A new paradigm for profiling testicular gene expression during normal and disturbed human spermatogenesis.

The aim of this study was to identify gene expression patterns of the testis that correlate with the appearance of distinct stages of male germ cells. We avoided the pitfalls of mixed pathological phenotypes of the testis and circumvented the inapplicability of using the first spermatogenic wave as done previously on rodents. This was accomplished by using 28 samples showing defined and highly homogeneous pathologies selected from 578 testicular biopsies obtained from 289 men with azoospermia (two biopsies each). The molecular signature of the different developmental stages correlated with the morphological preclassification of the testicular biopsies, as shown by resampling-based hierarchical clustering using different measures of variability. By using analysis of variance (ANOVA) and extensive permutation analysis, we filtered 1181 genes that exhibit exceptional statistical significance in testicular expression and grouped subsets with transcriptional changes within the pre-meiotic (348 genes), post-meiotic (81 genes) and terminal differentiation (38 genes) phase. Several distinct molecular classes, metabolic pathways and transcription factor binding sites are involved, depending on the transcriptional profile of the gene clusters that were built using a novel clustering procedure based on not only similarity but also statistical significance.