Dual‐targeting therapy against HER3/MET in human colorectal cancers

Abstract Background Colorectal cancer (CRC) is the most common malignancy in the world, and novel molecular targeted therapies for CRC have been vigorously pursued. We searched for novel combination therapies based on the expression patterns of membrane proteins in CRC cell lines. Results A positive correlation was observed between the expression of human pidermal growth factor receptor (HER) 3 and mesenchymal‐to‐epithelial transition factor (MET) on the cell surface of CRC cell lines. The brief stimulation of HER3/MET‐high SW1116 CRC cells with both neuregulin‐1 (NRG1) and hepatocyte growth factor enhanced ERK phosphorylation and cell proliferation more than each stimulation alone. In addition, a prolonged NRG1 stimulation resulted in the tyrosine phosphorylation of MET. In this context, the Forkhead Box protein M1 (FOXM1)‐regulated tyrosine phosphorylation of MET by NRG1 was demonstrated, suggesting the existence of a signaling pathway mediated by FOXM1 upon the NRG1 stimulation. Since the co‐expression of HER3 and MET was also demonstrated in in vivo CRC tissues by immunohistochemistry, we investigated whether the co‐inhibition of HER3 and MET could be an effective therapy for CRC. We established HER3‐and/or MET‐KO SW1116 cell lines, and HER3/MET‐double KO resulted in the inhibition of in vitro cell proliferation and in vivo tumor growth in nude mice by SW1116 cells. Furthermore, the combination of patritumab, an anti‐HER3 fully human mAb, and PHA665752, a MET inhibitor, markedly inhibited in vitro cell proliferation, 3D‐colony formation, and in vivo tumor growth in nude mice by SW1116 cells Conclusion The dual targeting of HER3/MET has potential as CRC therapy.


| INTRODUCTION
Colorectal cancer (CRC) is one of the leading causes of human cancer death worldwide. 1,2 Surgery and classical chemotherapy are generally the main treatment for CRC. Furthermore, molecular targeted therapy against human epidermal growth factor receptor (HER) 1 with therapeutic monoclonal antibodies (mAb) such as cetuximab and panitumumab are effectively administered to patients with CRC, 3 and other HER family proteins may also be attractive targets for CRC therapy. The HER family consists of four receptor-type tyrosine kinases: HER1 (EGFR), HER2 (erbB-2), HER3 (erbB-3), and HER4 (erbB-4), and these receptors form homo-or heterodimers, leading to the activation of receptors. The activation of HER family proteins stimulates intracellular events, such as the MAPK and PI3K-AKT pathways, 4 resulting in cancer initiation and progression.
The phosphor-transfer reaction of HER3 is catalytically impaired, resulting in very weak kinase activity; however, HER3 may be activated by forming heterodimers with other HER family receptors. By the binding to HER3 ligands, like neuregulin 1 (NRG1), HER3 plays important roles, such as in the differentiation and proliferation of normal and cancer cells. [5][6][7][8] The kinase activity of the HER2/HER3 heterodimer is stronger than that of other HER family heterodimers. 9,10 Regarding molecular targeted cancer therapies, HER3 was found to be crucial for limiting the effects of HER kinase inhibitors. 11,12 Mesenchymal-to-epithelial transition factor (MET or c-MET) is a receptor-type tyrosine kinase that is activated by the binding of hepatocyte growth factor (HGF). The HGF/MET pathway is essential in cell proliferation, motility, and the development of resistance to cancer therapy. 13,14 Growing evidence has suggested that the presence of different cell populations within a single tumor may limit the effectiveness of therapies targeting a single molecule. 15,16 Therefore, the efficacy of cancer therapy may be enhanced by simultaneously inhibiting multiple therapeutic targets. We searched for novel combination therapies based on an expression analysis of membrane proteins on CRC cell lines.
In the present study, we demonstrate that the cell surface expression of HER3 and MET proteins positively correlates in human CRC, the NRG1-stimulated tyrosine phosphorylation of MET is regulated by the transcription regulator, Forkhead Box protein M1 (FOXM1), and the coinhibition of HER3 and MET represents an effective therapy for human CRC.

| Animals
Female F344 rats and Male KSN nude mice were obtained from Shimizu Animal Farm (Kyoto, Japan). Animals were individually raised in plastic cages and maintained under specific pathogen-free conditions and housed in under a standard light/dark cycle at a constant temperature of 23 ± 1°C. Experiments were approved by the Committee for the Care and Use of Laboratory Animals at Kindai University (KAPS-23-004) and performed following the institutional guidelines and the United States National Institutes of Health Guide for the Care and Use of Laboratory Animals.

| Immunohistochemistry (IHC) of human colon tissue sections
Colon adenocarcinoma with a matched adjacent colon tissue array (CO487, US Biomax, Inc.) was used. Following deparaffinization and rehydration, tissue sections were treated with solution (Target Retrieval solution S1699) at 98°C for 5 min 3 times by a microwave to retrieve the antigen. After three washes with PBS ("washing" hereafter), these sections were incubated with 3% H 2 O 2 -methanol at room temperature for 10 min to quench the effects of endogenous peroxidase activity. After washing, they were incubated with Block Ace (DS Pharma Biomedical Co., Ltd.) at room temperature for 2 h. They were then incubated with Ab4 anti-HER3 (10 μg/mL) or Ab57 anti-MET mAb (10 μg/mL) primary rat mAb diluted in 1% BSA-PBS at 4°C overnight. After washing, sections were incubated with biotinylated goat anti-rat IgG (Vector Laboratories) diluted 1:200 in 2% BSA-PBS at room temperature for 30 min. After washing, samples were treated with avidin-biotin-peroxidase complex (ABC) reagent (Vector Laboratories) diluted 1:50 in 2% BSA-PBS at room temperature for 30 min. After washing, sections were incubated with 0.05% 3,3′-diaminobenzidine (Dojin Chemicals) and 0.01% H 2 O 2 in 0.1 M Tris-HCl (pH 7.4), and then counterstained with hematoxylin. Sections were dehydrated, cleared in xylene, and mounted in Permount (Fisher Scientific). IHC scores for HER3 or MET expression were set based on the standard HER2 test. IHC gave a score of 0-3, indicating the amount of HER3/MET in tissue samples. HER3/MET expression in normal (adjacent to or distant from CRC) and cancer tissues was estimated as score 0 (negative), score 1 (weak or borderline), score 2 (intermediate), or score 3 (strong) by two pathologists.

| Cell cycle analysis
The cell cycle distribution was analyzed by FCM. Briefly, cells were incubated for 96 h, washed with cold PBS, fixed in 70% ethanol overnight at −20°C, and stained with DAPI (Nacalai) reagent at 4°C for 15 min. The fluorescence intensity of the final mixture was analyzed by BD LSR Fortessa.

| Cell proliferation analysis
Cell proliferation was assessed by WST-8 (Cell Count Reagent SF, Nacalai). After cells were cultured in 96well plates, WST-8 (10 μL /well) was added to each well at different time points and incubated at 37°C for 2 h. Absorbance (measurement: 450 nm/reference: 600 nm) was measured using a fluorescence spectrophotometer (Infinite 200 PRO series; Tecan, Switzerland).

| Anchorage-dependent growth in 3D cultures
Matrigel® (CORNING®) was added at a volume of 150 μL to 8-well chamber slides (WATSON Bio Lab) and incubated at 37°C for 15 min, after which 5000 cells were seeded. Patritumab and the MET inhibitor were added the next day, and after a 4-day incubation, 50 cells were measured as one colony after imaging with BIOZERO (KEYENCE) in nine fields of view/one well.

| Statistical analysis, heat map analysis, principal component analysis (PCA), and assessment of additive or synergistic effects
Regarding the expression levels of cell surface proteins, S/N or ΔMFI was calculated. The results obtained were transformed into log 2 values and subtracted from the median. A clustered heat map analysis and PCA were conducted with the R Project for Statistical Computing (https://www.r-proje ct.org/). Data for correlations were analyzed using Pearson's correlation coefficient. Experimental data were analyzed using Prism 5 for Win (GraphPad Software). The criterion for significance was at least.

| Comprehensive cell-surface protein analysis of human colon cell lines
To identify targets for simultaneous inhibition, we initially examined the expression of cancer-associated cell-surface proteins in 24 human colon cell lines and analyzed the expression levels of 14 cell-surface proteins with specific mAb by FCM. Protein (vertical axis)/cell (horizontal axis) types were classified by the clustered heatmap ( Figure 1A). Cell lines were coarsely classified into two groups, SNU-C1/CCD841 and the other CRC cell lines. These two cell lines have different characteristics from the other cell lines, namely, SNU-C1 has the morphology of floating cells with round cell aggregates, and CCD841 is the only normal colon-derived cells among the 24 colon cell lines. Surface proteins were separated into three groups. The first group included cells expressing HER1, HER2, HER3, and MET oncogene products. The second group included cells expressing ASCT2, LAT1, CAT1, CD98hc, and xCT amino acid transporters. The third group included cells expressing CD44s, CD44v8, and CD44v9 hyaluronan receptors. Cells expressing EpCAM or HER4 were located in two gaps adjacent to three groups. A positive correlation was observed between the cell-surface expression of HER3 and MET in CRC cell lines ( Figure 1B), and we estimated 13 CRC lines in 24 cell lines (54%) as a HER3/MET double-high (DH) group. HER3 and MET protein expression was low in SNU-C1 and CCD841 cells, uniquely characterized in Figure 1A. Furthermore, 3D-PCA highlighted that SNU-C1/CCD841 and the other cell lines clustered separately ( Figure 1C). The higher expression of both HER3 and MET in some human CRC tissues than in normal colon tissues was demonstrated by IHC ( Figure 1D). The HER3 or MET-high (IHC score: 2 and 3) ratio was 86 or 66%, respectively, and the HER3/MET-DH group was estimated as more than half of the CRC lines. These results indicated that an analysis of cell-surface protein expression has the potential to classify CRC, and the HER3/ MET status may extract a unique population in human CRC. We selected SW1116 as a representative HER3/ MET-high CRC in this study, because this cell line was the most suitable CRC for the establishment of KO cell lines by CRISPR/Cas9-based gene disruption. Therefore, we used mainly SW1116 in subsequent analyses. We also used HT29 as the second HER3/MET-DH CRC ( Figure 1A-C) in some experiments.

| Effects of NRG1 and HGF on the proliferation of SW1116 cells
We examined the effects of NRG1 and HGF on the phosphorylation of HER3 and MET and on the proliferation of SW1116 cells. A brief ligand stimulation resulted in the phosphorylation of the receptor for each ligand (Figure 2A). After a prolonged ligand stimulation, the potential activation of additional downstream signaling was verified. The phosphorylation of AKT was only induced by the NRG1 stimulation and was not further enhanced by the combined use of HGF. On the contrary, the phosphorylation of ERK was enhanced by the combined dual stimulation than by the single stimulation ( Figure 2B). Furthermore, the combined dual stimulation with NRG1 and HGF significantly enhanced cell proliferation more than the stimulation with each ligand alone ( Figure 2C).

| NRG1-induced MET phosphorylation in human CRC cells
To analyze potential crosstalk between HER3 and MET, we initially examined MET activation by NRG1 in SW1116 cells. NRG1 induces the formation of the HER2/ HER3 heterodimer and the activation of HER3. We confirmed the formation of HER2/HER3 heterodimers and demonstrated MET phosphorylation by the NRG1 stimulation for 12 h ( Figure 3A). MET phosphorylation by the prolonged (12 h) NRG1 stimulation was also observed in HT29 cells ( Figure S1). To elucidate the mechanisms underlying NRG1-induced MET phosphorylation, we examined whether a direct interaction occurred between HER3 and MET. Heterodimers between MET and HER2 or MET and HER3 were not detected, whereas MET and HER1 formed heterodimers with or without NRG1 ( Figure 3B). We then focused on FOXM1, a transcription factor that regulates MET expression. 36 HER3 phosphorylation was induced 5 min after the addition of NRG1, whereas MET phosphorylation was only weakly observed at this point. The prominent phosphorylation of MET and induction of FOXM1 were detected 12 h after the addition of NRG1 and were maintained at least until 24 h ( Figure 3C). Although total MET expression seems unchanged by FOXM1 induction in Figure 3C, however, elevation of the cell-surface MET by NRG1 was observed with FCM analysis ( Figure S2). Since the formation of HER2/HER3 heterodimer was demonstrated in the presence of NRG1 ( Figure 3A), effects of anti-HER2 pertuzumab 37 or anti-HER3 patritumab, 32 which respectively recognizes HER2-epitope or HER3-epitope and inhibits heterodimerization between HER2 and HER3, on the NRG1-induced HER3 activation (phosphorylation) and the induction of FOXM1, were analyzed. Pertuzumab and patritumab inhibited the NRG1-induced phosphorylation of HER3 and MET and suppressed increases in the expression of FOXM1 ( Figure 3D). FOXM1-KO SW1116 cells were then established to investigate the contribution of FOXM1 to the NRG1-induced phosphorylation of MET. In FOXM1-KO SW1116, NRG1 did not induce the phosphorylation of MET, whereas the phosphorylation of HER3 was induced by NRG1 ( Figure 3E). We used the MEK inhibitor (U0126) and FOXM1 inhibitor (thiostrepton) to examine the effects of MET phosphorylation induced by the NRG1 stimulation on the MAPK pathway. The NRG1-induced phosphorylation of MET was inhibited by U0126 or thiostrepton. The inhibition of the MAPK pathway by U0126 was previously shown to decrease FOXM1 expression, 38,39 which is consistent with the present results ( Figure 3F). We also examined the effects of NRG1 or HGF on the MAPK pathway in FOXM1-KO SW1116 cells to demonstrate the importance of FOXM1-mediated signaling induced by the NRG1 stimulation. The NRG1 or HGF stimulation induced the phosphorylation of ERK in FOXM1-KO SW1116 cells ( Figure 3G,H).

MET-KO on the cell proliferation, cell-cycle progression, and tumor growth of CRC
To analyze the effects of HER3-and/or MET-KO on cell growth, we established HER3 and/or MET-KO SW1116  Figure 4B). Additionally, FOXM1 expression in HER3/MET-dKO SW1116 cells was significantly decreased, as compared to HER3-KO or MET-KO, suggesting the role of both HER3 and MET in the regulation of FOXM1 expression. Cell proliferation was significantly lower in HER3/MET-dKO cells than in other SW1116 cells ( Figure 4C). We then examined the effects of HER3-, MET-, or HER3/MET-KO on the cell-cycle progression of SW1116 cells by FCM. In HER3/MET-dKO SW1116 cells, cell-cycle arrest was observed in the G0/G1 phase ( Figure 4D,E). In MET-KO cells that showed enhanced cell proliferation in Figure 4C, increases in the S-phase fraction and decreases in the G0/G1-phase fraction were observed ( Figure 4D,E). We also investigated the effects of HER3-and/or MET-KO on in vivo tumor growth by xenografted SW1116 cells in nude mice. Periodic calculation of growing tumor volumes revealed that tumor growth was strongly inhibited in HER3/MET-dKO SW1116 tumors, as compared to parental SW1116 tumors ( Figure 4F). From photo image of each excised tumor in four groups at day 28, we also obtained similar result as in Figure 4F (Data not shown).

PHA665752 on the in vitro cell proliferation and in vivo tumor growth of human CRC
To examine the efficacy of the combined treatment of anti-HER3 and anti-MET drugs against human CRC cells, we evaluated the effects of patritumab and PHA665752 against SW1116 cells using in vitro colony formation in 3D-culture and in vivo tumor growth in nude mice. The combined treatment of patritumab and PHA665752 resulted in the significantly stronger inhibition of colony formation ( Figure 5A) and tumor growth ( Figure 5B) than the single treatment with patritumab or PHA665752 alone. In view of the importance to use multiple cell lines, HT29 (HER3/MET-high CRC) was selected for the in vitro and in vivo evaluation of dual-targeting therapy against HER3 and MET. Although in vitro proliferation of HT29 CRC cells by singly treated PHA665752 was inhibited, inhibitory effects were significantly obvious in the combined use of patritumab and PHA665752 ( Figure S3). Furthermore, patritumab and PHA665752 significantly inhibited in vivo tumor growth of HT29, as compared to singly treated patritumab or PHA665752 ( Figure S4).

| DISCUSSION
Due to the heterogeneity of cancer cells or cancer tissues, single-agent therapy often resulted in limited efficacy 40 ; therefore, combination therapy has become the mainstream 41 in recent years. We have developed mAb toward effective combination therapy against specific target molecules, including oncogene products, 18,19,[28][29][30][31] amino acid transporters , 20,21,23,[25][26][27][28][29]33 and adhesion molecules. 17,21,22,24,28,29 In this context, HER3 has been associated with the acquisition of drug resistance to anti-EGFR (HER1) and anti-HER2 cancer therapy, 11,42 and MET is involved in the proliferation, migration, invasion, and motility 43-45 of cancer cells. Therefore, HER3 and MET have been anticipated as potential therapeutic targets. We produced novel anti-HER3 mAb that exerted anti-tumor effects on human colon and breast cancer cells in xenografted mouse models, 18 and recently developed specific anti-MET mAb 28,29 in addition to transporters 20,25,26,28,29 and adhesion molecules. 17,21,28,29 In the present study, we analyzed the expression of 14 cancer-associated proteins in 24 CRC cell lines by FCM. A positive correlation was reported between the expression of HER2 and HER3 in primary colorectal cancer. 46 Similar results were obtained in the present study, namely, the correlation coefficient (r) of HER2 versus HER3 = 0.711 (data not shown). A positive correlation between the expression of HER3 and MET ( Figure 1B Regarding the effects of HER3 and MET on cancer cell proliferation, the combined stimulation of both ligands enhanced cell proliferation more than either ligand alone ( Figure 2C), suggesting that HER3 and MET contribute to cell proliferation in a concerted manner. Since NRG1 induces not only HER3 phosphorylation but also MET phosphorylation ( Figure 3A), a pathway in which HER3 regulates MET signaling in CRC was postulated. The formation of the HER3/MET heterodimer was reported in human lung cancer cells. 47 We did not detect a direct relationship between HER3 and MET; however, HER1/MET heterodimers were identified in the presence or absence of NRG1 ( Figure 3B). The phosphorylation of F I G U R E 5 Effects of patritumab and PHA665752 on the in vitro cell proliferation and in vivo tumor growth of human CRC cells. (A) SW1116 were cultured in Matrigel-coated slides in the presence of patritumab (30 μg/mL) and/or PHA665752 (0.5 μM) for 4 days. Cells were cultured with no ligand in this experiment. Using colonies containing more than fifty cells, results were expressed as the means ± SEM (n = 3). *p < 0.05, **p < 0.01, one-way ANOVA followed by Tukey's post hoc multiple comparison tests. (B) After the confirmation of tumor development of SW1116, mice were injected i.p. with Patritumab (12 mg/kg) and/or PHA665752 (0.6 mg/kg) on days 1, 7, 13, and 19. Tumor volumes were measured every 3 days using digital calipers and were quantified using the formula: volume [mm 3 Results are expressed as means ± SEM (n = 6). *p < 0.05, two-way ANOVA followed by Tukey's post hoc multiple comparison tests.
MET, but not HER3, and the induction of FOXM1 were diminished in FOXM1-KO SW1116 cells ( Figure 3E), indicating that FOXM1 is essential for the NRG1induced phosphorylation of MET. In this context, Figure 3G and 3H shows that HER3 and MET can transmit FOXM1-independent individual proliferation signals, in addition to FOXM1-dependent HER3/MET-mediated proliferation signal. These results suggest the existence of a pathway through which HER3 regulates MET signaling via FOXM1. Although we investigated the involvement of HGF as a mechanism for MET phosphorylation by NRG1, there was no increase in HGF expression at the mRNA or protein level in SW1116 cells from the NRG1 stimulation (data not shown).
Effects of NRG1 and HGF on colon cancer cell proliferation shown in Figures 2 and 3 are summarized in Figure 6. We assumed that at least three MAPK-based signals may be induced by the stimulation with NRG1 and HGF. The first is the activation of FOXM1-independent MAPK signaling by the NRG1 stimulation ( Figure 3G); the second is a signal that induces the activation of MET via FOXM1 upon the NRG1 stimulation ( Figure 3E), which was newly discovered in the present study; and the third is a signal that induces the activation of MAPK by the HGF stimulation ( Figure 3H). The activation of these three MAPK signals enhanced the phosphorylation of ERK by the combined use of ligands, which we assume is the reason why HER3 and MET contributed to cell proliferation in a concerted manner ( Figure 6).
Contrary to expectations, MET-KO promoted in vitro cell proliferation and in vivo tumor growth of SW1116. Although a marked increase in the percentage of cells in the S phase may be involved in these phenomena, the underlying mechanisms remain unclear. In HER3/MET-dKO SW1116 cells, in vitro cell proliferation and in vivo tumor growth were significantly inhibited and G0/G1 arrest was induced, indicating that simultaneous HER3/MET targeting was effective ( Figure 4B-F). Since FOXM1-KO markedly inhibited cell proliferation ( Figure S5), decreased expression of FOXM1( Figure 4B) may be an important factor in suppressed cell proliferation ( Figure 4C) and tumor growth ( Figure 4F) of HER3/MET-dKO cells.
We showed that HER3 and MET indirectly affected each other in the presence of each ligand; however, essential or ultimate effectors transcriptionally regulated by FOXM1 that are responsible for the NRG1-induced phosphorylation of MET and anti-tumor effects by the dual targeting of HER3/MET remains to be solved. As to possible mechanism of "NRG1-stimulated and FOXM1mediated MET phosphorylation," we remark cell-surface expression of MET. Total MET expression seems not altered in FOXM1 induction in Figure 3C, however, elevation of the cell-surface MET by NRG1 was observed with FCM analysis ( Figure S2). It has been reported that both the proportion of MET dimers and the level of MET autophosphorylation increased at high MET protein density, 48 therefore, possible MET homodimerization promoted by high density of cell-surface MET in this study could result in the efficient phosphorylation of MET in SW1116 cells.
In recent years, the combination therapy has been attracting attention, because of the possible additive or synergistic potential for effective therapeutic effects through independent mechanisms of action. 49 However, it is not easy to search for effective molecules for cotargeting from huge combinations, as in the example of co-targeting of IGF1R with aromatase inhibitor, estrogen receptor antagonist, 50,51 or HER2 inhibitor, 52 which has resulted in the poor clinical response. In this context, dual targeting of HER3 or MET seems to be one of the most promising mechanisms, which could overcome the resistance of cancer cells against anti-HER1 and anti-HER2 therapy. 11,12,[53][54][55] The present results demonstrated that HER3 and MET were not directly associated with each other in CRC but cooperated for the cellular and tumor growth of CRC ( Figure 6). The combination of anti-HER3 (Patritumab) and anti-MET (PHA665752) drugs was effective against HER3/MET-high SW1116 cells, indicating the potential of the dual inhibition of HER3 and MET as the possible targeting therapy overcoming the resistance of cancer cells against anti-HER1 and anti-HER2 therapy in CRC and various human epithelial cancers.

ACKNOWLEDGMENTS
We thank Dr. Shimada-Takaura K (Kindai University) and Dr. Yoshioka T (Akita University) for their technical advice on 3D-PCA and IHC judgments. The corresponding author is grateful to Drs. Nishikawa T (Nishikawa ENT Clinic), Saya H (Fujita Cancer Center), Uda E (Higashiosaka City Medical Center), Shibamoto Y (Narita Memorial Proton Center), Kusano J, Ishigaki T, Kitaura R (Japanese Red Cross Osaka Hospital), Yazaki Y, Nishimura T (Seren Clinic), Myojo T (Edogawa Plus Clinic), Yamamoto K (Osaka Hospital, Japan) and Hamada Y (Hamada Internal Medicine Clinic) for their clinical advice and radio-immunotherapy on the pharynx cancer of TM, which was detected during the process of preparing this manuscript.

FUNDING INFORMATION
This work was supported by MEXT/JSPS KAKENHI Grant Number 18H05463 (Grant-in-Aid for Scientific Research on Innovative Areas, Toward new frontiers: encounter and synergy of state-of-the-art astronomical detectors and exotic quantum beams).

CONFLICT OF INTEREST STATEMENT
Masuko T and Endo Y were supported by commissioned research fund to Kindai University (J20084, 2020-2021) from HEALIOS K.K. (Japan). The other authors declare that they have no conflicts of interest.

DATA AVAILABILITY STATEMENT
The data used and analyzed during this study are available from the corresponding author on request.