Oncogenic Roles of Laminin Subunit Gamma‐2 in Intrahepatic Cholangiocarcinoma via Promoting EGFR Translation

Abstract Intrahepatic cholangiocarcinoma (iCCA) is a highly lethal biliary epithelial cancer in the liver. Here, Laminin subunit gamma‐2 (LAMC2) with important oncogenic roles in iCCA is discovered. In a total of 231 cholangiocarcinoma patients (82% of iCCA patients) across four independent cohorts, LAMC2 is significantly more abundant in iCCA tumor tissue compared to normal bile duct and non‐tumor liver. Among 26.3% of iCCA patients, LAMC2 gene is amplified, contributing to its over‐expression. Functionally, silencing LAMC2 significantly blocks tumor formation in orthotopic iCCA mouse models. Mechanistically, it promotes EGFR protein translation via interacting with nascent unglycosylated EGFR in the endoplasmic reticulum (ER), resulting in activated EGFR signaling. LAMC2‐mediated EGFR translation also depends on its interaction with the ER chaperone BiP via their C‐terminus. Together LAMC2 and BiP generate a binding “pocket” of nascent EGFR and facilitate EGFR translation. Consistently, LAMC2‐high iCCA patients have poor prognosis in two iCCA cohorts. LAMC2‐high iCCA cells are highly sensitive to EGFR tyrosine kinase inhibitors (TKIs) treatment both in vitro and in vivo. Together, these data demonstrate LAMC2 as an oncogenic player in iCCA by promoting EGFR translation and an indicator to identify iCCA patients who may benefit from available EGFR‐targeted TKIs therapies.


Introduction
Intrahepatic cholangiocarcinoma (iCCA) is one of the most lethal malignancies, with an overall 5-year survival rate of 5-10%.[3] Liver resection and liver transplantation are potentially curative treatment options for very early stage iCCA.However, the majority of iC-CAs are often diagnosed at an unresectable stage.As for patients with advanced-stage or unresectable iCCA tumors, the available standard systemic chemotherapy (gemcitabine and cisplatin) provides only minimal benefits. [1]Most recently, two types of targeted therapeutic agents have been approved as second-line treatment for patients harboring key iCCA oncogenic genomic alterations, i.e., IDH1 mutation and FGFR fusion.They are Ivosidenib (mutant IDH1 inhibitor) as well as pemigatinib and infigratinib (FGFR fusion inhibitors), showing encouraging improvement in increasing patients' survival rates. [4,5]However, IDH1 mutation and FGFR fusion only occur in a limited iCCA population and the resistance typically develops within months. [6]In this case, more efforts are needed to continue investigating the key oncogenic events of iCCA and to explore their potential therapeutic applications.
EGFR has been recognized as an effective therapeutic target in various cancer types with approved EGFR tyrosine kinase inhibitors (TKIs) for clinical use.However, EGFR amplifications and mutations occurred in less than 5% of iCCA cases.Nonetheless, over-expression of EGFR exhibited in 10-44% of iCCA patients. [7,8]Additionally, a subtype of iCCA patients was previously reported with active EGFR signaling. [9]Clinical trials have been conducted to evaluate the efficacy of EGFR TKIs and neutralizing antibodies in iCCA patients, although the focus has primarily been on combining these treatments with chemotherapy drugs rather than using EGFR TKIs alone.Unfortunately, early phase clinical trials ended with negative results of EGFR TKIs in iCCA patients.[12] Till now, it remains unclear whether certain molecular subtypes of iCCA patients might benefit from clinical EGFR-targeted therapy.
We aimed to identify key oncogenic events specific to iCCA and to explore the potential therapeutic strategies for this condition.In this study, our findings revealed LAMC2 as a key oncogenic molecule in iCCA patients via LAMC2/BiP/EGFR axis and as a potential indicator in suggesting iCCA patients for available EGFR TKIs therapies.

LAMC2 Exhibited a Specific High Expression Level in iCCA Tumor Tissues
Liver, the anatomical location of iCCA, consists primarily of hepatocytes with a minority of cholangiocytes and is also the site of hepatocellular carcinoma (HCC, the most common primary liver cancer).To globally identify iCCA-specific oncogenic genes, we thus integrated transcriptome data from three clinical cohorts, including tissues from normal bile ducts as well as tumor and non-tumor liver tissues from cholangiocarcinoma patients (iCCA, 82%) and HCC patients (Figure 1A).Principle component analysis revealed that non-tumor liver tissues from iCCA and HCC patients were tightly clustered together while their tumor tissues were two distinct groups (iCCA and HCC groups) (Figure S1A, Supporting Information, Cohort 1), highlighting the necessity of including HCC and bile ducts as controls.Through class comparisons in Cohorts 1-3, Laminin Subunit Gamma 2 (LAMC2) and KRT19 were identified as key candidates upon the stringent criteria, i.e., 8 times higher expression level in CCA tumor versus non-tumor liver tissues (p < 0.001) and normal bile duct cells (p < 0.001), but no difference in HCC tumor versus non-tumor liver tissues (p > 0.05) (Figure 1A,B).KRT19 is a wellknown biliary differentiation marker and presents a high expression level in iCCA. [13]For LAMC2, a few articles reported that it was up-regulated in CCA, contributed to invasive features of CCA cells, and was related to poor prognosis of patients. [14]Therefore, LAMC2 was chosen for further exploration since its function remained largely unknown in iCCA carcinogenesis.
The specific high expression of LAMC2 was noticed in three iCCA cell lines in comparison with four HCC cell lines and 293T cells, at both the mRNA (Figure 1C) and protein levels (Figure 1D).Single-cell RNA-sequencing data of human liver cancers, including HCC and iCCA, [15] portrayed a noticeably higher expression of LAMC2 in iCCA tumor cells compared to other pathological types of liver cancer cells and tumor immune microenvironment cells (Figure S1B, Supporting Information).Furthermore, from hydrodynamic tail vein injection (HDTV) orthotopic liver cancer mouse models driven by various oncogenes, [16] RNA sequencing data showed that LAMC2 exhibited a significantly higher level in tumors from the iCCA-like tumor subtype than HCC subtype and normal livers (Figure 1E).Such an iCCA-specific expression panel of LAMC2 was similar to KRT19, if not even better.Comparable data were also obtained via the qRT-PCR method in mouse orthotopic liver cancers (Figure 1F).
Immunohistochemistry (IHC) staining of LAMC2 was then performed in FFPE tissues in Cohort 4, which included 33 iCCA patients and 20 HCC patients (Table S2, Supporting Information).As shown in Figure 1G, LAMC2 IHC staining was very strong in most iCCA tumor tissues, but no staining in normal bile duct cells or other non-tumor environment cells.Furthermore, only faint LAMC2 staining was noticed in hepatocytes, while weak and medium LAMC2 staining was observed in HCC tumor tissues.Quantitative data consistently showed that LAMC2 staining was significantly higher in iCCA tumor tissues versus bile duct cells, hepatocytes, and HCC tumors, respectively (p < 0.001 for each comparison, Figure 1H).Taken together, these data demonstrated a specific high expression of LAMC2 in iCCA tumor tissues within the liver.

LAMC2 Gene Amplification Contributed to the High Expression of LAMC2 in iCCA
The genetic alteration of LAMC2 gene was investigated.Analysis was performed with genetic data from cBioportal, including a total of 10 953 patients from 33 different cancer types (TCGA PanCancer Atlas Studies).It revealed LAMC2 gene amplification in various cancer patients among all 33 cancer types, with the highest amplification frequency (8.3%) residing in CCA patients (Figure 2A).No LAMC2 mutations were noticed in CCA.
The LAMC2 amplification in iCCA was then validated in Cohort 4. Genomic DNA was extracted from FFPE tissues of 19 iCCA tumors in Cohort 4 and gene copy number assay was performed with a copy number ≥4 as DNA amplification.In this cohort, five out of nineteen iCCA tumors (26.3%) presented LAMC2 copy number ≥4, indicating a noticeable LAMC2 amplification in iCCAs (Figure 2B).Furthermore, iCCA tumors with LAMC2  amplification exhibited strong LAMC2 IHC staining (IHC score ≥6).As shown in Figure 2C, 80% of iCCAs with LAMC2 amplification (4 out of 5) had strong LAMC2 IHC staining, while 57.1% of iCCAs without LAMC2 amplification (8 out of 14) exhibited strong LAMC2 staining.Thus, LAMC2 gene amplification occurred in iCCA, which partially contributed to the high expression of LAMC2 in iCCA tumors.

Silencing LAMC2 Inhibited iCCA Malignancy Features In Vitro and Blocked Orthotopic iCCA Formation In Vivo
In iCCA cell lines, the silencing of LAMC2 using siRNA significantly inhibited cell proliferation and colony formation of iCCA cells (Figure 2D,E).Similar results were acquired when using LAMC2 shRNA in these iCCA cells (Figure S2A-D, Supporting Information).Moreover, LAMC2 knockdown also reduced cell migration (Figure S2E,F, Supporting Information), as reported in extrahepatic CCA. [17]n two HDTV orthotopic iCCA mouse models driven by AKT/NICD and AKT/YapS127A, silencing mouse LAMC2 using shRNA (Figure S2G, Supporting Information) significantly reduced iCCA tumor formation (Figure 2F,G).In both models, all mice in control groups developed massive iCCA tumors (≥ 50 nodules), while mice in all shLAMC2 groups displayed reduced or even no iCCA tumor formation.Quantitatively, silencing LAMC2 largely reduced the iCCA tumor burden, shown through remarkable decreases in liver/body ratios, tumor numbers, and tumor sizes in both iCCA models (p < 0.01 for each comparison).These results further support the oncogenic role of LAMC2 in iCCA.
LAMC2 is a subunit of the extracellular matrix protein laminin332 and a secretory protein.In iCCA cells, both endogenous and exogenous LAMC2 could be secreted (Figure S3A,B, Supporting Information).Previous studies have reported LAMC2 protein cleavage in its protein domain iii [18,19] (Figure S3C, Supporting Information).Therefore, we investigated the functional form of LAMC2 protein in iCCA and found the full-length intracellular LAMC2 (rather than the secreted form) as the major functional form in regulating iCCA malignancy features (Figures S3 and S4, Supporting Information).First, endogenous LAMC2 presented two main protein bands, a major band at ≈150 kDa (the full-length) and a ≈105 kDa band (the long-cleaved C-terminal form) (Figure 1D).The major secreted LAMC2 form was at ≈150 kDa (Figure S3A,B, Supporting Information).Second, LAMC2 containing Flag-tags at its domains v and iii consistently showed ≈150 kDa LAMC2 as its major intracellular and secreted form (Figure S3D, Supporting Information).The presence of ≈105 kDa LAMC2 in cell lysates was at least partially due to the mixed extracellular matrix on the cell membrane, as shown by a lower ≈105 kDa LAMC2 level in cell lysates from the trypsin method (less extracellular matrix) than those from the scraping method (more extracellular matrix) (Figure S3E, Supporting Information).The data align with LAMC2 cleavage occurring extracellularly. [18,19]Third, LAMC2-high RBE cells showed increased cell proliferation and colony formation compared to LAMC2-low RBE cells (Figure S4A,B, Supporting Information).However, the exposure of RBE cells to LAMC2-high and LAMC2low conditioned medium did not lead to differences in cell prolif-eration and colony formation (Figure S4C, Supporting Information).

LAMC2 Enhanced EGFR Protein Level by Promoting EGFR Translation
To investigate the molecular mechanism of LAMC2 in iCCA oncogenesis, Gene Set Enrichment Analysis (GSEA) was performed in iCCA tumors from Cohorts 1 and 5 (with >90 iC-CAs in both cohorts) between the LAMC2-high group and the LAMC2-low group based on the median-LAMC2 cut-off (Figure 3A).Among the top 20 signatures identified from GSEA, several cancer-related signatures, including the CCA signature and EGFR & KRAS signaling signatures, were significantly enriched in both cohorts (Figure 3A).Moreover, when iCCA patients were classified into EGFR signaling activation and inactivation groups based on EGF/EGFR signaling gene sets from literature, [20] LAMC2-high expression iCCA cases notably assembled in the EGFR signaling activation group of both cohorts (p < 0.001) (Figure 3B).Consistent results were obtained in mass spectrometry (MS) proteomic analysis with RBE cells with/without silencing LAMC2 (Figure S5A,B, Supporting Information).GSEA analysis was performed with a significantly altered expression between the two groups.Among the top identified 20 signatures, two EGF/EGFR-related signatures were enriched in the control group compared to the LAMC2 silencing group (Figure S5C, Supporting Information).Meanwhile, protein intensities of genes presented in the EGF/EGFR signaling gene set including EGFR were also mainly reduced in the siLAMC2 group (Figure S5D, Supporting Information).
Comparably, in both RBE and HUCCT1 cells, the silencing of LAMC2 decreased the EGF-mediated EGFR signaling activation as shown by reduced levels of phosphorylated EGFR and phosphorylated ERK (Figure 3C).More importantly, silencing LAMC2 also reduced the EGFR protein level and the baseline of EGFR signaling activation considerably, shown by a reduced level of phosphorylated EGFR (Figure 3C,D).Comparably, LAMC2 overexpression significantly increased the EGFR expression in both RBE and HUCCT1 cells (Figure 3E).In Cohort 5, EGFR and LAMC2 protein levels were significantly positively correlated in 214 iCCA tumors with proteome data (p < 0.001, Figure S6A, Supporting Information).On the other hand, neither the overexpression of EGFR nor the EGF treatment induced LAMC2 expression (Figure S6B,C, Supporting Information).Secreted LAMC2 had no effect on EGF-mediated activation of EGFR signaling either (Figure S6D,E, Supporting Information).
In vivo, overexpressing either the human LAMC2 or the active human EGFR L858R both partially rescued shLAMC2-mediated tumor suppression in the AKT/YapS127A-induced HDTV iCCA mouse model (Figure 3F).Collectively, these results suggested that intracellular LAMC2 increased EGFR protein expression and EGFR was important in LAMC2-mediated iCCA formation.
Mechanistically, LAMC2 did not induce EGFR mRNA expression or increase its protein stability in iCCA cells (Figure S7A-E, Supporting Information).EGFR protein translation was thus examined using Boncat assay in combination with the click chemistry reaction, which could enable the detection of newly synthesized proteins (Figure 3G).This assay revealed that LAMC2 overexpression increased the amount of newly synthesized EGFR (Figure 3H), while silencing LAMC2 reduced it (Figure 3I).Thus, LAMC2 increased EGFR protein expression by promoting its translation.

LAMC2 Interacted with the Unglycosylated EGFR in Endoplasmic Reticulum
In two iCCA cell lines and 293T, exogenously expressed LAMC2-HA interacted with an undersized EGFR, which was ≈40 kDa smaller than the expected 180 kDa EGFR (Figure 4A).Similarly, endogenous LAMC2 also interacted with an undersized endogenous EGFR (≈140 kDa) in two iCCA cell lines (Figure 4B).Moreover, such an interaction was revealed in four out of five tested non-iCCA cancer cell lines, and silencing LAMC2 also reduced the EGFR protein translation, especially in pancreatic cancer cell line Panc-1, lung cancer cell line H1975 and colorectal cancer cell line HCT-116 (Figure S8A-C, Supporting Information).
EGFR is a transmembrane receptor tyrosine kinase that undergoes extensive asparagine (N)-linked glycosylation in its extracellular domain, with 13 N-glycosylation sites [21] (Figure 4C).Each N-linked glycosylation site contributes ≈3 kDa to the molecular weight of the modified protein.Therefore, the unglycosylated EGFR would be ≈40 kDa smaller compared to mature EGFR.We thus tested this possibility using several inhibitors targeting the different steps of the N-linked glycosylation process, including peptide N-glycosidase F (PNGase F) for removing N-glycan chains from proteins, tunicamycin for blocking N-glycan synthesizing, and NGI-1 for inhibiting STT3 complex function in transferring N-Glycans (Figure 4C).
As expected, upon the removal of N-Glycans via PNGase F, the protein size of EGFR was ≈140 kDa.Meanwhile, a strong interaction was detected between exogenous LAMC2 and the N-glycanremoved EGFR (≈140 kDa) (Figure 4D).Similar results were obtained when cells were treated with tunicamycin (Figure 4E) and NGI-1 (Figure 4F).Treatment of tunicamycin and NGI-1 led to the appearance of unglycosylated EGFR (≈140 kDa) and the LAMC2-interacted EGFR remained at the size of ≈140 kDa.Meanwhile, blocking EGFR glycosylation with tunicamycin and NGI-1 yielded a very strong interaction between LAMC2 and EGFR (Figure 4E,F).Similar results were also obtained with the interaction of endogenous LAMC2 and unglycosylated EGFR upon the treatment of tunicamycin and NGI-1 (Figure 4G).
N-linked glycosylation on proteins initiates in the endoplasmic reticulum (ER) lumen.Consistently, LAMC2 and EGFR were clearly co-localized with the ER marker calreticulin around the nucleus, shown by immuno-fluorescence assay (Figure 4H).To further confirm the interaction of LAMC2 with unglycosylated EGFR, an EGFR 13Q -flag construct was generated with N→Q mutations of all 13 glycosylation sites to mimic unglycosylated EGFR.The expressed EGFR 13Q -flag protein had a size of 140 kDa and strongly interacted with LAMC2 (Figure 4I).Mature EGFR undergoes lysosomal-related degradation upon ligand stimulation, [22] while abnormally glycosylated proteins are commonly degraded via the ER-associated protein degradation (ERAD) mechanism. [23]Consistently, an ERAD inhibitor CB-5083, but not a lysosome inhibitor Bafilomycin A1 rescued the degradation of EGFR 13Q -flag (Figure S9A,B, Supporting Information).As a control, Bafilomycin A1 but not CB-5083 partially rescued mature EGFR degradation (Figure S9C, Supporting Information).Thus, the EGFR 13Q -flag fairly represented a nascent unglycosylated situation of EGFR and showed a strong interaction with LAMC2.
Comparably, neither knocking down LAMC2 nor LAMC2 overexpression affected the degradation of EGFR 13Q -flag (Figure S9D,E, Supporting Information).Nonetheless, LAMC2 overexpression significantly increased the protein synthesis of EGFR 13Q -flag (Figure 4J).These data collectively suggested that LAMC2 interacted with nascent or immature EGFR (≈140 kDa) and enhanced its translation.

LAMC2 N-Terminus Interacted with Extracellular Domain of EGFR, Promoting EGFR Translation
We further mapped the interaction regions of LAMC2 and EGFR.Co-IP assay revealed that the N-terminus of LAMC2 (N-LAMC2, domains iii-v), not the C-terminus of LAMC2 (C-LAMC2, domains i-ii), interacted with immature EGFR (Figure 5A).Moreover, when domains iii-v of N-LAMC2 were discretely removed (LAMC2-ΔDv, LAMC2-ΔDiv, LAMC2-ΔDiii), each deletion resulted in a decreased interaction of LAMC2 with immature EGFR (Figure 5B).The interaction reduction was particularly significant when domain iii or domain iv of LAMC2 was removed.Thus, the N-terminus of LAMC2 was essential for interacting with EGFR.Next, Co-IP assay showed that the extracellular region of EGFR interacted with LAMC2, while the cytoplasmic region of EGFR did not (Figure 5C).This interaction aligns with the current understanding that the EGFR extracellular region resides in the ER lumen.
Furthermore, although the intact LAMC2 significantly increased the EGFR translation in both RBE and HUCCT1 iCCA cells, either LAMC2 C-terminus or LAMC2-ΔDiii mutant (as an extra test) did not (Figure 5D).Thus, the interaction between LAMC2 N-terminus and EGFR was necessary for promoting EGFR translation.Consistent results were also obtained when EGFR 13Q -flag was co-transfected with different LAMC2 vectors (Figure 5E).Comparably in the AKT/YapS127A-induced HDTV iCCA mouse model, overexpressing the N-terminus of LAMC2   partially rescued the suppressed iCCA carcinogenesis and progression caused by knocking down mouse LAMC2, but overexpressing LAMC2 C-terminus did not (Figure 5F).Taken together, the LAMC2 N-terminus interacted with the extracellular domain of EGFR during its immature status.This then promoted EGFR translation, contributing to iCCA development.

LAMC2 Promoting EGFR Translation Was Partially Dependent on BiP, an ER Chaperon
A tandem IP followed by mass spectrometry (MS) was carried out to decode mechanisms of LAMC2 in promoting EGFR translation.The top three identified LAMC2 interacting proteins were LAMB1, BiP, and LAMB3 (Figure 6A).LAMB1 and LAMB3 are the subunits of the laminin complex, with LAMB3 being a subunit of laminin332 along with LAMC2.BiP is a resident protein of the ER lumen.It is involved in assisting protein translation via binding to newly synthesized proteins as they are translocated into the ER lumen, and via counteracting the translation inhibitory effects induced by ER chaperon ERdj1, ERdj2/Sec62 and the ERdj6/PERK axis as BiP interacts with ERdj proteins with its nucleotide-binding domain. [24]We thus chose BiP for further investigation.
Co-IP assays revealed a strong interaction between LAMC2 with BiP upon their overexpression (Figure 6B,C).Moreover, silencing BiP in iCCA cell lines led to a reduction in EGFR protein expression promoted by LAMC2 (Figure 6D).Comparable data were obtained when EGFR 13Q was used (Figure 6E).Boncat assay further showed that LAMC2-mediated protein translation of EGFR and EGFR 13Q was visibly suppressed upon BiP silencing in both iCCA cells (Figure 6F,G).Thus, the LAMC2-promoted EGFR translation was partially dependent on BiP.
Besides assisting in protein translation, BiP also recognizes unfolded/misfolded proteins in ER for ERAD degradation or for initiating the unfolded protein response, which reduces the level of targeted proteins. [25,26]However, overexpressing BiP did not reduce, but increased the protein levels of both EGFR and EGFR 13Q in iCCA cells (Figure 6H).In the same set of cells, BiP also increased the protein translation of both EGFR and EGFR 13Q (Figure 6H).Consistent data were obtained when endogenous BiP was silenced (Figure 6I).In this case, BiP mainly promoted EGFR translation.
In addition, when LAMC2 was silenced, BiP-promoted EGFR translation was observably weakened (Figure 6J,K), indicating that the promotion of EGFR translation by BiP was also dependent on LAMC2.Collectively, LAMC2 and BiP increased EGFR protein translation interdependently.

LAMC2, EGFR, and BiP Interacted with Each Other, Contributing to EGFR Translation
The Co-IP assay revealed two EGFR bands interacting with BiP, a full-size EGFR and an undersized ≈140 kDa EGFR, which was similar to EGFRs interacting with LAMC2 (Figure 7A).Concurrently, ≈140 kDa EGFR remained as the major band to interact with BiP when cells were exposed to tunicamycin (Figure 7A), and BiP also strongly interacted with EGFR 13Q (Figure 7B).In this case, not only LAMC2 but also BiP interacted with nascent EGFR without N-glycosylation.
BiP contains a signal peptide (1-18aa), a nucleotide-binding domain (NBD, 125-280aa), a substrate-binding domain (SBD, 420-500aa), and an ER retention KDEL motif.Co-IP assay with BiP truncations revealed that the removal of either SBD or 501-650aa region of BiP completely abrogated the interaction of BiP with LAMC2 (Figure 7D).These two regions were also crucial for BiP to interact with EGFR (Figure 7E).Moreover, the Cterminus of LAMC2, yet not the N-terminus, interacted with BiP (Figure 7F).In line with these findings, the extracellular domain of EGFR interacted with BiP (Figure S10, Supporting Information).These suggested the importance of BiP C-terminus in EGFR translation.Based on these data, a possible interacting model of LAMC2, BiP, and EGFR was proposed (Figure 7G).LAMC2 C-terminus interacts with the BiP C-terminal region (including its SBD domain and region 501-650aa) in the ER, forming a "pocket" composed of the BiP C-terminus and LAMC2 Nterminus.This "pocket" binds to nascent EGFR at its extracellular domain, leading to an increased EGFR translation.
This model was further validated via several Boncat assays after this "pocket" was disrupted.As expected, overexpression of wild-type LAMC2 increased the levels of newly synthesized EGFR or EGFR 13Q in the presence of high levels of BiP expression, whereas LAMC2 lacking C-terminus did not have this effect (Figure 7H; Figure S11A, Supporting Information).Moreover, wild-type BiP enhanced the levels of newly synthesized EGFR or EGFR 13Q , but BiP mutants lacking the region 501-650aa or the SBD domain could not (Figure 7I; Figure S11B, Supporting Information).

LAMC2-High iCCA Tumors Had Poor Prognosis but Were Sensitive to EGFR TKIs Treatment
With the limited number of iCCA patients in Cohort 4, LAMC2high cases (IHC score ≥6) and LAMC2-amplification cases (copy number ≥4) appeared to have shorter overall survival with borderline statistical P-values compared to their corresponding control groups (Figure 8A).Whereas, neither LAMC2 staining nor LAMC2 amplification in this cohort was significantly related to other clinical parameters (Table S3, Supporting Information).Comparable and much more significant data were obtained in Cohort 5 with over 200 iCCA patients.In this cohort, patients with high-LAMC2 protein levels in their iCCA tumors showed significantly shorter overall survival compared to iCCA cases having low-LAMC2 levels, based on various cut-offs of LAMC2 protein level (median, tertile, or quartile division) (Figure 8B; Figure S12A-C, Supporting Information, p < 0.001 for each comparison).
Consistent with LAMC2 promoting EGFR translation, EGFR IHC staining in Cohort 4 showed a significantly higher staining score in iCCA tumors compared to bile ducts (p < 0.001) and a positive correlation with LAMC2 staining in iCCA tumor tissues (p = 0.05) (Figure 8C).We then tested the sensitivity of LAMC2-high iCCA cells to EGFR TKIs treatment.In response to EGFR TKI Gefitinib and EGFR neutralizing antibody Cetuximab, cell viability was lower in LAMC2-high cells (shLAMC2+LAMC2) in comparison to LAMC2-low cells (shLAMC2+Ctrl Vec) of   two iCCA cell lines (Figure 8D,E), indicating that LAMC2-high iCCA cells were more sensitive to EGFR TKIs.This result was further evaluated in vivo using AKT/YapS127A-induced HDTV iCCA mouse model (Figure 8F).Two sets of mice were prepared, i.e., one injected with AKT/YapS127A, and the other injected with AKT/YapS127A and LAMC2.Five weeks after injection, each set was assigned randomly to the treatment group accepting Gefitinib and the control group accepting the sterilized ultrapure water.The results showed that mice in the AKT/YapS127A/LAMC2 control group developed observable massive tumors, while AKT/YapS127A/LAMC2 mice with Gefitinib treatment had much less tumor formation.Significant quantitative data were also obtained on the liver/body ratios and tumor numbers (Figure 8F).This difference was not noticed in AKT/YapS127A-induced HDTV iCCA tumors.Taken together, LAMC2-high iCCA tumors with poor prognosis were sensitive to EGFR TKIs treatment.

Discussion
In this study, our results revealed LAMC2 as a new oncogenic player in iCCA patients and illustrated a signaling axis of LAMC2→ LAMC2/BiP/nascent EGFR→EGFR translation→iCCA carcinogenesis (Figure 8G).In this axis, a high level of LAMC2, partially caused by LAMC2 amplification in iCCA, increased EGFR translation, which in turn promoted iCCA tumorigenesis and demonstrated the sensitivity of LAMC2-high iCCAs to EGFR TKIs treatment.
As a laminin family member, LAMC2 was known to regulate cell invasion, migration, and tumor metastasis in several cancers including CCA. [14,17,27] Meanwhile, several recent studies reported its role in promoting cell proliferation in ovarian cancer and pancreatic cancer. [28,29][32] However, the mechanism remained unknown on whether and how EGFR signaling and LAMC2 regulated each other.Here we found that silencing LAMC2 reduced cell migration but more significant results were observed on its blocking iCCA formation in vivo.Mechanistically, we revealed thoroughly that LAMC2 significantly promoted EGFR translation.LAMC2 and BiP interacted via their C-terminus, creating a "pocket" composed of LAMC2 Nterminus and BiP C-terminus.This pocket captured newly synthesized EGFR without glycosylation and promoted EGFR translation.Meanwhile, it was not only in iCCA cells but also in several non-iCCA cancer cell lines that LAMC2 promoted EGFR translation via interacting with ≈140 kD EGFR.Thus, our findings attributed LAMC2 to a new function in promoting EGFR translation across different cancer cell lines.Moreover, it is likely that LAMC2 might promote not only EGFR translation but also other proteins.Further in-depth investigations are ongoing to uncover more breakthrough findings in this regard.
BiP was previously known to support protein translation via its NBD domain, by which it interacted with ERdj proteins and abolished the translation inhibition mediated by ERdj1, ERdj2/Sec62, or the ERdj6/PERK axis. [24]Consistently, these known mechanisms of BiP contributed to EGFR translation too (Figure S13, Supporting Information), i.e., BiP's NBD domain was important for BiP promoting EGFR translation; overexpression of ERdj1 or ERdj2 consistently reduced BiP-mediated EGFR protein translation; silencing PERK rescued the decrease in EGFR protein translation caused by BiP silencing.We have revealed here the importance of the BiP C-terminal region in EGFR translation (Figure 7), which thus extended the mechanism of BiP in protein translation.It will be interesting to further investigate whether the BiP C-terminal region functions in other proteins' translation and whether such a mechanism mainly relies on LAMC2.In iCCA, LAMC2 was significantly up-regulated whereas BiP expression did not seem to be deregulated across different CCA cohorts (Figure S14A, Supporting Information).BiP levels in iCCA tumors were not related to iCCA prognosis either or EGFR signaling activation (Figure S14B,C, Supporting Information).Therefore, LAMC2 appeared to be the key leading factor and BiP was jointly involved in promoting EGFR protein synthesis, at least in iCCA.
In clinics, several clinical trials have been performed to investigate the use of EGFR TKIs in treating CCA patients, but the results have been disappointing.In this study, a tight relationship was established between the level of LAMC2 and EGFR signaling activation.In this case, LAMC2 might be a valuable indicator to guide the EGFR TKI treatment in clinical practice for iCCA patients.More interestingly, LAMC2 could be secreted from iCCA cells.Therefore, it is worthwhile to further determine and consider the serum LAMC2 level as a non-invasive biomarker to stratify iCCA patients with high LAMC2 levels either for EGFR TKIs therapy alone or concurrently with systemic chemotherapy.
Moreover, the current strategies for targeting EGFR mainly focus on targeting the kinase domain of EGFR or binding the extracellular domain of EGFR to prevent ligand binding or receptor dimerization.Our finding of LAMC2 interacting with an immature EGFR without glycosylation highlighted the presence of nascent EGFR before its undergoing glycosylation and maturation in ER.It therefore offers a new opportunity to develop methods that target nascent EGFR or block its translation, consequently reducing the amount of mature EGFR and the EGFR signaling activation.
Exploring the potential of targeting LAMC2 as a therapeutic approach in iCCA is an intriguing avenue for further research.However, it is important to approach this with caution.Mutations in all three laminin332 subunit chains caused skin disease junctional epidermolysis bullosa (JEB).Knockout mice lacking any of the three chains exhibited symptoms similar to human JEB and died within a few days after birth. [33,34]Considering these phenotypes observed in LAMC2-knockout animals, it might be feasible to further investigate knocking down LAMC2, as we have done in this manuscript, or suppressing LAMC2's function in ER via other methods as the safe and suitable strategies to treat iCCA.

Conclusion
In summary, this study identified LAMC2 as a key oncogenic molecule in iCCA and highlighted its potential as an indicator for guiding EGFR TKI treatment in clinical practice.Amplification of LAMC2 gene in iCCA led to increased levels of LAMC2 protein.Within the ER, the increased LAMC2 together with BiP interacted with newly synthesized EGFR, promoting EGFR translation.Consequently, the LAMC2/EGFR axis contributed to iCCA carcinogenesis, and iCCA tumors with high levels of LAMC2 expression exhibited sensitivity to EGFR TKIs treatments.
TaKaRa).Quantitative reverse transcription polymerase chain reaction (qRT-PCR) was performed with the TB Green Premix Ex Taq II (Cat#RR420, TaKaRa).18S was used as a reference gene.All primer sequences are listed in Table S5 (Supporting Information).
Mouse Studies: All mouse procedures were conducted under the guidelines and the institutional animal care protocol approved by the Experimental Animal Committee at Zhejiang University.ICR mice were purchased from Shanghai SLAC Laboratory Animal Co.Ltd.FVB/N mice were from Beijing Vital River Laboratory Animal Technology.All mice were housed in Zhejiang University Laboratory Animal Center in laminarflow cabinets under specific pathogen-free conditions at room temperature with a 24-h night-day cycle.Oncogene-induced orthotopic iCCA and HCC mouse models were used.For these models, we performed hydrodynamic tail vein injection as we did before in six-week-old ICR or FVB/N mice [39] with the related oncogenes and the sleeping beauty (SB) transposon system.Briefly, the combination of pT3-EF1-myr-AKT and pT3-EF1-YapS127A (AKT/YapS127A), or the combination of pT3-EF1-myr-AKT and pT3-EF1-NICD (AKT/NICD) along with pCMV/SB was introduced to induce iCCA formation through hydrodynamic tail vein injection.pT3-EF1-Myc or the combination pT3-EF1-myr-AKT/NRasV12/pT2-CAGGS (AKT/Ras) along with pCMV/SB was introduced to induce HCC formation.
For each injection, the combined plasmids were diluted in 2 mL saline (0.9% NaCl), filtered through a 0.22 μm filter, and injected into the lateral tail vein of mice in 5-7 s.The detail plasmid combination and amount are listed in Table S6 (Supporting Information).
Conditioned Medium Preparation: When the cultured cells reached ≈90% confluency, the medium was replaced with fresh serum-free medium.Twelve hours later, the conditioned medium containing cell secretome was collected and centrifuged at 800 rpm for 5 min, to remove cell debris.The collected conditioned medium was either immediately used accordingly or stored at −80 °C to be used within 2 weeks.
For colony formation assay, RBE (1000 cells per well) or HUCCT1 (500 cells per well) cells were seeded in 6-cm dishes and cultured for 12 days.Colonies were fixed with methanol, stained with crystal violet, and counted.
For wound healing assay, RBE or HUCCT1 cells were seeded in 6-well plates and infected with the corresponding shRNA virus, artificial would tracks were generated in confluent monolayer cells by scraping with a 20 μL pipette tip.After removal of the detached cells by gently washing with PBS, the cells were incubated with a fresh complete medium.Images were acquired from 6 different fields for each group at the initial time and the later indicated time points.The remaining wound was measured and compared.

Figure 1 .
Figure 1.LAMC2 expressed a specific high level in iCCA tumors.A) Screening of iCCA-specific genes in Cohorts 1-3.B) LAMC2 and KRT19 expression levels in Cohorts 1-3.C) LAMC2 mRNA expression was examined by qRT-PCR.D) LAMC2 protein level was examined by Western blot.E) LAMC2 and KRT19 mRNA levels in tumors from 18 liver cancer mouse models and in normal mouse livers.F) qRT-PCR examination of LAMC2 mRNA level in tumor and non-tumors from various liver cancer mouse models.G) Representative images of LAMC2 IHC staining in Cohort 4. H) LAMC2 IHC staining score in tumors, bile duct cells, and hepatocytes from iCCA and HCC patients.B,F,H) Student's t-test was used.NS, not significant.T, tumor; NT, non-tumor.

Figure 2 .
Figure 2. LAMC2 was amplified in iCCA and silencing LAMC2 inhibited iCCA both in vitro and in vivo.A) LAMC2 amplification frequency of 10 953 patients from 33 cancer types in cBioPortal database.B) LAMC2 copy number detection in iCCA tumors from Cohort 4 (n = 19), 3 iCCA cell lines (HUCCT1, RBE, Huh28) and PBMCs.C) LAMC2 IHC staining score in iCCA patients with or without LAMC2 copy number ≥ 4 from Cohort 4. D) Cell viability and colony formation in RBE cells transfected with control or LAMC2 siRNA.E) Cell viability and colony formation in HUCCT1 cells transfected with control or LAMC2 siRNA.F) iCCA formation in AKT/NICD-induced iCCA mouse model with or without silencing LAMC2 by shRNAs.G) iCCA formation in AKT/YapS127A-induced iCCA mouse model with or without silencing LAMC2 by shRNAs.D,E) Two-way ANOVA was used for cell viability assay.Student's t-test was used for colony formation assay.F,G) Student's t-test was used.PBMC, peripheral blood mononuclear cell.

Figure 3 .
Figure 3. High LAMC2 associated with EGFR signaling activation and LAMC2 promoted EGFR translation.A) GSEA analysis with significantly altered genes (p < 0.01) between LAMC2-high and LAMC2-low iCCA patients.The top 20 enriched signatures were listed.B) The enrichment of LAMC2-high patients in EGFR signaling activation and non-activation groups subclassified by an EGF/EGFR signaling gene set.C) Western blot analysis in RBE and HUCCT1 cells transfected with control siRNA, siLAMC2 #1, or siLAMC2 #2 and treated with EGF.D) Western blot analysis in RBE and HUCCT1 cells transfected with control siRNA as well as siLAMC2 #1 and #2.E) Western blot analysis of RBE and HUCCT1 cells transfected with Ctrl vector or LAMC2-HA, along with EGFR-flag.F) iCCA tumor formation in AKT/YapS127A-induced iCCA mouse model with or without silencing LAMC2 by shRNAs, upon with or without LAMC2/EGFR L858R overexpression.Student's t-test was used.G) The flow chart of Boncat Assay with L-AHA to detect the newly synthesized proteins.H) Boncat assay in RBE and HUCCT1 cells with LAMC2 overexpression.I) Boncat assay in RBE and HUCCT1 cells with LAMC2 silencing.

Figure 4 .
Figure 4. LAMC2 interacted with unglycosylated EGFR.A) Cells were co-transfected with indicated vectors and IP was performed with anti-HA beads.B) IP assay with anti-LAMC2 antibody in iCCA cells.C) N-linked glycosylation sites of EGFR (up panel) and the flow chart of the N-glycosylation process (bottom panel).The related N-glycosylation inhibitors and N-glycan removing enzyme were indicated in red color.D) IP assay with anti-HA beads in cells co-transfected with LAMC2-HA and EGFR-flag.Both cell lysates and IP products were treated with or without PNGase F for 1 h before analysis.E) Anti-HA IP in cells co-transfected with LAMC2-HA and EGFR-flag and treated with or without tunicamycin (0.5 μg mL −1 ) for 24 h.F) Anti-HA IP in cells co-transfected with LAMC2-HA and EGFR-flag and treated with or without NGI-1 (10 μm) for 24 h.G) Anti-LAMC2 IP in RBE and HUCCT1 cells treated with tunicamycin (0.5 μg mL −1 ) or NGI-1 (10 μm) for 24 h.H) Confocal microscopy images of endogenous LAMC2, EGFR, and ER marker Calreticulin and their co-localization in iCCA cells.I) Construction of EGFR 13Q -flag and anti-HA IP assay in cells co-transfected with LAMC2-HA and EGFR 13Q -flag.J) Boncat assay in cells co-transfected with LAMC2-HA and EGFR 13Q -flag.SP, signal peptide; TM, transmembrane; N, Asparagine; Q, Glutamine.

Figure 5 .
Figure 5. LAMC2 N-terminus interacted with the extracellular domain of EGFR, promoting EGFR translation.A) Mapping LAMC2 regions involved in EGFR binding via IP in cells co-transfected with EGFR-flag and LAMC2-HA deletion mutants.B) Mapping LAMC2 N-terminus regions involved in EGFR binding via IP in cells co-transfected with EGFR-flag and LAMC2-HA deletion mutants.C) Mapping EGFR regions involved in LAMC2 binding via IP in cells co-transfected with LAMC2-HA and different EGFR-flag vectors.D) Boncat assay in cells co-transfected with LAMC2-HA deletion mutants and EGFR-flag.E) Boncat assay in cells co-transfected with LAMC2-HA deletion mutants and EGFR 13Q -flag.F) iCCA tumor formation in AKT/YapS127Ainduced iCCA mouse model with or without silencing LAMC2 by shRNAs, upon with or without C-LAMC2/N-LAMC2 overexpression.Student's t-test was used.

Figure 6 .
Figure 6.LAMC2 promoting EGFR translation was partially dependent on BiP.A) The flow chart of tandem IP followed by MS, and the top 4 candidates were listed based on the quantity of unique peptides.B) Anti-HA IP in cells co-transfected with LAMC2-HA and BiP-flag.C) Anti-flag IP in cells cotransfected with LAMC2-HA and BiP-flag.D) EGFR-flag expression in cells co-transfected with LAMC2-HA and EGFR-flag with or without silencing BiP.E) EGFR 13Q -flag expression in cells co-transfected with LAMC2-HA and EGFR 13Q -flag with or without silencing BiP.F) Boncat assay in cells co-transfected with LAMC2-HA and EGFR-flag with or without silencing BiP.G) Boncat assay in cells co-transfected with LAMC2-HA and EGFR 13Q -flag with or without silencing BiP.H) Boncat assay in cells transfected with BiP-flag and EGFR-flag (left panel) or EGFR 13Q -flag (right panel).I) Boncat assay in RBE and HUCCT1 cells upon BiP silencing.J,K) Boncat assay in cells co-transfected with BiP-HA and EGFR-flag (J) or EGFR 13Q -flag (K) with or without silencing LAMC2.

Figure 7 .
Figure 7. LAMC2, EGFR, and BiP interacted with each other, contributing to EGFR translation.A) Anti-flag IP in cells co-transfected with BiP-flag and EGFR-HA with or without tunicamycin treatment.B) Anti-HA IP in cells co-transfected with BiP-HA and EGFR 13Q -flag.C) Schematic diagram of BiP functional domain and a group of BiP truncations.D) Mapping BiP regions involved in LAMC2 binding via IP in cells co-transfected with LAMC2-HA and different BiP-flag vectors.E) Mapping BiP regions involved in EGFR binding via IPs in cells co-transfected EGFR-HA and different BiP-flag vectors with tunicamycin treatment.F) Anti-HA IP in cells co-transfected with BiP-flag and different LAMC2 vectors.G) An illustrated interacting model of LAMC2, BiP, and nascent EGFR in promoting EGFR translation in ER.H) Boncat assay in cells co-transfected with BiP-flag/EGFR-flag and an intact LAMC2, or N-LAMC2.I) Boncat assay in cells co-transfected with LAMC2-HA /EGFR-flag and an intact BiP, or BiP ΔSBD and BiP Δ501-650 .SP, signal peptide; NBD, nucleotide-binding domain; SBD, substrate binding domain.

Figure 8 .
Figure 8. LAMC2-high iCCA tumors had poor prognosis, but were sensitive to EGFR TKIs treatment.A) Kaplan-Meier survival analysis of iCCA patients from Cohort 4 based on LAMC2 IHC staining score and LAMC2 copy number.B) Kaplan-Meier survival analysis of iCCA patients from Cohort 5 based on LAMC2 protein level (the median cut-off).C) EGFR IHC staining and the spearman correlation of EGFR and LAMC2 in Cohort 4. Student's t-test was used for group comparison.D,E) Cell viability of iCCA cells with different LAMC2 levels under treatment of EGFR TKI Gefitinib (C) or EGFR neutralizing antibody Cetuximab (D).Two-way ANOVA analysis was used.F) Representative images and quantitative analysis of orthoptic iCCA tumor formation in AKT/YapS127A-induced iCCA mouse model with or without LAMC2, upon with or without Gefitinib treatment.Student's t-test was used.NS, not significant.G) The schematic model summarized LAMC2 as a key oncogenic event in iCCA.