β-catenin attenuation leads to up-regulation of activating NKG2D ligands and tumor regression in BrafV600E -driven thyroid cancer cells

Introduction BRAFV600E mutations frequently occur in papillary thyroid cancer (PTC). β-catenin, encoded by CTNNB1, is a key downstream component of the canonical Wnt signaling pathway and is often overexpressed in PTC. BRAFV600E -driven PTC tumors rely on Wnt/β-catenin signaling to sustain growth and progression. Methods In the present study, we investigated the tumorigenicity of thyroid cancer cells derived from BRAFV600E PTC mice following Ctnnb1 ablation (BVE-Ctnnb1null ). Results Remarkably, the tumorigenic potential of BVE-Ctnnb1null tumor cells was lost in nude mice. Global gene expression analysis of BVE-Ctnnb1null tumor cells showed up-regulation of NKG2D receptor activating ligands (H60a, H60b, H60c, Raet1a, Raet1b, Raet1c, Raet1d, Raet1e, and Ulbp1) and down-regulation of inhibitory MHC class I molecules H-2L and H-2K2 in BVE-Ctnnb1null tumor cells. In vitro cytotoxicity assay demonstrated that BVE-Ctnnb1wt tumor cells were resistant to NK cell-mediated cytotoxicity, whereas BVE-Ctnnb1null tumor cells were sensitive to NK cell-mediated killing. Furthermore, the overexpression of any one of these NKG2D ligands in the BVE-Ctnnb1wt cell line resulted in a significant reduction of tumor growth in nude mice. Conclusions Our results indicate that active β-catenin signaling inhibits NK cell-mediated immune responses against thyroid cancer cells. Targeting the β-catenin signaling pathway may have significant therapeutic benefits for BRAF-mutant thyroid cancer by not only inhibiting tumor growth but also enhancing host immune surveillance.


Introduction
Papillary thyroid cancer (PTC) is the most common type of thyroid cancer, accounting for more than 80% of all thyroid cancer cases (1). The BRAF V600E mutation is the most frequent genetic alteration in PTC ranging from 29% to 83% depending on the cohort studied with overall rate of approximately 45% (2,3). Nearly 60% (59.7% of 496 cases) of PTC patients harbor BRAF V600E mutations from the TCGA thyroid cancer consortium (4). In some studies, BRAF V600E mutation has been shown to be one of the factors contributing to thyroid cancer recurrence and mortality (5,6). The mutation constitutively activates the RAS-RAF-MEK-ERK mitogen-activated protein kinase (MAPK) signaling cascade and promotes the initiation and progression of PTC. It has been demonstrated that Braf V600E mutation drives oncogenic transformation of thyroid epithelial cells and development of PTC in transgenic mouse models (7,8).
b-catenin is a downstream component of the Wnt/b-catenin pathway and plays an important role in embryonic development and tissue homeostasis (9). Dysregulation of its signaling is involved in many types of tumors, including thyroid cancer (10)(11)(12)(13). Functional cross-talk between b-catenin and MAPK, PI3K/AKT, or CREB (cAMP-response element binding protein) signaling pathways has been demonstrated to maintain proliferation in thyroid cancer cell lines (12,14). Recently, we have shown that BRAF-driven PTC requires active b-catenin signaling to sustain its growth (15). Deletion of b-catenin results in tumor regression and differentiation as well as increased sensitivity to the BRAF V600E inhibitor PLX4720.
In this study, we investigated the tumorigenicity of thyroid cancer cells derived from Braf V600E -induced PTC mice with wild-type Ctnnb1 (BVE-Ctnnb1 wt ) and deleted Ctnnb1 (BVE-Ctnnb1 null ). We found that b-catenin ablation led to complete loss of tumorigenicity of BVE-Ctnnb1 null cells in nude mice, with increased expression of NKG2D ligands in BVE-Ctnnb1 null cells. Our data strongly suggest that b-catenin exerts oncogenic effects by promoting tumor growth and evading NK cell-mediated immune surveillance.

Animals
Athymic BALB/c-nu/nu (nude mice) and C57BL/6J mice (6-10 weeks of age) were acquired from Jackson Laboratory. Mice were provided with autoclaved food and water ad libitum. The study was approved by the Animal Care and Use Committee of the institution (RAC# 2190004) and was conducted in compliance with the Public Health Service Guidelines for the Care and Use of Animals in Research.
Thyroid tumor cell lines BVE-Ctnnb1 wt and BVE-Ctnnb1 null were established from thyroid tumors collected aseptically from donor mice, as described previously (15) and maintained in DMEM/F12 growth medium containing 10% fetal bovine serum, 100 units/ml penicillin, and 100 mg/ml streptomycin.

RNA sequencing for quantification of differentially expressed genes
Total RNA from BVE-Ctnnb1 wt and BVE-Ctnnb1 null cell lines was isolated using TRI Reagent® solution (#T9424, Sigma-Aldrich, St. Louis, MO, USA). Libraries were constructed using an Illumina TruSeq RNA Library Prep kit (San Diego, CA, USA) according to the manufacturer's instructions. All sequencing was performed on an Illumina HiSeq 4000 platform with at least 20 million clean reads. Significant DEGs were selected based on the following criteria: Log2 fold change>2, false discovery rate (FDR) <0.001, and p-value from the difference test<0.01. Gene list annotation and enrichment of biological pathways were performed using Metascape (16).

Cytotoxicity assay and phenotyping of effector cells
A label-free cell-based assay (ACEA Biosystems, San Diego, CA, USA) was used to measure the cytotoxicity of spleen cells from C57BL/6J (effector) mice against tumor cells (target) at an effector/ target ratio of 25:1, as previously described (17). The effector cells were immune phenotyped for the presence of surface markers: CD4, CD8, and asialo GM1 (aGM1) using complement-mediated (Cederlane Low-Tox®-M Rabbit Complement, #CL3051, 1:20 dilution) cytotoxicity assay in the presence or absence of appropriate blocking antibodies (18). Briefly, target cells (2000 cells/well) were seeded in electrode strip-coated 96-well electrodeintegrated microplates. Spleen cells were harvested from two normal C57BL/6J mice. Splenocytes were added to the target cells at 50,000 cells/well in the presence of 5 µg/ml phytohemagglutinin (PHA) and 100 U/ml IL-2 for the activation of T and NK cells with or without anti-CD8 (#CL169AP, 1:20 dilution, plus 5ul complement), anti-CD4 (#CL012AP, 1:20 dilution, plus 5ul complement), or asialo GM1 (anti-NK cell) antibodies (#CL8955, 1:100 dilution, plus 5ul complement, Cedarlane, Ontario, Canada). The percentage of specific cytotoxicity was calculated based on the cell index (CI), which was derived from the changes in electrode impedance due to cell growth and attachment relative to that of the background (without cells). The following formula was used: Cytotoxicity = [(experimental CIcontrol CI)/(untreated CIcontrol CI)] × 100.

Cloning and expression of inhibitory Ly49 ligands in BVE-Ctnnb1 null cells
The cDNA for each Ly49 ligand H2-L, H2-K2, and H2-K1 was cloned into the pcDNA3.1 expression vector and transfected into the BVE-Ctnnb1 null cell line using Lipofectamine (Invitrogen, CA, USA) and selected for 6 weeks with 100 µg/ml zeocin. Stable clones expressing each ligand were subcutaneously injected into nude mice to observe tumor growth. The overexpression of H2-L, H2-K2, and H2-K1 was confirmed by qRT-PCR.

Statistical analysis
Student's t-test (two-tailed) was used to compare two groups and the area under curve (AUC) calculation was performed using GraphPad Prisim 8. A P value < 0.05 was considered significant.

Tumorigenicity of BVE-Ctnnb1 null cells
Tumorigenicity of BVE-Ctnnb1 null and BVE-Ctnnb1 wt cells was investigated in athymic nude mice. No tumor growth was observed after 4 weeks of subcutaneous injection of 1 × 10 6 BVE-Ctnnb1 null cells into the right flank of the mice (n=8), except that a small nodule (0.5 × 0.5 cm) was formed at the injection site in one mouse (n=8). However, tumors (approximately 2 × 2 cm) developed in all mice 4 weeks following injection of 1 × 10 6 BVE-Ctnnb1 wt cells. The nodule, which formed at the injection site of BVE-Ctnnb1 null cells, was confirmed to be a tertiary lymph node by histology, with characteristic features of lymphoid follicles and germinal centers ( we compared the gene expression profiles of BVE-Ctnnb1 wt and BVE-Ctnnb1 null cells using RNA-Seq analysis of three biological replicates. Pathway analysis showed 833 up-regulated and 1407 down-regulated genes with a log2 fold-change >2 (Supplementary  Table 1). Importantly, the genes involved in NK cell-mediated immunity were among the top 20 enriched ontology clusters that were significantly up-regulated in the BVE-Ctnnb1 null cells ( Figure 2A). The genes typically involved in cancer growth and progression, such as extracellular matrix organization, chemotaxis, vascular development, integrin cell surface interactions and signaling, and ERK-1/ERK2 signaling, were among the top 20 enriched ontology clusters that were significantly down-regulated in the BVE-Ctnnb1 null cells ( Figure 2B). These changes in biological pathways are consistent with our previous study showing that the bcatenin signaling pathway is required for the growth and progression of BRAF V600E -driven thyroid cancer (15). The gene list for each cluster is presented in Supplementary Tables 2 and 3. We further analyzed the list of DEGs involved in NK cell activation. As shown in Figure 3, BVE-Ctnnb1 null cells showed up-regulation of NKG2D activating ligands (H60a, H60b, H60c, Raet1a, Raet1b, Raet1c, Raet1d, Raet1e and Ulbp1) and downregulation of inhibitory MHC class I molecules (H2-L, H2-K2, and H2-D1), which may lead to NK cell activation and tumor cell clearance. In contrast, the up-regulation of H2-L and H2-K2 and down-regulation of NKG2D ligands in BVE-Ctnnb1 wt tumor cells would probably result in evasion of NK cell-mediated immune surveillance and promotion of tumor growth.
NK-mediated cytotoxicity of BVE-Ctnnb1 wt and BVE-Ctnnb1 null cells Next, we investigated the sensitivity of BVE-Ctnnb1 null and BVE-Ctnnb1 wt cells to NK cell-mediated cytotoxicity using in vitro cytotoxicity assay. As shown in Figure 4, activated splenocytes completely inhibited the replication of both BVE-Ctnnb1 wt and BVE-Ctnnb1 null tumor cells (top panel). However, BVE-Ctnnb1 wt tumor cells were able to replicate at 67% efficiency at 120h in the presence of anti-CD8 T cell neutralizing antibodies, indicating their resistance to NK cell-mediated cytotoxicity. The area under curve (AUC) of BVE-Ctnnb1 null , BVE-Ctnnb1 wt , and BVE-Ctnnb1 wt + splenocytes + anti-CD8 were 210.4 ± 3.2, 128.7 ± 2.4, and 42.5 ± 1.9, respectively. In contrast, BVE-Ctnnb1 null tumor cells were unable to grow in the presence of anti-CD8 T cell neutralizing antibodies, indicating their sensitivity to NK killing (middle panel). Both BVE-Ctnnb1 wt and BVE-Ctnnb1 null tumor cells were unable to replicate in the presence of anti-NK cell neutralizing antibodies, indicating their sensitivity to CD8 T cell-mediated cytotoxicity (bottom panel).

Ctnnb1 wt cells reduces tumor growth
To further confirm the role of NKG2D activating ligands in tumor regression, we overexpressed each of these ligands by stable transfection of their cDNAs into BVE-Ctnnb1 wt cells and then injected these ligand-expressing tumor cells (1 × 10 6 ) into nude mice (n=5) to observe tumor growth for 4 weeks. As shown in Figure 5, overexpression of these ligands caused a significant reduction in tumor growth (p<0.001). In contrast to the complete loss of tumorigenicity in BVE-Ctnnb1 null cells, over-expression of a single NKG2D ligand could not eliminate BVE-Ctnnb1 wt tumor cell growth in nude mice. This might be due to high expression of inhibitory H2-L and H2-K2 ligands in BVE-Ctnnb1 wt tumor cells and/or tumor cell clearance may require the simultaneous expression of multiple NKG2D ligands for optimal NK cell activation.

Overexpression of Ly49 ligands in BVE-Ctnnb1 null cells partially rescue tumor growth
Since BVE-Ctnnb1 null cells express high levels of NKG2D activating ligands and low levels of H2-L and H2-K2 inhibitory ligands, we investigated whether the loss of tumorigenicity in BVE-  Ctnnb1 null cells could be partially rescued by overexpression of H2-L or H2-K2 ligands in these cells. As shown in Figure 6, overexpression of these ligands resulted in tumor growth in nude mice, thus confirming the role of H2-L and H2-K2 ligands in NK cell inhibition. H2-K2 appeared to be more effective than H2-L in inhibiting NK cells. The potential mechanisms of NK cell-mediated rejection of BVE-Ctnnb1 null cells are summarized in Figure 7.

Discussion
In the present study, we provided evidence of b-catenin in modulating NK cell activity against Braf-mutant PTC in mice. Down-regulation of b-catenin increased the expression of NKG2D ligands and decreased the expression of inhibitory MHC class I molecules, such as H-2L and H-2K2, in tumor cells, which triggered NK cell-mediated tumor rejection.
It has been reported that oncogenic activation of KRAS/BRAF/ MEK signaling stimulates Wnt/b-catenin pathway by b-catenin/ TCF4 complex activation via Frizzled co-receptor LRP6 phosphorylation, which in turn promotes tumor growth and invasion (19,20). The LRP6 may be the link between MAPK and Wnt/b-catenin signaling cross-talk during oncogenesis (20). In the BrafV600E murine thyroid cancer, b-catenin expression is upregulated and its knockout results in down-regulation of MARK signaling pathway, indicating a positive crosstalk between MAPK and Wnt/b-catenin signaling pathways. Active Wnt/b-catenin signaling in turn promotes Braf V600E -mediated tumor growth (15). Genetic alterations in b-catenin have not been identified in PTC. However, aberrant b-catenin expression or localization, such as the reduction of b-catenin plasma membrane levels and its aberrant nuclear localization, has been reported and is associated with c-Myc and cyclin D1 overexpression, loss of tumor differentiation, and poor prognosis in human PTC samples (13,21).
The activation of tumor-intrinsic Wnt/b-catenin signaling is frequently associated with poor prognosis in most human cancers. Besides its intrinsic effects on tumor cell proliferation and apoptosis, the WntT/b-catenin signaling pathway has been found to act on the tumor microenvironment and lead to immune evasion by reducing tumor infiltration of dendritic cells (DCs), NK cells, and T cells via multiple mechanisms, such as inhibition of CCL4 expression on cytotoxic T lymphocytes, CXCL10 expression on DCs, or increased expression of dickkopf WNT signaling pathway inhibitor 1 (DKK1), which prevents NK cell-dependent cancer cell lysis by reducing the expression of NK cell-activating ligands (22-24).
Lee et al. reported that incubation of NK cells and lymphokineactivated killer cells with TGF-b1 resulted in a dramatic reduction in surface NKG2D expression associated with impaired NK cytotoxicity (25). They further demonstrated that the modulation of NKG2D by TGF-b1 was specific, as the expression of other NK receptors was not affected by the presence of TGF-b1. TGF-b1mediated impairment of NK cell cytotoxicity was not due to alteration of other lytic moieties, such as perforin, Fas, or the apoptotic pathway. TGF-b1 signaling is activated in BRAFV600Einduced thyroid cancer (26). In our previous studies, we showed that TGF-b inhibited the anti-tumor activity of NK and CD8+ T cells in Braf V600E thyroid cancer mouse model. The immune suppression could be reversed by IL-12 treatment with increased lymphocyte and macrophage infiltration, resulting in a significant reduction in tumor load and an increase in survival (18). Increased infiltration of lymphocytes and macrophages has also been observed in the same thyroid cancer model following b-catenin ablation (15). In the present study, we have identified a novel mechanism by which Braf V600E -mutant thyroid cancer cells evade NK cellmediated immune surveillance. This is achieved by downregulation of stimulating NKG2D ligands and up-regulation of inhibitory MHC class I molecules. Given that the activation of Wnt/b-catenin signaling leads to increased TGF-b signaling, the down-regulation of stimulating NKG2D ligands and up-regulation of inhibitory MHC class I molecules may be due to increased TGF-b signaling. However, in contrast to the reduction in surface NKG2D expression reported by Lee et al, NKG2D expression was not significantly different between BVE-Ctnnb1 wt and BVE-Ctnnb1 null tumor cells.
NK cells play an important role in cancer immune surveillance because of their ability to recognize tumor cells through the interaction of several distinct cell surface receptors, with the net effect of NK cell activation and cytotoxic attack on tumor cells (29,30). The activation of NK cells (CD56+ in humans and CD49b+ in mice) is balanced by anti-and pro-activation signals through inhibitory and activating receptors, such as Ly49 in mouse or killer cell Ig-like receptor (KIR in human) family receptors, NKG2 (CD159) family of C-type lectin-like receptors, including activating NKG2D and inhibitory NKG2A, and natural cytotoxicity receptors (NKp46, NKp44, and NKp30) (31)(32)(33). NKG2D is expressed in all NK cells and activated CD8+ T cells. Its ligands are not or rarely expressed in normal tissues, but are expressed in tumor cells and viral-infected tissues (34). Previous studies have shown that the expression of NKG2D ligands on tumor cells renders them susceptible to killing by NK cells in vitro and the rejection of tumor xenografts in vivo (35, 36). NKG2D-mediated killing is negatively regulated by MHC class I molecules on tumor cells, which bind to inhibitory Ly49 receptors on NK cells to maintain a balance between activating and inhibitory signals. Thus, high expression of MHC class I molecules in tumor cells prevents efficient NKG2D-mediated cytotoxicity against tumor cells (37,38). Our data indicate that active b-catenin signaling in Braf V600E -mutant thyroid cancer cells inhibits NK cell-mediated immune response. Attenuation of Wnt/b-catenin signaling boosts the host immune response against thyroid cancer. To our knowledge, this is a first report describing the inhibitory effect of b-catenin expression on NK cell activity, linking oncogene activation to innate immune suppression. Given that overexpression of NKG2D ligands results in regression of thyroid cancer xenografts in vivo, these ligands could be used for cancer vaccine development or gene therapy to generate anti-cancer immune responses.
In summary, our results provide insights into b-catenin signaling in the NK cell-mediated immune response against BRAF-mutant thyroid cancer cells. b-catenin ablation results in increased expression of activating NKG2D ligands and decreased expression of inhibitory MHC class I molecules, H2-L and H2-K2. Targeting the b-catenin signaling pathway could not only inhibit tumor growth but also enhance host immune surveillance. Schematic diagram of NK cell-mediated tumor rejection in nude mice. Top: BVE (BVE-Ctnnb1 wt ) tumor cells express high levels of MHC class I molecules H2-L and/or H2-K2, which bind to the inhibitory Ly49 receptor on the NK cells. This interaction results in NK cell inhibition and impaired immune surveillance. Bottom: BVECtn (BVE-Ctnnb1 null ) tumor cells express high levels of NKG2D receptor ligands H60 (a, b, and c) and Raet1 (a, b, c, d, and e), which bind to the activating NKG2D receptor. This binding leads to NK cell activation and tumor cell rejection.

Data availability statement
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ncbi.nlm.nih.gov/geo/ query/acc.cgi?acc=GSE207695.

Ethics statement
The animal study was reviewed and approved by The Animal Care and Use Committee of King Faisal Specialist Hospital and Research Centre.