The main goal of all analytic procedures in patients with non-obstructive azoospermia is to quickly obtain reliable data for successful prediction of testicular sperm retrieval. Some laboratories attempted to predict spermatogenesis with non-invasive techniques with differing success [14–16]. To date, the only generally accepted reliable predictor of successful TESE is testicular histology . Analysis of the germ cell-specific gene expression in testicular samples can provide an additional, supplementing approach to increase the prediction of positive TESE outcome.
Testicular transcriptome consists of gene expression patterns of both somatic and germ cells and has been intensively studied in recent years . The first studies were focused on describing the global testicular gene expression and identifying testicular genes in mice  and human . Shima et al.  took advantage of the first synchronous wave of spermatogenesis in pubertal mice to locate the gene products to specific testicular cells. In a different approach, germ cells were purified for high-throughput analysis of cell-specific gene expression studies on animal models [20, 21]. The data from the above-mentioned studies provides a vast number of possible gene candidates as markers of spermatogenesis that fulfil the criteria of testis-specific gene expression, are transcribed and translated at specific time points of spermatogenesis, and their presence indicates correct gamete development. In addition, it was shown recently that besides changes in mRNA levels in azoospermic men, the miRNA expression is also altered .
Another approach to TESE sample analysis is to verify whether the expression of a single gene or a couple of genes can be used as a simple indicator of positive sperm retrieval in patients undergoing treatment in infertility clinics. Detection of DAZ (deleted in azoospermia), DAZL (DAZ-like) and protamine 2 (PRM2) mRNA in testicular samples was shown to be an informative tool for spermatogenesis evaluation . Similar results were obtained with the BOULE mRNA occurrence . Ando et al. showed that expression of VASA, ODF1, ODF2 (outer dense fiber 1 and 2) and SMCP (sperm mitochondria-associated cysteine-rich protein) genes was significantly stronger in the successful TESE group . Other genes expressed in post-meiotic stage may also be good candidates for the prediction of successful fertilization.
In our study, we followed expression of five genes that are expressed at certain stages of the spermatogenesis process and are important for meiosis and sperm development. The MND1 and SPATA22 genes were selected for spermatogenesis characterization in azoospermic patients because of the known significant gene expression differences in different non-obstructive azoospermic patients . The Mnd1/Gaj protein plays an important role in homologous chromosome pairing and efficient cross-over during meiosis . The SPATA22 gene product was shown to be involved in meiotic progression of germ cells in mice . The ACR mRNA appears first in pachytene spermatocytes, reaching the maximum levels in round spermatids, and preproacrosin protein appears in spermatids . The reason for non-obstructive azoospermia in this case may be interrupted or incomplete spermiogenesis or sperm maturation. Nevertheless, the loss of acrosin protease activity does not lead to infertility in mice and spermatozoa from knock-out mice can penetrate zona pellucida of the oocyte . The gene encoding sperm-specific glyceradehyde-3-phosphate dehydrogenase, GAPDHS, was shown to be expressed solely in haploid round and elongating spermatids [30, 31] to replace the function of somatic GAPDH gene, whose expression ceases in germ cells. GAPDHS gene expression may be a good marker for spermatogenesis analysis, as its transcription and translation are temporarily separated and mRNA forms a complex with an RNA-binding protein, which results in translation and mRNA degradation delay . Therefore, expression of the GAPDHS gene might be detectable even in poor-quality or low-quantity testicular samples. Poor detection of gene expression in nine biopsies (10–12 from HS and 17–24 from MA groups) suggests that in the tissue analysed for RNA purification, spermatogenesis was either greatly reduced or RNA was probably purified mainly from somatic cells. An interesting pattern of gene expression was observed in patients 9, 22 and 23 with normal expression of MND1 and SPATA22 genes and residual levels of GAPDHS and ACR genes. This might indicate that spermatogenesis in these patients continues undisturbed until meiosis, but either meiosis or spermiogenesis is somehow impaired.
Fertilization and pregnancy rates in population of studied patients were in accordance to those from previous studies. Fertilization rates for subsets of samples with positive expression of studied genes showed uniform fertilization rates around 70%, only MND1 gene was expressed in samples from SCO group where sperm cells could not be retrieved. Surprisingly, for most promising markers of final steps of spermatogenesis, ACR and GAPDHS, pregnancy rate was below 10%. This indicates that expression analysis of present testicular genes cannot indicate successful pregnancy in studied couples. It is highly probable that in this process, oocyte and embryo quality have higher impact on the successful pregnancy. Moreover, low number of studied samples does not allow drawing any correlation between specific gene expression and fertilization outcome.
All four studied genes are expressed at different stages of spermatogenesis, and ACR, SPATA22 and GAPDHS gene expression might be a good predictor of successful TESE outcome. Nevertheless, analysis of a greater number of testicular biopsies is needed to confirm that changes in gene expression of the selected genes can serve as markers to justify repeated TESE. Another thing to consider is that spermatogenesis is a dynamic process and TESE sample analysis provides information about the gene expression and spermatogenesis state at a single time point only.
A novel non-invasive approach to prediction of the state of spermatogenesis and pathophysiology of testicular tissues via the detection of germ cell-specific mRNA traces in seminal plasma was introduced recently [33, 34]. Future analysis of germ cell-specific genes, including those from our study, or GAPDH/GAPDHS ratio in cell-free seminal plasma from azoospermic patients might become a promising non-invasive tool for TESE success prediction. The advantage of this technique is that the seminal analysis provides complex whole-testis physiology in comparison to the TESE sample representing a limited region of the analysed tissue.
To sum up, non-obstructive azoospermia is a complex pathophysiological state that leads to changes of gene expression in the testes, and understanding this process may lead to identification of the molecular markers of the spermatogenesis level.