Exploration of the Combination of PLK1 Inhibition with Immunotherapy in Cancer Treatment

Background PLK1 overexpression is oncogenic and is associated with poor prognosis in various cancers. However, the current PLK1 inhibitors have achieved limited clinical successes. On the other hand, although immunotherapies are demonstrating efficacy in treating many refractory cancers, a substantial number of patients do not respond to such therapies. The potential of combining PLK1 inhibition with immunotherapy for cancer treatment is worthy of exploration. Methods We analyzed the associations of PLK1 expression with tumor immunity in 33 different cancer types. Moreover, we analyzed the associations of the drug sensitivities of PLK1 inhibitors with tumor immunity in cancer cell lines. Furthermore, we explored the mechanism underlying the significant associations between PLK1 and tumor immunity. Finally, we experimentally verified some findings from bioinformatics analysis. Results The cancers with higher PLK1 expression levels tended to have lower immune activities, such as lower HLA expression and decreased B cells, NK cells and tumor-infiltrating lymphocytes infiltration. On the other side, elevated tumor immunity likely increased the sensitivity of cancer cells to PLK1 inhibitors. The main mechanism underlying the associations between PLK1 and tumor immunity may lie in the aberrant cell cycle and p53 pathways in cancers. Conclusions Our findings implicate that the PLK1 inhibition and immunotherapy combination may achieve a synergistic antitumor efficacy.


Introduction
PLK1 (Polo-like kinase 1) is a member of the Polo-like kinase family [1], which plays an important role in cell cycle regulation [2]. The role PLK1 plays in regulating cell cycle is diverse which includes controlling mitotic entry, harmonizing centrosome and cell cycles, regulating chromosome segregation, and mediating cytokinesis and meiosis [2]. Thus, the malfunction of PLK1 would result in cell cycle aberration that often incites cell proliferation. In fact, a substantial number of studies have revealed that PLK1 was overexpressed in a wide variety of cancers, and its overexpression correlated with unfavorable prognosis of cancer patients [3]. Hence, the inhibition of PLK1 has been suggested as a potential strategy for cancer therapy [4]. A number of PLK1 inhibitors have been explored in laboratory or clinical studies such as BI2536, volasertib, GSK461364, rigosertib, poloxin, poloxin-2, and RO3280 [3]. However, none of these exploratory agents have been used in clinics thus far [5].
On the other hand, recently cancer immunotherapy is demonstrating astonishing successes in treating various cancers [6,7]. Particularly, the blockade of immune checkpoints CTLA4 (cytotoxic T-lymphocyte-associated protein 4), PD1 (programmed cell death protein 1), and PD-L1 (programmed cell death 1 ligand) has clinical successes in various cancers including melanoma, lung cancer, renal cell cancer, bladder cancer, head and neck cancer, Hodgkin's lymphoma, and the cancers with MSI (microsatellite instability) or DNA mismatch-repair deficiency [6]. Another notable cancer immunotherapeutic strategy is the chimeric antigen receptor (CAR) T cell therapy that has been used to treat refractory leukemia and lymphoma successfully [7]. Despite these remarkable achievements of cancer immunotherapy, a substantial proportion of patients had limited or no response to such therapies [8]. To predict the patients responsive to cancer immunotherapy, some biomarkers have been explored such as tumor mutation burden (TMB) [9,10], neoantigens [11], MSI [12], and PD-L1 expression [13]. In addition, to improve the efficacy of cancer immunotherapy, the combination of immunotherapy with chemotherapy, radiotherapy, or targeted therapies has been explored [14]. For example, a recent study demonstrated that the combination of cyclin-dependent kinases 4 and 6 (CDK4/6) inhibitors with immunotherapy could promote antitumor immunity [15].
In this study, to explore the potential of combining PLK1 inhibitors with immunotherapy in treating cancers, we analyzed the associations of PLK expression with immune cell infiltration and immune activities in 33 different cancer types based on the Cancer Genome Atlas (TCGA) data (https://cancergenome.nih.gov/). Moreover, we analyzed the associations of the drug sensitivities of PLK1 inhibitors with immune cell infiltration and immune activities in cancer cell lines (CCLs) based on the Genomics of Drug Sensitivity in Cancer (GDSC) project data (http://www.cancerrxgene.org/). Furthermore, we explored the potential mechanisms that underlie the significant associations between PLK expression and tumor immunity.

Methods
. . Datasets. The TCGA data for gene expression profiles (RNA-Seq, Level 3) and gene somatic mutations (Level 3) were downloaded from the genomic data commons data portal (https://portal.gdc.cancer.gov/). The 33 TCGA cancer types analyzed are shown in Table 1. The GDSC data for gene expression profiles (Affymetrix Human Genome U219 array) and drug sensitivities (IC50) were downloaded from the Wellcome Sanger Institute website: Journal of Oncology 3 https://www.cancerrxgene.org/downloads. We analyzed the enrichment levels of 6 immune cell types and functions in cancers including B cells, natural killer (NK) cells, tumorinfiltrating lymphocytes (TILs), human leukocyte antigen (HLA), regulatory T (Treg) cells, and cancer-testis antigens (CTAs) based on the expression profiles of their gene signatures. These gene signatures are shown in the Supplementary  Table S1.
. . Evaluation of the Activity of an Immune Cell Type or Function in Cancers. We quantified the activity (or enrichment levels) of an immune cell type or function in a cancer sample using the single-sample gene-set enrichment analysis (ssGSEA) score [16,17]. The gene-set is the set of gene signatures of the immune cell type or function. The higher the ssGSEA score, the higher the activity of the immune cell type or function. In addition, we assessed the levels of immune infiltration in cancers by the ESTIMATE algorithm [18]. ESTIMATE output immune scores quantify the immune infiltration levels in cancers based on gene expression profiles data.
. . Reverse Transcription Quantitative PCR (qPCR) Analysis. BI2536 were purchased from Selleck. Cells were harvested after being treated with BI2536 (1 M, 48h). The total RNA was isolated by Trizol (Invitrogen, USA) and was reversely transcribed into cDNA by the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher, USA). Primer sequences used for qPCR were presented in Supplementary Table S2. Primers were diluted in nuclease-free water with the real-time PCR (RT-PCR) Master Mix (SYBR Green) (TOYOBO Co., LTD, JAPAN). Relative copy number was determined by calculating the fold-change difference in the gene of interest relative to -actin. The qPCR was performed on an ABI 7500 FAST and Applied Biosystems StepOnePlus RT-PCR machine.
. . Flow Cytometry. Cells were harvested after treatments for 48 hours by trypsinization and were washed with PBS. Cells were resuspended in labeling buffer (PBS supplemented with 10% FBS and 1% NaN3) to a final concentration of 5×105 per ml and were stained with W6/32 monoclonal antibody (1:20,eBIOSCIENCE: 12-9983-42) at 37 ∘ C without lights for 30 minutes. Cells were then washed by PBS for flow cytometric analysis using a LSRII 4-laser flow cytometer (Becton Dickinson, USA). The results were analyzed and MFI calculated by FlowJo.
. . Statistical Analyses. We calculated the correlation between the PLK expression levels and the expression levels of another gene using the Pearson method, and the correlations between the PLK expression levels and the other variables including the enrichment levels (ssGSEA scores) of a gene-set, tumor mutation counts, and drug sensitivities (IC50) using the Spearman method. In comparisons of TP mutation rates between the cancers with higher PLK expression levels (upper third) and the cancers with lower PLK expression levels (lower third), we used the Fisher's exact test. We adjusted for multiple tests using the false discovery rate (FDR) calculated by the Benjamini and Hochberg (BH) method [19]. The threshold of FDR < 0.1 was used to identify the statistical significance. All the computational and statistical analyses were implemented by R (https://www.rproject.org/). The experimental data were analyzed by Prism 5.0 software (GraphPad) and were presented as mean ± SD. The t test P < 0.05 was considered statistically significant.

. . PLK Expression Likely Correlates with Depressed Immune
Cell Infiltration and Immune Activities in Cancer. We found that the PLK expression levels were negatively associated with immune scores in 10 cancer types (LUSC, TGCT (Figure 1(c)). In addition, in 6 cancer types (TGCT, PRAD, CESC, LIHC, KICH, and STAD), the upregulation of PLK was associated with higher levels of NK cell infiltration, and in 1 cancer type (LUAD), we observed an opposite trend (Spearman correlation, FDR<0.1) (Figure 1(d)). Furthermore, we associated the PLK expression levels with the enrichment levels of TILs in cancers. We found that the PLK expression levels negatively correlated with the enrichment levels of TILs     tends to inhibit immune cell infiltration and antitumor immunity in a number of cancer types.
. . PLK Expression Likely Correlates with Depressed HLA Activity in Cancer. HLA genes encode the MHC (major histocompatibility complex) proteins that are important in the tumor immune regulation [20]. We found that the PLK  (Figure 2(b)). Taken together, these results suggest that the PLK expression likely inhibits the HLA activity in cancer. The neoantigens yielded by gene mutations are associated with antitumor immunity [11]. We found that the tumors with higher PLK expression levels had significantly higher total somatic mutation counts than the tumors with lower PLK expression levels in TCGA (Spearman correlation, R=0.46, P=2.57 * 10 −214 ) (Figure 2(c)). Moreover, the tumors more highly expressing PLK had significantly more mutations yielding predicted HLA-binding peptides [21] than the tumors more lowly expressing PLK (Spearman correlation, R=0.43, P=1.03 * 10 −186 ) (Figure 2(d)). It suggests that although the PLK upregulation correlates with higher TMB and more neoantigens, it inhibits antitumor immune response by repressing the HLA activity. [22]. We found that high PLK expression levels were associated with depressed Treg cell enrichment levels in 16 (Figure 3). The GDSC data analysis showed that high PLK expression levels were associated with decreased Treg cell enrichment levels in cancer cell lines (Spearman correlation, R=-0.13, P=3.29 * 10 −5 ) (Figure 3). Altogether, these data suggest that the PLK expression is negatively associated with the Treg cell activity in a wide range of cancers.

. . PLK Expression Likely Positively Correlates with Expression of Cancer-Testis Antigens in Cancer.
CTAs are the immunogenic proteins that are aberrantly activated in many cancers [23]. Strikingly, we found that higher PLK expression levels were significantly associated with higher CTA enrichment levels in 31 of the 33 cancer types (Spearman correlation, FDR<0.1) (Figure 4(a)). Markedly, the CTA genes ATAD , CEP , FANCA, KIF C, NUF , OIP , and PBK had significantly positive expression correlations with the PLK expression in 30 cancer types (Pearson correlation, FDR<0.1) (Figure 4(b)). Moreover, 166 (74%) of the 223 CTA genes showed positive expression correlations with the PLK expression in LIHC, and 145 (65%) CTA genes did in KIRC. Furthermore, the GDSC data analysis showed that the PLK expression levels were positively associated with the CTA enrichment levels in cancer cell lines (Spearman correlation, R=0.22, P=4.09 * 10 −12 ) (Figure 4(a)). Altogether, these data suggest that higher PLK expression is associated with higher CTA presentation. ATAD2 . . PLK Inhibits Antitumor Immunity via the Cell Cycle Regulation. PLK1 is one of the essential regulators of cell cycle progression [24]. As expected, the TCGA data analysis showed that the PLK expression levels strongly correlated with the cell cycle activity in a positive direction in all 33 cancer types (Supplementary Table S3). Furthermore, our analysis showed that the high cell cycle activity tended to inhibit antitumor immunity. For example, the cell cycle activity negatively correlated with the TILs enrichment in 20 cancer types versus in 3 cancer types positively correlating  Immune infiltration degree (ssGSEA score or immune score) Figure 5: Cancer immune activities positively correlate the sensitivity of cancer cells to PLK inhibitors (GW and BI-). ssGSEA: the single-sample gene-set enrichment analysis.

. . Elevated Immune Activities Tends to Enhance the Sen
immune score HLA B cell Spearman correlation (cell cycle activity vs. immune activity) Negative Positive Not significant with the TILs enrichment (Spearman correlation, FDR<0.1) ( Figure 6). The cell cycle activity negatively correlated with the HLA enrichment in 21 cancer types, while only in 1 cancer type showed a positive correlation ( Figure 6). Moreover, the cell cycle activity negatively correlated the B cell enrichment in 11 cancer types versus in 5 cancer types positively correlating with the B cell enrichment ( Figure 6). Interestingly, the cell cycle showed a positive correlation with the CTA enrichment in 30 of the 33 cancer types (Spearman correlation, FDR<0.1) (Supplementary Table S4). Furthermore, we found that in 20 cancer types the cell cycle activity was negatively associated with the immune score compared to in 4 cancer types the cell cycle activity being positively associated with the immune score ( Figure 6). Altogether, these results suggest that the PLK upregulation inhibits antitumor immunity via enhancing the cell cycle activity in cancer. This is in line with a recent study showing that the cell cycle inhibition promoted antitumor immunity [15].

Discussion
PLK1 is a master regulator of cell cycle, and its overexpression is oncogenic in various cancer types. Thus, targeting PLK1 could be promising in treating a wide range of malignancies. However, the current PLK1 inhibitors have achieved very limited clinical successes. On the other hand, although immunotherapies are achieving rapid clinical successes in treating many refractory cancers, a considerable number of patients do not respond to such therapies. To improve the clinical efficacy of both therapies, the combination of PLK1 inhibition and immunotherapy merits consideration.  To explore the possibility of combining both therapies, we analyzed the associations between PLK expression and tumor immunity in various different cancer types. Our bioinformatics analyses showed that PLK expression tended to inhibit antitumor immunity as the cancers with higher PLK expression levels often had lower HLA expression levels and TILs infiltration. Moreover, the in vitro experiment verified that the PLK1 inhibition significantly increased the expression of HLA molecules in various cancers. A main mechanism by which PLK1 inhibits antitumor immunity lies in that the PLK1 upregulation activates the cell cycle which may decrease tumor immunogenicity (Figure 8(a)). Besides, we found that the cancers with higher PLK expression levels had significantly higher frequency of TP mutations than the cancers with lower PLK expression levels in 12 cancer types (Fisher's exact test, FDR<0.05) (Figure 8(b)), suggesting that the PLK upregulation positively correlates with the prevalence of TP mutations. Hence, the higher TP mutation rates in the cancers with higher PLK expression levels may also be partly responsible for the depressed tumor immunity in these cancers (Figure 8(a)) since a prior study has demonstrated that wildtype p53 could promote tumor immunity [25].
The correlations of PLK expression with the immune signature could be affected by other factors such as patient age, gender, tumor stage, and grade. We re-analyzed the correlations of PLK expression with the immune signature (immune score) under the stratification of patients based on age (<60 and ≥60 years old), gender (male and female), stage (early stage (Stage I-II) and late stage (Stage III-IV)), and grade (low-grade (G1-2) and high-grade (G3-4)), respectively. We did not observe marked changes of the statistical correlations when these covariates were considered (Supplementary Table S5). In addition, we performed the multiple linear regression analysis of the correlations between PLK expression and the immune signature by adding the covariate "age". Our results showed that the correlations between PLK expression and the immune signature were unlikely affected by the variable "age" (Supplementary Table  S6).
It is rational to anticipate that the PLK1 inhibition and immunotherapy combination may improve the antitumor efficacy. First, PLK1 inhibition is capable of boosting  tumor antigen presentation and antitumor immune infiltration, which can be further augmented by the addition of immunotherapy (Figure 8(c)). Second, cancer immunotherapy may enhance tumor immunogenicity, which in turn increases the sensitivity of cancer cells to PLK1 inhibitors (Figure 8(c)). Hence, the PLK1 inhibition and immunotherapy combination could be promising in cancer treatment, although it needs to be proved by further experimental and clinical validations.

Conclusions
PLK1 likely inhibits antitumor immunity, and the elevated tumor immunity may enhance the sensitivity of cancer cells to PLK1 inhibitors. It implicates that the combination of PLK1 inhibition and immunotherapy may achieve a synergistic antitumor efficacy.