Akslen, L. A., Fluge, O., Pergamenschikov, A., Williams, C., Zhu, S. X., Lonning, P. E., Borresen-Dale, A. L., Brown, P. O. & Botstein, D. (2000) Nature 406, 747-52.

2.         Alizadeh, A. A., Eisen, M. B., Davis, R. E., Ma, C., Lossos, I. S., Rosenwald, A., Boldrick, J. C., Sabet, H., Tran, T., Yu, X., Powell, J. I., Yang, L., Marti, G. E., Moore, T., Hudson, J., Jr., Lu, L., Lewis, D. B., Tibshirani, R., Sherlock, G., Chan, W. C., Greiner, T. C., Weisenburger, D. D., Armitage, J. O., Warnke, R., Staudt, L. M. & et al. (2000) Nature 403, 503-11.

3.         Eisen, M. B., Spellman, P. T., Brown, P. O. & Botstein, D. (1998) Proc Natl Acad Sci U S A 95, 14863-8.

4.         Alter, O., Brown, P. O. & Botstein, D. (2000) Proc Natl Acad Sci U S A 97, 10101-6.

5.         Troyanskaya, O., Cantor, M., Sherlock, G., Brown, P., Hastie, T., Tibshirani, R., Botstein, D., & Altman, R. B. (2001) Bioinformatics 17, 520-5.

6.         Troyanskaya, O. G., Garber, M. E., Brown, P. O., Botstein, D. & Altman, R. B. (2002) Bioinformatics 18, 1454-61.