Background
Necroptosis is a form of programmed necrotic cell death [
1‐
3]. While it is first reported through studying death receptor-induced cell death [
4], necroptosis mostly happens under pathological conditions including viral infection in vivo. For death-receptor-induced necroptosis, the protein kinase receptor-interacting protein kinase 1, 3 (RIPK1, RIPK3) and mixed lineage kinase domain-like protein (MLKL) are the key components of the necroptosis machinery [
1‐
3]. When the activity of cellular inhibitor of apoptosis proteins (cIAPs) and Caspase-8 are inhibited in cells, the engagement of death receptor triggers RIPK1 to recruit RIPK3, which in turn recruits MLKL to form the death complex known as necrosome, to initiate necroptosis [
5‐
10]. In the necrosome, RIPK3 is autophosphorylated and subsequently, the activated RIPK3 recruits and phosphorylates MLKL [
9‐
11]. Then, MLKL oligomerizes and translocates to the plasma membrane to carry out the execution of necroptosis [
12‐
15]. For viral infection-induced necroptosis, RIPK3 and MLKL, but not RIPK1, are required and another protein, Z-DNA-binding protein 1 (ZBP1), also known as DNA-dependent activator of IFN regulatory factors (DAI), functions upstream of RIPK3-MLKL to initiate the formation of the necrosome [
16]. As a necrotic cell death, the rupture of cell plasma membrane of necroptotic cells results in the release of many cellular factors of necroptotic cells, which may trigger inflammation and immune responses [
17,
18]. Importantly, our recent study demonstrated that ZBP1, not RIPK1, mediates tumor necroptosis in breast cancer [
19].
The translocation of the oligomerized, phosphorylated MLKL also leads to the activation of the cell surface proteases, a disintegrin and metalloproteases (ADAMs), which cause ectodomain shedding of many cell surface proteins of necroptotic cells including cadherins, ErbB2, epithelial cellular adhesion molecule (EpCAM) and many others [
20,
21]. The shedding of cell surface proteins by ADAMs promotes necroptosis, cell migration and inflammation [
21]. Notably, the soluble forms of some of these proteins, such as E-cadherin, junctional adhesion molecule A (JAM-A) and syndecan-1, have been shown to play a key role in tumor progression by modulating tumor microenvironments and high levels of these proteins in the serum are an indicator of poor prognosis [
22‐
26]. For example, the circulating level of soluble E-cadherin (sE-cad) is known to be highly increased in cancer patients and is correlated with tumor metastasis [
26,
27]. Also, the soluble E-cadherin is a known ligand of the inhibitory receptor, killer cell lectin-like receptor subfamily G member 1 (KLRG1) on lymphocytes, and this interaction has been suggested to inhibit T cell activity [
28,
29].
In our previous studies, we found that tumor necroptosis promotes lung metastasis [
30], but, however, the underlying mechanism for the promoting effect of necroptosis on tumorigenesis is largely unknown. As it is well documented that the inhibition of the tumor-suppressing functions of T cells plays a key role in tumor metastasis [
31,
32], we investigated whether tumor necroptosis inhibits the anti-tumor functions of T cells in two preclinical mouse breast cancer models, a genetic modified MMTV-PyMT model and a orthotopic transplantation MVT-1 model [
33,
34]. In this study, we demonstrated that the ectodomain shedding of cell surface proteins of necroptotic cells promotes tumor metastasis through inhibiting the anti-tumor activity of T cells. We found that blocking tumor necroptosis by MLKL deletion in both MMTV-PyMT mice and MVT-1 tumors resulted in the dramatic reduction of tumor metastasis to the lungs and the increase in the anti-tumor activity of both tumor-infiltrating and peripheral blood T cells. We found that the levels of soluble cell surface proteins, E-cadherin, JAM-A, syndecan-1, are dramatically reduced in MLKL null tumors/mice. Importantly, we showed that the administration of the ADAMs pan inhibitor in mice with WT tumors reduces the levels of all these three soluble proteins and leads to the increased anti-tumor activity of T cells and the dramatic decrease in metastasis. Finally, we showed that sE-cad/KLRG1 pathway plays a major role in mediating necroptosis-triggered inhibition of the anti-tumor activity of T cells. Hence, our study reveals a novel mechanism of tumor necroptosis-promoted metastasis and suggests that tumor necroptosis is a potential target for controlling metastasis.
Methods
Reagents and antibodies
Anti-phospho-MLKL (ab196436), anti-MLKL (ab184718), anti-GSDMD (ab209845) antibodies from Abcam; anti-cleaved Caspase3 (9664), anti-Caspase3 (9662), anti-cleaved Caspase8 (8592), anti-Caspase8 (4927), anti-cleaved Caspase9 (9509), anti-Caspase9 (9508) antibodies from Cell signaling; β-actin (A3853) from Sigma. Anti-KLRG1 antibody (16–5893-85) and IgG isotype control (16–4914-85) from eBioscience. Anti-CD8 antibody (BP0117) and IgG isotype control (BP0090) from Bioxcell. GW280264X (AOB3632) from AOBIOUS.
Mice
FVB/NJ and MMTV-PyMT mice were purchased from The Jackson Laboratory. All animal experiments were performed under protocols approved by the National Cancer Institute Animal Care and Use Committee and followed NIH guidelines. For orthotopic model, MVT-1 cells (syngeneic mouse mammary cancer cell line derived from c-Myc/VEGF tumor explants) were suspended in 100 µl Matrigel Matrix (Corning) solution (diluted 1:1 with PBS) and then injected (2 × 106/mouse) into the right inguinal mammary fat pad of FVB/NJ mice. Tumor volume was monitored weekly. For Anti-KLRG1 antibody injection, the mice were injected intraperitoneally with Anti-KLRG1 antibody or IgG isotype at 12 weeks until 15 weeks by every 3 days (1 mg/kg). For Anti-CD8 antibody injection, the mice were injected intraperitoneally with Anti-CD8 antibody or IgG isotype at 12 weeks until 15 weeks every 5 days (500ug/mouse/injection). For GW280264X injection, the mice were injected intraperitoneally with vehicle (3% DMSO in PBS) or GW280264X (3ug/mouse/injection) at 12 weeks until 15 weeks every 3 days. To be comparable to the late stage necrotic MMTV-PyMT and MVT-1 tumors, the largest tumor from these models was collected as indicated for further analysis.
Generation of Mlkl knockout and MMTV-PyMT double transgenic mice
The Mlkl knockout mice were generated using the CRISPR/Cas9 method. Briefly, two single guide RNAs (sgRNAs) were designed to target the first coding exon of the mouse Mlkl gene, one (TTGGGACAGATCATCAAGTT) targeting shortly after the translation initiation codon (ATG) and the other one (GCACACGGTTTCCTAGACGC, in reverse orientation) targeting after the only other Met (in frame ATG) in this exon for eliminating its possibility of being used as an alternative translation initiation codon to make a truncated protein. The two sgRNA DNA constructs were made using OriGene’s gRNA Cloning Services (Rockville, Maryland) and were then used as templates to synthesize sgRNAs using MEGAshortscript T7 Kit (Invitrogen). The two sgRNAs (20ug/ml each) were co-microinjected with Cas9 mRNA (100ug/ml, purchased from TriLink BioTechnologies) into the cytoplasm of fertilized mouse eggs, which were collected from mating pairs between male MMTV-PyMT transgenic mice (JAX Stock # 002374) and female FVB/NJ mice (JAX Stock #001800). The injected zygotes were cultured overnight in M16 medium (Millipore Sigma) at 37 °C in 6% CO2. Those embryos reached 2-cell stage of development were implanted into the oviducts of pseudopregnant surrogate mothers. Mice born to these foster mothers were genotyped by PCR amplification of the region surrounding the CRISPR cutting sites followed by Sanger DNA sequencing. Founder mice with frameshift deletions were bred with MMTV-PyMT mice to establish Mlkl KO and MMTV-PyMT double transgenic lines. For use in experiments, MMTV-PyMT/MLKL heterozygous male mice were bred with female MLKL heterozygous mice to obtain littermates MLKL wild type and Mlkl KO carrying MMTV-PyMT. To verify Mlkl genotype by PCR, the following primers were used: Forward 5′-TATGGATAAATTGGGACA-3′, Reverse: 5′-CGTCTTCCTCTGCATCCT-3′.
Generation of Mlkl knockout MVT-1 cells
Mouse mammary cancer cell line MVT-1 was cultured in DMEM containing 10% FBS, 2 mM l-glutamine, 100 U/ml penicillin and 100 mg/ml streptomycin. For targeting Mlkl with CRISPR/Cas9, lentiviral sgRNA vector targeting Mlkl was constructed by ligation of hybridized oligos into lentiCRISPR (pXPR-001, GeCKO) vector: Oligo1 (5′-caccgcgtctaggaaaccgtgtgca-3′) and Oligo2 (5′-aaactgcacacggtttcctagacgc-3′). The lentiCRISPR plasmid (with sgRNA cloned) and packaging plasmids pVSVg (Addgene, #8454) and psPAX2 (Addgene, #12260) were transfected into HEK293T cells for 48 h to generate lentivirus. MVT-1 cells were infected with the lentivirus for 24 h, followed by selection with puromycin (2 ug/ml, Life Technologies) for 5 days. Protein expression was determined by immunoblot analysis.
Cell sorting and FACS
Peripheral blood cells were separated with Lympholyte-Mammal (CL5115, Cedarlane Labs). Tumor-infiltrating lymphocytes (TILs) were isolated through magnetic isolation followed by fluorescence activated cell sorting and FACS. Briefly, tumor tissue was dissociated with gentleMACS Dissociator to get single-cell suspension. Cells were blocked with anti-mouse CD16/32 antibody (Clone 93; BioLegend) and stained APC anti-mouse CD45 (Clone 30-F11; BioLegend). Subsequently, total CD45 + tumor-infiltrating leukocytes were magnetically isolated by incubating with anti-APC magnetic beads (Miltenyi Biotec). Leukocytes were then stained with DAPI (Thermo Fisher) plus PE anti-mouse CD3 antibody (Clone 17A2; BioLegend), PE/Cyanine7 anti-mouse CD4 antibody (clone GK1.5; BioLegend), BB515 Anti-Mouse CD8a (Clone 53-6.7; BD Horizon). CD4 + lymphocytes and CD8+ lymphocytes were sorted from total CD45 + cells by using a BD FACSAria Fusion sorter (BD Biosciences).
For FACS, cells were blocked with anti-mouse CD16/32 antibody and stained with Aqua Dead cell staining dye (Thermo Fisher) plus APC anti-mouse CD45 (Clone 30-F11; BioLegend), PE anti-mouse CD3 antibody (Clone 17A2; BioLegend), FITC anti-mouse CD4 antibody (clone GK1.5; BioLegend), Percp/Cyanine5.5 Anti-Mouse CD8a (Clone 53–6.7; BioLegend). Intracellular cytokines staining was performed using the Fixation/Permeabilization Solution Kit (BD Biosciences), Pacific Blue anti-Granzyme B Antibody (Clone GB11; BioLegend) and PE/Cyanine7 anti-mouse IFNγ antibody (Clone XMG1.2; BioLegend). Flow cytometry was carried out on FACSCanto™ II flow cytometer (BD Biosciences). Data were analyzed using FlowJo v.10.7.1 (Treestar, Ashland, OR). Gating strategies for cell sorting and FACS can be found in Extended Data Figure S2a.
BMDMs active T cells with tumor antigen
Bone marrow–derived macrophages (BMDMs) were differentiated from mouse bone marrow cells with rmM-CSF (10 ng/ml, 416-ML-050/CF, R&D) for 7 days. LLC cells in culture medium (50 × 106 cells/ml) were subjected to 5 rapid freeze–thaw cycles in dry ice and 55℃, and then centrifuged at 5000 rpm to collect supernatant.
BMDMs plated on 24-well plate (2 × 105 cells/well) were treated with LLC cell lysate (12ul/well) for 18 h. Then, freshly isolated splenocytes (1 × 106 cells/well) were added to the BMDMs in fresh DMEM medium with CD40L (0.5 ug/ml, 34-8512-80, Invitrogen) and IL-2 (50 U/ml, 575402, BioLegend). Additionally, the following was used as indicated: rmE-cadherin (10 ug/ml, 8875-EC-050, R&D), anti-E-Cadherin antibody (5 ug/ml, U5885, Sigma-Aldrich), Anti-KLRG1 antibody (5 ug/ml, 16-5893-85, Invitrogen). After 4 days, splenocytes were analyzed by FACS.
Histology analysis
Tumors were bisected into two pieces in the middle of the tumor at the longest diameter orientation using a razor blade: One-half of the tumor was immediately placed in 4% buffered formalin (Z-fix) overnight, and the other half was frozen for protein extraction. The fixed tumors were embedded in paraffin and cut into 5-μm-thick serial sections staining using standard histological procedures. Every 3rd slide was routinely stained with hematoxylin and eosin as described [
35]. Tumor sections from the center of excised tumors of similar size were used for analysis. Tumor necrosis was designated on H&E-stained slides as areas of dark-hematoxylin-stained necrotic tumor cells immediately adjacent to light-hematoxylin-stained viable tissues [
36]. The quantitation of tumor necrotic/death area was counted using Image J and represented as the percentage of tumor necrotic/death area within whole tumor. Mice tumor sections (4 µm) were deparaffinized by incubation at 56 °C for 30 min and subsequent xylene washes then rehydrated with a graded ethanol. The paraffin sections were subjected to antigen retrieval with retrieval buffer (Dako) at 95 °C for 10 min and cooled down until room temperature. The slides were then treated with 3% H
2O
2 for 5 min washed with phosphate-buffered saline (PBS). The slides then were blocked with 2% normal goat serum, followed by overnight incubation with primary antibodies against p-MLKL (1:5000) or cl.Casp-3 (1:1000). Signals were developed using VECTASTIN ABC Elite kit (Vector Laboratories) and DAB Substrate Kit (Vector Laboratories) followed by manufacturer’s instructions. The slides were counter stained with Hematoxylin (Vector Laboratories) to detect nucleus. For the quantitation of metastatic burden, paraffin-embedded lung tissues were sectioned 400 μm apart throughout the whole lung followed by H&E staining. The frequency of the metastatic foci was counted manually in a blinded fashion.
Western blot
Tumor tissues and in vitro cultured cells were lysed in RIPA buffer. Tumor lysates were separated by SDS-PAGE, followed by probing with anti-mouse MLKL antibody (Clone EPR17514; Abcam), anti-Actin antibody. Signals were developed by using enhanced chemiluminescence kit (Bio-Rad).
Quantitative RT-PCR
The sorted T cells were subjected to RNA extraction using RNeasy Mini Kit (Qiagen). cDNA synthesis was conducted using PrimeScript RT reagent Kit (Roche, RR037A). Predesigned primers and probes for β-Actin, IFN-γ, T-bet (Integrated DNA Technologies) were used for qPCR Assays, and relative mRNA expression was measured using SensiFAST Probe Hi-ROX Mix (Bioline) on QuantStudio 3 Real-Time PCR System (Thermo Fisher Scientific). 2−ΔΔCT method was used to quantify fold induction, and each cDNA was normalized by β-Actin expression.
ELISA
For serum samples, blood samples were clotted for 2 h at room temperature and centrifuged at 2000g for 20 min. For TIF, tumor tissue was put on 10 µm Nylon Net filter (NY1002500, Sigma), centrifuged at 60 g for 10 min and collect fluid. Samples were then analyzed for mouse JAM-A (ab277080, Abcam), mouse CD138 (NBP2-76610, Novus) and mouse E-cadherin (ab197751, Abcam) according to manufacturer’s instruction.
The soft agar colony formation assay was performed as previously described [
37]. 12-well plate was coated with a 1 ml base layer containing 1% agarose (Lonza 50111 SeaPlaque GTG Agarose). Peripheral blood was first depleted of red blood cells (RBCs) by ACK lysing buffer (118–156-101, Quality Biological). After washing and centrifugation, cell pellets were resuspended in 600ul DMEM containing 10% FBS, 2 mM l-glutamine, 100 U/ml penicillin, 100 mg/ml streptomycin and 0.5% agarose at 37 °C and added to 12-well plates. After solidified at room temperature, 500ul DMEM was added containing 10% FBS, 2 mM l-glutamine, 100 U/ml penicillin and 100 mg/ml streptomycin and cultured at 37 °C with 5% CO
2 for 2 weeks. Colonies were stained with crystal violet and quantified.
Statistical analysis
All data were analyzed with the GraphPad Prism 8 software. Student’s t test was used to determine the statistical significance of differences between groups. Differences with p values < 0.05 were considered significant.
Discussion
Tumor metastasis is the main cause of cancer mortality and is a complex process modulated by many different factors including the interplays between tumor cells and host immunity [
47,
48]. Particularly, host T cells play a critical role in controlling metastasis and can exert their anti-tumor activity in tumors as tumor-infiltrating T cells or attack circulating tumor cells peripherally as peripheral T cells [
31]. Inhibiting the anti-tumor activity of T cells is a major feature of the interplay of tumor cells and host immunity during tumor development and metastasis [
31,
32]. In the current study, we found that the ectodomain shedding of cell surface proteins of necroptotic tumor cells leads to the accumulation of soluble cell surface proteins in tumor microenvironments, resulting in the inhibition of the anti-tumor activity of tumor-infiltrating and peripheral T cells and the promotion of tumor metastasis.
Although evading apoptosis is one of the hallmarks of cancer, tumor cell death happens during tumor development. Foci of cell death, referred as tumor necrosis, are commonly observed in core regions of solid tumors as a result of inadequate vascularization and subsequent metabolic stresses such as hypoxia and nutrient deprivation [
49,
50]. Tumor necrosis is often associated with aggressive tumor development and metastasis [
51,
52]. Unlike apoptosis in which cells have intact membranes and are rapidly removed by host macrophages, tumor necrosis often results in the release of cellular components to the tumor microenvironment and may lead to the alteration of the tumor microenvironment [
17,
53]. We recently demonstrated that necroptosis is the major form of cell death that causes tumor necrosis in MVT-1 breast cancer and blocking of necroptosis dramatically abolished lung metastasis [
19,
30]. On the other hand, apoptosis is accounted for the remaining tumor cell death in MLKL KO tumors [
19,
30]. We found that both intrinsic and extrinsic apoptotic pathways are engaged in MLKL KO tumors as both cleaved Casp-8 and Casp-9 are detected (Additional file
1: Fig. S5i). Some papers show that pyroptosis also involves in the alteration of the tumor microenvironment [
54]. However, we did not observe pyroptosis in MMTV-PyMT tumors (Additional file
1: Fig. S5i).
In the current study with the genetically modified MMTV-PyMT breast cancer model, we have confirmed that blocking necroptosis by deletion of the key necroptosis effector protein, MLKL, dramatically inhibits lung metastasis, but has no or little effect on tumor initiation and growth at the stages of tumor development before necroptosis occurs. More importantly, we found that blocking necroptosis significantly elevated the anti-tumor activity of both tumor-infiltrating and peripheral blood T cells in both MMTV-PyMT and MVT-1 breast cancer models. Our findings suggest that necroptosis has a suppressive effect on the anti-tumor activity of T cells in breast cancer. Additionally, necroptosis may promote the survival of circulating tumor cells (CTCs) through inhibiting the activity of peripheral blood T cells in MVT-1 breast cancer model (Additional file
1: Fig. S5j).
While recent studies reported that necroptosis plays an important role in regulating tumor immunity [
38‐
40], the exact role of necroptosis in tumor development may need to be further evaluated in different types of tumors and under different conditions, as current studies suggest that chronic and spontaneous necroptosis may promote tumor development due to its suppressive effect on anti-tumor immunity in certain types of cancer and that acute and massive induction of necroptosis by chemotherapy or radiation treatment may undermine tumor growth and triggers immunogenic responses [
55]. However, the underlying mechanisms of necroptosis on tumor immunity remain elusive [
56,
57]. Recently, we showed that necroptosis triggers the activation of cell surface proteases, ADAMs, which lead to the shedding and the release of cell surface proteins to become soluble proteins [
21]. As several of the soluble proteins, sE-cad, JAM-A and CD138, are known to promote metastasis [
22‐
24], we examined the levels of these proteins in TIFs and serum from WT and Mlkl KO tumors or mice and found that there is a marked increase in the levels of these proteins in both TIFs and serum from WT tumors or mice bearing WT tumors. Particularly, administration of ADAM inhibitor, GW280264X, dramatically reduced the levels of these molecules in MMTV-PyMT WT mice. More importantly, GW280264X led to the elevated anti-tumor activity of both tumor-infiltrating and peripheral T cells and the marked reduction of lung metastasis (Fig.
4). These findings suggest that necroptosis-mediated shedding of the surface proteins of tumor cells play a key role in inhibiting the anti-tumor activity of T cells and promoting metastasis.
Previous reports suggest that one of these proteins, sE-cad, functions as a ligand for the co-inhibitory receptor, KLRG1 [
29]. As the well-known co-inhibitory factor of immune response, PD-1, does in regulating T cell activity, KLRG1 also plays a role in keeping T cell activation in check [
43,
44]. While the expression levels of KLRG1 and PD-1 in T cells are normally increased during viral, bacterial or parasite infections [
58‐
60], we found that the levels of KLRG1 and PD-1 in T cells remain the same regardless of MLKL status in MMTV-PyMT model (Extended Data. Figure
5). However, as the ligands of KLRG1 or PD-1, respectively, we found that the level of sE-cad, not PD-L1, is significantly increased in TIFs and serum of WT tumors or mice of MMTV-PyMT model (Fig.
5). Importantly, neutralizing KLRG1 with antibody significantly elevated the anti-tumor activity of tumor-infiltrating and peripheral T cells and dramatically reduced lung metastasis in MMTV-PyMT model (Fig.
5). Because the effect of KLRG1 neutralization on T cell activity and lung metastasis is comparable to that as due to MLKL deletion or ADAM inhibition, KLRG1, not the PD-1, pathway likely plays a major role in necroptosis-mediated inhibition of T cell anti-tumor activity and the promoting effect on metastasis.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.