*2.3. Functional Characterization of DONSON In Vitro*

In order to investigate the functional role of DONSON in vitro, we used the antisense locked nucleic acid (LNA) GapmeR system to induce efficient and specific DONSON-knockdowns in established PCa cell culture models. The prostate cancer cell lines PC-3, LNCaP, C4-2B, and DU-145 were screened for their DONSON baseline expression under standard conditions (Figure 4A). As LNCap and PC-3 expressed the highest DONSON protein levels, they have been chosen for further investigations. Thus, via transfection of the specific antisense oligonucleotides, we were able to induce efficient DONSON-depletion assessed by qRT-PCR, Western blotting, and immunocytochemistry (Figure 4B,C, Figure S2).

**Figure 4.** Effect of specific DONSON-depletionin the PCa celllines LNCaP and PC-3. (**A**), ScreeningWestern Blot for DONSON in four broadly used PCa cell lines. (**B**,**C**), Induction of efficient Antisense LNA GapmeR-mediated DONSON knockdowns in LNCaP and PC-3 with subsequent validation via qPCR (**B**) and Western Blotting (**C**). (**D**,**E**) DONSON-depletion did not affect cell viability but specifically reduced the cellular motility in a Boyden Chamber Migration Assay. (**F**), Membranes depicted in 10× objective magnification. Each experiment was performed in biological triplicates. \* *p* < 0.05, \*\*\* *p* < 0.001.

After establishing efficient DONSON-depletion in both cell culture models, we aimed to investigate the dependence of important parameters of malignancy towards DONSON. In the conducted cell proliferation and cytotoxicity assay, no growth effects were evident in die DONSON-depleted PCa cells compared to the negative control (Figure 4D). Next, we explored the impact of DONSON-knockdown on the migration capacity of the investigated metastasizing PCa cells via Boyden chamber migration assays. Of note, a strong impairment of their migration capacity was seen after DONSON-knockdown (Figure 4E,F), which is thought to be an essential trait for metastatic spread and an important attribute conferring to an aggressive phenotype.

#### **3. Discussion**

To date, the role of DONSON in PCa has not been explored. In this study, we were able to identify the relatively unknown gene DONSON as a promising risk stratifier with oncogenic properties in the PCa cell culture model. DONSON was an independent predictor of a shortened PFS in the comprehensive PCa TCGA cohort and correlated with the clinical-pathological parameters (pT-stage, lymphonodal status, ISUP/Gleason score). In the group of Gleason 7 carcinomas, which plays a crucial role clinically due to the intermediate aggressiveness with regard to the prognosis and need for therapy, DONSON also shows an additive prognostic potential in the multivariate Cox analysis.

The prognostic potential of DONSON has been validated at the protein level in a large PCa TMA cohort, highlighting its potential as a robust biomarker. Of note, the DONSON protein was localized in the cytoplasm of the PCa samples, which was in accordance with the staining pattern observed in The Human Protein Atlas and as described previously for clear cell renal cell carcinoma tissue [7,8]. Staining specificity was confirmed via immunocytochemistry in PC-3 cells with and without DONSON knockdown. Nevertheless, due to its function in DNA replication and repair, an additional nuclear expression would have been expected. During the S phase, nuclear DONSON foci were observed [9]. However, the DNA replication and S phase only describes a small part of the cell cycle, and thus the localization of DONSON could differ during the G1 phase [24]. Furthermore, as the overall knowledge regarding DONSON is sparse, it may have additional functions, also inside the cytoplasm. As this was not the scope of our study, further investigations regarding its subcellular localization, trafficking, and exact biological function are needed to clarify this.

Interestingly, the PCa TMA cohort showed a heterogeneous picture, with some tumors being DONSON-negative while others, especially Gleason 8 and higher carcinomas, strongly overexpressed DONSON. It has to be mentioned that only a strong statistical trend was seen for DONSON to be an independent predictor of unfavorable PFS in this cohort (HR 1.46, 95% CI; 0.86–2.48; *p* = 0.17), which may be due to a relatively low sample size compared to the PCa TCGA cohort (PFS Follow-up PCa TMA cohort *n* = 103 (29 events); PCa TCGA cohort *n* = 497 (93 events)).

In addition, two independent PCa progression cohorts showed a significant increase in DONSON expression in the metastatic samples compared to pPCA, which was particularly evident in distant metastases and androgen-deprivation resistant metastases. The crucial step in PCa progression is displayed by the development of metastases and a castration-resistant status during androgen-deprivation therapy (ADT). Among the different mechanisms of CRPC development, aberrant androgen receptor (AR) signaling is thought to be a major player [22,25]. An association between DONSON and AR expression was observed in the PCa progression and the PCa tissue microarray (TMA) cohorts on both transcriptional and translational levels. However, the exact interaction of DONSON and the AR signaling pathway and a possible link between DONSON and the development of castration-resistance requires further functional investigations. In addition, the proliferative activity measured by Ki67 expression, which is also an established prognostic biomarker in PCa and other cancers [18,19], was significantly correlated with DONSON expression, which seems comprehensible due to the predicted function of DONSON as part of the replisome [10,26]. Thus, renal cell carcinoma cell lines showed decreased proliferative capacity after oligonucleotide-mediated DONSON knockdown [7,8]. However, in our PCa cell culture model, no influence on proliferation could be detected after DONSON-depletion, which suggests an additional unknown function of DONSON, but this requires further investigation. In our cell culture model, DONSON-depletion led to potent inhibition of cell motility, which is recognized as a surrogate for the metastatic capacity in vitro. This provides evidence that DONSON plays a role during the metastatic process, which could ultimately explain its significant upregulation in the metastatic samples in both PCa progression cohorts and the N+ pPCa samples (PCa TCGA).

Taxane-based therapy is a backbone of PCa therapy and preferentially attacks tumor cells with an increased cell division rate as well as limited DNA damage repair capacity. As DONSON plays a pivotal role in both cellular processes, replication, and maintaining genome stability, it could be an interesting therapeutic target for combination therapies [10,11]. Therefore, we think that our study on DONSON in PCa, as well as the fact that DONSON overexpression seems to mediate tumor-independent oncogenic properties, could be a starting point for further basic and oncological research on DONSON.

Thus, the results of our comprehensive expression analyses, as well as the functional data obtained after DONSON-depletion, lead us to the conclusion that DONSON is a promising prognostic biomarker with oncogenic properties in PCa.
