Glycogen synthase kinase-3β (GSK-3β), a serine/threonine kinase, has been implicated in the pathogenesis of many cancers, with involvement in cell-cycle regulation, apoptosis, and immune response. Small-molecule GSK-3β inhibitors are currently undergoing clinical investigation. Tumor sequencing has revealed genomic alterations in GSK-3β, yet an assessment of the genomic landscape in malignancies is lacking. This study assessed >100,000 tumors from two databases to analyze GSK-3β alterations. GSK-3β expression and immune cell infiltrate data were analyzed across cancer types, and programmed death-ligand 1 (PD-L1) expression was compared between GSK-3β–mutated and wild-type tumors. GSK-3β was mutated at a rate of 1%. The majority of mutated residues were in the kinase domain, with frequent mutations occurring in a GSK-3β substrate binding pocket. Uterine endometrioid carcinoma was the most commonly mutated (4%) tumor, and copy-number variations were most commonly observed in squamous histologies. Significant differences across cancer types for GSK-3β–mutated tumors were observed for B cells (P = 0.018), monocytes (P = 0.002), dendritic cells (P = 0.005), neutrophils (P = 0.0003), and endothelial cells (P = 0.014). GSK-3β mRNA expression was highest in melanoma. The frequency of PD-L1 expression was higher among GSK-3β–mutated tumors compared with wild type in colorectal cancer (P = 0.03), endometrial cancer (P = 0.05), melanoma (P = 0.02), ovarian carcinoma (P = 0.0001), and uterine sarcoma (P = 0.002). Overall, GSK-3β molecular alterations were detected in approximately 1% of solid tumors, tumors with GSK-3β mutations displayed a microenvironment with increased infiltration of B cells, and GSK-3β mutations were associated with increased PD-L1 expression in selected histologies. These results advance the understanding of GSK-3β complex signaling network interfacing with key pathways involved in carcinogenesis and immune response.

Glycogen synthase kinase-3 (GSK-3), a serine/threonine kinase, is a regulator of multiple signaling pathways (1). One of its isoforms, GSK-3β, acts as both a tumor suppressor and a proto-oncogene, depending on the downstream target (2). This kinase is implicated in the pathogenesis of many cancers, including breast (3), lung (4), pancreas (5), colorectal (6), and leukemia (7). Notably, overexpression of GSK-3β has been shown to promote both tumor growth and chemotherapy resistance in multiple disease states (1, 8).

GSK-3β is constitutively active and phosphorylates over 100 protein targets that include pivotal pathways regulating cellular proliferation and survival such as the Wnt/β-catenin and PI3K/AKT/mTOR pathways (9, 10). The role of GSK-3β in cell-cycle regulation, apoptosis, and tumor metastasis has also been well described (11–13). Furthermore, GSK3 plays a role in both innate and adaptive immune responses, modulating inflammation, T-cell differentiation, and metabolic regulation of B cells (14–16). The interaction between GSK-3β and programmed death-ligand 1 (PD-L1) has also been described, raising interest in the cross-talk between GSK-3β–mediated pathways and regulation of immune function. GSK-3β phosphorylates nonglycosylated forms of PD-L1 leading to its proteasomal degradation (17, 18).

Previous research suggests a relationship between GSK-3β and the activation of NFκB (2, 19). As NFκB activation promotes cancer progression, chemoresistance, and metastasis, it has been hypothesized that inhibiting GSK-3β could lead to downstream inhibition of NFκB, in turn decreasing cancer cell survival (2). With this rationale, potent and specific small-molecule GSK-3β inhibitors were developed and are currently undergoing clinical investigation (NCT03678883; ref. 20). These inhibitors have demonstrated significant antitumor activity observed in preclinical models for a variety of malignancies (21–24). Available tumor-sequencing data have revealed genomic alterations in the GSK-3β gene among many cancer types (25). However, a comprehensive, rigorous analysis of these alterations is lacking.

To assess the genomic landscape of GSK-3β alterations, we utilized two large tumor-sequencing datasets spanning a wide range of cancer histologies. We took a pan-cancer approach as an initial study of the overall genomic landscape to investigate alterations present among different cancer types. Moreover, we assessed differences in PD-L1 expression based on GSK-3β tumor mutational status to further elucidate the relationship between these two proteins. We hypothesized that characterization of molecularly defined subgroups of malignancies harboring GSK-3β alterations will advance the understanding of the pleiotropic GSK-3β signaling network and inform the development of biomarkers associated with anticancer activity of GSK-3β inhibitors and combinations with other agents. In addition, we hypothesized that GSK-3β tumor alterations will impact both the tumor immune cell microenvironment and the expression of PD-L1 in tumor cells.

Data collection

Primary cohort data were obtained from The cBioPortal for Cancer Genomics (26, 27). Data from all available studies, curated by cBioPortal to exclude duplicate tumor samples, were analyzed. References for studies comprising this sample is available in Supplementary File S1. Tumor samples with a GSK-3β genomic alteration, including mutations or copy-number variations (CNV), were included for analysis. For each tumor with a GSK-3β alteration, histology and GSK-3β protein change or type of CNV were obtained. The predicted functional impact of each mutation was assessed using the in silico prediction models Polyphen-2 (28) and SIFT (29).

Secondary cohort data were obtained from Caris Life Sciences Precision Oncology Alliance, referred to as CLSPOA database. A cohort of 73,324 tumors that had been analyzed by Caris Life Sciences from February 2015 to August 2019 were included in this study. The vast majority of these tumor specimens were obtained from patients with metastatic malignancies. This study was conducted in accordance with guidelines of the Declaration of Helsinki, Belmont report, and U.S. Common rule. In keeping with 45 CFR 46.101(b) (4), this study was performed utilizing retrospective, deidentified clinical data. Therefore, this study is considered IRB exempt and no patient consent was necessary from the subject. Tumors containing a GSK-3β mutation were included for analysis. CNV were not included in data obtained from this secondary cohort. For each tumor with a GSK-3β mutation, histology and protein change were obtained. Varsome, a search engine for human genomic variation that contains information from 30 external databases, was used to predict potential functional relevance for select mutations (30).

GSK-3β mRNA expression data were also obtained from CLSPOA. mRNA data were not obtained from cBioPortal. Gene expression was evaluated on mRNA isolated from a formalin-fixed paraffin-embedded (FFPE) tumor sample using the Illumina NovaSeq platform (Illumina, Inc.) and Agilent SureSelect Human All Exon V7 bait panel (Agilent Technologies); transcript per million (TPM) was reported. Microenvironment cell population-counter (MCP-counter) was used for quantification of the abundance of immune and stromal cell population using transcriptomic data as described previously (31).

PD-L1 expression data, assessed via SP-142 antibody using the Dako platform, were also obtained from tumors in the CLSPOA database. PD-L1 expression was assessed by IHC on full FFPE sections of glass slides. Slides were stained using automated staining techniques, per the manufacturer's instructions, and were optimized and validated per Clinical Laboratory Improvement Amendments (CLIA) and International Organization for Standardization requirements. The staining was deemed positive for SP-142 if its intensity on the membrane of the tumor cells was >2+ and the percentage of positively stained cells was >5%. A board-certified pathologist evaluated all IHC results independently.

Data analysis

The most common histologies with a GSK-3β mutation was assessed by combining tumor samples from both datasets. Discrepancies among histology names were settled between the first and senior authors and representatives from Caris Life Sciences to determine which data were able to be reliably combined. Protein changes from each dataset were also combined to ascertain the loci most frequently containing a GSK-3β mutation. Histologies with a GSK-3β CNV were assessed from the cBioPortal dataset.

GSK-3β mRNA expression data in the CLSPOA dataset were assessed via median TPM values. The distribution across cancer types was tested for significance via nonparametric Wilcoxon/Kruskal–Wallis tests and χ2 approximation (P < 0.05 was considered significant). Similarly, nonparametric Wilcoxon/Kruskal–Wallis tests were used to assess differences in median abundance values estimated by MCP-counter for immune and stromal cell subpopulations across cancer types for GSK-3β–mutated tumors, with P < 0.05 considered significant. MCP-counter P values are corrected for multiple comparison using Benjamini/Hochberg method (q < 0.05).

PD-L1 expression data by IHC for tumors containing a GSK-3β mutation in the CLSPOA dataset were compared with that of tumors that did not contain a GSK-3β mutation (GSK-3β wild-type). PD-L1–positive GSK-3β–mutated tumors for a given histology were compared with PD-L1–positive GSK-3β wild type of the same histology using a χ2 test (P < 0.05 was considered significant). Benjamini/Hochberg method was used to calculate q values to correct for multiple comparison. The association between mismatch repair (MMR) deficiency and differences in PD-L1 expression between GSK-3β–mutated and GSK-3β wild-type tumors was also investigated in a cohort of selected histologies (i.e., colorectal adenocarcinoma, endometrial cancer, ovarian epithelial carcinoma) with available results regarding their microsatellite status using Fisher exact tests. A combination of multiple test platforms was used to determine the microsatellite instability MSI or MMR status of the tumors profiled, including fragment analysis (FA, Promega), IHC [MLH1, M1 antibody: MSH2, G2191129 antibody: MSH6, 44 antibody; and PMS2, EPR3947 antibody (Ventana Medical Systems, Inc.)], and next-generation sequencing (NGS; 7,000 target microsatellite loci were examined and compared with the reference genome hg19 from the University of California, Berkeley, CA). Tumors were considered MMR deficient according to IHC if complete absence of protein expression of any of the four proteins (MLH1, MSH2, MSH6, PMS2) was observed; for NGS, tumors were classified as MSI-high (MSI-H) if ≥46 altered loci per tumor was identified; for FA the tumor was classified as MSI-H if ≥2 mononucleotide out of the five markers included in the assay were abnormal. The MSI or MMR status of the tumor was determined in the order of FA, IHC, and NGS.

GSK-3β genomic alterations—cBioPortal cohort

In total, data combined from cBioPortal comprised our primary cohort of 46,237 unique tumor samples. This combined sample included both localized and metastatic tumors. Of these, 430 tumors (1%) had either a GSK-3β mutation or CNV. Specifically, 227 tumors had mutations, while 217 had CNV, defined as a GSK-3β amplification or homozygous deletion. Of the mutations, 183 were unique, with the majority (71%) being missense mutations. Nonsense mutations and gene fusions were seen at rates of 12% and 9%, respectively. Furthermore, frameshift deletions/insertions and splice site/region mutations were each 4% of observed mutations. Of note, 50% of all observed gene fusions occurred in prostate adenocarcinoma, and gene fusions were the most frequent type of GSK-3β alteration found in prostate adenocarcinoma (53%). Finally, of the 217 samples with GSK-3β CNV, 86% of tumors had amplifications, while the remainder had homozygous deletions.

Figure 1 displays the genomic loci of each of the 227 mutations in the cBioPortal dataset. The most frequently mutated loci included R111 (n = 6), R96 (n = 5), R167 (n = 5), R180 (n = 5), R308 (n = 5), and R328 (n = 5). R96 and R180 are key GSK-3β residues that form a binding pocket for the phosphate group of a primed substrate (32). The third residue comprising the binding pocket, K205, was not found to be mutated in the dataset. Other key residues implicated in kinase activation (E97, Y216; ref. 33) and ATP binding (G63, N64, S66, G68, V70, A83, V110, V135, R141, and C199; refs. 34, 35) were found to be mutated and are noted on the figure. Interestingly, Y216, which is phosphorylated leading to constitutive activation of GSK-3β, was mutated in only one tumor specimen in the dataset. A mutation of this residue, however, has been previously shown to reduce GSK-3β activity by 5- to 10-fold (36). No mutations occurred in the residue responsible for inactivation of the GSK-3β (S9). Furthermore, 71% of mutations occurred in the functional kinase domain of GSK-3β.

Figure 1.

Top GSK-3β–mutated loci of the cBioPortal dataset. Loci of GSK-3β mutations are depicted in the figure. Notably, two of the top mutated loci comprise a binding pocket for GSK-3β substrates (highlighted in green). Residues implicated in activation and relevant to ATP binding are also highlighted in blue and purple, respectively.

Figure 1.

Top GSK-3β–mutated loci of the cBioPortal dataset. Loci of GSK-3β mutations are depicted in the figure. Notably, two of the top mutated loci comprise a binding pocket for GSK-3β substrates (highlighted in green). Residues implicated in activation and relevant to ATP binding are also highlighted in blue and purple, respectively.

Close modal

To predict the functional impact of the missense mutations (n = 129), Polyphen-2 and SIFT in silico models were used. SIFT criteria predicted 67% of these mutations to be “deleterious,” while the remainder were classified as “tolerated.” Similarly, Polyphen-2 predicted that 44% of these mutations were “probably damaging,” 19% were “possibly damaging,” and 36% were “benign.”

GSK-3β mutations—CLSPOA cohort

In total, 73,324 tumors were assessed from the CLSPOA database, 819 (1%) of which had GSK-3β mutations. In this cohort, most tumors analyzed were from patients with metastatic disease. Given our findings in the cBioPortal cohort, we specifically assessed mutations at loci in the aforementioned GSK-3β substrate binding pocket (R96, R180, and K205). Missense mutations were observed at R96 (n = 6) and K205 (n = 1), both of which were predicted to be pathogenic by Varsome. Additional substrate binding pocket mutations included R96Q (n = 3), R180Q (n = 4), R180W (n = 3), and K205N (n = 1), all of which were predicted to be of uncertain significance. Of note, 8 (44%) of these substrate binding pocket mutations were found to be in uterine neoplasms.

The most common GSK-3β residue changes identified in cBioPortal and CLSPOA datasets are summarized in Fig. 2. Transcript reference variants used in sequencing varied slightly between the two databases but were largely homologous. cBioPortal data were analyzed again using the same transcript reference variant as Caris to reliably assess the most common residue changes. Overall, the most frequent residue changes observed were R396Q (n = 45) and H310Q (n = 40). Although R396Q is outside of the kinase domain and its functional significance is unknown, this residue change has been previously reported in the literature in a case of hairy cell leukemia (37).

Figure 2.

Most common GSK-3β residue changes. The figure shows the most frequent residue changes from the combined cBioPortal and CLSPOA datasets. cBioPortal data were aligned to the same transcript variant reference sequence used by Caris for analysis. Of note, in this reference sequence, residues 304–316 are part of an insertion, which was not sequenced by cBioPortal. SBP, substrate binding pocket.

Figure 2.

Most common GSK-3β residue changes. The figure shows the most frequent residue changes from the combined cBioPortal and CLSPOA datasets. cBioPortal data were aligned to the same transcript variant reference sequence used by Caris for analysis. Of note, in this reference sequence, residues 304–316 are part of an insertion, which was not sequenced by cBioPortal. SBP, substrate binding pocket.

Close modal

GSK-3β expression across tumor types

Median TPM values of GSK-3β expression were obtained across tumor types using RNA sequencing from CLSPOA data (Fig. 3). The highest expression values were observed in melanoma, followed by squamous non–small cell lung cancer (NSCLC), whereas the lowest values were seen in pancreatic cancer and adenocarcinoma NSCLC. A statistically significant difference of expression across the studied histologies was observed (P < 0.0001). Notably, when comparing squamous cell and adenocarcinoma subtypes of NSCLC, a significant difference in TPM values was also observed (P < 0.0001).

Figure 3.

GSK-3β mRNA expression across tumor types. Using CLSPOA data, GSK-3β expression was compared between select histologies. A statistically significant difference of expression among multiple histologies was observed (P < 0.0001). A significant difference in TPM values was also observed between squamous cell and adenocarcinoma subtypes of NSCLC (P < 0.0001).

Figure 3.

GSK-3β mRNA expression across tumor types. Using CLSPOA data, GSK-3β expression was compared between select histologies. A statistically significant difference of expression among multiple histologies was observed (P < 0.0001). A significant difference in TPM values was also observed between squamous cell and adenocarcinoma subtypes of NSCLC (P < 0.0001).

Close modal

Most common histologies with GSK-3β alterations

Data from both cohorts (cBioPortal and CLSPOA) were used to assess which histologic subtypes, or histologies, most frequently contained tumors with GSK-3β mutations. Table 1 displays the most commonly mutated histologies. When combining the data from both cohorts, the histologies with most frequent GSK-3β mutations were nonmelanoma skin cancer (3%), uterine neoplasms (as a combined group; 3%), and melanoma (2%). Notably, the subtype of uterine endometrioid carcinoma was mutated at a rate of 4%.

Table 1.

Top histologies with GSK-3β mutations.

HistologyGSK3B mutantTotal samples for given histologyPercent mutated
Uterine endometrioid carcinoma 87 2,366 3.7% 
Nonmelanoma skin cancer 17 618 2.8% 
Uterine neoplasms 171 6,564 2.6% 
Melanoma 70 3,205 2.2% 
Non–small cell lung cancer 203 16,590 1.2% 
Cervical squamous cell carcinoma 17 1,480 1.1% 
Bladder cancer 29 2,528 1.1% 
Colorectal adenocarcinoma 141 12,424 1.1% 
Prostatic adenocarcinoma 33 3,367 1.0% 
Head and neck cancer 13 1,736 0.7% 
Small-cell lung cancer 960 0.7% 
Breast carcinoma 58 8,179 0.7% 
Esophageal cancer 17 2,492 0.7% 
Ovarian epithelial carcinoma 60 10,001 0.6% 
Pancreatic cancer 21 4,564 0.5% 
Renal cancer 1,947 0.4% 
Hepatocellular carcinoma 1,335 0.3% 
HistologyGSK3B mutantTotal samples for given histologyPercent mutated
Uterine endometrioid carcinoma 87 2,366 3.7% 
Nonmelanoma skin cancer 17 618 2.8% 
Uterine neoplasms 171 6,564 2.6% 
Melanoma 70 3,205 2.2% 
Non–small cell lung cancer 203 16,590 1.2% 
Cervical squamous cell carcinoma 17 1,480 1.1% 
Bladder cancer 29 2,528 1.1% 
Colorectal adenocarcinoma 141 12,424 1.1% 
Prostatic adenocarcinoma 33 3,367 1.0% 
Head and neck cancer 13 1,736 0.7% 
Small-cell lung cancer 960 0.7% 
Breast carcinoma 58 8,179 0.7% 
Esophageal cancer 17 2,492 0.7% 
Ovarian epithelial carcinoma 60 10,001 0.6% 
Pancreatic cancer 21 4,564 0.5% 
Renal cancer 1,947 0.4% 
Hepatocellular carcinoma 1,335 0.3% 

Note: Histologies were included if present in both cBioPortal and Caris. Combined cBioPortal and Caris cohorts reveal uterine endometrioid carcinoma, nonmelanoma skin cancer, uterine neoplasms as a group, and melanoma as top mutated histologies.

Histologies most frequently containing a GSK-3β CNV were assessed using the cBioPortal dataset (Table 2). Various squamous cell carcinomas were among the most common histologies with GSK-3β CNV: cervical (6%), lung (5%), head and neck (3%), and ovarian epithelial carcinoma (3%). These histologies differed from those with GSK-3β mutations.

Table 2.

Top histologies with GSK-3β CNV.

HistologyNumber of samples with CNVPercent of total cBio samples with CNV for given histology
Cervical squamous cell carcinoma 15 6% 
Lung squamous cell carcinoma 36 5% 
Head and neck squamous cell carcinoma 15 3% 
Ovarian epithelial carcinoma 15 3% 
Prostate adenocarcinoma 40 2% 
Breast invasive ductal carcinoma 18 1% 
Uterine endometrioid carcinoma <1% 
Melanoma <1% 
Lung adenocarcinoma <1% 
HistologyNumber of samples with CNVPercent of total cBio samples with CNV for given histology
Cervical squamous cell carcinoma 15 6% 
Lung squamous cell carcinoma 36 5% 
Head and neck squamous cell carcinoma 15 3% 
Ovarian epithelial carcinoma 15 3% 
Prostate adenocarcinoma 40 2% 
Breast invasive ductal carcinoma 18 1% 
Uterine endometrioid carcinoma <1% 
Melanoma <1% 
Lung adenocarcinoma <1% 

Note: Notably, top histologies with CNV are distinct from top mutated histologies.

Tumor microenvironment in GSK-3β–mutated tumors

The cellular profile of tumor microenvironment was assessed using MCP-counter method for quantification of the abundance of immune and stromal cell populations using transcriptomic data. MCP counts were obtained for 79 GSK-3β–mutated tumors across seven histologies (breast, colorectal adenocarcinoma, endometrial, melanoma, NSCLC, ovarian, and prostate adenocarcinoma). Specifically, counts of T cells, CD8 T cells, cytotoxic lymphocytes, natural killer cells, B lineage cells, monocytic lineage cells, myeloid dendritic cells, neutrophils, endothelial cells, and fibroblasts were obtained in these GSK-3β–mutated tumors. Each cell population was assessed individually for differences across the aforementioned cancer types. Significant differences were observed for B lineage cells (P = 0.018), monocytic lineage cells (P = 0.002), myeloid dendritic cells (P = 0.005), neutrophils (P = 0.0003), and endothelial cells (P = 0.014).

For B lineage cells, the highest median MCP counts, representing greatest B-cell infiltration among GSK-3β–mutated tumors, were observed in melanoma, while the lowest median, representing the least B-cell infiltration, was observed in breast cancer. For monocytic lineage cells, myeloid dendritic cells, and endothelial cells, the highest median MCP, or greatest infiltrate, was also observed in melanoma, while the lowest was seen in colorectal adenocarcinoma (see Supplementary Fig. S1). Prostatic adenocarcinoma had the highest median MCP for neutrophils, while the lowest median for this cell population was observed in colorectal adenocarcinoma. A comparison of immune and stromal cells infiltrates between wild-type and GSK-3β–mutated tumors was performed in colorectal carcinoma, endometrial cancer, and NSCLC, the three histologies with most frequent number of GSK-3β–mutated tumors. The comparison of immune infiltrate between colorectal cancer wild-type (n = 1,638) and GSK-3β–mutated tumors (n = 12) showed an increased CD8 T-cell population in tumors with GSK-3β mutations (P = 0.0049). There were no significant differences in immune infiltrate and stromal cell populations between endometrial cancers wild type (n = 648) and those with GSK-3β mutations (n = 14). Among NSCLC specimens, the only significant differences observed were lower infiltration of myeloid dendritic cells (P < 0.0001) and neutrophils (P < 0.0001) in tumors with GSK-3β mutations (n = 79) compared with wild-type tumors (n = 2,178).

GSK-3β–mutated tumors and PD-L1 expression

As GSK-3β phosphorylates nonglycosylated forms of PD-L1 leading to its degradation, this study sought to explore the impact of GSK-3β mutations on PD-L1 expression (17). Using CLSPOA database, we assessed differences in PD-L1 expression between GSK-3β–mutated tumors and GSK-3β wild-type tumors (Table 3). This analysis was performed on the subset of tumor specimens that had results available for both PD-L1 expression and GSK-3β mutation status. GSK-3β–mutated tumors were associated with a higher frequency of PD-L1 expression in colorectal cancer (8% vs. 4%, P = 0.03), endometrial cancer (12% vs. 7%, P = 0.05), melanoma (42% vs. 24%, P = 0.02), ovarian epithelial carcinoma (21% vs. 8%, P = 0.0001), and uterine sarcoma (40% vs. 10%, P = 0.002). Among these five histologies, 50 tumors had GSK-3β mutations and were positive for PD-L1 (>5% using SP-142 antibody), and the top mutated residues were H310Q (n = 5) and S215L (n = 3). The remaining assessed histologies producing nonsignificant results are shown in Supplementary Table S1.

Table 3.

Differences in PD-L1 expression between GSK-3β–mutated tumors and GSK-3β wild-type tumors.

HistologyFrequency of PD-L1 positive tumors (GSK-3β wild type)Frequency of PD-L1–positive tumors (GSK-3β mutant)PQ value
Colorectal adenocarcinoma 3.8% (355/9,437) 8.1% (8/99) 0.03 0.24 
Endometrial cancer 7.2% (399/5,541) 12.0% (15/125) 0.05 0.32 
Melanoma 24.3% (345/1,417) 41.9% (13/31) 0.02 0.24 
Ovarian epithelial carcinoma 7.6% (691/9,087) 21.4% (12/56) 0.0001 0.004 
Uterine sarcoma 9.7% (73/756) 40.0% (4/10) 0.002 0.03 
HistologyFrequency of PD-L1 positive tumors (GSK-3β wild type)Frequency of PD-L1–positive tumors (GSK-3β mutant)PQ value
Colorectal adenocarcinoma 3.8% (355/9,437) 8.1% (8/99) 0.03 0.24 
Endometrial cancer 7.2% (399/5,541) 12.0% (15/125) 0.05 0.32 
Melanoma 24.3% (345/1,417) 41.9% (13/31) 0.02 0.24 
Ovarian epithelial carcinoma 7.6% (691/9,087) 21.4% (12/56) 0.0001 0.004 
Uterine sarcoma 9.7% (73/756) 40.0% (4/10) 0.002 0.03 

Note: 38 total histologies were assessed, and those with significant results are shown in the table.

We also investigated the association between MMR deficiency and differences in PD-L1 expression between GSK-3β–mutated and GSK-3β wild-type tumors in a cohort of selected histologies for which microsatellite status information was available (colorectal adenocarcinoma, endometrial cancer, ovarian epithelial carcinoma). Considering the described association between MSI-H and higher expression of PD-L1 (38), we assessed whether microsatellite status influenced the higher frequency of PD-L1 expression among tumors with GSK-3β mutations. There was no significant difference in PD-L1 expression between GSK-3β wild-type and mutated tumors with MSI-H. In our microsatellite stable (MSS) cohort, we found a significantly greater number of PD-L1–positive GSK-3β–mutated tumors compared with PD-L1–positive GSK-3β wild-type tumors for colorectal adenocarcinoma (GSK-3β mutant: 7.0% vs. GSK-3β wild type: 2.5%, P = 0.03), endometrial cancer (13.2% vs. 5.7%, P = 0.02), and ovarian epithelial carcinoma (22% vs. 6.8%, P = 0.0005). These results suggest that the higher frequency of PD-L1 expression observed among colorectal, endometrial, and ovarian cancers was not driven by MMR deficiency (Supplementary Table S2).

We assessed the landscape of GSK-3β genomic alterations, a serine/threonine kinase involved in the pathogenesis of many cancers. Using over 100,000 tumor samples, we identified mutations in approximately 1% of tumors involving functionally relevant protein domains. We documented the most common histologies containing alterations and differential expression levels between tumor types. We also characterized the microenvironment profile of immune cell populations in GSK-3β–mutated tumors. Interestingly, in histologies with the highest GSK-3β alteration rates, we found differential expression in PD-L1 between GSK-3β–mutated tumors and GSK-3β wild-type tumors. As GSK-3β inhibitors are now in clinical trials (20), this work lays the foundation for advancing the understanding of the complex GSK-3β signaling network that interfaces with key oncologic pathways, such as NFκB and PI3K/AKT/mTOR.

Among the GSK-3β mutations described in the assessed cohorts, several were determined to be of potential functional relevance. cBioPortal data revealed that the majority of mutations were in the kinase domain. Among the three GSK-3β residues that form a binding pocket for the phosphate group of a primed substrate (32), two (R96 and R180) were among the most frequently mutated loci, and other mutated loci have been implicated in the regulation of protein activation or ATP binding (35, 39). While the majority of missense mutations were predicted to be either “deleterious” or “damaging,” the true functional impact of these alterations is unknown, warranting additional in vitro studies. The most frequent residue change found in the combined dataset was R396Q. As this residue is located outside the kinase domain and little is known about its functional relevance, future in silico modeling will investigate the role of this residue. In this case, a change of R to Q (arginine to glutamine) results in a change in the charge from positive to neutral at that position and could impact on the protein folding, interactions, or function.

It is plausible that specific mutations will have a distinct impact on the role of GSK-3β as a tumor suppressor or proto-oncogene depending on its downstream target and tumor type (2). For instance, a mutation that inactivates GSK-3β could lead to increased cytoplasmic levels of β-catenin that ultimately could promote tumor growth (32, 40). This indeed supports the role of GSK-3β as a tumor suppressor that may become pro-oncogenic in the presence of a certain mutation. Conversely, a mutation could compromise the ability of GSK-3β to bind to one of its substrates, such as NFκB. GSK-3β activates NFκB leading to chemoresistance, cancer progression, and metastasis (19). However, in the setting of GSK-3β mutation, lack of NFκB activation may actually lead to tumor suppression. Mutations may also impact specific histologies differently, as GSK-3β has been shown to act as a tumor suppressor in breast models (32, 40), while acting as a proto-oncogene in other malignancies, such as pancreatic and colorectal cancers (25). Taken together, these ideas further highlight the complexity of GSK-3β signaling.

GSK-3β mutations were frequently detected in gynecologic malignancies, with uterine endometrioid carcinoma, specifically, being the most frequently mutated histology, and uterine neoplasms, as a group, being mutated at a higher rate than other histologies. GSK-3β has been implicated in the transdifferentiation of endometrial carcinoma cells (41). In addition, GSK-3β inhibition reduced cellular proliferation and tumor growth in endometrial cancer cell lines and in vivo models (42). Future studies investigating the impact of these mutations in prognosis and clinical outcomes of endometrial cancer can help elucidate their role as potential therapeutic targets or biomarkers for treatment response.

Most frequently, we identified GSK-3β CNVs in cervical squamous cell carcinoma. GSK-3β has been involved in the carcinogenic effect of HPV16 in cervical cancer (43, 44). Additional squamous cell carcinomas, including lung and head and neck, frequently contained GSK-3β CNV. Notably, GSK-3β was previously identified as part of a proto-oncogenic network in head and neck squamous cell carcinoma (45) and has been implicated in oral and esophageal squamous cell carcinomas as well (46, 47).

In addition to studying the most frequently mutated histologies, we also assessed GSK-3β mRNA expression levels across different tumor types. Expression levels were highest in melanoma and squamous NSCLC and lowest in pancreatic and adenocarcinoma NSCLC. Importantly, a significant difference in expression was observed among studied cancer types and also specifically between squamous and adenocarcinoma NSCLC. This finding is noteworthy, as previous research has shown that GSK-3β overexpression leads to worse prognosis in NSCLC, with GSK-3β blockade reducing NSCLC cell proliferation (48). Our finding suggests that perhaps the squamous subtype of NSCLC may be a more appropriate target for GSK-3β inhibitors. In addition, our finding that the highest GSK-3β expression was observed in melanoma may also be relevant to both pathogenesis and treatment. Published data show that GSK-3β is implicated in melanoma cell migration and invasion, which are both prevented by GSK-3β blockade (49). Of note, a patient with a BRAF V600K–mutated melanoma metastatic to the brain, lung, bone and mediastinal lymph nodes achieved a complete response in an ongoing clinical trial with 9-ING-41, a small-molecule inhibitor of GSK-3β (20).

Given the role of GSK-3β in immune response, studying the tumor microenvironment in GSK-3β–mutated tumors may help elaborate upon the impact of mutations and potential downstream effects. The results revealed significant differences in immune cell populations across cancer types harboring GSK-3β mutations. While T-cell infiltration was not significantly different across cancer types, GSK-3β–mutated melanoma showed the highest counts for B cells, monocytic lineage cells, myeloid dendritic cells, and endothelial cells. This may be particularly relevant, as recent results have shown that tumor-associated B cells both sustain melanoma inflammation and may predict response to immune checkpoint inhibitors (50, 51).

Among the many signaling pathways regulated by GSK-3β, perhaps one of the most relevant to the current therapeutic landscape in oncology is the programmed death 1 (PD-1)/PD-L1 pathway. GSK-3β phosphorylates non-glycosylated forms of PD-L1, inducing its proteasomal degradation (17). Our results describe that GSK-3β mutations were associated with a significantly higher frequency of PD-L1–positive tumors among colorectal adenocarcinoma, endometrial cancers, melanoma, ovarian epithelial carcinoma, and uterine sarcoma as compared with GSK-3β wild-type tumors. Moreover, the majority of these histologies are among the top histologies containing a GSK-3β mutation or CNV. We hypothesize that GSK-3β genomic alterations may lead to inactivity of GSK-3β and downstream stabilization of tumor PD-L1, whereby tumors then evade the host immune system and further progress. If confirmed, this hypothesis could be relevant to the selection of patients with tumors carrying inactivating GSK-3β mutations for treatment with anti-PD-1/PD-L1 antibodies, and deserves further investigation. Conversely, if certain GSK-3β alterations are not present, GSK-3β inhibitors could block the degradation of PD-L1 and increase its expression and make the tumor sensitive to checkpoint inhibitors. In fact, preliminary results from experiments in prostate cancer cell lines showed that the GSK-3β inhibitor 9-ING-41 increased the expression of PD-L1 in DU145 cancer cells (52). This highlights the importance of future functional assessment of mutations and the potential therapeutic implications.

Our study has several limitations. First, by using a combined cohort from cBioPortal, nonuniform genotyping methodologies were utilized across these studies and histologic classifications were not standardized. However, this concern is mitigated by the standardization and uniformity in the larger CLSPOA dataset. Although predictions about functional relevance were made, the exact functional impact of the described GSK-3β genomic alterations remain to be discovered. Additional studies are also warranted to assess variable functions of certain mutations between different cancer types. In addition, limited clinical outcome data are currently available, and our future studies will be focused on the correlation between GSK-3β alterations/mRNA expression and survival and treatment sensitivity. Furthermore, mRNA expression may not correlate with GSK-3β protein expression, suggesting investigation at the protein levels in advanced malignancies is warranted. mRNA expression data were also not available for all tumor types, and our dataset did not allow for a comparison of mRNA expression in GSK-3β mutated versus wild-type tumors. Future studies are planned to address these important questions.

To conclude, our study is the first to comprehensively assess the landscape of GSK-3β genomic alterations. Our findings regarding the most commonly mutated loci, top mutated histologies, and differential expression levels between tumor types may help to unravel the understanding of the complex GSK-3β signaling network, which might carry histology blueprints. This may be especially true for histologies that most frequently contain GSK-3β alterations, including uterine neoplasms, melanoma, and nonmelanoma skin cancers, and squamous histologies. Furthermore, our findings that GSK-3β–mutated tumors have significantly varied immune cell populations and higher PD-L1 expression for select histologies support the potential role of GSK-3β as an immune regulator. As mutations were frequently found in the kinase domain and substrate binding pocket, ongoing work is focused on defining the functional impact of such mutations. Future investigations will focus on refining GSK-3β transcriptional signatures and exploring the impact of mutations in clinical outcomes.

B.A. Borden reports grants from Brown University (received a Summer Assistantship Grant from the Warren Alpert Medical School of Brown University) during the conduct of the study. J. Xiu reports other from Caris Life Sciences (salary) outside the submitted work. B.A. Weinberg reports personal fees from Lilly, Taiho, Bayer, and Sirtex outside the submitted work. A.M. VanderWalde reports personal fees and nonfinancial support from Caris Life Sciences during the conduct of the study; personal fees from Bristol Myers Squibb, Elsevier, Compugen, Roche/Genentech, and Immunocore; nonfinancial support from AstraZeneca; and grants from Amgen outside the submitted work. W.M. Korn reports other from Caris Life Sciences and Caris Life Sciences; and personal fees from Merck outside the submitted work. A.P. Mazar reports other from Actuate Therapeutics, Inc. (founder and shareholder in company that is developing GSK-3 inhibitors) outside the submitted work. F.J. Giles is consultant for Actuate Therapeutics. B.A. Carneiro reports other from Actuate Therapeutics (institutional research support for clinical trial), AstraZeneca (institutional research support for clinical trial), Bayer (institutional research support for clinical trial), Pfizer (institutional research support for clinical trial), MedImmune (institutional research support for clinical trial), and Astellas (institutional research support for clinical trial) outside the submitted work. No disclosures were reported by the other authors.

B.A. Borden: Conceptualization, data curation, formal analysis, methodology, writing-original draft, writing-review and editing. Y. Baca: Data curation, formal analysis, methodology, writing-original draft, writing-review and editing. J. Xiu: Data curation, formal analysis, methodology, writing-original draft, writing-review and editing. F. Tavora: Formal analysis. I. Winer: Data curation, writing-review and editing. B.A. Weinberg: Data curation, writing-review and editing. A.M. VanderWalde: Data curation, writing-review and editing. S. Darabi: Data curation, writing-review and editing. W.M. Korn: Data curation, writing-review and editing. A.P. Mazar: Formal analysis, writing-review and editing. F.J. Giles: Formal analysis, writing-review and editing. L. Crawford: Formal analysis, writing-review and editing. H. Safran: Formal analysis. W. El-Deiry: Conceptualization, formal analysis, methodology, writing-review and editing. B.A. Carneiro: Conceptualization, data curation, formal analysis, methodology, writing-original draft, writing-review and editing.

B.A. Borden received a grant from The Warren Alpert Medical School Summer Assistantship Program.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Domoto
T
,
Pyko
IV
,
Furuta
T
,
Miyashita
K
,
Uehara
M
,
Shimasaki
T
, et al
Glycogen synthase kinase-3β is a pivotal mediator of cancer invasion and resistance to therapy
.
Cancer Sci
2016
;
107
:
1363
72
.
2.
Walz
A
,
Ugolkov
A
,
Chandra
S
,
Kozikowski
A
,
Carneiro
BA
,
O'Halloran
TV
, et al
Molecular pathways: revisiting glycogen synthase kinase-3β as a target for the treatment of cancer
.
Clin Cancer Res
2017
;
23
:
1891
7
.
3.
Ugolkov
A
,
Gaisina
I
,
Zhang
J-S
,
Billadeau
DD
,
White
K
,
Kozikowski
A
, et al
GSK-3 inhibition overcomes chemoresistance in human breast cancer
.
Cancer Lett
2016
;
380
:
384
92
.
4.
Vincent
EE
,
Elder
DJE
,
O′Flaherty
L
,
Pardo
OE
,
Dzien
P
,
Phillips
L
, et al
Glycogen synthase kinase 3 protein kinase activity is frequently elevated in human non-small cell lung carcinoma and supports tumour cell proliferation
.
PLoS One
2014
;
9
:
e114725
.
5.
Ougolkov
AV
,
Fernandez-Zapico
ME
,
Bilim
VN
,
Smyrk
TC
,
Chari
ST
,
Billadeau
DD
. 
Aberrant nuclear accumulation of glycogen synthase kinase-3beta in human pancreatic cancer: association with kinase activity and tumor dedifferentiation
.
Clin Cancer Res
2006
;
12
:
5074
81
.
6.
Shakoori
A
,
Ougolkov
A
,
Yu
ZW
,
Zhang
B
,
Modarressi
MH
,
Billadeau
DD
, et al
Deregulated GSK3beta activity in colorectal cancer: its association with tumor cell survival and proliferation
.
Biochem Biophys Res Commun
2005
;
334
:
1365
73
.
7.
Song
EY
,
Palladinetti
P
,
Klamer
G
,
Ko
K-H
,
Lindeman
R
,
O’Brien
TA
, et al
Glycogen synthase kinase–3beta inhibitors suppress leukemia cell growth
.
Exp Hematol
2010
;
38
:
908
21e1
.
8.
Zhu
Q
,
Yang
J
,
Han
S
,
Liu
J
,
Holzbeierlein
J
,
Thrasher
JB
, et al
Suppression of glycogen synthase kinase 3 activity reduces tumor growth of prostate cancer in vivo
.
Prostate
2011
;
71
:
835
45
.
9.
Hermida
MA
,
Dinesh Kumar
J
,
Leslie
NR
. 
GSK3 and its interactions with the PI3K/AKT/mTOR signalling network
.
Adv Biol Regul
2017
;
65
:
5
15
.
10.
Beurel
E
,
Grieco
SF
,
Jope
RS
. 
Glycogen synthase kinase-3 (GSK3): regulation, actions, and diseases
.
Pharmacol Ther
2015
;
148
:
114
31
.
11.
Wang
Z
,
Smith
KS
,
Murphy
M
,
Piloto
O
,
Somervaille
TC
,
Cleary
ML
. 
Glycogen synthase kinase 3 in MLL leukaemia maintenance and targeted therapy
.
Nature
2008
;
455
:
1205
9
.
12.
Beurel
E
,
Jope
RS
. 
The paradoxical pro- and anti-apoptotic actions of GSK3 in the intrinsic and extrinsic apoptosis signaling pathways
.
Prog Neurobiol
2006
;
79
:
173
89
.
13.
Kroon
J
,
in 't Veld
LS
,
Buijs
JT
,
Cheung
H
,
van der Horst
G
,
van der Pluijm
G
. 
Glycogen synthase kinase-3beta inhibition depletes the population of prostate cancer stem/progenitor-like cells and attenuates metastatic growth
.
Oncotarget
2014
;
5
:
8986
94
.
14.
Beurel
E
,
Michalek
SM
,
Jope
RS
. 
Innate and adaptive immune responses regulated by glycogen synthase kinase-3 (GSK3)
.
Trends Immunol
2010
;
31
:
24
31
.
15.
Beurel
E
. 
Regulation of inflammation and T cells by glycogen synthase kinase-3: links to mood disorders
.
Neuroimmunomodulation
2014
;
21
:
140
4
.
16.
Jellusova
J
,
Cato
MH
,
Apgar
JR
,
Ramezani-Rad
P
,
Leung
CR
,
Chen
C
, et al
Gsk3 is a metabolic checkpoint regulator in B cells
.
Nat Immunol
2017
;
18
:
303
12
.
17.
Li
C-W
,
Lim
S-Oe
,
Xia
W
,
Lee
H-H
,
Chan
Li-C
,
Kuo
C-W
, et al
Glycosylation and stabilization of programmed death ligand-1 suppresses T-cell activity
.
Nat Commun
2016
;
7
:
12632
.
18.
Hsu
JM
,
Li
CW
,
Lai
YJ
,
Hung
MC
. 
Posttranslational modifications of PD-L1 and their applications in cancer therapy
.
Cancer Res
2018
;
78
:
6349
53
.
19.
Hoeflich
KP
,
Luo
J
,
Rubie
EA
,
Tsao
MS
,
Jin
O
,
Woodgett
JR
. 
Requirement for glycogen synthase kinase-3beta in cell survival and NF-kappaB activation
.
Nature
2000
;
406
:
86
90
.
20.
Carneiro
BA
,
Cavalcante
L
,
Bastos
BR
,
Powell
SF
,
Ma
WW
,
Sahebjam
S
, et al
Phase I study of 9-ING-41, a small molecule selective glycogen synthase kinase-3 beta (GSK-3B) inhibitor, as a single agent and combined with chemotherapy in patients with refractory tumors
.
J Clin Oncol
38: 15s, 
2020
(suppl; abstr 3507).
21.
Ito
H
,
Makiyama
K
,
Kawahara
T
,
Osaka
K
,
Izumi
K
,
Yokomizo
Y
, et al
The impact of gender difference on operative time in laparoscopic partial nephrectomy for T1 renal tumor and the utility of retroperitoneal fat thickness as a predictor of operative time
.
BMC Cancer
2016
;
16
:
944
.
22.
Ugolkov
A
,
Qiang
W
,
Bondarenko
G
,
Procissi
D
,
Gaisina
I
,
James
CD
, et al
Combination treatment with the GSK-3 inhibitor 9-ING-41 and CCNU cures orthotopic chemoresistant glioblastoma in patient-derived xenograft models
.
Transl Oncol
2017
;
10
:
669
78
.
23.
Ding
L
,
Liou
GY
,
Schmitt
DM
,
Storz
P
,
Zhang
JS
,
Billadeau
DD
. 
Glycogen synthase kinase-3beta ablation limits pancreatitis-induced acinar-to-ductal metaplasia
.
J Pathol
2017
;
243
:
65
77
.
24.
Ding
L
,
Madamsetty
VS
,
Kiers
S
,
Alekhina
O
,
Ugolkov
A
,
Dube
J
, et al
Glycogen synthase kinase-3 inhibition sensitizes pancreatic cancer cells to chemotherapy by abrogating the TopBP1/ATR-mediated DNA damage response
.
Clin Cancer Res
2019
;
25
:
6452
62
.
25.
Sahin
I
,
Eturi
A
,
De Souza
A
,
Pamarthy
S
,
Tavora
F
,
Giles
FJ
, et al
Glycogen synthase kinase-3 beta inhibitors as novel cancer treatments and modulators of antitumor immune responses
.
Cancer Biol Ther
2019
;
20
:
1047
56
.
26.
Gao
J
,
Aksoy
BA
,
Dogrusoz
U
,
Dresdner
G
,
Gross
B
,
Sumer
SO
, et al
Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal
.
Sci Signal
2013
;
6
:
pl1
.
27.
Cerami
E
,
Gao
J
,
Dogrusoz
U
,
Gross
BE
,
Sumer
SO
,
Aksoy
BA
, et al
The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data
.
Cancer Discov
2012
;
2
:
401
4
.
28.
Adzhubei
IA
,
Schmidt
S
,
Peshkin
L
,
Ramensky
VE
,
Gerasimova
A
,
Bork
P
, et al
A method and server for predicting damaging missense mutations
.
Nat Methods
2010
;
7
:
248
9
.
29.
Ng
PC
,
Henikoff
S
. 
SIFT: Predicting amino acid changes that affect protein function
.
Nucleic Acids Res
2003
;
31
:
3812
4
.
30.
Kopanos
C
,
Tsiolkas
V
,
Kouris
A
,
Chapple
CE
,
Albarca Aguilera
M
,
Meyer
R
, et al
VarSome: the human genomic variant search engine
.
Bioinformatics
2019
;
35
:
1978
80
.
31.
Becht
E
,
Giraldo
NA
,
Lacroix
L
,
Buttard
B
,
Elarouci
N
,
Petitprez
F
, et al
Estimating the population abundance of tissue-infiltrating immune and stromal cell populations using gene expression
.
Genome Biol
2016
;
17
:
218
.
32.
Patel
P
,
Woodget
JR
. 
Glycogen synthase kinase 3: a kinase for all pathways?
Curr Top Dev Biol
2017
;
123
:
277
302
.
33.
Lu
S
,
Jiang
Y
,
Lv
J
,
Zou
J
,
Wu
T
. 
Mechanism of kinase inactivation and nonbinding of FRATide to GSK3beta due to K85M mutation: molecular dynamics simulation and normal mode analysis
.
Biopolymers
2011
;
95
:
669
81
.
34.
UniProtKB - P49841 (GSK3B_HUMAN)
; 
2020
.
Available from:
https://www.uniprot.org/uniprot/P49841.
35.
Bertrand
JA
,
Thieffine
S
,
Vulpetti
A
,
Cristiani
C
,
Valsasina
B
,
Knapp
S
, et al
Structural characterization of the GSK-3beta active site using selective and non-selective ATP-mimetic inhibitors
.
J Mol Biol
2003
;
333
:
393
407
.
36.
Dajani
R
,
Fraser
E
,
Roe
SM
, et al
Structural basis for recruitment of glycogen synthase kinase 3beta to the axin-APC scaffold complex
.
EMBO J
2003
;
22
:
494
501
.
37.
Pawloski
JR
,
Randolph
B
,
Bajio
P
. 
Foundation-one-heme next generation sequencing of hairy cell leukemia variant lymphocytes
.
Blood
2016
;
128
:
5289
.
38.
Kim
ST
,
Klempner
SJ
,
Park
SeH
,
Park
JOh
,
Park
YS
,
Lim
HoY
, et al
Correlating programmed death ligand 1 (PD-L1) expression, mismatch repair deficiency, and outcomes across tumor types: implications for immunotherapy
.
Oncotarget
2017
;
8
:
77415
23
.
39.
Palomo
V
,
Soteras
I
,
Perez
DI
,
Perez
C
,
Gil
C
,
Eugenia
N
, et al
Exploring the binding sites of glycogen synthase kinase 3. Identification and characterization of allosteric modulation cavities.
J Med Chem
2011
;
54
:
8461
70
.
40.
Dembowy
J
,
Adissu
HA
,
Liu
JC
,
Zacksenhaus
E
,
Woodgett
JR
. 
Effect of glycogen synthase kinase-3 inactivation on mouse mammary gland development and oncogenesis
.
Oncogene
2015
;
34
:
3514
26
.
41.
Saegusa
M
,
Hashimura
M
,
Kuwata
T
,
Hamano
M
,
Okayasu
I
. 
Crosstalk between NF-kappaB/p65 and beta-catenin/TCF4/p300 signalling pathways through alterations in GSK-3beta expression during trans-differentiation of endometrial carcinoma cells
.
J Pathol
2007
;
213
:
35
45
.
42.
Yin
Y
,
Kizer
N
,
Thaker
P
,
Chiappinelli
K
,
Trinkaus
K
,
Goodfellow
P
, et al
Glycogen synthase kinase 3beta inhibition as a therapeutic approach in the treatment of endometrial cancer
.
Int J Mol Sci
2013
;
14
:
16617
37
.
43.
Ma
C
,
Zeng
C
,
Jin
L
,
Yang
Y
,
Li
P
,
Chen
L
, et al
GSK3beta mediates the carcinogenic effect of HPV16 in cervical cancer
.
Sci Rep
2015
;
5
:
16555
.
44.
Rath
G
,
Jawanjal
P
,
Salhan
S
,
Nalliah
M
,
Dhawan
I
. 
Clinical significance of inactivated glycogen synthase kinase 3beta in HPV-associated cervical cancer: relationship with Wnt/beta-catenin pathway activation
.
Am J Reprod Immunol
2015
;
73
:
460
78
.
45.
Georgy
SR
,
Cangkrama
M
,
Srivastava
S
,
Partridge
D
,
Auden
A
,
Dworkin
S
, et al
Identification of a novel proto-oncogenic network in head and neck squamous cell carcinoma
.
J Natl Cancer Inst
2015
;
107
:
djv152
.
46.
Gao
S
,
Li
S
,
Duan
X
,
Gu
Z
,
Ma
Z
,
Yuan
X
, et al
Inhibition of glycogen synthase kinase 3 beta (GSK3beta) suppresses the progression of esophageal squamous cell carcinoma by modifying STAT3 activity
.
Mol Carcinog
2017
;
56
:
2301
16
.
47.
Pramanik
KK
,
Nagini
S
,
Singh
AK
,
Mishra
P
,
Kashyap
T
,
Nath
N
, et al
Glycogen synthase kinase-3beta mediated regulation of matrix metalloproteinase-9 and its involvement in oral squamous cell carcinoma progression and invasion
.
Cell Oncol
2018
;
41
:
47
60
.
48.
Zeng
J
,
Liu
D
,
Qiu
Z
,
Huang
Yi
,
Chen
B
,
Wang
L
, et al
GSK3beta overexpression indicates poor prognosis and its inhibition reduces cell proliferation and survival of non-small cell lung cancer cells
.
PLoS One
2014
;
9
:
e91231
.
49.
John
JK
,
Paraiso
KHT
,
Rebecca
VW
,
Cantini
LP
,
Abel
EV
,
Pagano
N
, et al
GSK3beta inhibition blocks melanoma cell/host interactions by downregulating N-cadherin expression and decreasing FAK phosphorylation
.
J Invest Dermatol
2012
;
132
:
2818
27
.
50.
Helmink
BA
,
Reddy
SM
,
Gao
J
,
Zhang
S
,
Basar
R
,
Thakur
R
, et al
B cells and tertiary lymphoid structures promote immunotherapy response
.
Nature
2020
;
577
:
549
55
.
51.
Griss
J
,
Bauer
W
,
Wagner
C
,
Simon
M
,
Chen
M
,
Grabmeier-Pfistershammer
K
, et al
B cells sustain inflammation and predict response to immune checkpoint blockade in human melanoma
.
Nat Commun
2019
;
10
:
4186
.
52.
Tavora
F
,
Lotan
T
,
Alves
M
, et al
Glycogen synthase kinase-3β expression in prostate cancer (PCa) correlates with aggressive pathological features and its blockade with 9-ING-41 inhibits viability of PCa cell lines [abstract]
.
In:
Proceedings of the Annual Meeting of the American Association for Cancer Research 2020
; 
2020 Apr 27–28 and Jun 22–24
.
Philadelphia (PA)
:
AACR
;
Cancer Res 2020
;
80
(
16 Suppl
):
Abstract nr 2959
.