β‐Catenin alterations in testicular Leydig cell tumour: a immunohistochemical and molecular analysis

Testicular Leydig cell tumours (LCTs) are the most common type of sex cord–stromal tumour in men, representing 1%–3% of all testicular neoplasms. Among testicular sex cord–stromal tumours, CTNNB1 mutations and nuclear expression of β‐catenin have been typically associated with Sertoli cell tumour. Recent genomic analyses have shown that CTNNB1 variants are also identified in a subset of LCTs; however, the frequency and clinicopathologic associations of β‐catenin alterations remain incompletely understood in this tumour type.


Introduction
In the testis, Leydig cell tumours (LCTs) are the most common sex cord-stromal tumour, accounting for ~3%-5% of all testicular neoplasms. 1Their incidence is bimodal, with peaks in prepubertal children and in adults 30-60 years of age, respectively. 1,2A small subset of LCTs (~10%) are clinically malignant, and these occur exclusively in postpubertal patients. 3][4][5] Among testicular sex cord-stromal tumours, pathogenic CTNNB1 variants and diffuse nuclear expression of b-catenin are characteristic of many Sertoli cell tumours, with nuclear b-catenin positivity reported in ~70% of cases overall. 6,7In a recent study, gain-of-function CTNNB1 variants or inactivating APC variants were identified in most (~90%) benign and in half of malignant Sertoli cell tumours. 8Genomic analyses of a limited number of LCTs have demonstrated that pathogenic CTNNB1 variants are also present in a subset of these neoplasms, including cases with aggressive histopathologic features and malignant clinical behaviour. 5owever, the frequency, patterns, and clinicopathologic correlates of b-catenin alterations in LCT remain incompletely described.In this study, we sought to assess the prevalence of nuclear b-catenin expression in LCTs and explore its clinicopathologic and genomic correlates.

I D E N T I F I C A T I O N O F C A S E S A N D P R O C U R E M E N T O F T I S S U E
This research was approved by the Institutional Review Board of the Indiana University School of Medicine (IRB: #18697, 2023) and the remaining participating institutions (when applicable).Institutional pathology archives (Indiana University, Indianapolis, IN, USA), personal consultation files (TMU), and research files (AMA) were queried to identify cases diagnosed as LCTs in the testis or in metastatic sites (in male patients).Available haematoxylin and eosin-stained (H&E) slides were retrieved and centrally reviewed (Y.K. and A.M.A.) to confirm the diagnosis; cases with archival, formalin-fixed, paraffin-embedded tissue were further selected for inclusion.

A S S E S S M E N T O F C L I N I C O P A T H O L O G I C F E A T U R E S A N D I M M U N O H I S T O C H E M I S T R Y ( I H C )
Electronic medical records, pathology reports, and available H&E slides were reviewed to collect relevant clinical and pathologic data.The following parameters were recorded for each case: patient age (dichotomized as <40 versus ≥40 years), type of tumour (nonmetastasizing versus metastasizing), tumour size (dichotomized as ≤5 cm versus >5 cm), mitotic activity (dichotomized as ≤3 versus >3 per 10 high-power fields), marked nuclear atypia/pleomorphism, tumour necrosis, lymphovascular invasion, and destructive/invasive growth.Metastasizing LCTs were defined as tumours with biopsy-proven metastases.
Immunohistochemistry was performed on a single representative section per case with a monoclonal anti-b-catenin antibody (Clone 14, Cell Marque, Rocklin, CA, USA; RTU).Antigen retrieval was performed using a low-pH method, and the signal was generated using the Envision Plus DAB detection system (Dako, Carpinteria, CA, USA). 7Immunohistochemistry was interpreted as positive when there was at least focal nuclear b-catenin expression; tumours showing only membranous and/or cytoplasmic staining were considered negative.In positive cases, b-catenin expression was qualitatively interpreted as focal/multifocal if there was one or more discrete (i.e.nonconfluent) foci with nuclear reactivity or diffuse if there were extensive confluent areas of nuclear reactivity comprising more than half of the tumour present in the tissue section.Positive staining was further assessed semi-quantitively as <10%, 10%-25%, >25%-50%, and >50%, based on the fraction of tumour cells exhibiting nuclear b-catenin expression.
Sequencing was performed on an Ion S5 Prime machine and the results were analysed with the Ion Reporter tools (v5.18-ThermoFisher Scientific).As described in the validation study of the assay, 10 single nucleotide variants/indels present in at least 5% of the generated reads and observed in both strands were considered for variant calling.The Varsome tool (https://varsome.com/) 9 was used to evaluate the ACMG (American College of Medical Genetics and Genomics) classification, AMP score, and ClinVar classification of each reported variant.Sequencing results were further assessed for clinical and biologic relevance by one of the authors with expertise in molecular pathology (DdB).

S T A T I S T I C S
Differences between groups were assessed using Fisher's exact test (R Project for Statistical Computing, The R Foundation, Vienna, Austria).A P-value of <0.05 was considered the threshold for statistical significance.

b-catenin in Leydig cell tumour 77
Comparison of b-catenin-positive and b-cateninnegative cases showed no significant differences in patient age, tumour size, mitotic activity, and frequency of nuclear pleomorphism, invasive growth, lymphovascular invasion, tumour necrosis, and clinical behaviour between positive and negative cases (Table 1).
DNA sequencing was performed on a subset of seven LCTs with additional formalin-fixed paraffin-embedded material available to assess whether nuclear b-catenin expression correlated with underlying pathogenic CTNNB1 variants.Two additional b-catenin-positive LCTs had been previously sequenced by our group in a prior study. 5Sequencing demonstrated that 4/7 (57%) of the LCTs analysed de novo harboured exon 3 CTNNB1 variants expected to be pathogenic.All these variants involved a region of the N-terminal domain targeted by kinases that mark b-catenin for recognition by ubiquitin ligases and subsequent proteasomal degradation.The specific variants identified included CTNNB1 p.S33Y (variant allele frequency [VAF] 26%), CTNNB1 p.T41I (VAF 30%), CTNNB1 p.D32V (VAF 9%), and CTNNB1 p.H36_S37del (VAF 33%).In the remaining three cases sequenced de novo, no pathogenic CTNNB1 variants were identified.Of note, the LCT with CTNNB1 p.D32V (VAF 9%) also harboured a RET variant (p.R844L, VAF 33%).This change results in the replacement of a charged amino acid residue (R) for an uncharged aliphatic amino acid residue (L) in the tyrosine kinase domain of the protein and is expected to have functional impact.However, this variant has not been previously annotated in COSMIC and it has been reported as a germline event 11 ; therefore, its pathogenic role in this context remains uncertain.The two LCTs sequenced as part of a prior study harboured CTNNB1 p.G34R (VAF 28%) and CTNNB1 p.H36_T42del (VAF 7%). 5 Pairwise comparisons between the VAFs of the CTNNB1 variants and the percentage of cells with nuclear b-catenin expression in the seven cases with positive sequencing results were: IHC <10%/VAF 9%, IHC 10%-25%/VAF 33%, IHC10-25%/VAF 30%, IHC10-25%/VAF 26%, IHC 25%-50%/VAF 7%, and IHC 25-50%/VAF 28%.In the three cases without CTNNB1 variants, the percentage of cells with nuclear b-catenin expression was <10%, <10%, and 10%-25%, respectively.

Discussion
Among testicular sex cord-stromal tumours, pathogenic CTNNB1 variants resulting in nuclear b-catenin expression are characteristically associated with Sertoli cell tumour. 6,7,12In this tumour type, gain-offunction CTNNB1 variants likely represent the main oncogenic driver and correlate with diffuse nuclear expression of b-catenin (Figure 3).Recently, pathogenic CTNNB1 variants have also been identified in LCTs, possibly as subclonal (rather than main driver) alterations. 5In this context, subclonal CTNNB1 mutations may represent stochastic (i.e.random) passenger events without major biologic significance or, alternatively, they may confer an evolutionary advantage and promote progression to a more aggressive neoplastic phenotype.However, the clinicopathologic and genomic correlates of nuclear b-catenin expression remain incompletely described.The present study demonstrates that a significant proportion of LCTs (~50%) express nuclear b-catenin, and this correlates with the presence of underlying pathogenic CTNNB1 variants in most of them.In this study 3/7 LCTs (43%) sequenced de novo (3/9, 33% of all cases sequenced) did not harbour pathogenic CTNNB1 variants.In these tumors, the presence of nuclear b-catenin expression may correlate with variants in regions of the gene that were not assessed by the panel.Alternatively, b-catenin may be functionally upregulated as a consequence of other events, including alterations in genes that are key regulators of the protein (such as APC ) but are not covered by the assay. 5,8,13Moreover, the possibility that nuclear b-catenin represents a "false positive" result in a subset of tumours cannot be entirely excluded, especially when positivity for this marker is limited (<10%).
As suggested in a prior study, 5 the pattern of nuclear b-catenin expression seen in the LCTs analysed in this series is consistent with subclonal activation of the protein (and Wnt signalling).Of note, the frequency of nuclear b-catenin expression did not differ between cases with and without malignant clinical behaviour and/or adverse histopathologic findings, suggesting that b-catenin alterations may represent "passenger" events.However, their high frequency (~50%) suggest that they are not random/ stochastic alterations in this tumour type.It is possible that, for unknown reasons, LCTs are predisposed to acquire CTNNB1 alterations that do not confer a subclonal evolutionary advantage; however, further studies are needed to explore this hypothesis.
In conclusion, our study demonstrates that nuclear b-catenin expression and CTNNB1 variants are highly prevalent in LCTs.The subclonal nature of these variants, which correlates with focal or multifocal expression of nuclear b-catenin demonstrated by IHC, can be used to distinguish LCTs from Sertoli cell tumour (Figure 3).The results presented herein suggest that b-catenin alterations and gain-of-function CTNNB1 variants are not associated with aggressive behaviour in LCT, although definitive conclusions in this regard are limited by the small number of malignant cases included in the series.Additional studies are needed to clarify any possible oncogenic role of b-catenin alterations in this tumour type.

Figure 1 .
Figure 1.Benign (nonmetastasizing) Leydig cell tumours with b-catenin expression.A: This tumour was sequenced as part of a prior study and harboured CTNNB1 p.G34R at a variant allele frequency of 28%.B: Immunohistochemistry demonstrated nuclear b-catenin expression in >25%-50% of the cells.C: This tumour was sequenced de novo and harboured CTNNB1 p.H36_S37del at a variant allele frequency of 33%.D: Immunohistochemistry demonstrated nuclear b-catenin expression in 10%-25% of the cells.

Figure 3 .
Figure 3. Nuclear b-catenin expression in Sertoli cell tumour.A,B: Unlike Leydig cell tumour, Sertoli cell tumour typically demonstrates diffuse nuclear b-catenin expression in >95% of tumour cells.