Unconventional Secretion of PKCδ Exerts Tumorigenic Function via Stimulation of ERK1/2 Signaling in Liver Cancer

Leaderless proteins, which lack an N-terminal signal peptide, are generally encompassed inside living cells. Interestingly, a number of studies have recently emerged that describe the extracellular release of living cell–derived leaderless proteins (1–3). This active release of leaderless proteins, often termed as unconventional protein secretion (UPS), is characterized to have alternative exocytic routes, different from a conventional endoplasmic reticulum (ER)-Golgi–mediated secretion pathway, and has been implicated in some physiologic processes, such as inflammatory response, vascularization, and organogenesis (4, 5). Furthermore, the presence of extracellular leaderless proteins is also recognized in various diseases, especially in cancers (6), but whether UPS occurs in cancer cells remains fully elucidated.

In tumor tissue, general leaderless proteins detected in the extracellular space are considered to be a result of leakage from dying cells or loss of cell integrity due to the harsh microenvironments. Nonetheless, in some case, the extracellular detection of these leaderless proteins does not appear to simply reflect a cancer cell–derived leakage. For example, the extracellular expression of nucleolin and importin α1, both of which are leaderless proteins, was observed in specific types of cancer cells (7–9). In addition, soluble importin α1 is cumulatively stored in cancer cell conditioned media (CM), especially in liver cancer cells, at their growing condition (8), suggesting that unconventional secretion of leaderless proteins is likely to occur in liver cancer cells in a living state.

Heparan sulfate proteoglycans (HSPG) are glycoproteins consisting of heparan sulfate glycosaminoglycan chains and a protein core, and have been known to be ubiquitously expressed at the cell surface of several type of normal and abnormal cells, including cancer cells (10, 11). HSPGs can regulate the activation of several cell surface receptors through their association with heparin-binding ligands such as FGF2, human immunodeficiency virus-Tat protein (HIV-Tat), and hepatocyte growth factor (5, 8, 12, 13). Notably, this association depends on the commonly carried basic amino acid consensus sequences of the heparin-binding proteins, which allow their invariable binding to sulfate groups of HSPGs (14). Interestingly, this basic site is often characterized by the presence of a nuclear localization signal (NLS) at the internal side of the cell membrane, which permits nuclear trafficking proteins to pass through the nuclear pore toward the nucleoplasm (15–17). It is therefore likely that secreted nuclear trafficking proteins would interact with the cell surface HSPGs. Indeed, our recent study has shown that HSPGs are essential for the cell surface localization of importin α1 (8). Moreover, the cell surface importin α1 is involved in proliferative signaling in liver cancer cells (8). These lines of evidence tempt us to hypothesize that some of extracellular heparin-binding leaderless proteins may have tumorigenic functions.

Protein kinase C delta (PKCδ) is a leaderless serine/threonine kinase, which has been characterized to localize at cytoplasm and nucleus. Previous studies have shown that expression level of PKCδ is associated with some types of cancers, including liver cancer (18–20). It has also been reported that intracellular activation of PKCδ is involved in invasive capacity and survival in liver cancer cells (21, 22). These studies suggested that PKCδ might play a critical role in tumor progression of liver cancer. However, the mechanistic relevance of PKCδ to tumor promotive phenotypes of liver cancer has not been poorly understood.

In this study, we carried out a secretome analysis based on a strategy of using extracellular leaderless proteins that interacted with HSPGs, and identified PKCδ as an important secretory leaderless protein. We found that secreted PKCδ was highly detected in CM of liver cancer cell lines and serum of patients with liver cancer. Furthermore, we demonstrated that the extracellular PKCδ induced activation of IGF-1 receptor (IGF1R), and its downstream signaling such as ERK1/2, which leads to an accelerated proliferation in liver cancer cells.

Human liver cancer lines HepG2, Hep3B, HuH7, and HLF; human embryonic kidney cell line HEK293; and human gastric cancer line AGS were obtained from the Japanese Collection of Research Bioresources in 2017. Primary human hepatocytes were purchased from Lonza in 2020. TRE-inducible HepG2 cells were purchased from Takara in 2018. HepG2, Hep3B, HuH7, HLE, TIG-1, and HEK293 were maintained in DMEM (Sigma) supplemented with 10% FBS (Sigma). Hepatocytes were maintained in HBM basal Medium (Lonza) supplemented with SignalQuots Kit (Lonza). AGS cells were maintained in RPMI1640 (Sigma) supplemented with 10% FBS. TRE-inducible HepG2 cells were maintained in α-MEM (Nacalai) supplemented with 0.1 mmol/L nonessential amino acid, 500 μg/mL G418, and 10% Tet system approved FBS (Takara). Cell lines were routinely monitored for Mycoplasma (4A Biotech Co.). Authentication of the cell lines was confirmed by short tandem repeating profiling every 6 months. The cell used for experiments were passaged within 10 times after thawing.

Serum samples were obtained from 19 patients with liver cancer, 12 patients with chronic hepatitis, 4 patients with liver cirrhosis, and 8 healthy donors. The protocol used in this study conformed to the ethical guidelines of the 1975 Declaration of Helsinki and were approved by the Jikei University School of Medicine Ethics Review Committee [Ethics Approval License: 29-135 (8751)], and written informed consent was obtained from all patients. Venous blood was collected and centrifuged for serum collection. The serum samples collected were stored at −80°C until further analysis.

The following 2 PKCδ and 2 GPC3 gRNAs were designed by CRISPRdirect (https://crispr.dbcls.jp).

• PKCδ gRNA 1: CACCGATGCGCAGGAACGGCGCCA

• PKCδ gRNA 2: CACCGCAAACAGTCTATGCGCAGTG

• GPC3 gRNA 1: CACCGCACCAAGCACGCGGTGCGCA

• GPC3 gRNA 2: CACCGACTGCAGCCCGGACTCAAGT

The gRNAs were cloned into the lentiviral lentiCRISPR v2 vector. All constructs were confirmed by DNA sequencing.

Cells were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton-X-100, and blocked with 1% BSA. Cells were then incubated with indicated antibodies. The fluorescence signals were detected by the Duolink In Situ proximity ligation assay (PLA) probe according to the manufacturer's instructions and visualized using a Zeiss LSM 880 laser microscope.

PKCδ protein levels in culture media and human serum were determined using a Sandwich ELISA Kit (MyBioSource).

All animal studies were approved by the Animal Care and Use Committee of Jikei University School of Medicine (Tokyo, Japan; the ethical approval number: 28-16) and conducted in accordance with the guidelines for animal experimentation established at Jikei University School of Medicine (Tokyo, Japan). Male nude mice (6–7 weeks old) were obtained from CLEA. The animals were maintained in a pathogen-free animal facility of Jikei University (Tokyo, Japan). Mice were randomized into indicated groups. Cells in 100 mL of Matrigel were implanted subcutaneously in the back flank of the mice. Mice were injected intratumorally with isotype control IgG or anti-PKCδ mAb (0.5 mg/kg per injection). Tumor size was determined by caliper measurement of the largest (x) and smallest (y) perpendicular diameters and was calculated according to the formula V = π/6 × xy².

Data are represented as mean ± SD from the indicated number of replicates. Two-sided differences between two samples were analyzed using the Mann–Whitney U test or Student t test. Repeated measure ANOVA analysis was also used if there were more than two data groups to compare. P < 0.05 were considered significant. Statistical analyses were performed using SPSS 25 (SPSS Inc.) or Prism 8 software (Graphpad Inc.).

Additional or detailed methods are described in the Supplementary Materials and Methods.

To explore putative extracellular leaderless proteins involved in liver cancer, we precipitated CM of two liver cancer cell lines (HepG2 and Hep3B) with heparin, a HSPG (Fig. 1A). We then performed unbiased secretome analysis of these obtained precipitates, and identified 1,245 common proteins, of which, 859 (69%) were leaderless proteins that were characterized as intracellular proteins in the Database for Annotation, Visualization and Integrated Discovery (DAVID) database (https://david.ncifcrf.gov; Fig. 1A; Supplementary Fig. S1A–S1C; Supplementary Table S1), indicating that a relatively large number of heparin-binding leaderless proteins are present in the extracellular fluid of liver cancer cells. Among these leaderless proteins, we detected PKCδ, an NLS-bearing intracellular protein (23), at a high identical score and specificity as the only PKC isoform. The extracellular presence of PKCδ was validated by immunoblotting analysis and PKCδ kinase activity measurements (Fig. 1B,–D; Supplementary Fig. S1D). Immunoblotting analysis also revealed that full-length form of PKCδ was conspicuously present in the CM of HepG2 cells (Fig. 1B,–D). To exclude the possibility that extracellular PKCδ is derived from extracellular vesicles, such as exosomes, we removed the extracellular vesicles by ultracentrifugation procedures. Full-length PKCδ was detected in the extracellular vesicle-free CM in HepG2 cells (Supplementary Fig. S1E and S1F). Furthermore, PKCδ isoform was specificity confirmed in immunoblotting analysis using antibodies for representative PKC isoforms (PKCα and PKCβ; Fig. 1B). These results indicate that PKCδ molecule is extracellularly released from liver cancer cells.

To show the cell-type specificity of the extracellular PKCδ, we examined several types of cell lines that endogenously express PKCδ: four liver cancer lines (HepG2, Hep3B, HuH7, HLE), a gastric cancer line (AGS) and human embryonic kidney line (HEK293), and primary human normal hepatocytes. Although intracellular PKCδ levels at both mRNA and protein expression were comparable among all tested cell lines, a large amount of PKCδ was detected in the CM of all of liver cancer cell lines, whereas the AGS and HEK293 cell lines showed no or low PKCδ levels in their CM (Fig. 1E; Supplementary Fig. S2A-S2D). In addition, we found that no PKCδ levels was detectable in the CM of hepatocytes (Fig. 1E). These results indicate that the extracellular release of PKCδ preferentially occurs in liver cancer cells.

We next examined the clinical impact of extracellular PKCδ in liver cancer by using human serum samples. Elevated levels of serum PKCδ were detected in most patients with liver cancer, compared with those in patients with chronic hepatitis or liver cirrhosis, or healthy individuals (Fig. 1F and G; Supplementary Fig. S3), suggesting that extracellular PKCδ may be an indicator for patients with liver cancer. From these findings, we concluded that the presence of extracellular PKCδ was a character of liver cancer.

Given that many heparin-binding proteins are anchored at the target cell surface via HSPGs (10, 24), we investigated whether the extracellular release of PKCδ was accompanied with its cell surface anchoring. Ultrastructural and flow cytometric studies showed that cell surface expression of PKCδ was observed in tested all of liver cancer cell lines, whereas less or no cell surface PKCδ was detected in AGS cells (Fig. 2A and B). We confirmed that both the C-terminal and N-terminal regions of PKCδ were detectable at the side of the extracellular space (Supplementary Fig. S4A–S4E), indicating that full-length PKCδ is present at the cell surface of liver cancer cells. We next examined the relevance of HSPGs to the cell surface localization of PKCδ. The abundance of cell surface PKCδ was apparently reduced when several liver cancer cells were treated with heparinase, an enzyme that digests heparan sulfate groups of HSPGs (Fig. 2C), suggesting that HSPGs is involved in the cell surface localization of PKCδ in liver cancer cells.

It has been reported that glypican 3 (GPC3), one of cell surface HSPGs, is specifically expressed in hepatocellular carcinoma (25). Hence, we challenged whether GPC3 is involved in the cell surface localization of PKCδ in liver cancer cells. Both immunoprecipitation and ultrastructural analysis revealed that PKCδ was associated with GPC3 (Fig. 2D and E). To further visualize the interaction of extracellular PKCδ with GPC3, we performed the Duolink In Situ PLA. A number of signals of the PKCδ–GPC3 interaction were observed in cultured cells and spheroids of liver cancer cell lines (Fig. 2F and G). Furthermore, to investigate the impact of GPC3 expression on the cell surface localization of PKCδ, we established GPC3-knockout HepG2 cells using CRISPR/Cas9 approach. In GPC3-knockout HepG2 cells, the cell surface expression level of PKCδ was attenuated compared with that in wild-type HepG2 cells (Fig. 2H and I). These results indicate that PKCδ localizes at the cell surface of liver cancer cells through association with GPC3.

It has been accepted that UPS can occur in living cells (26). Kinetic study showed that an incremental PKCδ accumulation was observed in the liver cancer cell CM under some culture conditions (0.1 or 10% FBS; Fig. 3A; Supplementary Fig. S5), while most cells were alive (trypan blue exclusion; > 97%, n = 3), suggesting that the PKCδ release occurred from living liver cancer cells. To exclude the possibility that the PKCδ release is due to leakage, we monitored the kinetics of lactate dehydrogenase (LDH), a quantitative cytotoxic leakage marker. LDH abundance in the CM was invariable throughout the incubation periods under these culture conditions (Fig. 3B; Supplementary Fig. S5A and S5B).

To clear whether the PKCδ release is classified as one of UPS, we compared some properties of the UPS with those of the PKCδ release. We first employed some ER-Golgi secretion inhibitors (brefeldin A, monensin, and nigericin) (27). These inhibitors drastically reduced the extracellular release of α-fetoprotein (AFP), a secretory protein with an N-terminal signal peptide, whereas no effect on the PKCδ release was observed by treatment with these inhibitors (Fig. 3C; Supplementary Fig. S6), indicating that the PKCδ release is ER-Golgi–independent. Next, we treated liver cancer cells with ATP, which is known as an inducer of UPS (28). An increased PKCδ level was observed in the CM of liver cancer cells treated with ATP (Fig. 3D). We confirmed no change in LDH release in the presence of ATP (Fig. 3D). Taken together, these findings indicate that the PKCδ release in liver cancer cells is one of UPS.

To further confirm that the PKCδ release is originated from living cells, we established doxycycline-inducible PKCδ-eGFP–expressing HepG2 cells, which allowed the living cell to express the PKCδ-eGFP fusion protein only upon doxycycline treatment (Fig. 3E). In such cells, an unambiguous PKCδ-eGFP accumulation was observed in the CM upon doxycycline treatment (Fig. 3E). Doxycycline treatment showed no effect on neither cell proliferation nor LDH release levels in the CM (Fig. 3F and G). Furthermore, in a xenograft tumor model of the cells, both serum level of PKCδ and the number of PKCδ-GPC3 colocalization in tumor tissues were apparently increased after doxycycline administration (Fig. 3H,–J). These results strongly support that the extracellular PKCδ release occurs in living tumor cells.

It is generally accepted that inactive PKCδ preferentially resides in the cytosol, whereas activated PKCδ is translocated to the cytoplasmic side of the plasma membrane and other organelle membrane (29). Therefore, we examined whether the inactive/activated state is involved in the PKCδ release. To this end, we utilized phorbol-12-myristate-13-acetate (PMA), an activator of PKC family. PMA treatment markedly declined amount of PKCδ both in the CM and at the cell surface of liver cancer cells (Figs. 4A and 3E; Supplementary Fig. S7A). PMA treatment also cancelled the enhancement effect on PKCδ release following ATP treatment (Fig. 3D), suggesting that extracellular release of PKCδ occurs at its inactive state. Furthermore, PMA treatment had no effect on the release of both importin α1 and LDH (Fig. 4A and B), indicating that the inhibitory effect of PMA on the extracellular release is PKCδ specific.

To examine the involvement of inactive state of PKCδ in its extracellular release, we evaluated the Thr507 and Tyr313 phosphorylation levels of PKCδ, both of which are highly phosphorylated by Src kinases in an active state (30–34). Low phosphorylation levels were detected both at these residues of PKCδ both in the CM and lysates of normal cultured HepG2 cells, as reported previously (Fig. 1C and D; refs. 30–34). In contrast, PMA treatment markedly increased the phosphorylation level at both Thr507 and Tyr313 on intracellular PKCδ (Fig. 4C and D), as reported previously (35). Furthermore, declined phosphorylation of the both sites by Src inhibitors (PP1 and HβCD; ref. 35) appeared to resume the PKCδ release in a dose-dependent manner (Fig. 4C and D). These findings strongly indicated that the inactive PKCδ was translocated into the extracellular space from the inside of the cell.

We next examined the relation of the PKCδ release to its subcellular localization. In a normal culture of HepG2 cells, PKCδ was preferentially observed at the cytosol (36), whereas PMA treatment led to a reduction of cytosolic PKCδ localization with its concurrent translocation to the plasma membrane (Fig. 4E; Supplementary Fig. S7B and S7C). It is shown that PKCδ translocation to the cell membrane upon activation is responsible for its lipid-binding C1 domain (18). Therefore, we investigated whether the C1 domain is involved in the PKCδ release. Transfection experiments showed a persistent release of the C1-deleted PKCδ mutant to the CM even after the PMA treatment (Fig. 4F and G). Furthermore, it is also shown that PKCδ can be translocated to the nucleus (23, 37, 38). To address a possibility that the PKCδ release could be initiated from the nucleus, we introduced HepG2 cells with expression vectors of several PKCδ mutants fused with NLS of SV40 large T antigen or a point mutant (PPxxP to AAxxA; NLS active mutant; ref. 38). Neither mutant could be detected in the CM of the cells (Fig. 4H; Supplementary Fig. S7D). These results suggest that the PKCδ release is initiated from the cytosol.

To examine the functional relevance between extracellular PKCδ and liver cancer cells, we treated liver cancer cells with purified recombinant PKCδ protein (rPKCδ). Because wild-type liver cancer cells constantly secreted PKCδ to the extracellular space, we first established PKCδ-knockout HepG2 cells using the CRISPR/Cas9 system (Fig. 5A). Proliferation assay demonstrated that a significant increase after rPKCδ treatment was noted in not only wild-type, but also PKCδ-knockout HepG2 cells (Fig. 5B), suggesting that extracellular PKCδ functions as an inducer of cell growth. Interestingly, when the similar experiment was done in GPC3-knockdown HepG2 cells, the rPKCδ-stimulated proliferative enhancement was disappeared (Fig. 5B), indicating that cell growth effect of extracellular PKCδ was dependent on GPC3.

Next, we studied how extracellular PKCδ affects proliferative phenotype. Given that GPC3 regulates several receptor activities as a coreceptor by facilitating concentration of their ligands (39, 40), we asked whether the cell surface–accumulated PKCδ triggers receptor phosphorylation and activation because the tyrosine kinase cascades are associated with tumorigenesis. We performed an unbiased antibody array for human phospho-RTKs (receptor tyrosine kinases; Fig. 5C). rPKCδ treatment markedly increased phosphorylation of IGF1R in HEK293 and HepG2 cells (Fig. 5C and D). Previous studies have shown that GPC3 is associated with IGF1R signaling in liver cancer (39, 40). Indeed, in GPC3-knockout HepG2 cells, the increased phosphorylation of IGF1R after rPKCδ treatment was apparently cancelled (Fig. 5D), indicating that the activation of IGF1R by extracellular PKCδ was mediated by GPC3 expression.

To further investigate which intracellular signaling molecules were activated after rPKCδ treatment, we carried out a phosphorylation array, and found elevated phosphorylation of both ERK1/2 and STAT3 (Tyr705 and Ser727) in HepG2 cells treated with rPKCδ, which are all known as IGF1R signaling molecules (Supplementary Fig. S8A; refs. 40–42). The elevated phosphorylations of ERK1/2 and STAT3 (Tyr705), but not STAT3 (Ser727), were confirmed in HuH7 and parental and PKCδ-knockout HepG2 cells (Fig. 5E; Supplementary Fig. S8B–S8D). We also confirmed that both phosphorylation levels of ERK1/2 and STAT3 (Tyr705) were involved in GPC3 expression (Supplementary Fig. S8E and S8F). Furthermore, pretreatment with IGF1R inhibitors (linsitinib or GSK1904529A) blocked rPKCδ-mediated IGF1R, ERK1/2, and STAT3 (Tyr705) activation (Fig. 5F), indicating that extracellular PKCδ regulates IGF1R-related signaling pathways. To confirm the impact of IGF1R activation on liver cancer cell growth, we treated liver cancer cells with linsitinib. Cell proliferation was apparently decreased in linsitinib-treated liver cancer cell lines and three-dimensional (3D) spheres of HepG2 cells (Supplementary Fig. S8G and S8H). Taken together, these results suggest that the extracellular PKCδ stimulates cell growth mediated by IGF1R signaling.

To address whether extracellular PKCδ is required for the cell growth of liver cancer, we employed antibodies specific for PKCδ. We found that treatment with an anti-PKCδ mAb led to decreased proliferation of liver cancer cells (Fig. 6A and B; Supplementary Fig. S9A–S9D). We confirmed the specificity of this mAb for extracellular PKCδ by using AGS and PKCδ-knockout HepG2 cells, which were negative for extracellular PKCδ (Fig. 6A and B). We also found that the anti-PKCδ mAb treatment reduced the phosphorylation level of IGF1R and ERK1/2, but not STAT3 in PKCδ-positive liver cancer cell lines (Fig. 6C; Supplementary Fig. S9E), suggesting that activation of ERK1/2 by extracellular PKCδ plays a critical role in liver cancer cell proliferation. We further confirmed that c-myc expression was decreased in liver cancer cell lines treated with the anti-PKCδ mAb (Supplementary Fig. S9F), as previously reported in GPC3-deficient cells (39). These findings indicate that the anti-PKCδ mAb could inhibit the effect of extracellular PKCδ on cancer cell proliferation through impairment of IGF1R-ERK1/2 signaling.

We further investigated the involvement of extracellular PKCδ in tumorigenesis by using the anti-PKCδ mAb. An in vitro spheroid formation assay showed that inhibitory effects by the anti-PKCδ mAb were observed on both the spheroids size and the cell number in wild-type HepG2 cells, whereas no change was observed on that in PKCδ-knockout HepG2 cells (Fig. 6D and E). In addition, when we generated a xenograft tumor model of liver cancer cells, the tumor volume was apparently diminished in xenografted mice administrated with the anti-PKCδ mAb, compared with that in the mice injected with isotype control IgG (Fig. 6F and G). Immunostaining study showed that the anti-PKCδ mAb treatment markedly diminished the number of Ki67-positive cells in tumor tissues (Fig. 6H). Similarly, the anti-PKCδ mAb treatment attenuated phosphorylation levels of IGF1R and ERK1/2 in the tumor tissues (Fig. 6H). Taken together, these results suggest that extracellular PKCδ may contribute to tumor growth in liver cancer.

PKCδ has been well characterized as an intracellular serine/threonine kinase. In this study, we demonstrate that PKCδ is localized at the extracellular space in liver cancer cells (Figs. 1 and 2). We found that this extracellular localization of PKCδ is based on unconventional secretion from living cells (Fig. 3). Furthermore, extracellular PKCδ is found to behave as like one of growth factors in liver cancer cells. Importantly, we found that extracellular PKCδ stimulates phosphorylation of IGF1R signaling (Figs. 5 and 6), whose activation is important in liver cancer growth (40). This proliferative effect was suppressed by anti-PKCδ mAb treatment in cultured cells, 3D-cultured spheroids, and a xenograft mouse model (Fig. 6). These findings provide the rationale that PKCδ might become a target for potential antibody therapy for liver cancer, although PKCδ-targeted small compounds and peptides to penetrate inside the cell have often been developed for the treatment of various diseases.

Remarkably, high levels of PKCδ were observed in serum of patients with liver cancer compared with those in patients with chronic hepatitis and hepatic cirrhosis and healthy donors (Fig. 1F and G). This finding suggests that serum PKCδ may be a promising biomarker for diagnosis of liver cancer. Up to date, we could not determine which clinical factors correlated with the abundance of serum PKCδ levels; thus, further investigation is required to resolve this issue.

PKCδ expression has been reported to be involved in both tumor progressive and suppressive phenotypes in several cancers, including liver cancer (19, 43). Notably, upon DNA damage, induced nuclear localization of PKCδ from the cytosol is known to lead to apoptosis mediated through the activation of Jun-N-terminal kinase (JNK) in several models of cancer cells (23, 44, 45). In contrast to this tumor suppressive function of PKCδ, we found that extracellular PKCδ stimulates cancer cell proliferation (Fig. 5). We also showed that extracellular release of PKCδ was inhibited by the induction of nuclear entry of PKCδ (Fig. 4). These findings indicate that the distinct destination of PKCδ at the nucleus or the extracellular space results in reciprocal functions on cell survival. Taken together with our finding that extracellular movement of PKCδ appeared to initiate from the cytosol (Fig. 4), the switching by which the destination of PKCδ from the cytosol is determined is likely to be a critical factor for tumorigenesis.

We observed that rPKCδ treatment induced activation of not only IGF1R and ERK1/2, but also STAT3 (Fig. 5). Conversely, the anti-PKCδ mAb treatment was appeared to decrease in phosphorylations of IGF1R and ERK1/2 (Fig. 6). However, phosphorylation level of STAT3 was less change in liver cancer cells treated with the anti-PKCδ mAb (Fig. 6). This discrepancy on STAT3 may be likely to be explained by compensation of STAT3 signaling by other humoral factors such as IGF2 and IL6. Therefore, ERK1/2 activation may be key role in extracellular PKCδ-induced liver cancer cell/tumor growth.

In this study, we found that activation of IGF1R signaling by extracellular PKCδ depends on GPC3 expression (Fig. 5). Generally, GPC3 is a cell surface receptor, but could not be a direct regulator in the intracellular signaling cascades because of its defect in cytoplasmic expression (46). Instead, it is functionally known to fine tune the activation of cell surface receptors through association with the receptors and their ligands (11, 47, 48). It is reported that GPC3 is associated with IGF1R, and regulates IGF2-IGF1R signaling (40). Our current study identified IGF1R as an extracellular PKCδ-stimulated receptor (Fig. 5). These results suggest that association of GPC3 with PKCδ is likely to induce concentration of PKCδ at IGF1R-resided sites, and consequently enhance the IGF1R signaling, similar to IGF2-IGF1R axis. Up to date, the direct relation between extracellular PKCδ and IGF1R, especially at its extracellular domain, has remained unclear. This issue is needs to be resolved as a next important study.

We suggested that anti-PKCδ mAb had antitumor effect in for serum PKCδ- and GPC3/IGF1R-positive liver cancer cells (Fig. 6). Until now, many therapeutic drugs have been developed. In the near future, we will compare the antitumor effects of the standard therapeutic drugs, such as sorafenib, and the anti-PKCδ mAb in animal models.

Tumor development is achieved by intercellular communication through extracellular materials (49). The importance of autocrine or paracrine secretions by cancer cells in tumor development suggests that secreted leaderless proteins through UPS may be correspondingly important in forming the tumor microenvironment. Indeed, we and others have reported the involvement of leaderless proteins in extracellular functions in cancer cells. In our current and previous studies, the extracellular PKCδ and importin α1 (8) are shown to contribute to proliferation of liver cancer cells. As another example, cell surface nucleolin is reported to mediate angiogenesis and tumorigenesis to direct the formation of tumor microenvironment (9, 50). However, the machinery by which these leaderless proteins are translocated to the extracellular space has not yet been explained. As one possible explanation, these extracellular leaderless proteins may have other consensus signals that direct their secretion, instead of the N-terminal signal peptide. Taken together, we speculate that extracellular localization of these leaderless proteins in cancer cells might provide functional diversity for their survival in the harsh microenvironment of tumors.

M. Saruta reports grants from EA Pharma Co., Ltd., Zeria Pharmaceutical Co., Ltd., Kissei Pharmaceutical Co., Ltd., Mochida Pharmaceutical Co., Ltd; personal fees from AbbVie GK, Mitsubishi Tanabe Pharma, Janssen Pharma K.K, and Takeda Pharmaceutical Co., Ltd. outside the submitted work. No disclosures were reported by the other authors.

K. Yamada: Conceptualization, resources, data curation, funding acquisition, validation, investigation, methodology, writing-original draft, writing-review and editing. T. Oikawa: Conceptualization, resources, data curation, funding acquisition. R. Kizawa: Investigation, methodology. S. Motohashi: Investigation. S. Yoshida: Investigation. T. Kumamoto: Investigation. C. Saeki: Investigation. C. Nakagawa: Investigation. Y. Shimoyama: Investigation. K. Aoki: Investigation. T. Tachibana: Resources, investigation. M. Saruta: Supervision. M. Ono: Data curation, investigation, methodology. K. Yoshida: Conceptualization, resources, data curation, supervision, funding acquisition, validation, writing-original draft, project administration, writing-review and editing.

We thank N. Tago for secretarial and technical assistance. We also thank H. Saito, and Y. Takemura, N. Oikawa, M. Tatsushiro, and R. Shimokawa for technical assistance. This work was supported by grants from the Japan Society for the Promotion of Science (KAKENHI grant numbers 16K18434, 18K19484, and 20K07621 to K. Yamada; 17H03584, 18K15253, and 20H03519 to K. Yoshida; and 19K16781 to S. Yoshida), AMED under grant number B326TS to K. Yamada, The Jikei University Graduate Research Fund to K. Yamada, T. Oikawa, and S. Yoshida, and The Science Research Promotion Fund to K. Yoshida.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.