Targeting TYRO3 inhibits epithelial–mesenchymal transition and increases drug sensitivity in colon cancer


Colon cancer is the third leading cause of death from cancer worldwide with less than 10% survival rate at the late stage. Although mutations of certain genes have been implicated in familial colon cancer development, the etiology of the majority of colon cancer remains unknown. Herein, we identified TYRO3 as a potential oncogene. Immunohistochemical staining results demonstrated that levels of TYRO3 were markedly elevated in polyps and colon cancer cells and were negatively correlated with prognosis. Overexpression of TYRO3 enhanced cell motility, invasion, anchorage-independent growth and metastatic ability, while knockdown of TYRO3 impaired all these processes. Results from meta-analysis showed that TYRO3 was associated with epithelial–mesenchymal transition (EMT) signatures. Gain-of-function and loss-of-function experiments demonstrated that expression of SNAI1, the master regulator of EMT, was regulated by TYRO3 and played a major role in mediating TYRO3-induced EMT processes. The murine model also demonstrated that Tyro3 and Snai1 were upregulated in the early stage of colon cancer development. To provide a proof-of-concept that TYRO3 is a druggable target in colon cancer therapy, we raised anti-TYRO3 human antibodies and showed that treatment with the human antibody abolished TYRO3-induced EMT process. More importantly, administration of this anti-TYRO3 antibody increased drug sensitivity in primary cultured colon cancer cells and xenografted mouse tumors. These findings demonstrate that TYRO3 is a novel oncogene and a druggable target in colon cancer.


Colorectal cancer is a worldwide disease, of which incidence continues to rise each year. Despite the efforts in improving treatment regimen and clinical care, 50% of patients with colorectal cancer die within 5 years. The death rate of colon cancer further increases to 90% at the late stage.1 The development of colorectal cancer is a complex and multiple-step process. In 1990, Fearon and Vogelstein2 first proposed a genetic model of colorectal carcinogenesis to explain the formation of colon cancer from normal colonic tissue. This model states that colorectal cancer is a result of mutant gene accumulation, including tumor suppressor genes and oncogenes. It is well known that adenomatous polyposis coli gene (APC), K-ras and P53 are involved in colorectal cancer progression. APC serves as a tumor suppressor gene to suppress Wnt/β-catenin signaling from activation. Mutant APC are usually found in polyposis and familial colorectal cancer.3 Mutant K-ras provides neoplastic properties by constitutive activation of Ras signaling pathway4 while mutant P53 induces higher transcription activity and leads to aberrant proliferation.5 These are commonly known oncogenes/tumor suppressor genes involved in colorectal cancer formation. However, the hereditary colorectal cancer accounts for only 40–50% of all colorectal cancer cases; the majorities (>50%) are sporadic colon cancers arising from unknown mechanisms.

In an attempt to investigate novel genes contribute to colon cancer development and progression, we employed an in-house bioinformatic platform (Tumor Associated Gene, TAG)6 to identify potential oncogenes and/or tumor suppressor genes. TYRO3 was identified as the top candidate in the oncogene category. TYRO3, also known as Tif, Sky, BYK and Dtk, belongs to the TAM (TYRO3, AXL and MER) family of receptor tyrosine kinase (RTK). The TAM family is different from other RTKs by a conserved sequence within the kinase domain and adhesion molecule-like domain in the extracellular region. The structure of TAM family has two immunoglobin-like domains, two fibronectin type III domains at the extracellular region, one transmembrane domain and one kinase domain at the cytoplasmic tail. The expressions of TAM receptors are broadly found in cells of the mature immune, nervous, reproductive and vascular systems.7

The ligands of TAM family are growth arrest-specific 6 (Gas6) and protein S, which share structural homology with each other.8 When ligand binds to TAM receptors, it induces typical activation of the receptors. The TAM receptors become homodimer or heterodimer and subsequently autophosphorylate the tyrosine residue within the kinase domain.9 Besides, the TAM can go through the ligand-independent activation pathway when expressing at high levels.10 The activation of TAM initiates a cascade of signal transduction processes, resulting in a variety of cellular responses and activities. In contrast, aberrant expression of TAM members, especially AXL, is associated with cancers, including myeloid leukemia, uterine leiomyoma, gastric cancer and ovarian cancers.11, 12, 13, 14, 15, 16 Studies show that AXL is an oncogene, which enhances cell survival, promotes metastasis and is important for epithelial–mesenchymal transition (EMT) in breast cancer.16, 17, 18, 19 Opposite to the well-studied AXL, the function of TYRO3 is poorly understood.

TYRO3 was first identified in 1993 and was found to regulate embryonic differentiation.20 Expression level of TYRO3 was detected in the ovary, testis, nervous system, breast, lung, kidney, osteoclasts and retina as well as a number of hematopoietic cells, including monocytes, macrophages and platelets.21 Besides the above-mentioned functions in immune, central nervous and reproductive systems, the pathological roles of TYRO3 is largely unknown. Recently, some evidence suggests that TYRO3 is a potential oncogene. Aberrant expression of TYRO3 has been found in leukemia, melanoma and thyroid cancer.22, 23, 24 However, the detail mechanisms of how TYRO3 regulates malignant phenotype and whether it plays any role in colon cancer pathogenesis are not clear. Herein, we investigated the functions of TYRO3 in tumorigenesis, cancer progression and metastasis. Our findings demonstrate that TYRO3 is an oncogene and is a druggable molecular target in colorectal cancer.


TYRO3 is overexpressed in polyps and colon cancer

To investigate the pathological roles of TYRO3 in colon cancer, we examined the expression levels of TYRO3 by immunohistochemical (IHC) staining in clinical biopsies, including 78 polyps and 265 pairs of normal and cancer samples. Results demonstrated that levels of TYRO3 were low or undetectable in normal colon tissues but highly expressed mainly in the cytoplasm and plasma membrane of polyps and cancer cells (Figures 1a and b). Clinicopathological analysis revealed that levels of TYRO3 were positively associated with tumor status and stages of disease (Supplementary Table S1). Further analysis revealed that patients with a high TYRO3 level had poor prognosis (Figure 1c).

Figure 1

TYRO3 is overexpressed in colorectal cancer. (a) Representative immunohistochemical staining for TYRO3 in polyp and paired normal and cancerous colon tissue. (b) Accumulated percentage of TYRO3 staining intensity and percentage in normal (N) and tumor (T) tissues. N=11, 107, 107 and 40 for Duke’s stage A to D, respectively. (c) Kaplan–Meier plot of the colon cancer patients with low TYRO3 expression (staining intensity equal to and lower than 2, N=82) and high TYRO3 expression (staining intensity greater than 2, N=158). (d) Schematic drawing indicates the experimental protocol to create AOM/DSS-induced colon cancer in mouse. Arrow indicates point of AOM injection and arrowheads indicate points of tissue collection. DSS was given to mice in weeks 1, 4 and 7. (e) Representative IHC images show the immunoreactivity of Tyro3 in AOM/DSS-treated mouse colon. Control tissue was collected from age-matched saline-injected mice. Normal rabbit IgG was used for negative control.

To investigate the temporal expression pattern of TYRO3 during colon cancer development, we used the azoxymethane/dextran sodium sulfate (AOM/DSS) protocol to induce colon cancer in a mouse model (Figure 1d). At 1 week after AOM/DSS treatment, mucosal epithelia were damaged due to severe inflammation. Regeneration of luminal epithelia was observed at 2 weeks after AOM/DSS treatment and aberrant crypt focus was evident at 3 weeks after AOM/DSS treatment. Adenoma and adenocarcinoma were easily identified after 1 month of AOM/DSS treatment (Figure 1e). Expression of Tyro3 was detected at 2 weeks after AOM/DSS treatment and markedly elevated in aberrant crypt focus and cancer lesions (Figure 1e). Taken together, these data demonstrate that aberrant expression of TYRO3 occurs at very early stage of colon cancer development likely due to inflammation.

TYRO3 induces cell transformation of normal cells

Given that TYRO3 was overexpressed in polyps, we then tested whether forced expression of TYRO3 would transform normal cells into cancer-like cells. For this purpose, we overexpressed TYRO3 in NIH3T3 cells and found that forced expression of TYRO3 in NIH3T3 cells significantly increased cell proliferation, migration and anchorage-independent cell growth (Figures 2a–d). When injected into the tail vein of mice, NIH3T3 cells with TYRO3 overexpression established many tumor nodules in the lung and some in the liver (Figures 2e and f). In contrast, NIH3T3 cell expressing GFP only (as a control) failed to grow in lung or liver (Figure 2f). These data demonstrate that overexpression of TYRO3 promotes cell transformation and tumor development.

Figure 2

Overexpressed TYRO3 induces non-tumorigenic cells transformation. (a) Representative western blots of TYRO3 expression level in NIH3T3 carrying human TYRO3 minigene. (b) Stable clones of NIH3T3 cells carrying TYRO3 gene or GFP vector were cultured for 1, 2, 3, 4 and 5 days and numbers of cells were calculated by trypan blue staining. Data represent mean and s.e.m. from three independent experiments using different batches of cells. **P<0.01. (c) Transwell migration assay was performed for 8 h using cells described in (a). Data represent mean and s.e.m. from three independent experiments using different batches of cells. *P<0.05. (d) Soft agar colony formation assays were performed for 2 weeks. Asterisks indicate significant difference compared with GFP alone group. Data represent mean and s.e.m. from three independent experiments using different batches of cells. **P<0.01. (e) Representative pictures show the lung of mice injected with NIH3T3 cells carrying stable TYRO3 and GFP genes, respectively. (f) Representative pictures show the hematoxylin and eosin staining of lung and liver of mice injected with NIH3T3 cells carrying stable TYRO3 or GFP gene for 14 days. * indicates the tumor in lung and liver.

We then asked whether overexpression of TYRO3 in cancer cell would promote tumor malignancy. For this purpose, we evaluated the endogenous level of TYRO3 in several colon cancer cell lines. We chose HCT116 and HT29 for subsequent experiments since the endogenous level of TYRO3 is elevated but not enormously high (Supplementary Figure S1A), which enables us to perform both knockdown and overexpression experiments. First, we performed overexpression experiment using HT29 cells. Induction of TYRO3 expression by addition of doxycycline significantly increased cell proliferation, migration and invasion abilities (Supplementary Figures S1B–E). These data demonstrate that overexpression of TYRO3 in cancer cells can further enhance cancer malignancy.

TYRO3 promotes cancer progression

We next employed the loss-of-function approach to investigate whether TYRO3 is necessary for tumor progression. For this purpose, we knocked down TYRO3 in HCT116 colon cancer cells (Figure 3a). Knockdown of TYRO3 inhibited cell migration, invasion and colony formation (Figures 3b–d). Similarly, knockdown of TYRO3 in HT29 cells also inhibited cell migration (Supplementary Figures S1F and G). When inoculated into NOD/SCID mice subcutaneously, tumor derived from TYRO3-knockdown cells grew significantly slower than that derived from control cells (Figures 3e and f). Notably, 4 out of 12 mice injected with TYRO3-knockdown cancer cells failed to develop tumor, indicating the important role of TYRO3 in promoting tumor growth. When injected into NOD/SCID mice through tail vein, the tumor lesions in lung were significantly reduced in TYRO3-knockdown cancer cells (Figures 3g and h).

Figure 3

Loss of function of TYRO3 inhibits motility and tumorigenesis. (a) Representative western blots show levels of TYRO3 and β-actin in control and shTYRO3-knockdown cells. shTYRO3A–E represents five different shRNAs targeting different areas of TYRO3. (bd) Migration (b), invasion (c) and colony formation (d), abilities of HCT116 cells with TYRO3-knockdown (shTYRO3#B and shTYRO3#E) and control (shLuciferase) cells. Data represent mean and s.e.m. from three independent experiments using different batches of cells. (e, f) Growth curves (e) and weight of tumors (f) in xenografted mice inoculated with TYRO3-knockdown (shTYRO3#B and shTYRO3#E) or control (shLuciferase) cells. There were six mice per treatment group. (g) The number of metastatic colonies in lung of mice injected with TYRO3-knockdown (shTYRO3#B and shTYRO3#E) or control (shLuciferase) cells. There are six mice per group. Ten fields per mouse were examined. (h) Representative pictures show the hematoxylin and eosin staining of lung and liver of mice inoculated with control or TYRO3-knockdown HCT116 cells for 14 days. Scale bar indicated 200 μm. *P<0.05, **P<0.01, ***P<0.001.

TYRO3 promotes EMT process through SNAI1

To investigate the mechanism responsible for TYRO3-induced tumor progression, we performed meta-analysis of a microarray data set (E-MTAB-990), which collects gene expression information from 688 colon cancer patients. We found that the expression of TYRO3 positively correlates with mesenchymal markers including SNAI1, N-cadherin and fibronectin (mesenchymal markers), but negatively correlates with epithelial markers like E-cadherin and Occludin (Figure 4a). To validate the correlation of TYRO3 and EMT process in clinical samples, we performed IHC staining of SNAI1, a master regulator of EMT, to evaluate its expression level in colorectal cancer samples. The result showed that colorectal cancer had high SNAI1 expression (Figure 4b), which was positively correlated with levels of TYRO3. Consistent with this notion, the Snai1 expression level was markedly elevated in AOM/DSS-induced murine colon cancer cells (Figure 4c). In contrast, the epithelial marker, E-cadherin, was reduced in AOM/DSS-induced mouse tumors (Figure 4d). Next, levels of epithelial and mesenchymal markers were detected. Results demonstrated that overexpression of TYRO3 reduced the overall levels of epithelial markers such as β-catenin, E-cadherin and ZO-1 (Figure 4e and Supplementary Figure S2A) but significantly upregulated mesenchymal markers such as N-cadherin and α-smooth muscle actin (Figure 4e and Supplementary Figure S2A) in HCT116 cells. Similar results were also observed in HT29 cells (Supplementary Figure S2B). Immunofluorescent staining showed that the expression of β-catenin, E-cadherin and ZO-1 was reduced in the cell–cell junction when TYRO3 was overexpressed (Figure 4f). In contrast, the mesenchymal marker, N-cadherin was upregulated by forced expression of TYRO3 (Figure 4f).

Figure 4

TYRO3 positively correlates with EMT process. (a) Meta-analysis of TYRO3 expression levels with EMT-associated genes in 688 colorectal cancer samples. (b) Representative images show IHC staining for TYRO3 and SNAI1 in paired normal (N) and cancerous (T) colon tissues. The N and Ca in No. 78T indicate normal part and cancer part in the same slide, respectively. (c, d) Representative images show IHC staining of SNAI1 (c) and E-cadherin (d) in AOM/DSS-treated colon tissues. (e) Representative western blots show levels of N-cadherin (N-cad), α-smooth muscle actin (α-SMA), E-cadherin (E-cad), ZO-1, β-catenin and β-actin in control and TYRO3-overexpressed HCT116 cells. (f) Confocal photomicrographs showing immunofluorescence stained β-catenin (left three panels), E-cadherin (second left central three panels), ZO-1 (second right three panels) and N-cadherin (right three panels) in control (con) and TYRO3-overexpressed (oeTYRO3) HCT116 cells. Scale bar: 10 μm.

We then performed gain-of-function and loss-of-function experiments, which showed levels of SNAI1 mRNA and protein were upregulated in TYRO3-overexpressed cells and downregulated in TYRO3-knocked down cells (Figures 5a and b and Supplementary Figures S2B–D), suggesting its expression is regulated at the transcriptional level. Since SNAI1 needs to translocate to nucleus in order to perform its biological function, we then checked the level of nuclear SNAI1. Western blot data showed that the level of nuclear SNAI1 was markedly upregulated in the TYRO3-overexpressed cells but downregulated in TYRO3-knocked down cells (Figure 5b). Given SNAI1 was regulated by TYRO3, we then asked whether TYRO3-induced EMT was mediated by upregulation of SNAI1. To probe this question, we used shRNA to knockdown SNAI1 in TYRO3-overexpressed cells. Western blotting and immunofluorescence staining demonstrated that knockdown of SNAI1 abolished TYRO3-induced N-cadherin expression and prevented TYRO3-mediated downregulation of β-catenin, E-cadherin and ZO-1 (Figures 5c and d), indicating that knockdown of SNAI1 prevents TYRO3-induced EMT. Indeed, knockdown of SNAI1 abolished TYRO3-induced cell migration and invasion (Figures 5e and f). Taken together, these data demonstrate that TYRO3-induced EMT process is mediated by upregulation of SNAI1.

Figure 5

Knockdown of SNAI1 reverses TYRO3-induced EMT. (a) Levels of SNAI1 mRNA in TYRO3-overexpressed (left panel) and knockdown (right panel) HCT116 cell. Data represent mean and s.e.m. from four independent experiments using different batches of cells. **P<0.01; ***P<0.001 by t-test. (b) Representative western blots show levels of TYRO3, SNAI1 and β-actin in control, TYRO3-overexpressed and TYRO3-knockdown HCT116 cells (upper panel). Lower panel shows levels of SNAI1 and Lamin A/C in the nuclear fraction of same cells. WCL, whole-cell lysate, NF, nuclear fraction. (c) Representative western blots show levels of TYRO3, β-caternin, E-cadherin, SNAI1, N-cadherin, ZO-1 and β-actin in control and TYRO3-overexpressed HCT116 cells with or without SNAI1 knockdown. (d) Confocal photomicrographs show HCT116 cells stained for β-catenin (left three panels), E-cadherin and (second left three panels), ZO-1 (second right three panels) and N-cadherin (right three panels) in control and TYRO3-overexpressed HCT116 cells with or without SNAI1 knockdown. (e, f) Migration (e) and invasion (f) abilities of control and TYRO3-overexpressed (TYRO3) cells with or without SNAI1 knockdown. Data represent mean and s.e.m. from three independent experiments using different batches of cells. **P<0.01; ***P<0.001 compared with control.

Blocking TYRO3 signaling by human antibody abolishes TYRO3-induced EMT process and increases drug sensitivity

Since TYRO3 induced tumor growth and metastasis through EMT process, we thought to develop a strategy to reverse the EMT process by blocking TYRO3. To this end, we generated human anti-TYRO3 antibody (TYRO3-hIgG) and used it for a series of studies. Results showed that treatment with TYRO3-hIgG inhibited cell migration and invasion abilities (Figure 6a), and reversed TYRO3-mediated EMT marker expression (Figures 6b–f). These data suggest that TYRO3-hIgG treatment reverses TYRO3-induced EMT.

Figure 6

Human anti-TYRO3 antibody inhibits TYRO3-induced EMT and tumor growth. (a) Migration (upper panel) and invasion (lower panel) abilities of cells carrying control or TYRO3 genes with or without TYRO3 hIgG treatment. Data represent mean and s.e.m. from three independent experiments using different batches of cells. (be) Confocal photomicrographs show immunofluorescence staining of β-catenin (b), E-cadherin (c), ZO-1 (d) and N-cadherin (e) in HCT116 cells carrying control (pcDNA) or TYRO3 (oeTYRO3) genes and treated with control IgG (con hIgG) or human anti-TYRO3 IgG (TYRO3-hIgG). (f) Representative western blots show levels of TYO3, SNAI1, N-cadherin, α-SMA, E-cadherin, β-catenin, ZO-1 and β-actin in control or TYRO3-overexpressed cells treated with control IgG (Con-hIgG) or human anti-TYRO3 IgG (TYRO3-hIgG). (g) Percentage of apoptotic cells in HCT116 cells treated with or without paclitaxel (left panel), oxaliplatin (central panel) and 5FU (right panel) in the presence (TYRO3-hIgG) or absence (con hIgG) of TYRO3 hIgG. Data represent mean and s.e.m. from three independent experiments using different batches of cells. Asterisks indicate comparison between con hIgG and TYRO3 hIgG and pound signs indicate comparison between drug and vehicle groups. # and *: P<0.05, ## and **: P<0.01, ### and ***: P<0.001. (h, i) Growth curves (h), representative image (i, left panel) and quantification data (i, right panel) of xenografted tumors in mice inoculated with HCT116 cells (n=6 mice per group) treated with/without 5FU and control or TYRO3 hIgG, respectively. *P<0.05, **P<0.01, ***P<0.001.

We next asked whether anti-TYRO3 antibody could be used as a therapeutic agent. The cells were treated with TYRO3-hIgG in the presence or absence of anticancer drugs, 5-FU, paclitaxel and oxaliplatin. Results showed that treatment with TYRO3-hIgG enhanced cell apoptosis by two-folds (Figure 6g). More importantly, TYRO3-hIgG markedly increased drug sensitivity (Figure 6g). These in vitro results were replicated by in vivo xenograft mouse experiment (Figures 6h and i). Finally, we tested the cell-killing and drug-enhancing effects of TYRO3-hIgG using colon cancer cells isolated from patients. Results showed that some primary cancer cells were resistant to anticancer drugs (such as Patient #4), but all responded to TYRO3 IgG (Figures 7a–c). Again, co-treatment with TYRO3 IgG and anticancer drugs significantly induced cell apoptosis in all four patents. Taken together, these data indicated that human anti-TYRO3 antibody can be used as a potent adjuvant for anticancer therapy.

Figure 7

Human TYRO3 antibody enhances drug sensitivity in primary colon cancer cells. Percentage of apoptotic cells in primary colon cancer cells treated with vehicle, paclitaxel (a), oxaliplactin (b) or 5FU (c) in the presence (TYRO3 hIgG) or absence (con hIgG) of TYRO3 hIgG for 72 h. The number of apoptotic cells were determined by flow cytometry as described in Materials and methods. Each experiment was repeated three times using different batches of primary colon cancer cells isolated from the same patient. Primary colon cancer cells isolated from four different patients were labeled as pt#1–4, respectively. Asterisks indicate comparison between con hIgG and TYRO3 hIgG and pound signs indicate comparison between drug and vehicle groups. # and *: P<0.05, ## and **: P<0.01, ### and ***: P<0.001.


Despite intensive efforts on cancer research and treatment improvement, 50% of patients with colon cancer ultimately die from this disease. One of the important reasons for the poor performance of therapeutic treatments is the heterogeneity of cancer etiology. Thus, expanding our understanding of the mechanism of colon cancer development shall provide important information for cancer therapy. In this study, we identified a novel oncogene, TYRO3, that contributes to colon cancer development and progression. We demonstrated that TYRO3 is highly expressed in polyp and colon cancer cells. The expression level positively correlates with disease stages and negatively correlates with overall survival. Results from gain-of-function and loss-of-function experiments demonstrate that TYRO3 promotes cancer cell proliferation, migration, invasion and EMT process. More importantly, we raised human monoclonal antibody against TYRO3 and provided compelling evidence to demonstrate that this antibody reverses TYRO3-induced EMT and effectively enhances the treatment efficacy of currently used first-line drugs. Our result not only is the first report to demonstrate that TYRO3 is a novel oncogene for colon cancer but also provides an alternative approach for designing new therapeutic regimens.

RTKs have been implicated in malignancy of many cancers; however, most studies focused on classical RTK families such as EGFR and VEGFR. By using the bioinformatics algorithm to predict potential oncogenes and tumor suppressor genes, we found that TYRO3, a member of the unique RTK family, possesses several signature motifs resembling previously identified oncogenes6 and thus hypothesized that it may play critical roles in colon cancer malignancy. Indeed, IHC staining of clinical biopsies demonstrated that levels of TYRO3 were markedly increased in most polyps and cancer samples compared with the normal counterparts. Further analysis revealed that the level of TYRO3 is associated with cancer malignancy and poor prognosis. These data provide important information to support our hypothesis and serve as a solid foundation for further investigation of the pathological roles of TYRO3 in colon cancer tumorigenesis, progression and malignancy.

The entire TAM family member was originally cloned from cancer cells, which suggests they all possess oncogenic potential. Surprisingly, no activating mutation of human TAM family member was found to be associated with development of cancer. This notion indicates that the oncogenic potential of TAM receptors is related to aberrant regulation of the same cellular processes in which these receptors normally play a role. TAM receptor signaling pathways have been linked to regulation of the cytoskeleton organization, cell–cell contact and immune modulation.25, 26 In this study, we identified that overexpression of TYRO3 in NIH3T3 cells enhances cell proliferation, migration, anchorage-independent growth and metastatic ability while knockdown of TYRO3 in colon cancer cells inhibits all these abilities. Furthermore, forced expression of TYRO3 in colon cancer cell reduces β-catenin, E-cadherin and ZO-1 in cell membrane, indicating the promotion of EMT process. Taken together, these lines of data demonstrate that aberrant expression of TYRO3 drives normal cells into cancer-like cells while reduction of TYRO3 in cancer cells reverses cancer characteristics. More importantly, all these alterations are mediated via the cellular functions normally regulated by TYRO3, demonstrating that the disease status and malignancy primarily depends on the level of TYRO3 and subsequently its downstream signaling.

By taking the advantage of public database, we re-analyzed the microarray data set established by Budinska et al.27 and found that the level of TYRO3 was positively correlated with SNAI1, a master regulator of EMT. We then tested whether overexpression of TYRO3-induced EMT is mediated by SNAI1. Immunostaining of clinical samples revealed that the level of SNAI1 was positively correlated with that of TYRO3. Cellular experiments demonstrated that expression of SNAI1 in colon cancer cells was controlled by TYRO3. We then employed the AOM/DSS mouse model to recapitulate the developmental sequence of colon cancer and found the expression of SNAI1 was temporally correlated with that of TYRO3. Functionally, we demonstrated that knockdown of SNAI1 abolished TYRO3-induced cell migration, invasion and loss of epithelial markers. These lines of data demonstrate that overexpression of TYRO3-induced tumor malignancy is likely mediated by SNAI1.

Our data provide evidence that aberrant expression of TYRO3 plays critical roles in colon cancer development and progression. As a simple thumb-of-rule that if cancer cells expressed elevated levels of TYRO3, they should be addicted to TYRO3 signaling-mediated cellular functions and, thus, blockage of this signaling pathway should be able to ameliorate cancer malignancy. Since there is no TYRO3 inhibitor available, we raised human monoclonal antibodies against TYRO3. Monoclonal antibodies are wildly used in cancer immunotherapy. For example, Trastuzumab (anti-Her-2/neu), Gemtuzumab (anti-CD33), Cetuximab (anti-Her-1) and Bevacizumab (anti-VGFR) had been used in clinical therapy.28, 29, 30 Our results showed that treatment with human anti-TYRO3 antibody repressed cell migration and invasion, increased cell apoptosis and enhanced tumor-killing effect of anticancer drugs. Since human antibodies are much less immunogenic than murine monoclonal antibodies, this anti-TYRO3 hIgG may have great potential in future clinical application.

In conclusion, we demonstrated that TYRO3 is overexpressed in the early stage of colon cancer development and aberrant expression of TYRO3 promotes tumorigenesis and induces EMT through the regulation of SNAI1. Blocking TYRO3 signaling by human anti-TYRO3 antibody ameliorates cancer malignancy and increased sensitivity to anticancer drugs. These lines of evidence demonstrate that TYRO3 plays a critical role in colon cancer development and is a druggable target in cancer therapy.

Materials and methods

Antibodies, cell lines and plasmids

Antibodies used in this study include anti-TYRO3 antibody (Cell Signaling Technology, Danvers, MA, USA, #5585), β-catenin (Cell Signaling Technology, #8480), E-cadherin (Cell Signaling Technology, #3195), ZO-1 (Cell Signaling Technology, #8193), SNAI1 (Cell Signaling Technology, #3879) and N-cadherin (Cell Signaling Technology, # 13116). The cells (HCT116, HT29 and NIH3T3) were purchased from ATCC (Manassas, VA, USA) and maintained in a 37oC incubator with 5% CO2. HCT116 and HT29 were cultured in McCoy’s 5A (Invitrogen, Grand Island, NY, USA) supplemented with 10% fetal bovine serum while NIH3T3 was cultured in minimum essential medium supplemented with 10% fetal bovine serum. All cell lines were tested for free of mycoplasma contamination and were authenticated by Center of Genomic Medicine at the National Cheng Kung University (HCT116, NIH3T3) or Mission Biotech (HT29). The human TYRO3 expression vector contains full-length cDNA (BC051756; GenDiscovery Biotenology Inc., Taipei, Taiwan, ROC) were purchased from ORIGENE (Rockville, MD, USA) and further ligated to pcDNA5/To vector (Invitrogen). All shRNAs were also purchased from the National RNAi Core Facility Platform (Taipei, Taiwan, ROC).

Clinical samples

The normal colon tissues and tumor specimens (n=265) were obtained from patients with colon cancer who underwent surgery at the Department of Surgery at the National Cheng Kung University Hospital. Specimens of tumor tissue and adjacent tissue at 5 cm away from the tumor site were collected. Polyps (n=78) were collected from patients who underwent colonoscopic surgery. The postsurgical stages of each tumor and polyp were classified and histologically confirmed by pathologists. This study was approved by Institutional Reviewing Board of National Cheng Kung University Hospital and informed consent was obtained from each patient.

Immunohistochemical staining

The procedures for IHC were described previously.31 In brief, sections were incubated with anti-TYRO3 antibody (Sigma-Aldrich, St Louis, MO, USA) at 1:100 dilutions at 4oC overnight. After washing with phosphate-buffered saline (PBS), incubating with horseradish peroxidase-conjugated goat anti-rabbit IgG (Leica Biosystems, Newcastle Upon Tyne, UK), color was developed by AEC kit (Bio-SB Inc., Santa Barbara, CA, USA). Ten high-power fields were examined per section. Score of 0 to 3 were assigned to samples with staining intensity with 0: negative; 1: weak staining; 2: medium staining; 3: strong staining. The positive percentage was also classified to: 0: no positive cells; 1: less than 30% positive cells; 2: 30–70% positive cells; 3: more than 70% positive cells. The sum of scores from intensity and percentage were further grouped into 0, 2, 4 and 6.

Create TYRO3 expression and knockdown stable clones

Full-length cDNA of TYRO3 was inserted into pcDNA5/TO vector (Invitrogen), which contain hygromycin B resistant sequence. TYRO3-containing and vector only plasmids were transfected into HCT116 cells, respectively. After transfection, the medium was replaced by fresh medium containing hygromycin B (50 μg/ml) to select resistant cells. The selective medium was changed every 3–4 days until hygromycin-resistant colonies were established.

To establish TYRO3-knockdown clones, shTYRO3 viral particles were generated by co-transfecting human embryonic kidney 293T cells with pAS4.1w.Ppuro-aOn-TYRO3 plasmids, PAX2 and pMD2 using lipofectamine 2000 following the manufacturer’s protocol. After 48 h, the supernatants were collected, filtered and concentrated by ultracentrifugation (100 000 × g) in a Beckman ultracentrifuge. The viral particles were used to infect HCT116 cells and infected cells were incubated in selection media containing puromycin for 1 month to obtain TYRO3-knockdown stable clones. Sequences of shTYRO3 are listed in Supplementary Table S2.

Colony formation assay

Cells (10 000 cells/well) suspended in 0.3% agar/media were plated into each well of a six-well plate precoated with 0.6% agar/media. Cells were cultured at 37 °C for 2 weeks until colonies became visible. The colonies were visualized and counted by 0.05% crystal violet staining.

Migration and invasion assays

The 1.5 × 104 cells were plated into the upper area of the chamber of transwell (Millipore, Billerica, MA, USA) in serum-free medium. For invasion assay, the transwells were precoated with 5% Matrigel before plating. The bottom wells were filled with medium supplemented with 10% fetal bovine serum. After 8 (for migration assay) and 24 h (for invasion assay) incubation, non-migrated cells were removed from the upper chamber and migrated cells were quantified by light microscopy after staining with crystal violet. Ten different fields were taken by the image system. The total cell number in 10 fields was counted to represent the migrated cell number.

Tumor growth and metastasis assay

In tumor growth assay, 5 × 105 NIH3T3 cells in a volume of 100 μl were injected into each BALB/C mouse subcutaneously. Tumor sizes were measured once every 4 days. The volume was calculated as length × width × height. In metastatic assay, 1 × 106 cells in a volume of 100 μl were injected into the tail vein of each BALB/C mouse. The health of mice was monitored every other days. Mice were killed once they showed signs of sickness and lung and liver were collected. The metastatic colonies were determined by hematoxylin and eosin staining. All animal study procedures were performed under protocols approved by the Institutional Animal Care and Use Committee of the National Cheng Kung University and no blindness design was applied.

Confocal microscopy

HCT116 cells were plated on μ-slides VI0.4 (ibidi, Munich, Germany) and transiently transfected with pcDNA5/to or TYRO3 for 72 h and then fixed in 4% paraformaldehyde. After washing, cells were treated with 0.5% Triton X-100 and then washed again in 1 × PBS. After blocking by Superblock buffer (Thermo Fisher Scientific Inc., Rockford, IL, USA), cells were incubated with antibodies against TYRO3 (1:100), β-catenin (1:100), E-cadherin (1:100) or ZO-1 (1:100), respectively, at 4 °C overnight. After washing in PBST for three times, cells were incubated with secondary antibody (Alexa Fluor 594–goat anti-rabbit; Invitrogen) and Hoechst (1:1000) for 1 h at room temperature and then washed again in PBST five times. Finally, the cells were observed on a scanning confocal laser microscope (OLYMPUS FV1000, Center Valley, PA, USA).

Isolation of phages binding to TYRO3 from a phage-displayed scFv library

A human naïve phage-displayed scFv library with 2 × 1010 complexity previously established in our laboratory32 was used for selection. The scFv library was subtracted nonspecific binding in protein G Dynabeads (Invitrogen), and subsequently incubated with TYRO3-Fc recombinant protein (R&D Systems, Minneapolis, MN, USA) immobilized Dynabeads. After washing with PBS containing 0.1% Tween 20, the phages that bound to TYRO3-Fc were recovered by infection with E. coli TG1 cells. Some of the infected cells were serial diluted to determine titer, and the others were rescued by the M13KO7 phage. After determination of rescued phages titer, the next round of biopanning was performed. In the fourth and fifth round of biopanning the phage clones were randomly selected to culture for enzyme-linked immunosorbent assay screening. Positive phage clones against TYRO3 were screened and identified by enzyme-linked immunosorbent assay.

Construction and expression of anti-TYRO3 human antibody

The VH region of anti-TYRO3 human antibody was cloned separately into expression vector pcDNA5-FRT-Gamma1 with a signal peptide and human IgG1 constant region, using AgeI and NheI sites. In addition, the VL region of anti-TYRO3 human antibody was separately cloned into modified expression vector p-Kappa-HuGs, using AgeI and EcoRV sites. Both heavy- and light-chain gene-containing plasmids were co-transfected into FreeStyle 293 cells. Cultured media of stable clones was collected, centrifuged and filtered through a 0.45 μm membrane. The supernatant was then subjected to protein G column chromatography (GE Healthcare, Pittsburgh, PA, USA) for purification of anti-TYRO3 human IgG. After dialysis of eluents with PBS, the concentration of antibody was assessed using Bradford reagent (Thermo Fisher Scientific Inc.) and spectrophotometry.

Apoptosis assay

Cells were scraped from a culture dish, and centrifuged 400 g for 10 min. The cell pellet were then washed by flow buffer (1% fetal bovine serum/PBS) and then stained with 7-AAD and Annexin-V for 20 min (559763; BD Pharmingen, BD Biosciences, San Jose, CA, USA). The stained cells were then diluted by binding buffer and analyzed by flow cytometry (FACS Canto II).

AOM/DSS-induced colon carcinogenesis

On day 1, the male B6 mice (6–8 weeks old) in the experimental group (n=6 mice/time point) were injected with AOM working solution (10 mg/kg body weight) while those in the control group (n=6 mice/time point) received sterile isotonic saline injection. The DSS (MP Biomedicals, Santa Ana, CA, USA)-containing water (2%) was given at weeks 1, 4 and 7 to mice in the AOM group. Mice were killed at 1, 2, 3, 4, 8 and 12 weeks after AOM injection.

In vivo drug sensitivity test

HCT116 cells (5 × 105) suspended in 100 μl of serum-free medium with 20% Matrigel were inoculated subcutaneously in the hind flank of male NOD/SCID mice (6–8 weeks old). For the anticancer drug and antibody treatment experiment, the mice were randomly assigned into experimental or control groups (n=6 mice/group) once the tumor reached 30 mm3. Mice in different groups received control human IgG (10 μg/ml), TYRO3-hIgG (10 μg/ml), 5-FU (25 mg/kg body weight) and 5-FU+TYRO3-hIgG, respectively. Antibodies were given by subcutaneous injection around the tumor region on days 3 and 6 of a 7-day cycle for three cycles. Tumors were measured with calipers every 3 days during the experimental period.

Statistical analysis

The correlation of IHC intensity and percentage with clinical prognostic factors were calculated and analyzed by χ2. The numeric data were expressed as mean±standard error of the mean. Two-tailed Student’s t-test was used if there were only two groups, while one-way analysis of variance followed by post-test was applied if there were three or more groups for comparison. Statistical analysis was performed by using commercial software, GraphPad Prism 4.02 (GraphPad Software, La Jolla, CA, USA). Assumptions of normal distribution and equal variance were examined for each statistical analysis when applicable.


  1. 1

    Golan T, Urban D, Berger R, Lawrence YR . Changing prognosis of metastatic colorectal adenocarcinoma: differential improvement by age and tumor location. Cancer 2013; 119: 3084–3091.

    Article  Google Scholar 

  2. 2

    Fearon ER, Vogelstein B . A genetic model for colorectal tumorigenesis. Cell 1990; 61: 759–767.

    CAS  Article  Google Scholar 

  3. 3

    Groden J, Thliveris A, Samowitz W, Carlson M, Gelbert L, Albertsen H et al. Identification and characterization of the familial adenomatous polyposis coli gene. Cell 1991; 66: 589–600.

    CAS  Article  Google Scholar 

  4. 4

    Takayama T, Katsuki S, Takahashi Y, Ohi M, Nojiri S, Sakamaki S et al. Aberrant crypt foci of the colon as precursors of adenoma and cancer. N Engl J Med 1998; 339: 1277–1284.

    CAS  Article  Google Scholar 

  5. 5

    Petitjean A, Achatz MI, Borresen-Dale AL, Hainaut P, Olivier M . TP53 mutations in human cancers: functional selection and impact on cancer prognosis and outcomes. Oncogene 2007; 26: 2157–2165.

    CAS  Article  Google Scholar 

  6. 6

    Chan HH, Tsai SJ, Sun HS . TAG: a comprehensive database for cancer research. Nucleic Acids Res 2007, Molecular Biology Database Collection: 922.

  7. 7

    Lemke G, Rothlin CV . Immunobiology of the TAM receptors. Nat Rev Immunol 2008; 8: 327–336.

    CAS  Article  Google Scholar 

  8. 8

    Stitt TN, Conn G, Gore M, Lai C, Bruno J, Radziejewski C et al. The anticoagulation factor protein S and its relative, Gas6, are ligands for the Tyro 3/Axl family of receptor tyrosine kinases. Cell 1995; 80: 661–670.

    CAS  Article  Google Scholar 

  9. 9

    Schlessinger J . Cell signaling by receptor tyrosine kinases. Cell 2000; 103: 211–225.

    CAS  Article  Google Scholar 

  10. 10

    Taylor IC, Roy S, Varmus HE . Overexpression of the Sky receptor tyrosine kinase at the cell surface or in the cytoplasm results in ligand-independent activation. Oncogene 1995; 11: 2619–2626.

    CAS  PubMed  PubMed Central  Google Scholar 

  11. 11

    Crosier KE, Crosier PS . New insights into the control of cell growth; the role of the AxI family. Pathology 1997; 29: 131–135.

    CAS  Article  Google Scholar 

  12. 12

    Neubauer A, Burchert A, Maiwald C, Gruss HJ, Serke S, Huhn D et al. Recent progress on the role of Axl, a receptor tyrosine kinase, in malignant transformation of myeloid leukemias. Leuk Lymphoma 1997; 25: 91–96.

    CAS  Article  Google Scholar 

  13. 13

    O'Bryan JP, Frye RA, Cogswell PC, Neubauer A, Kitch B, Prokop C et al. axl, a transforming gene isolated from primary human myeloid leukemia cells, encodes a novel receptor tyrosine kinase. Mol Cell Biol 1991; 11: 5016–5031.

    CAS  Article  Google Scholar 

  14. 14

    Sun WS, Fujimoto J, Tamaya T . Clinical implications of coexpression of growth arrest-specific gene 6 and receptor tyrosine kinases Axl and Sky in human uterine leiomyoma. Mol Hum Reprod 2003; 9: 701–707.

    CAS  Article  Google Scholar 

  15. 15

    Sawabu T, Seno H, Kawashima T, Fukuda A, Uenoyama Y, Kawada M et al. Growth arrest-specific gene 6 and Axl signaling enhances gastric cancer cell survival via Akt pathway. Mol Carcinog 2007; 46: 155–164.

    CAS  Article  Google Scholar 

  16. 16

    Sun W, Fujimoto J, Tamaya T . Coexpression of Gas6/Axl in human ovarian cancers. Oncology 2004; 66: 450–457.

    CAS  Article  Google Scholar 

  17. 17

    Gjerdrum C, Tiron C, Hoiby T, Stefansson I, Haugen H, Sandal T et al. Axl is an essential epithelial-to-mesenchymal transition-induced regulator of breast cancer metastasis and patient survival. Proc Natl Acad Sci USA 2010; 107: 1124–1129.

    CAS  Article  Google Scholar 

  18. 18

    Song X, Wang H, Logsdon CD, Rashid A, Fleming JB, Abbruzzese JL et al. Overexpression of receptor tyrosine kinase Axl promotes tumor cell invasion and survival in pancreatic ductal adenocarcinoma. Cancer 2011; 117: 734–743.

    CAS  Article  Google Scholar 

  19. 19

    Asiedu MK, Beauchamp-Perez FD, Ingle JN, Behrens MD, Radisky DC, Knutson KL . AXL induces epithelial-to-mesenchymal transition and regulates the function of breast cancer stem cells. Oncogene 2014; 33: 1316–1324.

    CAS  Article  Google Scholar 

  20. 20

    Biesecker LG, Gottschalk LR, Emerson SG . Identification of four murine cDNAs encoding putative protein kinases from primitive embryonic stem cells differentiated in vitro. Proc Natl Acad Sci USA 1993; 90: 7044–7048.

    CAS  Article  Google Scholar 

  21. 21

    Linger RM, Keating AK, Earp HS, Graham DK . TAM receptor tyrosine kinases: biologic functions, signaling, and potential therapeutic targeting in human cancer. Adv Cancer Res 2008; 100: 35–83.

    CAS  Article  Google Scholar 

  22. 22

    De Vos J, Couderc G, Tarte K, Jourdan M, Requirand G, Delteil MC et al. Identifying intercellular signaling genes expressed in malignant plasma cells by using complementary DNA arrays. Blood 2001; 98: 771–780.

    CAS  Article  Google Scholar 

  23. 23

    Avilla E, Guarino V, Visciano C, Liotti F, Svelto M, Krishnamoorthy G et al. Activation of TYRO3/AXL tyrosine kinase receptors in thyroid cancer. Cancer Res 2011; 71: 1792–1804.

    CAS  Article  Google Scholar 

  24. 24

    Zhu S, Wurdak H, Wang Y, Galkin A, Tao H, Li J et al. A genomic screen identifies TYRO3 as a MITF regulator in melanoma. Proc Natl Acad Sci USA 2009; 106: 17025–17030.

    CAS  Article  Google Scholar 

  25. 25

    Duijkers FA, Meijerink JP, Pieters R, van Noesel MM . Downregulation of Axl in non-MYCN amplified neuroblastoma cell lines reduces migration. Gene 2013; 521: 62–68.

    CAS  Article  Google Scholar 

  26. 26

    Gerloff J, Korshunov VA . Immune modulation of vascular resident cells by Axl orchestrates carotid intima-media thickening. Am J Pathol 2012; 180: 2134–2143.

    CAS  Article  Google Scholar 

  27. 27

    Budinska E, Popovici V, Tejpar S, D'Ario G, Lapique N, Sikora KO et al. Gene expression patterns unveil a new level of molecular heterogeneity in colorectal cancer. J Pathol 2013; 231: 63–76.

    CAS  Article  Google Scholar 

  28. 28

    Goldenberg MM . Trastuzumab, a recombinant DNA-derived humanized monoclonal antibody, a novel agent for the treatment of metastatic breast cancer. Clin Ther 1999; 21: 309–318.

    CAS  Article  Google Scholar 

  29. 29

    Caron PC, Schwartz MA, Co MS, Queen C, Finn RD, Graham MC et al. Murine and humanized constructs of monoclonal antibody M195 (anti-CD33) for the therapy of acute myelogenous leukemia. Cancer 1994; 73 (Suppl): 1049–1056.

    CAS  Article  Google Scholar 

  30. 30

    Prewett M, Rockwell P, Rockwell RF, Giorgio NA, Mendelsohn J, Scher HI et al. The biologic effects of C225, a chimeric monoclonal antibody to the EGFR, on human prostate carcinoma. J Immunother Emphasis Tumor Immunol 1996; 19: 419–427.

    CAS  Article  Google Scholar 

  31. 31

    Chien C-W, Lin S-C, Lai Y-Y, Lin B-W, Lin S-C, Lee J-C et al. Regulation of CD151 by hypoxia controls cell adhesion and metastasis in colorectal cancer. Clin Cancer Res 2008; 14: 8043–8051.

    CAS  Article  Google Scholar 

  32. 32

    Lu RM, Chang YL, Chen MS, Wu HC . Single chain anti-c-Met antibody conjugated nanoparticles for in vivo tumor-targeted imaging and drug delivery. Biomaterials 2011; 32: 3265–3274.

    CAS  Article  Google Scholar 

Download references


We thank Miss Yi-Hsuan Yeh and Miss Yi-Chen Tang for technical assistance with IHC and animal study. This work was supported by grants from National Science Council of Taiwan (NSC 101-2321-B-006-020 and NSC 102-2321-B-006-011) and by Top University Grant of National Cheng Kung University (grant # D103-35A17).

Author information



Corresponding authors

Correspondence to H S Sun or S-J Tsai.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Author contributions

CWC and PCH performed most experiments; HCW and YLC produced and tested human monoclonal antibody; SCL, BWL, JCL and YJC participated and conducted experiments using clinical samples; HSS and SJT conceived the project; and CWC, HSS and SJT wrote the manuscript. All authors read and approved the final version of manuscript.

Supplementary Information accompanies this paper on the Oncogene website

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Chien, CW., Hou, PC., Wu, HC. et al. Targeting TYRO3 inhibits epithelial–mesenchymal transition and increases drug sensitivity in colon cancer. Oncogene 35, 5872–5881 (2016).

Download citation

Further reading


Quick links