Prostate cancer does not appear to respond to immune checkpoint therapies where T-cell infiltration may be a key limiting factor. Here, we report evidence that ablating the growth regulatory kinase Erk5 can increase T-cell infiltration in an established Pten-deficient mouse model of human prostate cancer. Mice that were doubly mutant in prostate tissue for Pten and Erk5 (prostate DKO) exhibited a markedly increased median survival with reduced tumor size and proliferation compared with control Pten-mutant mice, the latter of which exhibited increased Erk5 mRNA expression. A comparative transcriptomic analysis revealed upregulation in prostate DKO mice of the chemokines Ccl5 and Cxcl10, two potent chemoattractants for T lymphocytes. Consistent with this effect, we observed a relative increase in a predominantly CD4+ T-cell infiltrate in the prostate epithelial and stroma of tumors from DKO mice. Collectively, our results offer a preclinical proof of concept for ERK5 as a target to enhance T-cell infiltrates in prostate cancer, with possible implications for leveraging immune therapy in this disease. Cancer Res; 77(12); 3158–68. ©2017 AACR.

Prostate cancer is a significant public health problem. In males worldwide, prostate cancer is second in incidence to skin cancer, and it is the second most common cause of cancer-related deaths in men after lung cancer in the United States and the United Kingdom (1, 2). Recent progress has been made in the treatment of prostate cancer, through the development of novel inhibitors of the androgen receptor pathway and chemotherapy, which have led to improved patient survival. However, advanced prostate cancer remains incurable, indicating a clear need to better understand the molecular events associated with disease progression to facilitate the development of novel therapies.

ERK5 (MAPK7) is a member of the mitogen-activated protein kinase (MAPK) family. MAPKs function to regulate and integrate cellular signaling cascades in order to control multiple intracellular events, including cell proliferation, differentiation, migration, and survival. ERK5 is activated in response to extracellular stimuli, including mitogens and cellular stress, via sequential activation of MEKK2/3 and then MEK5, which is the sole known MAPK kinase (MAP2K) of ERK5. Phosphorylation of ERK5 by MEK5 in turn causes translocation of ERK5 to the nucleus, where it can regulate transcription of its target genes. The uniquely large C-terminus of ERK5 also possesses a transcriptional activation domain and contains the nuclear localization and export signals required for dynamic shuttling of ERK5. Despite its unique domain, it is noteworthy that the N-terminus of ERK5 shares ∼66% sequence homology with ERK1/2 (3–6).

The ERK5 pathway has a critical role in development, as demonstrated by the identical embryonic lethal phenotype of Mek5 and Erk5 knockout mice due to cardiovascular defects (7–10). Furthermore, ERK5 has been implicated in endothelial cell function (9, 11), skeletal muscle cell fusion (12), neuronal cell differentiation (13), the survival and proliferation of immune cells (14, 15), and carcinogenesis (16). In vivo and clinical data suggest that aberrant or altered ERK5 signaling is implicated in skin cancer (17), colorectal cancer (18), squamous cell lung, and esophageal cancer (19). ERK5 is amplified in hepatocellular carcinoma (20) and is considered a potential therapeutic target in triple negative breast cancer (21).

Previous studies have demonstrated that MEK5 and ERK5 signaling is upregulated in prostate cancer and is associated with metastases and reduced patient survival (22–24). Therefore, we investigated the in vivo role of Erk5 in a genetically engineered mouse (GEM) model of prostate cancer. In the Pbsn (PB)-Cre4:Ptenfl/fl (Pten null; herein referred to as Ptenfl/fl) model of prostate cancer (25, 26), we show for the first time that Erk5 deletion delayed prostate cancer formation, extending the survival of these mice, and reducing tumor size. We also observed elevated Ccl5 and Cxcl10 expression, along with enhanced tumoral T-cell infiltration.

Mouse strains and breeding

Ptenfl/fl [Pbsn (PB)-Cre4:Ptenfl/fl (Pten null)] mice (25, 26) were intercrossed with mice harboring the conditional inactivatable Erk5 allele (where Exon 4 is flanked by LoxP sites; ref. 27) to generate double mutant Ptenfl/fl Erk5fl/fl (PB-Cre4:Ptenfl/fl Erk5fl/fl) mice. Mice were genotyped by PCR by Transnetyx. Mice were of a mixed background (C57BL/6J) with strain-matched littermates (either Cre+ve WT or mice not expressing Cre) used as control mice. Mice were taken at specific time points or aged until end point, which was defined as two of the following three criteria being met: palpable tumor; evidence of blood in the urine; mouse becoming thin/weight loss and “hunched.” All small animal (murine) experiments were approved by the Animal Care and Use Committee at the University of Glasgow.

Sample processing and histologic analysis

Samples for histologic analysis were fixed in 10% formalin for 48 hours, then transferred to 70% ethanol before being processed. Murine prostates were divided into two halves, with one half fixed and processed for histology whereas the other half was snap frozen by placing on dry ice for storage at −80°C. Routine hematoxylin and eosin staining was performed on formalin-fixed, paraffin-embedded (FFPE) prostate samples.

Immunohistochemistry

IHC for Ki67 and CD3 was performed on FFPE sections from a minimum of three individual murine prostate samples for each genotype on Dako Autostainer. Briefly, samples were dewaxed in Xylene then rehydrated through graded alcohols to dH2O. Antigen retrieval was performed using sodium citrate (pH6) buffer in a PT module (Thermo Fisher Scientific) at 98°C for 25 minutes (Ki67, CD45, CD3). Slides were then washed in Tris-buffered TWEEN (TbT) prior to incubation in 3% H2O2 for 5 minutes to block endogenous peroxidases. Next, after washing in TbT, slides were either blocked with ImmPRESS normal goat blocking serum (Vector Laboratories) for 30 minutes then directly incubated with appropriate primary antibody for 35 minutes. Following antibody incubation, slides were washed twice with TbT then staining was visualized using DAKO envision HRP-tagged secondary antibodies and DAB chromogen kit. Slides were counterstained with Hematoxylin Z prior to mounting in DPX media. Antibodies used were: Ki67 [Thermo Fisher Scientific, (SP6) MA5-14520, 1:100], CD3 [Abcam, (SP7) ab16669, 1:50].

In situ hybridization

Single in situ hybridization (ISH) detection for Erk5 [Advanced Cell Diagnostics (ACD) Probe: 433911] was performed manually using RNAscope 2.5 HD reagent kit (brown; ACD, 322310) according to manufacturer's instructions. Single ISH detection for Ccl5 (ACD Probe: 469608), Cxcl10 (ACD Probe: 408928), Ppib (ACD Probe: 313911), and Dapb (ACD Probe: 310043) was performed using RNAscope 2.5 LS reagent kit (brown; ACD, 322100) on the Lecia Bond Rx Autostainer according to manufacturer's instructions. For all single stains, positive staining was indicated by brown punctate dots present in the nucleus and/or cytoplasm.

Dual ISH detection for Cd4 (fast red) and Cd8 (DAB) mRNA was performed by an automated method using RNAscope 2.5 HD Duplex Assay Kit (ACD, 322440) according to the manufacturer's instructions on the Lecia Bond Rx Autostainer. Positive staining was indicated by red and brown punctate dots present in the nucleus and/or cytoplasm.

Microscopy

Light microscopy was carried out using the Olympus BX51 microscope using Olympus UPlanFLN UIS2 10× and 40× objectives, Olympus DP71 camera and CellF software or Zeiss Axio Imager.A1 microscope using Zeiss EC Plan-NEOFLUAR 10× and 40× objectives, Zeiss Axiocam 10S camera, and ZEN software.

Image analysis of IHC and ISH staining

Automated scoring of Ki67 IHC staining was performed using an algorithm designed within SlidePath's (Leica Biosystems) Tissue IA system to count nuclear Ki67-positive cells in scanned “virtual” slides (×20 magnification). Prior to image analysis, 12 representative regions of prostate epithelium per slide were annotated. A minimum of 3 slides per genotype being studied were assessed. All annotated regions on each slide were subsequently submitted for batch image analysis. The image analysis results for these slides were exported to Microsoft Excel. The overall percentage of Ki67 for each sample was calculated by dividing the total number of Ki67-positive cells identified in the 12 annotated areas by the total number of cells identified in the 12 annotated areas on each slide.

For scoring of Ccl5 and Cxcl10 RNA ISH (RNA scope), an algorithm within HALO software was designed and used to count the number of probe copies (dots) and number of cells in annotated areas of scanned slides (×40 magnification). A minimum of 12 representative areas of prostate epithelium per slide were annotated and 4 (Ptenfl/fl) or 5 (Ptenfl/fl Erk5fl/fl) slides were assessed for each genotype. The image analysis results for these slides were exported to Microsoft Excel, which included a value for the average number of probe copies per cell across the entire areas studied.

For epithelial CD3 scoring, SlidePath software was used to design an algorithm to automatically count the number of cells in annotated areas of scanned slides (×20 magnification). Again, 12 representative regions of prostate epithelium per slide were annotated and 3 slides per genotype were assessed. The number of CD3-positive cells within each annotated area was counted manually. The percentage of CD3-positive cells for each sample was calculated by dividing the total number of CD3-positive cells in the 12 annotated areas by the total number of cells identified by the software in the 12 annotated areas on each slide.

For stromal CD3 scoring, lymphoid aggregate areas where intense CD3 staining was present were annotated and their area was calculated using SlidePath Software in scanned slides (×20 magnification). The total area of prostate present on each slide was also evaluated using Leica software. The total combined area of stained lymphoid aggregates on an individual slide per total prostate area on that slide was calculated for each sample. Three slides were evaluated for each genotype.

Deconvolution of Cd4/Cd8 dual ISH (RNA scope) staining in scanned slides (×40 magnification) was performed using HALO software.

Laser capture microdissection

Frozen sections of mouse prostate tissue were cut using a cryostat (2 × 20-μm sections per slide) and put onto Leica FrameSlides (PET-membrane)—5 slides were prepared for each sample; 3 independent samples for each genotype. Slides were stored at −80°C until needed.

Prior to laser capture microdissection (LCM), slides were stained with cresyl violet immediately after being taken out of −80°C. The staining procedure was as follows: 70% ethanol (EtOH; 30 seconds), 50% EtOH (30 seconds); Cresyl violet (1% in 50% EtOH, filtered; 45 seconds), wash twice in DEPC-treated dH2O, 50% EtOH (30 seconds), 70% EtOH (30 seconds), 95% EtOH (30 seconds), 100% EtOH (30 seconds), 100% EtOH (30 seconds). Note that 1× RNase Inhibitor (Sigma) was added to all solutions after the cresyl violet stain.

LCM was carried out using Leica LMD6000 Laser Microdissection System Z6018. Captured prostate epithelial material was collected into the lids of 0.5 mL Eppendorf tubes containing 20 μL RLT lysis buffer (Qiagen). After cutting [for a maximum of 30 min per slide to minimize RNA degradation), a further 30 μL RLT lysis buffer [containing β-mercaptoethanol (1%)] was added to the lystaes before they were stored at −80°C until needed.

RNA isolation and quantitation

For LCM samples, RNA was isolated using RNeasy Micro Kit (Qiagen) as per the manufacturer's instructions including a DNase treatment step. RNA was then quantified using Qbit Assay (Invitrogen) according to the manufacturer's instructions, and quality assessed by running on 2100 Bioanalyzer (Agilent) to generate RNA electropherograms, with calculation of RNA integrity number (RIN; range, 7–10).

For non-LCM samples, RNA was isolated using RNeasy Mini Kit (Qiagen) as per manufacturer's instructions including a DNase treatment step. RNA samples were quantified by spectrophotometry using the NanoDrop 2000 spectrophotometer (Thermo Scientific).

RNA sequencing and bioinformatic analysis

Two hundred nanograms of LCM RNA samples (3 samples for each genotype) was processed and libraries prepared for sequencing using the TruSeq Stranded mRNA kit (Illumina, RS-122-2101) with Poly(A) selection according to the manufacturer's instructions. The amplified libraries were sequenced on the Nextseq 500 (Illumina) with a paired-end sequencing strategy. The read length was set at 75 bp with an expected total library size of ∼330 bp. The mean number of reads per sample was over 30 million.

Quality checks on the raw RNASeq data files were done using fastqc (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and fastq_screen (http://www.bioinformatics.babraham.ac.uk/projects/fastq_screen). RNA sequencing (RNA-Seq) reads were aligned to the GRCm38 (28) version of the mouse genome using tophat2 version 2.0.10 (29) with Bowtie version 2.1.0 (30). Expression levels were determined and statistically analyzed by a combination of HTSeq version 0.5.4p3 (http://www-huber.embl.de/users/anders/HTSeq/doc/overview.html), the R 3.1.1 environment, utilizing packages from the Bioconductor data analysis suite and differential gene expression analysis based on a generalized linear model using the DESeq2 (31). Pathway analysis of genes with a fold change >1.5 and an adjusted P value <0.05 was performed using GeneGo Pathways Software (MetaCore; https://portal.genego.com/version 6.22.67265).

Quantitative real-time PCR

First-strand cDNA was prepared from RNA samples using the High Capacity cDNA Transcription Kit (Applied Biosystems). Quantitative PCR (qPCR) was performed to evaluate relative transcript expression levels in cDNA samples using the Taqman technique. Primers were designed using the Universal Probe Library (UPL) Assay design center (Roche). Specific primer/probe combinations (Supplementary Table S7) and Taqman Universal PCR Mastermix (Applied Biosystems) were added to cDNA samples (according to the manufacturer's instructions) before being run on an Applied Biosystems 7500 Fast Real-Time PCR System machine. The qPCR program conditions were: 20 seconds at 50°C, 10 minutes at 95°C, followed by 40 cycles of 15 seconds at 95°C and 1 minute at 60°C. Technical replicates (2 or 3 wells) for each sample and three independent samples for each genotype/cell line were included. Casc3 was used as the reference gene (chosen as it had similar expression level across various genotypes) and 7500 Software v2.0.5 (Applied Biosystems) used 2−ΔΔCT method to determine the relative gene expression.

Culture of human prostate cancer cell lines and generation and culture of murine tumor-derived cell lines

The human prostate cancer cell line CWR22 was obtained from the laboratory of Dr. Thomas Pretlow, Case Western University, Cleveland, Ohio, in 2002. CWR22 cells were maintained in RPMI medium containing 10% FBS. These cells were last authenticated in-house using Promega GenePrint 10 Kit, according to manufacturer's instructions, in January 2016 and had a >95% match with ATCC and DSMZ (German collection of cell cultures) databases. All cell lines in the laboratory are routinely tested every 6 months for mycoplasma in-house using Lonza MycoAlert Mycoplasma Detection Kit, according to manufacturer's instructions. The length of time between thawing of cells and their use in the experiments described was within 6 weeks.

Tumor-derived cell lines from Ptenfl/fl (P1) and double mutant PB-Cre4:Ptenfl/fl Erk5fl/fl (PE) prostate tumors were generated from a small piece of murine prostate tumor tissue. Prostate tumors taken at the point of dissection were dipped in 70% EtOH, rinsed with dH2O then placed into DMEM (Gibco) medium containing 20% fetal bovine serum (FBS; Gibco). The tumor tissue was chopped as finely as possible then incubated at 37°C for 30 minutes with the addition of collagenase (1 mg/mL; Sigma) and hyaluronidase (2 mg/mL; Sigma) to the medium at 37°C for 1 hour. Tissue was then pelleted by centrifugation at 241 × g for 5 minutes prior to resuspension in fresh DMEM medium containing 20% FBS and antibiotics [penicillin/streptomycin (1%; Gibco) and gentamicin (33 μg/mL; Sigma)]. The following day, adherent cells were washed with PBS then cells were maintained in DMEM medium containing 10% FBS N.B. antibiotics were removed after 1 week of culture.

Statistical analysis

All statistical analyses (namely, Mann–Whitney, Pearson correlation coefficient, t test and Kaplan–Meier survival analysis) were performed using GraphPad Prism v5.0c. For all graphs, mean ± SEM (error bars) are presented.

Erk5 deletion extended the survival of Pten-null mice and reduced prostate tumor weight

Consistent with published literature implicating elevated ERK5 signaling in tumorigenesis, recent The Cancer Genome Atlas sequencing data (32) reported on cBioportal (33, 34) demonstrate altered (predominantly upregulated) MEK5 and ERK5 status in around 11% of 333 human prostate cancer cases (Fig. 1A). Therefore, we sought to investigate whether ablating Erk5 expression in the prostate could have an antitumorigenic effect in vivo using a murine model of prostate cancer. The Ptenfl/fl model (26) is widely used in prostate cancer research. Ptenfl/fl mice develop preneoplastic prostatic intraepithelial neoplasia (PIN) at around 3 months of age, followed by slow progression to prostate adenocarcinoma at >10 months, with no appreciable evidence of metastases up to 18 months of age (35).

Figure 1.

Ptenfl/fl Erk5fl/fl mice survived longer and had smaller prostate tumors compared with Ptenfl/fl mice. A, Oncoprint of genetic alteration of ERK5 (MAPK7) and MEK5 (MAP2K5) in a human prostate tumor; The Cancer Genome Atlas dataset is available on cBioportal (The Cancer Genome Atlas Research Network, 2015). “Deep deletion” refers to homozygous deletion. B, Kaplan–Meier survival curve demonstrating increased survival in Ptenfl/fl Erk5fl/fl (n = 11) compared with Ptenfl/fl (n = 13) mice (***, P = 0.0006; log-rank Mantel–Cox test; see Supplementary Table S1). C, Representative images of isolated prostates from WT, Erk5fl/fl, Ptenfl/fl, and Ptenfl/fl Erk5fl/fl mice taken at endpoint. Highlighted in the WT image are anterior prostate (AP; red arrows), dorsal–lateral prostate (DLP), ventral prostate (VP), and seminal vesicles (SV). D–F, Scatter plots comparing differences between Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostate weights at 3 months (D), 9–10 months (E), and endpoint (11–15 months for Ptenfl/fl; 15 months for Ptenfl/fl Erk5fl/fl; F). Also included in F are weights of WT and Erk5fl/fl prostates (age 15 months). In E and F, because of the cystic nature of the tumors, isolated prostates were weighed prior to removal of the cystic fluid (termed wet weight, WW), then reweighed to assess the solid tumor mass with the cysts drained (termed dry weight, DW). Cystic component was also calculated [((WW-DW)/WW)*100]. Shown in D–F are individual data points; long horizontal line, mean; error bars, SEM; t test (unpaired, two-tailed) was used to calculate P values; key significant changes (P < 0.05) are shown (see also Supplementary Tables S2–S4).

Figure 1.

Ptenfl/fl Erk5fl/fl mice survived longer and had smaller prostate tumors compared with Ptenfl/fl mice. A, Oncoprint of genetic alteration of ERK5 (MAPK7) and MEK5 (MAP2K5) in a human prostate tumor; The Cancer Genome Atlas dataset is available on cBioportal (The Cancer Genome Atlas Research Network, 2015). “Deep deletion” refers to homozygous deletion. B, Kaplan–Meier survival curve demonstrating increased survival in Ptenfl/fl Erk5fl/fl (n = 11) compared with Ptenfl/fl (n = 13) mice (***, P = 0.0006; log-rank Mantel–Cox test; see Supplementary Table S1). C, Representative images of isolated prostates from WT, Erk5fl/fl, Ptenfl/fl, and Ptenfl/fl Erk5fl/fl mice taken at endpoint. Highlighted in the WT image are anterior prostate (AP; red arrows), dorsal–lateral prostate (DLP), ventral prostate (VP), and seminal vesicles (SV). D–F, Scatter plots comparing differences between Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostate weights at 3 months (D), 9–10 months (E), and endpoint (11–15 months for Ptenfl/fl; 15 months for Ptenfl/fl Erk5fl/fl; F). Also included in F are weights of WT and Erk5fl/fl prostates (age 15 months). In E and F, because of the cystic nature of the tumors, isolated prostates were weighed prior to removal of the cystic fluid (termed wet weight, WW), then reweighed to assess the solid tumor mass with the cysts drained (termed dry weight, DW). Cystic component was also calculated [((WW-DW)/WW)*100]. Shown in D–F are individual data points; long horizontal line, mean; error bars, SEM; t test (unpaired, two-tailed) was used to calculate P values; key significant changes (P < 0.05) are shown (see also Supplementary Tables S2–S4).

Close modal

We targeted ablation of Erk5 expression to the prostate by crossing mice harboring the conditional inactivatable Erk5 allele (where Exon 4 is flanked by LoxP sites; ref. 27) to Ptenfl/fl mice to generate double mutant PB-Cre4:Ptenfl/fl Erk5fl/fl (herein referred to as Ptenfl/fl Erk5fl/fl) mice. Analyzing isolated prostatic epithelium by LCM from end point mice, qPCR assay for Erk5 confirmed significantly reduced Erk5 expression in prostates null for Erk5 (PB-Cre4:Erk5fl/fl; herein referred to Erk5fl/fl) compared with WT and in Ptenfl/fl Erk5fl/fl compared with Ptenfl/fl prostates (Fig. 2C, left). Similarly, we found reduced expression of Pten mRNA in Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostates compared with WT (Fig. 2C, right).

Figure 2.

Histopathologic analysis revealed that Ptenfl/fl Erk5fl/fl prostate tumors were less proliferative compared with Ptenfl/fl prostate tumors and that Ptenfl/fl prostate tumors expressed more Erk5 compared with WT prostate. A, Hematoxylin and eosin (H&E) staining was performed to examine the tissue histology and Ki67 IHC was employed to investigate the level of cellular proliferation in prostates from WT, Erk5fl/fl, Ptenfl/fl, and Ptenfl/fl Erk5fl/fl mice (minimum of 3 for each genotype). Representative micrographs of hematoxylin and eosin and Ki67 staining for each genotype are shown (main images ×10; insets are higher magnification (×40) view of prostate epithelium). B, Percentage of Ki67-positive cells in each sample was scored using an automated method (Leica software). Twelve representative fields of view per sample were scored for 6 samples for each genotype at endpoint. (WT, all 15 months; Erk5fl/fl all 15 months; Ptenfl/fl age range 11–15 months; Ptenfl/fl Erk5fl/fl, all 15 months. C, qPCR analysis of RNA isolated from mouse prostate epithelium by LCM in WT, Erk5fl/fl, Ptenfl/fl, and Ptenfl/fl Erk5fl/fl mice (n = 3 for each genotype) for Erk5 Exon 4 (the floxed allele) and Pten. D, RNA ISH analysis by RNA scope was utilized to examine the level of Erk5 expression in WT prostate and Ptenfl/fl prostate tumors (n = 3 for each genotype). Representative micrographs (×100) of RNA scope staining for Erk5, Ppib (positive control for RNA integrity), and Dapb (negative control for probe specificity) of WT prostate and Ptenfl/fl tumors at endpoint are presented. Shown in B and C are the means; error bars, SEM; t test (unpaired, two-tailed) was used to calculate P values; key significant changes (P < 0.05) are shown.

Figure 2.

Histopathologic analysis revealed that Ptenfl/fl Erk5fl/fl prostate tumors were less proliferative compared with Ptenfl/fl prostate tumors and that Ptenfl/fl prostate tumors expressed more Erk5 compared with WT prostate. A, Hematoxylin and eosin (H&E) staining was performed to examine the tissue histology and Ki67 IHC was employed to investigate the level of cellular proliferation in prostates from WT, Erk5fl/fl, Ptenfl/fl, and Ptenfl/fl Erk5fl/fl mice (minimum of 3 for each genotype). Representative micrographs of hematoxylin and eosin and Ki67 staining for each genotype are shown (main images ×10; insets are higher magnification (×40) view of prostate epithelium). B, Percentage of Ki67-positive cells in each sample was scored using an automated method (Leica software). Twelve representative fields of view per sample were scored for 6 samples for each genotype at endpoint. (WT, all 15 months; Erk5fl/fl all 15 months; Ptenfl/fl age range 11–15 months; Ptenfl/fl Erk5fl/fl, all 15 months. C, qPCR analysis of RNA isolated from mouse prostate epithelium by LCM in WT, Erk5fl/fl, Ptenfl/fl, and Ptenfl/fl Erk5fl/fl mice (n = 3 for each genotype) for Erk5 Exon 4 (the floxed allele) and Pten. D, RNA ISH analysis by RNA scope was utilized to examine the level of Erk5 expression in WT prostate and Ptenfl/fl prostate tumors (n = 3 for each genotype). Representative micrographs (×100) of RNA scope staining for Erk5, Ppib (positive control for RNA integrity), and Dapb (negative control for probe specificity) of WT prostate and Ptenfl/fl tumors at endpoint are presented. Shown in B and C are the means; error bars, SEM; t test (unpaired, two-tailed) was used to calculate P values; key significant changes (P < 0.05) are shown.

Close modal

Compared with Ptenfl/fl mice, Ptenfl/fl Erk5fl/fl mice have significantly prolonged median survival (Fig. 1B; Supplementary Table S1). At study end point (Ptenfl/fl: age range 9–15 months, average 11.54 months; Ptenfl/fl Erk5fl/fl: age range 9.5–15.5 months, average 14.67 months), prostate tumors in Ptenfl/fl mice looked similar to our previous report (35) whereas prostate tumors in double mutant Ptenfl/fl Erk5fl/fl mice were visibly smaller in appearance compared with Ptenfl/fl control mice (Fig. 1C). Furthermore, throughout the study period at 3 and 9–10 months of age and at end point, Ptenfl/fl Erk5fl/fl prostates weighted significantly less than those of the Ptenfl/fl control mice (Figs. 1D–F; Supplementary Tables S2–S4).

Ptenfl/flErk5fl/fl prostate tumors were less proliferative than Ptenfl/fl prostate tumors

To further investigate the phenotypes of isolated prostates from Ptenfl/fl and Ptenfl/fl Erk5fl/fl mice, histopathologic analysis was carried out. Significantly, no adverse prostate phenotype was observed in Erk5fl/fl mice – these prostates had similar histology to WT control mice (Fig. 2A). At 3 months of age, prostates from both Ptenfl/fl and Ptenfl/fl Erk5fl/fl mice developed preneoplastic PIN lesions, with intact glandular structures (Supplementary Fig. S1A). By 9 to 10 months of age (Supplementary Fig. S1B) and at study end point (Fig. 2A), adenocarcinomas were present in both Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostates with glandular invasion into the surrounding stroma and merging of glands. There was no evidence of proliferative inflammatory atrophy in any of the samples studied.

At the three time points studied, Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostate tumors are highly proliferative, as evidenced by Ki67 IHC staining [Supplementary Figs. S1A (3 months), S1B (9–10 months), and S2A (end point)]. Although we did not see a difference in the percentage of Ki67-positive cells at 3 months in Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostates (Supplementary Fig. S1C), importantly at 9 to 10 months (Supplementary Fig. S1D) and study end point (Fig. 2B), we observed significantly fewer Ki67-positive cells in Ptenfl/fl Erk5fl/fl when compared with Ptenfl/fl prostate tumors. Collectively, double mutant Ptenfl/fl Erk5fl/fl mice survived longer and harbored smaller prostate tumors with reduced proliferation than Ptenfl/fl mice.

Erk5 expression is increased in Ptenfl/fl prostate tumors compared with WT prostate

Interestingly, when analyzing LCM isolated prostatic epithelium by qPCR for Erk5 expression, we discovered that Erk5 expression is significantly increased in Ptenfl/fl prostate tumors compared with WT prostate at end point (Fig. 2C). As an additional confirmation of this result, ISH analysis using RNA scope for Erk5 mRNA was performed and we saw increased ISH signal for Erk5 in Ptenfl/fl prostate tumors compared with WT prostate (Fig. 2D).

Elevated expression of the cytokines Ccl5 and Cxcl10 in Ptenfl/flErk5fl/fl prostate tumors

To gain insight into the molecular differences between Ptenfl/fl, and Ptenfl/fl Erk5fl/fl prostate tumors, RNA-Seq was carried out using RNA samples prepared from prostatic epithelium isolated by LCM from the respective end point tumors (n = 3 for each genotype). Bioinformatic analysis of the data revealed that the expression of 70 genes was significantly altered (P ≤ 0.05) between Ptenfl/fl Erk5fl/fl and Ptenfl/fl prostate tumors. Furthermore, 67 of these genes had >1.5 fold change in expression (Fig. 3A). Among these genes, pathways relating to inflammatory responses, immune responses and regulation of cytokine, and chemokine expression were over-represented (Supplementary Table S6).

Figure 3.

Validation of RNA-Seq data confirmed an increase in the expression of T-cell recruiting cytokines, Ccl5 and Cxcl10, in Ptenfl/fl Erk5fl/fl compared with Ptenfl/fl prostate tumors. A, Heatmap of RNA-Seq data. Presented are the 67 genes that were identified from analysis of RNA-Seq data as having significantly altered expression (P < 0.05; >1.5-fold change) between Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostate tumors (n = 3 for each genotype). In the heatmap, blue represents downregulation of gene expression (row Z-score <0) and red represents upregulation of gene expression (row Z-score>0). B and C, qPCR analysis of LCM prostate epithelial RNA from Ptenfl/fl and Ptenfl/fl Erk5fl/fl tumors for Ccl5 and Cxcl10 at 9–10 month time point (n = 3 for each genotype; B) and at endpoint (n = 3 for each genotype; C). D, qPCR analysis of RNA from tumor-derived cell lines, P1 (Ptenfl/fl) and PE (Ptenfl/fl Erk5fl/fl), for Ccl5 and Cxcl10. E, RNA ISH by RNA scope was used to examine Ccl5 and Cxcl10 expression in Ptenfl/fl (n = 4) and Ptenfl/fl Erk5fl/fl (n = 5) prostate tumors at endpoint. Representative micrographs [main images, ×20; inset (malignant epithelium), ×100] of RNA scope staining for Ccl5, Cxcl10, Ppib, and Dapb are shown. F, Average number of Ccl5 and Cxcl10 probe copies per cell was calculated using an automated method (HALO software). A minimum of 12 representative areas of prostate epithelium per slide were analyzed for each sample [Ptenfl/fl (n = 4) and Ptenfl/fl Erk5fl/fl (n = 5)]. Shown in B–D and F are means; error bars, SEM; t test (unpaired, two-tailed) was used to calculate P values and those with significance (P < 0.05) are specified.

Figure 3.

Validation of RNA-Seq data confirmed an increase in the expression of T-cell recruiting cytokines, Ccl5 and Cxcl10, in Ptenfl/fl Erk5fl/fl compared with Ptenfl/fl prostate tumors. A, Heatmap of RNA-Seq data. Presented are the 67 genes that were identified from analysis of RNA-Seq data as having significantly altered expression (P < 0.05; >1.5-fold change) between Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostate tumors (n = 3 for each genotype). In the heatmap, blue represents downregulation of gene expression (row Z-score <0) and red represents upregulation of gene expression (row Z-score>0). B and C, qPCR analysis of LCM prostate epithelial RNA from Ptenfl/fl and Ptenfl/fl Erk5fl/fl tumors for Ccl5 and Cxcl10 at 9–10 month time point (n = 3 for each genotype; B) and at endpoint (n = 3 for each genotype; C). D, qPCR analysis of RNA from tumor-derived cell lines, P1 (Ptenfl/fl) and PE (Ptenfl/fl Erk5fl/fl), for Ccl5 and Cxcl10. E, RNA ISH by RNA scope was used to examine Ccl5 and Cxcl10 expression in Ptenfl/fl (n = 4) and Ptenfl/fl Erk5fl/fl (n = 5) prostate tumors at endpoint. Representative micrographs [main images, ×20; inset (malignant epithelium), ×100] of RNA scope staining for Ccl5, Cxcl10, Ppib, and Dapb are shown. F, Average number of Ccl5 and Cxcl10 probe copies per cell was calculated using an automated method (HALO software). A minimum of 12 representative areas of prostate epithelium per slide were analyzed for each sample [Ptenfl/fl (n = 4) and Ptenfl/fl Erk5fl/fl (n = 5)]. Shown in B–D and F are means; error bars, SEM; t test (unpaired, two-tailed) was used to calculate P values and those with significance (P < 0.05) are specified.

Close modal

Besides being members of the over-represented pathways, the cytokines chemokine (C-C motif) ligand 5 (Ccl5; Rantes) and C-X-C motif chemokine 10 (Cxcl10) were among the most highly upregulated genes in Ptenfl/fl Erk5fl/fl prostate tumors (Fig. 3A; Supplementary Table S5). Using prostate epithelial RNA samples prepared using LCM, we found a trend for increased Ccl5 and Cxcl10 expression at 9- to 10-month time point in Ptenfl/fl Erk5fl/fl compared with Ptenfl/fl prostate tumors by qPCR (Fig. 3B). At end point, Cxcl10 expression was significantly upregulated in the prostates of Ptenfl/fl Erk5fl/fl mice, along with a trend for increased Ccl5 expression (Fig. 3C). Furthermore, in cell lines derived from Ptenfl/fl (P1) and Ptenfl/fl Erk5fl/fl (PE) prostate tumors, which were validated to lack expression of Pten and both Pten and Erk5, respectively (Supplementary Fig. S2A and S2B), we observed a dramatic increase in Ccl5 and Cxcl10 mRNA expression in PE compared with P1 cells (Fig. 3D).

CCL5 (rantes) and CXCL10 are members of the CC and CXC cytokine families, respectively. They are produced by a wide range of cell types, including endothelial cells, epithelial cells, and fibroblasts, in response to stimulation by CD40L, IL15 (for CCL5), and IFNγ (for CXCL10). CCL5 and CXCL10 are potent chemo-attractants for T lymphocytes upon binding to the appropriate cell surface receptors on T cells: CCR5 for CCL5 and CXCR3 for CXCL10 (36, 37). ISH revealed upregulated expression of Ccl5 and Cxcl10 mRNA in end point Ptenfl/fl and Ptenfl/fl Erk5fl/fl tumors (Fig. 3E), with near-significant increase and significant increase in the number of probe copies per cell within the malignant epithelium respectively (Fig. 3F), which is in keeping with qPCR data (Fig. 3C). IHC staining was also performed to examine the pattern of expression of CCL5, CXCL10, and their cognate receptors, CCR5 and CXCR3, respectively (Supplementary Fig. S2C). CCL5 and CXCL10 immunoreactivity was elevated in Ptenfl/fl Erk5fl/fl prostate tumors, and in infiltrating immune cells in the stroma. While CCR5 expression was largely negative in Ptenfl/fl prostate tumors, patchy CCR5 immunoreactivity was detected in Ptenfl/fl Erk5fl/fl prostate epithelium and immune cells in the stroma. CXCR3 expression was also identified in tumor epithelium and immune cells present in the stroma in Ptenfl/fl Erk5fl/fl prostates.

CD3-positive cells within the epithelium and stromal aggregates were elevated in Ptenfl/fl Erk5fl/fl compared with Ptenfl/fl prostate tumors

To investigate the functional correlates of increased CCL5 and CXCL10 expression in the prostate, we studied T lymphocytes in Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostate tumors by IHC for CD45 (marker for general lymphocyte) and CD3 (a component of the T-cell receptor complex). CD45 staining revealed areas of intense immunoreactivity in the stroma of both Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostate tumors at end point, indicating significant lymphoid infiltration (Supplementary Fig. S3A). Similar to CD45, areas of CD3-positive T cells were observed in the stroma (referred to herein as stromal lymphoid aggregates) as well as within the malignant epithelium at 9- to10-month time point (Supplementary Fig. S1B) and at endpoint (Fig. 4A) in both Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostate tumors. At end point, there were significantly more CD3-positive cells in the prostate tumor epithelium of Ptenfl/fl Erk5fl/fl than in Ptenfl/fl samples (Fig. 4B). Furthermore, CD3-positive T cells were also statistically more abundant in the tumor stroma of Ptenfl/fl Erk5fl/fl than in Ptenfl/fl prostate tumor samples at study end point (Fig. 4C). In keeping with these data, we observed a trend for increased CD3 immunoreactivity in both tumor epithelium and stroma at the earlier 9- to 10-month time point in Ptenfl/fl Erk5fl/fl prostates when compared with the control Ptenfl/fl prostates (Fig. 4B and C). Interestingly, we did not detect substantial CD3-positive cell infiltration in Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostates taken at 3 months of age (Supplementary Fig. S1A), suggesting that lymphoid infiltration may take place as the tumor progresses.

Figure 4.

Infiltration of CD3-positive cells was increased in Ptenfl/fl Erk5fl/fl compared with Ptenfl/fl prostate tumors and T-cell infiltrates were predominantly CD4 positive. A, IHC for the T-cell marker, CD3, was used to investigate the presence and localization of T cells in Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostate tumors (n = 3 for each genotype) and representative micrographs (top panel, ×4; bottom panel, ×40) of the staining are shown. Top, red arrows indicate CD3-stained stromal lymphoid aggregate areas in the stroma; bottom, yellow arrows highlight the presence of CD3-positive cells within the prostate tumor epithelium. B, Percentage CD3-positive cells within the tumor epithelium at 9- to 10-month time point and endpoint in Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostate tumors was determined by automated scoring of total number of cells (Leica software), followed by manual counting of the number of CD3-positive cells per field studied; 12 fields of view per sample for n = 3 mice of each genotype at each time point were analyzed. C, The proportion of CD3 stromal lymphoid aggregate area per total area of prostate on slide in Ptenfl/fl and Ptenfl/fl Erk5fl/fl tumors was calculated by using Leica software to measure the area of lymphoid stromal aggregates and total area of prostate per slide (n = 3 samples for each genotype at each time point). Shown in B and C are means; error bars, SEM; t test (unpaired, two-tailed) was used to calculate P values; key significant changes (P < 0.05) are shown. D, Dual RNA ISH was used to investigate the levels of T-cell markers, Cd4 (Fast Red) and Cd8 (DAB), in Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostate tumors (n = 3 for each genotype) and representative micrographs of the staining (×40) are shown. Blue arrows, Cd8 staining.

Figure 4.

Infiltration of CD3-positive cells was increased in Ptenfl/fl Erk5fl/fl compared with Ptenfl/fl prostate tumors and T-cell infiltrates were predominantly CD4 positive. A, IHC for the T-cell marker, CD3, was used to investigate the presence and localization of T cells in Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostate tumors (n = 3 for each genotype) and representative micrographs (top panel, ×4; bottom panel, ×40) of the staining are shown. Top, red arrows indicate CD3-stained stromal lymphoid aggregate areas in the stroma; bottom, yellow arrows highlight the presence of CD3-positive cells within the prostate tumor epithelium. B, Percentage CD3-positive cells within the tumor epithelium at 9- to 10-month time point and endpoint in Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostate tumors was determined by automated scoring of total number of cells (Leica software), followed by manual counting of the number of CD3-positive cells per field studied; 12 fields of view per sample for n = 3 mice of each genotype at each time point were analyzed. C, The proportion of CD3 stromal lymphoid aggregate area per total area of prostate on slide in Ptenfl/fl and Ptenfl/fl Erk5fl/fl tumors was calculated by using Leica software to measure the area of lymphoid stromal aggregates and total area of prostate per slide (n = 3 samples for each genotype at each time point). Shown in B and C are means; error bars, SEM; t test (unpaired, two-tailed) was used to calculate P values; key significant changes (P < 0.05) are shown. D, Dual RNA ISH was used to investigate the levels of T-cell markers, Cd4 (Fast Red) and Cd8 (DAB), in Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostate tumors (n = 3 for each genotype) and representative micrographs of the staining (×40) are shown. Blue arrows, Cd8 staining.

Close modal

CD4-positive T cells were abundant in Ptenfl/flErk5fl/fl tumors

We next studied the subtypes of T lymphocytes in the Ptenfl/fl Erk5fl/fl tumors, focusing on those that may have an antitumor immunity function, namely CD8-positive cytotoxic T lymphocytes (TC) and CD4-positive T lymphocytes, which include helper T lymphocytes (TH) and natural killer (NK) T cells (TNK). ISH analysis using dual RNA scope for Cd4 and Cd8 mRNA was performed. Although ISH signals for both Cd4 (fast red) and Cd8 (DAB) were detected (Fig. 4D), Cd4-positive T cells were more abundant, particularly within the lymphoid aggregates of Ptenfl/fl Erk5fl/fl tumors. As an additional validation to the ISH experiment, dual IHC staining for CD3 and CD4 protein was also done. In keeping with the Cd4 ISH signal that we observed in the lymphoid aggregates in Ptenfl/fl Erk5fl/fl prostate tumors, we observed strong immunoreactivity for CD4-positive cells (Supplementary Fig. S3B). At higher magnification, there were cells that were positive for both CD3 and CD4 as well as cells expressing either CD3 or CD4.

The current standard-of-care treatment for patients with advanced prostate cancer remains unsatisfactory, with significant risk of premature death among these patients. For the first time, we demonstrated that Erk5 deletion in the prostates of PB-Cre4:Ptenfl/fl mice suppressed prostate tumorigenesis, resulting in smaller tumors and prolonged mouse survival (Figs. 1, 2A and B). Interestingly, we also observed increased Erk5 expression in Ptenfl/fl compared with WT prostate at end point by both qPCR and ISH analysis (Figs. 2C and D), which has not previously been reported.

Our RNA sequencing analysis to compare Ptenfl/fl and Ptenfl/fl Erk5fl/fl prostate epithelial tumors revealed pathways relating to inflammatory responses, immune responses and regulation of cytokine and chemokine expression to be significantly altered. There are several reports in the literature indicating a role for ERK5 in inflammation, with both anti-inflammatory (38–40) and pro-inflammatory effects (41–43) being documented. ERK5 signaling has also been implicated in inflammation-driven skin cancer (17).

The cytokine encoding genes Ccl5 and Cxcl10 were highly upregulated in their levels of expression as well as being members of significantly enriched pathways in Ptenfl/fl Erk5fl/fl prostate tumors when compared with Ptenfl/fl control prostate tumors (Fig. 3; Supplementary Tables S5 and S6). There are emerging roles of the CCL5/CCR5 and CXCL10/CXCR3 axes in cancer [reviewed in (36, 37)]. Importantly, CCL5 and CXCL10 can direct the innate and adaptive immune response to mediate an antitumor effect via their ability to attract leukocytes to inflammatory sites, including cytotoxic (CD8+), helper (CD4+), and NK T lymphocytes. Recent reports have found that local secretion of CCL5 and CXCL10 correlates with CD8+ T-cell infiltration in colorectal cancer (44) and esophageal squamous cell carcinoma (45). In metastatic melanoma, overexpression of CXCL10 and CCL5 was associated with responsiveness to adoptive therapy with tumor-infiltrating lymphocytes (46). Similarly, epigenetic silencing of CXCL9 and CXCL10 appeared to promote tumor immune-evasion in ovarian cancer (47). Restoring the expression of these cytokines enhanced effector T-cell infiltration, reduced tumor progression and increased tumor response to immune-based therapy in vivo. Overexpression of CXCL10 has previously been found to inhibit the proliferation of LNCaP prostate cancer cells, an effect that was accompanied by induced CXCR3 expression (48). A strong positive correlation was observed between CXCL10 and local CD8 expression in human prostate tumor explants (49). Our data demonstrating an increase in T-cell infiltrate to the tumor microenvironment of Ptenfl/fl Erk5fl/fl prostate tumors that have enhanced local bioavailability of CCL5 and CXCL10 (Fig. 4) is in keeping with these reports in the literature.

Although we observed some Cd8 mRNA ISH signal in the stromal aggregates of Ptenfl/fl Erk5fl/fl tumors, the levels of Cd4 mRNA ISH signal were striking (Fig. 4D). Moreover, from CD3/CD4 dual IHC protein analysis, we also saw high levels of CD4 immunoreactivity (Supplementary Fig. S3B), indicating a predominantly CD4-positive cell population. Interestingly, it can be seen from the dual IHC staining that there are CD3/CD4 dual-positive cells, which could incorporate TH and NK T cells. There are also cells that are positive for just CD3 (these could include CD8-positive cells and NK T cells) or CD4 (these could comprise regulatory T cells).

In conclusion, our data support ERK5 as a target for therapy in prostate cancer with the loss of Erk5 in a Pten-null background impairing prostate tumorigenesis and extending the survival of the study animals. Shown in Fig. 5 is a schematic summary of our other key findings that Erk5 deletion in combination with Pten deletion resulted in upregulation of the T-cell recruiting cytokine genes, Ccl5 and Cxcl10, and that this was associated with enhanced recruitment of CD4 (predominantly) and CD8-positive T lymphocytes to the tumor epithelium itself and distinct sites within the surrounding stroma. The role of the ERK5 pathway in the host immune response warrants further investigation.

Figure 5.

Schematic representation summarizing the compartmentalization and potential role of the cytokines, CCL5 and CXCL10, in the context of Pten and Erk5 loss in prostate epithelium. The cytokines CCL5 and CXCL10 are produced by a variety of cell types, including epithelial cells, endothelial cells, fibroblasts, in response to external stimuli such as IFNγ, CD40L, and IL15. CCL5 and CXCL10, can exert autocrine and paracrine effects on the cells from which they were produced. However, the major function of CCL5 and CXCL10 is to recruit T lymphocytes to their site of production via recognition of the cytokines by their corresponding receptors on the surface of the T lymphocytes (CCR5 for CCL5 and CXCR3 for CXCL10). T cells can also produce CCL5 and CXCL10. We propose that loss of Erk5 in combination with loss of Pten leads to increased CCL5 and CXCL10 expression within the malignant tumor epithelium. This in turn is associated with recruitment of T cells (predominantly CD4+) to the tumor epithelium and distinct sites within the stroma (lymphoid aggregates).

Figure 5.

Schematic representation summarizing the compartmentalization and potential role of the cytokines, CCL5 and CXCL10, in the context of Pten and Erk5 loss in prostate epithelium. The cytokines CCL5 and CXCL10 are produced by a variety of cell types, including epithelial cells, endothelial cells, fibroblasts, in response to external stimuli such as IFNγ, CD40L, and IL15. CCL5 and CXCL10, can exert autocrine and paracrine effects on the cells from which they were produced. However, the major function of CCL5 and CXCL10 is to recruit T lymphocytes to their site of production via recognition of the cytokines by their corresponding receptors on the surface of the T lymphocytes (CCR5 for CCL5 and CXCR3 for CXCL10). T cells can also produce CCL5 and CXCL10. We propose that loss of Erk5 in combination with loss of Pten leads to increased CCL5 and CXCL10 expression within the malignant tumor epithelium. This in turn is associated with recruitment of T cells (predominantly CD4+) to the tumor epithelium and distinct sites within the stroma (lymphoid aggregates).

Close modal

No potential conflicts of interest were disclosed.

Conception and design: C.J. Loveridge, O. Sansom, H.Y. Leung

Development of methodology: C.J. Loveridge, R. Patel, I. Ahmad, M. Welsh, J. Galbraith, H.Y. Leung

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C.J. Loveridge, E.J. Mui, E.H. Tan, I. Ahmad, M. Welsh, J. Galbraith, A. Hedley, K. Blyth, O. Sansom, H.Y. Leung

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C.J. Loveridge, M. Welsh, A. Hedley, H.Y. Leung

Writing, review, and/or revision of the manuscript: C.J. Loveridge, M. Welsh, A. Hedley, O. Sansom, H.Y. Leung

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): C.J. Loveridge, A. Hedley, C. Nixon

Study supervision: O. Sansom, H.Y. Leung

Other (advised on in vivo experiments): K. Blyth

We thank the CRUK Beatson Institute core research services, including the biological services unit and the Beatson histology department. Mice harboring the conditional inactivatable Erk5 allele (where Exon 4 is flanked by LoxP sites) were provided by Dr. Kathy Tournier, University of Manchester. We acknowledge Dr. Jonathan Salmond, consultant pathologist NHS Greater Glasgow and Clyde, for histopathologic examination of tissues.

This work was supported by Cancer Research UK Beatson Institute (C596/A17196; Prof. Hing Y. Leung, CRUK A15151; Prof. Hing Y. Leung) and Prostate Cancer UK (PG10-10; Prof. Hing Y. Leung).

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.

3.
Simões
AE
,
Rodrigues
CM
,
Borralho
PM
. 
The MEK5/ERK5 signalling pathway in cancer: a promising novel therapeutic target
.
Drug Discov Today
2016
;
21
:
1654
63
.
4.
Nishimoto
S
,
Nishida
E
. 
MAPK signalling: ERK5 versus ERK1/2
.
EMBO Rep
2006
;
7
:
782
6
.
5.
Wang
X
,
Tournier
C
. 
Regulation of cellular functions by the ERK5 signalling pathway
.
Cell Signal
2006
;
18
:
753
60
.
6.
Drew
BA
,
Burow
ME
,
Beckman
BS
. 
MEK5/ERK5 pathway: the first fifteen years
.
Biochim Biophys Acta
2012
;
1825
:
37
48
.
7.
Regan
CP
,
Li
W
,
Boucher
DM
,
Spatz
S
,
Su
MS
,
Kuida
K
. 
Erk5 null mice display multiple extraembryonic vascular and embryonic cardiovascular defects
.
Proc Natl Acad Sci U S A
2002
;
99
:
9248
53
.
8.
Yan
L
,
Carr
J
,
Ashby
PR
,
Murry-Tait
V
,
Thompson
C
,
Arthur
JS
. 
Knockout of ERK5 causes multiple defects in placental and embryonic development
.
BMC Dev Biol
2003
;
3
:
11
.
9.
Hayashi
M
,
Kim
SW
,
Imanaka-Yoshida
K
,
Yoshida
T
,
Abel
ED
,
Eliceiri
B
, et al
Targeted deletion of BMK1/ERK5 in adult mice perturbs vascular integrity and leads to endothelial failure
.
J Clin Invest
2004
;
113
:
1138
48
.
10.
Wang
X
,
Merritt
AJ
,
Seyfried
J
,
Guo
C
,
Papadakis
ES
,
Finegan
KG
, et al
Targeted deletion of mek5 causes early embryonic death and defects in the extracellular signal-regulated kinase 5/myocyte enhancer factor 2 cell survival pathway
.
Mol Cell Biol
2005
;
25
:
336
45
.
11.
Roberts
OL
,
Holmes
K
,
Müller
J
,
Cross
DA
,
Cross
MJ
. 
ERK5 and the regulation of endothelial cell function
.
Biochem Soc Trans
2009
;
37
:
1254
9
.
12.
Sunadome
K
,
Yamamoto
T
,
Ebisuya
M
,
Kondoh
K
,
Sehara-Fujisawa
A
,
Nishida
E
. 
ERK5 regulates muscle cell fusion through Klf transcription factors
.
Dev Cell
2011
;
20
:
192
205
.
13.
Cavanaugh
JE.
Role of extracellular signal regulated kinase 5 in neuronal survival
.
Eur J Biochem
2004
;
271
:
2056
9
.
14.
Rovida
E
,
Spinelli
E
,
Sdelci
S
,
Barbetti
V
,
Morandi
A
,
Giuntoli
S
, et al
ERK5/BMK1 is indispensable for optimal colony-stimulating factor 1 (CSF-1)-induced proliferation in macrophages in a Src-dependent fashion
.
J Immunol
2008
;
180
:
4166
72
.
15.
Sohn
SJ
,
Lewis
GM
,
Winoto
A
. 
Non-redundant function of the MEK5-ERK5 pathway in thymocyte apoptosis
.
EMBO J
2008
;
27
:
1896
906
.
16.
Lochhead
PA
,
Gilley
R
,
Cook
SJ
. 
ERK5 and its role in tumour development
.
Biochem Soc Trans
2012
;
40
:
251
6
.
17.
Finegan
KG
,
Perez-Madrigal
D
,
Hitchin
JR
,
Davies
CC
,
Jordan
AM
,
Tournier
C
. 
ERK5 is a critical mediator of inflammation-driven cancer
.
Cancer Res
2015
;
75
:
742
53
.
18.
Simões
AE
,
Pereira
DM
,
Gomes
SE
,
Brito
H
,
Carvalho
T
,
French
A
, et al
Aberrant MEK5/ERK5 signalling contributes to human colon cancer progression via NF-κB activation
.
Cell Death Dis
2015
;
6
:
e1718
.
19.
Gavine
PR
,
Wang
M
,
Yu
D
,
Hu
E
,
Huang
C
,
Xia
J
, et al
Identification and validation of dysregulated MAPK7 (ERK5) as a novel oncogenic target in squamous cell lung and esophageal carcinoma
.
BMC Cancer
2015
;
15
:
454
.
20.
Zen
K
,
Yasui
K
,
Nakajima
T
,
Zen
Y
,
Gen
Y
,
Mitsuyoshi
H
, et al
ERK5 is a target for gene amplification at 17p11 and promotes cell growth in hepatocellular carcinoma by regulating mitotic entry
.
Genes Chromosomes Cancer
2009
;
48
:
109
20
.
21.
Ortiz-Ruiz
MJ
,
Álvarez-Fernández
S
,
Parrott
T
,
Zaknoen
S
,
Burrows
FJ
,
Ocaña
A
, et al
Therapeutic potential of ERK5 targeting in triple negative breast cancer
.
Oncotarget
2014
;
5
:
11308
18
.
22.
Mehta
PB
,
Jenkins
BL
,
McCarthy
L
,
Thilak
L
,
Robson
CN
,
Neal
DE
, et al
MEK5 overexpression is associated with metastatic prostate cancer, and stimulates proliferation, MMP-9 expression and invasion
.
Oncogene
2003
;
22
:
1381
9
.
23.
McCracken
SR
,
Ramsay
A
,
Heer
R
,
Mathers
ME
,
Jenkins
BL
,
Edwards
J
, et al
Aberrant expression of extracellular signal-regulated kinase 5 in human prostate cancer
.
Oncogene
2008
;
27
:
2978
88
.
24.
Ramsay
AK
,
McCracken
SR
,
Soofi
M
,
Fleming
J
,
Yu
AX
,
Ahmad
I
, et al
ERK5 signalling in prostate cancer promotes an invasive phenotype
.
Br J Cancer
2011
;
104
:
664
72
.
25.
Lesche
R
,
Groszer
M
,
Gao
J
,
Wang
Y
,
Messing
A
,
Sun
H
, et al
Cre/loxP-mediated inactivation of the murine Pten tumor suppressor gene
.
Genesis
2002
;
32
:
148
9
.
26.
Wu
X
,
Wu
J
,
Huang
J
,
Powell
WC
,
Zhang
J
,
Matusik
RJ
, et al
Generation of a prostate epithelial cell-specific Cre transgenic mouse model for tissue-specific gene ablation
.
Mech Dev
2001
;
101
:
61
9
.
27.
Wang
X
,
Finegan
KG
,
Robinson
AC
,
Knowles
L
,
Khosravi-Far
R
,
Hinchliffe
KA
, et al
Activation of extracellular signal-regulated protein kinase 5 downregulates FasL upon osmotic stress
.
Cell Death Differ
2006
;
13
:
2099
108
.
28.
Church
DM
,
Schneider
VA
,
Graves
T
,
Auger
K
,
Cunningham
F
,
Bouk
N
, et al
Modernizing reference genome assemblies
.
PLoS Biol
2011
;
9
:
e1001091
.
29.
Kim
D
,
Pertea
G
,
Trapnell
C
,
Pimentel
H
,
Kelley
R
,
Salzberg
SL
. 
TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions
.
Genome Biol
2013
;
14
:
R36
.
30.
Langmead
B
,
Salzberg
SL
. 
Fast gapped-read alignment with Bowtie 2
.
Nat Methods
2012
;
9
:
357
9
.
31.
Love
MI
,
Huber
W
,
Anders
S
. 
Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2
.
Genome Biol
2014
;
15
:
550
.
32.
Network CGAR
. 
The molecular taxonomy of primary prostate cancer
.
Cell
2015
;
163
:
1011
25
.
33.
Patel
R
,
Gao
M
,
Ahmad
I
,
Fleming
J
,
Singh
LB
,
Rai
TS
, et al
Sprouty2, PTEN, and PP2A interact to regulate prostate cancer progression
.
J Clin Invest
2013
;
123
:
1157
75
.
34.
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
.
35.
Ahmad
I
,
Patel
R
,
Singh
LB
,
Nixon
C
,
Seywright
M
,
Barnetson
RJ
, et al
HER2 overcomes PTEN (loss)-induced senescence to cause aggressive prostate cancer
.
Proc Natl Acad Sci U S A
2011
;
108
:
16392
7
.
36.
Aldinucci
D
,
Colombatti
A
. 
The inflammatory chemokine CCL5 and cancer progression
.
Mediators Inflamm
2014
;
2014
:
292376
.
37.
Liu
M
,
Guo
S
,
Stiles
JK
. 
The emerging role of CXCL10 in cancer (Review)
.
Oncol Lett
2011
;
2
:
583
9
.
38.
Clark
PR
,
Jensen
TJ
,
Kluger
MS
,
Morelock
M
,
Hanidu
A
,
Qi
Z
, et al
MEK5 is activated by shear stress, activates ERK5 and induces KLF4 to modulate TNF responses in human dermal microvascular endothelial cells
.
Microcirculation
2011
;
18
:
102
17
.
39.
Woo
CH
,
Massett
MP
,
Shishido
T
,
Itoh
S
,
Ding
B
,
McClain
C
, et al
ERK5 activation inhibits inflammatory responses via peroxisome proliferator-activated receptor delta (PPARdelta) stimulation
.
J Biol Chem
2006
;
281
:
32164
74
.
40.
Li
L
,
Tatake
RJ
,
Natarajan
K
,
Taba
Y
,
Garin
G
,
Tai
C
, et al
Fluid shear stress inhibits TNF-mediated JNK activation via MEK5-BMK1 in endothelial cells
.
Biochem Biophys Res Commun
2008
;
370
:
159
63
.
41.
Mackesy
DZ
,
Goalstone
ML
. 
Extracellular signal-regulated kinase-5: Novel mediator of insulin and tumor necrosis factor α-stimulated vascular cell adhesion molecule-1 expression in vascular cells
.
J Diabetes
2014
;
6
:
595
602
.
42.
Kim
S
,
Lim
JH
,
Woo
CH
. 
ERK5 inhibition ameliorates pulmonary fibrosis via regulating Smad3 acetylation
.
Am J Pathol
2013
;
183
:
1758
68
.
43.
Wilhelmsen
K
,
Xu
F
,
Farrar
K
,
Tran
A
,
Khakpour
S
,
Sundar
S
, et al
Extracellular signal-regulated kinase 5 promotes acute cellular and systemic inflammation
.
Sci Signal
2015
;
8
:
ra86
.
44.
Zumwalt
TJ
,
Arnold
M
,
Goel
A
,
Boland
CR
. 
Active secretion of CXCL10 and CCL5 from colorectal cancer microenvironments associates with GranzymeB+ CD8+ T-cell infiltration
.
Oncotarget
2015
;
6
:
2981
91
.
45.
Liu
J
,
Li
F
,
Ping
Y
,
Wang
L
,
Chen
X
,
Wang
D
, et al
Local production of the chemokines CCL5 and CXCL10 attracts CD8+ T lymphocytes into esophageal squamous cell carcinoma
.
Oncotarget
2015
;
6
:
24978
89
.
46.
Bedognetti
D
,
Spivey
TL
,
Zhao
Y
,
Uccellini
L
,
Tomei
S
,
Dudley
ME
, et al
CXCR3/CCR5 pathways in metastatic melanoma patients treated with adoptive therapy and interleukin-2
.
Br J Cancer
2013
;
109
:
2412
23
.
47.
Peng
D
,
Kryczek
I
,
Nagarsheth
N
,
Zhao
L
,
Wei
S
,
Wang
W
, et al
Epigenetic silencing of TH1-type chemokines shapes tumour immunity and immunotherapy
.
Nature
2015
;
527
:
249
53
.
48.
Nagpal
ML
,
Davis
J
,
Lin
T
. 
Overexpression of CXCL10 in human prostate LNCaP cells activates its receptor (CXCR3) expression and inhibits cell proliferation
.
Biochim Biophys Acta
2006
;
1762
:
811
8
.
49.
Muthuswamy
R
,
Corman
JM
,
Dahl
K
,
Chatta
GS
,
Kalinski
P
. 
Functional reprogramming of human prostate cancer to promote local attraction of effector CD8(+) T cells
.
Prostate
2016
;
76
:
1095
105
.