DARPP-32 and t-DARPP promote lung cancer growth through IKKα-dependent cell migration and Akt/Erk-mediated cell survival

Lung cancer is the leading cause of cancer-related death worldwide. In this study, we demonstrate that elevated expression of dopamine and cyclic adenosine monophosphate-regulated phosphoprotein, Mr 32000 (DARPP-32) and its truncated splice variant t-DARPP promotes lung tumor growth, while abrogation of DARPP-32 expression in human non-small cell lung cancer (NSCLC) cells reduces tumor growth in orthotopic mouse models. We observe a novel physical interaction between DARPP-32 and inhibitory kappa B kinase-α (IKKα) that promotes NSCLC cell migration through non-canonical nuclear factor kappa-light-chain-enhancer of activated B cells 2 (NF-κB2) signaling. Bioinformatics analysis of 513 lung adenocarcinoma patients reveals elevated t-DARPP isoform expression is associated with poor overall survival. Histopathological investigation of 62 human lung adenocarcinoma tissues also showed that t-DARPP expression is elevated with increasing tumor (T) stage. Our data suggest that DARPP-32 is a negative prognostic marker associated with increasing stages of NSCLC and may represent a novel therapeutic target.


Introduction
Lung cancer is the leading cause of cancer deaths among both men and women (Torre et al., 2016). In 2017, an estimated 160,420 lung cancer deaths will occur in the United States (Siegel et al., 2017). Nonsmall cell lung cancer (NSCLC) represents 85-90% of all cases of lung cancer and carries a very poor survival rate with less than 15% of patients surviving more than five years (Cetin et al., 2011;Molina et al., 2008). Despite administration of standard chemotherapeutic agents with evolving systemic cancer therapies directed at driver mutations (EGFR, BRAF and ALK), inhibiting angiogenesis (anti-VEGF therapy) and immune-checkpoint blockade (anti-PD-1 antibody), these statistics remain dismal due to the large number of patients diagnosed with advanced stage disease and the primary and secondary resistance to current therapies. A better understanding of the mechanisms that regulate lung tumor growth, metastasis and drug resistance will result in new diagnostic tools and therapeutic strategies to improve the clinical outlook and quality of life of patients afflicted with this deadly disease.
In the early 2000s, El-Rifai and colleagues discovered DARPP-32 is frequently amplified and upregulated in gastric cancer (Belkhiri et al., 2005;El-Rifai et al., 2002). Cloning and sequence assembly analysis revealed a novel transcriptional splice variant of DARPP-32 is also overexpressed in gastric cancer. The N-terminally truncated isoform of DARPP-32, termed t-DARPP, was found to utilize a unique alternative first exon located within intron 1 of DARPP-32 and to lack the first 36 amino acids of DARPP-32, including the T34 phosphorylation residue required for DARPP-32-mediated PP-1 inhibition (El-Rifai et al., 2002). Overexpression of both DARPP-32 and t-DARPP has been observed in 68% of gastric cancers (Belkhiri et al., 2005;El-Rifai et al., 2002). Elevated expression levels of DARPP-32 and t-DARPP have also been associated with many adenocarcinomas, including stomach, colon, prostate and breast cancers (Beckler et al., 2003;Belkhiri et al., 2012;Christenson et al., 2015;Gu et al., 2009;Vangamudi et al., 2010;Wang et al., 2005). Reports have implicated DARPP-32 and t-DARPP in cancer cell proliferation, survival, invasion and angiogenesis . Several studies have demonstrated that DARPP-32 and t-DARPP protect cancer cells from drug-induced apoptosis, which is dependent upon their T75 phosphorylation residue (Belkhiri et al., 2005;El-Rifai et al., 2002) and involves upregulation of Akt and Bcl2 proteins (Belkhiri, Dar, Zaika, et al., 2008;Belkhiri et al., 2012;Zhu et al., 2011). To date, the role of DARPP-32 isoforms in lung cancer remains unexplored. However, we recently described the role of dopamine signaling in NSCLC by demonstrating dopamine D2 receptor agonists inhibit lung cancer growth by reducing angiogenesis and tumor infiltrating myeloid derived suppressor cells in preclinical orthotopic murine models (Hoeppner et al., 2015). Given the role of dopamine signaling in lung cancer and the oncogenic nature of DARPP-32 isoforms in a variety of tumor types, we sought to determine whether DARPP-32 and t-DARPP contribute to lung cancer growth, progression and drug resistance.
Nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) is a transcription factor that regulates numerous biological processes, such as immunity, inflammation, cell growth, differentiation, migration, tumorigenesis and apoptosis (Hayden et al., 2008). The family of NF-κB proteins is comprised of structurally homologous transcription factors, including NF-κB1 (p105/50), NF-κB2 (p100/52), RelA (p65), RelB and c-Rel (Caamano et al., 2002). In the absence of external stimuli, NF-κB proteins are sequestered in the cytoplasm by specific inhibitory proteins, inhibitors of NF-κB (IκBs) (Hayden et al., 2008). When a cell receives appropriate stimuli, IκB kinase (IKK) phosphorylation is initiated, leading to proteasome-mediated processing of p105 and p100. This cleavage event generates their respective mature proteins, p50 and p52, resulting in the nuclear translocation of previously sequestered NF-κB members (Beinke et al., 2004). NF-κB signaling has been categorized into canonical and non-canonical (i.e. alternative) pathways. Recent studies have shown that both canonical and non-canonical NF-κB pathways are capable of promoting oncogenesis by interacting with other cellular pathways in breast cancer, pancreatic ductal adenocarcinoma (PDAC) and glioblastomas (Bang et al., 2013;Kendellen et al., 2014;Rinkenbaugh et al., 2016). NF-κB1 pathway activation causes induction of the IKK complex that contains two catalytic subunits, IKKα and IKKβ and one scaffold subunit called nuclear factor κB essential modulator (NEMO) or IKKγ (Z. J. Chen et al., 1996;DiDonato et al., 1997). Dysregulation of the IKK complex can initiate constitutive activation of the NF-κB1 pathway in cancer cells (Baldwin, 2001). Noncanonical NF-κB2 signaling requires IKKα to mediate p100 cleavage into p52, but does not depend upon IKKβ and NF-κB essential modulator (NEMO), which are essential for canonical NF-κB1 signal transduction (Dejardin et al., 2002;Senftleben et al., 2001). A recent finding has suggested that constitutive activation of KRAS and IKK/NF-κB1 pathways expedites tumorigenesis and worsens survival in PDAC patients (Ling et al., 2012). Ablation of constitutive IKK activity by small molecule inhibitor reduces cellular NF-κB1 activity and melanoma cell survival in vitro and in vivo (Yang et al., 2006). A recent report has suggested that proinflammatory H. pylori infection and canonical NF-κB1 activation play a significant role in the regulation DARPP-32 expression, which has been shown to counteract infection-induced cell death and promote cell survival in gastric carcinogenesis .
We aimed to investigate the role of DARPP-32 isoforms in NSCLC. Here we demonstrate that DARPP-32 and t-DARPP promote cell survival and non-canonical NF-κB2 p52-mediated cell migration in lung cancer. In NSCLC patients, elevated expression of t-DARPP was found to be associated with tumor stage and worsened patient survival.

DARPP-32 and t-DARPP promote cell survival in NSCLC through activation of Akt and Erk signaling
Given the oncogenic role of DARPP-32 in gastric and breast cancer progression (Belkhiri et al., 2005;Christenson et al., 2014;Vangamudi et al., 2010), we sought to determine whether DARPP-32 proteins regulate cell survival in NSCLC. First, we stably silenced endogenous DARPP-32 protein expression through lentiviral shRNA-mediated knockdown in A549 and H1650 human lung adenocarcinoma cells as well as H226 human lung squamous cell carcinoma cells (Fig. 1a, b, c). Two shRNAs targeting distinct regions of DARPP-32 were utilized to decrease the likelihood of potentially confounding off-target effects ( Fig. 1a, b, c). To determine the role of DARPP-32 in regulation of cell survival, we first assessed apoptosis through flow cytometry-based annexin V assays. We observed increased annexin V positive cells in DARPP-32 knockdown cell lines compared to controls (Fig. 1d, e, f) suggesting that DARPP-32 inhibits apoptosis in lung cancer cells. Based on this finding, we next performed a colorimetric cell viability assay in A549 and H226 cells stably transduced with retrovirus to overexpress exogenous DARPP-32 proteins (Fig. 1g, h). Cell viability was significantly increased in DARPP-32 overexpressing cells compared to controls (Fig. 1i,j). An N-terminally truncated isoform and transcriptional variant of DARPP-32, called t-DARPP, lacks the protein phosphate inhibitory (PP-1) domain, which is phosphorylated at threonine 34 (T34) and important for dopamine signaling function (El-Rifai et al., 2002). Overexpression of t-DARPP in A549 and H226 lung cancer cells increased viability (Fig. 1i, j), suggesting that the N-terminal T34-dependent PP-1 regulatory function of DARPP-32 (Huang et al., 1999) does not contribute to regulation of cell viability. Given the role of t-DARPP in promoting cellular proliferation in gastrointestinal cancer (Vangamudi et al., 2011), we sought to determine whether DARPP-32 and t-DARPP proteins regulate proliferation of NSCLC cells. We found modulation of DARPP-32 isoforms does not alter proliferation of lung cancer cells using flow cytometry-based BrdU cell proliferation assays upon silencing endogenous DARPP-32 and overexpression of DARPP-32 and t-DARPP ( Supplementary Fig. 1). Taken together, our findings suggest that DARPP-32 and t-DARPP promote lung tumor cell survival by regulating apoptosis but do not control cellular proliferation.

DARPP-32 promotes lung cancer cell migration
DARPP-32 is upregulated in various cancers including breast and gastric cancer, in which expression of DARPP-32 is associated with increased migration and invasion (Hansen et al., 2009;Zhu et al., 2013). To determine the role of DARPP-32 in NSCLC motility, we performed scratch wound healing assays using A549 and H1650 lung adenocarcinoma cells. We observed a significant decrease in cellular migration of DARPP-32 shRNA silenced A549 and H1650 cells compared to controls (Fig. 3a, b). We next investigated whether overexpression of DARPP-32 enhances cell motility in lung cancer cells. DARPP-32, as well as t-DARPP and the T34A DARPP-32 mutant, promoted increased migration in A549 and H1650 cells (Fig. 3c, d). To validate this result using a more physiologically relevant three-dimensional culture system, we performed Matrigel spot assays to assess lung tumor cell migration. A549 and H1650 lung adenocarcinoma cells stably transduced with lentivirus encoding control or DARPP-32 shRNAs were mixed with Matrigel and spotted on a cell culture plate followed by addition of medium. Similar to our previous findings, DARPP-32 knockdown substantially decreased tumor cell migration in A549 and H1650 cells ( Supplementary Fig. 2a, b). Moreover, A549 and H1650 cells stably overexpressing DARPP-32, t-DARPP or mutant DARPP-32 (T34A) increased cell migration compared to control ( Supplementary   Fig. 3a, b). Taken together, our results suggest DARPP-32 promotes lung tumor cell migration.

DARPP-32 interacts with IKKα to activate non-canonical NF-ĸB2 signaling
We sought to determine the molecular signaling through which DARPP-32 promotes migration. Lung tumor cell migration has been previously shown to be regulated by non-canonical NF-ĸB2 signaling (Yeudall et al., 2012). Thus, we hypothesized that DARPP-32 stimulates cell migration through modulation of non-canonical NF-ĸB2 signaling. In stimulated cells, NF-ĸB inducing kinase (NIK) activates inhibitory kappa B kinase-α (IKKα), which in turn, phosphorylates cytosolic NF-ĸB2 p100 causing its cleavage to NF-ĸB2 p52, which translocates to the nucleus to transcriptionally regulate gene expression (Cildir et al., 2016). We found DARPP-32 knockdown decreased cytosolic phosphorylated NF-ĸB2 p100 and nuclear NF-ĸB2 p52 protein expression in A549 and H1650 cells (Fig. 4a, b). Given our immunoblotting result suggesting DARPP-32 promotes nuclear p52 expression, we sought to determine whether elevated DARPP-32 increases nuclear NF-ĸB2 p52 localization using immunofluorescence. Indeed, we observed significantly greater nuclear localization of p52 in A549 and H1650 lung cancer cells overexpressing DARPP-32, t-DARPP, or DARPP-32 T34A relative to controls (Fig. 4c,d). In accordance with these results, western blot data confirmed that overexpression of DARPP-32 activates NF-κB2 signaling by increasing the expression of cytosolic phosphorylated p100 and nuclear p52 protein expression ( Supplementary Fig. 4). Interestingly, knockdown (Fig. 4a, b) or overexpression ( Supplementary Fig. 4b, d) of DARPP-32 had no effect on phosphorylation of cytosolic IKKα, suggesting activation of NF-κB2 signaling is regulated by DARPP-32 in an NIK-independent manner. Thus, we sought to determine whether DARPP-32 is capable of activating NF-κB2 signaling in an NIKindependent manner through a direct interaction with IKKα. We demonstrate a physical association between DARPP-32 and IKKα through co-immunoprecipitation studies in A549 and H1650 human lung adenocarcinoma cells (Fig. 4e, f). The interaction of DARPP-32 and IKKα was significantly decreased upon DARPP-32 ablation (Fig. 4e, f). Taken together, our findings suggest that DARPP-32 activates noncanonical NF-κB2 signaling by interacting with IKKα.

DARPP-32 positively regulates cell migration by promoting nuclear translocation of NF-κB2 p52
We next aimed to further investigate how DARPP-32 and IKKα regulate cell migration. The IKKαdependent non-canonical NF-ĸB2 pathway has a well-documented role in cell motility (Kew et al., 2012).
To confirm the role of IKKα in the non-canonical NF-ĸB2-mediated regulation of tumor cell migration, we performed the scratch wound healing assay with control or IKKα-depleted NSCLC cells ( Fig. 5a and Supplementary Fig. 5a, b). We observed a substantial decrease in migration of IKKα shRNA transduced cells compared to those expressing control shRNA (Fig. 5a). Next, we examined migration in A549 and signaling. To test whether DARPP-32 requires downstream IKKα and NF-κB2 signaling to promote migration, we overexpressed DARPP-32 upon shRNA-mediated knockdown of IKKα or NF-ĸB2 in human NSCLC cells ( Supplementary Fig. 6a, b). Migration was not significantly altered in IKKα-or NF-ĸB2-depleted NSCLC cells upon overexpression of DARPP-32, but migration was significantly increased when DARPP-32 was overexpressed in the absence of IKKα or NF-ĸB2 knockdown (Fig. 5c, d). Thus, our findings suggest that DARPP-32 acts specifically through IKKα and NF-ĸB2 signaling to induce lung tumor migration ( Supplementary Fig. 7).

DARPP-32 promotes lung tumor growth in orthotopic murine models
Based on our findings that DARPP-32 promotes lung cancer survival and migration, combined with previous studies implicating DARPP-32 as an oncogenic factor contributing to breast cancer and gastric tumor progression (Z. Chen et al., 2016;Vangamudi et al., 2010), we sought to determine whether DARPP-32 drives lung cancer growth in vivo. To this end, we tested whether DARPP-32 ablation reduces lung tumor growth in an orthotopic xenograft mouse model. Briefly, we injected luciferase-labeled human A549 NSCLC cells into the left thorax of anesthetized SCID mice, allowed establishment of the lung tumor and then xenogen imaged the mice regularly over the course of five to six weeks. We observed a significant decrease in lung tumor growth in mice challenged with DARPP-32 ablated A549 cells compared to mice challenged with cells transduced with control LacZ shRNA (Fig. 6a). Correspondingly, we found decreased tumor growth in mice orthotopically injected with DARPP-32 ablated H1650 ( Fig.   6b) or H226 (Fig. 6c) human lung cancer cells. These findings support our hypothesis that DARPP-32 depletion inhibits human lung tumor growth. We next sought to determine whether overexpression of DARPP-32 promotes lung tumor growth in vivo. To address this question, we injected luciferase-labeled human A549 NSCLC cells stably overexpressing exogenous DARPP-32 or t-DARPP into the left thorax of anesthetized SCID mice. We demonstrated that DARPP-32 and t-DARPP overexpression promotes lung tumor growth in mice (Fig. 6d). Taken together, our data suggest DARPP-32 proteins drive lung tumorigenesis and inhibition of DARPP-32 reduces lung cancer growth.

Elevated t-DARPP expression positively correlates with tumor (T) staging score in a large cohort of lung adenocarcinoma patients
We aimed to elucidate the clinical relevance of DARPP-32 given its role in promoting tumor growth in mouse models of human NSCLC. Correspondingly, previous studies have linked upregulation of DARPP-32 and t-DARPP with breast, gastric and colorectal cancer (Z. Chen et al., 2016;Christenson et al., 2014;El-Rifai et al., 2002;Hong et al., 2012;Kopljar et al., 2015;Wang et al., 2005). To assess DARPP-32 and t-DARPP expression in NSCLC patients, we obtained tissue specimens from 62 lung adenocarcinoma patients and performed differential immunohistochemistry to detect the expression of DARPP-32 and t-DARPP. Specifically, we individually immunostained serial whole tissue sections of formalin-fixed paraffin embedded tissue blocks corresponding to each patient with two distinct DARPP-32 antibodies that: 1) detects both DARPP-32 and t-DARPP via a C-terminal epitope present in both isoforms, or 2) exclusively detects DARPP-32 through an N-terminal epitope absent in the N-terminally truncated t-DARPP isoform. Because most of the patients in our cohort had Stage III lung adenocarcinoma (Supplementary Table 1), we used the tumor (T) staging score (i.e. from the 7 th edition of the lung cancer TNM staging system (Mirsadraee et al., 2012)), which represents the size of the primary tumor and whether it has grown into nearby areas, as a metric of tumor progression and growth. A pulmonary pathologist (ACR) scored the percentage of positive tumor cells and their staining intensity of DARPP-32 only and both isoforms (DARPP-32 and t-DARPP) using a scale of 0-3 (i.e. 0= none, 1= weak, 2= moderate, 3= strong expression). Using the resulting pathological scoring, we calculated an immune reactive (IR) score for each specimen based on the percentage of tumor cells staining positive and the staining intensity in those cells (IR score = percentage of tumor cells x staining intensity). We found that high relative expression of t-DARPP correlates with worsening T staging score in the 62 lung adenocarcinoma specimens examined by immunohistochemistry (Fig. 7a, b). Our results suggest that a subset of patients with advanced lung adenocarcinoma exhibit elevated levels of t-DARPP protein and that upregulation of t-DARPP appears to be associated with T staging score.

Upregulation of t-DARPP, NF-ĸB2 and IKKα in lung adenocarcinoma is associated with decreased patient survival
We utilized a bioinformatics approach to validate our finding that high relative t-DARPP expression correlates with tumor growth in lung adenocarcinoma patients. We assessed relative DARPP-32 and t-DARPP transcript expression in specimens corresponding to 513 human lung adenocarcinoma patients cataloged in The Cancer Genome Atlas (TCGA). Interestingly, we found that expression of t-DARPP increases with advancing tumor (T) stages in lung adenocarcinoma (Fig. 8a). As assessed by Kaplan-Meier survival curve, we observed that patients with high t-DARPP expression showed substantially decreased survival relative to lung adenocarcinoma patients with low t-DARPP expression (Fig. 8b). Our findings indicate that t-DARPP expression is an important determinant of survival in lung adenocarcinoma patients.
Given our findings that DARPP-32 isoforms regulate non-canonical NF-ĸB2-mediated cell migration, we asked whether expression of NF-ĸB2 or IKKα is associated with overall survival of lung adenocarcinoma patients. RNA-Seq expression data from 201 human lung adenocarcinoma tissue samples was used to generate Kaplan-Meier survival curves. Our results reveal significantly decreased survival in the patients with high expression of NF-ĸB2 and IKKα transcripts compared to low expressers of those mRNAs (Fig.   8c, d). Thus, upregulation of NF-ĸB2 and IKKα expression is associated with decreased overall patient survival and may predict poor clinical outcome in lung adenocarcinoma patients.

Discussion
For the first time, we demonstrate DARPP-32 and its splice variant t-DARPP stimulate lung cancer cell survival and migration to promote oncogenesis, and we show elevated t-DARPP isoform levels in NSCLC patients are associated with increased tumor staging and worsened patient survival. The role of DARPP-32 and t-DARPP in cancer has emerged beyond their classical function as modulators of dopamine-mediated neurotransmission, highlighting their importance in the regulation of physiological and pathological effects. For example, alternations in expression of DARPP-32 and t-DARPP have been implicated in schizophrenia, bipolar disorder and Alzheimer's disease (Cho et al., 2015;Kunii et al., 2014) as well as numerous types of tumors, including breast, gastric, prostate, esophageal and colon cancers (Beckler et al., 2003;El-Rifai et al., 2002;Hamel et al., 2010;Vangamudi et al., 2010). Since investigation of the frequent amplification at the 17q12 locus in gastric cancers implicated DARPP-32 and t-DARPP in oncogenesis (Belkhiri et al., 2005;El-Rifai et al., 2002), numerous studies have demonstrated the role of these proteins in cancer cell survival, drug resistance, migration, invasion and angiogenesis .
Our results suggest DARPP-32 and t-DARPP promote NSCLC cell survival through activation of Akt and Erk1/2 signaling by protecting cells from apoptotic cell death (Figs. 1-2). Correspondingly, overexpression of DARPP-32 and t-DARPP in human gastrointestinal adenocarcinoma cells was shown to cause a four-fold reduction in apoptosis (Belkhiri et al., 2005). The T75 phosphorylation residue shared by DARPP-32 and t-DARPP was attributed to promoting cell survival (Belkhiri et al., 2005), and a follow-up report by the same group suggested increased activation of Akt and Bcl2 is mechanistically responsible for t-DARPP-mediated cancer cell survival (Belkhiri, Dar, Zaika, et al., 2008). Evasion of apoptosis is a major underlying mechanism in the ability of cancer cells to acquire resistance to molecular targeted therapies (Rotow et al., 2017). El-Rifai and colleagues demonstrated that DARPP-32 promotes cell survival and gefitinib resistance in gastric cancer cells by stimulating EGFR phosphorylation and activating PI3K/Akt signaling . DARPP-32 stimulated resistance to pro-apoptotic proteins through induction of pro-survival molecule Bcl-xL through Src/STAT3 signaling cascades . Numerous reports have implicated t-DARPP in breast cancer patients acquiring resistance to trastuzumab (Herceptin), a monoclonal antibody targeting the ERBB2 (Her2/neu) receptor.
Collectively, these studies demonstrated that t-DARPP drives breast cancer cell resistance to trastuzumab through inhibition of apoptotic caspase-3 and activation of pro-survival Akt signaling through its T75 residue, common among both DARPP-32 isoforms (Belkhiri, Dar, Peng, et al., 2008;Gu et al., 2009;Hamel et al., 2010). Another report showed t-DARPP promotes trastuzumab resistance in esophageal adenocarcinoma cells through similar mechanisms (Hong et al., 2012). In both breast and esophageal cancer, t-DARPP physically interacted with ERBB2 in a protein complex to mediate trastuzumab resistance (Belkhiri, Dar, Peng, et al., 2008;Hong et al., 2012). Like previous studies in gastric, breast and esophageal cancers, our studies suggest DARPP-32 and t-DARPP promote cell survival through upregulation of Akt signaling. Despite this observation and the common underlying mechanistic evidence, future studies beyond the scope of this manuscript are necessary to determine whether t-DARPP promotes resistance to specific molecular targeted NSCLC therapies.
Like pro-survival mechanisms, increased cell migration also contributes to cancer cell growth and resistance to molecular targeted therapies. Given the well-established association between DARPP-32 isoforms and acquired drug resistance in cancer, it is unsurprising that several reports and detailed reviews have described the role of DARPP-32 in breast and gastric cancer cell migration and invasion Hansen et al., 2006;Zhu et al., 2013). Correspondingly, we provide evidence that DARPP-32 and t-DARPP promote NSCLC cell migration based on in vitro scratch (Fig. 3) (Kopljar et al., 2015;Zhu et al., 2013). Conversely, DARPP-32 has been shown to inhibit breast cancer cell migration through a dopamine D1 receptor-dependent mechanism (Hansen et al., 2006). A subsequent in vitro study has revealed that PP-1 inhibition regulated by phosphorylation of DARPP-32 at residue T34 is critical for modulating cell migration in breast cancer (Hansen et al., 2009). Taken together, the regulation of cancer cell migration by DARPP-32 is likely cell and tumor type dependent.
We identify a novel physical interaction between DARPP-32 and IKKα that suggests DARPP-32 regulates non-canonical NF-κB2 signaling to control NSCLC migration (Fig. 4). Knockdown of IKKα, as well as independently silencing NF-κB2, decreased migration of human lung adenocarcinoma cells (Fig.   5a, b). Based on our findings, we propose that DARPP-32 activates IKKα through an unknown NIKindependent mechanism that leads to IKKα-mediated phosphorylation of NF-ĸB2 p100, ubiquitination and partial degradation of p100 to p52, and translocation of NF-ĸB2 p52 to the nucleus where it acts as a transcription factor to modulate expression of genes involved in cell migration ( Supplementary Fig. 7). A recent report has demonstrated that Helicobacter pylori infection induces canonical NF-ĸB1-mediated transcriptional upregulation of DARPP-32 mRNA and protein, which counteracts Helicobacter pylorimediated cell death through activation of Akt . Therefore, we investigated whether non-canonical NF-κB2 signaling altered DARPP-32 protein expression, but observed no effect ( Supplementary Fig. 5). While canonical NF-ĸB1 pathway activation has been linked to the growth and survival of many solid and hematological malignancies, the role of non-canonical NF-ĸB2 signaling in cancer is still emerging (Cildir et al., 2016;Hayden et al., 2008). However, studies suggest the NF-ĸB2 pathway is activated in cancer through viral oncogenes, mutations in pathway components, and upregulation of upstream components of the pathway (Cildir et al., 2016), the latter of which is supported by our results, suggesting DARPP-32 promotes activation of non-canonical NF-ĸB2 signaling in lung cancer through an interaction with IKKα.
We demonstrate stable overexpression of DARPP-32 and t-DARPP in human NSCLC cells orthotopically implanted into the thoracic cavity of SCID mice promotes tumor growth (Fig. 6d). Correspondingly, mice that received an orthotopic xenograft of shRNA-mediated DARPP-32 silenced NSCLC cells exhibited decreased tumor growth relative to controls (Fig. 6a, b, c). El-Rifai and colleagues have shown overexpression of t-DARPP in human OE19 esophageal adenocarcinoma subcutaneously xenografted into athymic nude mice stimulates tumor growth (Hong et al., 2012). Using a similar xenograft mouse model, they have subsequently demonstrated shRNA-mediated knockdown of DARPP-32 reduces gastric tumorigenesis  and overexpression of DARPP-32 in AGS human gastric adenocarcinoma cells promotes in vivo tumor growth (Z. Chen et al., 2016). To the best of our knowledge, our study is the first to assess DARPP-32 knockdown as well as DARPP-32 and t-DARPP overexpression in an orthotopic cancer xenograft mouse model. Importantly, our in vivo results showing DARPP-32 and t-DARPP promote NSCLC oncogenesis coincide with similar findings in esophageal and gastric cancer subcutaneous xenograft models.
Based on differential immunostaining of over 60 human NSCLC specimens, we describe that high relative expression of t-DARPP correlates with tumor staging in lung adenocarcinoma patients (Fig. 7).
Similar differential immunohistochemistry approaches in serial tissue sections have been previously used to distinguish between detection of DARPP-32 only (N-terminal antibody) versus both isoforms (C-terminal antibody). Two independent studies have demonstrated a subset of primary human breast cancer specimens exhibit elevated t-DARPP protein levels relative to DARPP-32 (Hamel et al., 2010;Vangamudi et al., 2010). Using a genetic spontaneous murine model of breast cancer, Christenson and Kane have found DARPP-32 was expressed in normal mammary tissue and in some breast tumors, whereas t-DARPP was detected exclusively in tumors, typically at higher or equal levels as DARPP-32 (Christenson et al., 2014). This transition from DARPP-32 to t-DARPP observed during breast tumorigenesis corresponds to our pathological and bioinformatics findings linking upregulation of t-DARPP expression with increased NSCLC growth and worsened patient survival. The DARPP-32 to t-DARPP isoform shift in cancer may be directed by the SRp20 splicing factor, which has been shown to physically associate with DARPP-32 . The upregulation of t-DARPP in NSCLC progression suggests its expression stimulates oncogenesis. Thus, t-DARPP may represent a promising molecular target in NSCLC as well as possess prognostic value.

Cell Culture
Human NSCLC cell lines A549, H1650 and H226 as well as the transformed human embryonic kidney epithelial cell line, HEK-293T, were purchased from American Type Culture Collection (Manassas, VA) and maintained according to the manufacturer's instructions. HEK-293T cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Corning; Manassas, VA) and lung cancer cell lines were cultured in RPMI-1640 medium (Corning). Media was supplemented with 10% fetal bovine serum (FBS; Millipore; Burlington, MA) and 1% Penicillin/Streptomycin antibiotics (Corning). All cell lines were certified by the indicated cell bank and routinely authenticated by morphologic inspection.

Transient transfections
5x10 5 A549 or H1650 cells were seeded in 60-mm cell culture plates and incubated for 24 h in RPMI-1640 medium. Cells were then washed with PBS, suspended in OPTI-MEM reduced serum medium (Gibco; Grand Island, NY), and transfected with 2 µg of pMMP-LacZ or pMMP-DARPP-32 plasmids using Polyfect transfection reagent (Qiagen; Hilden, Germany) according to instruction from the manufacturer. After 4 h, antibiotic-containing complete RPMI-1640 medium was added and cells were grown until they had established a confluent monolayer.

Generation of stable cell lines
Expression constructs of human DARPP-32, t-DARPP and DARPP-32 T34A cDNA in pcDNA3.1 were a generous gift from Dr. Wael El-Rifai at Vanderbilt University Medical Center (Z. Chen et al., 2016). The Flag-tagged coding sequence of DARPP-32, t-DARPP and T34A DARPP-32 were subcloned into the retroviral (pMMP) vector. The pMMP plasmid and its corresponding pMMP-LacZ control construct were kindly provided by Dr. Debabrata Mukhopadhyay at Mayo Clinic in Jacksonville, Florida (Zeng et al., 2001). Production of retrovirus and transduction of A549, H1650 and H226 lung cancer cell lines were performed as previously described (Zeng et al., 2001).
Four to five different lentiviral shRNA pLKO.1 constructs (Sigma-Aldrich; St. Louis, MO) were used to silence protein expression of each target, including DARPP-32, NF-ĸB2 and IKKα. pLKO.1-LacZ shRNA (Sigma-Aldrich) was used as a corresponding control. Generation of lentivirus and transduction of A549, H1650 and H226 lung cancer cell lines were accomplished as previously described (Alam et al., 2016).

Cell survival assay
A549, H226 and H1650 human NSCLC cell lines were each plated in a 96-well microplate at a concentration of 3000 cells/well. Cell viability was assessed after 72 h of incubation using CellTiter 96 ® AQueous One System (Promega; Madison, WI). Absorbance was recorded at 490 nm using an Epoch microplate spectrophotometer (Biotek; Winooski, VT). The average of three independent experiments has been reported.

Cell proliferation analysis by BrdU labeling
Human NSCLC cells were seeded at a density of 1x10 5 cells per 60-mm plate. The following day, bromodeoxyuridine (BrdU; 30 µM; Sigma-Aldrich) diluted in fresh medium was administered to the cells for 30 minutes. The cells were harvested, fixed, and processed for incubation with primary mouse anti- Technology)] and incubated on ice for 15 minutes. Nonionic detergent NP-40 (10%; Sigma Aldrich) was then added to the cell suspension, which was mixed vigorously. Next, the cell homogenate was centrifuged at 5,000 rpm for 10 minutes at 4°C. The supernatant was collected as the cytoplasmic fraction, and the pellet was suspended in cell extraction buffer (Thermo Fisher Scientific; Waltham, MA) supplemented with protease inhibitor cocktail (Roche) and 1 mM PMSF. The suspension was incubated on ice for 30 minutes with intermittent vortexing. Finally, the sample was centrifuged at 14,000 g for 30 minutes at 4°C, and the supernatant was collected as nuclear extract.

Immunoprecipitation
Human lung cancer cells were homogenized and lysed in RIPA buffer (Millipore) supplemented with protease inhibitor cocktail (Roche). Protein concentration was measured using the Quick Start Bradford protein assay (Bio-Rad) and 500 µg of protein lysate was loaded into the supplied spin column (Catch and Release Immunoprecipitation Kit; Millipore). Immunoprecipitation was achieved by following manufacturer's protocol (Cat no.:17-500; Millipore).
The mean nuclear fluorescence was calculated and plotted in GraphPad Prism software (Version 7).

Immunohistochemistry
Human lung adenocarcinoma tissue specimens were obtained from 62 NSCLC patients at Mayo Clinic in Rochester, MN in accordance with IRB approved protocols. We performed differential immunohistochemistry using an N-terminal antibody that extensively recognizes DARPP-32 (Abcam; Cat No.:ab40801). We used another C-terminal antibody that recognizes both DARPP-32 and t-DARPP (Santa Cruz Biotechnology; Cat No.:sc-11365;). Formalin-fixed, paraffin-embedded whole tissues were serially sectioned and immunostained for DARPP-32 using a Bond Autostainer (Leica; Wetzlar, Germany) as previously described (Hoeppner et al., 2015). H&E staining was also performed. In each lung tumor specimen, the intensity and prevalence of DARPP-32 staining in various cell types was scored by a pulmonary pathologist (ACR).

Scratch wound assay
A549 and H1650 cells were seeded in 60-mm culture dishes at an appropriate density to achieve a confluent monolayer. After 16h, a linear scratch wound was generated using a sterile 20 µl pipette tip.
Cells were imaged at time 0 and 14h post-scratch induction. All the images were captured using a 4X Plan S-Apo 0.16 NA objective on an EVOS FL cell imaging system (Thermo Fisher Scientific). The images were analyzed using ImageJ software and cell migration was quantified as previously described (Liang et al., 2007).

Spot Assay
A549 and H1650 human lung cancer cells were trypsinized and suspended in RPMI-1640 medium (Corning) at a concentration of 5x10 4 cells per µl. Cells (2.5x10 5 in 5 µl) were then mixed with Matrigel® Basement Membrane Matrix (Corning) in 1:1 ratio and pipetted as a spot in a 60-mm culture dish.
Matrigel containing cell suspension (i.e. the spot) was allowed to solidify by incubating at 37°C for 5 minutes. Thereafter, medium was added and images were captured using a 4X Plan S-Apo 0.16 NA objective on an EVOS FL cell imaging system (Thermo Fisher Scientific). After a 96h incubation, the spots were imaged again and cell migration was calculated as previously described (Kaur et al., 2012).

In vivo orthotopic lung cancer model
Six to eight-week-old pathogen-free SCID/NCr mice were purchased from the Charles River Laboratories. Mice were allowed one week to acclimate to their surroundings, maintained under specific pathogen-free conditions in a temperature-controlled room with alternating 12h light/dark cycles and fed a standard diet. Mice were orthotopically injected with 1×10 6 luciferase-labeled human A549, H226 and H1650 lung cancer cells suspended in 80 μl PBS and Matrigel. After establishment of the lung tumor, mice were imaged using an In-Vivo Xtreme xenogen imaging system (Bruker; Billerica, MA) to measure luciferase intensity. To determine tumor growth, luciferase intensity was calculated using Bruker molecular imaging software and plotted over time in GraphPad Prism 7 software. All animal studies were performed in accordance with protocols approved by the University of Minnesota Institutional Animal Care and Use Committee. The NF-ĸB2 and IKKα expression data and the clinical variables of 203 human lung adenocarcinoma tissue specimens were obtained from cBioPortal for Cancer Genomics (http://cbioportal.org) (Cerami et al., 2012;Gao et al., 2013). Patients were categorized into 2 separate groups based on the mRNA expressions (normalized read count) and Kaplan-Meier survival curve was generated by using GraphPad Prism 7 software.

Statistics
Data are expressed as mean ± SEM and representative of at least three independent experiments.
Statistical significance was determined using one-way analysis of variance (ANOVA) and a value of P < 0.05 was considered significant.
at Mayo Clinic in Jacksonville, FL for his valuable support and contributions to the development of this project, including sharing reagents and advice. We appreciate the contributions of Todd Schuster, the shared facilities manager at The Hormel Institute, to flow cytometry-based apoptosis and BrdU assays.