Keratin 17 enhances tumor aggression in a murine orthotopic model of PDAC and detergent-soluble Keratin 17 is negatively prognostic in human and murine PDACs.

A. Immunoblot of lentiviral-generated KPC cells transduced to express K17 or empty vector (EV) for orthotopic implantation into the pancreata of immunocompetent C57BL/6 mice. B. Immunofluorescent labeling of K17 protein in lentiviral-generated clonal KPC cells transduced to express K17 or EV. K17 is shown in green, and the nucleus in blue (DAPI). Scale = 100μm. C. Kaplan-Meier curves show the overall survival in mice implanted with EV or K17-expressing PDACs. Significance determined using Log-rank Mantel–Cox test. D. Tumor size measured by 3D ultrasound. K17 tumors were on average larger than EV tumors, determined by ultrasound imaging. Representative IHC images indicating that K17 tumor cells implanted into mice retain expression of K17 in vivo. Scale = 20μm. Data are mean ± s.d, n = 8 mice per group. Student’s t-test. E. Percent of tumor necrosis in tumors. K17 overexpression increases the extent of tumor necrosis in vivo, representative H&E images. Scale = 200μm. Data are mean ± s.d, n = 5 mice per group. Student’s t-test. F. Percentage of hepatic metastases. K17 overexpression increases metastasis, representative H&E images. Scale = 50μm. Data are percentages, n = 5 mice per group. G. Western blot of detergent soluble (S) and detergent-insoluble (I) fractions of K17 from isogenic tumors with K17 overexpression. Molecular weight shift of K17 may be due to the hyperphosphorylation in the detergent-soluble fraction. n = 3. H. Kaplan-Meier curves demonstrate overall survival of mice with absent K17 (EV), low detergent-soluble-K17 and high detergent-soluble-K17 PDACs. Significance determined using Log-rank Mantel–Cox test. I. Representative images of K17 IHC of human PDACs, demonstrating that K17 disassembly is independent of K17 IHC score. Scale = 50μm. J. Representative blots of detergent soluble (S) and detergent-insoluble (I) fractions from human PDACs, showing a range of K17 disassembly. n = 3. K. Kaplan-Meier curves show the overall survival of patients with low detergent-soluble-K17 and high detergent-soluble-K17 PDACs. Significance determined using Log-rank Mantel–Cox test. L. Kaplan-Meier curves of PDAC cases from the Clinical Proteomic Tumor Analysis Consortium (CPTAC) stratified by either low or high K17-phosphorylation status. Significance determined using Log-rank Mantel–Cox test. s.d, standard deviation; n, number of repetitions. Coomassie Brilliant Blue (CBB). Hazard Ratio (HR). *P< 0.05; **P< 0.01; ***P< 0.001.

K17 is differentially phosphorylated in its detergent-soluble state, and N-terminal serine phosphorylation promotes disassembly.

A. Western blot of human PDAC samples showing three high-K17 expressing tumors chosen for mass spectrometry (indicated with a red *). Equal micrograms of protein were loaded for each sample. B. Illustration of results from liquid chromatography mass spectroscopy of the three PDACs from (A), which identified serine phosphorylation on the N-terminal residues indicated (head of the keratin). C. Alignment of K17 polypeptide in different vertebrate species show identification of a conserved serine hotspot (Ser10-13). D. Western blot depicting detergent-soluble and detergent-insoluble K17 after treatment with okadaic acid (OA). Treatment leads to disassembly of K17 in L3.6 and BxPC3 PDAC cell lines, compared to vehicle control (Veh), shown by fold change (FC). Data are mean ± s.d, n = 3 independent repetitions. E. Western blot depicting detergent-soluble and detergent-insoluble K17 in cytoplasm (Cyt.) and nuclei (Nuc.) after treatment with okadaic acid (OA). K17 is preferably localized to the nucleus upon treatment of PDAC cells with okadaic acid, as shown by fold change of total K17 level. Data are mean ± s.d, n = 3 independent repetitions. F. Subcellular localization of K17 after okadaic acid (OA) treatment by immunofluorescence in SW13 vim-cells. Scale =5μm. 200 cells were quantified in each condition. G. Illustration of K17 edited by CRISPR-Cas9 technique used to generate K17 knockout (KO) in L3.6 cells. Western blot validation of the knockout in three L3.6 cell clones and immunofluorescence images of L3.6 expressing K17 and the KO clone. Filament form of K17 is shown in green, and the nucleus was stained by DAPI, in blue. Scale = 100 μm. H. Western blot depicting total K17 expression following lentiviral transfection of K17 phospho-mutants into L3.6 clone cells. I. Percent Detergent-soluble K17 in functional phospho-K17 mutants in original knockout cells. S10-13 hotspot is sufficient to regulate K17 disassembly, as shown by fold change in detergent-soluble K17 level. Data are mean ± s.d, n = 3 independent repetitions. J. Immunofluorescence of functional phospho-K17 mutants shows that L3.6 gain-of-function S10-13D and S1013-E K17 mutants have a distinctly punctate, not filamentous, cytoplasmic K17 organizational pattern. Filament and punctate forms of K17 are shown in green, and the nucleus was stained by DAPI, in blue. Scale = 20 μm. s.d, standard deviation; n, number of repetitions. Coomassie Brilliant Blue (CBB). *P< 0.05; **P< 0.01; ***P< 0.001.

K17 is phosphorylated by a PKC-mediated signaling cascade.

A. Bioinformatic prediction from NetPhods and GPS 5.0, for serine, threonine, or tyrosine phosphorylation in K17 indicates that the S10-13 region is likely to be phosphorylated by a PKC-mediated pathway. B. Western blot of detergent-soluble and detergent-insoluble K17 in L3.6 cells with TPA treatment (T, 400nM) strongly induces K17-disassembly after 1h, shown by fold change of detergent-soluble K17. Data are mean ± s.d, n = 5 independent repetitions. Student’s t-test. C. Western blot of detergent-soluble and detergent-insoluble K17 in L3.6 cells with TPA treatment (T, 400nM) in loss of function mutant and wt K17. Data are mean ± s.d, n = 3 independent repetitions. Student’s t-test. Data are mean ± s.d, n = 3 independent repetitions. Student’s t-test. D-F. Treatment of L3.6 cells with sub-IC50 doses of inhibitors for PKC (D), MEK1/2 (E), and RSK (F) kinases for 1h, followed by 1h-TPA stimulation, reduced the levels of detergent-soluble K17, shown by fold change of detergent-soluble K17 and western blot. Data are mean ± s.d, n = 3 independent repetitions. Student’s t-test. G. Illustration of the PKC and downstream kinases proposed to phosphorylate K17 at the N-terminal S10-13 region. s.d, standard deviation; n, number of repetitions. Coomassie Brilliant Blue (CBB), detergent-insoluble (insol), detergent-soluble (sol), 12-O-Tetradecanoylphorbol-13-acetate (TPA). *P<0.05, **P<0.01, ***P<0.001.

K17-phosphorylation impacts tumor growth and metastases in vivo.

A. Kaplan-Meier curves for immunodeficient NU/J mice implanted orthotopically with L3.6 cells stably expressing K17 with loss of phosphorylation function (LOPF), S10-13A. Significance determined using Log-rank Mantel–Cox test. B. Representative images of K17 IHC in murine tumors expressing wtK17 or phospho-mutant K17. Scale = 100μm. C. Relative tumor mass. Tumors of immunodeficient NU/J mice implanted orthotopically with L3.6 cells stably expressing gain of phosphorylation function (GOPF) S10-13D K17 are larger than those expressing loss of phosphorylation function S10-13A, as determined by tumor mass relative to body weight. Data are mean ± s.d, n = 4 mice per group. Student’s t-test. D. K17-positive hepatic metastasis assessed by IHC. Mice implanted with L3.6 cells expressing S10-13D or S10-13E K17 have more hepatic metastasis. Data are mean ± s.d, n = 5-8 mice per group. Student’s t-test. E. Protein expression of metastasis-associated genes from CPTAC PDAC samples annotated with K17 protein expression and total K17-phosphorylation status. Expression expressed in fold change. F. qRT-PCR mRNA expression of metastasis-associated genes in L3.6 cells expressing phospho-K17 mutants, relative to wild-type (wt) K17. Data expressed as fold change of wt K17. Data are mean ± s.d, n = 3 repetitions. Student’s t-test. *P<0.05, **P<0.01, ***P<0.001.

Detergent-soluble K17 is negatively prognostic in human and murine PDACs.

A-B. Linear regression showing that K17 disassembly is independent of PCNA expression in mouse tumors n= 24 tumors (A) and human tumors n= 23 tumors (B). Significance determined by Spearman correlation test. C. Kaplan-Meier curves show the overall survival of human patients included in this study, according to clinical stage. Significance determined using Log-rank Mantel–Cox test.

K17 is differentially phosphorylated in its detergent-soluble state, and N-terminal serine phosphorylation promotes disassembly.

A. Alignment of type I human keratins polypeptide show similarity in the S10-13 hotspot. B. Serine 44 phosphorylation is highly enriched in detergent-soluble portions of the K17 protein in 3 representative human PDACs. C-D. Subcellular localization of K17 after okadaic acid (OA) treatment. Immunofluorescence stains (C) and quantifications (D) in L3.6 cells. Scale =5μm. E. Western blot depicting detergent-soluble and detergent-insoluble K17 after treatment with okadaic acid (OA) treatment. Treatment induces disassembly of K17 in L3.6 and BxPC3 cells. Data are mean ± s.d, n = 3 independent repetitions. Student’s t-test. F. Western blot depicting detergent-soluble and detergent-insoluble K17 after treatment with sodium orthovanadate (OV). Data are mean ± s.d, n = 3 independent repetitions. Student’s t-test. Coomassie Brilliant Blue (CBB), *P<0.05, **P<0.01, ***P<0.001.

K17-phosphorylation impacts tumor growth and metastases in vivo.

A. Percent detergent-soluble K17 in the tumors derived from GOPF S10-13E K17 wtK17 and LOPF S10-13A K17. Data are mean ± s.d, n = 4 independent PDACs. Student’s t-test. B. Percent tumor necrosis from PDACs derived from GOPF S10-13E K17 wtK17 and LOPF S10-13A K17. Data are mean ± s.d, n = 3 independent repetitions. Student’s t-test. C. Proliferation curve of K17-phosphorylation gain-of-function mutants (S10-13D and S10-13E), loss of function mutant (S10-13A) and wild type (WT) K17. D-E. Wound-closure assay (4x-D) and respective quantification (E) show that phosphorylation of K17 in S10-13 region enhances migration of PDAC cells. Coomassie Brilliant Blue (CBB), detergent-insoluble (ins, I), detergent-soluble (sol, S). *P<0.05, ***P<0.001, ****P<0.0001.