An inhibitor of death-associated protein kinase 3 (DAPK3) disrupts Hippo signaling and intestinal epithelial regeneration in murine DSS-induced colitis

Abstract

Various signaling modules that affect epithelial wound healing are dysregulated in ulcerative colitis. Hippo signaling, acting downstream of cytoskeletal remodeling, is necessary for intestinal epithelial regeneration. Death-associated protein kinase 3 (DAPK3) is a regulator of actin cytoskeleton organization that also controls cellular proliferation and apoptosis responses. Recent genetic linkages between DAPK3 and the Hippo pathway suggest signaling coordination that has not been empirically evaluated. The impact of HS38, a DAPK3 inhibitor, on epithelial wound healing was examined using the dextran-sodium-sulphate (DSS) murine model of experimental colitis and Caco-2 human intestinal epithelial cell (IEC) monolayers. DAPK3 was significantly elevated in IECs isolated from DSS-treated mice, with cytoplasmic staining observed in epithelial crypts. The administration of HS38 to mice receiving DSS impeded the resolution of intestinal injury with attenuated IEC proliferation. HS38 treatment in DSS-colitis was also accompanied with decreased pS127, but not pS397, phosphorylation of Yes-associated protein (YAP). The data suggest that HS38 treatment, and hence DAPK3 inhibition, attenuates Hippo pathway signaling with subsequent nuclear enrichment of YAP. Additional analysis of single-cell RNAseq transcriptomic data from mucosal biopsies of UC patients reveals strong positive correlations between DAPK3 and YAP1 in the epithelial compartment. In conclusion, this study establishes DAPK3 as a novel factor in intestinal epithelial regeneration and ulcerative colitis progression by way of Hippo/YAP signaling. Nevertheless, the role that DAPK3 plays in different cell types will need further investigation to decipher the full consequence of DAPK3 involvement in epithelial homeostasis.

1. INTRODUCTION

Intestinal epithelial cells (IECs) that line the gastrointestinal tract provide a physical and biochemical barrier that separate the host’s internal milieu from foreign antigens and microbial pathogens present in the lumen (Kagnoff, 2014; Peterson and Artis, 2014; Patankar and Becker, 2020). Moreover, the epithelium serves as a communication hub for the intestinal mucosa where it participates in the coordination of mucosal defense and homeostasis by disseminating luminal events to the sub-epithelial immune compartment (Kagnoff, 2014). Not surprisingly, altered IEC function is frequently detected in ulcerative colitis (UC) patients (Issenman et al., 1993; Chang et al., 2017), and has been proposed as a potential etiologic factor for UC (Pastorelli et al., 2013). Epithelial integrity may be provisionally weakened by physiological inflammation or minor trauma (Hageman et al., 2020). But, as the balance between pro- and anti-inflammatory stimuli is perpetually disturbed in UC, epithelial regeneration may be altered so that normal epithelial barrier function is not restored (Strober and Fuss, 2011; Brazil et al., 2019). Inefficacious wound healing can give rise to altered epithelial architecture, impaired barrier function, and intestinal neoplasia which further aggravate the inflammatory state (Koch and Nusrat, 2012; Leoni et al., 2015; Boal Carvalho et al., 2016). Therefore, rapid, and complete epithelial wound healing is essential for intestinal health and resolution of UC.

While much is known about the spatial and temporal signaling events necessary for epithelial repair (Kagnoff, 2014; Brazil et al., 2019; Hageman et al., 2020), additional intricacies remain to be explored. The death-associated protein kinases (DAPKs) have emerged as important contributors to intestinal mucosal defense and epithelial homeostasis (Chen and MacDonald, 2023). Members of this kinase family display distinct roles in the intestinal epithelium to control: cell death and neoplastic transformation (DAPK1), anoikis of malignant intestinal epithelium (DAPK2), and UC-associated dysplasia-carcinoma progression (DAPK3). DAPK3 also has fundamental roles in cellular processes integral to the maintenance of epithelial homeostasis (Bovellan et al., 2010; Lai and Chen, 2014; Steinmann et al., 2015) wherein it contributes to cell growth and proliferation of colon carcinoma cell lines (Togi et al., 2011) as well as autophagy induction (Li et al., 2021). A reduction in focal adhesion kinase (FAK) activity found with ectopic expression of DAPK3 (Nehru et al., 2013) suggests the kinase may also regulate cellular adhesion during epithelial wound repair and influence IEC proliferation and survival processes during colitis progression (Owen et al., 2011). However, the molecular actions of DAPK3 have yet to be assessed in experimental models of colitis.

Intriguingly, DAPK3 was recently identified as a potential key factor in ulcerative colitis (UC) progression (Chen and MacDonald, 2022). Core to this finding was the identification of a novel linkage between DAPK3 and a downstream effector of the Hippo signaling pathway (i.e., Yes-associated protein 1, YAP; a proliferation-activating transcriptional coactivator). A differential co-expression analysis revealed a regulatory partnership for DAPK3 and YAP, conditioned on UC extent, with dysregulation of the actin cytoskeleton as a core biological association (Chen and MacDonald, 2022). Research also identifies the Hippo/YAP pathway to be an important regulator of epithelial cell growth, proliferation and survival (Yui et al., 2018; Hong et al., 2016; Sprangers et al., 2021; Xie et al., 2021). Within the nucleus, YAP may bind with TEAD1–4 (TEA domain transcription factors) to activate the expression of pro-proliferative and pro-survival genes (Zhao et al., 2023). Additional evidence suggests that Hippo/YAP is associated with inflammatory processes (i.e., NF-kb) as well as immune-regulating pathways (Xie et al., 2021). Dysregulation of Hippo/YAP was found in UC patients (Nterma et al., 2020) as well as experimental models of colitis (Deng et al., 2018).

Although the impairment of several conserved signaling pathways is associated with intestinal epithelial dysfunction in UC (Koch and Nusrat, 2012; Leoni et al., 2015; Xie et al., 2021), the intricate coordination of the many networks involved in the restoration of epithelial homeostasis after injury and/or inflammation remain to be discerned. We investigated whether DAPK3 plays a role in epithelial repair and/or colitis progression given that this protein kinase is linked to a key signaling pathway that coordinates epithelial regeneration after injury.

2. MATERIALS AND METHODS

Materials –

The Click-iT EdU Cell Proliferation Kit for Imaging, Alexa Fluor^™ 488 (#C10337) and 4’,6-diamidino-2-phenylindole (DAPI; #62248) were purchased from Invitrogen-ThermoFisher Scientific. The Phalloidin-iFluor 647 reagent (#176759) was from Abcam (Cambridge, MA). Multiple antibodies were obtained from Cell Signaling Technology (Danvers, MA), including YAP (D8H1X) XP^® rabbit mAb (#14074), phospho[Ser127]-YAP (D9W2I) rabbit mAb (#13008), phospho[Ser397]-YAP (D1E7Y) rabbit mAb (#13619), YAP/TAZ (D24E4) rabbit mAb (#8418), phospho[Thr35]-MOB1 (D2F10) rabbit mAb (#8699), MOB1 (E1N9D) rabbit mAb (#13730), LATS1 (C66B5) rabbit mAb, (#3477), GAPDH (14C10) rabbit mAb (#2118), Ki-67 (D3B5) rabbit mAb (#12202), MST1 rabbit IgG (#3682) and MST2 rabbit IgG (#3952). DAPK3 rabbit IgGs #2928 and #TA313745 were obtained from Cell Signaling and Origene (Rockville, MD), respectively. Cy3-conjugated AffiniPure donkey anti-rabbit IgG (#711-165-152) was from Jackson ImmunoResearch, and dextran sodium sulphate (DSS, 36–50 kDa; #0216011090) was from MP Biomedicals (Solon, OH).

DAPK3 Inhibitor (HS38) –

The small molecule inhibitor HS38 (CAS No. 1030203-81-6) was synthesized as described previously (Carlson et al., 2013). HS38 is an ATP-competitive, small molecule DAPK inhibitor that displays high specificity and great potency while bypassing concerning off-target liabilities that are inherent to other compounds. For example, HS38 lacks the concerning off-target effect of the TC-DAPK6 molecule (CAS No. 315694-89-4) on Rho-associated coiled-coil containing kinase (ROCK) (Al-Ghabkari et al., 2016). Several uncommon structural features in the ATP-binding pocket of DAPK3 underwrite the high selectivity offered by HS38 when compared against other protein kinase clades (Carlson et al., 2018). In brief, HS38 displays highest potency towards DAPK3 (and DAPK1) and the structurally related PIM3 along with diminishing action toward IRAK4, DYRK2 and PIM1 (10-fold less), PIM2 and MLCK (100-fold less), and ROCK (1000-fold less).

Mouse model of dextran sodium sulphate (DSS)-induced colitis -

All animal protocols were approved by the Animal Care and Use Committee at the University of Calgary and conform to the guidelines set by the Canadian Council of Animal Care (Protocol# AC16–0046). Animal experiments were completed by randomly assigning 10-week-old male C57BL/6 mice (Charles River Laboratories, Montreal, QC, Canada) to one of four experimental treatment groups: Cohort 1 (H2O + Vehicle/DMSO, n=12), Cohort 2 (H2O + HS38, n=12), Cohort 3 (DSS + Vehicle/DMSO, n=12) and Cohort 4 (DSS + HS38, n=12). Mouse Cohorts 1/2 were provided regular drinking water while Cohorts 3/4 were provided 2.5% (w/v) DSS in drinking water ad libitum for seven days. At the end of 7 days, all mice were provided regular drinking water for two days, then the animals were euthanized by cervical dislocation after anesthesia by isoflurane inhalation. Daily subcutaneous injection of HS38 (500 μg/kg) or vehicle (DMSO) began three days prior to the introduction of DSS to drinking water and continued until euthanasia on DSS+2. The HS38 dosage was selected based on in vitro binding kinetics (Carlson et al., 2013) and pharmacokinetic observations from previous studies of in vivo tolerance among treatment animals (Carlson et al., 2018). Body weight was tracked daily and normalized to day one of the DSS administration. A disease activity index (DAI) score was calculated daily based on weight, activity/lethargy, stool consistency, and blood in feces; investigator-blinded macroscopic scoring was also performed on colon tissue dissected postmortem (Hirota et al., 2011).

Histological Examination of Epithelial Damage and Inflammation -

Paraffin-embedded tissues (distal colon), sectioned at 5-μm thickness, were deparaffinized, rehydrated with gradient ethanol, then stained with hematoxylin and eosin (H&E) for investigator-blinded microscopic assessment of colitis severity. The grading system considered the degree of epithelial damage, infiltration of inflammatory cells, as well as the percentage of involvement (Hirota et al., 2011).

Immunohistochemistry -

Deparaffinized colon tissue was rehydrated with gradient ethanol, then subjected to heat-induced antigen retrieval in 10 mM sodium citrate (pH 6.0). Sections were subsequently washed with PBST, then blocked with 10% normal donkey serum in PBS. Overnight incubation was carried out at 4 °C with anti-Ki-67, anti-YAP or anti-DAPK3 diluted 1:400 in PBS. The Cy3-conjugated AffiniPure donkey anti-rabbit IgG was the secondary antibody and was applied to samples at 1:200 dilution. DAPI was used at 10 μg/mL to mark nuclei. Images were taken using an Olympus FV10i confocal microscope with 60X objective (oil). Ki-67⁺ positive nuclei were manually counted, then expressed as a percentage of total nuclei per crypt. Between 19–30 well-oriented crypts from 2–3 sections were evaluated per animal. For sections probed with anti-DAPK3 or anti-YAP, images were taken with Nikon A1R laser scanning confocal microscope on 40X objective (oil) at Nyquist (xy-plane). For image deconvolution, stacks of 10 equidistant (0.5 μm) z-planes were evaluated for image deconvolution.

Isolation of Colonic Crypts and DAPK3 expression analyses –

Colonic crypts were isolated by chelation agent treatment (Booth and O’Shea, 2002). Mouse colons were dissected, extensively washed, and then macerated; tissue fragments were incubated on ice for 60 min in Hank’s balanced salt solution (HBSS) without calcium and magnesium, containing 0.5 mM DTT and 8 mM EDTA. Crypts were mechanically released from the lamina propria by vigorous pipetting. The released epithelium was filtered through a 70-μm mesh and centrifuged at 100 × g for 5 min at 4 °C. In some cases, the isolated epithelial samples were further digested with Collagenase/Dispase Enzyme solution to release individual cells for flow cytometry analyses.

qRT-PCR Analyses -

RNA was extracted from colonic crypt samples with TRIzol^™ (Cat#15596, Invitrogen). Total RNA (1.5 μg) was then prepared into cDNA using the RT2 First Strand Kit (Cat#330404, Qiagen) by following procedures outlined in the manufacturer manual. PCR reactions were run using the CFX96 Touch Real-Time PCR Detection System (Bio-Rad). Dapk3 was amplified with forward primer 5’-CTGGAGTGGAAGCTGCTGGAAG-3’ and reverse primer 5’-AACTTGGCTGCGTACTCCTTGC-3’, ribosomal protein lateral stalk subunit P0 (Rplp0) was amplified with forward primer 5’-TGGTCATCCAGCAGGTGTTCGA-3’ and reverse primer 5’- ACAGACACTGGCAACATTGCGG-3’, peptidylprolyl isomerase A (Ppia) was amplified with forward primer 5’-GGCAAATGCTGGACCCAACACA-3’ and reverse primer 5’- TGCTGGTCTTGCCATTCCTGGA-3’. The PCR reaction components were: 20 μL total volume, 2.5 μL of cDNA, 250 nM of each primer, and 10 μL of SsoAdvanced^™ Universal SYBR^® Green Supermix (Cat#1725271, Bio-Rad). The PCR cycling parameters were 95 °C 30 seconds, (95 °C 10 seconds, 61.5 °C 30 seconds) × 40 cycles, melt curve 65 °C to 95 °C (0.5 °C/5 seconds). Melt curves were inspected to rule out amplification of unrelated fragments during PCR. Cycle threshold was determined with the LinRegPCR analysis program (v11.0). Gene expression was calculated by the comparative threshold cycle (2-ΔΔCt) method. Briefly, ΔCt was calculated vs. the geometric mean of Rplp0 and Ppia, as stable expression of these two genes was previously found in human IBD (Krzystek-Korpacka et al., 2014). ΔΔCt was calculated against the ΔCt of an inter-run calibrator, represented by the normal samples of H2O + vehicle mice. 2-ΔΔCt was then log- transformed to arrive at the relative fold expression for each sample.

Immunoblotting –

Colonic crypt isolates or midsections from murine colon were homogenized in lysis buffer (20 mM Tris-HCl (pH 8.0), 100 mM NaCl, 0.1% (w/v) SDS, 0.9 mM EDTA, 0.5% (v/v) Triton X-100, 7 mM Na₃VO₄, 50 mM NaF, and cOmplete^™ Protease Inhibitor Cocktail). Homogenates were sonicated for 10 minutes at 4 °C and then clarified by centrifugation at 14,000 rpm for 10 minutes. Proteins were resolved by SDS-PAGE (12%) gel electrophoresis then transferred onto 0.2 μm nitrocellulose membranes. Blocking was done with 5% (w/v) non-fat milk in TBST, then membranes were incubated overnight in primary antibodies at 4 °C. Immunoreactive proteins were visualized using the ChemiDoc Imaging System (Bio-Rad Laboratories, Hercules, CA).

Analysis of single-cell RNAseq –

Single-cell RNAseq transcriptomic data were obtained from a recent genome-wide association study of risk alleles for UC (Smillie et al., 2019). The dataset represents transcripts identified in 366,650 cells from the colon mucosa of 18 UC patients and 12 healthy individuals. Transcript reads were processed and normalized using the Matrix (v1.7–2), SingleCellExperiment (v1.28.1), and scuttle (v1.16.0) R packages. Log-normalization was performed using the `logNormCounts()` command with a pseudo-count of one. Expression of DAPK3 and YAP1 across intestinal epithelial cell types was assessed using custom visualization functions built with ggplot2 (v3.5.2), applying global scales to compare expression patterns across health states. Non-zero expression was defined based on log-normalized counts. Spearman’s correlation (ρ ± 95% CI) between DAPK3 and YAP1 expression was calculated using the base `stats` package in R (v4.4.3), restricted to cell types containing at least 10 cells with non-zero expression of both genes. P-values were adjusted for multiple testing using the Benjamini– Hochberg method.

Statistical Analyses -

Data are presented as mean ± SEM, and the statistical analyses were performed with GraphPad Prism 8.3.0 (GraphPad Software, La Jolla, CA). Unless otherwise noted, the Student’s unpaired t-test was used to compare between 2 groups. Two-way ANOVA (repeated measures when applied to time course data) followed by Šídák or Tukey’s multiple comparison test was used to assess more than 2 groups. The Mann-Whitney U test was utilized to analyze non-parametric datasets (e.g., macroscopic scores). P values <0.05 were considered significant.

3. RESULTS

In vivo HS38 administration to mice increased susceptibility to DSS-induced colitis –

Mice exposed to HS38/DSS showed greater weight loss than the vehicle/DSS cohort (p<0.0001; Fig. 1B). Indeed, two of the mice co-treated with HS38/DSS were removed from the study due to severe weight loss that exceeded the humane endpoint criteria of the animal care protocol (i.e., weight loss of >15%). While the DAI (disease activity index) was not markedly impacted by HS38 treatment over the time-course of DSS provision, post hoc multiple comparison with repeated measures showed significant HS38-dependent change in DAI at DSS+2 (i.e., seven days of DSS treatment plus two days of recovery on normal drinking water in the presence or absence of HS38 treatment; p=0.039, Fig. 1C). Colon lengths (Fig. 1D) and macroscopic scores of colon injury (Fig. 1E) exhibited subtle, but non-significant, differences between HS38- and vehicle-treated mice at DSS+2. Histological analyses (Fig. 1F) revealed a small but significant increase in colitis severity for the HS38-treated mice, as per measures of epithelial damage and inflammation (Fig. 1G). These data suggest that HS38 can disrupt the protective signaling networks that maintain animal health in DSS-induced colitis. Interestingly, a pilot study designed to interrogate HS38 impact on recovery from DSS-induced colitis also found an HS38-dependent restriction on body weight restitution (Supplemental Fig. S1).

Figure 1. In vivo administration of HS38 increased susceptibility to DSS-induced colitis.

Male mice were subjected to experimental colitis (A) by providing 2.5% DSS in drinking water ad libidum for 7 days while receiving daily injections of HS38 (500 μg/kg) or vehicle (DMSO). Animals were then sacrificed after two days recovery on normal water (DSS+2). The percentage weight change (B) and disease activity index, DAI (C) were monitored over the 8 days when mice were co-treated with DSS and HS38 (n=12) or vehicle control (n=10). Statistical significance was determined via Šídák multiple comparison analysis after 2 way-ANOVA with repeated measures; error bars represent ± SEM, ** p<0.01, *** p<0.001. Colons were removed for length measurements (D) and macroscopic scoring (E). Transverse sections of colons were stained with H&E (F), and then histological scoring (G) was completed. Scale bars: upper panels, 300 μm; lower magnified image panels, 50 μm). The Mann-Whitney U test was used to determine statistical significance in macroscopic and histologic scores; * p<0.05.

Increased DAPK3 in IECs is identified with DSS-induced colitis –

The epithelial expression of Dapk3 in mouse IEC isolates was not impacted by DSS-induced colitis or HS38 treatments (Fig. 2A); however, DAPK3 protein abundance was significantly elevated in IECs from colonic crypts isolated from animals subjected to experimental colitis (DSS+2; Fig. 2B). This elevation in DAPK3 was not affected by HS38 treatment. Flow cytometry and gating with EpCaM (epithelial), CD31 (endothelial) and CD45 (hematopoietic) confirmed that >74% of cells in the cellular analytes were epithelial in nature (Supplemental Fig. S2). DAPK3 protein localization within the distal colon of mice was examined with immunohistochemistry (Fig. 2C). As expected, crypt atrophy and other morphological alterations were apparent in the colonic tissue of DSS+2 animals. Marked DAPK3 immunoreactivity was observed for epithelial cells in tissue sections dissected from both healthy (H2O) and colitis (DSS) cohorts. Punctate DAPK3 immunoreactivity, most prominently observed in the cytoplasm of IECs in the upper crypt zone and the columnar cells at the crypt base, was observed for the colonic epithelium of healthy (vehicle/H2O) animals. For vehicle/DSS+2 mice, demonstrable DAPK3 immunoreactivity was detected in the crypts of the damaged epithelium and was characterized by diffuse cytoplasmic staining. Taken together, these data support the concept that DAPK3 functions as a responding element within murine IECs during DSS-induced colitis.

Figure 2. Effect of HS38 and DSS-induced colitis on the colonic expression of DAPK3 in mice.

Mice were provided regular drinking water (H₂O) or DSS followed by two days recovery on normal water (DSS+2) while concurrently receiving either HS38 or vehicle (DMSO). Colonic crypts were isolated, then cell lysates were prepared and subjected to qRT-PCR for Dapk3 expression (A) or immunoblot analysis for DAPK3 protein (B). DAPK3 immunoreactivity was identified in the intestinal epithelium of mice (C). Cross-sections of the distal colon were fixed and probed with anti-DAPK3 and counterstained with DAPI. Scale bars, 50 μm.

IEC proliferation is attenuated with HS38 administration during DSS-colitis –

Crypts provide the niche for proliferating epithelial cells, which are responsible for the active self-renewal of the intestinal epithelium (Clevers, 2013). Distal-colon tissues of mice were paraffin-embedded, sectioned and probed with antibodies targeting Ki-67 as a marker of cellular proliferation (Fig. 3A, 3B). Healthy mice, provided normal drinking water instead of DSS, showed no significant differences in the Ki-67⁺ percentage score regardless of vehicle or HS38 treatment (Fig. 3C). In these animals, Ki-67⁺ cells were restricted to the crypt base and encompassed ~25% of total crypt IECs. In mice with colitis (DSS+2), the proportion of Ki-67⁺ cells increased to ~57% of total IECs per crypt, and coincident exposure to HS38 significantly decreased the number of proliferating (Ki-67⁺) epithelial cells by ~50% (Fig. 3C). A small decrease in absolute crypt length was observed in colons isolated from DSS-treated mice, but no significant differences for the HS38 treatments (regardless of DSS) were revealed for the average crypt length (Fig. 3D).

Figure 3. In vivo HS38 administration suppressed colonic IEC proliferation in DSS-induced injury.

Representative Ki-67 immunofluorescence images of colon tissues are provided. Mice were provided regular drinking water (A: Healthy, H₂O) or DSS followed by two days recovery on normal water (B: Colitis, DSS+2) while concurrently receiving HS38 or vehicle (DMSO). Scale bars = 50 μm. In (C), IEC proliferation was quantified and is provided as the percentage of Ki-67⁺ cells related to the total DAPI⁺ cells per crypt. In (D), the crypt length was measured in histological sections. Statistical significance was determined by 2-way ANOVA with Šídák post hoc multiple comparisons test ** p<0.01, *** p<0.001; n=3–8 animals. Error bars represent ± SEM.

HS38 administration was associated with decreased pS127-YAP phosphorylation in DSS-colitis –

The Hippo pathway provides essential regulatory control over the proliferative capacity of the intestinal epithelium (Xie et al., 2021), and previous reports suggest a putative genetic linkage between DAPK3 and the Hippo/YAP pathway (Chen and MacDonald, 2022). So, the effects of HS38 treatment on the Hippo effector proteins and upstream regulators were further examined in DSS-colitis. YAP levels appeared elevated in response to HS38 treatment in two-thirds of healthy mice (Fig. 4A); however, this difference did not reach statistical significance. YAP levels were not affected by HS38 and DSS co-treatment. Canonical Hippo signaling regulates YAP phosphorylation (Kwon et al., 2022), and the LATS1/2-dependent phosphorylation of the S127 phosphorylation site is associated with nuclear exclusion of YAP and reduced target gene expression. The phosphorylation of S397 promotes proteasomal-dependent degradation of YAP. HS38 administration was accompanied with decreased pS127-YAP levels (Fig. 4B), significantly so under conditions of DSS co-treatment. However, there was no significant effect of HS38 on pS397-YAP phosphorylation (Fig. 4C). A decrease in pT35-MOB1 phosphorylation was also observed with DSS-induced injury; however, no HS38-dependent effect on this MOB1 phosphorylation was apparent (Supplemental Fig. S3). None of the other upstream regulators showed significant changes in protein abundance upon HS38 treatment regardless of injury status, including MST1/2, LATS1 as well as MOB1.

Figure 4. HS38 treatment was associated with decreased pYAP-S127 phosphorylation in DSS-induced injury.

Colon tissue was collected from mice with DSS-induced colitis after two days recovery on normal water (DSS+2). Animals concurrently received HS38 or vehicle (DMSO). Tissue lysates were immunoblotted for YAP (A), pS127-YAP (B), and pS397-YAP (C). GAPDH was used as loading control for YAP whereas the pan-YAP signals were used to normalize the pS127-YAP and pS397-YAP levels. Statistical significance between vehicle and HS38 treatments was determined via unpaired Student’s t-test, * p<0.05. Error bars represent ± SEM, n=5–7.

Epithelial YAP nuclear localization is increased with HS38 treatment during DSS-colitis –

The reductions in pS127-YAP levels observed from immunoblots (Fig. 4B) provided motive to examine the effect of HS38 on the subcellular localization of YAP during DSS-colitis. Paraffin-embedded distal-colon tissues were probed with antibodies targeting total YAP to evaluate its subcellular localization (Fig. 5). In DSS+2 mice provided HS38 co-treatment, colonic IECs showed increased nuclear accumulation of YAP when compared to DSS+2 animals receiving vehicle (DMSO). This nuclear accumulation of YAP was primarily observed in IECs located at the crypt bottom, although significant nuclear immunofluorescence for YAP was also identified in the transit-amplifying compartment for cells moving from the crypts toward the lumen and crypt apex (Fig. 5, insets iii and iv, respectively).

Figure 5. Nuclear accumulation of YAP is found in colonic IECs with HS38 treatment during DSS-induced injury.

Colon tissue was collected from mice with DSS-induced colitis after two days recovery on normal water (DSS+2). Animals concurrently received HS38 or vehicle (DMSO). Tissue sections were immunostained for YAP (yellow) along with the nuclear marker DAPI (cyan). Scale bars: upper panels, 100 μm; magnified insets (i-iv), 20 μm.

4. DISCUSSION

Our findings advance DAPK3 as a putative regulator of intestinal epithelial repair and suggest a unique functional linkage between this kinase and the Hippo pathway effector YAP. In our study, prominent DAPK3 immunoreactivity was associated with crypt IECs, where the kinase may regulate processes including epithelial-mesenchymal transition (EMT), cell-cell adhesions, cell migration and actin cytoskeleton reorganization. In vivo administration of HS38, a novel structure-based pharmacophore for DAPK inhibition, impeded the normal epithelial proliferative response to injury and was associated with attenuated Hippo signaling (i.e., decreased S127-YAP phosphorylation and increased nuclear YAP localization). These molecular and cellular changes were accompanied by exacerbation of disease severity in the murine DSS-colitis model. Although this signaling profile aligns with YAP activation, epithelial proliferation was reduced in this inflammatory setting, suggesting that nuclear YAP can drive in vivo transcriptional programs that are distinct from canonical proliferation-associated outputs.

Additional evidence from patient biopsies supports the coordinated regulation of DAPK3 and YAP1 at the colonic mucosal surface. In the human colon, DAPK3 was broadly expressed across the epithelial compartment (Fig. 6A); the highest proportion of expressing cells was observed in inflamed enterocytes (17%), inflamed immature enterocytes 2 (17%), and stem cells (17%). YAP1 expression was most prominent in inflamed stem cells (20%) and non-inflamed stem cells (13%). Strong positive correlations between DAPK3 and YAP1 expression were observed in Best4+ enterocytes, immature enterocytes 2, stem cells, and transit-amplifying populations (Fig. 6B). Altogether, these findings suggest the existence of a DAPK3–YAP1 axis that influences epithelial repair outcomes in both health and disease.

Figure 6. Correlation of DAPK3 and YAP1 expression across intestinal epithelial cell types.

Single-cell RNAseq data were obtained from the Single Cell Portal (SCP#259, Smillie et al., 2019). The dataset includes transcriptional profiling of colonic mucosa biopsies from 18 UC patients and 12 healthy individuals. In (A), bubble plots show the percentage of cells with non-zero expression of DAPK3 and YAP1 across epithelial cell types in healthy, non-inflamed, and inflamed samples. In (B), the Spearman’s correlation (ρ ± 95% CI) between DAPK3 and YAP1 expression is provided for cell types containing at least 10 cells with non-zero expression of both genes.

The essential role of Hippo signaling in the repair of intestinal epithelial injury is well-documented (reviewed in: Hong et al., 2016; Xie et al., 2021). As the downstream effector of the Hippo pathway, the transcriptional regulator YAP is an important mediator of cell cycle entry and proliferation during crypt regeneration. The loss of YAP in IECs leads to impaired epithelial repair and worsened DSS colitis outcomes (Cai et al., 2010; Taniguchi et al., 2015). Yui and colleagues further demonstrated that colonic epithelium undergoes a YAP-dependent transition to a fetal-like state following injury to facilitate regeneration (Yui et al., 2018). Canonical Hippo activation involves LATS-dependent phosphorylation of YAP at S127, which stimulates its nuclear export and limits transcriptional activity (Kwon et al., 2022). In our study, HS38 treatment reduced pS127-YAP without affecting pS397-YAP, consistent with suppressed Hippo signaling, stabilization of nuclear YAP, and enhanced transcriptional potential. Despite observing a molecular profile favoring nuclear YAP accumulation, epithelial proliferation was reduced, suggesting that nuclear YAP does not always engage TEAD-driven proliferative programs in the inflamed epithelium (Zhao et al., 2023).

YAP transcriptional output is highly context-dependent. In keratinocyte and squamous epithelial models, YAP has been shown to cooperate with p73 to induce pro-apoptotic gene expression (Strano et al., 2005; Lapi et al., 2008). Consistent with this, JNK-dependent phosphorylation of YAP can shift its role from promoting survival in HaCaT keratinocytes to promoting apoptosis in squamous carcinoma cells (Tomlinson et al., 2010), highlighting the context dependence of YAP function. In gastric epithelial cells, optogenetic activation of YAP promotes wound closure predominantly through enhanced migration rather than proliferation (Toh et al., 2023). Under inflammatory conditions, nuclear YAP may engage alternative cofactors such as p73, RUNX, or SMADs to favor stress-response and pro-apoptotic programs (p73), pro-oncogenic programs (SMAD), or tumor-suppressive programs (RUNX) rather than canonical TEAD-driven proliferation (Ghomlaghi et al., 2024).

Past studies have shown that actomyosin integrity, comprising F-actin and non-muscle myosin II, is an important regulator of non-canonical Hippo/YAP signaling (Das et al., 2016; Wada et al., 2011; Nardone et al., 2017; Seo and Kim, 2018). Treatment of NIH-3T3 cells with F-actin destabilizers (cytochalasin D or latrunculin A) promoted cytoplasmic localization of YAP, while inhibition of actomyosin contractility reduced stress fiber formation and limited nuclear YAP retention (Wada et al., 2011). Because DAPK3 has established roles in actomyosin organization, focal adhesion assembly, and cell geometry (Komatsu and Ikebe, 2004; Turner et al., 2023), we examined its inhibition in Caco-2 monolayers. HS38 treatment qualitatively disrupted the intracellular F-actin network and increased the co-localization of YAP with cortical F-actin, while YAP nuclear localization was unchanged (Supplemental Fig. S4 & S5). HS38 also accelerated wound closure in the Caco-2 monolayer model (Supplemental Fig. S6) and was associated with increased proliferative responses (Supplemental Fig. S7).

Our in vitro findings indicate that inhibition of DAPK3 alters cytoskeletal architecture, influencing Hippo/YAP signaling and modifying epithelial behavior. This relationship also appears to operate in vivo, even though the phenotypic outcomes differ between systems. Such divergence is not unexpected since Caco-2 cells are derived from a colon adenocarcinoma and display transcriptomic and signaling profiles distinct from normal intestinal epithelium (Abbas et al., 2023; Sääf et al., 2007; Tadjali et al., 2002). In addition, the constraints of 2D monolayer culture likely affect YAP dynamics, as the translocation of YAP to the nucleus is greatly suppressed at high cell densities in a variety of cells, including IECs (Furukawa et al., 2017). This density dependence likely limits HS38-induced nuclear accumulation in our monolayer model. Taken together, these factors help explain why the phenotypic direction of YAP-associated outputs diverges between Caco-2 monolayers and the in vivo intestinal epithelium. Building on this context, our findings support a model in which DAPK3 regulates epithelial behavior through cytoskeleton-associated control of Hippo/YAP signaling, with its transcriptional and proliferative consequences shaped by the cellular and tissue environment.

We acknowledge limitations in the current study. Male mice were used exclusively, and future studies may wish to explore whether the contribution of DAPK3 signaling to epithelial repair mechanisms displays sex dimorphism (Bábíčková et al., 2015). The available human expression data for male and female samples from normal colon tissue would suggest that DAPK3 is not impacted by sex (Supplemental Fig. S8). However, an intriguing sex-dependency for the survival probabilities of high vs. low DAPK3 expression is observed for colon adenocarcinoma patients (Supplemental Fig. S8). Consistent with previously published findings, lower DAPK3 expression was associated with improved overall survival. When analyzed separately by sex, males exhibited a stronger survival benefit and greater statistical significance with low DAPK3 expression compared with the unstratified cohort, whereas the association between DAPK3 expression and overall survival was not statistically significant in females. Thus, empirical studies will be required to determine if sex hormone effects integrate with DAPK3 to alter UC disease activity and intestinal regeneration processes.

Importantly, it is necessary to consider the off-target impacts of HS38 on DAPK1 during IEC wound healing in vivo. Previous research applied a different DAPK inhibitor (i.e., TC-DAPK6) in the DSS-colitis model; mice were more susceptible to DSS-induced colitis as gauged by disease activity score, colon length, and weight loss (Lopes et al., 2018). Although no specific investigations of the epithelium were provided, bacterial translocation into the colonic mucosa was significantly increased by TC-DAPK6 co-treatment. So, these results also generically support DAPK family involvement in the maintenance of epithelial barrier; however, the TC-DAPK6 molecule displays a very concerning off-target liability. Namely, the molecule has equal potency toward ROCK (Al-Ghabkari et al., 2016), and ROCK is a well-documented contributor to intestinal epithelial barrier (Holvoet et al., 2017; Kangawa et al., 2017). Unfortunately, with regard to DAPK1, all currently available small molecule inhibitors demonstrate comparable potency. Thus, the crosstalk between DAPK3 and DAPK1 and the potential for overlapping molecular impact will need to be assessed in future studies.

While a genetic approach might be prioritized to dissect specific roles for DAPK1 and DAPK3 in the intestinal epithelium, global Dapk3 knockout is embryonically lethal for mice (Kocher et al., 2015). Similar lethality was observed in attempts to generate endothelial-cell specific Dapk3 knockout (Zhang et al., 2019). While inducible-Dapk3 global knockout and smooth-muscle specific Dapk3 knockout mice were successfully generated (Zhang et al., 2019), Dapk3 knockout cell lines created with CRISPR-Cas9 gene editing could not be maintained, suggesting an essential role for DAPK3 in cell growth and/or survival (Takahashi et al., 2021). In the absence of transgenic animal lines, other opportunities may exist to provide future in vivo manipulation of DAPK3 at the mucosal surface. Genetic engineering with intrarectal administration of siRNA (or shRNA) is possible with naked-siRNA as well as nanoparticle- or liposomal-complexed formulations (Chevalier, 2019). Alternatively, nonpathogenic-invasive E. coli (InvColi) was previously employed as an in vivo delivery vehicle for siRNA/shRNA in the mouse colon (Spisni et al., 2015). While these techniques could be used for the titration of DAPK3 at the colonic mucosal surface, they would not allow for interrogation of the cell-population specific roles of DAPK3 in epithelial homeostasis in vivo.

Indeed, the in vivo impact of HS38 on mice recovering from DSS-induced injury may be confounded by the functioning of Hippo/YAP in other cell types. A recent study suggests that YAP can drive macrophages toward M1 polarization while restricting M2 polarization during UC development (Zhou et al., 2019). While IEC proliferation was not specifically examined, myeloid-specific ablation of YAP was previously found to attenuate the severity of DSS-induced colitis; YAP impeded M2 polarization, promoted M1 polarization, and increased production of IL-6. Given that IL-6 signaling can trigger YAP activation via stimulation of gp130-associated Src family kinase Yes (Taniguchi et al., 2015), a positive feedback loop between YAP and IL-6 production by macrophages may exist. Potentially, DAPK3 inhibition with HS38 could elicit YAP activation in macrophages, which could subsequently impair M2 polarization to depress IEC proliferation post-DSS-induced injury. The regulation of macrophage programming by DAPK3 was also previously demonstrated (Mukhopadhyay et al., 2008); DAPK3 restrained interferon (IFN)-γ-induced gene expression in macrophages through its phosphorylation of the ribosomal protein L13a. The phosphorylation of L13a activated the IFN-γ-activated inhibitor of translation complex to repress translation of inflammatory genes (Mukhopadhyay et al., 2008). Thus, DAPK3 may also coordinate the macrophage-dependent resolution of inflammation in vivo and take part in a yet undefined counteracting mechanism that resets the positive feedback loop between YAP and IL-6 in macrophages.

Taken together, our findings suggest that DAPK3 contributes to epithelial regeneration at least in part by modulating Hippo/YAP signaling, but the transcriptional outputs of nuclear YAP, and thus the net proliferative outcome, are strongly shaped by the inflammatory and mechanical context. Future studies will be required to determine how DAPK3 influences YAP cofactor recruitment and transcriptional wiring in IECs to define the conditions under which YAP promotes regeneration versus alternative programs.