Differential processing of let-7a precursors influences RRM2 expression and chemosensitivity in pancreatic cancer: role of LIN-28 and SET oncoprotein

Abstract

Overexpression of ribonucleotide reductase subunit M2 (RRM2), involved in deoxyribonucleotide synthesis, drives the chemoresistance of pancreatic cancer to nucleoside analogs (e.g., gemcitabine). While silencing RRM2 by synthetic means has shown promise in reducing chemoresistance, targeting endogenous molecules, especially microRNAs (miRNAs), to advance chemotherapeutic outcomes has been poorly explored. Based on computational predictions, we hypothesized that the let-7 tumor suppressor miRNAs will inhibit RRM2-mediated gemcitabine chemoresistance in pancreatic cancer. Reduced expression of the majority of let-7 miRNAs with an inverse relationship to RRM2 expression was identified in innately gemcitabine-resistant pancreatic cancer cell lines. Direct binding of let-7 miRNAs to the 3' UTR of RRM2 transcripts identified post-transcriptional regulation of RRM2 influencing gemcitabine chemosensitivity. Intriguingly, overexpression of human precursor-let-7 miRNAs led to differential RRM2 expression and chemosensitivity responses in a poorly differentiated pancreatic cancer cell line, MIA PaCa-2. Defective processing of let-7a precursors to mature forms, in part, explained the discrepancies observed with let-7a expressional outcomes. Consistently, the ratios of mature to precursor let-7a were progressively reduced in gemcitabine-sensitive L3.6pl and Capan-1 cell lines induced to acquire gemcitabine resistance. Besides known regulators of let-7 biogenesis (e.g., LIN-28), short hairpin RNA library screening identified several novel RNA binding proteins, including the SET oncoprotein, to differentially impact let-7 biogenesis and chemosensitivity in gemcitabine-sensitive versus -resistant pancreatic cancer cells. Further, LIN-28 and SET knockdown in the cells led to profound reductions in cellular proliferation and colony-formation capacities. Finally, defective processing of let-7a precursors with a positive correlation to RRM2 overexpression was identified in patient-derived pancreatic ductal adenocarcinoma (PDAC) tissues. These data demonstrate an intricate post-transcriptional regulation of RRM2 and chemosensitivity by let-7a and that the manipulation of regulatory proteins involved in let-7a transcription/processing may provide a mechanism for improving chemotherapeutic and/or tumor growth control responses in pancreatic cancer.

Figure 1. An inverse relation of RRM2 and let-7 in human pancreatic cancer cells.

A, RRM2 mRNA expression in pancreatic cancer cell lines relative to expression identified in HPDE. Columns, mean of triplicate; bars, SD. n = 3. B, Western blotting analysis of RRM2 (∼45 kDa) and β-actin (45 kDa) in whole cell lysates of HPDE and pancreatic cancer cell lines. C, Expression of let- 7 family members in pancreatic cancer cell lines relative to expression in HPDE. Columns, mean of triplicate; bars, SD. n = 3; *P<0.05. D, RRM2 is a direct target of let-7. 293TA and MIA PaCa-2 cells were virally infected for expression of precursors of let-7a-1, let-7a-2, let-7a-3, let-7b, and miR-214 (negative control) and subsequently transfected with a RRM2 3′ UTR luciferase reporter construct. Luciferase activities measured 36 h after transfection (normalized relative to renilla activity) were plotted. Columns, mean of triplicate; bars, SD. n = 3. *p<0.05, **p<0.01.

Figure 2. Differential RRM2 expression and gemcitabine chemosensitization by let-7 precursors in MIA PaCa-2.

A, Western blotting analysis of RRM2 (∼45 kDa) and β-actin (45 kDA) in whole cell lysates of MIA PaCa-2 overexpressing precursors of let-7 family members. Ratios of RRM2 to β-actin band intensities (normalized to control) from three experiments are indicated (top). Asterisks indicate significant reductions (p<0.05) in RRM2 levels compared with control. B, Immunocytochemical detection of RRM2 in exponentially growing MIA PaCa-2 overexpressing pre-let-7 family members. Original magnification, x20. C, MIA PaCa-2 cells stably overexpressing pre-let-7 family members (red) or vector alone (blue) were treated with gemcitabine (0.1 nM to 100 µM), and percent inhibition of cellular proliferation was measured using an MTT assay. Points, mean of triplicate; bars, SE. n = 3. Gemcitabine IC₅₀ estimations indicated (parentheses).

Figure 3. Defective processing of pre-let-7a-1, but not pre-let-7a-3, into let-7a in MIA PaCa-2.

A, Relative expression of mature forms of let-7 in MIA PaCa-2 stably expressing pre-let-7 family members. Columns, mean of triplicate; bars, SD. n = 3. B, Relative expression of precursor (filled bars; right axis) and mature (open bars; left axis) let-7 forms in pancreatic cancer cells transiently expressing let-7a precursors. Columns, mean of triplicate; bars, SD. n = 3. C, Schematic representation of the structures of pre-let-7a-1 and pre-let-7a-3. Sequences of mature and passenger let-7a strands within the precursors are boxed in continuous or broken lines, respectively. D, Lack of complete cleavage of pre-let-7a-1-GFP mRNA in MIA PaCa-2 cells. MIA PaCa-2 cells were transiently transfected with either pmirGLO-G-Fud (control), pmirGLO-GFP-pre-let-7a-1, or pre-pmirGLO-GFP-pre-let-7b constructs, and GFP fluorescence was captured. Original magnification, x20. *p<0.05, **p<0.01, ***p<0.001.

Figure 4. Acquired gemcitabine resistance is accompanied with RRM2 overexpression and defective let-7a precursor processing.

A, Western blotting analysis of hENT1 (∼50–55 kDa), hENT2 (50 kDa), hCNT1 (72 kDa), hCNT3 (77 kDa), CDA (55 kDa), dCK (30 kDa), RRM1 (94 kDa), RRM2 (45 kDa), and β-actin (45 kDA) levels in whole cell lysates of Capan-1 and Capan-1-GR cells. B, RRM2 protein increased in a gemcitabine dose-dependent fashion in Capan-1-GR cells. C, Western blotting analysis of RRM2 (∼45 kDa) and β-actin (45 kDA) levels in cells with acquired gemcitabine resistance. Ratios of RRM2 to β-actin band intensities (normalized to untreated cells) from three experiments are indicated (top). Asterisk indicates significantly higher RRM2 expression in gemcitabine-resistant cells (p<0.05) compared with untreated cells. D and E, Relative expression of precursor and mature let-7a in Capan-1 (D) and L3.6pl (E) cells induced to acquire gemcitabine resistance. Columns, mean of triplicate; bars, SD. n = 3. F, Differential miRNA expression in Capan-1-GR compared with Capan-1. Putative RRM2-modulating miRNAs and their extent of reduction in Capan-1-GR cells are shown (right).

Figure 5. Screening for putative let-7a biogenesis regulators in MIA PaCa-2.

A, Western blotting analysis of LIN-28A (26 kDa), hnRNP-A1 (36 kDa), KHSRP (∼66–70 kDa), and β-actin (45 kDA) in whole cell lysates of normal and cancerous pancreatic cells. Asterisk indicates significantly higher LIN-28 expression (p<0.05) compared with that identified in HPDE from three experiments. B–E, Screening for regulators of let-7a biogenesis. MIA PaCa-2 cells were infected with lentiviruses harboring shRNAs for putative RNA processing proteins, and the relative levels of mature let-7a (B) and three precursors of let-7a (C–E) in stable clones were plotted. Data are mean±SD. n = 3. Arrows indicate the candidates showing reductions in precursor let-7a with a concomitant increase in mature let-7a.

Figure 6. Silencing of LIN-28 and SET showed differential let-7 biogenesis, growth, and gemcitabine chemosensitivity effects.

A, Overexpression of SET in pancreatic cancer. Western blotting analysis of SET (39 kDa) and β-actin (45 kDa) in HPDE and pancreatic cancer cell lines. B, Western blotting analysis of SET (39 kDa) or LIN-28 (26 kDa) in MIA PaCa-2 expressing shRNAs for the respective proteins. Mock-transduced cells were used for comparison, and β-actin (45 kDa) was used as a loading control. C and D, Relative expression of mature let-7 members in LIN-28- (filled bars) or SET-silenced (open bars) MIA PaCa- 2 (C) and L3.6pl (D). Columns, mean of triplicate; bars, SD. n = 3. E and F, Relative cellular proliferation (G) and colony-formation (H) capacities of control and LIN-28- or SET-silenced MIA PaCa-2 compared with mock-transduced MIA PaCa-2. G and H, 3×10³ control and LIN-28- or SET-silenced MIA PaCa-2 (E) and L3.6pl (F) were treated with gemcitabine (0.1 nM to 100 µM), and percent inhibition of cellular proliferation measured by an MTT was plotted. Points, mean of triplicate; bars, SD. n = 3. *p<0.05, **p<0.01, ***p<0.001.

Figure 7. Defective processing of let-7a precursors and RRM2 overexpression in resected human PDAC tissues.

A and B, Relative expression of primary let-7a transcripts (A) and mature let-7a (B) in 2 normal pancreatic tissues and 10 PDAC samples representing various tumor stages. C, Ratios of mature to precursor let-7a transcripts calculated from data points in A and B. D, Western blotting analysis of RRM2 (45 kDa) in total lysates (50 µg) of 6 matched normal-PDAC pairs. Ratios of RRM2 band intensities in PDACs compared to matched normal tissues indicated (top). Asterisk indicates significantly higher RRM2 expression in PDAC tissues (p<0.05) compared with matched normal tissues. E and F, Relative expression of primary let-7a transcripts (E) and mature let-7a (F) in matched normal-PDAC pairs. G, Ratios of mature to precursor let-7a transcripts calculated from data points in E and F. Asterisk indicates significantly lower mature to precursor let-7a ratio in PDAC tissues (p<0.05) compared with matched normal tissues.