INPP4B promotes PDAC aggressiveness via PIKfyve and TRPML-1-mediated lysosomal exocytosis

Abstract

Aggressive solid malignancies, including pancreatic ductal adenocarcinoma (PDAC), can exploit lysosomal exocytosis to modify the tumor microenvironment, enhance motility, and promote invasiveness. However, the molecular pathways through which lysosomal functions are co-opted in malignant cells remain poorly understood. In this study, we demonstrate that inositol polyphosphate 4-phosphatase, Type II (INPP4B) overexpression in PDAC is associated with PDAC progression. We show that INPP4B overexpression promotes peripheral dispersion and exocytosis of lysosomes resulting in increased migratory and invasive potential of PDAC cells. Mechanistically, INPP4B overexpression drives the generation of PtdIns(3,5)P2 on lysosomes in a PIKfyve-dependent manner, which directs TRPML-1 to trigger the release of calcium ions (Ca2+). Our findings offer a molecular understanding of the prognostic significance of INPP4B overexpression in PDAC through the discovery of a novel oncogenic signaling axis that orchestrates migratory and invasive properties of PDAC via the regulation of lysosomal phosphoinositide homeostasis.

Graphical abstract

Figure 1.

High levels of INPP4B expression are a common feature in primary PDAC patient samples and affects growth of PDAC cell lines. (a) INPP4B transcript expression from normal pancreas (N) versus PDAC patient specimens (P) in GEO datasets. (b) Kaplan–Meier survival analysis performed on four patient datasets where INPP4B^high versus INPP4B^low are compared using a median cutoff. (c and d) Depicted are cells with stable overexpression of INPP4B in INPP4B^low (c) PANC-1 or INPP4B KO in (d) INPP4B^high HPAC cell lines. Cells were tested for INPP4B protein expression by immunoblot (top left panels). Representative images of crystal violet–stained colonies and quantitation of average colony number and average colony size in clonogenic assay of INPP4B overexpressing (c) PANC-1 or INPP4B KO in (d) HPAC cells, with respective empty controls for PANC-1 cells or indicated control CRISPR-Cas9 KO for HPAC cells 19 days after seeding. Proliferation assays (bottom left images) and representative crystal violet–stained wells (bottom right graph) at day 0 and on the indicated day of the proliferation assay along with crystal violet quantification for the indicated days. Data are representative of three individual experiments, with unpaired two-tailed parametric t test performed for a and colony formation assays for c, log-rank (Mantel-Cox) test for b, one-way ANOVA Tukey’s post hoc test for growth curve assays from c and colony formation assays from d, two-way ANOVA and multiple unpaired t tests for growth curve assays from d. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Source data are available for this figure: SourceData F1.

Figure S1.

Effect of INPP4B expression on PDAC cell growth. (a) Differential analysis of INPP4B protein expression in normal and primary PDAC were assessed using Clinical Proteome Tumor Analysis Consortium (CPTAC) data accessed from the UALCAN data portal. (b) Dot blot of INPP4B transcript expression of INPP4B in normal (N) and tumor specimens (T) across 33 TCGA tumor samples and paired normal tissues. (c) Ranked means and total mean (inset) of INPP4B expression in PDAC cell lines compared to all other cancer cell lines in the Cancer Cell Line Encyclopedia. (d and e) Images representing cells with stable overexpression of INPP4B in INPP4B^low (d) PK-1 or INPP4B KO in (e) INPP4B^high PK8 cell lines. Cells were tested for INPP4B protein expression by IB (top left panels). Representative images of crystal violet–stained colonies and quantitation of average colony number and average colony size in clonogenic assay, with respective empty controls for PK-1 cells (d) or control CRISPR-Cas9 KO in (e) PK8 cells 19 days after seeding. Proliferation assays (bottom left images) and representative crystal violet–stained wells (bottom right graph) at day 0 and on the final day of the proliferation assay along with crystal violet quantification for the indicated days for PK-1 (d) and PK8 (e) cells. (f–h) Anchorage-independent growth capacity in soft agar (representative images, left images) was quantitated (right graphs) for empty control or INPP4B overexpressing PANC-1 (f) and PK-1 (g) cells, control or INPP4B CRISPR-Cas9 KO for PK8 (h) cells. (i) pCW-INPP4B PANC-1 and PK-1 cells were treated with Dox at the indicated concentrations, and HPAC cells stably expressing indicated sgRNAs were immunoblotted for INPP4B, LAMP1, and actin. (j) Quantitation of total LAMP1 intensity per cell for indicated cell lines and treatments. Data are representative of ± SD from three individual experiments for d–h, with 100–120 cells counted per treatment per condition for j, with unpaired two-tailed parametric t test performed for c, and colony formation assays for d, e, f–h, and j, one-way ANOVA Tukey’s post hoc test for j, and two-way ANOVA and multiple unpaired t tests for growth curve assays for d and e. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Source data are available for this figure: SourceData FS1.

Figure 2.

INPP4B expression regulates lysosomal localization in PDAC cells. (a) Top 20 KEGG gene sets enriched in INPP4B^high TCGA-PAAD patients as determined using GSEA. (b) KEGG lysosome, autophagy, and endocytosis enrichment plots from analysis of INPP4B^high GSEA analysis. NES, normalized enrichment score. (c) Micrographs of indicated PDAC cell lines immunostained for LAMP1 (green) and DAPI nuclear stain (blue). Scale bar: 10 µm. (d) Whole-cell lysates from indicated PDAC cell lines were subjected to IB for INPP4B and GAPDH. Quantification of INPP4B expression levels normalized to GAPDH is shown in red. (e) Illustration of image analysis technique used to segment cell areas for nucleus, inner shell (perinuclear), and outer shell (peripheral). (f) Quantitation of LAMP1 intensity distribution for images from c as measured by peripheral (outer shell)/perinuclear (inner shell) LAMP1 intensity ratio. (g) Scatterplot correlation of INPP4B protein expression in various PDAC cell lines versus outer/inner shell LAMP1 intensity. (h and j) LAMP1 immunostaining and DAPI nuclear staining of (h) PANC-1 and (j) PK-1 stably expressing pCW-INPP4B with and without 500 nM Dox. Scale bar: 20 µm for (h) PANC-1 cells and 15 µm for (j) PK-1 cells. (i and k) Quantitation of LAMP1 intensity distribution as measured by peripheral (outer shell)/perinuclear (inner shell) LAMP1 intensity ratio for (i) PANC-1 and (k) PK-1 cells. (l) HPAC cells stably expressing CRISPR-Cas9 and indicated sgRNAs for sgCDY1B, sgINPP4B-1, or sgINPP4B-2 immunostained for LAMP1 and nuclear stained with DAPI. Scale bar: 10 µm. (m) Quantitation of LAMP1 intensity distribution for l HPAC cells. Data represent ± SD from 100 to 120 cells from three independent experiments assessed per treatment per condition, with one-way ANOVA Tukey’s post hoc test performed for f and m, and unpaired two-tailed parametric t test for i and k. ***P < 0.001, ****P < 0.0001. Source data are available for this figure: SourceData F2.

Figure S2.

Effect of INPP4B expression on lysosome function. (a–l) PANC-1 (a–f) or PK-1 (g–l) cells transiently expressing pmCherry or pmCherry-INPP4B were stained with Lysotracker Green (a and g) or DQ Green BSA (b and h) or Lucifer Yellow (c and i). Scale bar: 25 µm. Flow cytometry used to quantitate Lysotracker Green (lysosome acidity probe) (d and j) or DQ-BSA (probe for lysosomal proteolytic function) (e and k) or Lucifer Yellow (to assess endocytosis toward lysosomes) (f and l) intensity in non-transfected, mCherry, or INPP4B-mCherry expressing PANC-1 (a–f) or PK-1 (g–l) cells. Data represent ± SD from three independent experiments, with unpaired two-tailed parametric t test performed for statistical measures. *P < 0.05.

Figure S3.

Effect of INPP4B expression on lysosome gene transcription and TFEB function. (a) Stably expressing pCW-INPP4B PANC-1 and PK-1 cells treated with or without 500 nM Dox, or HPAC cells stably expressing the indicated gRNAs sgCDY1B (C), sgINPP4B-1 (1), or sgINPP4B-2 (2), and assessed for gene expression through qRT-PCR for the select lysosome genes as indicated and normalized to GAPDH. (b) pCW-INPP4B PANC-1 or PK-1 cells without or with 500 nM Dox immunoblotted for total and phosphorylated states for the proteins TFEB, mTOR, AKT, S6, 4EBP1, INPP4B, and β-actin. (c and d) pCW-INPP4B (c) PANC-1 cells or (d) PK-1 cells without or with 500 nM Dox or EBSS serum starvation and immunostained for endogenous TFEB (green) and DAPI nucleus (blue). Scale bar: 20 µm. Quantification of nuclear TFEB over cytosol TFEB intensity for PANC-1 (c) and PK-1 (d) cells. (e) Immunoblot of the isolated nucleus and cytosol fractions of no Dox or 500 nM Dox treated pCW-INPP4B PANC-1 cells or PK-1 cells for TFEB, MITF, and TFE3. (f and g) pCW-INPP4B PANC-1 cells treated with siCTRL or siTFEB and no Dox or 500 nM DOX and (f) immunoblot for INPP4B TFEB and actin, (g) examined for transcript expression of select lysosome genes through qRT-PCR. (h) pCW-INPP4B PANC-1 cells treated with siCTRL or siTFEB and without or with 500 nM Dox immunostained for LAMP1 (green) and DAPI nucleus (blue) and assessed for LAMP1 outer/inner shell intensity distribution. Scale bar: 20 µm. Data represent ± SD from three independent experiments for each treatment condition with 80–100 cells examined for c, d, and h, with unpaired two tailed parametric t test performed for PANC-1 and PK-1 cells for a, one-way ANOVA Tukey’s post hoc test for HPAC cells for a, c, d, g, and h. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Source data are available for this figure: SourceData FS3.

Figure 3.

INPP4B expression regulates cellular Ca²⁺ levels in PDAC cells through TRPML-1 function. (a) pCW-INPP4B PANC-1 cells without or with 500 nM Dox or ionomycin were co-stained for lysosomes with Lysotracker Red (red) and cellular Ca²⁺ with Fluo-4 AM (green). Scale bar: 15 µm. (b and c) Quantitation of (b) Fluo-4 AM intensity on Lysotracker Red–positive structures or (c) cytosolic Fluo-4 AM intensity. (d–f) HPAC cells stably expressing indicated sgRNAs were (d) co-stained for lysosomes with Lysotracker Red (red) and cellular Ca²⁺ with Fluo-4 AM (green) and quantitated (e and f) as above. Scale bar: 10 µm. (g–i) (g) pCW-INPP4B PANC-1 without or with 500 nM Dox or Gly-Phe β-naphthylamide (GPN) treatment, transfected with pBoBi-hLAMP2-C-GC6s (LAMP2-GCaMP6 [green]), labeled for lysosomes with Lysotracker Red (red), and quantitated (h and i) as above. Scale bar: 5 µm. (j–l) (j) pCW-INPP4B PANC-1 without or with 500 nM Dox were treated with vehicle or ML-SI3, co-stained for lysosomes with Lysotracker Red (red) and cellular Ca²⁺ with Fluo-4 AM (green), and quantitated (k and l) as above. Scale bar: 15 µm. Data represent ± SD from 100 to 120 cells (a–f and j–l) or 50–60 cells (g–i) from three independent experiments assessed per treatment per condition, with one-way ANOVA Tukey’s post hoc test performed for statistical measures. **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure S4.

INPP4B expression regulates cellular Ca²⁺ levels in PDAC cells. (a) pCW-INPP4B PK-1 cells treated without or with 500 nM Dox or ionomycin were co-stained for lysosomes with Lysotracker Red (red) and cellular Ca²⁺ with Fluo-4 AM (green). Scale bar: 15 µm. (b and c) Quantitation of Fluo-4 AM intensity on Lysotracker Red–positive structures (b) or cytosolic Fluo-4 AM intensity (c). (d–f) Representative intensity distribution false-color images and intensity surface plots (inset) of pCW-INPP4B PANC-1 (d) or PK-1 (e) cells treated without or with 500 nM Dox or ionomycin, or HPAC cells stably expressing indicated gRNAs (f), and labeled with Fluo-4 AM for cellular calcium. (g–i) (g) pCW-INPP4B PK-1 cells treated without or with 500 nM Dox or GPN, transfected with pBoBi-hLAMP2-C-GC6s (LAMP2-GCaMP6 [green]), labeled for lysosomes with Lysotracker Red (red), and quantitated (h and i) as above. Scale bar: 5 µm. (j–l) pCW-INPP4B PK-1 cells treated without or with 500 nM Dox and vehicle or ML-SI3 (j), labeled for cellular Ca²⁺ with Fluo-4 AM and lysosomes with Lysotracker Red, and quantitated as above (k and l). Scale bar: 10 µm. (m–o) (m) pCW-INPP4B PK-1 cells treated with vehicle or ML-SA1 or MK6-83, labeled for cellular Ca²⁺ with Fluo-4 AM and lysosomes with Lysotracker Red, and quantitated as above (n and o). Scale bar: 10 µm. Data represent ± SD from 80 to 90 cells (a–c and j–o) or 50–60 cells (g–i) from three independent experiments assessed per treatment per condition, with one-way ANOVA Tukey’s post hoc test performed for statistical measures. ****P < 0.0001.

Figure 4.

INPP4B-regulated lysosome positioning is mediated by TRPML-1–associated Ca²⁺. (a and b) PANC-1 (a) or PK-1 (b) was treated with vehicle or ionomycin and immunostained for LAMP1 (green) and DAPI nuclear stain (blue). Scale bar: 15 µm. LAMP1 intensity distribution as measured by peripheral (outer shell)/perinuclear (inner shell) LAMP1 intensity ratio. (c and d) pCW-INPP4B (c) PANC-1 or (d) PK-1 cells without or with 500 nM Dox were treated with vehicle or BAPTA followed by immunostaining with LAMP1 (green) and DAPI nuclear stain (blue) and LAMP1 intensity distribution as indicated above. Scale bar: 20 µm. (e and f) pCW-INPP4B (e) PANC-1 or (f) PK-1 cells without or with 500 nM Dox were treated with vehicle or ML-SI3. Cells were immunostained as above and LAMP1 intensity distribution was measured. Scale bar: 20 µm. (g) HPAC cells stably expressing indicated sgRNAs were treated with vehicle, MK6-83, or ML-SA1. (h) Cells were immunostained as above and LAMP1 intensity distribution was measured. Data represent ± SD from 100 to 120 cells from three independent experiments assessed per treatment per condition, with unpaired two-tailed parametric t test for a and b, and one-way ANOVA Tukey’s post hoc test performed for c–h. *P < 0.05, ***P < 0.001, ****P < 0.0001.

Figure 5.

INPP4B regulates lysosomal exocytosis in a TRPML-1–dependent manner. (a–d and e–h) pCW-INPP4B PANC-1 (a–d) or PK-1 (e–h) cells without or with 500 nM Dox were treated with vehicle, ionomycin, BAPTA, or ML-SI3 followed by flow cytometry analysis of cell surface LAMP1. Representative immunostained images of cell surface LAMP1 and nuclear DAPI stain for PANC-1 (a) or PK-1 (e) cells. Scale bar: 15 µm. (i and j) HPAC cells stably expressing indicated sgRNAs were treated with vehicle, ML-SA1, or MK6-83 TRPML-1 agonists followed by flow cytometry analysis of cell surface LAMP1 (j). Representative immunostained images of cell surface LAMP1 and nuclear DAPI stain for HPAC cells (i). Scale bar: 15 µm. (k–m) Indicated cell models were cultured for cells for 6 days followed by measurement of extracellular lysosomal hexosaminidase activity. (n) pCW-INPP4B PK-1 cells were grown for 5 days, then exposed to vehicle or ML-SI3 for 24 h followed by measurement of extracellular lysosomal hexosaminidase activity. Data represent ± SD from three independent experiments, with one-way ANOVA Tukey’s post hoc test performed for b–j and m and n, and unpaired two-tailed parametric t test for k and l. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 6.

INPP4B affects cellular processes and surface F-actin related to migration. The most enriched GO terms among the top 200 correlated genes with INPP4B expression in TCGA-PAAD. (a–c) The top 20 GO terms for (a) Molecular Function, (b) Biological Process, and (c) Cellular Component, ranked by significance, with the top five terms colored. FDR, false discovery rate. (d) Image analysis technique to quantitate cell edge F-actin intensity. (e and g) pCW-INPP4B (e) PANC-1 and (g) PK-1 cells without or with 500 nM Dox-induced and vehicle or ML-SI3 treatment were assessed for cell edge F-actin expression through phalloidin staining. Scale bar: 20 µm. (f and h) Edge F-actin intensity analysis for (f) PANC-1 or (h) PK-1 cells. (i) HPAC cells stably expressing the indicated gRNAs were treated with vehicle or ML-SA1 or MK6-83 to assess the effect of TRPML-1 activation on cell edge F-actin. (j) Edge F-actin intensity analysis for i HPAC cells. Scale bar: 15 µm. Data represent ± SD from 90 to 100 cells from three independent experiments assessed per treatment per condition, with one-way ANOVA Tukey’s post hoc test performed for statistical measures. *P < 0.05, **P < 0.01, ****P < 0.0001.

Figure 7.

INPP4B affects PDAC cell migration and invasion. (a–i) Depicted are cells with stable overexpression of empty control (pSMAL) or INPP4B (pSMAL-INPP4B) in INPP4B^low (a, d, and e) PK-1 or (b, f, and g) PANC-1 cells, or INPP4B KO with the indicated gRNAs in c, h, and i INPP4B^high HPAC cell lines. Respective cell models were tested for wound healing (a–c), transwell migration (d, f, and h), and transwell invasion (e, g, and i) assays. (j–m) pCW-INPP4B PANC-1 treated with siCTRL or siMCOLN-1 and no Dox or 500 nM Dox and assessed for transwell migration (j), transwell invasion (k), protein levels of MCOLN-1 (l), and transcripts of MCOLN-1 gene expression (m). Data represent ± SD from three independent experiments, with unpaired two-tailed parametric t test performed for a, b, and d–g, and one-way ANOVA Tukey’s post hoc test for c and h–m. **P < 0.01, ***P < 0.001, ****P < 0.0001. Source data are available for this figure: SourceData F7.

Figure 8.

INPP4B promotes PDAC cell migration and invasion through TRPML-1 function. (a–d) pCW-INPP4B PANC-1 (a and b) or PK-1 (c and d) cells without or with 500 nM Dox were treated with vehicle or ML-SI3 and assessed for cell migration (a and c) or invasion through a Matrigel basement membrane (b and d), followed by crystal violet staining of cells attached to ThinCert membrane. (e and f) HPAC cells stably expressing indicated sgRNAs were treated with vehicle ML-SA1 or MK6-83 and assessed for cell migration (e) or invasion (f) as above. Quantitation of crystal violet staining is displayed for each condition. Data represent ± SD from four independent experiments, with one-way ANOVA Tukey’s post hoc test performed for statistical measures. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure S5.

INPP4B regulates phosphoinositide levels in PDAC cells for migration. (a–d) pCW-INPP4B PANC-1 (a and c) or PK-1 (b and d) cells without or with 500 nM Dox immunostained for PtdIns(3,4)P₂ (a,b) or PtdIns(3,5)P₂ (c and d) and transiently expressing LAMP1-mCherry, with intensity quantification of respective phosphoinositide immunostain on LAMP1-mCherry–positive structures or total cell phosphoinositide intensity. (e) PANC-1 cells treated with vehicle or 20 nM apilimod, transfected with GFP-SnxA and stained for lysosomes with Lysotracker Red, and quantified as above. Scale bar for a–e: 5 µm. (f–h) pCW-INPP4B PK-1 treated with or without 500 nM Dox and vehicle or 10 nM apilimod assessed for (f) cell surface LAMP1 through microscopy imaging (left image panel, scale bar: 20 µm) and flow cytometry (right graph), and transwell migration (g) and invasion (h). Data represent ± SD from three independent experiments assessed per treatment per condition, with 80–90 cells (a–d) or 50–60 cells (e) counted, with unpaired two-tailed parametric t test performed for a–e, one-way ANOVA Tukey’s post hoc test for f–h. *P < 0.05, ***P < 0.001, ****P < 0.0001.

Figure 9.

INPP4B promotes the generation of lysosomal PtdIns(3,5)P2 in PDAC cells. (a–d) pCW-INPP4B PANC-1 or PK1 cells without or with 500 nM Dox were transfected with GFP-2xFYVE (green) (a and b) to assess PtdIns(3)P or GFP-SnxA (green) (c and d) to assess PtdIns(3,5)P₂ and labeled for lysosomes with Lysotracker Red (red), with quantification for the respective probes overlayed on Lysotracker Red–positive structures. Scale bar: 5 µm. Data represent ± SD from three independent experiments assessed per treatment per condition with 50–60 cells counted for a–d, with unpaired two-tailed parametric t test performed for statistical measures. ****P < 0.0001.

Figure 10.

INPP4B promotes lysosomal phenotypes, migration, and invasion migration through PtdIns(3,5)P₂ generation. (a–d) pCW-INPP4B PANC-1 cells treated without or with 500 nM Dox and with vehicle or 10 nM apilimod and assessed for (a) LAMP1 intensity distribution through LAMP1 immunostain (green) and DAPI nuclear stain (blue) (scale bar: 20 µm), (b) cell surface LAMP1 through microscopy imaging (left panel images, scale bar: 20 µm) and flow cytometry (right graph), (c) transwell migration, and (d) transwell invasion. Data represent ± SD from three independent experiments assessed per treatment per condition with 100–120 cells counted for a with one-way ANOVA Tukey’s post hoc test performed for statistical measures. **P < 0.01, ****P < 0.0001.