Coordinated downregulation of Spinophilin and the catalytic subunits of PP1, PPP1CA/B/C, contributes to a worse prognosis in lung cancer

Oncotarget105196 www.impactjournals.com/oncotarget Coordinated downregulation of Spinophilin and the catalytic subunits of PP1, PPP1CA/B/C, contributes to a worse prognosis in lung cancer Eva M. Verdugo-Sivianes1,2, Lola Navas1,2, Sonia Molina-Pinelo1,2, Irene Ferrer2,3, Alvaro Quintanal-Villalonga3, Javier Peinado1,4, Jose M. Garcia-Heredia1,2,5, Blanca Felipe-Abrio1,2, Sandra Muñoz-Galvan1,2, Juan J. Marin1,2,6, Luis Montuenga2,7, Luis Paz-Ares2,3 and Amancio Carnero1,2 1Instituto de Biomedicina de Sevilla (IBIS), Hospital Universitario Virgen del Rocío, Universidad de Sevilla, Consejo Superior de Investigaciones Científicas, Sevilla, Spain 2CIBER de Cáncer, Instituto de Salud Carlos III, Pabellón 11, Planta 0, Madrid, Spain 3H120-CNIO Lung Cancer Clinical Research Unit, Instituto de Investigación Hospital 12 de Octubre and CNIO, Madrid, Spain 4Radiation Oncology Department, Hospital Universitario Virgen del Rocío, Sevilla, Spain 5Department of Vegetal Biochemistry and Molecular Biology, University of Seville, Seville, Spain 6Department of Predictive Medicine and Public Health, Universidad de Sevilla, Sevilla, Spain 7Program in Solid Tumors and Biomarkers, Center for Applied Medical Research (CIMA), Pamplona, Spain Correspondence to: Amancio Carnero, email: [email protected] Keywords: Spinophilin; PP1; biomarker; lung cancer; therapy Received: May 13, 2017 Accepted: September 03, 2017 Published: October 26, 2017 Copyright: Verdugo-Sivianes et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License 3.0 (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. ABSTRACT The scaffold protein Spinophilin (Spinophilin, PPP1R9B) is one of the regulatory subunits of phosphatase-1 (PP1), directing it to distinct subcellular locations and targets. The loss of Spinophilin reduces PP1 targeting to pRb, thereby maintaining higher levels of phosphorylated pRb. Spinophilin is absent or reduced in approximately 40% of human lung tumors, correlating with the malignant grade. However, little is known about the relevance of the coordinated activity or presence of Spinophilin and its reported catalytic partners in the prognosis of lung cancer. In the present work, we show that the downregulation of Spinophilin, either by protein or mRNA, is related to a worse prognosis in lung tumors. This effect is more relevant in squamous cell carcinoma, SCC, than in adenocarcinoma. Downregulation of Spinophilin is related to a decrease in the levels of its partners PPP1CA/B/C, the catalytic subunits of PP1. A decrease in these subunits is also related to prognosis in SCC and, in combination with a decrease in Spinophilin, are markers of a poor prognosis in these tumors. The analysis of the genes that correlate to Spinophilin in lung tumors showed clear enrichment in ATP biosynthesis and protein degradation GO pathways. The analysis of the response to several common and pathway-related drugs indicates a direct correlation between the Spinophilin/PPP1Cs ratio and the response to oxaliplatin and bortezomib. This finding indicates that this ratio may be a good predictive biomarker for the activity of the drugs in these tumors with a poor prognosis. www.impactjournals.com/oncotarget/ Oncotarget, 2017, Vol. 8, (No. 62), pp: 105196-105210 Research Paper Oncotarget105197 www.impactjournals.com/oncotarget INTRODUCTION The Spinophilin (Spn, PPP1R9b) gene is located at 17q21.33, a cytogenetic area frequently associated with microsatellite instability and loss of heterozygosity (LOH). LOH in chromosome 17q21.3 has been observed in different tumors, including breast, ovarian, prostate, colorectal, gastric, renal and lung carcinomas as well as in salivary gland carcinosarcoma, an extremely aggressive neoplasm [1–6]. Low levels of Spinophilin expression have been found in lung adenocarcinoma [7], head and neck cancer [8], hepatocellular carcinoma [9], human gastric, small intestine and colorectal adenocarcinoma [10, 11], glioblastoma [12] and breast cancer [13, 14]. In all cases, downregulation of Spinophilin correlated with a higher malignant grade, more aggressive biological behavior and resistance to therapies, leading to faster relapse and poorer patient survival. Furthermore, the loss of Spinophilin also correlated with p53 mutations. Analysis of human breast tumors showed that Spinophilin downregulation increased the stemness properties and the expression of stem-related genes (Sox2, KLF4, Nanog and OCT4) [14]. Breast tumor stem cells appeared to have low levels of Spinophilin mRNA, and Spinophilin loss correlated with increased stem-like cell appearance in breast tumors, as indicated by an increase in CD44+/CD24cells. A reduction of the levels of PPP1CA mimicked the cancer stem-like cell phenotype of Spinophilin downregulation, suggesting that the mechanism of Spinophilin involves PP1a [14]. Spinophilin is a scaffolding protein interacting with more than 30 partner proteins, including protein phosphatase 1 (PP1) and F-actin [15–17]. However, the physiological relevance of some of these interactions remains to be determined. Spinophilin performs important functions in the nervous system where it is implicated in regulating spine morphology and density, synaptic plasticity and neuronal migration [15]. Spinophilin also regulates seventransmembrane receptors and may link these receptors to intracellular mitogenic signaling events that are dependent on p70S6 kinase and the small G protein-GEF Rac. Spinophilin also interacts with doublecortin, an actinbinding protein with an established role in the subcellular targeting of PP1. Spinophilin enhances PP1-mediated dephosphorylation of doublecortin [18] concomitant with PP1 localization in the cytosol [19, 20]. Thus, localization of the doublecortin–Spinophilin–PP1 complex in the cytosol inhibits PP1 phosphatase activity, leading to glioma cell death effects via a mitotic spindle catastrophe. At the functional level, Spinophilin regulates PP1 activity, thereby maintaining higher levels of phosphorylated pRb [21]. This effect contributes to an increase in p53 activity through the increase in ARF protein. However, in the absence of p53, reduced levels of Spinophilin increase the tumorigenic properties of cells. Spinophilin knockout mice have a reduced lifespan, an increased number of tumors and increased cellular proliferation in some tissues, such as the mammary ducts. In addition, the combined loss of Spinophilin and p53 activity in mouse models leads to an increase in mammary carcinomas, confirming the functional relationship between p53 and Spinophilin [22]. The data suggest that the regulation of PP1 activity is the mechanism by which Spinophilin acts to produce its tumor suppressor activity [23]. However, little is known about the relevance of the catalytic subunits of PP1 in cancer prognosis and specifically its coordinated regulation with Spinophilin. PP1 serine/threonine phosphatases are multimeric enzymes assembled from a small number of catalytic subunits with one of hundreds of regulatory subunits. The regulatory subunit provides precision and specificity to the target [24]. These phosphatase regulators do not share extensive sequence conservation. Instead, they are identified by their physical interaction and function [24]. There are 4 different catalytic PP1 isoforms, PPP1CA/B/C, derived from 3 different genes, plus an alternate splicing of PPP1CC [25–27]. Due to the broad spectrum of activity of each catalytic subunit, not much information regarding their role in tumors or their pathological value has been published. In glioblastoma, PP1A protein expression showed no correlation with prognosis in all cases or on stratification based on IDH1 or ATRX expression. However, nuclear PP1A expression is a strong independent predictor of poor overall survival in p53-positive GBMs only [28]. PPP1CA urinary content was also associated with recurrence in bladder tumors [29]. Finally, analysis of human tumors suggests that one of the PPP1CA alleles might be lost in a high percentage of kidney and colorectal carcinomas [30]. Because the specificity and precision of each PPP1C isoform is given by the regulatory subunit and we detected downregulated Spinophilin in a subgroup of lung tumors, we investigated the effect of Spinophilin regulation of the PP1 catalytic subunits in human lung tumors. RESULTS Loss of Spinophilin in lung tumors Spinophilin downregulation in human lung tumors is a causal event that triggers an increase in tumorigenic properties, contributing to their malignant status [7]. We have previously described [7, 10] that the cut off for downregulated or lower levels of SPINOPHILIN protein is equal or below 50% of the protein levels found in non-tumoral cells of the same tissue. Downregulation of SPINOPHILIN can be observed by immunostaining (Figure 1A) comprising around 30% samples. On the other hand, we quantified the levels of Spinophilin by mRNA expression levels and detected approximately 30% of Oncotarget105198 www.impactjournals.com/oncotarget tumors with levels of Spinophilin lower than average (= 0.0113) (Figure 1B). This percentage broadly correlated with the percentage of samples showing downregulation of Spinophilin by immunohistochemistry [7]. A more in-depth analysis using public transcriptomic databases shows a similar percentage of tumor samples with downregulated Spinophilin (Supplementary Figure 1). Furthermore, this downregulation of Spinophilin is more pronounced in squamous tumors than in adenocarcinoma (Figure 1C-1E). A similar percentage of cases showed reduced Spinophilin levels independent of whether the analysis was performed on mRNA or protein, suggesting that Spinophilin downregulation occurs either by the loss of the 17p21 locus or in most cases through the regulation of mRNA levels. To further explore this result, we analyzed Spinophilin promoter methylation in matched lung tumor samples vs. non-tumoral samples from the same patient [31] (Supplementary Table 2). We found that this gene showed increased methylation in tumoral samples vs. non-tumor samples (Table 1). The increased methylation, however, occurs regardless of whether the samples are from adenocarcinoma or squamous carcinoma. Figure 1: Loss of Spinophilin in human tumors. (A and B) Representative photo of different lung tumors with different Spinophilin levels. (C) The levels of Spinophilin mRNA in a cohort of 72 human lung tumors described in reference [7] were analyzed according to the procedure described in M&M. The mean value of mRNA levels of Spinophilin in all samples calculated and tumors distributed according to this mean value. High Spinophilin > mean value; Low Spinophilin > mean value. (D) Analysis of Spinophilin mRNA levels in samples from the cohort of [64]. (E) Analysis of Spinophilin mRNA levels in samples from the cohort of [65]. (E) Analysis of Spinophilin mRNA levels in samples from the cohort of [66]. In all three cases (D, E and F) the differences of the Spinophilin mRNA levels between ADC and SCC were statistically significant (p<0.05) ADC: Lung Adenocarcinoma; LCLC: Large Cell Lung Carcinoma; SCC: Squamous Cell Lung Carcinoma. Oncotarget105199 www.impactjournals.com/oncotarget Decreased Spinophilin levels predicted poor outcome in lung cancer patients To evaluate whether SPINOPHILIN levels were associated with clinical outcome, we correlated SPINOPHILIN immunohistochemical staining with patient disease-free interval (DFI) and overall survival (OS). Decreased SPINOPHILIN levels were associated with a poorer OS (p=0,022) and DFI (P=0.020) in patients with lung cancer (Figure 2A and 2B). Multivariate analyses confirmed ECOG and stage as independent predictive factors for PFS. Regarding overall survival, only stage Table 1: Methylation of Spinophilin promoter in lung tumors Spinophilin methylation NON-tumor Tumor Student’s T-test P value Adenocarcinoma 0.44 (n=13) 0.52 (n=13) <0.0001 SCC 0.44 (n=10) 0.51 (n=10) 0.009 The data show the average mean of methylation (see Methods) in the analyzed samples (n=num of samples analyzed). Figure 2: Survival probability of patients with lung cancer according to Spinophilin levels. (A) Overall survival (OS) and (B) Disease-Free Interval (DFI). (C) Overall survival (OS) and (D) Disease-Free Interval (DFI) in patients with low Spinophilin according to p53 levels. Oncotarget105200 www.impactjournals.com/oncotarget retained independent prognostic significance in the Cox multiple regression model (Supplementary Table 3). The analysis of concurrent molecular alterations showed a correlation with nuclear p53 staining [7]. Therefore, we measured the predictive value of SPINOPHILIN in human lung tumors with nuclear accumulation of p53 (Figure 2C and 2D). We found that low SPINOPHILIN correlated with high levels of nuclear p53 (>10%), as an indicator of mutant p53, and this pattern (low SPINOPHILIN and mutant p53) correlated with a worse prognosis in both OS and DFI (Figure 2C and 2D). To further assess this hypothesis, we evaluated the association of Spinophilin mRNA levels with tumor response to patient survival in independent cohorts of publicly available databases (Supplementary Figure 1). Low levels of Spinophilin mRNA in lung tumor tissue samples of different databases were always predictive of worse survival probability, correlating with the findings of immunohistochemistry on the SPINOPHILIN protein. Consequently, low levels of Spinophilin mRNA were also associated with a shorter disease-free interval (DFI) and OS in these series of patients with lung tumors, highlighting the relevance of Spinophilin as a predictive factor (Figure 2 and Supplementary Figure 1). Decreased levels of PPP1Cs are related with high risk and predicted poor outcome in lung SCC cancer patients SPINOPHILIN is a regulatory subunit of PP1 and binds one of the catalytic subunits, PPP1CA, B or C, forming a heterodimer with PP1 phosphatase activity. Therefore, we measured whether the levels of these catalytic subunits are related to survival probability in tumors of the lung. To this end, we selected the TCGA databases for adenocarcinoma or SCC specific tumors. We observed that for adenocarcinoma, the patients with low levels of PPP1CA or B have a significantly higher risk of decreased survival than patients with high levels of expression of these genes (Figure 3A). The opposite effect is observed for PPP1CC. However, the survival probability of the patients does not change significantly when all three PPPC catalytic subunits are considered (Figure 3B). When we studied SCC tumors, we observed a more homogeneous behavior and the patients with low levels of PPP1CA, B or C have a significantly higher risk than patients with high levels of these genes (Figure 3C). Furthermore, patients with low levels of these genes have a significantly poorer survival probability (Figure 3D). Figure 3: Higher risk and lower survival probability of patients with lung cancer according to lower mRNA levels of the catalytic subunits of PP1. (A and C) Risk of worse survival probability according to mRNA levels in patients with adenocarcinoma (left, A) or squamous cell carcinoma (right, C). (B and D) Survival probability (log rank) of patients with lung cancer according to the mRNA levels of the joint catalytic subunits of PP1. Values were taken above or below the average for each subunit evaluated. High or low risks were taken according to the values of figure (A) and (C). (B) Lung adenocarcinoma; (D) Squamous cell carcinoma. The TCGA cohort was used. Oncotarget105201 www.impactjournals.com/oncotarget As a regulatory subunit, we studied the correlation of Spinophilin mRNA expression with those of the different catalytic subunits. We analyzed the levels of PPP1Cs mRNA in samples from Supplementary Table 1. We detected a direct relationship between the levels of Spinophilin and those of each catalytic subunit (Figure 4A, Figure 4: Survival probability of patients with lung cancer according to the joint mRNA levels of the individual catalytic subunits of PP1 and Spinophilin. (A) Correlation between the mRNA levels of Spinophilin and the catalytic subunits of PP1 (Student’s T-test; ***=p<0.001). We analyzed the levels in a cohort of 70 samples from our cohort from Supplementary Table 1, for which we had mRNA (see Materials and Methods). We quantitated by Q-RT-PCR the levels of PPP1CA, PPP1CB and PPP1CC and plotted according to the low or high levels of Spinophilin mRNA (graph). The graph shows the distribution of the PP1 catalytic subunit according to the categorization of samples in high or low Spinophilin according to its average. Furthermore we plotted one to one correlation of PPP1CA/B/C mRNA levels to Sphinophilin mRNA levels in each sample. The index of correlation (r) and statistical significance (p) for the Pearson correlation is included in the inset in each graph. (B) Survival probability (log rank) of patients with lung cancer according to the mRNA levels of the joint individual catalytic subunits of PP1 and Spinophilin. Values were taken above or below the average for each subunit evaluated. High or low risks were taken according to the values of Figure 3A. The TCGA cohort was used. Oncotarget105202 www.impactjournals.com/oncotarget box graph was analyzed by Student’s T-test; inset also shows Pearson r). Therefore, we next measured whether the combination of regulatory Spinophilin and either of the catalytic subunits might have some predictive capability. We found that in adenocarcinoma tumors, the combination had no clear effect on survival prognosis (Figure 4B). However, the prognosis capability of low Spinophilin combined with low levels of PPP1CA/B or C is clearer in patients with SCC tumors of the lung (Figure 4B). Finally, we combined low Spinophilin and low PPP1CA/B and C levels and analyzed the predictive capability of survival in patients with adenocarcinoma or SCC tumors. We observed a clear and significant poor prognosis in patients with tumors with low levels of combined Spinophilin/PPP1CA, B and C only in patients with SCC tumors (Figure 5). While the analysis of the methylation of Spinophilin showed increased methylation, the analysis of PPPCs subunits methylation showed a decreased methylation mean, in tumors vs. non-tumor samples, in PPP1CA and PPP1CB, and increased methylation mean in PPP1CC (See Supplementary Table 4). Therefore, the mechanism of regulation must be different for all three isoforms. It may include transcriptional regulation or cell adaptation throughout the growth of the tumor. Analysis of the GO terms correlating to Spinophilin Patients with lung tumors with low Spinophilin levels showed clear and significant poor prognosis. Therefore, new therapeutic alternatives for these patients are needed. To explore this point, we looked for genes that correlated positively and negatively to Spinophilin (PPP1R9B) in the TCGA database. We selected genes with a correlation r>0.350 or r<-0.350, to identify genes that correlate positively or negatively to Spinophilin levels in tumors (Supplementary Table 5). Next, we identified the GO terms related to these genes using the Enrichr web portal (Supplementary Table 6A and 6B). The GO terms that correlated positively to Spinophilin were enriched in several biological processes, such as chromatin modification, ATP biosynthetic processes, embryo development and the regulation of GTPase activity (Figure 6). Alternatively, GO terms that correlated negatively to Spinophilin, and therefore may be enriched in tumors with low Spinophilin, were the regulation of protein degradation, ATP biosynthetic processes and the regulation of the cell cycle (Figure 6). These GO term enrichment assays suggest that several pathways interfere with the aim of finding alternative therapies. Processes such as the regulation of protein degradation, ATP biosynthetic processes and the regulation of the cell cycle being negatively correlated seemed more suitable because they are altered in the absence of Spinophilin. Therefore, we tested the effect of metformin, a regulator of the ATP biosynthetic process, and bortezomib, an inhibitor of protein degradation, in a panel of lung cancer cell lines and their correlation to Spinophilin. Analysis of the correlation between Spinophilin and PPP1Cs in a panel of lung cancer cell lines and their relationship with the drug response First, we analyzed the levels of expression of Spinophilin in a panel of 17 lung cancer cell lines (Supplementary Table 7), and we observed a different pattern of expression (Figure 7A). Therefore, we classified the panel as high Spinophilin cell lines (H1437, H1781, Figure 5: Survival probability of patients with lung cancer according to the joint mRNA levels of the catalytic subunits of PP1 and Spinophilin. Survival probability (log rank Cox) of patients with lung cancer according to the mRNA levels of the joint catalytic subunits of PP1 and Spinophilin. Values were taken above or below the average for each subunit evaluated. High or low risks were assessed according to the values of Figure 3A. The TCGA cohort was used. Oncotarget105203 www.impactjournals.com/oncotarget H2009, H358, Calu3 and Nuli1) and low Spinophilin cell lines (H1650, H1975, H2228, H226, H3122, H460, H520, HCC827, Calu1, A549 and NL20). Then, we studied the correlation of Spinophilin expression with those of the different catalytic subunits in the panel of 17 different lung cancer cell lines (Supplementary Table 8). As in the case of tumors, we detected a direct relationship between the levels of Spinophilin and the levels of each catalytic subunit (Figure 7B, inset shows also Pearson r; Supplementary Table 8). As in tumors, cell lines with low Spinophilin also contained low levels of each catalytic subunit. With the aim of finding a drug that may be active in lung cell lines with low Spinophilin levels, we subjected this panel of 17 cell lines to different treatments to obtain the IC50 for the response in each cell line (Supplementary Table 9). We specifically tested cisplatin and etoposide as common treatments for lung tumors as well as oxaliplatin as a platinum-derived compound with a different spectrum of activity. Finally, we also tested metformin as a modulator of ATP biosynthesis and bortezomib as an inhibitor of proteasomal degradation [32, 33] because these two mechanisms seemed to be highly related to genes that correlated to Spinophilin levels. Then, we correlated the IC50 for each drug with the levels of Spinophilin, PPP1Cs or the ratio among them and calculated the correlation index (Pearson r). We found that none of the activities of these drugs correlated with the levels of expression of any of the tested genes individually (Table 2). However, we found a clear correlation between the ratio between Spinophilin and PPP1CA or PPP1CB and the activity of oxaliplatin or bortezomib (Table 2, Supplementary Figure 2). Alternatively, we did not observe a correlation between metformin activity and Spinophilin or PP1 values. These data suggest that the lower ratio between the levels of expression of these genes may be a good marker for the response to oxaliplatin or bortezomib. Therefore, tumors with lower levels of Spinophilin might respond better to oxaliplatin and/or bortezomib depending on the levels of PPP1CA/B. However, this functional hypothesis needs more research and to be validated in animal models. DISCUSSION Downregulation of Spinophilin, either in protein or mRNA, is related to worse prognosis in lung tumors. This effect is more relevant in squamous cell carcinoma than in adenocarcinoma. Downregulation of Spinophilin is related to a decrease in the levels of PPP1CA/B/C, the catalytic subunits of PP1 and partners of Spinophilin. A decrease in these subunits is also related to a poor prognosis in SCC and is observed more clearly in combination with Figure 6: GO term enrichment by genes that correlate positively or negatively to Spinophilin levels. The analysis by the Enrichr portal was performed on the genes from Supplementary Table 5. The genes were obtained by R2 analysis in data from the TCGA database. Oncotarget105204 www.impactjournals.com/oncotarget a decrease in Spinophilin. PP1 has been identified as the major enzyme that dephosphorylates pRb during mitosis [34, 35] and plays an important role in the G1/S transition [36]. Although there is literature supporting the role of PP1 regulation of pRb in vivo [37–39], our work is the first to support the downregulation of the components of the PP1 heterodimer as a direct contribution to lung cancer and as a predictor of a worse prognosis for these patients. Figure 7: Correlation between the mRNA levels of Spinophilin and the catalytic subunits of PP1 in a panel of tumor cell lines. We analyzed a cohort of 17 cell lines described in Supplementary Table 7. (A) We analyzed the expression levels of Spinophilin in the cohort of 17 cell lines described in Supplementary Table 3. We divided the panel into high and low Spinophilin, considering high Spinophilin those cell lines with expression > 0.01: H1437, H1781, H2009, H358, Calu3 and Nuli1 cell lines; and low Spinophilin the ones with expression < 0.01: H1650, H1975, H2228, H226, H3122, H460, H520, HCC827, Calu1, A549 and NL20 cell lines. (B) We detected a direct relationship between the levels of Spinophilin and the levels of each catalytic subunit. The graph shows the distribution of the PP1 catalytic subunit according to the categorization of samples with high or low Spinophilin according to its average. We considered high Spinophilin: H1431, H1781, H2009, H358, Calu3 and Nuli1 cell lines, and low Spinophilin: H1650, H1975, H2228, H226, H3122, H460, H520, H827, Calu1, A549 and NL20 cell lines. Furthermore, the inset shows Pearson’s r and its statistical significance. Table 2: Pearson correlation (r) between the IC50 of the different treatments and the levels of the indicated genes Spinophilin PPP1CA PPP1CB PPP1CC Spinophilin/ PPP1CA* Spinophilin/ PPP1CB* Spinophilin/ PPP1CC* metformin -0.2567736 -0.0305505 0.0471795 0.0072612 -0.1737827 -0.2997706 -0.3934922 oxaliplatin 0.1098795 -0.0758366 -0.2116024 0.2138328 0.5907283 0.6886702 0.0679746 cisplatin 0.0367228 0.06338384 0.0771211 -0.1303417 -0.1979426 -0.1891476 -0.0018704 etoposide 0.0134608 -0.0350167 -0.0008127 -0.0196913 -0.0707570 -0.0349190 -0.0275979 bortezomib -0.0652384 -0.3027230 -0.0553075 0.2275598 0.7996544 0.4329986 0.0972166 * indicates the ratio between the levels of Spinophilin and the catalytic subunit indicated.