PTK6 drives HNRNPH1 phase separation to activate autophagy and suppress apoptosis in colorectal cancer

ABSTRACT

Macroautophagy/autophagy is the principal mechanism that mediates the delivery of various cellular cargoes to lysosomes for degradation and recycling, and has been reported to play a crucial role in colorectal cancer (CRC) pathogenesis and progression. Targeting autophagy may be a promising therapeutic strategy for CRC. However, the specific functions and potential mechanisms of autophagy in CRC remain unclear. In the present study, we discovered that PTK6 (protein tyrosine kinase 6) could activate autophagy and inhibit CRC apoptosis. PTK6 physically interacted with HNRNPH1 and mediated tyrosine phosphorylation at Y210 of HNRNPH1, which promoted the latter’s liquid-liquid phase separation (LLPS). Furthermore, LLPS of HNRNPH1 formed biomolecular condensates and triggered splicing-switching of the NBR1 exon 10 inclusion transcript, thereby activating autophagy and suppressing apoptosis of CRC. Additionally, PDO and CDX models indicated that tilfrinib, an inhibitor targeting PTK6, could inhibit CRC growth. Overall, our findings reveal the novel PTK6-HNRNPH1-NBR1 regulatory autophagy axis and provide a potential therapy target for CRC.

Abbreviation: 1,6HD: 1,6-hexanediol, CQ: chloroquine, CRC: colorectal cancer, DFS: disease-free survival, FRAP: fluorescence recovery afterphotobleaching, GSEA: Gene Set Enrichment Analysis, GTEx: Genotype-Tissue Expression, HNRNPH1: heterogeneous nuclearribonucleoprotein H1, IDRs: intrinsically disordered regions, IHC: immunohistochemical, KEGG: Kyoto Encyclopedia of Genes and Genomes,LLPS: liquid-liquid phase separation, NBR1: NBR1 autophagy cargoreceptor, OS: overall survival, PDO: patient-derivedorganoid, PTK6: protein tyrosine kinase 6, PTMs: post-translationalmodifications, SE: skipped exon, TCGA: The Cancer Genome Atlas, TEM: transmission electron microscopy, TMA: tissue microarray, TyrKc: tyrosine kinase catalytic.

Introduction

Over the past two decades, numerous studies suggest that activation of autophagy may promotes malignant progression in cancer [1,2]. Under normal conditions, autophagy could recycle damaged components to maintain homeostasis [2]. However, in the context of cancer, autophagy could provide nutrients to tumor cells, thereby promoting invasion, migration and adaptation to nutrient scarcity and hypoxia [3]. Concurrently, research has increasingly focused on the significance of autophagy in colorectal cancer (CRC) and its potential as a target for therapeutic interventions [4]. Although chloroquine (CQ) and other autophagy-targeting drugs have shown promise in inhibiting autophagosome degradation in CRC, they have not yet been approved for clinical use [5]. Therefore, there is an urgent need to further explore the molecular mechanisms of autophagy in CRC to develop new therapeutic strategies and improve patient prognosis.

PTK6 (protein tyrosine kinase 6) is a non-receptor tyrosine kinase characterized by the presence of a SRC homology 3 (SH3), a SRC homology 2 (SH2) domain, and a tyrosine kinase catalytic (TyrKc) domain, which enable it to phosphorylate various substrate proteins involved in signal transduction processes [6]. PTK6 has been implicated in regulating cell proliferation, apoptosis, and migration across different cancers [6]. For instance, PTK6 directly phosphorylates tyrosine 845 in the EGFR kinase domain, thereby promoting breast cancer cell proliferation [7]. Previous study founded that PTK6 could phosphorylate JAK2 to promote stemness and chemoresistance in CRC [8]. However, the molecular mechanisms underlying PTK6’s role in CRC autophagy remains unclear.

Phase separation refers to the dynamic liquid-liquid phase separation (LLPS) structures formed by specific protein or nucleic acid molecules through their interactions, which play crucial roles in cellular signaling and gene expression regulation [9]. LLPS relies on intrinsically disordered regions (IDRs) within proteins, which drive the formation of membraneless cellular compartments through multivalent weak interactions [10]. Previous studies have demonstrated that proteins undergo LLPS can regulate tumor migration and promote drug resistance in CRC [11,12]. However, the specific functions and mechanisms of LLPS in CRC autophagy remain inadequately studied.

In this study, we determined that upregulated PTK6 is associated with poor prognosis and promotes autophagy while inhibiting apoptosis in CRC. Further research revealed that PTK6 phosphorylates HNRNPH1, facilitating its LLPS. Additionally, phase-separated HNRNPH1 affects the inclusion of NBR1 exon 10, thereby modulating the autophagy process. Our findings uncover a novel PTK6-HNRNPH1-NBR1 regulatory axis in autophagy, offering a potential therapeutic target for CRC.

Results

Elevated PTK6 expression in CRC is associated with clinical-pathological features and patient prognosis

To investigate the role of autophagy in the development of CRC, we identified 222 autophagy-related genes using the Human Autophagy Database (HADb, http://www.autophagy.lu/). Then, we utilized the GEPIA2 database in conjunction with the publicly available dataset GSE138202 to identify genes that are upregulated in CRC. By integrating the data from these sources, four key molecules were discovered: PTK6, SERPINA1, NKX2–3, and BIRC5 (Figure 1A). Further analysis of their relationship with clinical prognosis was performed using univariate Cox regression analysis, which revealed that PTK6 exhibits a higher level of malignancy compared to the other genes (Figure 1B).

Figure 1.

Elevated PTK6 expression in CRC is linked to clinical-pathological features and patient prognosis. (A) Venn diagram showing the intersection of crc-related genes from GEPIA2 and GSE138202 databases with autophagy-related genes from HADb. (B) Forest plot illustrating the prognosis of PTK6, SERPINA1, NKX2–3, and BIRC5. (C) Expression levels of PTK6 in tumor and adjacent non-tumor tissues observed using TCGA combined with GTEx databases. (D) Expression levels of PTK6 across different clinical stages. (E) Overall survival (OS) analysis of PTK6 using Kaplan-Meier (K-M) method. (F) PTK6 expression levels in 53 pairs of cancerous and adjacent adjacent non-tumor tissues measured by western blot analysis. (G,H) IHC scores and representative images of PTK6 in tumor and adjacent non-tumor tissues from 73 paired tissue microarrays (TMAs). (I) analysis of OS of patients in TMAs based on PTK6 expression levels. (J,K) Univariate and multivariate analyses performed for CRC patients in TMAs. Bars indicate mean ± SE, *p < 0.05, **p < 0.01, and ***p < 0.001 compared with the control.

The Cancer Genome Atlas (TCGA) and the Genotype-Tissue Expression (GTEx) datasets, in conjunction with the GEPIA2 database (http://gepia.cancer-pku.cn/), demonstrated that PTK6 was significantly upregulated in colon adenocarcinoma/COAD and rectum adenocarcinoma/READ (Figure 1C and Figure S1A). Moreover, Kaplan-Meier (K-M) analysis indicated that patients exhibiting high PTK6 expression have a poorer prognosis, with its expression correlating with clinical staging (Figure 1D,E).

Subsequently, we examined the expression of PTK6 in 53 pairs of fresh CRC and adjacent normal tissue samples. The results showed that both protein and mRNA expression levels of PTK6 were significantly higher in CRC tissues compared to normal tissues (p < 0.001) (Figure 1F and Figure S1B). Additionally, we conducted immunohistochemical (IHC) analysis on a tissue microarray (TMA) comprising 73 pairs of tissues (Figure 1G). The results revealed that PTK6 IHC scores for tumor tissues were significantly higher than those for adjacent normal tissues (Figure 1H).

To further investigate the clinical significance of PTK6 in the TMA cohort, we performed Chi-square tests to analyze the relationship between PTK6 expression and various clinicopathological features. The results demonstrated that high PTK6 expression was positively correlated with tumor size (p = 0.001), depth of invasion (p < 0.001), lymph node metastasis (p = 0.011), distant metastasis (p = 0.017), and TNM stage (p = 0.006) (Supplementary Table S1). Additionally, we employed K-M survival curves and Cox regression analyses to evaluate the prognostic value of PTK6 in CRC. The K-M analysis indicated that patients with high PTK6 expression had poor overall survival (OS) and disease-free survival (DFS) compared to those with low PTK6 expression (p = 0.0012; p = 0.0001) (Figure 1I and Figure S1C). Univariate Cox regression analysis revealed that PTK6 expression, TNM stage, lymph node metastasis, and distant metastasis were significant risk factors affecting the prognosis of CRC patients (Figure 1J). Moreover, multivariate Cox analysis established that PTK6 expression and TNM stage were functioned as independent prognostic factors (Figure 1K).

Simultaneously, we performed Gene Set Enrichment Analysis (GSEA) on the gene expression matrix of CRC patients obtained from the TCGA database, comparing groups with high and low PTK6 expression. The GSEA results indicated that PTK6 expression was positively correlated with autophagy regulation pathways (Figure S1D). These findings suggest that PTK6 is upregulated in CRC, and this overexpression is associated with the clinicopathological features and prognosis of CRC patients.

PTK6 inhibits apoptosis in CRC by activating autophagy both in vitro and in vivo

To explore the biological function of PTK6 in CRC, we initially analyzed its expression across various cell lines. Our results indicated that HCT116 exhibited the most pronounced increase in PTK6 expression, while DLD1 displayed the least significant increase (Figure S2A). Consequently, we selected these two cell lines for subsequent experiments involving both knockdown and overexpression of PTK6. We generated two shRNA constructs targeting PTK6 (sh-PTK6#1,sh-PTK6#2) and a PTK6 overexpression vector (PTK6[OE]).

Transmission electron microscopy (TEM) revealed a marked increase in the number of autophagosome upon PTK6 overexpression in both HCT116 and DLD1 cell lines (Figure 2A). Western blot analysis further showed that PTK6 overexpression led to a significant increase in the LC3B-II:LC3B-I ratio and BECN1 levels, accompanied by a decrease in SQSTM1/p62 expression, indicating enhanced autophagic flux. Conversely, PTK6 knockdown resulted in a reduction autophagic activity (Figure 2B and Figure S2B). Immunofluorescence staining for LC3B corroborated these findings, demonstrating elevated LC3B puncta in PTK6-overexpressing cells (Figure 2C and Figure S2C). To further evaluate the role of PTK6 in autophagy, we assessed autophagic flux using the mCherry-GFP-LC3 reporter with or without chloroquine (CQ), a lysosomal inhibitor that blocks autophagic degradation. The data showed that PTK6 enhanced autophagic flux, both in the absence and presence of CQ (Figure 2D and Figure S2D).

Figure 2.

PTK6 inhibits apoptosis in CRC by activating autophagy both in vitro and in vivo. (A) TEM observation of autophagosomes in HCT116 and DLD1. (B) Western blot analysis of the effects of PTK6 on BECN1, SQSTM1, and LC3B in HCT116. (C) immunofluorescence analysis of LC3B in HCT116 under different treatments. (D) Examination of autophagic flux with the mCherry-EGFP-LC3 reporter in HCT116. (E,F) Analysis of the role of PTK6 in apoptosis via flow cytometry and statistical analysis in HCT116. (G) Analysis of the role of PTK6 in apoptosis via western blot analysis in HCT116. (H,I) Analysis of the effect of PTK6 on apoptosis after the addition of CQ (20 μM) via flow cytometry and statistical analysis in HCT1116. (J) Analysis of the effect of PTK6 on apoptosis after the addition of CQ (20 μM) v via western blot analysis in HCT116. (K,L) Establishment of a xenograft tumor model experiment in mice using treated HCT116 (n = 5), with tumor weight and volume monitored over a 24-day period. (M) Immunohistochemistry for PTK6, LC3B, SQSTM1, cleaved CASP3 and MKI67. **p < 0.01, and ***p < 0.001 compared with the control.

Previous studies have indicated that tumor cells under stress conditions activate autophagy to provide nutrients and inhibit apoptosis [13]. Flow cytometry and western blot analysis revealed that PTK6 overexpression significantly inhibited apoptosis, whereas PTK6 knockdown had the opposite effect (Figure 2E–G and Figure S2E-G). Additionally, flow cytometry and western blot analysis confirmed that PTK6 inhibits apoptosis by promoting autophagy (Figure 2H–J and Figure S2H-J).

To further validate the function of PTK6 in vivo, we performed a xenograft tumor model experiment in mice, harvesting the tumors after 24 days for IHC analysis. We generated stable cell lines with either PTK6 knockdown or overexpression, which were subsequently subcutaneously injected into nude mice. Two independent experimental groups, along with their corresponding control groups, were established for comparison. Tumor volumes were measured periodically. The results showed that PTK6 significantly promoted tumor growth, with both tumor volume and weight being markedly higher than those in the control groups (p < 0.001) (Figure 2K). Conversely, PTK6 knockdown resulted in a significant reduction in tumor growth (p < 0.001) (Figure 2L). TUNEL and IHC analyses further substantiated our in vitro findings, revealing that PTK6 directly enhances autophagy while concurrently suppressing apoptosis in tumor cells (Figure 2M and Figure S2K,L).

Previous studies have shown that under protective autophagy, the majority of cleaved CASP8 (caspase 8) is sequestered by autophagosomes and subsequently degraded by lysosomes [14]. To investigate whether autophagy inhibits apoptosis in CRC by selectively degrading pro-apoptotic proteins, we used CQ to block autolysosome degradation. The apoptosis of CRC cells following CQ treatment was assessed through flow cytometry and TUNEL assays (Figure S2M,N). The results demonstrated that blocking autolysosome degradation significantly induced CRC apoptosis. Additionally, we analyzed the expression levels of cleaved CASP9, CASP3, cleaved CASP3, cleaved CASP8, BAX, and BCL2 in CRC cells with CQ treatment (Figure S2O). We observed that CQ treatment increased the expression of cleaved CASP3 and cleaved CASP8, but it had no significant effect on cleaved CASP9 levels. Then, the results of western blot indicated that PTK6-mediated autophagy decreased the expression of cleaved CASP8 to inhibit CRC apoptosis (Figure S2P).

Overall, these findings demonstrate that PTK6 enhances autophagy and subsequently inhibit apoptosis in both in vitro and in vivo models. Furthermore, PTK6-mediated autophagy selectively degrades cleaved CASP8, effectively blocking the apoptotic cascade.

PTK6 activates autophagy and inhibits apoptosis through HNRNPH1

To further elucidate the role of PTK6 in promoting autophagy, we utilized immunoprecipitation-mass spectrometry (IP-MS) technology (Figure 3A), identifying 1363 potential PTK6-binding proteins. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis revealed that PTK6 is enriched in the spliceosome pathway (Figure S3A). As a result, we identified 53 proteins belonging to various key splicing factor families, including the SRSF (serine and arginine rich splicing factor) family, HNRNP (heterogeneous nuclear ribonucleoprotein) family, U small nuclear ribonucleoprotein/U snRNP family, polypyrimidine tract-binding/PTB family, poly (rC)-binding protein/PCBP family, as well as other notable splicing factors. By integrating the IP-MS data with splicing factor information, we identified 19 splicing factors associated with PTK6. Among these, HNRNPH1, which exhibited the highest Sequest HT score, was selected as the most reliable splicing factor (Figure 3B). Co-immunoprecipitation (co-IP) assays further validated the interaction between PTK6 and HNRNPH1 (Figure 3C,D).

Figure 3.

PTK6 activates autophagy and inhibits apoptosis through HNRNPH1. (A) coomassie brilliant blue staining demonstrating the protein profiles isolated by immunoprecipitation using an anti-PTK6 antibody. (B) Venn diagram (left) showing intersections between PTK6 mass spectrometry results and splicing factors. HNRNPH1 identified as a potential interacting protein through sequest HT score prioritization (right). (C,D) Co-immunoprecipitation of PTK6 and HNRNPH1 (HNR) in HCT116 and DLD1. (E) Differential splicing events associated with HNRNPH1 were analyzed through rna-seq, followed by KEGG analysis to identify the genes involved in these alternative splicing events. (F) Examination of autophagic flux with the mCherry-EGFP-LC3 reporter in HCT116. (G) Western blot analysis of the effects of HNRNPH1 on BECN1, SQSTM1, and LC3B in HCT116. (H,I) Analysis of the effect of HNRNPH1 on apoptosis after the addition of CQ (20 μM) via flow cytometry and statistical analysis in HCT1116. (J) Flow cytometry analysis of apoptosis in HCT116, comparing control, sh-PTK6, and sh-PTK6+HNRNPH1[OE] groups. (K) Statistical analysis of apoptosis in HCT116 by flow cytometry. (L) Western blot analysis of HCT116 divided into three groups: control, sh-PTK6 and sh-PTK6+HNRNPH1[OE], detecting for BECN1, SQSTM1, LC3B, cleaved CASP3, BAX, and BCL2 to assess the effects on autophagy and apoptosis. (M) Analysis of autophagic flux using the mCherry-EGFP-LC3 reporter in HCT116, comparing three groups: control, sh-PTK6 and sh-PTK6+HNRNPH1[OE]. **p < 0.01, and ***p < 0.001 compared with the control.

To determine the function of HNRNPH1, we performed RNA-seq in HCT116 transfected with sh-HNRNPH1 and sh-NC (Figure S3B). Notably, KEGG analysis of cellular processes revealed a significant associated between HNRNPH1 and both autophagy and apoptosis (Figure 3E). To further verify the role of HNRNPH1 in autophagy, we conducted the mCherry-EGFP-LC3 dual fluorescence reporter assay, demonstrating that HNRNPH1 increases autophagic flux (Figure 3F and Figure S3C). This observation was corroborated by western blot analysis (Figure 3G and Figure S3D). Subsequently, we used flow cytometry to assess the role of HNRNPH1 in apoptosis. Following the addition of CQ, we found that HNRNPH1 promotes autophagy, consequently inhibiting apoptosis (Figure 3H,I and Figure S3E,F). To investigate the relationship between PTK6 and HNRNPH1 in regulating autophagy and apoptosis, we used western blot analysis, mCherry-EGFP-LC3 dual fluorescence reporter assay, and flow cytometry (Figure 3J–M and Figure S3G-J). These results demonstrated that HNRNPH1 overexpression can rescue the effects of PTK6 knockdown on autophagy and apoptosis. In summary, these findings suggest that PTK6 May promote autophagy and inhibit apoptosis through the action of HNRNPH1.

PTK6 phosphorylates HNRNPH1 at the Y210 site

To investigate the underlying mechanisms, we first performed a Mn²⁺-Phos-tag SDS-PAGE experiment to verify whether PTK6, a protein tyrosine kinase, can phosphorylate HNRNPH1, thereby promoting its biological functions in CRC. The results indicated that the overexpression of PTK6 significantly increased the phosphorylation level of HNRNPH1, while the knockdown of PTK6 yielded the opposite effect (Figure 4A). Using PhosphoSitePlus (https://www.phosphosite.org) and NetPhos3.1 (https://services.healthtech.dtu.dk/services/NetPhos-3.1/), we identified 19 potential phosphorylation sites and 8 tyrosine phosphorylation sites on HNRNPH1, with 4 overlapping tyrosine phosphorylation sites (Y210, Y253, Y276, Y306) identified (Figure 4B). Consequently, we mutated the four potential phosphorylation sites Y210, Y253, Y276, and Y306 of HNRNPH1 to alanine, creating the Y210A, Y253A, Y276A, and Y306A mutants (Figure 4C). Following this, we co-transfected PTK6[OE] with the various HNRNPH1 mutants. After transfection, we purified the HNRNPH1 proteins using IP and then incubated the purified HNRNPH1 with a pan-tyrosine antibody to evaluate the phosphorylation levels. Western blot analysis showed a significant increase in HNRNPH1 phosphorylation in cells co-transfected with PTK6[OE] and WT HNRNPH1. Interestingly, the three mutants (HNRNPH1^Y253A, HNRNPH1^Y276A, and HNRNPH1^Y306A) did not affect PTK6-mediated phosphorylation of HNRNPH1 (Figure 4D,E). However, in cells co-transfected with the HNRNPH1^Y210A, the effect of PTK6 on the phosphorylation of HNRNPH1 disappeared, indicating that Y210 is a crucial site for PTK6-mediated phosphorylation of HNRNPH1 (Figure 4D,E). Additionally, in vitro kinase assays corroborated that Y210 is the essential site for PTK6-mediated phosphorylation of HNRNPH1, while mutations at Y253, Y276, and Y306 did not affect its phosphorylation level (Figure 4F).

Figure 4.

PTK6 phosphorylates HNRNPH1 at Y210 site. (A) Phos-tag SDS-PAGE analysis (upper image) demonstrated the impact of PTK6 on the phosphorylation of HNRNPH1, while conventional SDS-PAGE (lower image) showed no bands at the same position. (B) Phosphorylation sites on HNRNPH1 were predicted using PhosphoSitePlus and NetPhos 3.1 tools. (C) Diagram illustrating the HNRNPH1 protein with wild type and the mutant variants (tyrosine replaced by alanine). (D,E) Analysis of the effects of PTK6 on various HNRNPH1 mutants using western blot in HCT116 and DLD1. (F) an in vitro phosphorylation assay was employed to explore the effects of PTK6 on various HNRNPH1 mutants. (G) Diagram illustrating the truncated form of the PTK6 protein. (H) Western blot analysis of IP assay in HCT116 and DLD1 showing the interactions between PTK6 and HNRNPH1.

As previously described, PTK6 consists of three domains: SH3, SH2, and tyrosine kinase catalytic domain (TyrKc). To identify the domain involved in the binding of PTK6 to HNRNPH1, we cloned a series of Flag-tagged truncated PTK6 proteins, including full-length PTK6, PTK6[ΔSH2] (lacking the SH2 domain), and PTK6[ΔSH3] (lacking the SH3 domain) (Figure 4G). Immunoprecipitation and immunoblotting results revealed that both WT PTK6 and PTK6[ΔSH3] significantly bound to HNRNPH1, while PTK6[ΔSH2] exhibited no detectable binding (Figure 4H).

Combining these results with the phosphorylation mechanism of PTK6, these findings demonstrate that PTK6 binds to HNRNPH1 via its SH2 domain and mediates HNRNPH1 Y210 phosphorylation through its TyrKc.

HNRNPH1 exhibits LLPS capabilities both in vitro and in vivo

Phosphorylation, as a crucial post-translational modification (PTM), significantly affects protein localization by adding phosphate groups to serine, threonine, or tyrosine residues [15]. Interestingly, upon transfection of cells with EGFP-HNRNPH1 and EGFP-HNRNPH1^Y210A, both bright-field and fluorescence microscopy revealed that HNRNPH1 formed droplet-like structures, with a notable reduction in droplet number observed in the EGFP-HNRNPH1^Y210A (Figure 5A and Figure S4A).

Figure 5.

HNRNPH1 exhibits phase separation capabilities both in vivo and in vitro. (A) EGFP-HNRNPH1 and EGFP-HNRNPH1^Y210A were observed under fluorescence and brightfield microscopy in HCT116. (B) 3D-reconstructed live cell images of EGFP-HNRNPH1 were observed using a confocal laser scanning microscope (CLSM, Leica STELLARIS 5, Germany) in HCT116. (C) Two EGFP-HNRNPH1 droplets fuse to form a larger droplet in HCT116. (D) EGFP-HNRNPH1 droplets were disrupted by exposure to 10% 1,6-hexanediol and subsequently recovered after the removal of 1,6-hexanediol in HCT116. (E) Left images show the recovery of EGFP-HNRNPH1 fluorescence in HCT116 after bleaching with high-intensity laser light over time; the right plot shows the quantification of fluorescence intensity. Data shown as mean ± SE. (n = 3 droplets). (F) Representative fluorescence images demonstrate the phase separation behaviors of purified EGFP-HNRNPH1 at various NaCl concentrations, supplemented with 5% PEG8000. (G) 3D-reconstructed images of in vitro purified EGFP-HNRNPH1 were observed using a confocal laser scanning microscope. (H) Two in vitro purified EGFP-HNRNPH1 droplets fuse to form a larger droplet. (I) Images (left) and quantification (right) of FRAP for in vitro purified EGFP-HNRNPH1. Data shown as mean ± SE (n = 3 droplets). (J) If analysis (left) and quantification (right) of HNRNPH1 puncta in CRC and normal tissues. (K) If analysis (left) and quantification (right) of HNRNPH1 puncta in primary cell lines from CRC and normal colonic epithelial tissues. **p < 0.01, and ***p < 0.001 compared with the control.

To confirm the liquid-liquid phase separation phenomenon of HNRNPH1, we noted that EGFP-HNRNPH1 formed droplet-like puncta in both HCT116 and DLD1 cell lines, with individual droplets merging into larger droplets (Figure 5B,C and Figure S4B,C). Fluorescence recovery after photobleaching (FRAP) further demonstrated that EGFP-HNRNPH1 exhibited liquid-like dynamic properties (Figure 5E and Figure S4G). To validate these observations, we introduced 1,6-hexanediol (1,6-HD) as a LLPS inhibitor. Research showed that 1,6-HD interferes with weak hydrophobic interactions between proteins or between protein and RNA, thereby inhibiting LLPS [16]. Upon treatment with 1,6-HD, the number of HNRNPH1 droplets significantly decreased or even disappeared, and the droplets gradually reformed after the removal of 1,6-HD (Figure 5D and Figure S4D-F).

We purified His-tagged EGFP-HNRNPH1 to investigate whether HNRNPH1 undergoes LLPS in vitro. The results demonstrated that the purified protein formed droplet-like puncta in salt solutions, with the number of droplets decreasing as the salt concentration increased (Figure 5F,G). Additionally, the number of droplets showed a positive correlation with the protein concentration (Figure 5F). Observations from both in vitro experiments and FRAP analyses confirmed that these droplets displayed dynamic, liquid-like behavior (Figure 5H,I).

It is crucial to clarify whether there are differences in HNRNPH1 phase separation between CRC and normal tissues. Therefore, we used tissue sections and IF experiments to study the number of HNRNPH1 puncta in CRC and normal tissues. The results showed that HNRNPH1 puncta was significantly more abundant in CRC tissues compared to normal tissues (Figure 5J). Additionally, we established primary cell lines from CRC and normal colonic epithelial tissues and performed IF experiments. The results indicated that HNRNPH1 puncta was significantly more abundant in CRC primary cells than in normal colonic epithelial primary cells (Figure 5K). These findings may be attributed to the high expression of PTK6, which mediates increased phosphorylation of HNRNPH1, leading to different phase separation states.

Overall, our findings suggest that HNRNPH1 can undergo phase separation both in vitro and in vivo, which may be associated with PTK6 phosphorylation.

PTK6 regulates the LLPS capability of HNRNPH1 by phosphorylating Y210

To investigate the relationship between LLPS and phosphorylation, we first identified the IDRs of HNRNPH1 using IUPred2A (https://iupred2a.elte.hu/) (Figure 6A). Notably, we found that Y210 is situated within the IDRs region. To assess the impact of Y210 on LLPS, we constructed four plasmids encoding different variants of EGFP-tagged HNRNPH1: the full-length HNRNPH1 protein (EGFP-HNRNPH1), a deletion mutant lacking amino acids 206–246 within the LC1 domain (EGFP-HNRNPH1[Δ206–246]), a point mutation with tyrosine 210 replaced by alanine within the LC1 domain (EGFP-HNRNPH1^Y210A), and a chimeric construct with the IDRs region of hnRNPA1 added to the EGFP-HNRNPH1^Y210A mutant (EGFP-HNRNPH1[IDR]) (Figure 6B). Transfection experiments in cells revealed that both EGFP-HNRNPH1[Δ206–246] and EGFP-HNRNPH1^Y210A displayed similar LLPS droplet behavior, while the addition of the IDRs restored droplet formation (Figure 6C and Figure S4H,I). Additionally, in vivo observations and FRAP experiments demonstrated that HNRNPH1 devoid of the IDRs exhibited diminished phase separation and prolonged recovery time (Figure 6D,E and Figure S4J).

Figure 6.

PTK6 regulates the phase separation capability of HNRNPH1 by phosphorylating Y210. (A) The disordered region of HNRNPH1 was analyzed with IUPred2A. (B) Schematic representation of four plasmids encoding different variants of egfp-tagged HNRNPH1. (C) Confocal microscopy images of cells transfected with different HNRNPH1-expressing plasmids in HCT116. (D,E) Images and quantification of EGFP-HNRNPH1 and EGFP-HNRNPH1[Δ206–246] FRAP in HCT116. Data are presented as mean ± SE (n = 3 droplets). (F,G) Images and quantification of EGFP-HNRNPH1, EGFP-HNRNPH1^Y210A, and EGFP-HNRNPH1[IDR] FRAP, as well as their FRAP following PTK6 overexpression in HCT116. Data are presented as mean ± SE (n = 3 droplets).

To further investigate the role of PTK6 in regulating the LLPS ability of HNRNPH1 via phosphorylation at the Y210 site, we conducted several FRAP experiments and quantified the number of condensates (Figure 6F,G and Figure S4K,L). The results indicated that the fluorescence recovery time of EGFP-HNRNPH1^Y210A was prolonged after bleaching, indicating that this mutation adversely affected the LLPS capability of HNRNPH1 (Figure 6F,G and Figure S4K,L). Notably, when co-expressed with PTK6, the fluorescence recovery exhibited minimal change, thereby reinforcing the notion that phosphorylation at the Y210 site is essential for PTK6’s regulation of HNRNPH1 LLPS (Figure 6F,G and Figure S4K). PTK6 and EGFP-HNRNPH1 co-expression results in the highest number of condensates, whereas PTK6 does not significantly influence the LLPS of EGFP-HNRNPH1^Y210A (Figure S4L). Furthermore, in the case of EGFP-HNRNPH1[IDR], HNRNPH1 regained its LLPS capability, with PTK6 showing no effect on this process (Figure 6F,G and Figure S4K,L). These findings collectively suggest that PTK6 modulates the LLPS ability of HNRNPH1 through the phosphorylation of the Y210 residue.

PTK6 promotes HNRNPH1 LLPS, which leads to NBR1 exon retention, resulting in activation of autophagy and inhibition of apoptosis

To further elucidate the mechanism by which the PTK6-HNRNPH1 axis drives autophagy in CRC, we conducted HNRNPH1-RNA-seq analysis. Considering the role of HNRNPH1 in RNA splicing, we analyzed its alternative splicing events, a key mechanism that generates diverse mRNA variants and protein isoforms (Figure S5A). RNA-seq analysis identified a total of 81,178 hNRNPH1-related alternative splicing events, including 60,280 skipped exon (SE) events 12,400 mutually exclusive exon/MXE events, 929 intron retention/IR events, 4757 alternative 3’splice site/A3SS events, and 2812 alternative 5’ splice site/A5SS events (Figure S5B). Thus, we focused on SE events, which are the predominant splicing events regulated by HNRNPH1.

Firstly, we identified 11,496 SE-related events (p < 0.05) from the HNRNPH1-regulated alternative splicing events (n = 16375; p < 0.05) (Figure 7A). By integrating data from autophagy databases and key autophagy genes, we determined 83 relevant genes [17]. Among these, we further merged 14 genes regulated by HNRNPH1 through SE events (Figure 7A). Notably, the SE event with the highest inclusion level difference score was the skipping of exon 10 in NBR1. Furthermore, differential alternative splicing analysis conducted with rMATS indicated showed that HNRNPH1 knockdown significantly increased the expression of the exon 10 skipping variant of NBR1 (Figure 7B). Additionally, we designed primers to investigate the impact of HNRNPH1 on this splicing event (Figure 7C and Figure S5C). The RT-PCR validation products were consistent with the sequencing results.

Figure 7.

PTK6 promotes HNRNPH1 LLPS, which leads to NBR1 exon retention, resulting in activation of autophagy and inhibition of apoptosis. (A) Flow chart for screening HNRNPH1 candidate target genes. (B) Sashimi plots illustrating the alternatively spliced transcripts of NBR1 in HCT116, analyzed using rMATS. (C) Analysis of the percent of NBR1 exon 10+ transcripts in HCT116 after HNRNPH1 knockdown and overexpression using semi-quantitative PCR. (D) Analysis of the percent of NBR1 exon 10+ transcripts in HCT116 expressing EGFP-HNRNPH1 and EGFP-HNRNPH1[Δ206–246] using semi-quantitative PCR. (E) Semi-quantitative PCR analysis of the percent of NBR1 exon 10+ transcripts in HCT116 across the five experimental groups. (F) If analysis of HNRNPH1, SRRM2, and SON in HCT116 and DLD1. (G) The motif plot (up) shows the preferred binding sequence of HNRNPH1, and primers designed for potential binding regions (down). (H) Clip-qPCR analysis of HNRNPH1 association with NBR1 pre-rna at the indicated locations. (I) the diagram of the NBR1 (WT NBR1 and NBR1[ΔExon 10]). (J-P) Western blot analysis, mCherry-EGFP-LC3 dual fluorescence reporter assay, and flow cytometry were utilized to assess the effects of WT NBR1 and NBR1[ΔExon 10] on PTK6-mediated autophagy and apoptosis in HCT116 and DLD1. Bars represent mean ± SE; ns indicates no significance, and ***p < 0.001 compared to the control.

To further determine whether LLPS affects alternative splicing events, RT-PCR analysis revealed that HNRNPH1 promoted the inclusion of NBR1 exon 10, whereas the EGFP-HNRNPH1[Δ206–246] did not (Figure 7D and Figure S5D). Moreover, we conducted co-transfection experiments to validate the role of the PTK6-HNRNPH1 axis in NBR1 alternative splicing. The results showed that PTK6 knockdown increased the skipping of NBR1 exon 10. However, overexpression of WT HNRNPH1 rescued this effect on NBR1 splicing (Figure 7E and Figure S5E). In contrast, the expression of the mutant HNRNPH1^Y210A failed to rescue this splicing change. Notably, restoring the LLPS function of HNRNPH1 also rescued the skipping of NBR1 exon 10 (Figure 7E and Figure S5E). These findings collectively indicate that PTK6 promotes the LLPS of HNRNPH1 through phosphorylation at the Y210 site, thereby regulating NBR1 alternative splicing.

It is known that nuclear speckle is a membraneless organelle, which is vital to the splicing of mRNAs [18]. Previous studies have confirmed that SRRM2 and SON are involved in the formation of nuclear speckles [19]. Therefore, we used these two proteins as markers for nuclear speckles. IF experiments showed that HNRNPH1, as an alternative splicing factor, co-localizes with SRRM2 and SON (Figure 7F). To identify the HNRNPH1 binding motif in NBR1 exon 10, we performed motif analysis using Jalview software on all SE events, revealing that HNRNPH1 appears to prefer a specific sequence pattern (Figure 7G). To investigate the binding motifs of HNRNPH1 in NBR1 pre-RNA, we conducted in vivo crosslinking followed by immunoprecipitation and RT-PCR assays (Figure 7H). Our results revealed that HNRNPH1 predominantly binds to motifs within the region centered around exon 10 of NBR1.

We further constructed plasmids for WT NBR1 and NBR1 with exon 10 deletion (NBR1[ΔExon 10]) (Figure 7I). As demonstrated by the mCherry-EGFP-LC3 dual fluorescence reporter assay, flow cytometry, and western blot analysis, WT NBR1 significantly promoted autophagic flux, while NBR1 lacking exon 10 was unable to perform this function (Figure S5F-H). Furthermore, WT NBR1 rescued the effects of PTK6 on autophagy and apoptosis, whereas NBR1[ΔExon 10] could not (Figure 7J–P and Figure S5I).

To investigate how NBR1 participates in autophagy regulation, we first examined the cellular localization of WT NBR1 and NBR1[ΔExon 10] and found no significant differences in their localization (Figure S5J). In addition, our study demonstrates that HNRNPH1 mediates the retention of NBR1 exon 10, which encodes amino acids 288–347 and fully encompasses the CC1 domain (288–329). Consequently, based on the domains of NBR1, we generated NBR1 truncation mutants and conducted immunoprecipitation assays with SQSTM1/p62 (Figure S5K). Deletion of the CC1 domain significantly impaired the binding of NBR1 to SQSTM1 (Figure S5L). In-depth analysis revealed that NBR1 can bind to SQSTM1, while the binding capacity of NBR1[ΔExon 10] is significantly reduced (Figure S5M). Furthermore, overexpression of WT NBR1 led to a decrease in the number of SQSTM1 bodies (Figure S5N). These findings suggest that the interaction between NBR1 and SQSTM1, mediated by the CC1 domain, is critical for autophagy. Furthermore, NBR1-SQSTM1 interactions are essential for maintaining cellular homeostasis through autophagy and indirectly inhibit apoptosis.

In conclusion, our findings suggest that the PTK6-HNRNPH1 axis promotes the inclusion of exon 10 in NBR1, thereby enhancing autophagy and subsequently inhibiting apoptosis.

Tilfrinib inhibits PTK6-activated autophagy, thereby promoting apoptosis both in vivo and in vitro

Tilfrinib, a potent and selective PTK6 inhibitor, exhibits significant antitumor activity, although its effects on CRC have not been previously reported (Figure 8A). To address this, we employed western blot analysis to assess the impact of various concentrations of tilfrinib on the expression of autophagy-related proteins in CRC cells, aiming to determine the optimal treatment concentration. Our results indicate that treatment with 20 μM tilfrinib for 24 h is the optimal condition (Figure 8B). TEM revealed a marked reduction in the number of autophagosomes under these conditions (Figure 8C). MCherry-EGFP-LC3 dual fluorescence assays demonstrated that tilfrinib significantly inhibits PTK6-mediated promotion of autophagy (Figure 8D and Figure S6A). Additionally, the results of flow cytometry demonstrated that tilfrinib significantly attenuates the anti-apoptotic effects mediated by PTK6 (Figure 8E and Figure S6B-D). Additionally, western blot analysis further confirmed that tilfrinib inhibits PTK6-mediated autophagy, thereby promoting CRC apoptosis (Figure 8F and Figure S6E).

Figure 8.

Tilfrinib inhibits PTK6-activated autophagy, thereby promoting apoptosis both in vivo and in vitro. (A) Molecular structure of tilfrinib. (B) Western blot analysis of BECN1, SQSTM1, and LC3B in HCT116 cultured for 24 h with different concentrations of tilfrinib. (C) Observation of autophagosome changes in HCT116 and DLD1 after tilfrinib treatment using electron microscopy. (D) The mCherry-EGFP-LC3 reporter assay was conducted to investigate the impact of tilfrinib on PTK6-mediated autophagic flux in HCT116. (E) Flow cytometry was conducted to investigate the impact of tilfrinib on PTK6-mediated apoptosis in HCT116. (F) Western blot analysis of BECN1, SQSTM1, LC3B, cleaved CASP3, BAX, and BCL2 in HCT116 to investigate the impact of tilfrinib. (G) Treatment of patient-derived CRC organoids with tilfrinib (20 µm), with detailed experimental procedures provided in the materials and methods. (H-J) Establishment of a xenograft tumor model in mice using HCT116 (n = 5), with tumor weight and volume monitored over a 24-day period. The treatment group received intraperitoneal injections of tilfrinib (10 mg/kg) every 3 days. (K) Immunohistochemistry for LC3B, SQSTM1, MKI67 and cleaved CASP3. The rightmost column displays the probe specific to NBR1 exon 10. ***p < 0.001 compared with the control.

To further validate its therapeutic potential, we established patient-derived organoid (PDO) models of CRC. The experiments showed that PTK6 significantly promotes PDO formation, whereas tilfrinib substantially inhibits PTK6-induced tumor proliferation (Figure 8G and Figure S6F). Then, we established a subcutaneous xenograft model in mice, and the results showed that tilfrinib treatment significantly inhibited the increase in both tumor volume and weight. Moreover, tilfrinib exhibited a pronounced effect in suppressing PTK6-induced tumor proliferation (Figure 8H–J). IHC analysis of the tumors revealed that tilfrinib significantly inhibited PTK6-mediated promotion of autophagy within the cells (Figure 8K). Additionally, we applied a probe targeting NBR1 exon 10 in subcutaneous tumor tissues. The results indicated a significant increase in NBR1 exon 10 skipping events following tilfrinib treatment (Figure 8K). The TUNEL assay results demonstrated that the anti-apoptotic effects induced by high PTK6 expression in CRC cells were effectively reversed by tilfrinib treatment (Figure S6G,H). At the same time, we examined the LLPS of HNRNPH1 in HCT116 and DLD1 following tilfrinib treatment and found that the LLPS of HNRNPH1 was significantly inhibited (Figure S6I). Overall, our experiments demonstrate that tilfrinib effectively inhibits PTK6-activated autophagy, thereby promoting apoptosis both in vivo and in vitro (Figure 9).

Figure 9.

Schematic illustration depicting the molecular mechanism underlying autophagy mediated by the novel PTK6-hnRNPH1-NBR1 regulatory axis. In brief, phosphorylation of HNRNPH1 at Y210 by PTK6 promotes its phase separation, which enhances the retention of exon 10 in NBR1, thereby facilitating autophagy and inhibiting apoptosis in CRC. Additionally, we discovered that tilfrinib, a potent and selective PTK6 inhibitor, effectively induces apoptosis by targeting autophagy in CRC. The image is drawn by Figdraw.

Discussion

In this study, we elucidated the molecular mechanism underlying PTK6-mediated autophagy activation that suppresses CRC apoptosis. PTK6 is generally upregulated in CRC and plays a significant role in the regulation of autophagy within this context. We further demonstrated that PTK6 facilitates the phosphorylation of HNRNPH1 at Y210, which mediates the LLPS of HNRNPH1, subsequently promoting the inclusion of NBR1 exon 10. Functional assays revealed that the inclusion of NBR1 exon 10 is essential for regulating autophagy in CRC. Notably, the suppression of CRC apoptosis via PTK6-mediated autophagy underscores the potential of PTK6 as a therapeutic target in CRC. Consequently, we introduced tilfrinib, a selective PTK6 inhibitor, as a promising therapeutic agent for suppressing CRC growth. Overall, our findings reveal the novel PTK6-HNRNPH1-NBR1 regulatory autophagy axis and provide a potential therapy target for CRC.

Autophagy is an important catabolic process involved in tumorigenesis and cancer progression [3]. Over recent decades, research has highlighted the dual role of autophagy in cancer, with ongoing studies continuously enhancing our understanding [20]. It is now widely accepted that autophagy serves to inhibit cancer initiation [21], however, accumulating evidence indicates that autophagic processes are also necessary for cancer progression [22,23]. Therefore, the role of autophagy appears to be contingent upon the stage of cancer, its biological characteristics, and the surrounding microenvironment. Understanding the detailed function of autophagy is essential for CRC treatment. Several lines of evidence suggest that autophagy plays a pivotal role in CRC malignant progression. For instance, Xie et al. demonstrated that the SLC16A1/MCT1-STK11/LKB1-AMPK signaling pathway activates protective autophagy in osimertinib/OSI-treated CRC cells, and inhibition of autophagy significantly promotes osimertinib-induced CRC apoptosis [24]. Some studies have shown that autophagy exerts a tumor-suppressive function [25,26]. Consequently, the precise role of autophagy in CRC remains inadequately understood. Autophagy is a dynamic and complex process that plays context-dependent roles, either inhibiting or promoting CRC. Here, we demonstrated that protective autophagy degraded cleaved CASP8 to inhibit apoptosis in CRC. Regrettably, previous research has not yielded effective treatment strategies targeting autophagy in CRC. In our current study, we indicated that PTK6 is correlated with the autophagy process in CRC, and we propose a promising application of tilfrinib for treating autophagy-mediated anti-apoptosis.

Previous studies have reported on the expression level of PTK6 in CRC. Mathur et al. demonstrated that PTK6 has an inhibitory role in CRC [27]. In contrast, Zhao et al. provided evidence of a stimulatory role for PTK6 in CRC progression [8]. Consistent with the findings of Zhao et al., we identified that PTK6 is upregulated in CRC, and that its expression correlates with poor patient prognosis. Additionally, we detected an enrichment of PTK6 in the autophagy and apoptosis pathways in PTK6-overexpressing CRC tissues sourced from open-access databases. Moreover, a series of functional experiments demonstrated that PTK6 promotes autophagy while inhibiting apoptosis in CRC. The autophagy pathway inhibitor CQ dramatically reverses PTK6-mediated anti-apoptosis in CRC. Thus, our finding suggests that PTK6 inhibits CRC apoptosis through the regulation of the autophagy pathway.

As a member of the SRC family kinases, PTK6 possesses a SH3 domain, a SH2 domain, and a TyrKc domain [6]. Research indicates that PTK6 functions through interactions with other tumorigenic factors, facilitated by its SH2/SH3 domains and/or tyrosine phosphorylation via the tyrosine kinase domain [6]. Similar to other SH2 domains within the SRC family kinases, the SH2 domain of PTK6 contains a consensus α/β-fold and a Tyr(P) peptide binding surface, playing an important role in substrate recognition [28]. In our study, we generated a PTK6 mutant wherein the SH2 domain was deleted, and this type of mutation resulted in the loss of interaction between PTK6 and HNRNPH1. Therefore, our results prove that PTK6 interacts with HNRNPH1 through the SH2 domain and phosphorylates HNRNPH1 to enhance CRC autophagy.

Generally, substrates are bound and phosphorylated by PTK6, and this phosphorylation leads to re-localization or activation. In breast cancer, PTK6 phosphorylates the splicing factor PSF, affecting its ability to bind RNA and regulating its cellular localization [29]. However, this study did not establish that PTK6-mediated phosphorylation of PSF regulates alternative splicing. Thus, the mechanism by which PTK6 regulates alternative splicing remains unclear. In this manuscript, we observed that PTK6 phosphorylates Y210 of the splicing factor HNRNPH1. The phosphorylation of HNRNPH1 at Y210 mediates LLPS of HNRNPH1, promoting the inclusion of NBR1 exon 10. Our study revealed that WT HNRNPH1 forms phase-separated condensates with liquid-like behavior in CRC, whereas HNRNPH1^Y210A lacks LLPS. This suggests that PTK6 promotes NBR1 exon 10 inclusion via phosphorylation of HNRNPH1 Y210.

LLPS elucidates how cells efficiently organize various molecules to perform specific reactions or activities in a temporal and spatial manner [30–32]. It has been proven that RNA-binding proteins/RBPs possessing IDRs are likely to undergo LLPS [33]. IDRs exhibit low amino acid sequence complexity and consist of a limited set of amino acid types, including glycine, serine, glutamine, and aromatic residues, such as phenylalanine and tyrosine [34–36]. Recent evidence suggests that the aromatic residues in IDRs are particularly important for promoting LLPS [37]. In addition, PTMs induce a wide range of alterations in the physicochemical characteristics of the regulated amino acids in phase-separated proteins, such as valency, electric charge, or volume, which can enhance or decrease the process of LLPS [35]. For instance, the Arg residues of RGG/RG motifs in FUS were largely modified by the deposition of asymmetric dimethyl groups by PRMT1 (protein arginine methyltransferase 1) or PRMT8 [38,39], which in turn decreased the LLPS rate of FUS and enhanced condensate dynamics [40]. As an RNA-binding protein, HNRNPH1 is comprised of one potential IDRs. Given the role of IDRs in protein LLPS, we employed the IDRs deletion mutation to interfere with its LLPS ability and found that the mutation can effectively impact HNRNPH1 LLPS. Furthermore, we noticed that Y210, located in the IDRs, undergoes phosphorylation, which alters the charge of the Y210 side chain by introducing two negative charges (PO₄^2-) [41]. This drastically alters the steric and chemical properties of HNRNPH1 and provides intermolecular electrostatic interactions, mediating LLPS. Aromatic mutation of HNRNPH1 (HNRNPH1^Y210A) prevents phosphorylation, disrupting intermolecular electrostatic interactions, which effectively inhibits the intra- and intermolecular interactions that maintain pathological aggregation in CRC. Consequently, our findings indicate that phosphorylation of Y210 plays a crucial role in HNRNPH1 LLPS, suggesting that future research could focus on targeting PTMs to impede the carcinogenic LLPS process.

As a splicing factor, HNRNPH1 can directly bind to the pre-mRNAs of target genes for alternative splicing [42]. HNRNPH1 interacts with G‐rich motifs in the RBM3 poison exon to mediate poison exon skipping and increase RBM3 mRNA levels [42]. Our data suggest that NBR1 is a potential target of HNRNPH1, with HNRNPH1 mediating the inclusion of NBR1 exon 10 in CRC. NBR1 has been identified as a selective autophagy receptor; it interacts with SQSTM1 and directly binds to ATG8 proteins and ubiquitin [43,44]. In addition, NBR1 stabilizes SQSTM1 and promotes the formation of SQSTM1 body, further participating in the autophagic degradation of SQSTM1 bodies [45,46]. Johansen et al. indicated that NBR1 forms oligomers via its CC1 domain, which is required for the formation of SQSTM1 bodies [45]. In the present study, we generated NBR1 truncation mutants and, consistent with the findings of Johansen et al. demonstrated that NBR1 interacts with SQSTM1 in a CC1 domain-dependent manner, underscoring the pivotal role of the CC1 domain in mediating this interaction. Furthermore, the abolished droplet formation ability of HNRNPH1 led to NBR1 exon 10 skipping, which in turn generated a truncated subtype lacking amino acids 288–347, resulting in CC1 domain inactivation. Importantly, our results revealed that HNRNPH1 promotes the inclusion of exon 10 in NBR1, thereby facilitating the formation of SQSTM1 bodies and inducing selective autophagy in CRC.

Conclusion

In this study, we demonstrated that PTK6 upregulation enhances autophagy while suppressing apoptosis in CRC, contributing to poorer prognosis in CRC patients. Mechanistically, PTK6 interacts with and phosphorylates HNRNPH1 at Y210, facilitating its LLPS and promoting the HNRNPH1-mediated exon 10 inclusion of NBR1, which drives selective autophagy. Furthermore, we demonstrated that tilfrinib, a PTK6 inhibitor, exhibits promising preclinical efficacy in PDO and CDX models. Collectively, our findings underscore the pivotal role of PTK6-mediated autophagy in CRC progression and provide a potential therapeutic strategy for targeting PTK6 in CRC treatment.

Materials and methods

Patients and clinical specimens

In this study, 73 pairs of CRC and normal adjacent tissue samples were collected from the Department of Gastrointestinal Surgery at Xuzhou Medical University Affiliated Hospital between June 2015 and December 2015. These samples were utilized for constructing tissue microarrays (TMAs). Additionally, 53 pairs of fresh CRC and adjacent non-tumorous tissue samples were collected from surgical resections performed between December 2021 and December 2023 to validate the protein and mRNA expression levels of PTK6 in CRC and adjacent tissues.

Ethical approval for the study was obtained from the Ethics Committee of Xuzhou Medical University Affiliated Hospital (approval number: XYFY2024-KL-245-01), and informed consent was received from all participants. All patients included in the study had primary CRC and underwent surgery for the first time without any prior antitumor treatment.

Cell lines and culture

Human CRC cell lines (DLD1, SW480, HCT116, RKO, SW620, LoVo) were obtained from the Cell Bank of the Chinese Academy of Sciences (SCSP-5241, SCSP-5033, SCSP-5076, SCSP-5236, TCHu101, SCSP-514). Normal colonic epithelial cells (NCM460) were purchased from Guangzhou Xin Yuan Technology Company, China. To maintain cell viability and growth characteristics, HCT116 cells were cultured in McCoy’s 5A medium (Gibco 12,330,031) supplemented with 10% fetal bovine serum/FBS (Gibco, 10099141C). DLD1, SW480 and RKO cells were cultured in RPMI 1640 medium (Gibco 12,633,020), while NCM460, SW620 and LoVo cells were cultured in DMEM/high-glucose medium (Gibco 11,965,092). All cell lines were maintained in a humidified incubator at 37°C with 5% CO₂ and underwent regular mycoplasma testing to ensure absence of contamination prior to experiments.

Mutagenesis at the phosphorylation site and plasmid construction

The plasmids were developed by HeWu Biotechnology Co., Ltd. The recombinant plasmids, containing either the wild-type or mutant HNRNPH1 genes, were constructed by inserting these genes into the pcDNA3.1⁽⁺⁾ (Invitrogen, V79020). In this context, Y210A Y253A Y276A Y306A indicates that tyrosine (Y) at positions 210, 253, 276, and 306 was mutated to alanine (A) through site-directed mutagenesis.

qRT-pcr, RT-PCR and western blotting

Total RNA was extracted using Trizol Reagent (Vazyme, R401–01). It was then reverse transcribed into cDNA with the First Strand cDNA Synthesis Kit (Servicebio, G3330–50) and used as a template for amplification with SYBR Green qPCR Master Mix (Servicebio, G3322–05). DNA amplification was carried out using Fast sTaq PCR Master Mix (Servicebio, G3304–05) according to the manufacturer’s instructions. The primers used for RT-PCR and qRT-PCR is listed in Table S1.

To extract proteins, RIPA lysis buffer (KeyGEN BioTECH, KGP702) was employed, and protein quantification was performed using the Enhanced BCA Protein Assay Kit (KeyGEN BioTECH, KGP903). Equivalent amounts of protein were separated by 10% SDS-PAGE and transferred onto nitrocellulose membranes (Pall 66,485). Membranes were blocked with 5% nonfat dry milk or BSA (Yeasen, 36101ES50) for 2 h at room temperature, followed by incubation with specific antibodies. GAPDH served as an internal control. Protein bands were visualized using Tanon High-sig ECL western Blot Substrate (Tanon, 5200), and band intensities were analyzed with Tanon’s Image Analysis Software and ImageJ.

Mn²⁺-Phos-tag SDS-PAGE

HCT116 and DLD1 cells were cultured to the logarithmic growth phase and treated to overexpress or knock down PTK6. Cells were lysed in RIPA buffer containing protease and phosphatase inhibitors (Thermo Scientific 78,440), and protein concentration was determined using a BCA assay (Thermo Scientific 23,225). The protein samples were mixed with SDS-PAGE loading buffer and boiled for 5 min. Mn²⁺-Phos-tag SDS-PAGE was performed using a separating gel with 50 µM Mn²⁺-Phos-tag and MnCl₂, prepared with a 29:1 acrylamide to N,N’-methylenebisacrylamide ratio. Proteins were separated in running buffer (25 mm Tris, 192 mm glycine, 0.1% SDS, pH 8.4) and the gel was washed in 1 mm EDTA for 10 min, then in transfer buffer without EDTA for 10 min. Proteins were transferred to PVDF membranes, blocked with 5% milk in TBST (Servicebio, G0004), and incubated with primary antibodies specific to HNRNPH1 (Proteintech 14,774–1-AP). After washing, membranes were incubated with HRP-conjugated secondary antibodies (Proteintech, SA00001–2), developed using an ECL substrate, and visualized. Band intensity was quantified using image analysis software to assess HNRNPH1 phosphorylation levels.

In vitro phosphorylation assay

The phosphorylation assay utilized the Homogeneous Time-Resolved Fluorescence/HTRF KinEASE-TK kit (Cisbio, 62TK0PEB) as per the manufacturer’s guidelines. This kit is effective for assessing tyrosine kinase activity, including PTK6. Reagents were prepared following the provided protocol. The 1× kinase buffer was made by diluting the 5× buffer with distilled water and adding 1 mm DTT (Sigma, DTT-RO) and 5 mm MgCl2 (Sigma 00,457). The TK substrate-biotin and synthetic peptides were prepared in a 50 µM solution with distilled water. Streptavidin-XL665 was diluted to 500 nM. Recombinant human PTK6 protein was stored at − 80°C. The detection buffer was reconstituted with distilled water to make the working solutions of Streptavidin-XL665 and TK-antibody-cryptate. ATP (Sigma, A6559) was diluted to 5 mm with 1× kinase buffer.

Immunofluorescence (IF)

The cells were fixed with paraformaldehyde for 20 min at room temperature and permeabilized with 0.5% Triton X-100 (Sigma, T8787) for 30 min. Blocking was performed using goat serum (ZSGBBIO, ZLI-9022) for 30 min. The cells were incubated overnight at 4°C with LC3B antibody (1:500; Proteintech 14,600–1-AP), followed by a 1 h incubation with CoraLite488-conjugated secondary antibody (1:100; Proteintech, SA00013–2) after three washes with PBST (PBS [Servicebio, G4202] with 0.1% Triton X-100). Nuclei were stained with DAPI (Biosharp, BL105A). Fluorescence imaging was conducted using a Leica STELLARIS 5 laser scanning confocal microscope (Germany).

RNA-seq and data analysis

RNA was isolated from HCT116 cells treated with sh-NC and sh-HNRNPH1 using Trizol reagent. RNA concentration and purity were determined using a Nanodrop 2000 (Thermo Fisher Scientific, USA). Libraries were constructed, and transcriptome sequencing was conducted by BGI Genomics (Shenzhen, China). Differential alternative splicing (AS) events were analyzed using rMATS (version 3.2.5). rMATS identifies AS events as exon skipping, intron retention, alternative 5′ splice site, alternative 3′ splice site, and mutually exclusive exon, and performs differential AS analysis with biological replicates.

Transmission electron microscopy (TEM)

HCT116 and DLD1 cell lines were cultured and harvested at the logarithmic growth phase. The cells were then fixed with 3% glutaraldehyde and 2% paraformaldehyde in 0.1 M cacodylate buffer (pH 7.3) to preserve their morphological structure. After fixation, the cells were washed in 0.1 M sodium cacodylate buffer and treated with 0.1% Millipore-filtered cacodylate buffered tannic acid, followed by post-fixation with 1% osmium tetroxide to ensure preservation of cellular lipids. Next, the cells were dehydrated through a graded ethanol series (50%, 70%, 90%, and 100%) and then transitioned to acetone. The dehydrated cells were infiltrated with a liquid epoxy resin (Sigma 45,345), which were then polymerized at 60°C for 48 h to form solid blocks. Once the resin blocks were solidified, ultra-thin sections (about 70–90 nm) were cut using an ultramicrotome and collected onto copper grids. The sections were stained with uranyl acetate and lead citrate to enhance contrast. Finally, the prepared samples were examined using TEM to observe and capture images of the ultrastructural details.

ANXA5-FITC PI staining

CRC cells were seeded at an equal density in 6-well plates and allowed to adhere. After treatment with 20 μM chloroquine (CQ; MCE, HY-17589A) for 24 h, cells were digested with trypsin without EDTA, centrifuged at 300 g for 5 min at 4°C, and washed twice with pre-chilled PBS. The cells were then resuspended in 100 μL of 1× Binding Buffer, stained with 5 μL Annexin V-FITC and 10 μL PI Staining Solution (Yeasen, 40302ES50) for 15 min in the dark. After adding 400 μL of 1× Binding Buffer, apoptosis was analyzed by flow cytometry.

Immunohistochemistry

The harvested subcutaneous tumors were fixed in 4% formalin, embedded in paraffin, and then sectioned into 4-μm slices. Immunohistochemistry (IHC) of the TMAs and subcutaneous tumors was performed using a streptavidin-peroxidase (SP) Kit (ZSGBBIO, PV-9001) following the manufacturer’s instructions. The slides were incubated with antibodies specific for PTK6 (1:200; Proteintech 18,697–1-AP), LC3B (1:500; Proteintech 14,600–1-AP), SQSTM1/p62 (1:500; Proteintech 18,420–1-AP) and MKI67/Ki-67 (1:2000; Proteintech 28,074–1-AP). IHC images were captured with an Olympus microscope (Tokyo, Japan).

For IHC scoring method: staining scores were evaluated by combining staining intensity and the percentage of immunoreactive cells, quantified by the immunoreactivity score. Intensity was scored as 0–3 (0, negative; 1, weak; 2, moderate; 3, strong), and the percentage of positive cells was graded as 1 (0–25%), 2 (26–50%), 3 (51–75%), and 4 (76–100%). The score was calculated by multiplying the intensity score by the percentage score.

Protein purification

BL21 (DE3) E. coli cells were transformed with the HNRNPH1-expressing plasmid containing an EGFP tag. The transformed cells were cultured in Luria-Bertani/LB medium with ampicillin at 37°C until the OD600 reached 0.5–0.7. Induction was carried out by adding isopropyl β-D-1-thiogalactopyranoside (IPTG; Beyotime, ST098) to a final concentration of 1 mm, followed by incubation for 4–5 h. Cells were harvested, and protein purification was performed using the His-tag Protein Purification Kit (Beyotime, P2226). The purified protein was snap-frozen in liquid nitrogen and stored at −80°C.

Droplet assay

In vitro phase separation assays were carried out using a physiological LLPS buffer containing 20 mm Tris-HCl, pH 7.5, 15 mm NaCl, 130 mm KCl, 5 mm KH₂PO₄, 1.5 mm MgCl₂, and 1 mg/mL BSA, with PEG8000 (Yeasen, 60304ES76) added to a final concentration of 10% (w:v). BL21 (DE3) E. coli cells were transformed with an HNRNPH1-expressing plasmid containing an EGFP tag. The transformed cells were grown in Luria-Bertani medium supplemented with ampicillin at 37°C until the optical density at 600 nm (OD₆₀₀) reached 0.6–0.8. To induce protein expression, IPTG was added to a final concentration of 1 mm, and the culture was incubated for an additional 4–5 h to ensure sufficient protein translation. Subsequently, the E. coli cells were harvested, and the protein was purified using a His-tag Protein Purification Kit (Beyotime, P2226). The purified protein was then snap-frozen in liquid nitrogen and stored at −80°C. For droplet analysis, the phase-separated droplets were transferred to a confocal dish with a chamber, and imaging was performed using a laser scanning confocal microscope.

Fluorescence recovery after photobleaching (FRAP)

FRAP was performed using a laser scanning confocal microscope (Leica STELLARIS 5, Germany) with the 488-nm laser. An appropriate region of interest/ROI was selected and photobleached 8–10 times. The bleached area was then imaged to record the recovery. GraphPad Prism was used to plot and analyze FRAP data.

Patient-derived organoid (PDO) culture model

To establish the PDO model, tumor tissues were obtained from CRC patients undergoing surgery at the Department of Gastrointestinal Surgery, Affiliated Hospital of Xuzhou Medical University. Freshly excised tumor tissues were immediately minced into 1–3 mm³ fragments and washed multiple times with antibiotic-containing PBS to remove debris. The tissue fragments were enzymatically digested using 200 U/ml collagenase (Sigma, SCR103) and 100 U/ml hyaluronidase (Sigma, B20222) at 37°C for 30 min. The resulting cell suspension was filtered through a 100-μm cell filter and centrifuged to remove the supernatant. The isolated cells were resuspended in Matrigel (MCE, HY-K6001).

For PDO culture, 50 µL of Matrigel-cell suspension was added to each well of a 48-well plate and allowed to solidify. Subsequently, 450 µL of CRC organoid medium (bioGenous, K2103-CR) was added to each well. The culture medium was replenished every 2–3 days, and organoids were passaged every 10–15 days. Organoid morphology was monitored using an Olympus FSX100 microscope (Olympus, Tokyo, Japan).

When organoids reached a diameter of 200 μm, they were harvested and dissociated into single cells using organoid digestion buffer at 37°C for 10 min. The reaction was terminated with organoid culture medium, and cells were centrifuged at 300 g for 3 min, washed, and resuspended in medium at a concentration of 5 × 10⁵ cells/mL. To prepare for lentiviral transduction, 80 μL of pre-cooled Matrigel was added to each well of a 12-well plate and incubated at 37°C for 30 min to solidify. Single-cell suspensions (250 μL) were mixed with lentiviral particles (250 μL) and polybrene in a 1.5-mL tube, and the mixture was seeded onto the solidified Matrigel. After overnight incubation at 37°C, the medium containing virus was removed, and 60 μL of fresh cold Matrigel was overlaid onto the cells. The plate was incubated at 37°C for 20 min to solidify the Matrigel, followed by the addition of organoid culture medium for further growth.

For tilfrinib treatment, organoids with diameters ranging from 200 to 400 μm were harvested and replated into Matrigel. Organoids were embedded in Matrigel domes containing organoid culture medium supplemented with 20 µM tilfrinib. The organoids were cultured for 13 days with fresh medium containing 20 µM tilfrinib replaced every 3 days. Control wells were treated with an equivalent concentration of DMSO. During the treatment period, the organoids were monitored under a microscope, and their diameters were measured at regular intervals to evaluate changes in size.

In vivo crosslinking followed by immunoprecipitation

Cells were harvested and cross-linked under ultraviolet. The IP procedure was conducted using HNRNPH1 antibody (Proteintech 14,774–1-AP) or rabbit IgG polyclonal antibody (Proteintech 30,000–0-AP). Cell extracts were incubated with antibodies overnight at 4°C, followed by coincubation with 50 µl magnetic beads (MCE, HY-K0205) per sample for 1 h. Beads were washed three times with buffer containing 1% cocktail (Thermo Scientific 78,440) and 1 U/µl RNase inhibitor (Abclonal, RK21401), then suspended in 120 µl elution buffer with 1% cocktail and 1 U/µl RNase inhibitor at 30°C for 15 min. Supernatant was collected, and 5 µl 4.8 M NaCl with 1 U/µl RNase inhibitor was added, followed by shaking overnight at 65°C. Proteinase K (Sigma, P2308) was utilized to digest the proteins at 60°C for 1 h. Subsequently, RNA was extracted and reverse transcribed using random primers. PCR assays were then conducted with specifically designed primers to amplify the skipped cassette and the adjacent exons.

Animal study

Animal studies were approved by the Institutional Animal Care and Use Committee of Xuzhou Medical University (approval number:202406T005). Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China) provided female BALB/c nude mice between the ages of 4 and 6 weeks, which were housed in a specific pathogen-free facility and randomized into groups (n = 5 per group). A xenograft tumorigenesis model was constructed by injecting 5 × 10⁶ cells of transfected cells into the axillary region of mice. Growth of the tumor was measured every 3 days, and the tumor volume (V) was calculated as follows: V= (length × width²)/2.

To evaluate the effect of tilfrinib (a PTK6 inhibitor) on the growth of subcutaneous tumors in nude mice, drug treatment commenced 72 h after tumor cell inoculation to allow for cell adaptation and stable growth. Mice in the tilfrinib treatment group and the PTK6[OE] + tilfrinib group received intraperitoneal injections of tilfrinib (10 mg/kg) every 3 days. Control mice and those in the PTK6[OE] group received intraperitoneal injections of an equivalent volume of DMSO, serving as the vehicle control, every 3 days. Mice were euthanized, and tumors were excised for IHC analysis 2–3 weeks post-injection.

Statistical analysis

Data analysis was carried out using SPSS 19.0 (IBM, Armonk, NY, USA) and GraphPad Prism 8.2.1 (La Jolla, CA, USA). Results are expressed as means ± standard deviations (SD), and all statistical tests were two-sided. A p value less than 0.05 was considered to indicate statistical significance. Differences between groups were assessed using Student’s t-test or one-way ANOVA. Correlations were evaluated using Spearman’s correlation coefficient. The association between PTK6 expression and clinicopathological features of CRC patients was analyzed using chi-square or Fisher’s exact test. Overall survival (OS) was determined by the Kaplan – Meier method and compared using the log-rank test. The impact of PTK6 expression and other clinicopathological factors on survival and hazard ratio was assessed using univariate and multivariate Cox proportional hazard regression models.

Funding Statement

Our study was funded by grants from the National Natural Science Foundation of China [No. 82472731, 82203486], the Natural Science Foundation of Jiangsu Province [BK20231159], the Scientific Research of Jiangsu Health Committee [ZDA2020005], the Xuzhou Medical Leading Talents Training Project [XWRCHT20210034], Cultivation Plan for High level Scientific Research Projects of Xuzhou Medical University Affiliated Hospital [PYJH2024101, PYJH2024205], Medical Science and Technology Project of Xuzhou Municipal Health Commission [XWKYHT20220164], and Postgraduate Research & Practice Innovation Program of Jiangsu Province [KYCX24_3072, SJCX24_1570].

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15548627.2025.2481001