Loss of Ubiquitin Carboxy-Terminal Hydrolase L1 Impairs Long-Term Differentiation Competence and Metabolic Regulation in Murine Spermatogonial Stem Cells

Spermatogonia are stem and progenitor cells responsible for maintaining mammalian spermatogenesis. Preserving the balance between self-renewal of spermatogonial stem cells (SSCs) and differentiation is critical for spermatogenesis and fertility. Ubiquitin carboxy-terminal hydrolase-L1 (UCH-L1) is highly expressed in spermatogonia of many species; however, its functional role has not been identified. Here, we aimed to understand the role of UCH-L1 in murine spermatogonia using a Uch-l1−/− mouse model. We confirmed that UCH-L1 is expressed in undifferentiated and early-differentiating spermatogonia in the post-natal mammalian testis. The Uch-l1−/− mice showed reduced testis weight and progressive degeneration of seminiferous tubules. Single-cell transcriptome analysis detected a dysregulated metabolic profile in spermatogonia of Uch-l1−/− compared to wild-type mice. Furthermore, cultured Uch-l1−/− SSCs had decreased capacity in regenerating full spermatogenesis after transplantation in vivo and accelerated oxidative phosphorylation (OXPHOS) during maintenance in vitro. Together, these results indicate that the absence of UCH-L1 impacts the maintenance of SSC homeostasis and metabolism and impacts the differentiation competence. Metabolic perturbations associated with loss of UCH-L1 appear to underlie a reduced capacity for supporting spermatogenesis and fertility with age. This work is one step further in understanding the complex regulatory circuits underlying SSC function.


Introduction
Mammalian spermatogenesis is a complex, highly regulated process of cell proliferation and differentiation sustained by spermatogonial stem cells (SSCs). These cells give rise to progenitors that undergo a finite number of synchronous cell divisions before entering meiosis as a spermatocyte [1]. Previously considered a cell population committed to differentiation, an early subset of progenitors can de-differentiate and regain stem cell capacity, which highlights the plasticity of the undifferentiated, mitotic germ cell state [2,3].
Clear identification of a definitive SSC has remained elusive, due to the lack of biochemical and molecular markers that unequivocally distinguish SSCs from other undifferentiated spermatogonia. Several putative SSC markers have been described, such as inhibitor of DNA binding 4 (Id4), paired box protein Pax-7 (Pax7), glial cell-line derived neurotrophic factors receptor alpha 1 (Gfra1) and promyelocytic leukemia zinc finger protein (Plzf ). However, germ cell populations that express these markers are generally regarded as a mixed population containing SSCs among other undifferentiated spermatogonia [4][5][6]. Unlike in other stem cells, tyrosine-kinase receptor c-Kit neither is expressed in SSCs in vivo nor plays a role in their maintenance, but it is critical to the primordial germ cells' movement towards the basement membrane, where they give rise to SSCs [7]. A study showed, however, that c-KIT positive and c-KIT negative cultured germ cells displayed comparable SSC activity after transplantation [8]. On the other hand, stimulated by retinoic acid gene 8 (Stra8), which is the receptor for retinoic acid and a key molecule in initiating meiosis [9], also plays a role in SSC maintenance, proliferation and self-renewal [10].
The process of spermatogenesis through mitotic, meiotic, and post-meiotic phases is governed by several cellular processes. Evidence underlines the key role of ubiquitination in the derivation of SSCs and differentiation of spermatogonia from gonocytes [11], in genetic recombination [12] and in sex chromosome silencing during meiosis [13]. Ubiquitination is a post-translational modification that regulates several cellular processes such as protein turnover, DNA damage response and intracellular signaling [10][11][12], which can be reversed through the action of de-ubiquitinase enzymes [14].
Ubiquitin carboxy-terminal hydrolase 1 (UCH-L1), also known as PGP9.5, is classified as a de-ubiquitinase enzyme [15] that is primarily expressed in the nervous system and testes [15,16]. Beside its role in catalyzing the hydrolysis of ubiquitin from its target and thus preventing protein degradation [15], UCH-L1 can perform a variety of other functions, including binding and sequestering mono-ubiquitin, signaling for protein translocation within the cell, and stabilizing ubiquitin modifications [17,18].
The role of UCH-L1 in maintaining cell homeostasis under normal growth and oxidative stress conditions has been reported in vitro [19]. Evidence links the development of gracile axonal dystrophy [20,21], Parkinson's and Alzheimer's diseases to mutations or deletions in the Uch-l1 gene [22][23][24][25]. Hussain et al. (2013) demonstrated the importance of UCH-L1 in regulating the balance between mechanistic Target of Rapamycin (mTOR) complexes by disrupting mTOR complex 1 (mTORC1) and promoting mTORC2 assembly [26], which plays a role in preventing brain degeneration and development of malignancies [27]. UCH-L1 deficiency, associated with enhanced mTORC1 activity, causes an early increased protein turnover, which leads to progressive neurodegeneration in Uch-l1 deficient mice [28].
To date, attempts to describe the potential function for UCH-L1 in the testis are limited to a few studies of mutant testes that either carry a spontaneous in-frame deletion of Uch-l1 [19] or overexpress Uch-l1 in all tissues [29]. The testis of a mouse model that constitutively over-expresses UCH-L1 showed a meiotic block at the early pachytene spermatocyte stage and a reduction in proliferating spermatogonia, suggesting a role for UCH-L1 in spermatogonial proliferation and differentiation [29]. Conversely, the testes of a gracile axonal dystrophy (gad) mouse that had a spontaneous in-frame deletion of Uch-l1 showed a decrease in proliferating cells, tubule size and cell numbers along with an increase in apoptosis and size of the remaining cells within the seminiferous tubules at 25 weeks of age [19]. Subsequent studies reported gad testes being resistant to cryptorchidinduced injury and early apoptotic first wave, suggesting a role in BAX/BCL-2 (BCL2 associated X, apoptosis regulator, B-cell lymphoma 2)-regulated apoptosis [30]. Yet the role of UCH-L1 and the site of the ubiquitination target in the testes remain to be further determined [20]. Loss of UCH-L1 does not result in compensatory up-regulation of other isoforms, suggesting a distinct, non-redundant cellular role of UCH-L1 [20].
Here we show that UCH-L1 is expressed in undifferentiated spermatogonia (SSC/ progenitors) as well as in early differentiating spermatogonia. Taking advantage of the Uch-l1 −/− mouse model that lacks all UCH-L1 function, we found an increased number of degenerative seminiferous tubules in Uchl1 −/− compared to Uch-l1 +/+ , a condition that is preceded by a dysregulation of normal metabolic transitions in undifferentiated spermatogonia, as revealed by single cell RNA sequencing analysis. We further demonstrate a loss of differentiation competence of SSCs after transplantation in vivo and accelerated mitochondrial activity in vitro. These data point towards an important role of UCH-L1 in murine spermatogenesis. Loss of UCH-L1 is associated with reduced capacity of SSCs to support spermatogenesis and accompanied by increased oxidative phosphorylation (OXPHOS), mitochondrial dysfunction and metabolic perturbations. The disturbance of the tight regulation of normal metabolic transitions necessary for murine spermatogenesis associated with the loss of UCH-L1 could ultimately result in reduced fertility with age.
This work provides novel insight into the functional implication of UCH-L1 for the SSC metabolic homeostasis that is required for long-term metabolic plasticity and stem cell function.

Mice
Uch-l1 tm1Dgen (Deltagen Inc., Menlo Park, CA, USA), B6/SV129 and ROSA26 reporter line (B6.129S7-Gt(ROSA)26Sor/J) mice were obtained from the Jackson Laboratory. All animal experiments were carried out in compliance with the Canadian Council of Animal Care guidelines and approved by the University of Calgary's Health Sciences Animal Care Committee (AC20-0058, Renewal 5 August 2020).

TUNEL Assay
The DeadEnd TM Fluorometric TUNEL system (488) (G3250, Promega, Madison, WI, USA) was used to detect and quantify cell apoptosis in tissue sections. The labeling was performed according to the manufacturer's recommendations and nuclei were stained with DAPI. The assay catalytically incorporates fluorescein-12-dUTP at 3 -OH ends of DNA, which allows direct visualization of labeled DNA (in green) through fluorescence microscopy.

Quantitative PCR
RNA was isolated from 7, 17 and 120-day old mice testis and primary cultured cells using RNeasy Mini Kit (74106, Qiagen, Toronto, ON, Canada). RNA was reversetranscribed into cDNA using ABI High Capacity cDNA synthesis kit (4368814, Fisher Scientific, Ottawa, ON, Canada) according to the manufacturer's instructions. cDNA amplification was carried out in an ABI 7500 Fast qPCR machine (721BRO1344, Thermo Fisher Scientific) using Bio-Rad Sso Fast Eva Green master mix with low Rox (172-5212, Bio-Rad laboratories, Mississauga, ON, Canada). The qPCR results were baseline-corrected using LinReg PCR. Primer specificity was confirmed by sequencing RT-PCR amplified products and melt curve analysis (Primers list: Table S1).

Germ Cell Transplantation
To enable the identification and quantification of cells and colonies transplanted into a germ cell-depleted recipient testis, Uch-l1 −/− mouse (B6;129P2-Uchl1tm1Dgen/Mmnc) was crossed with a ROSA26 reporter line (B6.129S7-Gt(ROSA)26Sor/J) expressing the lacZ transgene. Spermatogonia collected from Uch-l1 +/+ , Uch-l1 −/− or heterozygous gtROSA 6 to 8-day old pups, which ubiquitously express the LacZ reporter gene, were cultured for 3 to 5 passages prior to transplantation. Approximately one million cells contained in 8 µL were transplanted into six-week old B6/SV129 mice via the efferent duct during week 6 after busulfan treatment (50 mg/kg) (B2635, Sigma) as previously described [31]. Three independent cell preparations per genotype were used for transplantation. Testes from recipient B6/SV129 males were collected two and 10 months post-transplantation. To visualize donor-derived spermatogenesis, testes were fixed in 4% PFA for 2 h at 4 • C, incubated with X-gal substrate at 4 • C overnight, then kept for another 2-4 h at 37 • C until the donor-derived cells turned deep blue. Colonies were counted and individual colony length measured. The blue intensity of colonized tubules was qualitatively scored against an open reference of three color intensities (light, medium and dark blue). Each colony was measured for the percentage that was light, medium and dark blue. Seminiferous tubules were collected, fixed in Bouin's solution for 2 h at room temperature, processed and embedded in paraffin. The embedded sections were used for histological analysis of donor-derived colonies.

Single-Cell RNA-Sequencing
Testicular cells were collected via enzymatic digestion from two 8-week old mice from each genotype. Cells were stained with goat anti GFRα 1 protein (GT15004, Neuromics, Edina, MN, USA) and rat anti c-KIT protein (ab65525, Abcam) and subjected to fluorescence-activated cell sorting (FACS) to enrich for spermatogonia using a FACSAria III (BD Biosciences). Briefly, previous experiments have identified a population of cells based on light scatter properties that is enriched in spermatogonia ( Figure S1A). All c-KIT+ and GFRa1+ cells (c-KIT+ GFRa1-, c-KIT+ GFRa1+, and c-KIT-GFRa1+) are within this spermatogonia-enriched cell population (Figure S1B-D). Isolated primary spermatogonia samples were resuspended in PBS supplemented with 1% BSA and 6000-10,000 cells per sample were processed using Chromium TM Single Cell 3 v2 Chemistry library prep kit according to the manufacturer's instructions (10X Genomics, San Francisco, CA, USA). Samples were sequenced on an Illumina HiSeq4000 with paired end 50 base pair reads by Genome Quebec (Montreal, QC, Canada). The raw counts were subsequently processed with a 10x Genomics CellRanger pipeline using default settings, after which the outputs from the four individual samples were combined using the CellRanger aggr function to normalize the sequencing depth and produce aggregated gene-cell UMI count matrix. A total of 7874 single germ cells (spermatogonia to spermatids) was used for the downstream analysis.

Single Cell RNA Sequencing Analyses
The concatenated gene-cell barcode matrix was imported into Seurat3.1.0 [34]. A prefiltration step to remove low quality cells was performed and only cells expressing ≥ 200 detected genes and genes expressed in ≥3 cells were kept for the downstream analysis. The individual samples were then subjected to the standard Seurat pre-processing workflow to select and filter high quality cells based on the QC metrics (genes detected, UMI counts and mitochondrial genes). To perform a comparative analysis between the two genotypes (Uch-l1 −/− vs. Uch-l1 +/+ ), Seurat V3 integration workflow was used to merge, correct for the batch effect and generate a new integrated Seurat matrix that was used for further analysis. To identify the dimensionality of the dataset, the "ElbowPlot" function was applied, and the first 15 Principal components (PCA) were selected to run the "FindNeighbors", "FindClusters" and "RunUMAP/RunTSNE" functions with a resolution of 0.5. Once clusters were generated, the "FindMarkers" function was applied to identify differentially expressed genes (DEG) with p values lower than 0.01 and fold change of 1 or higher. The "VlnPlot", "FeaturePlot", and "Dimheatmap" functions were used to visualize genes of interest.

Gene Set Enrichment Analysis (GSEA)
GSEA was performed with the javaGSEA application (version 4.0.3) available online (javaGSEA. http://www.broadinstitute.org/gsea/downloads.jsp, accessed on 24 August 2021) with default settings [35] (http://software.broadinstitute.org/gsea/, accessed on 24 August 2021). A pre-ranked list consisting of DEGs and their corresponding foldchange values was used as an input to determine the enrichment scores for pathways. The curated gene sets used for reference were downloaded from Molecular Signatures Database (MSigDB) [36]. The enriched pathways were displayed as a heat map using the "heatmap.2" function in R studio. The gene network was built with STRING [37].

Statistical Analysis
Quantitative data are displayed as mean + standard deviation. Non-parametric t-tests (Mann-Whitney U tests) were performed when comparing two data sets and Kruskal-Wallis ANOVA for three data sets according to statistical guidelines. Significance was set at p < 0.05. For the Seahorse experiment, the results are presented as mean ± SEM, if not otherwise stated. Normality was tested with Shapiro-Wilk test (if possible, by sample size via Anderson-Darling test, D'Agostino-Pearson test and Kolmogorov-Smirnov test); normality was assumed if all performed tests indicated a normal distribution. Normally distributed data was analyzed with an unpaired parametric t-test, otherwise a non-parametric t-test (Mann-Whitney test) was used to compare ranks. Unequal standard deviation was accounted for with Welch's correction. A value of p < 0.05 was considered statistically significant. GraphPad Prism 8.0 software was used for all statistical analyses.

UCH-L1 Is Expressed in Mouse Spermatogonia and Sertoli Cells
To examine the expression pattern of UCH-L1 in testes, immunohistochemistry was performed on PFA-fixed, paraffin-embedded sections from Uch-l1 +/+ mouse testes collected at different stages of development. As in other mammalian species, UCH-L1 in mice is expressed in cells located along the basement membrane of seminiferous tubules but not in further differentiating germ cells ( Figure 1; Figure S2). Figure 1A illustrates the expression of UCH-L1 and SRY-box 9 (SOX9) from day 0 to day 120, along with the expression of synaptonemal complex protein (SPC3), a marker specifically expressed during meiosis in spermatocytes. At day 30, the expression of UCH-L1 was detected in Sertoli cells as well ( Figure 1B-M), which is an expression specific to rodents [38]. In addition, UCH-L1 is co-expressed with LIN28A, a marker largely expressed in the spermatogonial progenitor population [39]. All LIN28A positive spermatogonia were UCH-L1 positive and all UCH-L1 positive cells that were not Sertoli cells were LIN28A positive ( Figure 1B-M).

Uch-l1 −/− Mice Have Decreased Testis Weight and Increased Degeneration of Seminiferous Tubules
To determine whether UCH-L1 deficiency affects the morphology and cell composition of the seminiferous tubules, we histologically examined testes collected from Uch-l1 +/+ , Uch-l1 +/− and Uch-l1 −/− mice at different time points of testis development. Due to the strong neurodegenerative phenotype of the Uch-l1 −/− mice, 120 days of age was ethically dictated as a final time point. In 7-day-old mice, only spermatogonia were detected in the seminiferous epithelium, followed by the first wave of spermatogenesis, with the appearance of primary spermatocytes at day 17 of age. By day 30, the first wave of spermatogenesis was completed and the 60, 90 and 120-day time points captured subsequent spermatogenic waves. Testes generally appeared histologically normal and achieved full spermatogenesis regardless of the genotype (Figure 2A-C).  However, a significant difference in body weight and testis weight was detected at days 90 and 120 between the two genotypes, while they were indistinguishable at earlier stages. The body weight of 90 and 120-day old Uch-l1 −/− males was significantly lower than Uch-l1 +/− and Uch-l1 +/+ males, which was reflective of their progressive neurodegeneration ( Figure 2D). There was also a significant decrease in testis weight at 120 days of age in Uch-l1 −/− compared to Uch-l1 +/− and Uch-l1 +/+ genotypes ( Figure 2E). Furthermore, changes in testicular weight were consistent with an increase in the number of degenerative seminiferous tubules in Uch-l1 −/− compared to Uch-1l +/+ testes at days 90 and 120 of age ( Figure  2F-H). However, a significant difference in body weight and testis weight was detected at days 90 and 120 between the two genotypes, while they were indistinguishable at earlier stages. The body weight of 90 and 120-day old Uch-l1 −/− males was significantly lower than Uch-l1 +/− and Uch-l1 +/+ males, which was reflective of their progressive neurodegeneration ( Figure 2D). There was also a significant decrease in testis weight at 120 days of age in Uch-l1 −/− compared to Uch-l1 +/− and Uch-l1 +/+ genotypes ( Figure 2E). Furthermore, changes in testicular weight were consistent with an increase in the number of degenerative seminiferous tubules in Uch-l1 −/− compared to Uch-1l +/+ testes at days 90 and 120 of age ( Figure 2F-H). wild-type females. There was a significantly lower number of confirmed matings in the Uch-l1 −/− males compared to Uch-l1 +/+ , which may indicate an issue with copulation ( Figure  2I). However, the average size of the few Uch-l1 −/− litters produced remained normal with healthy pups. The progressive decrease in testis weight together with an increase in the degeneration of seminiferous tubules may impact the fertility potential in Uch-l1 −/− mice as they age.  To assess whether these morphological defects impacted the ability of Uch-l1 knockout mice to reproduce, Uch-l1 +/+ and Uch-l1 −/− males at 2-3 months of age were mated with wild-type females. There was a significantly lower number of confirmed matings in the Uch-l1 −/− males compared to Uch-l1 +/+ , which may indicate an issue with copulation ( Figure 2I). However, the average size of the few Uch-l1 −/− litters produced remained normal with healthy pups. The progressive decrease in testis weight together with an increase in the degeneration of seminiferous tubules may impact the fertility potential in Uch-l1 −/− mice as they age.
As the only cells detected in degenerated seminiferous tubules were spermatogonia and Sertoli cells, we used the TUNEL assay on testes sections and qPCR for Bcl2 and Bax gene expression to assess if cell apoptosis may account for the loss of differentiated germ cells. No difference was observed in the number of TUNEL-positive cells per seminiferous tubules ( Figure S3A-D) or in the Bcl2/Bax ratio between the two genotypes at any given time point. First, our data do not corroborate the implication of UCH-L1-induced apoptosis during the first wave of spermatogenesis, as previously described in the gad mouse model. The Uch-l1 −/− mouse model used in our study did not show any resistance to early wave apoptosis during the pre-pubertal stage of development, which is essential for functional spermatogenesis in adults [30]. Second, our results might be suggestive of a decrease in differentiation competence of Uch-l1 −/− germ cells, rather than a loss of germ cells by apoptosis, since no difference was observed in cell apoptosis between the two genotypes at 120 days. Differences were also detected in relative expression of germ cell associated genes reflective of differentiation status during development ( Figure 3A). At 7 days, a significant increase in stem cell-associated gene expression for Id4 and a decrease in Gpr125 gene expression was present in Uch-l1 −/− compared to Uch-l1 +/+ testes. At 17 days old, there was an overall decrease in gene expression associated with stem cells (Id4) as well as differentiating germ cells (Dazl, Uch-l3, Scp3, Pgk2, Vasa). At 120 days, a significant increase was detected in progenitor cell differentiation-associated Sall4 gene expression, associated with a significant decrease in differentiation-associated c-Kit expression in Uch-l1 −/− compared to Uch-l1 +/+ testes ( Figure 3A). These observations were further supported by the flow cytometry analysis ( Figure 3B,C) showing fewer PLZF and c-KIT positive cells in Uch-l1 −/− testis cells ( Figure 3D-F).

Single Cell RNA Sequencing Analysis Reveals Similar Path Reflecting Stages of Spermatogenesis in Both Genotypes
To capture any molecular changes that may emerge in germ cells due to the absence of Uch-l1, single-cell RNA sequencing was performed on 5000-8000 sorted germ cells (GFRa1 + /c-KIT + cells) collected from two Uch-l1 −/− and two Uch-l1 +/+ mice. Seurat integrated analysis workflow was first used to merge the replicates from each genotype, correct for batch effect and perform the downstream analysis. Unsupervised, unbiased clustering projected onto t-distributed stochastic neighbor embedding (tSNE) revealed 10 and 12 cell clusters for Uch-l1 +/+ ( Figure 4A) and Uch-l1 −/− ( Figure 4B), respectively. In both genotypes, cell clusters were arranged in a continuous path that reflected different stages of spermatogenesis, including spermatogonia (SPG), spermatocytes and spermatids. SPG was further divided into four cell states, previously identified in different studies [40][41][42][43] and species [44]. The differentially expressed genes (DEG) that distinguish between these clusters in both genotypes are depicted in supplemental Figure S4A,B. As shown in Figure 4C, SPG-1 that encompassed SSC and progenitors were characterized by the expression of Gfra1, Tcl1, Etv5, and Nanos3 and lack of expression of differentiation markers. SPG-2, identified as early differentiating cells, expressed both c-Kit and Stra8 and lacked meiotic gene expression. SPG-3, which was further divided into three sub-clusters, were c-Kit + /Stra8 − , and expressed the meiotic gene Sycp3 at low levels. Cells in SPG-4 were c-Kit − /Stra8 + and expressed high levels of Sycp3 [40][41][42][43]. Of note, SPG-4 was organized into two distinct clusters in Uch-l1 −/− , while it showed only one cluster in Uch-l1 +/+ . Although the number of cells present in clusters 7, 8, 9, 10, 11 and 12 in both genotypes was low, the analysis showed a clear distinction between the early and late meiotic process (spermatocytes and spermatids). Overall, there was no difference between the two genotypes in the expression of genes that mark SSC/progenitors, differentiating and differentiated cells. However, this assessment represents only a snapshot of a single time point, which might not represent further phenotypic heterogeneity that can emerge due to UCH-L1 deficiency with aging.

Single-Cell RNA Sequencing Analysis Shows an Altered Profile of Metabolic Transitions in Uch-l1 −/− SSC/Progenitors Populations
To determine whether germ cells displayed any functional change that would precede or accompany the degenerative phenotype of the seminiferous tubules, we used GSEA to identify enriched pathways whose genes were differentially expressed between cell clusters. Heat maps that display the enriched pathways for the two genotypes are shown in Figure

Single-Cell RNA Sequencing Analysis Shows an Altered Profile of Metabolic Transitions in Uch-l1 −/− SSC/Progenitors Populations
To determine whether germ cells displayed any functional change that would precede or accompany the degenerative phenotype of the seminiferous tubules, we used GSEA to identify enriched pathways whose genes were differentially expressed between cell clusters. Heat maps that display the enriched pathways for the two genotypes are shown in Figure 4D. The data showed an upregulation of gene sets associated with OXPHOS and cMyc signaling pathways in Uch-l1 −/− SPG1, while eight gene sets were significantly upregulated in Uch-l1 +/+ SPG-1, including hypoxia, p53 and cMyc signaling pathways. These data suggest the knockout of Uch-l1 caused the SSCs/progenitors (SPG1) to shift their metabolism towards an oxidative metabolism as compared to the wild-type. Interestingly, the early-differentiating cell population (SPG2) showed an abnormal increase of glycolytic and hypoxic response-related genes in Uch-l1 −/− , while there was an upregulation of OXPHOS-related genes in the wild type. This indicates an imbalance of normal metabolic transition after the loss of UCH-L1 in murine germ cells.
An enrichment of gene sets related to a higher cell proliferative state, namely G2M checkpoint, E2F_targets, and mitotic spindle, was observed in SPG3 from both genotypes. However, gene sets associated with the unfolded proteins pathway were upregulated in the late differentiating state (SPG4) in the knockout, while they were enriched early in the SPG3 population. The unfolded protein response (UPR) is a conserved cell process, induced in response to an accumulation of unfolded or misfolded proteins in the endoplasmic reticulum [45], which plays a role in cell homeostasis [46].
To further examine whether the difference observed in enriched pathways in SPG1 was not due to batch effects between the two genotypes, and to achieve higher statistical power for differential expression analysis of SSC/progenitors, Seurat integrated analysis workflow was used to merge the four replicates (two from each genotype), correct for batch effect and perform the downstream analysis. Unsupervised clustering grouped the four transcriptomes into 12 distinct clusters, as illustrated by Uniform Manifold Approximation and Projection (UMAP) (Figure 5A,B). As expected, the clusters were organised into a continuous path as described above, showing the presence of four states of spermatogonia (SPG1-4), spermatocytes and spermatids ( Figure 5A, Figure S4C). When extracted and reanalysed with a higher resolution, SPG1 showed further subdivision into seven individual clusters ( Figure 5C,D), where both genotypes contributed to clusters 0, 1, 3, 4, 5 and 6, while cluster 2 was exclusively represented by the Uch-l1 −/− genotype in one replicate ( Figure S5A). Comparing the expression of some SSC/progenitor markers between the knockout and wild-type showed no differences, except for Dmrt1 ( Figure S5B). DMRT1 plays a role in maintaining the SSC pool under normal conditions and allowing progenitors to help restore the SSC pool when the germ cells are depleted [47]. The significantly identified DEGs in Uch-l1 −/− (as compared to Uch-l1 +/+ ) were loaded into GSEA to search for the enriched gene sets. Here again we confirmed the upregulation of gene sets related to OXPHOS in cluster 2 of SSCs/progenitors ( Figure 5E), which was found only in the knockout. The 11 genes that significantly contributed to the enriched score of OXPHOS are presented as a network in Figure 5F. Collectively, these data suggest that UCH-l1 deficiency promotes a shift towards an oxidative metabolism in undifferentiated spermatogonia and a loss of normal metabolic transitions during SSC maintenance and differentiation [37,39,45]. Cultured SSCs derived from Uch-l1 +/+ or Uch-l1 −/− mice were subsequently transplanted into the seminiferous tubules of busulfan-treated mouse testes. Testes were collected at two or 10 months post-transplantation to determine stem cell number (colony number), stem cell capacity (colony length) and score of a complete spermatogenesis of the transplanted cell population ( Figure 6A-D). As illustrated in Figure 6E,F, no differences were found between Uch-l1 +/+ and Uch-l1 −/− colony numbers or colony length at two months post-transplantation. At 10 months post-transplantation, there was no difference in the colony number between the two genotypes ( Figure 6G). Due to the increased length of colonies at the 10-month collection time point, the incidence of merged colonies was high; therefore, individual colony length could not be assessed accurately. When the total length of colonized seminiferous

Cultured Uch-L1 −/− SSCs Show a Decreased Capacity for Regeneration of Full Spermatogenesis
The neurodegenerative phenotype of the Uch-l1 −/− mice limited the in vivo examination of spermatogonia to 120 days. Additionally, the expression of UCH-L1 in Sertoli cells in adult mice precludes interpretation of differences found between the Uch-l1 −/− and Uch-l1 +/+ testes, as it cannot be unequivocally determined if any observed phenotype was due to the absence of UCH-L1 in Sertoli cells or spermatogonia. Therefore, transplantation of Uch-l1 +/+ and Uch-l1 −/− spermatogonia into syngeneic mice was used to overcome these limitations.
First, we examined the impact of Uch-l1 knockout on the growth and survival of spermatogonia in vitro. Testicular cells isolated from Uch-l1 +/+ and Uch-l1 −/− 6 to 8-day old mice were maintained in culture for several passages and then collected for analysis on days 3 and 6 from passages 3, 5 and 8 ( Figure S6A,B). Cell number, viability and proliferation were not significantly different between the two genotypes ( Figure S6C,D). Gene expression for several spermatogonia-associated genes (Id4, Plzf, Lin28a, c-Kit and Stra8) was similar between genotypes ( Figure S6E).
Cultured SSCs derived from Uch-l1 +/+ or Uch-l1 −/− mice were subsequently transplanted into the seminiferous tubules of busulfan-treated mouse testes. Testes were collected at two or 10 months post-transplantation to determine stem cell number (colony number), stem cell capacity (colony length) and score of a complete spermatogenesis of the transplanted cell population ( Figure 6A-D). As illustrated in Figure 6E,F, no differences were found between Uch-l1 +/+ and Uch-l1 −/− colony numbers or colony length at two months post-transplantation. At 10 months post-transplantation, there was no difference in the colony number between the two genotypes ( Figure 6G). Due to the increased length of colonies at the 10-month collection time point, the incidence of merged colonies was high; therefore, individual colony length could not be assessed accurately. When the total length of colonized seminiferous tubules was measured, Uch-l1 −/− recipient testes had significantly higher total colonized tubule length compared to Uch-l1 +/+ recipients ( Figure 6H). The intensity of the blue X-gal staining was scored in three categories: light, medium and dark blue ( Figure 6I), reflecting the number of differentiating germ cell layers [48]. Dark blue colonies had complete donor-derived spermatogenesis, medium blue colonies had fewer germ cell layers and light blue colonies had only one or very few layers of donorderived cells. The intensity of the blue X-gal staining was lighter in the Uch-l1 −/− colonies than the Uch-l1 +/+ colonies (represented in Figure 6B,D,J). A larger proportion of Uch-l1 −/− colonies were medium blue and a smaller proportion were dark blue as compared to the Uch-l1 +/+ colonies, indicating less differentiation and lower capacity to regenerate full spermatogenesis in Uch-l1 −/− colonies ( Figure 6J).

Loss of Regenerative Capacity Is Accompanied with Decreased Mitochondrial Respiration in
Cultured Uch-l1 −/− SSCs SSCs undergo a metabolic switch from glycolysis to OXPHOS with differentiation [49]. Anaerobic metabolism is required for maintenance of stemness in vivo and in vitro and decreased glycolysis is associated with a lower differentiation potential of SSCs [33,49]. Transcriptome data revealing dysregulation of metabolic transitions in Uch-l1 −/− as compared to Uch-l1+/+, together with a decline in spermatogenic differentiation potential of Uch-l1 −/− after transplantation, prompted us to elucidate the impact of UCH-L1 loss on mitochondrial respiration. Two different biological samples each for Uch-l1 −/− and Uch-l1 +/+ were assessed with the Seahorse flux analyzer after 21 cell passages (150 days). Similar to the shift observed at the transcriptional level ( Figure 4D), we detected significantly higher basal respiration and maximum respiration capacity in Uch-l1 −/− (Figure 7). This underlines that the loss of UCH-L1 altered the metabolic homeostasis of SSCs and therefore stem cell maintenance. It explains the decrease in the SSC differentiation potential, as assessed by scoring of full spermatogenesis described above, and indicates a role for UCH-L1 in maintenance of SSC metabolic integrity and low aerobic metabolism. There was, however, a lack of correlation between the OCR data and OXPHOS-associated gene expression as assessed by qPCR. As shown in Table S2, no significant difference in ∆Ct values was observed between the two genotypes, except for the superoxide dismutase (Sod)1 gene (Table S2). This data is not surprising if a feedback loop leading to mRNA degradation exists when OXPHOS-related proteins are synthesized at high quantities. Collecting samples for qPCR at earlier time points may have provided better insight into the pattern of OXPHOS-associated gene expression in cultured SSCs.
blue colonies had only one or very few layers of donor-derived cells. The intensity of the blu X-gal staining was lighter in the Uch-l1 −/− colonies than the Uch-l1 +/+ colonies (represented i Figure 6B,D,J). A larger proportion of Uch-l1 −/− colonies were medium blue and a smalle proportion were dark blue as compared to the Uch-l1 +/+ colonies, indicating less differen tiation and lower capacity to regenerate full spermatogenesis in Uch-l1 −/− colonies (Figur 6J).  as assessed by qPCR. As shown in Table S2, no significant difference in ∆Ct values was observed between the two genotypes, except for the superoxide dismutase (Sod)1 gene (Table S2). This data is not surprising if a feedback loop leading to mRNA degradation exists when OXPHOS-related proteins are synthesized at high quantities. Collecting samples for qPCR at earlier time points may have provided better insight into the pattern of OXPHOS-associated gene expression in cultured SSCs.

Discussion
The role of UCH-L1 in the nervous system has become significantly better understood due to the contributions of Uch-l1 knockout mouse models; however, its function in spermatogonia has not been clearly identified. We report here that the loss of UCH-L1 impacts the differentiation competence of murine SSCs in vivo. The observed disruption of metabolic homeostasis in undifferentiated spermatogonia and loss of anaerobic metabolism of SSCs after long-term cell culture indicate a role of UCH-L1 in the maintenance of low OXPHOS and protection of stemness-associated metabolism in SSCs.

Discussion
The role of UCH-L1 in the nervous system has become significantly better understood due to the contributions of Uch-l1 knockout mouse models; however, its function in spermatogonia has not been clearly identified. We report here that the loss of UCH-L1 impacts the differentiation competence of murine SSCs in vivo. The observed disruption of metabolic homeostasis in undifferentiated spermatogonia and loss of anaerobic metabolism of SSCs after long-term cell culture indicate a role of UCH-L1 in the maintenance of low OXPHOS and protection of stemness-associated metabolism in SSCs.
UCH-L1 is highly expressed in undifferentiated spermatogonia and early differentiating spermatogonia, but not in further differentiating germ cells, and this expression pattern is phylogenetically conserved in mouse, pig and human testes [38,50,51]. However, whether UCH-L1 is required for normal spermatogenesis remains poorly understood. We used a Uch-l1 knockout mouse model to investigate morphological and functional changes associated with UCH-L1 function deficiency.
We report here a significant decrease in testis weight in Uch-l1 −/− compared to Uch-l1 +/+ mice at 120 days of age. Testis weight is an important reproductive trait in males that reflects both the spermatogenic capacity, as the sperm production rate positively correlates with the testis weight [52], and the mating potential [53]. Although Uch-l1 −/− mice remained fertile until at least 3 months of age, a reduction in confirmed matings and litters produced was observed in this study. It remains, however, plausible that the reduction in the Uch-l1 −/− mice' ability or desire to copulate may also have resulted from progressive neurodegenerative hind limb paralysis [54].
Of note, decreased testis weight was accompanied by a significant increase in the number of degenerative seminiferous tubules in Uch-l1 −/− compared to Uchl-1 +/+ at 120 day of age. Considering the progressive nature of the seminiferous tubule degeneration, this phenotype becomes more pronounced in Uch-l1 −/− mice with age. Previous findings using the gad mouse model corroborate this assumption as a slight decrease in the size of seminiferous tubules was reported in 25-week old homozygous gad mice compared to those of age-matched wild-type mice [20]. Since the final 120-day time point was dictated by the strong neurodegenerative phenotype, it was not possible to assess the Uch-l1 −/− germinal epithelium in older animals. Uch-l1-deficient Sertoli cells also may lose their ability to provide morphological and nutritional support needed for spermatogenesis, contributing to spermatogenic failure [55,56].
Maintaining a normal seminiferous epithelium relies on a regulated balance between proliferation, differentiation, and apoptosis of germ cells [1]. A potential role of UCHL1dependent apoptosis in germ cells and sperm maturation has been suggested [30]. Testes in the gad mouse model showed resistance to the early and massive wave of germ cell apoptosis during the pre-pubertal stage of development, a phase that is essential for functional spermatogenesis in adults. In contrast, there was no difference in cell apoptosis between Uch-l1 −/− and Uch-l1 +/+ mice in the current study, either in situ or in vitro from day 7 to day 120, a timeline that captures the first as well as subsequent spermatogenic waves. The discrepancy between the two studies might be explained by a potential phenotypical difference between the two mouse models. The gad model is the result of a spontaneous congenital mutation and might possibly display additional mutations within the ubiquitinproteasome pathway, while the model used in our study is a targeted mutant mouse model. Nonetheless, our data suggest that the progressive loss in the seminiferous tubule structure might be associated with a decrease in cell proliferation or differentiation rather than an increase in germ cell loss, since no difference in TUNEL-positive cells was apparent between the two genotypes at 90 or 120 days of age.
Long-term transplantation data support these conclusions. At 10 months post transplantation, the colonized area had expanded more in testes that received germ cells from Uch-l1 −/− mice as compared to Uch-l1 +/+ . Interestingly, these Uch-l1 −/− colonies contained fewer layers of donor-derived cells, indicating that Uch-l1 −/− SSCs had a reduced capacity to regenerate the entire spermatogenic lineage. Our data showing no difference in the cell number, viability, proliferation as well as in spermatogonia-associated gene expression (Id4, Plzf, c-Kit and Stra8) between Uch-l1 +/+ and Uch-l1 −/− in culture, suggest that the intrinsic loss of spermatogenic potential of Uch-l1 −/− SSCs did not affect normal SSC gene expression and viability.
It is becoming increasingly evident that maintaining metabolic integrity and plasticity is pivotal for SSC function [33,49,57]. Anaerobic metabolism is necessary for SSC maintenance in vivo and in vitro [33,49], while SSCs upregulate mitochondrial activity with differentiation [40,42,49,58]. The protection of this metabolic balance by preserving the metabolic integrity of SSCs is essential to maintaining the differentiation potential of SSCs, which gradually diminishes in aging cells [58].
We show here that Uch-l1 −/− SSCs kept in culture for 21 passages (150 days) had a higher basal respiration and maximum respiration capacity as compared to Uch-l1 +/+ , suggesting a loss of stem cell metabolic integrity in UCH-L1-deficient SSCs. These data indicate that the loss of UCH-L1 was detrimental to metabolic homeostasis regulating SSC function, therefore causing a decrease in SSC differentiation competence, all of which was reflected in the long-term outcome after transplantation of Uch-l1 −/− -derived SSCs.
Furthermore, our single-cell RNA sequencing data are consistent with a disturbed shift in cell metabolism, as an upregulation of genes associated with OXPHOS was observed in Uch-l1 −/− SSC/progenitors whereas genes associated with hypoxia were upregulated in Uch-l1 + l + SSC/progenitors. Upregulation of genes associated with glycolysis and hypoxia was detected in early differentiating spermatogonia of UCH-1 −/− , as opposed to the physiological switch towards OXPHOS in the wild-type. Hence, UCH-L1 deficiency altered the bioenergetic equilibrium between glycolysis and OXPHOS that is necessary for spermatogonial proliferation and differentiation [58]. These data suggest that UCH-L1 is required for maintenance of proper mitochondrial activity and its upregulation leads to impaired function of SSCs, impairing SSC metabolic and differentiation competence. Singlecell RNA sequencing captured these metabolic perturbations largely in one biological replicate. As these data represent a snapshot of a single time point, it is possible that these metabolic changes might not occur at similar rate or speed in mice, considering variability that might exist between individual mice of the same age group, and the compromising effect that the loss of UCH-L1 may additionally exert on the cellular oxidative defense mechanisms. In fact, UCH-L1 deficiency in the brains of knockout mice was accompanied by decreased levels of glutathione [59], which functions as an antioxidant and redox regulator [60]. The detrimental effect that UCH-L1 loss has on mitochondrial function in cultured SSC aligned with progressive metabolic perturbations that might interfere with SSC differentiation as mice age.
We detected metabolic dysregulation in murine Uchl1 −/− SSCs, characterized by an initial upregulation of OXHPOS-related genes followed by their decrease during differentiation, which is the inverse of the metabolic changes in the wild-type. Furthermore, mitochondrial activity also increased in Uchl1 −/− SSCs after long-term culture, while it remained low in wild-type SSCs. Similar metabolic changes have been described in neuroblastoma cells. A disruption of mitochondrial structure through the reduction of mitofusion 2 (MFN2) levels and an increased oxygen consumption rate were observed upon knockdown of UCH-L1 with shRNA within a few days [61].
Although further experiments are warranted to determine the mechanism by which UCH-L1 deficiency affects germ cell metabolism, one may speculate that UCH-L1 plays a role in regulating the function of mitochondria directly via the turn-over of mitochondrial proteins or indirectly via, among others, interference with the mechanistic target of rapamycin (mTOR) or hypoxia-inducible factor-1α (HIF-1α).
UCH-L1 is associated with mitochondrial quality control by regulating the expression of mitofusin-2 (MFN2), thus affecting several mitochondrial functions, especially ultrastructure in neural cells [61]. It also inhibits mTOR [28], potentially in synergy with PLZF [62], all of which might potentially help maintain the metabolic phenotype of SSCs. Studies also underlined the role of hypoxia-inducible factor (HIF) in regulating the expression of enzymes involved in glycolysis in response to hypoxia [63], but also in reducing metabolite entry into the TCA cycle, affecting OXPHOS as well [64,65]. UCH-L1 stabilizes HIF [66], and its depletion caused a significant decrease in HIF-1α protein level in a 3D culture model [67]. As UCH-L1 deficiency inhibits Retinoic acid (RA)-induced differentiation of neuroblastoma tumor cells by regulating RA-induced AKT and ERK1/2 signaling activation [68], and considering the critical role of RA-stimulated ERK1/2 pathway activity in Stra8 expression and meiotic progression in fetal germ cells [69], one may speculate that UCH-l1 deficiency can lead to a similar outcome in Uch-l1 −/− spermatogonia.
In summary, UCH-L1 plays a critical role in maintaining SSC differentiation competence in mice, as a dysregulation of normal metabolic transitions associated with SSC maintenance and differentiation occurred in Uch-l1 −/− -derived SSCs in vitro and in vivo. Additional studies are required to decipher the underlying mechanism by which UCH-L1 regulates metabolism in SSCs, yet the current study highlights that stage-dependent metabolic regulation is critical for male germ cell function and reveals a unique role of UCH-L1 in this complex system.

Data Availability Statement:
The data presented in this study are available in the article or its supplementary material.