PDCD10/CCM3, a potential target for pancreatic ductal adenocarcinoma?

So far, experiments utilized exogenously applied recombinant human TGF-β1. It was, therefore, of interest to determine if PDCD10 exerts the same effects to TGF-β1 produced and secreted in a paracrine manner by other neighbouring cell types in the PDAC microenvironment, i.e., cancer-associated fibroblasts (CAFs). Of note, indirect co-culture of CAF-PSC47 cells with each of the two PDAC cell lines induced a proliferative response in the parental PDAC cells (which was more pronounced in PaTu8902), but this effect was reduced in PDCD10^-/- cells. To confirm that the CAF-induced proliferation was indeed caused by CAF-secreted TGF-β1 rather than any other secreted factor, the authors employed a soluble human TGF-β type II receptor-Fc conjugate (TGFBRII-Fc). This ligand trap binds TGF-β1 and 3 and prevents them from binding to the receptors, thereby inhibiting their bioactivity. Indeed, these experiments showed that CAF-induced TGF-β accounted for the proliferative activity in both PaTu8902 and PaTu8988t-wt cells. However, it should be noted that pancreatic tumour cells themselves are capable of expressing and secreting TGF-β(s) in an autocrine manner [9,10]. It would, thus, have been interesting to also treat PaTu monocultures with the TGFBRII-Fc ligand trap to test if autocrine TGF-β1 signalling by these cells affects TGF-β1-mediated proliferation [9] in the same way as paracrine/CAF-derived TGF-β1. Since TGF-β2 cannot bind TGFBRII-Fc and is, thus, refractory to inhibition by this ligand trap, and PSC47 cells were essentially devoid of TGF-β3 mRNA, authors concluded that CAF-derived TGF-β1 is responsible for the pro-proliferative effect on the PDAC cell lines (Figure 1). The authors then subjected the PDCD10-deficient cells to the same co-culture procedure in the presence of TGFBRII-Fc. Surprisingly, only PDCD10-deficient PaTu8902, but not PaTu8988t cells, exhibited higher proliferative activity. Since both cell lines differ merely in the expression of functional SMAD4 with PaTu8988t being DPC4-deficient, the results suggested that both PDCD10 and SMAD4 are required for promoting proliferation in response to CAF-derived TGF-β1 (Figure 1). Authors also concluded that in the absence of PDCD10, but in the presence of functional SMAD4, TGF-β1’s function switches back to cell cycle/growth arrest. However, in the presence of the soluble TGFBRII-Fc, which suspends all TGF-β1 bioactivity, growth arrest by TGF-β1 in the DPC4-wt PaTu8902 cells was relieved and the cells’ mitotic activity restored. Therefore, the authors went on to study how PDCD10 blocks the growth-inhibitory effect of TGF-β1 (and turns it into a growth-stimulatory one) and how SMAD4 is involved here.

The Retinoblastoma 1 protein (pRB) is a well-known inhibitor of cell cycle progression and is induced by TGF-β1 to promote growth inhibition (Figure 1). Authors revealed in the PDCD10-KO cell lines that it was hypophosphorylated ( = activated) in DPC4^+/+ PaTu8902 but not DPC4^-/- PaTu8988t cells. They interpreted these results to indicate that PDCD10 promotes pRB hyperphosphorylation and thus its inactivation (Figure 1). This is thought to deprive TGF-β1 of its cell cycle arrestor, leading to loss of TGF-β1/SMAD-induced growth inhibition.

The MEK-ERK signalling pathway is involved in promoting cell proliferation and EMT in PDAC [9]. Moreover, in prostate cancer, PDCD10 has been shown to induce cell proliferation through modulation of the ERK pathway [5]. Hence, authors explored whether PDCD10 utilizes activated ERK to mediate the mitotic effect of TGF-β1. To this end, loss of PDCD10 decreased ERK1/2 activation in both PaTu8902 and PaTu8988t cells. In addition, a positive correlation between the mRNA levels of PDCD10 and ERK1/2 was revealed; however, since ERK1 and ERK2 are expressed constitutively and their activities controlled almost exclusively by differential phosphorylation, these data might not be very conclusive. Authors then hypothesized that PDCD10 promotes TGF-β1-dependent proliferation through the activation of the MEK-ERK pathway, independently of SMAD4. Using ERK inhibition in combination with TGF-β1 stimulation experiments, they were able to show that the activation of ERK by TGF-β1 is SMAD4-independent (Figure 1) and that TGF-β1 must activate ERK in order to induce proliferation.

Interestingly, when the same setup (ERK inhibitor + TGF-β1) was applied to the two (PDCD10-wt) PDAC cell lines for the analysis of cell motility, migration was enhanced in PaTu8902, but declined in PaTu8988t. These data imply that both PDCD10 and ERK are needed to drive migration but that an SMAD4-dependent mechanism exists that ensures TGF-β1’s ability to enhance migration even under the conditions of inactivated ERK.

To investigate these effects in a more translational, 3D setting, authors went on to generate organoids from tissue specimens of six PDAC patients. Once established, they sorted the organoids into two subsets according to basal expression levels of PDCD10, namely, PDCD10^low and PDCD10^high. After a 72-hour treatment with SCH772984, cell counts in the PDCD10^low group were strongly attenuated, while no effect was detected in the PDCD10^high group. This matches well with the authors' initial finding that PDCD10-KO cells exhibited reduced proliferative activity compared with their wt counterparts. These findings suggest that targeting PDCD10 in PDAC patients may be a useful adjunct to the clinical use of ERK inhibitors.

Prompted by the observation that PDCD10 interacts with the serine/threonine kinase MST4 to modulate the MEK-ERK pathway, authors hypothesized that PDCD10 forms a complex with MST4 to translocate it to a subcellular site where it can efficiently promote pRB and ERK phosphorylation. The authors detected some basal co-localization of PDCD10 and MST4 (using proximity ligation assays, PLAs) in the parental cell lines, which was enhanced by TGF-β1 (Figure 1, green arrow), but whether this TGF-β effect was SMAD4-dependent was not analysed. However, TGF-β1-mediated co-localization and, hence, potential physical interaction between MST4 and p-pRB in the PDCD10-wt lines were SMAD4-dependent. When repeating these assays with PDCD10-deficient cells, the TGF-β1-induced co-localization of MST4 and pRB was blunted, confirming the positive functional association between TGF-β1 signalling and PDCD10. In contrast, TGF-β1-mediated co-localization of MST4 and p-ERK in the parental cell lines did not exhibit any SMAD4 dependency. However, in cells with PDCD10 knocked out, the abundance of PLA signals for both MST4-p-pRB and MST4-p-ERK failed to increase upon TGF-β1 stimulation. Based on these findings, authors concluded that TGF-β1 enhances the complex formation of PDCD10 and MST4, thereby enabling PDCD10 to phosphorylate pRB to become inactive and ERK1/2 to become active (Figure 1).

To further validate their findings and to correlate PDCD10 expression with tumour progression in an in silico approach, authors conducted a series filtration process in the current single cell sequencing database of PDAC patients. Notably, according to the molecular signature, the ductal cell types fell into two classes, termed ‘ductal cell type 1’ and ‘ductal cell type 2’. Type 2 cells were more malignantly progressed than type 1 cells and, hence, can be considered the ductal tumour cells, while the ductal cell type 1 resembles normal (benign) ductal cells. Interestingly, type 2 cells presented with higher levels of PDCD10 than type 1 cells. The type 2 cells were then subdivided into two groups, PDCD10^high and PDCD10^low, while type 1 cells were classified into SMAD4^normal and SMAD4^low based on the median expression levels of PDCD10 and SMAD4, respectively. Authors found that normal SMAD4 expression segregated with high expression of the pRB target gene CDK4 (encoding cyclin-dependent kinase 4) but not the ERK target gene KRT19 (encoding cytokeratin 19). These findings are supportive of previous findings, suggesting that PDCD10 is unable to counteract the pRB pathway in tumour cells with low or absent SMAD4 expression, whereas it does so in tumour cells with physiological SMAD4 levels. In the two PaTu lines, activities of the MEK-ERK pathway remain unaffected by SMAD4 expression; however, this feature may be cell type-specific, since in pancreatic cancer cell lines other than PaTu8902/PaTu8988t, i.e., PANC-1, ERK1 and ERK2 activation has been shown to be negatively regulated by SMAD4, for example, in response to activation by the tumour suppressor protein TAp73 [11].

Mutations in DPC4 are among the most characteristic genetic alterations in PDACs. As one of the driver genes, DPC4 is mutated in half of PDAC cases, with homozygous deletions occurring in 30% and chromosome allelic loss in 20% of cases. While the paper discusses the relationship between PDCD10 and SMAD4 expression (or activity), it does not study the possibility of a more direct association between both proteins/genes at either the genetic level/status, e.g., by TCGA analysis, or by functional studies in cell culture.

A study involving 471 patients with resected PDAC evaluated the relationship of SMAD4 expression with overall survival and showed that loss of Smad4 expression was predictive for postoperative adjuvant chemotherapy benefit [12]. Likewise, DPC4 inactivation in established PDAC cell lines sensitized them by 2–4-fold to cisplatin and irinotecan, but slightly decreased their sensitivity to gemcitabine [13]. Unfortunately, no data are available for PDCD10 as a therapeutic target in PDAC. However, in other cancer types, a dual role in drug resistance has been observed. For instance, in therapy of glioblastoma, PDCD10 acted as a chemosensitizer, since the loss of PDCD10 caused resistance to temozolomide during treatment and promoted a rapid re-growth of tumour cells after treatment [14]. Moreover, PDCD10 expression is both cell and anti-cancer drug-specific. Down-regulation of PDCD10 expression in the breast cancer cell line, MCF7, increased the resistance to doxorubicin and docetaxel, while it decreased resistance in doxorubicin-resistant HeLa cervical carcinoma cells. Conversely, overexpression of PDCD10 in parental HeLa cells increased resistance to doxorubicin, while it re-sensitized doxorubicin-resistant MCF7 cells [15]. These apparently conflicting functions of PDCD10 might be due to its interactions with different signalling pathways specifically in certain cancer contexts.

Despite these findings, authors speculate that targeting PDCD10 in PDAC patients could represent a new and promising adjunct to TGF-β-targeted therapies. Here, PDCD10 may reverse drug resistance, an assumption that is given support by the author’s observation that PDCD10 was able to facilitate EMT of PDAC cells [7]. The possibility of targeting PDCD10 is intriguing in different contexts, although further studies are needed to identify the best method to be used. Direct pharmacological targeting of PDCD10 in patients is hampered by the unavailabilty of specific (small molecule) inhibitor(s) and likely by serious side effects that are expected due to the pleiotropic cellular effects of this protein [5]. Since PDCD10 expression is subject to regulation by several microRNAs [4], novel microRNA-based therapeutics in combination with newly designed gene therapy strategies could be powerful tools to interfere with PDCD10-dependent growth, EMT and chemoresistance.

Rather than PDCD10 itself, its crucial interaction partners, for example, MST4, may be drug targeted. It remains to be seen, however, if MST4 has a role as a target for metastatic PDAC as was observed for the most aggressive forms of prostate cancer. PDCD10 also promotes EMT and tumour progression of hepatocellular carcinoma (HCC) by interacting with PP2Ac to promote YAP activation. Hence, PDCD10 was suggested as a therapeutic target for HCC and the PP2Ac inhibitor, LB100, proposed to restrict tumour growth and metastasis of HCC with high PDCD10 expression [16].

Collectively, the study by Zhou et al. [2] established a novel oncogenic role of PDCD10 in TGF-β1-induced progression of PDAC, involving SMAD4-dependent inactivation of pRB and SMAD4-independent activation of ERK1/2 (Figure 1). Nevertheless, the role of PDCD10 in tumour progression needs further investigation to provide deeper and more comprehensive knowledge of its functions in cancer. This knowledge would also help to better define PDCD10 as a ‘druggable’ therapeutic target, either directly or indirectly.

The authors declare that they have no conflict of interest.

Hendrik Ungefroren: Writing—original draft, Writing—review and editing.

CAF

cancer-associated fibroblast

CCM

cerebral cavernous malformation

EMT

epithelial-mesenchymal transition

ERK

extracellular signal-regulated kinase

EndMT

endothelial-mesenchymal transition

PDAC

pancreatic ductal adenocarcinoma, PDCD10, programmed cell death 10

PLA

proximity ligation assay

TGFBRII-Fc

TGF-β type II receptor-Fc conjugate

TGF-β

transforming growth factor-beta