The production of glial cell line-derived neurotrophic factor by human sertoli cells is substantially reduced in sertoli cell-only testes

Abstract

Study question: Do human Sertoli cells in testes that exhibit the Sertoli cell-only (SCO) phenotype produce substantially less glial cell line-derived neurotrophic factor (GDNF) than Sertoli cells in normal testes?

Summary answer: In human SCO testes, both the amounts of GDNF mRNA per testis and the concentration of GDNF protein per Sertoli cell are markedly reduced as compared to normal testes.

What is known already: In vivo, GDNF is required to sustain the numbers and function of mouse spermatogonial stem cells (SSCs) and their immediate progeny, transit-amplifying progenitor spermatogonia. GDNF is expressed in the human testis, and the ligand-binding domain of the GDNF receptor, GFRA1, has been detected on human SSCs. The numbers and/or function of these stem cells are markedly reduced in some infertile men, resulting in the SCO histological phenotype.

Study design, size, and duration: We determined the numbers of human spermatogonia per mm2 of seminiferous tubule surface that express GFRA1 and/or UCHL1, another marker of human SSCs. We measured GFRA1 mRNA expression in order to document the reduced numbers and/or function of SSCs in SCO testes. We quantified GDNF mRNA in testes of humans and mice, a species with GDNF-dependent SSCs. We also compared GDNF mRNA expression in human testes with normal spermatogenesis to that in testes exhibiting the SCO phenotype. As controls, we also measured transcripts encoding two other Sertoli cell products, kit ligand (KITL) and clusterin (CLU). Finally, we compared the amounts of GDNF per Sertoli cell in normal and SCO testes.

Participants/materials setting methods: Normal human testes were obtained from beating heart organ donors. Biopsies of testes from men who were infertile due to maturation arrest or the SCO phenotype were obtained as part of standard care during micro-testicular surgical sperm extraction. Cells expressing GFRA1, UCHL1 or both on whole mounts of normal human seminiferous tubules were identified by immunohistochemistry and confocal microscopy and their numbers were determined by image analysis. Human GDNF mRNA and GFRA1 mRNA were quantified by use of digital PCR and Taqman primers. Transcripts encoding mouse GDNF and human KITL, CLU and 18 S rRNA, used for normalization of data, were quantified by use of real-time PCR and Taqman primers. Finally, we used two independent methods, flow cytometric analysis of single cells and ELISA assays of homogenates of whole testis biopsies, to compare amounts of GDNF per Sertoli cell in normal and SCO testes.

Main results and the role of chance: Normal human testes contain a large population of SSCs that express GFRA1, the ligand-binding domain of the GDNF receptor. In human SCO testes, GFRA1 mRNA was detected but at markedly reduced levels. Expression of GDNF mRNA and the amount of GDNF protein per Sertoli cell were also significantly reduced in SCO testes. These results were observed in multiple, independent samples, and the reduced amount of GDNF in Sertoli cells of SCO testes was demonstrated using two different analytical approaches.

Large scale data: N/A.

Limitations, reasons for caution: There currently are no approved protocols for the in vivo manipulation of human testis GDNF concentrations. Thus, while our data suggest that insufficient GDNF may be the proximal cause of some cases of human male infertility, our results are correlative in nature.

Wider implications of the findings: We propose that insufficient GDNF expression may contribute to the infertility of some men with an SCO testicular phenotype. If their testes contain some SSCs, an approach that increases their testicular GDNF concentrations might expand stem cell numbers and possibly sperm production.

Study funding/competing interest(s): This research was funded by the Eunice Kennedy Shriver National Institute for Child Health and Human Development, National Centers for Translational Research in Reproduction and Infertility Program (NCTRI) Grant 1R01HD074542-04, as well as grants R01 HD076412-02 and P01 HD075795-02 and the U.S.-Israel Binational Science Foundation. Support for this research was also provided by NIH P50 HD076210, the Robert Dow Foundation, the Frederick & Theresa Dow Wallace Fund of the New York Community Trust and the Brady Urological Foundation. There are no competing interests.

Keywords: GDNF; human; progenitor spermatogonia; spermatogonial stem cells.

Figure 1

Co-localization of GFRA1⁺ and/or UCHL1⁺ spermatogonia of human seminiferous tubules. Tubules were immunostained for both proteins and 2.3 μm optical sections were captured by confocal microscopy. Green (UCHL1) and red (GFRA1) channels were captured separately and then merged. Results for expression of UCHL1 (A, F, J) alone, GFRA1 (B, E, H, I) alone, and both proteins together (C,D, G) are shown. These images reveal considerable heterogeneity in the intensity of cellular expression of GFRA1. Note, too, that while many cells express both proteins (see box in 1C as well as D, E and F), some cells only express UCHL1 (see white arrows in C, G and H), and others only express GFRA1 (see yellow arrows in C, I and J). Inserts in upper right hand corner of A and B are representative images of negative controls for UCHL1 and GFRA1 immunostaining. Analysis of tubules from three different donors produced similar results.

Figure 2

A comparison of the expression of GDNF mRNA by human and mouse testes. GDNF mRNA in human and mouse testes (n = 6 for both species) were quantified by use of digital PCR (human) and real-time quantitative PCR (mouse). Data (mean + SEM) are expressed as amount of GDNF mRNA normalized to the amount of 18 s rRNA in the same sample. Species-specific Taqman primers were used for this analysis.

Figure 3

Expression of GFRA1 mRNA and DDX4 mRNA in human testes with normal spermatogenesis and in human testes diagnosed with non-obstructive azoospermia with a Sertoli cell-only phenotype, NOA(SCO). GFRA1 mRNA was assayed by digital PCR; DDX4 mRNA was assayed by quantitative real-time PCR, and the amounts of both transcripts were normalized to the amount of 18S rRNA in each sample. Results for DDX4 mRNA are presented on a log scale. Data (n = 5/group) are expressed as mean + SEM and demonstrate that expression of both transcripts are significantly reduced in SCO testes.

Figure 4

Expression of transcripts encoding GDNF (A) , kit ligand (KITL) (B) and Clusterin (CLU) (C) in normal human testes, in testes of men diagnosed with non-obstructive azoospermia with maturation arrest of spermatogenesis, NOA (MA), and in testes of men diagnosed with non-obstructive azoospermia with a SOC phenotype, NOA SCO). Data (n = 6/group; mean + SEM) are normalized to the amount of 18S rRNA in each sample. Bars marked with different letters differ statistically. The data presented here for expression of GDNF mRNA in normal human testes are also presented in Fig. 2.

Figure 5

Analysis of GDNF protein expression by Sertoli cells in testes with normal spermatogenesis and by Sertoli cells of men diagnosed with non-obstructive azoospermia and an SCO histological phenotype. (A) Flow cytometric analysis of GDNF expression by individual cells. Single cells were obtained from biopsies of normal testes and from biopsies of SCO testes. Cells were fixed, permeablized and incubated with antibodies to GDNF and SOX9 labeled with fluorochromes, AF-647 and AF-448, respectively. Fluorescence from single cells was analyzed by flow cytometry, with single cells identified by size. Results show that in normal testes, there is a single population of cells that express significant amounts of GDNF, and these cells also express the Sertoli cell marker, SOX9. In contrast, SCO testes contain two populations of GDNF⁺, SOX9⁺ cells. The larger of the two populations of Sertoli cells express substantially less GDNF than do Sertoli cells from normal testes. These results are representative of independent analyses of cells from three normal testes and four SCO testes. Similar results were obtained with two different antibodies to GDNF. (B) Quantification of GDNF in biopsies of normal human testes and biopsies of SCO testes. Data (n = 5/group; mean + SEM) are expressed as ng of GDNF/μg of vimentin, in the same biopsy, and confirm a significant reduction in the amount of GDNF per Sertoli cell in SCO testes (P = 0.014).