Therapeutic Targeting of Metadherin Suppresses Colorectal and Lung Cancer Progression and Metastasis

Lung cancer is the most common cancer type worldwide. Although the incidence of lung cancer continues to decline since the last decade, it still accounts for one-quarter of all cancer deaths (1). More than 80% of lung cancers are non–small cell lung cancer (NSCLC), of which, adenocarcinoma, squamous cell carcinoma, and large cell carcinoma are the most common subtypes (2). Aided by increasingly sophisticated therapeutic strategies, such as targeted therapies (3) and immunotherapies (4), the clinical outcomes in patients with NSCLC have significantly improved over the past two decades. However, given the molecular heterogeneity of the disease, the response rates of the current therapeutic strategies are usually low (5). Furthermore, treatment resistance is frequently observed (6), which reduces the clinical benefits of these treatments to patients. To overcome these challenges and increase the population that could potentially benefit from targeted therapies, identification of new driver genes that promote lung cancer progression and subsequent development of novel therapeutic strategies are urgently needed (7).

Following lung cancer, colorectal cancer is the second most common cause of cancer death in United States (8). More than 900,000 patients are predicted to die from this disease annually worldwide (9). Even worse, the incidence and mortality were gradually increasing in young patients (age <50) during the last decade (8). Similar to lung and other types of cancer, colorectal cancer is also a heterogeneous disease with multiple genetic and epigenetic aberrations (10). The earliest event that occurs in the development of colorectal cancer is the mutation of Adenomatous polyposis coli (APC) gene, which results in hyperactivation of Wnt signaling (11). About 90% of patients show APC mutations or allelic losses (12). Although Wnt signaling is essential for colorectal cancer progression, therapies that directly target key components in this pathway are not currently feasible given the broad side effects. Therapies that target angiogenesis or EGFR provide marginal durable benefits in combination with chemotherapies. However, the overall response rate of these therapies is limited (13). It is therefore imperative to further explore the mechanisms underlying colorectal cancer progression and metastasis with the goal of developing new and effective therapeutic agents.

Metadherin (MTDH, also called AEG-1) has been found to be involved in the development of many types of cancer, including breast, prostate, liver, lung, and colorectal cancers (14). Extensive studies of MTDH in liver (15, 16), breast (17–19), and prostate (20) cancers established its key role in promoting tumor initiation, progression, metastasis, and treatment resistance. Moreover, no apparent physiologic defects were observed upon Mtdh knockout (KO) in mouse models, making it an attractive therapeutic target for cancer treatment (19). Previous in vitro studies on MTDH in lung and colorectal cancers resulted in contradictory conclusions (21–23). In this study, we thoroughly investigated the role of MTDH in lung and colorectal cancers using robust genetically modified mouse tumor models as well as xenograft/allograft models. In addition to revealing distinct functions of MTDH in promoting colorectal and lung cancers, we developed locked nucleic acid (LNA)-modified MTDH-specific antisense oligonucleotides (ASO), and extensively assessed their anticancer therapeutic potentials in preclinical models.

All procedures involving mice and experimental protocols were approved by the Institutional Animal Care and Use Committees of Princeton University. Mouse strains used in this study are listed in Supplementary Materials and Methods. For allograft and xenograft studies, 8-week immunocompromised NSG or immunocompetent C57BL/6 mice were used. For ASO treatment, ASOs were reconstituted with PBS and injected into mice intraperitoneally at a dose of 20 mg/kg. The treatment schemes are indicated in figures. Additional details and references are provided in the Supplementary Materials and Methods.

Clinical data of colorectal adenocarcinoma and lung adenocarcinoma was downloaded from The Cancer Genome Atlas (TCGA; Firehose Legacy) dataset in cBioportal (RRID:SCR_014555) website. The MTDH expression was determined by Agilent microarray and RNA-seq V2 RSEM in colorectal adenocarcinoma and lung adenocarcinoma, respectively. Patients without disease-free survival (DFS) information or MTDH expression data were excluded. DFS of patients with colorectal adenocarcinoma and lung adenocarcinoma was then stratified by MTDH expression. The cutoffs were chosen automatically by R software with optimal P values.

HEK293T (RRID:CVCL_0063), SW620 (RRID:CVCL_0547), mouse breast cancer cell line 4T1 (RRID:CVCL_0125), mouse renal cell carcinoma cell line Renca (RRID:CVCL_2174), Lewis lung carcinoma cell line LLC1 (RRID:CVCL_4358), and L cells were grown in DMEM supplemented with 10% FBS and Pen/Strep. DLD1 (RRID:CVCL_0248) cells were grown in RPMI1640 supplemented with 10% FBS and Pen/Strep. MC-38 (RRID:CVCL_B288) cells were cultured with DMEM media supplemented with 10% FBS, 2 mmol/L glutamine, 0.1 mmol/L nonessential amino acids, 1 mmol/L sodium pyruvate, 10 mmol/L Hepes, 50 μg/mL gentamycin sulfate, and Pen/Strep. HEK293T, SW620, DLD1, 4T1, Renca, and LLC1 cells were purchased from ATCC. MC-38 cells were purchased from Kerafast. L cells were gift from the Devenport lab at Princeton University. All cell lines were authenticated through STR profiling and tested monthly for Mycoplasma by PCR. Cell lines were not passaged more than 30 times.

For the immunoprecipitation (IP) experiment, plasmids were transfected into HEK293T cells. The cells were then lysed and IP with indicated antibodies. Samples were subjected to Western blotting. For Western blot analysis, protein lysates were resolved with SDS-PAGE gel and immunoblotted with standard protocols. Details and references are provided in the Supplementary Materials and Methods.

3D culture was performed as described previously (24). Details and references are provided in the Supplementary Materials and Methods.

Paraffin-embedded intestine and lung tumor samples were cut at a thickness of 4 μm. The slides were baked overnight at 60°C. Next, the tissue slides were washed with PBS after deparaffinization and hydration and then boiled in citrate buffer at 100°C for 40 minutes. After treatment with 3% H₂O₂ for 30 minutes to block endogenous peroxidase, slides were incubated at 4°C overnight with indicated antibodies. Antibodies and staining details are listed in Supplementary Materials and Methods.

For paraffin-embedded liver samples, slides were prepared similarly as for IHC staining described above. For 3D cultured spheres, the samples were collected from dishes and Matrigel (Corning Matrigel Matrix) was digested as previously reported (25). Antibodies and staining details are listed in Supplementary Materials and Methods.

The infection was performed as described previously (26, 27). Details and references are provided in the Supplementary Materials and Methods.

Dissociated lungs were dissected, and single cell suspensions were prepared as described previously (28). Antibodies and staining details are listed in Supplementary Materials and Methods. Gating strategy is shown in Supplementary Fig. S3D.

Intestines were collected from experimental mice, opened longitudinally and washed with PBS. The intestines were then fixed overnight with 10% formalin at room temperature. Meanwhile, 0.5% Methylene Blue Stain was prepared by dissolving 15 mg Methylene Blue (Sigma-Aldrich, #M9140) in 30 mL of distilled water. Intestines were stained in 0.5% Methylene Blue Stain for 30 seconds. The samples were rinsed with distilled water and examined with a dissecting microscope.

15-, 16-, and 17-mer ASOs were selected on the basis of the sequence of the mouse MTDH gene. LNA-modified nucleotides were inserted into the flanks of ASOs. Main criterion for sequence selection was selectivity to avoid undesired off-target effects. LNA-modified ASOs were ordered from Microsynth or Axolabs. For in vitro testing, ASOs were resuspended in H₂O, for in vivo experiments ASOs were resuspended in PBS. ASOs were added to cells in vitro without the use of a transfection reagent and used for in vivo studies without a delivery system. Sequences of selected MTDH ASOs and control oligonucleotide used in this study are listed in Supplementary Table S1. Control oligonucleotide was derived from a previous study (29).

Target knockdown (KD) efficacy of mouse MTDH-specific LNA-modified ASOs was investigated in 4T1 and Renca cells in vitro. Details and references are provided in the Supplementary Materials and Methods.

HEK293T cells stably transfected with a mouse TLR9/NFκB/Luciferase reporter plasmid were treated with different concentrations of the respective ASO or a positive control (ODN1668, Invivogen) for 24 hours. For detection of luciferase activity, ONE-Glo EX reagent (Promega) was added to the cells and luminescence was measured on a BMG Clariostar reader.

Real-time RT-PCR was performed on an ABI 7900 96 HT series PCR machine (Applied Biosystems) using SYBR Green Supermix (Bio-Rad Laboratories). Primer sequences are listed in Supplementary Table S2.

Virus were produced using HEK293T cells that detailed in Supplementary Materials and Methods.

For Wnt reporter assay, the cells were treated with PBS, 500 ng/mL Wnt3a recombinant protein (R&D, #5036-WN), Wnt3a or control conditioned media that isolated from L cells as described previously (30). Six hours after the treatment, Wnt activity was determined by firefly luciferase signal.

Results are reported as mean ± SEM. For in vivo liver, lung metastasis, and primary tumor growth experiments, the number of mice in each group was indicated in each specific experiment. Survival curves were constructed using the Kaplan–Meier method and compared between subgroups with the log-rank test. All statistical analyses were done using GraphPad Prism 7 (RRID:SCR_002798). In instances where multiple comparisons were performed, Dunnett or Bonferroni correction was used, unless otherwise specified. Longitudinal models for analyzing the repeated measurements of tumor volume and body weight to test: (i) whether there is a significant difference in the overall trend with a polynomial model and best fitted correlation structure among the repeated measurements; and (ii) whether the difference is significant at the last time-point. Specific tests are described in the respective figure legends. P values <0.05 were considered statistically significant, with *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; n.s., not significant (P > 0.05), unless otherwise indicated in the figure.

To study the function of MTDH in colorectal cancer using genetically modified mouse models, we obtained the B6.Apc^min/+ strain from Jackson Laboratory, which has been widely used to mimic APC mutation-induced gastrointestinal (GI) tract cancers (31, 32). The strain was crossed to the B6.Mtdh^−/− strain that we previously generated (19, 20) to obtain B6.Apc^min/+ mice with and without endogenous Mtdh KO (Fig. 1A). Mtdh KO significantly extended the lifespan of mice as compared with the Mtdh wild-type (WT) strain (Fig. 1B). Consistently, Mtdh KO significantly reduced cancer development, which is evidenced by less tumor nodules across the intestines (Fig. 1C and D), suggesting that MTDH is critical for GI tract cancer development.

To further validate these findings, we generated a human colorectal cancer cell line SW620 with endogenous MTDH KD by transducing the cells with lentiviruses expressing a short hairpin RNA (shRNA) targeting MTDH (Fig. 1E). The control and MTDH-KD SW620 cells were then injected subcutaneously into NSG mice to investigate primary tumor growth as well as spontaneous metastasis. MTDH-KD significantly delayed primary tumor progression in vivo (Fig. 1F–H). Moreover, spontaneous liver metastasis was also notably attenuated upon MTDH KD (Fig. 1I). Next, we performed intrasplenic injections to directly investigate the role of MTDH in experimental liver metastasis. NSG mice that were injected with control SW620 cells generated significantly more liver metastatic nodules than MTDH-KD cells (Fig. 1J and K). The results were confirmed with another colorectal cancer cell line DLD1 (Supplementary Figs. S1A–S1F). These findings were consistent with prognosis analysis of MTDH in the TCGA colorectal cancer dataset. Patients with high level of MTDH expressed in their tumors experienced significantly worse DFS (Supplementary Fig. S1G). Overall, these data indicated that MTDH enhances colorectal cancer growth and metastasis.

Next, we investigated how MTDH enhances intestinal/colorectal cancer development. Intestinal samples collected from B6.Apc^min/+;Mtdh^+/+ (WT), B6.Apc^min/+;Mtdh^+/− (HET), or B6.Apc^min/+;Mtdh^−/− (KO) mice were subjected to IHC staining of Ki67. The results indicated that MTDH promotes tumor proliferation (Fig. 2A and B), as indicated by number of Ki67-positive cells. However, no difference in terms of cancer cell survival was observed, which was indicated by cleaved caspase-3 (CC-3) staining (Supplementary Figs. S2A and S2B).

To further explore the underlying mechanism, we established an in vitro 3D culture assay using human SW620 colorectal cancer cells. Consistent with the high proliferation rate that we observed in B6.Apc^min/+;Mtdh^+/+ mice, control SW620 cells formed more and bigger tumorspheres in our 3D culture system than MTDH-KD cells (Fig. 2C and D). Previous studies have established Wnt signaling as the most frequently altered and hyperactivated pathway in colorectal cancer (33, 34). We next investigated if MTDH is involved in Wnt signaling activation. To this end, we stably expressed the 7xTcf-FFluc Wnt reporter (35) in control and MTDH-KD SW620 cells. We observed a lower basal Wnt signaling activity in MTDH-KD cells than control cells, and moreover, less activation of the reporter in response to Wnt3a stimulation (Fig. 2E). These results suggested that MTDH may be required for optimal Wnt activation in colon cancer.

SW620 tumors from experiment in Fig. 1F were further analyzed to validate these in vitro findings. Many Wnt signaling downstream targets were found to be downregulated in MTDH-KD tumors (Fig. 2F; Supplementary Fig. S2C), suggesting reduced Wnt signaling activity in vivo upon MTDH loss. Previous studies indicated that MTDH could be involved in β-catenin nuclear translocation in other cancer types (36, 37), which is the key step of Wnt activation. Consistent with this notion, we observed significantly less nuclear β-catenin in MTDH-KD colorectal cancer 3D culture spheres (Fig. 2G; Supplementary Figs. S2D and S2E). The finding was also validated by our in vivo liver metastatic samples and spontaneous mouse model, which again showed weaker nuclear localization of β-catenin upon MTDH KD and KO (Supplementary Figs. S2F–S2I). Moreover, immunoprecipitation assay indicated that MTDH interacts directly with β-catenin (Fig. 2H). Of note, MTDH has stronger interaction with S33Y and S33A/S37A/S45A mutant forms of β-catenin, which mimic the activated forms of β-catenin (Fig. 2H; ref. 29). Collectively, these results show that MTDH binds to β-catenin, especially its activated forms, and facilitates its nuclear translocation. As a result, Wnt signaling was enhanced to promote colorectal cancer growth, progression, and metastasis.

Taking advantage of the same Mtdh KO mouse model, we also explored the function of MTDH in lung cancer development. To this end, we employed the B6.KRAS-G12D^fl/+ model, which is a widely used mouse model of K-Ras-induced lung adenocarcinoma (38). Endogenous Mtdh was knocked out in the strain by crossing it with our B6.Mtdh^−/− mouse model (Fig. 3A). The B6.KRAS-G12D^fl/+;Mtdh^+/+ (WT), B6.KRAS-G12D^fl/+;Mtdh^+/− (HET), and B6.KRAS-G12D^fl/+;Mtdh^−/− (KO) mice were randomized and were subjected to intranasal infection with adenovirus Cre (27) to specifically induce KRAS-G12D expression in the lungs and initiates lung cancer development. Mice that maintained WT Mtdh expression had a significantly worse survival rate after adenovirus Cre infection (Fig. 3B) as compared with Mtdh HET and Mtdh KO mice. Consistently, WT mice had heavier lung tumor burden compared with Mtdh KO and Mtdh HET mice (Fig. 3C–E), suggesting that MTDH promotes lung cancer progression.

We also used the mouse Lewis lung carcinoma cell line LLC1, which was established from the lung of a C57BL/6 mouse (39) to confirm the results that we observed in the genetic KRAS mouse model. This cell line is highly tumorigenic and elicits significant antitumor immune response, and has been widely used as a syngeneic mouse model to evaluate therapeutic efficacy for lung cancer in vivo (40–44). We generated LLC1 cells with endogenous MTDH KD (Fig. 3F) and injected these cells into C57BL/6 mice to investigate lung cancer progression. As expected, mice injected with MTDH-KD cells developed fewer tumors in lungs and had improved survival rates than mice injected with control cells (Fig. 3G and H). Consistently, lung adenocarcinoma patients with low MTDH expression in their tumors have significantly better DFS based on the analysis of the TCGA dataset (Supplementary Fig. S3A).

Next, we asked whether MTDH enhances lung cancer progression also through facilitating β-catenin nuclear translocation and enhancing Wnt signaling. To address this question, we performed Wnt reporter assay in LLC1 cells with or without MTDH KD. We did not observe different Wnt activation at basal level upon MTDH KD. However, MTDH KD significantly inhibited Wnt activation when the cells were treated with Wnt3a conditioned media (Supplementary Fig. S3B), suggesting that MTDH KD also attenuates Wnt activation in lung cancer cells in vitro. Moreover, we observed less tumor sphere formation after MTDH KD (Supplementary Fig. S3C). To further confirm these findings in vivo, lung tumors formed by control or MTDH KD LLC1 cells were collected for IHC staining to investigate the nuclear translocation of β-catenin. To our surprise, although there is a trend toward reduced nuclear β-catenin upon MTDH KD, this difference did not reach statistical significance (Supplementary Fig. S3D). Moreover, the overall ratio of nuclear localized β-catenin is low in the lung tumor sections (Supplementary Fig. S3D), indicating that Wnt/β-catenin signaling may not play a key role in LLC lung cancer progression in vivo. Consistent with this notion, it has been reported that Wnt hyperactivation is not the main cause of lung cancer in the clinic (45). Taken together, although MTDH may enhance Wnt activation upon induction in vitro, the lung cancer-promoting role of MTDH may be mediated by other mechanisms in vivo.

Given the facts that lung is an inflammatory organ with pronounced immune cell infiltration (46) and immune checkpoint blockade therapy has achieved great success in lung cancer (47), we asked if MTDH affects T-cell infiltration into tumors. To this end, lungs with tumors from the B6.KRAS-G12D^fl/+ model were collected to analyze T-cell populations. MTDH KO significantly increased T-cell infiltration, especially CD8⁺ cytotoxic T cells that are mainly responsible for tumor clearance (Fig. 4A, gating strategy in Supplementary Fig. S3E). The elevated T-cell infiltration in lung tumors with MTDH KO was also confirmed with IHC staining (Supplementary Figs. S3F and S3G). Moreover, lungs with MTDH KO have significantly fewer PD-1 expressing CD8⁺ T cells (Fig. 4A, right). On the other hand, we observed less PD-L1 expression in tumor samples upon MTDH KO (Fig. 4B), suggesting MTDH KO tumors may experience less T-cell exhaustion. These findings were further validated by immunostaining of LLC1 tumors, which again showed that MTDH KD enhanced T-cell infiltration, reduced PD-1⁺ cells, and downregulated PD-L1 expression in tumors (Fig. 4C and D). Collectively, our results indicate that MTDH promotes lung cancer progression by inducing PD-1/PD-L1 expression, and as a result, decreasing cytotoxic T-cell infiltration.

As observed above, MTDH-induced lung cancer is mediated by decreasing cytotoxic T-cell infiltration. In addition to Wnt/β-Catenin pathway, we asked if such alteration of immune cell population as compared with MTDH KO also contributes to colorectal cancer progression. To this end, we collected intestine samples from age-matched B6. Apc^min/+;Mtdh^+/+ (WT) and B6. Apc^min/+;Mtdh^−/− (KO) mice and performed IHC staining of CD3, CD8, PD-1, and PD-L1. We did not detect any significant difference between MTDH WT and KO mice in terms of the numbers of CD3⁺, CD8⁺, and PD-L1⁺ cells in tumors (Supplementary Figs. S3H and S3I). Interestingly, we did not observe any PD-1⁺ cells in either MTDH WT or KO samples (Supplementary Fig. S3H). Moreover, in general, we observed much less overall CD3⁺ and CD8⁺ T-cell infiltration compared with lung tumor models (Fig. 4C and D; Supplementary Figs. S3H and S3I), suggesting immune cell regulation may not be a major contributing factor to the tumor-promoting function of MTDH in colorectal cancer, at least in the mouse model used in our current study. Collectively, the results indicate that MTDH promotes lung and colorectal cancers progression and metastasis through two distinct mechanisms.

The tumor-promoting role of MTDH in colorectal and lung cancers in this study, as well as results from other cancer models (19, 20) established it as a potentially novel target for cancer therapy (14). To evaluate the therapeutic potential of targeting MTDH, we developed a set of LNA-modified ASOs with potent MTDH-targeting activity in vitro and in vivo. LNA-modified ASOs against MTDH were designed in silico with a focus on target selectivity to avoid undesired off-target effects. A set of 104 LNA-modified ASOs were screened with regard to target KD efficacy in 4T1 and Renca cells at a final concentration of 5 μmol/L without the use of a transfection reagent. Mtdh mRNA expression was potently inhibited in both cell lines by MTDH-specific LNA-modified ASOs (Supplementary Fig. S4A). There was a significant correlation in the activity of ASOs in both tested cell lines. Of note, treatment of cells with a control oligonucleotide did not result in any reduction of Mtdh mRNA expression (Supplementary Fig. S4A). We selected five top ASOs that inhibited Mtdh mRNA expression by >90% (represented by a residual Mtdh mRNA expression of <0.1) and investigated the dose-dependency of target KD in 4T1 cells. All candidates reduced Mtdh mRNA expression in a dose-dependent manner with IC₅₀ values in the low nanomolar range (Supplementary Fig. S4B). The KD efficacy of the ASOs was also validated at the protein level in colorectal and lung cancer cell lines, in addition to the 4T1 cell line (Fig. 5A; Supplementary Fig. S4C).

It has been shown in the past that ASOs can have pro-inflammatory potential, e.g., by activating TLR9 (48). We therefore investigated the pro-inflammatory potential of the five ASOs using a reporter assay and did not observe TLR9 activation (Fig. 5B), which supports the feasibility for in vivo treatment. Three ASOs, A34051Mi, A34068Mi, and A34082Mi were next selected for extensive assessment of in vivo tolerability. To this end, we designed a treatment scheme based on our experiences that we gathered in the past for the tolerability study (Supplementary Fig. S4D). Specifically, there is a loading phase with 5 consecutive injections in the first week (days 0–6) followed by maintenance in the second week (days 7–13), with a one-day break in between each of three injections. The loading phase is beneficial for the efficacy of ASOs as this leads to saturation of the liver and the kidney where most of the ASOs end up. The mice were then treated with these ASOs using the treatment regimen. Body weight was measured daily starting on the day of treatment initiation and liver damage was evaluated at endpoint to monitor potential toxicity. Among these ASOs, only A34068Mi induced significant body weight loss across the treatment (Fig. 5C). Consistently, high liver toxicity was observed in the mice treated with A34068Mi as measured by serum ALT and AST levels (Fig. 5D). Taken together, we selected A34051Mi and A34082Mi for further studies as they could effectively KD MTDH with minimal toxicity in vivo.

To evaluate the therapeutic potential of the two selected MTDH LNA-modified ASOs, an optimal in vivo treatment scheme was designed on the basis of a previously reported study (49). A34051Mi was selected to validate the MTDH KD efficiency in vivo with the treatment scheme (Supplementary Fig. S5A). Considering the extensive functional studies of MTDH that have been done on breast cancer (17–20, 50), we chose a breast cancer model to start with the in vivo treatment to determine the in vivo MTDH KD efficiency. Mice bearing 4T1 primary mammary gland tumors were treated as indicated in Supplementary Fig. S5A. At the end of the treatment, primary tumors were collected and MTDH KD efficiency was evaluated with qRT-PCR and Western blot analysis. As we expected, MTDH mRNA and protein levels were effectively reduced in vivo upon the treatment (Supplementary Fig. S5B and S5C). We therefore adapted the treatment scheme that includes five consecutive injections as a loading phase in the first week followed by a maintenance phase of three consecutive injections in each week for the subsequent experimental therapeutic studies (Supplementary Fig. S5D).

Next, we started the therapeutic evaluation in a breast cancer model. Mice bearing 4T1 primary tumors with size of ∼2 mm in diameter were randomized into three groups and treated with the negative control oligonucleotide or the two MTDH-specific ASOs, A34051Mi and A34082Mi. A34051Mi and A34082Mi treatments significantly inhibited primary tumor growth as well as spontaneous lung and liver metastasis (Supplementary Figs. S6A–S6C). MTDH KD was confirmed with Western blot analysis in tumors from mice that had been treated with A34051Mi and A34082Mi (Supplementary Fig. S6D). Next, we extended our treatments to colorectal and lung cancer models. Mice bearing MC-38 and LLC1 tumors, which represent colorectal and lung cancer respectively, were treated with control ASO, A34051Mi, or A34082Mi. A34051Mi and A34082Mi treatments significantly suppressed tumor progression (Fig. 6A, B, E, and F) as well as liver and lung metastasis (Fig. 6C; Supplementary Fig. S6E), and significantly extended the survival in the LLC1 model (Fig. 6G). To directly assess the anti-metastasis effect of ASOs, we performed tail vein injection of 4T1 cells and intrasplenic injection of MC-38 cells to develop experimental lung and liver metastasis, respectively. After metastases were established, mice were treated with control ASO or A34051Mi and A34082Mi with the same scheme as shown in Supplementary Fig. S5D. The development of metastasis was monitored weekly by in vivo BLI. Significant reduction of lung and liver metastasis was observed upon MTDH-ASOs treatments (Supplementary Figs. S6F–S6I), confirming their anti-metastasis therapeutic effects. Consistently, Western blot analysis and IHC staining of MC-38 and LLC1 tumor samples validated potent KD of MTDH in tumors upon A34051Mi and A34082Mi treatments (Fig. 6D and H). Collectively, our data established the therapeutic potential of targeting MTDH with LNA-modified ASOs.

To further validate that the tumor-suppressive effect upon A34051Mi and A34082Mi treatments were caused specifically by MTDH targeting, MC-38 cells treated with the control oligonucleotide or the MTDH-specific ASOs A34051Mi or A34082Mi were subjected to the Wnt reporter assay in vitro. Consistent with what we observed in stable MTDH KD cells (Fig. 2E), A34051Mi and A34082Mi treatment also attenuated Wnt activation (Fig. 7A). Consistently, MC-38 tumors from animals treated with A34051Mi and A34082Mi showed significantly lower expression of Wnt downstream genes (Fig. 7B). These results indicate that A34051Mi and A34082Mi suppress Wnt activation in colorectal cancer by targeting MTDH. Similar to what we observed in Mtdh KO mice, A34051Mi and A34082Mi decreased PD-L1 expression in lung tumors, reduced PD-1⁺ cells, and increased cytotoxic T-cell infiltration in LLC1 tumors (Fig. 7C and D). Taken together, these results confirm that MTDH-specific LNA-modified ASOs (A34051Mi and A34082Mi) suppress colorectal and lung cancer progression and metastasis by specifically reducing the expression of MTDH.

In this study, we extensively explored the functions of MTDH in lung and colorectal cancers with both genetically modified spontaneous tumor models and xenograft/allograft models. The results indicate that MTDH promotes lung and colorectal cancer progression and metastasis with distinct mechanisms. More importantly, we identified MTDH-specific LNA-modified ASOs that potently inhibit MTDH expression in vivo. Targeting MTDH with LNA-modified ASOs significantly inhibits lung and colorectal cancer progression and metastasis, suggesting MTDH-specific ASOs as a promising therapeutic strategy to treat lung, colorectal, breast and other cancers.

Previous studies indicated that MTDH is upregulated during colorectal cancer progression in patients (51). Its genetic variants have also been observed in colorectal cancer patient samples (52), indicating that MTDH is involved in colorectal cancer progression. In vitro studies also revealed that MTDH promotes the tumorigenesis and migration of colorectal carcinoma cells (53, 54). Consistent with these in vitro findings and clinical correlations, our genetic model of spontaneous mouse intestinal cancer models that mimic Wnt-induced colorectal cancer in patients showed a critical role of MTDH for cancer progression and metastasis in vivo (Fig. 1). Mechanistically, MTDH binds to and facilitates β-catenin nuclear translocation, and thus, elevates the activation of Wnt signaling. Consistent with our findings, positive correlation between MTDH and nuclear location of β-catenin has also been observed in patient samples (54). Moreover, MTDH high expression is correlated with worse survival in colorectal cancer patients (Supplementary Fig. S1G).

The involvement of MTDH in Wnt/β-catenin signaling has not only been identified in colorectal cancer, but also in other cancers, such as glioma (37), liver cancer (36, 55), and gastric cancer (56). However, in liver and gastric cancers, MTDH activates Wnt signaling by upregulating the expression of LEF1 or β-catenin (55, 56) rather than by directly binding to and facilitating the nuclear translocation of β-catenin. In our study, we observed a direct interaction of MTDH with β-catenin, which is consistent with previous studies (37, 54). Moreover, we found that MTDH binds even stronger to activate β-catenin (Fig. 2H). Taken together, the findings from our study and others established a key role of MTDH for Wnt activation by facilitating the nuclear translocation of activated β-catenin during colorectal cancer progression and metastasis.

In addition to Wnt/β-catenin signaling, other pathways, such as NF-κB (51), apoptosis (19, 22), and AMPK (53) have also been proposed to explain the MTDH-dependent promotion of cancer development. These findings suggest a multifaceted role of MTDH in affecting multiple cancer-related pathways and functions during tumor progression in diverse cancer types. In line with this notion, we found that MTDH promotes lung cancer progression by suppressing T cell infiltration, which is distinct from its role in colorectal cancer. Moreover, given the fact that we observed reduced PD-L1 expressions in MTDH-KD LLC1 tumors, in which the host MTDH has not been altered, we believe that MTDH-induced reduction of cytotoxic T infiltration is likely through a tumor intrinsic pathway. However, we cannot exclude the possibility that stroma MTDH may also be involved in regulating T cell response to tumors, which may be a subject of investigation in future studies.

After having established the role of MTDH in colorectal and lung cancer progression and metastasis, we sought to therapeutically target MTDH using LNA-modified ASOs. To this end, we designed a set of ASOs as Gapmers that contain a central “gap” of DNA monomers flanked by LNA-modified nucleotides. The central “gap” recruits the cellular enzyme RNase H, which cleaves the target RNA upon binding. The ASOs that we used have a fully phosphorothioate (PTO) backbone, which makes them resistant to enzymatic degradation (57). As described in a previous study (49), we also observed potent activity of MTDH-specific ASOs in vitro without the use of a transfection reagent and in vivo without a delivery system. Both lung and colorectal cancer progression and metastasis were significantly suppressed upon treatment with MTDH-specific ASOs in mouse models. In parallel, given the tumor-promoting role of MTDH in breast cancer that we identified in our previous studies (17–19), we also tested the susceptibility of a breast cancer model for treatment with MTDH-specific ASOs. As expected, the selected MTDH ASOs significantly inhibited primary tumor growth and metastasis, suggesting the therapeutic potential of these ASOs could extend to other cancer types in which MTDH is required for tumor progression and metastasis. Moreover, the suppression of Wnt signaling and elevation of T cell infiltration upon MTDH ASO treatments confirms the on-target effects of these ASOs mechanistically. Taken together, our study thoroughly elucidated the function of MTDH in lung and colorectal cancers. Furthermore, we validated the therapeutic potential of LNA-modified ASOs targeting MTDH, which could be further developed to benefit patients with lung, colorectal, breast and other types of cancer in the future.

M. Shen reports grants from Susan G. Komen Foundation and New Jersey Commission on Cancer Research during the conduct of the study. S. Michel reports personal fees from Secarna Pharmaceuticals GmbH & Co. KG during the conduct of the study and has a patent for inhibitor of metadherin expression pending. L. Wan reports a patent for MTDH–SND1 interaction as a target for therapeutic intervention in cancer, No. 10,357,539 B2 issued and licensed to Firebrand Therapeutics. J.F. Jin reports other funding from Firebrand Therapeutics during the conduct of the study and other funding from Firebrand Therapeutics outside the submitted work. F. Jaschinski reports personal fees from Secarna Pharmaceuticals GmbH & Co. KG during the conduct of the study and has a patent for inhibitor of metadherin expression pending. R. Klar reports personal fees from Secarna Pharmaceuticals GmbH & Co. KG during the conduct of the study and has a patent for Inhibitor of metadherin expression pending. Y. Kang reports other funding from Firebrand Therapeutics and nonfinancial support from Secarna Pharmaceuticals outside the submitted work and has patents for US 9,932,587 B2 and US10,357,539 issued and licensed to Firebrand Therapeutics. No disclosures were reported by the other authors.

M. Shen: Conceptualization, data curation, formal analysis, validation, investigation, methodology, writing—original draft, writing—review and editing. S. Xie: Data curation, formal analysis, validation, investigation, methodology. M. Rowicki: Data curation, formal analysis, investigation. S. Michel: Data curation, methodology. Y. Wei: Data curation, formal analysis. X. Hang: Data curation, investigation. L. Wan: Investigation, methodology. X. Lu: Investigation, methodology. M. Yuan: Data curation. J.F. Jin: Conceptualization, project administration. F. Jaschinski: Resources, data curation, investigation, methodology, writing—original draft. T. Zhou: Resources. R. Klar: Conceptualization, resources, data curation, formal analysis, investigation, methodology, writing—original draft. Y. Kang: Conceptualization, resources, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, writing—original draft, project administration, writing—review and editing.

We thank N. Aiello, R. Gunaratna, M. Esposito, and other lab members for technical supports and helpful discussions. We thank C. DeCoste and K. Rittenbach at the Molecular Biology Flow Cytometry Resource Facility of Princeton University for flow cytometry assays. This research was supported by the Brewster Foundation and grants from the U.S. Department of Defense (BC123187) and American Cancer Society (RP-19-180-01-CSM) to Y. Kang, and postdoctoral fellowships from Susan G. Komen (PDF17332118) and NJCCR (DFHS15PPCO21) to M. Shen. This research was also supported by the Pre-clinical Imaging and Flow Cytometry Shared Resources of the Rutgers Cancer Institute of New Jersey (P30CA072720).

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