Leydig cell injury as a consequence of an acute graft-versus-host reaction

Hematopoietic stem cell transplantation (HSCT) is the preferred therapy for malignant and nonmalignant disorders.^1,2  To maintain quality of life following HSCT, an appropriate reproductive status is required. A retrospective study conducted by the European Group for Blood and Marrow Transplantation (EBMT) determined, however, that the overall posttransplantation pregnancy rate is currently as low as 0.6%.³  These results suggested that female HSC transplant recipients become transiently or permanently infertile. Likewise, most male HSC transplant recipients are reversibly or irreversibly infertile due to germ cell injury and/or Leydig cell (LC) insufficiency.^4-6 

Posttransplantation gonadotoxicity has been attributed to the type and intensity of the pre-HSCT conditioning used. Here, myeloablative and nonmyeloablative reduced-intensity conditioning regimens can induce both spermatocyte injury and impaired LC function.^5,7,8  Gonadal dysfunction has, however, also been associated with the development of graft-versus-host disease (GVHD) because sperm counts in long-term survivors of allogeneic HSCT are much lower in GVHD patients.⁹  GVHD remains a major transplantation-related toxicity that is initiated by the recognition of host alloantigens by donor-derived mature T cells, leading to injury in a restricted set of target tissues (ie, skin, liver, gastrointestinal tract).^10,11  The mechanisms leading to diminished spermatocyte levels secondary to GVHD remain unknown. Here we tested the hypothesis that injury to intratesticular cells occurs following the infiltration of alloreactive donor T cells during an acute graft-versus-host reaction (GVHR).

Male B6.SJL-Ptprc^aPep3^b/BoyJ mice (B6.CD45.1; H-2^b, CD45.1⁺) were purchased from the Jackson Laboratories (Bar Harbor, ME) and male [C57BL/6 × DBA/2]F₁ (B6D2F₁, H-2^bd) mice were obtained from Charles River (Les Oncins, France). Animals between 6 and 10 weeks of age were kept according to federal regulations. GVHD was induced in nonirradiated B6D2F₁ mice by infusion of 60 × 10⁶ allogeneic splenocytes from B6.CD45.1 mice containing approximately 15 × 10⁶ T cells.^12,13  This transplantation model allowed us to examine in a longitudinal study the acute GVHR-mediated testicular pathology independent of tissue injury effected by chemotherapy/radiotherapy. Control mice received the same number of cells from syngeneic B6D2F₁ donors.

Testosterone in blood serum was measured by radioimmunoassay validated for murine samples.¹⁴  For analysis of the LC compartment, we used computer-aided quantitative microscopic morphometry. Transversal paraffin sections of paraformaldehyde-fixed testicular tissue (6 μm) were treated with 0.1% protease/0.5 M Tris (tris(hydroxymethyl)aminomethane) and labeled with an antibody directed against Cytochrome p450scc (Cy p450scc; Ab1244; VWR International Life Science, Lucerne, Switzerland). Cy p450scc is a key enzyme for steroid biosynthesis and, in the testis, it is exclusively localized to mitochondria in LCs.¹⁵  Bound antibody was revealed with biotin-conjugated goat antirabbit polyclonal antibody (Ab), ExtrAvidin (Sigma, Buchs, Switzerland), and aminoethylcarbazole. The aggregate area of the specific immunostaining within a random high-powered microscope field (758.6 × 508.4 μm²) was determined by detection of gray scale in digitized microscopic 2-dimensional (2-D) images (Leica DMRE microscope, Nidau, Switzerland; and Soft Imaging System, Lakewood, CO). The LC volume density (%) was then calculated by standard methods as described.^16,17 

For analysis of cytokine mRNA expression, quantitative polymerase chain reaction (PCR) was performed using SYBR Green (PE Biosystems, Rotkreuz, Switzerland), with glyceraldehyde phosphate dehydrogenase (GAPDH) as a reference.¹³  The following primers were used: interferon-γ (IFN-γ) sense: GCTGATGGGAGGAGATGTCTACAC, antisense: GACACATTCGAGTGCTGTCTGG; tumor necrosis factor-α (TNF-α) sense: AGGTCTACTTTGGAGTCATTGC, antisense: ACATTCGAGGCTCCAGTGAATTCGG; GAPDH sense: ACCATGTAGTTGAGGTCAATGAAGG, antisense: GGTGAAGGTCGGTGTGAACG. PCR data were analyzed using the comparative Ct method (ΔΔCt method) for relative quantitation of gene expression.

For immunohistochemical analyses of CD3ϵ and CD11b, frozen testicular sections (6 μm) were incubated with specific biotin-conjugated monoclonal antibodies (moAbs; clones 145-2C11 and M1/70, respectively; Pharmingen, San Diego, CA), ExtrAvidin, and aminoethylcarbazole. For confocal microscopy, frozen sections were simultaneously stained with biotin and fluorescein-conjugated moAbs to CD3ϵ and CD45.1 (clone A20), respectively. Strepatavidin-cyanin 3 (Cy3) was used as secondary reagent (Zymed Laboratories, Basel, Switzerland). Two-color immunofluorescent sections were analyzed using a Zeiss LSM510 confocal laser scanning microscope Axiovert 100 (Carl-Zeiss AG, Feldbach, Switzerland). A20 × objective lens with NA 0.5 was used. Data were analyzed with the LSM software version 3.2.

To evaluate gonadal insufficiency during acute GVHR, we assessed serum testosterone. Two to 3 weeks following the infusion of alloreactive T cells into unconditioned, haploidentical F₁ mice, median serum testosterone levels were diminished relative to mice that received syngeneic transplants (1.7; range, 0.1-4.3 vs 3.4; range, 1.2-63.5 nM; Figure 1A). As lower systemic testosterone did not per se prove injury to intratesticular cells, however, we then analyzed testosterone-producing cells using expression of Cy p450scc as marker for LC.¹⁵  In recipients of alloreactive T cells, expression levels of Cy p450scc were lower than in mice that received syngeneic transplants (Figure 1B-E). To quantify this reduction, computer-aided microscopic morphometry was employed to determine the LC volume density (Figure 1F). We found that 3 weeks after transplantation, the LC volume density significantly decreased from a normal level of 4.70% ± 1.06% to a value of 1.78% ± 1.28% in mice with acute GVHD. Since transplant recipients had not received pretransplantation conditioning, this result indicated injury to LCs (decreased LC number or, alternatively, LC dysfunction as a result of mitochondrial damage) due to an acute GVHR. This testicular injury was temporally related to GVHD-associated impairment of other organs (ie, gastrointestinal tract and liver) as evidenced by histopathology at 3 weeks after T-cell transfer (data not shown). Similar to the other target organs in mice surviving GVHD (data not shown), the LC compartment had recovered from the insult at 7 weeks after transplantation (Figure 1F).

Following activation by host professional antigen-presenting cells (APCs) in secondary lymphoid organs, local APCs recruit primed alloreactive T cells, where they elicit cutaneous and hepatic GHVD during the disease effector phase.^11,18-20  To establish whether donor T cells infiltrated the gonads, we examined testes from transplant recipients at 2 weeks after transplantation (Figure 2A-B). We found that CD3⁺ cells were very rare in normal testes but were abundant in the LC area of mice with GVHD (ie, the interstitial space separating the seminiferous tubules). Confocal microscopy confirmed that these T cells were CD45.1⁺ and were thus derived from the donor inoculum (Figure 2C). The infiltration of donor T cells peaked at 3 weeks after transplantation (data not shown), coinciding with the nadir in LC volume density. To further detail their function, testicular tissue was analyzed for cytokine expression because the activation of alloreactive T cells has been linked to secretion of the T-helper 1 (Th1) signature cytokine IFN-γ.^21,22  Quantitative reverse transcriptase-PCR (RT-PCR) analysis revealed that intratesticular transcription of IFN-γ mRNA was enhanced by more than 100-fold as a result of the GVHR (Figure 2D top panel). Expression levels were up-regulated for the entire observation period although they decreased with time.

Our results support the premise that LCs are lost or dysfunctional as a consequence of a local alloresponse, representing a mechanism that contributes to gonadal insufficiency following HSCT. Similar to other target organs,²⁰  tissue-resident professional or semiprofessional APCs²³  may recruit activated T cells, which initiate a local inflammatory immune response, to the testis. Evidence for such intratesticular inflammation was provided by enhanced numbers of intratesticular macrophages (Figure 2E-F) and increased transcripts for TNF-α (Figure 2D bottom panel) in mice with GVHD. Due to the spatial proximity of donor T cells to major histocompatibility complex (MHC)-bearing host LCs, it is conceivable that direct T-cell-mediated cytotoxicity is responsible for target cell injury. Alternatively, inflammatory cytokines such as IFN-γ and TNF-α, which impair LC function,²⁴  may affect LCs in an antigen-nonspecific mechanism similar to other target organs.²⁵  In our experimental system, we did not find evidence for T-cell infiltration into the seminiferous tubules, a fact that correlated with the absence of overt abnormalities in this compartment (data not shown). Thus, it remains to be investigated whether spermatogenesis is indirectly altered during acute GVHD as a consequence of loss and/or dysfunction of Leydig cells.

The authors thank K. Hafen, H. Schaller, and B. Erne for their expert help with morphometric analyses, confocal microscopy, and animal work. Special thanks is also extended to Dr M. Simoni and R. Sandhowe (Institute for Reproductive Medicine, University of Muenster, Muenster, Germany) for the determination of testosterone.