Involvement of clusterin expression in the refractory response of pancreatic cancer cells to a MEK inhibitor

Abstract

Constitutive activation of the mitogen‐activated protein kinase (MAPK) signaling pathway is essential for tumorigenesis of pancreatic ductal adenocarcinoma (PDAC). To date, however, almost all clinical trials of inhibitor targeting this pathway have failed to improve the outcome of patients with PDAC. We found that implanted MIA Paca2, a human PDAC cell line sensitive to a MAPK inhibitor, PD0325901, became refractory within a week after treatment. By comparing the expression profiles of MIA Paca2 before and after acquisition of the refractoriness to PD0325901, we identified clusterin (CLU) as a candidate gene involved. CLU was shown to be induced immediately after treatment with PD0325901 or expressed primarily in more than half of PDAC cell lines, enhancing cell viability by escaping from apoptosis. A combination of PD0325901 and CLU downregulation was found to synergistically or additively reduce the proliferation of PDAC cells. In surgically resected PDAC tissues, overexpression of CLU in cancer cells was observed immunohistochemically in approximately half of the cases studied. Collectively, our findings highlight the mechanisms responsible for the rapid refractory response to MEK inhibitor in PDAC cells, suggesting a novel therapeutic strategy that could be applicable to patients with PDAC using inhibitor targeting the MAPK signaling pathway and CLU.

Our findings highlight the mechanisms responsible for the rapid acquisition of resistance to MEK inhibitor in PDAC cells, suggesting a novel therapeutic strategy that could be applicable to patients with PDAC using inhibitor targeting the MAPK signaling pathway and CLU.

Abbreviations

CLU

clusterin

DUSP

dual‐specificity phosphatase

EGFR

epidermal growth factor receptor

IC50

half‐maximal inhibitory concentrations

KD

knockdown.

KO

knockout

MAPK

mitogen‐activated protein kinase

PDAC

pancreatic ductal adenocarcinoma

1. INTRODUCTION

Several reports have indicated that constitutive activation of the mitogen‐activated protein kinase (MAPK) signaling pathway is required for development of pancreatic ductal adenocarcinoma (PDAC). ¹ , ² During pancreatic carcinogenesis, mutation of the KRAS gene is the first and most frequent alteration detected in the precursor cells, ³ , ⁴ allowing activation of downstream molecules in the order BRAF, MEK, and then ERK. However, the activated ERK also induces MAPK phosphatases, such as dual‐specificity phosphatase (DUSP) 4 ⁵ and DUSP6, ⁶ which, in turn, inactivate ERK through their dephosphorylation activity, thus maintaining cellular homeostasis. ⁷ Therefore, disruption of this negative feedback loop between ERK and MAPK phosphatases causes hyperactivation of ERK, resulting in constitutive activation of the MAPK signaling pathway and progression of carcinogenesis. In fact, downregulation of DUSP4 and DUSP6 has been frequently observed in PDAC tissues and shown to enhance ERK activity in PDAC cells. ⁸ , ⁹ These findings suggest that therapies targeting this pathway would be effective for patients with PDAC.

To date, a variety of inhibitor targeting the MAPK signaling pathway have been developed. Among them, MEK inhibitor have been extensively investigated in clinical trials for their efficacy against PDAC. However, none of these recent trials demonstrated any significant effect of MEK inhibitor. ¹⁰ , ¹¹ In fact, the restricted inhibition of MEK by these inhibitor led to bypass signaling or activation of alternative pathways. Therefore, therapies using a MEK inhibitor in combination with other inhibitor targeting signal transduction molecules such as epidermal growth factor receptor (EGFR) or AKT have been attempted, but these showed only a modest therapeutic benefit for patients with PDAC. ¹² , ¹³ , ¹⁴ , ¹⁵ These results suggest that, in addition to the activation of compensatory signaling pathways, other unknown mechanisms that increase cell viability account for the ineffectiveness of therapies using MEK inhibitor.

In the present study, we attempted to clarify how PDAC cells respond to MEK inhibitor by using an orthotropic xenograft model of PDAC and an in vivo imaging system. By comparing the transcriptome profiles of xenografts with and without MEK inhibitor treatment, we obtained candidate genes involved in the refractory response to MEK inhibitor. Among these genes, we identified clusterin (CLU), which was rapidly induced by MEK inhibitor or expressed primarily in PDAC, as a key molecule conferring the refractoriness to MEK inhibitor on PDAC cells. Furthermore, we found that MEK inhibitor combined with CLU knockdown had a synergistic or additive effect in comparison with either treatment alone, suggesting that CLU may have potential as a new therapeutic target in PDAC.

2. MATERIALS AND METHODS

2.1. Cell lines

The human PDAC cell lines AsPC‐1 and BxPC‐3 were obtained from the American Type Culture Collection. PK‐45H, PK‐59, MIA Paca2, PANC‐1, and PK‐1 were obtained from RIKEN BRC through the National Bio‐Resource Project of MEXT. KP‐2, KP‐3L, KP‐4, and SUIT‐2 were obtained from the Japanese Collection of Research Bioresources. SNU‐213 and SNU‐324 were obtained from the Korean Cell Line Bank. TCC‐Pan2, PSN‐1, Sui65, Sui66, Sui67, Sui68, Sui69, Sui70, Sui71, Sui72, Sui73, Sui74, Sui75, and Sui76 were established by Yanagihara et al. at the National Cancer Center (Chiba, Japan). ¹⁶ , ¹⁷ , ¹⁸ All cell lines were authenticated by the providers or by Yanagihara et al. for the short‐tandem repeat profile. After arrival, all cell lines were propagated and frozen immediately, and cells recovered from the frozen stock were used within 10 weeks. MIA Paca2 constitutively expressing the luciferase gene (referred to as MIA/luc) was generated by introducing the pGL4.51[luc2/CMV/Neo] vector (Promega) into MIA Paca2. PD0325901‐resistant MIA/luc cells were established by culturing the parent MIA/luc cells for 1 month with gradually increasing concentrations of PD0325901 (from 20 to 80 nM).

2.2. Xenografts

Eight male NOD‐SCID mice (Charles River Laboratories Japan) aged 8 weeks were injected with 5 × 10⁵ MIA/luc cells suspended in 50 μL RPMI medium into the pancreas under anesthesia. Three weeks after implantation, the mice were divided into two groups, with four mice in each group, and then orally administered PD0325901 (LC Laboratories) or vehicle only, as described previously. ⁹ At 0, 4, 7, 10, 13, 17, 21, and 25 days after starting the treatment, D‐luciferin (150 mg/kg, Xenogen) dissolved in 200 μL PBS was injected intraperitoneally, and tumor growth was monitored using the IVIS Imaging System and Living Image Software (v4.2, Perkin Elmer). For transcriptome analysis, the two groups with four mice in each group were identically injected with MIA/luc cells followed by treatment with PD0325901 or vehicle only. At day 37, tumors were collected and divided into two pieces for preparation of a frozen block and a formalin‐fixed paraffin‐embedded block, respectively. To confirm the role of CLU in PDAC in vivo, the two groups with eight mice in each group were identically injected with MIA/luc cells or CLU‐knockout clone CLU‐KO1 followed by treatment with PD0325901 or vehicle only. At day 21, tumors were collected and weighed.

2.3. Expression microarray analysis

Three hundred nanograms of total RNA extracted from the frozen block of collected tumors using RNeasy Mini kit (Qiagen, Hilden, Germany) was subjected to expression microarray analysis, as described previously. ⁹ All the data obtained in the expression microarray analysis are available on the GEO database, under accession number GSE155549 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE155549). Extraction of the genes upregulated or downregulated by the treatment with PD0325901 was performed using the following filters: upregulated genes, >4000 of raw data in the PD0325901 group, and >4‐fold higher expression in the PD0325901 group than in the vehicle group; downregulated genes, >4000 of raw data in the vehicle group, and <0.25‐fold lower expression in the PD0325901 group than in the vehicle group.

2.4. Quantitative real‐time PCR

Reverse transcription and quantitative real‐time PCR analysis were performed, as described previously. ⁹ All the primer sets for genes of interest were designed by Roche Diagnostics (Indianapolis, IN, USA). All assays were normalized against GAPDH as an internal control.

2.5. RNA interference

Cells were transfected with pooled small interfering RNAs targeting CLU (ON‐TARGETplus SMARTpool, human CLU; Thermo Scientific, referred to as siCLU) or negative control small interfering RNAs (ON‐TARGETplus Non‐targeting Pool; Thermo Scientific, referred to as siCont) at a final concentration of 10 nM, using Lipofectamine RNAiMAX (Invitrogen).

2.6. Proliferation assay

Two thousand cells in 100 μL of medium were cultured in a 96‐well tissue culture plate (Corning) for the indicated periods at 37°C in 5% CO₂. Proliferation was analyzed by MTS assay, as described previously. ¹⁹

2.7. Half‐maximal inhibitory concentrations (IC50)

PD0325901‐resistant MIA/luc cells were established by culturing the parent MIA/luc cells for 1 month with PD0325901 at gradually increasing concentrations from 20 to 80 nM. The resistant cells and the parent cells were cultured in a 96‐well tissue culture plate (Corning) with 0–10 μM PD0325901. After 72 h of treatment, MTT assays were performed, and the IC50 values were calculated.

2.8. Establishment of clusterin knockout MIA Paca2 cell lines

The CLU‐knockout clones CLU‐KO1 and CLU‐KO2 were established using the CRISPR/Cas9 system and the SpCas9(BB)‐2A‐Puro (pX459) V2.0, which was a gift from Feng Zhang (Addgene plasmid #62988; http://n2t.net/addgene:62988; RRID:Addgene_62988), ²⁰ as described previously. ²¹ A set of oligonucleotides for guide RNAs, CLU‐CRISPR‐F: 5′‐CACCGGGGAAGTAAGTACGTCAATA‐3′, and CLU‐CRISPR‐R: 5′‐AAACTATTGACGTACTTACTTCCCC‐3′ was designed on the site of CRISPRdirect (https://crispr.dbcls.jp/).

2.9. Lentivirus expressing clusterin

Using the Human Gateway entry clone (FLJ94503AAAF, NBRC, NITE, Chiba, Japan) containing full‐length cDNA of CLU with the Gateway system (Invitrogen), lentivirus encoding CLU cDNA or lentivirus encoding no cDNA for transient transduction was generated as described previously ⁹ and designated LV‐CLU or LV‐Cont, respectively. Transduction of PDAC cells with these lentiviruses was performed at an optimized multiplicity of infection of 1 with Polybrene (Sigma) at a final concentration of 6 μg/mL. Forty‐eight hours after transduction, the cells were used for the following experiments.

2.10. Cell cycle analysis

Three days after transfection with siCont or siCLU, PANC‐1 cells were harvested and subjected to cell cycle analysis, as described previously. ²¹

2.11. Apoptosis assay

Two days after transfection with siCont or siCLU, induction of apoptosis in PANC‐1 cells was evaluated by measuring the fragmentation of cytoplasmic oligonucleosomes and the activation of Caspase 3/7 using a Cell Death Detection ELISA PLUS Kit (Roche Diagnostics) and a Caspase‐Glo 3/7 Assay System (Promega), respectively.

2.12. Western blotting and antibodies

The lysates from cultured cells or cryosections of frozen tissues were prepared and subject to SDS‐PAGE as described previously. ²² The primary antibodies used in this study were anti‐CLU antibody (Santa Cruz, #sc‐5289, Dallas, TX, USA), anti‐phosphorylated ERK antibody (Thr²⁰²/Tyr²⁰⁴, Cell Signaling Technology, #4370), anti‐ERK antibody (Cell Signaling Technology, #9102), and anti‐GAPDH antibody (Santa Cruz, #sc‐32,233). Detection was performed with ECL Western Blotting Detection Reagents (Amersham Biosciences, Piscataway, NJ, USA). The signal intensity of CLU or GAPDH in each cell line was quantified using ImageJ. Then, the relative expression of CLU was calculated by dividing the quantified value of CLU by that of GAPDH. CLU was considered to show positive expression when the relative expression was >0.01. In addition, induced expression was considered to have occurred when the fold‐induction, calculated by dividing the relative expression of CLU after the PD0325901 treatment by that before the treatment, was >2.

2.13. Patients and tissues

Surgical specimens were obtained from 91 patients who underwent pancreatectomy for PDAC at Oita University Hospital between 1992 and 2010. Among them, 68 patients received adjuvant chemotherapy using gemcitabine or 5‐FU. None of them received neoadjuvant chemotherapy. The tissues were fixed in 10% formalin and embedded in paraffin for histopathology and immunohistochemistry. Histopathologic diagnosis was performed in accordance with the TNM Classification of Malignant Tumors, 8th Edition. ²³

2.14. Immunohistochemistry

Immunohistochemistry was performed as described previously, ²⁴ with some modifications. The antigen retrieval was performed by autoclaving in 0.01 M citrate buffer (pH 6.0) for 10 min at 120°C. CLU expression was detected by incubation with anti‐CLU antibody (Santa Cruz, #sc‐166907) and biotinylated goat anti‐mouse IgG (Nichirei, Tokyo, Japan). Evaluation of CLU expression was performed by two independent pathologists (NH and TU) based on the following criteria. The staining intensity in normal islet cells was used as an internal control for CLU expression in each section. Compared to the internal control, the predominant populations of tumor cells showing equivalent or more intensity were judged as positive (++), and those with less intensity as positive (+). Tumor cells without any CLU signal intensity were judged as negative (−).

2.15. Double‐labeled immunofluorescence

After antigen retrieval, the sections were incubated with a mixture of anti‐CLU antibody and anti‐TP53 antibody (Cell Signaling Technology, #2527) for 16 h at 4°C, and subsequently incubated with a mixture of Alexa Fluor 488‐goat anti‐mouse IgG (Invitrogen, Eugene, OR, USA, #A11001), Alexa Fluor 568‐goat anti‐rabbit IgG (Invitrogen, #A11011) and DAPI Solution (Thermo Scientific) for 1 h at room temperature. The specimens were examined for fluorescence with a LSM5 Pascal confocal laser scanning microscope (Carl Zeiss; Oberkochen, Germany).

2.16. Statistics

The JMP statistical software package (SAS institute, Cary, NC, USA) was used for statistical analyses. Differences at p < 0.05 were considered statistically significant.

3. RESULTS

3.1. MIA/luc becomes refractory to MEK inhibitor shortly after the treatment

To determine the therapeutic effects of MEK inhibitor on the growth of PDAC in vivo, we generated an orthotropic xenograft model using MIA/luc (see Materials and Methods) and an in vivo imaging system (Figure 1A). As shown in Figure 1B,C, the tumors in mice treated with the vehicle grew continually in the abdomen. In contrast, the tumors in mice treated with PD0325901 were shown to be markedly reduced at day 4; thereafter, however, they began to regrow and expand, showing a lag of a few weeks relative to control tumors, suggesting that MIA/luc became refractory to PD0325901 within a week after treatment.

FIGURE 1.

Transplanted MIA/luc acquires refractoriness to MEK inhibitor shortly after treatment. (A) The schedule of treatment with PD0325901 or vehicle alone and in vivo imaging analysis in a MIA/luc‐transplanted xenograft model. (B) Representative images acquired with the IVIS imaging system for MIA/luc‐transplanted mice treated with PD0325901 or vehicle alone on days 0, 4, 7, 10, 13, 17, 21, and 25. (C) Tumor growth curve for MIA/luc‐transplanted mice treated with PD0325901 or vehicle alone. For each group, the photon count detected by the IVIS imaging system on day 0 was set to 1, and the relative intensities at different time points are shown as the mean ± SD. *p < 0.05 by ANOVA and Student's t test.

3.2. Clusterin expression is upregulated in PD0325901‐treated MIA/luc cells

To clarify the mechanisms responsible for the reduced susceptibility to PD0325901 in MIA/luc, we compared the expression profiles of xenograft tumors treated and untreated with PD0325901. We collected the tumors from both groups of mice at day 37, when the tumors in the treated mice had already begun to regrow (Figure 1C). The whole tumors recovered from the treated mice were clearly reduced in size (Figure 2A), and the level of ERK phosphorylation in their cancer cells was found to have drastically decreased (Figure 2B). A comprehensive analysis of gene expression revealed that seven genes (LY6D, ANXA8L2, TUBA4A, KRT16, LFNG, CLU, and JAG2) and five genes (TOX2, EGR1, MYBPH, SPRY2, and LDLR) were significantly upregulated more than 4‐fold and downregulated less than 0.25‐fold in the PD0325901‐treated tumors in comparison with the vehicle‐treated tumors, respectively (Table S1). The altered expressions of these 12 genes, except for MYBPH, were further validated by RT‐PCR (Figures S1A and 1B). Among these genes, we focused on CLU as a candidate gene involved in the refractory response to PD0325901, as the involvement of CLU in chemoresistance has been reported previously. ²⁵ Initially, we determined whether CLU protein was produced in PD0325901‐treated tumors and found that CLU protein expression tended to be restricted to a low level in the vehicle‐treated tumors, whereas it was strongly upregulated in the PD0325901‐treated tumors (Figure 2C). Immunohistochemical analysis revealed that populations of CLU‐positive cells in the vehicle‐treated tumors were small and tended to be focally distributed in the tumor tissue. In contrast, in the PD0325901‐treated tumors, populations of CLU‐positive cells were markedly increased and diffusely distributed (Figure 2D). To further confirm whether the CLU‐expressing cells in the PD0325901‐treated tumors were MIA/luc, we performed double immunostaining for CLU and TP53. The CLU‐positive cells exhibited TP53 positivity in the nucleus (Figure 2E), suggesting that the CLU‐positive cells were MIA/luc, as they have TP 53 mutation (R248W), leading to nuclear accumulation of the mutated TP53.

FIGURE 2.

Clusterin (CLU) is induced in MIA/luc cells by treatment with MEK inhibitor. (A) The gross appearance of xenograft tumors treated with PD0325901 or vehicle alone for 37 days. (B) Inhibitory effect of PD0325901 on phosphorylation level of ERK in MIA/luc tumors. (C) Inactivation of ERK and induction of CLU were observed in the tumor lysate from the PD0325901‐treated xenograft. (D) CLU was extensively expressed in tumors treated with PD0325901 compared to those treated with vehicle. (E) CLU‐positive cells exhibited TP53 positivity in the nucleus, suggesting that the CLU‐positive cells were MIA/luc. (F, G) Induced expression of CLU mRNA and protein in MIA/luc after addition of PD0325901 at 1 μM. (H) The IC50 of PD0325901 was increased in PD0325901‐resistant MIA/luc cells. (I) The expression levels of CLU and the phosphorylation levels of ERK in PD0325901‐resistant MIA/luc cells. Scale bars, 10 mm (A), 100 μm (B), and 200 μm (D). The data in H are presented as the mean ± SD. *p < 0.05 by Student's t test. The data in F, G, H, and I are representative of three independent experiments.

Next, we confirmed the induction of CLU in PD0325901‐treated MIA/luc in vitro. As shown in Figure 2F, expression of CLU mRNA was immediately induced in the cultured MIA/luc within 6 h after the addition of PD0325901. Production of CLU protein was also observed within 24 h after the start of incubation with PD0325901 and increased gradually until 72 h (Figure 2G).

Furthermore, we established PD0325901‐resistant MIA/luc cells by culturing the parent MIA/luc cells with gradually increasing concentrations of PD for 1 month. The IC50 of PD0325901 in the resistant cells was significantly increased, and it was markedly decreased by the downregulation of CLU (Figure 2H). The resistant cells were found to express a modest level of CLU even without the PD0325901 treatment, and the expression level increased more after PD treatment compared to that in parent cells (Figure 2I). Interestingly, the phosphorylation level of ERK in the resistant cells was not completely suppressed by PD0325901 treatment; However, it was markedly reduced by downregulation of CLU (Figure 2I), suggesting that the stably induced CLU by long‐term administration of PD0325901 is involved in refractoriness to PD0325901 treatment.

3.3. Clusterin induction is required for maintenance of refractoriness to MEK inhibitor in MIA/luc cells

Next, we investigated whether PD0325901‐induced CLU is functionally involved in the reduced sensitivity to PD0325901 in MIA/luc cells. As shown in Figure 3A, PD0325901 treatment alone or CLU knockdown (CLU‐KD) alone had little or no effect on the proliferation of MIA/luc, respectively. In contrast, a combination of PD0325901 treatment and CLU‐KD inhibited cell proliferation synergistically, indicating that the CLU induced after PD0325901 treatment indeed enhanced cell viability. To validate the role of CLU, we generated two CLU‐deficient MIA/luc cells, CLU‐KO1 and CLU‐KO2. They showed no induction of CLU after PD0325901 treatment, as expected (Figure 3B). Under normal culture conditions without PD0325901, CLU‐KO1 proliferated similarly to their parent cells, MIA/luc; however, in the presence of PD0325901, the proliferation activity of CLU‐KO1 was much more reduced in comparison with that of MIA/luc (Figure 3C). To further confirm whether the suppressed proliferation of CLU‐KO1 was due to the lack of CLU induction after PD0325901 treatment, CLU was transduced into CLU‐KO1 using lentivirus. As shown in Figure 3D,E, the proliferation activity of CLU‐KO1 treated with PD0325901 was found to be completely rescued by exogenous expression of CLU, reaching levels similar to that of MIA/luc. In contrast, the introduction of CLU had no effect on the growth of both MIA/luc and CLU‐KO1 cells in the absence of PD0325901 (Figure S2). Finally, the role of CLU in PDAC was determined using an orthotopic xenograft model (Figure 3F,G). In the mice receiving vehicle only, both CLU‐KO1 and MIA/luc grew similarly. In contrast, in the mice treated with PD0325901, tumor growth of CLU‐KO1 was significantly inhibited compared with that of MIA/luc. Taken together, these results confirmed that CLU induced after PD0325901 treatment is responsible for the acquisition of refractoriness to PD0325901 in MIA/luc.

FIGURE 3.

Induced clusterin (CLU) is necessary for acquisition of the refractoriness to MEK inhibitor in MIA/Luc cells. (A) Knockdown of CLU synergistically suppressed the proliferation of MIA/luc cells by PD0325901. (B) Two CLU‐knockout MIA/luc cell lines, CLU‐KO1, and CLU‐KO2, were established. (C) The growth of CLU‐KO1 was similar to that of MIA/luc in the absence of PD0325901 (left); however, it was more suppressed in comparison with MIA/luc in the presence of PD0325901 (right). (D, E) The loss of refractoriness to PD0325901 in CLU‐KO1 was recovered by exogenous expression of CLU using lentivirus. (F) The gross appearance of MIA/luc and CLU‐KO1 xenograft tumors treated with PD0325901 or vehicle alone for 28 days. Scale bars, 10 mm. (G) Inhibitory effect of PD0325901 on tumor growth was significantly higher in CLU‐KO1 than MIA/luc. The data in A, C, E, and G are presented as the mean ± SD. ns, not significant, *p < 0.05 by ANOVA and Student's t test. The data in A–E are representative of three independent experiments.

3.4. Clusterin induction suppresses apoptosis in pancreatic ductal adenocarcinoma cells

Next, we attempted to clarify how CLU endows PDAC cells with refractoriness to PD0325901. For further studies, PANC‐1 was selected because our previous transcriptome data had shown that it was one of the PDAC cell lines expressing CLU primarily under normal culture conditions. ⁹ Proliferation of PANC‐1 was found not to be altered upon treatment with PD0325901 alone; however, it was strongly suppressed by CLU‐KD alone (Figure 4A), suggesting that PANC‐1 is dependent on CLU expression for survival and proliferation. To clarify the mechanism of growth inhibition resulting from downregulation of CLU in PANC‐1, we performed flow cytometry analysis to determine the effects of CLU downregulation on cell cycle progression. As shown in Figure 4B,C, the cell population in sub‐G1 phase was markedly increased after CLU‐KD. In addition, DNA fragmentation and Caspase 3/7 activation were shown to be accelerated in PANC‐1 by downregulation of CLU (Figure 4D,E), indicating the induction of apoptosis. These results suggested that CLU, which is not only expressed primarily in PDAC cells but also induced secondarily after PD0325901 treatment, may play an important functional role in escape from apoptosis.

FIGURE 4.

Downregulation of clusterin (CLU) induces apoptosis in PANC‐1 cells. (A) The growth of PANC‐1 cells, which primarily express CLU, was reduced by knockdown of CLU but not by inhibition of MEK. (B, C) Knockdown of CLU in PANC‐1 cells increased the sub G1 fraction of the cell population. (D, E) Knockdown of CLU accelerated DNA fragmentation and Caspase 3/7 activation in PANC‐1, indicating the induction of apoptosis. The data in A, D, and E are presented as the mean ± SD. *p < 0.05 by ANOVA. All data are representative of three independent experiments.

3.5. Clusterin is commonly inducible in pancreatic ductal adenocarcinoma cells

To address whether CLU induction by PD0325901 treatment was detectable in PDAC cell lines other than MIA/luc, 27 PDAC cell lines, including MIA/luc, were analyzed (Figure 5A). The level of CLU production in each cell line was quantified and normalized against that of GAPDH (Table S2). The results in Table S2 are further summarized in Figure 5B. Unexpectedly, CLU production was primarily detectable in 15 of the 27 cell lines (55.6%). Of these 15 cell lines, 5 (18.5%) exhibited further induction of CLU expression upon treatment with PD0325901. In contrast, of the 12 cell lines (44.4%) not showing primary expression of CLU, 7 (25.9%) revealed induced expression of CLU. Collectively, only 5 of the 27 cell lines (18.5%) failed to show both primary expression and induced expression of CLU (Figure 5B). These results suggested that CLU tends to be expressed primarily and is detectable in more than half of PDAC cell lines and that CLU is inducible by PD0325901 treatment in some PDAC cell lines regardless of their primary expression levels.

FIGURE 5.

Treatment with MEK inhibitor induces clusterin (CLU) in a large proportion of pancreatic ductal adenocarcinoma (PDAC) cell lines. (A) The expression level of CLU protein in 27 PDAC cell lines with or without MEK inhibition by incubation with 1 μM PD0325901 for 48 h. (B) The primary expression of CLU and the induction of CLU by PD0325901 treatment in 27 PDAC cell lines were evaluated using the data in A (see Materials and Methods). All the data are representative of two independent experiments.

3.6. Synergistic or additive effect of combined treatment with PD0325901 and clusterin knockdown

Next, we analyzed 21 PDAC cell lines including MIA/luc to determine whether treatment with PD0325901 and CLU‐KD might exert suppressive effects on cell proliferation (Figure S3). The data are summarized in Table 1, which depicts the effectiveness of PD0325901 alone or CLU‐KD alone in terms of the proliferation of PDAC cells relative to that of cells treated with only control siRNAs. The effectiveness of PD0325901 and CLU‐KD in combination was judged in terms of the proliferation of PDAC cells relative to that of cells treated with PD0325901 alone or CLU‐KD alone, respectively. Concerning the growth suppression elicited by treatment with PD0325901 alone or CLU‐KD alone, the effect was significant in 17 of 21 cell lines (81.0%) or in 8 of 21 cell lines (38.1%), respectively. Additionally, in 7 (33.3%) or 6 (28.6%) of the 21 cell lines, cell growth was suppressed synergistically or additively, respectively, by combined treatment with both PD0325901 and CLU‐KD than by treatment with either alone.

TABLE 1.

The effects of MEK inhibition and CLU downregulation on cell growth in pancreatic ductal adenocarcinoma cell lines.

Cell line	Growth suppression a
	PD alone	CLU‐KD alone	Combination
MIA/luc	Yes	No	Yes
AsPC1	Yes	No	Yes
BxPC3	Yes	No	Yes
KP3L	Yes	Yes	Yes
PANC1	No	Yes	No
PK1	Yes	Yes	Yes
PK59	No	No	Yes
TCCPan2	Yes	No	No
PSN1	Yes	No	No
Sui65	Yes	No	Yes
Sui66	Yes	No	No
Sui67	Yes	No	Yes
Sui68	Yes	No	Yes
Sui69	Yes	Yes	Yes
Sui70	Yes	Yes	Yes
Sui71	Yes	No	No
Sui72	No	No	No
Sui73	Yes	Yes	Yes
Sui74	Yes	No	No
Sui75	No	Yes	No
Sui76	Yes	Yes	Yes
Number of Yes (%)	17 (81.0%)	8 (38.1%)	13 (61.9%)

Note: CLU, clusterin; CLU‐KD, knockdown of CLU with siRNA PD, PD0325901.

^a The judgments on the growth suppression of PDAC cell lines by PD, CLU‐KD, or a combination of the both were based on the results in Figure S2. “Yes” denotes statistically significant growth suppression by the treatment and is shaded.

Next, we compared the growth‐suppressive effect of PD0325901 alone with that of both PD0325901 and CLU‐KD together and found that cell growth was suppressed more significantly by the combined treatment than by PD0325901 alone (Figure 6A), suggesting that combined treatment with both PD0325901 and CLU‐KD synergistically or additively suppressed cell proliferation.

FIGURE 6.

Clusterin (CLU) expression is correlated with efficacy of CLU knockdown (KD) treatment and is immunohistochemically detectable in pancreatic ductal adenocarcinoma (PDAC) tissues. (A) The growth‐suppressive effect of treatment with PD0325901 alone was compared to that of combined treatment with PD0325901 and CLU‐KD in each PDAC cell line. The growth activity of each cell line treated with only control siRNA was set to 100%. (B) The proportion of PDAC cell lines with sensitivity to CLU‐KD treatment was compared between 18 PDAC cell lines showing either or both of primary CLU expression and CLU inducibility and three cell lines showing neither of them. (C) CLU expression was determined immunohistochemically in normal pancreas (a), ADM lesions (b), poorly differentiated PDAC (c, d), and well‐differentiated PDAC (e, f). The blue arrows in b and the red triangles in c–f show the ADM and the cancer cells, respectively. The intensity of CLU expression relative to the internal control (a) was judged as negative in c, positive (+) in d and e, and positive (++) in f. Scale bars, 50 μm. (D) Primary expression of CLU in cancer cells was observed in 50.6% of the PDAC cases. (E) Survival curves of patients with CLU‐positive PDAC and those with CLU‐negative PDAC were calculated using the Kaplan–Meier method, and differences were analyzed with the log‐rank test.

3.7. Clusterin detection by immunohistochemistry is useful for the prediction of treatment sensitivity

Next, we compared the efficacy of CLU‐KD on cell proliferation in PDAC cell lines showing either or both primary CLU expression and CLU inducibility by treatment with PD0325901 with that in PDAC cell lines showing neither primary CLU expression nor CLU inducibility (Figure 6B). In 18 cell lines showing either or both primary CLU expression and CLU inducibility, 15 cell lines (83.3%) were sensitive to CLU‐KD treatment. In contrast, all of three cell lines showing neither primary CLU expression nor CLU inducibility were refractory to treatment with CLU‐KD, indicating that the efficacy of CLU‐KD treatment is correlated with the expression status of CLU in PDAC cell lines.

Based on these results, we hypothesized that the detection of CLU in human PDAC tissues might predict the clinical efficacy of CLU‐targeted therapy. Therefore, we attempted to determine CLU expression by immunohistochemistry in surgically resected PDAC tissues. As shown in Figure 6C, in non‐cancerous regions, CLU expression was observed only in islet cells (Figure 6C‐a). Interestingly, however, acinar cells showing ductal architecture (i.e., so‐called acinar‐to‐ductal metaplasia [ADM] known to be a precancerous lesion ²⁶ and frequently detectable in cancer‐associated obstructive pancreatitis) were found to express CLU on their luminal surface (Figure 6C‐b). In tumor tissue, cancer cells with poor differentiation tended to show faint (Figure 6C‐c) or diffuse CLU expression in the cytoplasm (Figure 6C‐d), whereas those that were moderate to well differentiated showed higher expression on the luminal surface, similar to the ADM lesions (Figure 6C‐e,f). Among the 91 PDAC cases, 45 (49.4%) were CLU‐negative, whereas 40 (44.0%) and 6 (6.6%) cases exhibited CLU (+) and (++) positivity, respectively (Figure 6D). Interestingly, the patients with CLU‐positive tumors exhibited a better prognosis than those with CLU‐negative tumors (Figure 6E). Among the clinicopathologic factors analyzed, only differentiation status was found to be significantly correlated with CLU expression status (Table S3). These results suggest that at least half of patients with PDAC would potentially benefit from CLU‐targeted therapy and that those with poorly differentiated PDAC might potentially be refractory to such therapy. However, we cannot rule out the possibility that, in these latter cases, CLU might be induced after treatment with MEK inhibitor, thereby rendering CLU‐targeted therapy effective.

4. DISCUSSION

Clusterin is associated with resistance to anti‐tumor treatments, including androgen ablation therapy for prostate cancer, radiotherapy for breast, prostate, bladder, and cervical cancers, and chemotherapy for breast, prostate, lung, bladder, renal, endometrial, and esophageal cancers. ²⁷ Our findings showed that CLU plays an important role in the acquisition of early refractoriness to MEK inhibitor in PDAC. Furthermore, to date, despite several clinical trials of MEK inhibitor in patients with PDAC, none have demonstrated significant clinical benefits, ¹⁰ , ¹¹ suggesting that MEK inhibitor monotherapy would be insufficient to suppress the growth of PDAC. In this study, we found that the acquisition of early refractoriness to MEK inhibitor in PDAC might be caused by induction of CLU after treatment with MEK inhibitor, suggesting that CLU may be key to the development of a new therapeutic strategy.

We found that induction of CLU enhanced the proliferation of PDAC cells by increasing their resistance to apoptosis. Interestingly, CLU expression was observed in a small number of MIA/luc cells in xenograft tumors even before treatment with the MEK inhibitor (see Figure 2D), suggesting that CLU is inducible under specific microenvironmental conditions even without MEK inhibitor treatment. In contrast, the population of cells expressing CLU was markedly increased shortly after MEK inhibitor treatment, suggesting that CLU might be induced in response to MEK inhibitor rather than being expressed in newly generated clones harboring additional genomic mutations. Our finding that CLU functions as an anti‐apoptosis protein in MEK inhibitor‐treated PDAC is consistent with the results of previous studies. ²⁸ , ²⁹ Thus, the anti‐apoptosis mechanism exerted by CLU in PDAC warrants further investigation.

In a series of experiments to determine the therapeutic efficacy of MEK inhibitor, we found that approximately 60% of 21 PDAC cell lines showed synergistic or additive suppression of cell growth using a combination of MEK inhibitor and CLU knockdown, in comparison with either treatment alone. Considering that clinical trials of monotherapy with MEK inhibitor have failed to show significant efficacy in patients with PDAC, ¹⁰ , ¹¹ it is important to determine the therapeutic efficacy of MEK inhibitor and CLU‐targeted therapy used in combination. Although phase III trials of CLU‐targeted therapy using antisense oligonucleotides against CLU have already been performed in patients with prostate cancer, non‐small cell lung cancer, and breast cancer, the clinical outcomes were no better than those achieved with conventional therapies involving cytotoxic anti‐cancer drugs, such as docetaxel and gemcitabine. ³⁰ , ³¹ Based on these findings, we propose that combined treatment with MEK inhibitor and CLU‐targeted therapy might be worth testing in the near future as a priority goal for PDAC patients.

Immunohistochemical analyses of PDAC tissues surgically resected without receiving any preoperative treatment revealed that CLU was expressed in the cancer cells in more than half of the cases examined. As knockdown of CLU suppressed the growth of PDAC cell lines with a high level of primary CLU expression or that showed significant induction of CLU after treatment with MEK inhibitor (see Figure 6B), we propose that CLU‐targeted therapy could be applicable to patients with PDAC showing CLU expression. However, as shown in the present in vitro study, CLU might be inducible in some PDAC cases lacking primary CLU expression, thus possibly rendering them treatable with CLU‐targeted therapy. Therefore, to achieve personalized medicine including CLU‐targeted therapy for patients with PDAC, we need to develop a novel strategy for screening of patients who would potentially benefit from the therapy. For this purpose, a tissue culture system for individual patients might be useful for predicting the efficacy of the therapy. Currently, it is possible to generate a PDAC‐derived organoid culture from a portion of a biopsy specimen collected preoperatively for pathological diagnosis. ³² Use of this culture system should make it possible to determine the levels of CLU expression before and after treatment with MEK inhibitor in vitro and then utilize the results to decide whether a combination of MEK inhibitor and CLU‐targeted therapy would be feasible. Further research is needed to investigate this novel combination therapy for patients with PDAC.

FUNDING INFORMATION

This study was supported by Japan Society for the Promotion of Science (JSPS) KAKENHI Grant Numbers 18K06992 and the Research Fund at the Discretion of the President, Oita University.

CONFLICT OF INTEREST STATEMENT

Dr MT and Dr MA are editorial board members of Cancer Science. The other authors have no conflict of interest.

ETHICS STATEMENT

Approval of the research protocol by an Institutional Reviewer Board: The study was approved by the Oita University Ethics Committee (Approval number: 1724).

Informed consent: Written informed consent was obtained from all patients.

Registry and the registration no. of the study/trial: N/A.

Animal studies: All animal experiments were approved by the Committee on the Use of Live Animals for Teaching and Research (Approval number: K008001) and conducted in accordance with the Animal Care and Use Committee guidelines of Oita University.

Supporting information

Figure S1 Validation of microarray data by quantitative real‐time PCR. Seven upregulated genes (A) and five downregulated genes (B) identified by the transcriptome data were validated for their expression by quantitative real‐time PCR. For each gene, the expression level relative to GAPDH was compared between vehicle (veh, n = 4) or PD0325901 (PD, n = 4) treatment. p‐values were calculated using Student’s t test.

Figure S2 The effects of exogenous expression of CLU on cell proliferation in MIA/luc and CLU‐KO cells. CLU was transduced into MIA/luc and CLU‐KO1 cells using lentivirus. Forty‐eight hours after transduction, the cells were harvested and replated in a 96‐well plate (3000 cells/well in 100 μL medium, n = 6) (T = 0 day). The proliferation activities were determined on T = 0 and T = 2 day, and presented as fold change values, T = 2 day/T = 0 day. The data are represented as the mean ± SD. ns, not significant; *p < 0.05 by ANOVA.

Figure S3 The synergistic or additive effects of CLU downregulation and MEK inhibition on cell proliferation. The therapeutic effects of PD0325901 and CLU downregulation on cell growth of 21 PDAC cell lines were determined. The cells (2000 cells/well) were transfected with pooled siCont or siCLU at 10 nM, then further incubated for 3 days with or without PD0325901 at 1 μM. In each cell line, the relative proliferation compared to the siCont‐transfected cells without PD0325901 is shown. The data are represented as the mean ± SD. ni, not inhibited; ns, not significant; *p < 0.05 by ANOVA. All data are representative of two independent experiments.

Table S1 The genes upregulated or downregulated by PD0325901 treatment.

Table S2 Production levels of CLU in PDAC cell lines before and after PD0325901 treatment.

Table S3 Correlation of clinicopathological factors with CLU expression.

Amada K, Hijiya N, Ikarimoto S, et al. Involvement of clusterin expression in the refractory response of pancreatic cancer cells to a MEK inhibitor. Cancer Sci. 2023;114:2189‐2202. doi: 10.1111/cas.15735