A defect in the peroxisomal biogenesis in germ cells induces a spermatogenic arrest at the round spermatid stage in mice

Peroxisomes are involved in the degradation of very long-chain fatty acids (VLCFAs) by β-oxidation. Besides neurological defects, peroxisomal dysfunction can also lead to testicular abnormalities. However, underlying alterations in the testes due to a peroxisomal defect are not well characterized yet. To maintain all metabolic functions, peroxisomes require an import machinery for the transport of matrix proteins. One component of this translocation machinery is PEX13. Its inactivation leads to a peroxisomal biogenesis defect. We have established a germ cell-specific KO of Pex13 to study the function of peroxisomes during spermatogenesis in mice. Exon 2 of floxed Pex13 was specifically excised in germ cells prior to meiosis by using a transgenic mouse strain carrying a STRA8 inducible Cre recombinase. Germ cell differentiation was interrupted at the round spermatid stage in Pex13 KO mice with formation of multinucleated giant cells (MNCs) and loss of mature spermatids. Due to a different cellular content in the germinal epithelium of Pex13 KO testes compared to control, whole testes biopsies were used for the analyses. Thus, differences in lipid composition and gene expression are only shown for whole testicular tissue but cannot be limited to single cells. Gas chromatography revealed an increase of shorter fatty acids and a decrease of n-6 docosapentaenoic acid (C22:5n-6) and n-3 docosahexaenoic acid (C22:6n-3), the main components of sperm plasma membranes. Representative genes of the metabolite transport and peroxisomal β-oxidation were strongly down-regulated. In addition, structural components of the blood-testis barrier (BTB) were altered. To conclude, defects in the peroxisomal compartment interfere with normal spermatogenesis.

Peroxisomes are ubiquitous eukaryotic cell organelles that are essential to maintain cellular function and homeostasis. One of their major functions is the degradation of VLCFAs and their derivatives via β-oxidation. Even though still under debate, they at least partially contribute to the synthesis of steroids and cholesterol 1 . In addition, they are involved in the biosynthesis of glycerolipids and bile acids, the catabolism of purines and polyamines, as well as the degradation of reactive oxygen species (ROS) 1 .
Peroxisomes are generally formed by growth and division 2 or are derived de novo from the ER 3,4 . Peroxisomal membrane and matrix proteins are translated in the cytosol on free ribosomes or on ER-associated ribosomes and directly targeted to peroxisomes by means of e.g. PEX19 that acts as chaperone for newly synthesized peroxisomal membrane proteins in the cytosol and directs cargo to the peroxisomal membrane and thereby functions as shuttling receptor 5,6 . The formation of the peroxisomal membrane, peroxisome proliferation and compartmentalization of peroxisomal matrix proteins is maintained by peroxins (PEX proteins) 7,8 . All peroxisomal matrix proteins harbor a peroxisomal targeting signal type 1 (PTS1) or type 2 (PTS2) at the C-or N-terminus, respectively 9 . The targeting signals are recognized in the cytosol by the cognate peroxisomal import receptors (e.g. PEX5 and PEX7) that cycle between the cytosol and the peroxisomal membranes 10,11 . The receptor/cargo complex interacts with

Materials and Methods
Generation of gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− mice. Pex13loxP (loxP: locus of crossing over of the P1 phage) transgenic mice in C57Bl/6 J background were provided by Eveline Baumgart-Vogt. The germ cell-specific deletion of exon 2 of flanked Pex13 was achieved by crossing homozygous male (or female) Pex13 loxP/loxP mice to corresponding female (or male) animals expressing Cre recombinase. Cre recombinase expression was directed by a STRA8 (stimulated by retinoic acid gene 8) genomic promoter fragment. Stra8-Cre transgenic animals in FVB/N background were obtained from Jackson laboratory (Bar Harbor, Maine, USA). To generate a congenic mouse strain, mice were initially backcrossed into C57Bl/6 strain (Charles River Laboratories, Sulzfeld, Germany), based upon marker assisted selection protocol (MASP). Single nucleotide polymorphism (SNP) genotyping was carried out by LGC genomics GmbH (Herts, UK). After the fourth backcross into C57Bl/6 mice, the offspring were less than 1% original background strain and ≥ 99% C57Bl/6.
For visualization of the peroxisomal compartment, GFP-PTS1 transgenic mice were crossed into heterozygous floxed Pex13 mice carrying the Stra8-Cre transgene. In the GFP-PTS1 transgenic mice, a fusion protein of the green fluorescent protein (GFP) and PTS1 is frequently used for visualization of peroxisomes in living cells 28 .
The mouse line was generated in the laboratory of Professor Zimmer (Department of Neurobiology; University of Bonn, Germany) by injecting a GFP-PTS1 cDNA fragment under the control of the murine Rosa26 promoter into the pronucleus of CD1 mouse zygotes 29 .
All animals went through the embryo transfer at the transgenic animal facilities at the UKE Hamburg. Mice were housed under standard conditions with free access to standard laboratory food and water and a 12 hrs dark-/light-cycle. The use of mice was in accordance with the Guide for the Care and Use of Laboratory Animals from the Institute for Laboratory Animal Research. The experiment was supervised by the institutional animal welfare officer and approved by the local licensing authority (Behörde für Soziales, Familie, Gesundheit, Verbraucherschutz; Amt für Gesundheit und Verbraucherschutz; Billstr. 80, D-20539 Hamburg, Germany). All methods were performed in accordance with the relevant guidelines and regulations by the local authorities.
Processing of testes for cryo and paraffin embedding and sectioning. Mice were anaesthetized by intraperitoneal injection using a cocktail of 100 mg/kg ketamine and 10 mg/kg xylazine and euthanized by cervical dislocation. Testicles were aseptically removed from the scrotum. The Tunica albuginea was carefully dissected and testicles were perfused with 4% paraformaldehyde (PFA; pH 7.4). For paraffin embedding, testes were immersed in 4% PFA at 4 °C overnight. Testes were transferred into phosphate buffer until embedded into paraffin (Paraplast, Sigma-Aldrich, Missouri, USA). Paraffin blocks of testes were cut into sections of 2-4 µm thickness, electron microscopy of testis sections. Mice testes were fixed by perfusion with 4% depolymerized PFA (w/v, pH 7.4). Dissected testes fragments were post-fixed with 5.5% glutaraldehyde (Merck, Darmstadt, Germany) in 0.05 M phosphate buffer (v/v) overnight. Testes fragments were further treated with 1% aqueous osmium tetroxide (Roth, Karlsruhe, Germany) for 90 min and subjected to dehydration in a series of graded alcohol (35% to absolute) before they were embedded in a mixture of 1,2,3-propanetriol glycidyl ether (Epon) and propylene oxide (Serva, Heidelberg, Germany). 1 μm semithin sections were cut to determine the region of interest. For electron microscopy, 80 nm ultrathin sections were cut. The sections were examined on an electron microscope (Philips, Amsterdam, Netherlands). Lipid analysis by gas chromatography. Total testicular triglyceride and phospholipid levels were quantified by gas chromatography. Tissue lipid extracts were prepared according to Folch et al. (1957) and total tissue fatty acid profiling was performed as already described 32 , using 20 µl of solvent per mg of tissue. Lipid classes were separated on silica gel 60 plates: 100 µl of extract was spotted onto the plate and developed with an eluent containing hexane, diethylether and acetic acid (80:20:1.5). Visualization of lipid bands was performed with primuline (5 mg in 100 ml acetone:water 80:20). Fatty acid methyl esters were prepared from 25 µl Folch extract (total FA) or of the scratched bands of PL and TG fractions without further extraction, based on the method of Lepage and Roy 33 by adding 1 ml methanol/toluene (4:1), 100 μl heptadecanoic acid (200 μg/ml in methanol/toluene, 4:1), 100 μl acetyl chloride and heating in a capped tube for 1 hr at 100 °C. After cooling to room temperature, 3 ml of 6% sodium carbonate were added. The mixture was centrifuged (1800 × g, 5 min.). 5 μl of the upper layer was diluted 1:5 with toluene and transferred to auto sampler vials. Gas chromatography analyses were performed using an HP 5890 gas chromatograph (Hewlett Packard) equipped with flame ionization detector (Stationary phase: DB-225 30 m × 0.25 mm id., film thickness 0.25 μm; Agilent, Böblingen, Germany). Peak identification and quantification were performed by comparing retention times, respectively, peak areas to standard chromatograms. All calculations are based on fatty acid methyl esters values. Concentration of individual fatty acids was calculated as % of total fatty acids. evans Blue. The azo dye Evans Blue (Sigma-Aldrich, Missouri, USA) was applied by an intravenous injection at a dose of 100 µl (50 mg/ml PBS). Adult mice of all genotypes were used. The azo dye was circulating in the blood system for 1 hr until mice were anaesthetized and intracardially perfused with 1% BSA in NaCl to remove free dye. Testes were surgically removed and fixed overnight in an additional step of 3.7% buffered formalin. For cryo-preservation, the samples were first transferred to 15% sucrose (30 min), followed by 30% sucrose (30 min) and stored at −80 °C. Prior to use, tissue samples were directly embedded into a cryo-preservative solution (Optimal Cutting Temperature, OCT, Tissue-tek ® ). Cryo-sections of 6 µm thickness were obtained on a LEICA microtome (CM3050; Wetzlar, Germany). The level of incorporated Evans Blue was assessed at 620 nm using a confocal laser microscope (Nikon Eclipse Ti NIS-Elements). total RNA isolation from testes biopsies. Total RNA from testes was isolated by phenol-chloroform extraction. Testes biopsies were taken from all genotypes. The Tunica albuginea was removed. Decapsulated www.nature.com/scientificreports www.nature.com/scientificreports/ testicles were immediately transferred to liquid nitrogen. Cooled mortar and pistil were used to mince the tissue. 1 ml TRIzol (Qiagen, Hilden, Germany) was added to the pulverized tissue and kept at room temperature. Homogenate was transferred to an Eppendorf tube and chloroform was added at one fifth of the total volume. After centrifugation (14000 x g, 20 min at 4 °C), the upper liquid phase was carefully removed and transferred to another Eppendorf tube. RNA was precipitated with isopropyl alcohol, followed by cooling at −20 °C for 30 min and centrifugation (14000 x g, 20 min at 4 °C). The pellet was washed two times with 70% ethanol and solubilized in 100 µl RNase-free water. RNA concentrations were quantified using a nanodrop ND-1000 spectrophotometer (Thermo Fisher Scientific, Massachusetts, USA).
QRt-pCR and primers used. For cDNA synthesis, purified RNA of whole testes was reverse transcribed with the RT 2 First Strand Kit (Qiagen, Hilden, Germany). Real-time PCR amplification of genes of either cell-cell junctions or fatty acid synthesis was performed on 96-well plates in a LightCycler ® 480 Real-Time PCR System (Roche, Basel, Switzerland). The mRNA level of peroxisome-related genes was evaluated with a customized RealTime ready Panel in a 384 multi-well plate format (Roche, Basel, Switzerland). QRT-PCR was performed according to manufacturer´s instructions using the SYBR Green I Master kit. 5 ng of cDNA were applied per reaction. The samples were run in triplicate following the MIQE guidelines 34 . A no-template control with RNase-free water instead of cDNA was included. Expression levels were calculated as relative values using the mean of both reference genes Gapdh and β-actin. The difference of the C T -values (ΔC T ) from the target gene and the mean of the C T -values from both reference genes were determined to quantify the target gene expression. Expression levels were further related (ΔΔC T ) to control samples using the difference of the ΔC T -value from the sample (ΔC T sample) as well as the ΔC T -value from the control (ΔC T control) and the relative values were calculated as the 2^− ΔΔCT according to Livak and Schmittgen (2001) 35 . Oligonucleotides are listed in a supplemental Table T1. statistical analysis. GraphPad Prism v.5 software was used for statistical analyses. Statistically significant differences between groups were determined using Student's t-test. A p value < 0.05 was considered statistically significant.
peroxisomal proteins showed a different distribution in the gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− testes. To confirm the Pex13 KO, we performed indirect immunofluorescent analyses in testes from adult mice (8 weeks). In the seminiferous tubules of control testes, immunoreactivity for anti-PEX13 was shown in testis-specific somatic cells (Sertoli cells and peritubular cells) and germ cells (spermatogonia, primary and secondary spermatocytes, round and elongated spermatids), with the exception for mature spermatozoa in these mice. Pex13 immunoreactivity was most intense in spermatocytes and spermatids but weaker in Sertoli-, peritubular myoid and Leydig cells (Fig. 6a,b). In germ cells with a Pex13 KO, PEX13 was not localized as punctuate structures (Fig. 6c,d). Surprisingly, gene expression analysis showed a highly significant up-regulation of Pex13 mRNA in gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− mice (Table 1). Note, that the primers for the quantification of Pex13 mRNA were designed to specifically anneal in exon 4, which was not affected by the Cre-mediated excision.
The existence of peroxisomal remnants in Pex13-deficient cells was confirmed in GFP-PTS1 transgenic gcPex-13 ∆ex2/∆ex2 /Stra8-Cre +/− mice. Whereas the GFP-PTS1 fluorescence pattern within the germinal epithelium of control mice showed characteristic peroxisomal structures (Fig. 7a,c,g,i), we found a typical expression of the GFP-PTS1 transgene in peritubular myoid cells, in the basal epithelium and in Leydig cells but not in MNCs of gcPex13 ∆ex2/∆ex-2 /Stra8-Cre +/− mice (Fig. 7d,f,j,l). Although the presence of peroxisomes was already confirmed in Sertoli cells, GFP was not expressed in these cells (Fig. 7c,f,i,l). As PEX13 directly interacts with the core protein PEX14, the effects of the Pex13 KO on the translocation machinery were further substantiated in GFP-PTS1 transgenic gcPex13 ∆ex-2/∆ex2 /Stra8-Cre +/− mice by fluorescence analysis, using anti-PEX14. In control testes, PEX14 co-localized with the GFP-PTS1 transgene and was detected in all germ cells, peritubular myoid and Leydig cells (Fig. 7a-c). However, in Pex13-deficient cells of gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− testes, PEX14 was not found as punctuate structures (Fig. 7d-f). In concordance to that, Pex14 mRNA was significantly lowered in the gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− testes (Table 1). www.nature.com/scientificreports www.nature.com/scientificreports/ The expression of the peroxisomal biogenesis factor Pex19 was likewise but not significantly decreased (Table 1). To test for a defect in the import of metabolites including long-chain unsaturated acyl-CoAs, 2-methyl branched-chain acyl-CoAs and long-chain dicarboxylic CoA esters, the localization of the peroxisomal ABC transporter ABCD3 (70 kDa peroxisomal membrane protein; PMP70) was analyzed at protein level. In control seminiferous tubules, ABCD3 showed immunoreactivity to peroxisomes in Leydig cells, Sertoli cells and in germ cells of the basal part of the germinal epithelium (Fig. 7g-i). In MNCs of gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− mice, ABCD3 was not localized in a typical punctuate pattern (Fig. 7j-l). In qRT-PCR analyses, the metabolite transporter gene Abcd1, whose mutated form is related to X-ALD, was strongly but not significantly down-regulated.
Acox1, that is involved in the dehydrogenation step during β-oxidation, together with the gene that expresses the bifunctional enzyme involved in the peroxisomal β-oxidation pathway for fatty acids, Hsd17b4, were likewise down-regulated in the gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− testes compared to control ( Table 1).
To functionally assess the integrity of the BTB, mice were injected i.v. with Evans Blue, an azo dye that binds to serum albumin. Because serum albumin cannot cross the functional BTB barrier, the adluminal compartment of seminiferous tubules would remain unstained. In healthy seminiferous tubules, Evans Blue was clearly located around spermatogonia, peritubular myoid cells and in interstitial Leydig cells (Fig. 9c). In gcPex13 ∆ex-2/∆ex2 /Stra8-Cre +/− mice, the distribution of injected Evans Blue was only slightly different. Compared to control animals, Evans Blue was partially dislocated to the adluminal compartment and showed a more irregular pattern (Fig. 9d).  www.nature.com/scientificreports www.nature.com/scientificreports/ To test for triglycerides and cholesteryl esters, cryo-sections from control and gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− mouse testes were analyzed by an Oil Red O (ORO) staining. In control testes, lipids were mainly assigned to Leydig cells (Fig. 11a,b). At stage VI of the germinal epithelium, small deposits of lipid droplets were present apically of Sertoli cells in control testes (Fig. 11b). In the gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− testes, enlarged lipid droplets were accumulated in the basal part of Sertoli cells of stage IX to XI tubules (Fig. 11d), but were also detectable in the apical compartment of the germinal epithelium (Fig. 11c).

Accumulation of lipid droplets in gcPex13
Different concentrations of fatty acids in the gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− testes. Gene expression studies on Pex13-deficient germ cells already indicated a disturbance in the metabolism of fatty acids as shown by differentially expressed genes of the peroxisomal β-oxidation pathway (Table 1). Thus, a total fatty acid analysis was done on whole testes biopsies by gas chromatography (Table 2).

Discussion
The most severe forms of peroxisomal disorders are lethal, as lipids and cholesterol are a pre-requisite for cell structure. Moreover, ROS metabolism is essential to provide a toxic-free environment. In less severe forms of peroxisomal dysfunction, as described for patients with AMN and X-ALD, testicular alterations including degenerating Leydig cells, reduction of the seminiferous tubules or even spermatogenic arrest were diagnosed. In the present study, a conditional KO of the translocation machinery constituent Pex13 was induced in pre-meiotic germ cells by using the Cre-loxP system, to analyze the effect of dysfunctional peroxisomes on spermatogenesis. We hypothesized severe disturbances on cellular level due to abolished peroxisomal function and provide an initial basic characterization of the phenotype.
According to Santos et al., tissues from Zellweger Spectrum patients contained peroxisomal membrane structures, known as ghosts that lacked peroxisomal matrix proteins 45 . They report, that some peroxisomal proteins were synthesized normally in Zellweger syndrome but not assembled into peroxisomes. Whereas some proteins were rapidly degraded and thus not detectable in Zellweger cells, some peroxisomal enzymes, as e.g. catalase 46 , were located free in the cytoplasm 45 . This observation is in concordance to the peroxisomal protein pattern shown in our immunofluorescent analyses. In the gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− testes, immunostaining not only against catalase but also peroxisomal 3-ketoacyl CoA thiolase, an enzyme that is involved in the degradation of straight chain acyl-CoAs (including the CoA esters of dicarboxylic fatty acids and eicosanoids), did not show a typical punctuate peroxisomal pattern in Pex13-deficient germ cells of gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− mice (Fig. 8j-l), as a sign for a defect in peroxisomal matrix protein import. The effects of inactivated Pex13 were already reported by Maxwell et al. 26 . Mutant Pex13 mouse pups, exhibiting the Zellweger syndrome, lacked morphologically intact peroxisomes, displayed defective peroxisome biogenesis and showed deficient import of matrix proteins. www.nature.com/scientificreports www.nature.com/scientificreports/ The elongation of FAs ≥ C16, that are either derived by food intake or de novo, occurs in the endoplasmic reticulum 47,48 . Fatty acid metabolism of short and medium chain length FAs takes place in mitochondria whereas peroxisomes metabolize VLCFA. These will then be transported into mitochondria for subsequent oxidation to CO 2 and H 2 O 49 . The analyses of total fatty acids showed a reduction of n-6 DPA (C22:5n-6) and n-3 DHA (C22:6n-3) that can either be consumed by food intake or biosynthesized from corresponding precursor fatty acids, as 18:3n-3 and 18:2n-6 50 , of which the latter was slightly increased in the gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− testes. N-6 docosatetraenoic acid (C22:4n-6) and n-3 DPA (C22:5n-3) were also increased in the gcPex13 ∆ex2/∆ex-2 /Stra8-Cre +/− testes. Interestingly, in a study from Petroni et al. (1998), increased formation of the intermediate n-3 DPA (C22:5n-3) was found in skin fibroblasts of Zellweger patients when analyzing the β-oxidation of arachidonic acid (C20:4n-6). Moreover, they could show an accumulation of adrenic acid (C22:4) in these samples by  Table 2. Total fatty acid composition up to C24, analyzed by gas chromatography of control and gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− (gcPex13KO) testes. SFAs and MUFAs were significantly increased in the gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− testes, with the exception for C16:0 (palmitic acid). C22:6n-3 (DHA) and C22:5n-6 (DPA) were significantly decreased in the gcPex13KO testes. Units are percentage of total ± SEM. Statistical significance was determined using t-test (non-parametric Mann-Whitney U test). The data of 5 biological replicates per group were considered and represented as a mean ± SD. *p ≤ 0.05; **p0.001 < p ≤ 0.01; *** p ≤ 0.001. www.nature.com/scientificreports www.nature.com/scientificreports/ HPLC analyses. The authors conclude an impaired synthesis of DHA from EPA in the steps beyond the formation of the intermediates with carbon length C22 and C24. They assume that impaired peroxisomal β-oxidation, as characteristic of Zellweger patients, leads to a redirection of the conversion of arachidonic acid through elongation/desaturation pathway with consequent accumulation of the C22:4 product 51 .
Based on our results, we do not expect a problem of fatty acid transport into germ cells 56,57 as the fatty acid levels are not equally altered. One putative conclusion could be that the different concentrations of fatty acids in the gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− testes are due to altered degradation and to altered incorporation of fatty acids in e.g. triglycerides, phospholipids which needs to be tested in further studies. Taking all data together, we can state that fatty acids being elongated were increased, whereas fatty acids that are supposed to be oxidized via peroxisomal β-oxidation, were decreased which we assume might be associated with a defect of fatty acid import into the organelle as consequence of the Pex13 KO.
However, one limitation of the present study is the lipidomic and mRNA analyses on whole testes biopsies and not on single cells: It does not allow a conclusion whether the changes in FA metabolism were the consequence of the different cellular composition of the testis (absence of spermatozoa and presence of MNCs) or if the physical membrane characteristics were altered, as already described for patients with a peroxisomal biogenesis disorder 58 . Different cell separation techniques were performed, including BSA gradient, cell cytometry, laser capture microdissection (LCM) and counterflow centrifugation elutriation (CCE) to gain single cell populations. However, all cell separation techniques were not reliable to purify MNCs from gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− mice. They were still found in all samples and could only be enriched to a purification rate of 60%. Moreover, many cytoplasmic cells lacking nuclei were found in the sample, indicating either a disruption of MNCs during the procedure or even a high apoptotic rate during spermatogenesis.
A particular focus was placed on the polarization of Sertoli cells and tight junction proteins as their major function is protecting germ cells from the circulatory and lymphatic system and therefore providing them an immune-privileged microenvironment for meiosis 59,60 , by preventing trespassing of molecules larger than 1,000 Da. The BTB is assembled by specialized junctions between Sertoli cells, including basal ectoplasmic specialized (ES) protein complexes, gap junction protein complexes and actin-based tight junction protein complexes [61][62][63] . The main contributors to maintain the BTB integrity are claudin-3 64 , 5 65 and −11 66 as well as occludin 67 . Immunolabelling against claudin-11 already indicated a severe disturbance in the structural organization of the BTB in the gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− mice (Fig. 9b). Moreover, Cldn3 gene expression was significantly down-regulated, whereas ZO1 and Ocln were slightly up-regulated (Fig. 10). Overall, the tight junction proteins were only differently expressed but not fully absent in the gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− mice.
In X-ALD patients with mutated ABCD1, the endothelial tight junction proteins, e.g. CLDN5 and ZO-1, are likewise differentially regulated, leading to a disruption of the blood-brain barrier (BBB) [68][69][70] . The functional analysis based on Evans Blue injections revealed a diffusion of the dye into the adluminal compartment, surrounding some MNCs in the gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− testes (Fig. 9d). These data indicated a disruption of Sertoli cell polarity, as the cytoskeletal marker vimentin showed an irregular pattern in the gcPex13 ∆ex2/∆ex2 /Stra8-Cre +/− testes (Fig. 8f,l), displaying not only lateral but also cytoskeletal projections from the basal compartment to the lumen, whereby surrounding single germ cells, including MNCs, in the adluminal compartment of the germinal epithelium. However, the functional integrity of the BTB must be further assessed 71 .
To conclude, the present data gives evidence for a very complex mechanism to ensure spermatogenesis and to maintain the germinal epithelium in a structured organization. The defect in the peroxisomal compartment influenced the cellular architecture as well as cell differentiation within the germinal epithelium. Moreover, the Pex13 KO caused alterations in the fatty acid synthesis and thus different levels of SFAs, MUFAs and PUFAs. However, we were not able to examine single mechanisms that could explain the present phenotype. It still remains unclear whether the failure in peroxisomal biogenesis in pre-meiotic germ cells had an effect on peroxisomal β-oxidation, leading to an imbalance of metabolites for mitochondria, or whether the organelle interplay was disturbed inducing defects in mRNA expression and thus alterations in fatty acid synthesis. Moreover, one limitation of the study was that all analyses were performed on whole testis biopsies. So far, no suitable approach was found to isolate, purify and enrich MNCs to analyze pathways and fatty acid synthesis in single cells.

Data Availability
The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.