CDK4/6 Inhibition Suppresses p73 Phosphorylation and Activates DR5 to Potentiate Chemotherapy and Immune Checkpoint Blockade

Cyclin-dependent kinases (CDK) are serine/threonine kinases that play a key role in cell cycle regulation. CDKs 4 and 6 (CDK4/6) bind to D-type cyclins to form active kinases, which phosphorylate retinoblastoma 1 (RB1) to relieve its inhibition on E2F transcription factors leading to G₁–S transition (1). A hallmark of cancer is sustained cell proliferation due to abnormal cell cycle regulation (2). The CDK4/6-cyclin D-RB1-E2F axis is one of the most frequently dysregulated cell cycle pathways in cancer cells (3).

Pharmacologic inhibition of CDK4/6 has emerged as a promising anticancer strategy. To date, three small-molecule CDK4/6 inhibitors (CDK4/6i), including palbociclib, ribociclib, and abemaciclib, have been approved for the treatment of hormone receptor (HR)–positive and advanced breast cancer (4). CDK4/6i have several anticancer effects including inhibition of the cell cycle and modulation of cellular senescence, tumor cell metabolism, and tumor microenvironment (5). Recent studies have shown that CDK4/6 inhibition triggers antitumor immunity and potentiates anti–PD-1 immunotherapy (6, 7). Despite these studies, the anticancer mechanisms of CDK4/6i in different types of cancer cells, the molecular markers for predicting response and resistance to CDK4/6i, and the rationale for combining CDK4/6i with other anticancer agents, have yet to be determined.

CDK4/6i combined with other anticancer agents enhance apoptosis of cancer cells (8). Nonetheless, it is unclear how CDK4/6i promote cell death, and if such an effect is essential for the anticancer and immunogenic effects of CDK4/6i. Apoptosis is initiated via the extrinsic and/or intrinsic pathways. The extrinsic pathway is engaged upon activation of the tumor necrosis factor (TNF) family receptors such as death receptor 5 (DR5; TRAILR2), leading to activation of caspase-8 and effector caspases (9). DR5 is a receptor of immune cytokine TRAIL (10). DR5 can also be transcriptionally activated by p53, TAp73 (p73 thereafter), NF-κB, and CAAT/enhancer binding protein (C/EBP) homologous protein (CHOP) in response to various stresses (11, 12). The intrinsic pathway is controlled by the Bcl-2 proteins via mitochondrial dysfunction (13, 14). Different forms of cell death have distinct immunologic consequences (15, 16). For example, DR5-mediated apoptosis has characteristics of immunogenic cell death (ICD; ref. 10), which stimulates an immune response against dead-cell antigens (15, 16).

CDK4/6i have shown promising activity against different cancer types including colorectal cancer (17), a leading cause of cancer-related deaths in the United States (18). Patients with colorectal cancer are treated with 5-fluorouracil (5-FU)–based chemotherapy, targeted therapy, and recently, anti–PD-1 immunotherapy (19, 20). In this study, we analyzed the effects of CDK4/6i in colorectal cancer cells and found that CDK4/6 inhibition leads to induction of DR5 via a novel p73-dependent, but p53- and RB1-independent mechanism. Our results suggest a critical role of DR5 induction in mediating the antitumor and immunogenic effects of CDK4/6i in colorectal cancer cells.

Parental cell lines, including human colorectal cancer cell lines HCT116, RKO, DLD1, HT29, and LoVo, human breast cancer cell lines MCF7, T47D, and MDA-MB-231, and mouse colorectal cancer cell line CT26, were purchased from the ATCC. Lim2405 was a gift from Dr. Alberto Bardelli (University of Torino, Turin, Italy). NCM356 nontransformed human colonic epithelial cells were from INCELL. Isogenic derivative cell lines, including DR5-knockout (KO) HCT116, DLD1, RKO, and CT26, and p73-KO HCT116 and DLD1, were generated by homologous recombination or CRISPR/Cas9 as described in Supplementary Methods. FADD-KO, Bid-KO, Caspase-8–knockdown (KD), and Flag-p73 knock-in (Flag-KI) HCT116 cell lines were previously described (21, 22). p53-KO HCT116 was from Dr. Bert Vogelstein (Johns Hopkins University, Baltimore, MD).

Parental cancer cell lines were authenticated by analysis of known oncogenic drivers including DNA mismatch repair (MMR) status and p53, KRAS, BRAF, PIK3CA, FBW7 and other mutations (23–25). Breast cancer cell lines were also verified by the expression of estrogen receptor (ER), progesterone receptor (PR), and/or HER2 (25). NCM356 was verified by the normal function of APC tumor suppressor (26). Isogenic derivative cell lines were authenticated by sequencing of the targeted genomic regions and analysis of protein expression. All cell lines were routinely checked for Mycoplasma contamination by PCR. Cell lines with fewer than 15 passages from the original stocks were used.

Cell lines, chemicals, cell culture media, and supplements are listed in Supplementary Table S1. All cell lines were maintained at 37°C and 5% CO₂ atmosphere, and cultured in McCoy's 5A Modified Media, except for NCM356, which was cultured in M3 media (INCELL). All media were supplemented with 10% defined FBS, 100 units/mL penicillin, and 100 μg/mL streptomycin. For drug treatment, cells were plated in 12-well plates at 20% to 30% density 24 hours before treatment. DMSO stocks of agents used, including palbociclib (Apexbio) and ribociclib (Apexbio), were diluted to appropriate concentrations with cell culture media. TRAIL (Pepro Tech) was diluted with distilled water.

MTS assays were performed using the MTS Assay Kit (Promega) according to the manufacturer's instructions. Chemiluminescence was measured by a Wallac Victor 1420 Multilabel Counter (Perkin Elmer). Apoptosis was analyzed by staining adherent and floating cells with Hoechst 33258 (Invitrogen) and counting cells with condensed and fragmented nuclei (27). A minimum of 300 cells were counted for each sample. Apoptosis was also analyzed through Annexin V/propidium iodide (PI; Invitrogen) staining followed by flow cytometry as described (27). Long-term cell survival was analyzed by colony formation assays as described (27). Each assay was conducted in triplicate and repeated three times.

Expression constructs and siRNAs are described in Supplementary Methods and Supplementary Table S1. Transfection of expression constructs and siRNA was performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. siRNA transfection was performed 24 hours prior to drug treatment using 200 pmole of siRNA.

Purified recombinant proteins, including human CDK4, CDK6, TRAIL, wild-type (WT) p73, and T86A p73, are described in Supplementary Methods. In vitro kinase assay was performed as described in Supplementary Methods.

Western blotting was performed as previously described (28). Antibodies include those for DR5 (ProSci); DR4 (EMD Millipore); CDK4; CDK6; FLIP; FADD; FAS; phospho-RB1 (p-RB1; Ser780); p53; hemagglutinin (HA); Lamin A/C; p65 (Santa Cruz); RB1; Flag; p-p65 (Ser536); CHOP; cleaved caspases-3, 8, 9; Bid (Cell Signaling Technology), p73 (Bethyl), p-p73 (Thr86) (Affinity Biosciences), Noxa, p21, β-actin (Sigma), V5, phospho-Thr (Invitrogen), p16 (Abcam) and DNA methyltransferase 1 (DNMT1; Novus). Additional information on antibodies is in Supplementary Table S1.

Total RNA was isolated from cells using the Mini RNA Isolation II kit (ZYMO Research) according to the manufacturer's protocol. Total RNA (1 μg) was used to generate cDNA using SuperScript II reverse transcriptase (Invitrogen). PCR was performed using previously described conditions (27) and primers listed in Supplementary Table S2.

Targeting DR5 in HCT116 cells by homologous recombination, and DR5 in DLD1 and RKO cells, p73 in HCT116 and DLD1 cells, and DR5 in CT26 cells by CRISPR/Cas9 are described in Supplementary Methods.

All animal experiments were approved by the University of Pittsburgh (Pittsburgh, PA) Institutional Animal Care and Use Committee. Mice were housed in a sterile environment with micro isolator cages and allowed access to water and chow ad libitum. Cell line xenografts were established by subcutaneously injecting 4 × 10⁶ WT or DR5-KO HCT116 cells into both flanks of 5- to 6-week-old female Nu/Nu mice (Charles River).

Patient-derived xenograft (PDX) tumors were propagated in 5- to 6-week-old female NOD.Cg-Prkdc^scidIl2rg^tm1Wjl/SzJ (NOD/SCID gamma; NSG) mice (Jackson Laboratory). For PDX passage, tumor tissues were cut into 25-mg pieces and implanted subcutaneously into both flanks of NSG mice as described (29, 30). PDX models analyzed include PDX1 and PDX2, which were established using treatment-naïve primary tumors resected from patients with newly diagnosed colorectal cancers under written informed consent (29). The tissue samples and deidentified clinical information were obtained from the Biospecimen Core at the University of Pittsburgh. PDX1 was from a KRAS-mutant (G13D), NRAS-mutant (G12D), FBW7-mutant (R505C), and MMR-proficient tumor (T4N0M1) in the sigmoid colon of a 77-year-old male (29, 30). PDX2 was from an MMR-proficient tumor (T2N0) in the right colon of a 69-year-old female. Tumors passaged for two generations (P2) were analyzed.

Cell line xenograft or PDX tumors reached approximately 60 mm³ in size before treatment. Mice bearing tumors were randomized into different groups and treated with palbociclib (oral gavage; 150 mg/kg daily), TRAIL (100 μg/dose every other day i.p.), 5-FU (25 mg/kg every other day i.p.), or a combination of palbociclib with TRAIL or 5-FU for 10 days. Palbociclib was dissolved in sodium lactate buffer (pH 4.0). Tumor growth was monitored by calipers and tumor volume was calculated according to the formula 1/2 × length × width². Ethical endpoint represents a timepoint when tumors reached 2 cm or more in any dimension.

Tumor tissues were dissected and fixed in 10% formalin and embedded in paraffin. Immunostaining was performed for terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL), active caspase-3, active caspase-8, and Ki67 on 5-μm paraffin-embedded sections using kit and primary antibodies listed in Supplementary Table S1 as described (27). Signals were detected using AlexaFluor 488-conjugated secondary antibody (Invitrogen) with nuclear counter staining by 4′6-Diamidino-2-phenylindole (DAPI).

Induction of ICD in WT, p73-KO, and DR5-KO HCT116 cells treated with 1 μmol/L palbociclib or ribociclib for 24 hours was determined by analyzing cell-surface calreticulin (CRT) and phagocytosis of tumor cells by dendritic cells (DC) as described in Supplementary Methods.

Syngeneic tumors were established by subcutaneously injecting CT26 cells into the flanks of 5- to 6-week-old BABL/cJ mice (Jackson Laboratory). Each mouse was inoculated with 5 × 10⁵ WT or DR5-KO CT26 cells to establish a single tumor. Tumor-bearing mice were randomized and treated with palbociclib (oral gavage; 100 mg/kg), PD-1 blockade by InVivoPlus anti-mouse PD-1 (BioXcell; 200 μg/dose i.p.), or their combination as indicated in Fig. 8A. Some mice also received CD8 blockade by InVivoPlus anti-mouse CD8α (BioXcell; 400 μg/dose i.p.) as in Fig. 8A. Tumor growth and tumor volume were analyzed as done for cell line and PDX models. Tumor-infiltrating lymphocytes were analyzed by immunostaining followed by flow cytometry on tumors dissected 24 hours after the third treatment as described in Supplementary Methods. Antibodies included those for CD3, CD4, CD8α, CD11c, CD25, CD45, CD86, FoxP3, Granzyme B (GzmB), IFNγ, and MHC class II (MHCII; BioLegend). Tumor sections were also stained using antibodies for CD3 (Thermo Fisher Scientific) and CD8α (BioLegend). Additional information on antibodies is in Supplementary Table S1.

Statistical analysis was performed using Prism 9 software (GraphPad). For cell culture and immunostaining experiments, P values were calculated using Student t test. Means ± SD are indicated in the figures. For animal treatment experiments, P values were calculated by ANOVA with Fisher least significant difference (LSD) posthoc test for tumor volume analysis, and by log-rank test for survival analysis. Means ± SEM are indicated in the figures. Differences were considered significant if P < 0.05.

We analyzed RNA sequencing (RNA-seq) data from HCT116 colorectal cancer cells transfected with CDK4 siRNA or treated with palbociclib. Upregulation of DR5 was identified as a prominent and consistent change in the apoptotic pathways (Supplementary Fig. S1A–S1D). DR5 mRNA and protein were markedly induced in HCT116 cells upon CDK4/6 KD (Fig. 1A and B), or inhibition by palbociclib or ribociclib (Fig. 1C–E). In contrast, other apoptosis regulators were unchanged or only slightly induced (Fig. 1A and C; Supplementary Fig. S2A). Induction of DR5 by CDK4/6 KD or inhibition was observed in colorectal cancer cell lines with WT or mutant p53 (Fig. 1F and G; Supplementary Fig. S2B), and in MCF7, MDA-MB-231, and T47D breast cancer cells (Supplementary Fig. S2C–S2E). Interestingly, DR5 was not induced by palbociclib or ribociclib in NCM356 normal colonic epithelial cells (Supplementary Fig. S2F). CDK4/6 inhibition suppressed RB1 phosphorylation (Ser780) and downregulated RB1 (Fig. 1H). However, RB1 KD did not induce DR5 or affect DR5 induction and G₁ cycle arrest following palbociclib treatment (Fig. 1I; Supplementary Fig. S2G). Furthermore, ectopic expression of p16 in HCT116 cells induced DR5 (Supplementary Fig. S2H). CDK4/6 inhibition downregulated DNMT1 (Supplementary Fig. S2I; ref. 6), but did not affect CpG methylation in the DR5 promoter (Supplementary Fig. S2J). These results suggest that DR5 induction by CDK4/6i in colorectal cancer cells is mediated by transcriptional activation independent of p53, RB1, and DNMT1.

We then investigated the transcription factor responsible for DR5 induction. CDK4/6 KD or inhibition did not affect the levels of p53, p73, p65, p16, and CHOP (Supplementary Fig. S3A and S3B). DR5 induction was abrogated in p73-KO HCT116 cells (Fig. 2A–C), p73-KO DLD1 cells (Supplementary Fig. S3C and S3D), and p73-KD RKO cells (Supplementary Fig. S3E). In contrast, DR5 induction was reduced in p53-KO cells (Fig. 2B) and remained intact in p65-KD and CHOP-KD cells.

We then analyzed luciferase reporters containing different regions of the DR5 promoter (Supplementary Fig. S3F). CDK4/6 KD or inhibition strongly activated the DR5 reporters containing the p53 binding site (fragment A and C; Supplementary Fig. S3F) in a p73-dependent manner (Fig. 2D and E). Transfection of exogeneous p73 also induced DR5 and activated the DR5 promoter (Supplementary Fig. S3G and S3H). These effects were abolished by mutating key residues in the p53 binding site (Fig. 2F; Supplementary Fig. S3F and S3H). CDK4/6 inhibition also upregulated p73 targets p21 and Noxa in a p73-dependent manner (Supplementary Fig. S3I). Chromatin immunoprecipitation was performed by using HCT116 cells with Flag-KI (22). CDK4/6 KD or inhibition markedly enhanced the binding of p73 to the DR5 promoter (Fig. 2G). p73 is regulated by nuclear import to transcriptionally activate p53 targets (31, 32). CDK4/6 inhibition increased nuclear accumulation of p73 in Flag-KI cells (Fig. 2H and I). These results demonstrate that p73 directly binds to the p53 binding site in the DR5 promoter to activate its transcription upon CDK4/6 inhibition.

p73 subcellular localization and activity are regulated by posttranslational modifications such as tyrosine (Tyr) and threonine (Thr) phosphorylation (33, 34). CDK4/6 inhibition had no effect on the most common p73 modification, Tyr99 phosphorylation (p-Y99; ref. 33), which was undetectable in HCT116 cells (Supplementary Fig. S4A). In contrast, CDK4/6 KD or inhibition abrogated Thr phosphorylation of p73 in Flag-KI cells (Fig. 3A and B). Transfection of WT CDK4 or CDK6, but not dominant negative (DN) mutants, enhanced Thr phosphorylation (Fig. 3C; Supplementary Fig. S4B), which was suppressed by palbociclib treatment (Fig. 3D; Supplementary Fig. S4C). CDK4/6 transfection also retarded migration of Flag-p73 in gel electrophoresis (Fig. 3E), which was inhibited by λ protein phosphatase (λPPase) or palbociclib treatment (Fig. 3E and F). Immunoprecipitation (IP) analysis detected direct binding of CDK4 to endogenous p73 in Flag-KI cells (Supplementary Fig. S4D), which was markedly enhanced upon palbociclib or ribociclib treatment (Fig. 3G). In an in vitro kinase assay, incubating recombinant human CDK4 or CDK6 with purified p73 (Supplementary Fig. S4E) strongly stimulated kinase reaction (Fig. 3H).

We then determined the specific Thr residue of p73 that is phosphorylated by CDK4/6. Thr86 of p73 was shown to undergo CDK1/2-mediated phosphorylation (35, 36). We therefore analyzed Thr86 by testing phospho-dead T86A and phospho-mimic T86D mutants. In contrast to WT p73, CDK4/6 transfection had no effect on Thr phosphorylation and gel migration of T86A expressed in p73-KO cells (Fig. 3I and J; Supplementary Fig. S4F). Incubating CDK4 or CDK6 with purified T86A protein did not stimulate kinase reaction (Fig. 3H). Compared with WT p73, T86A mutant had increased nuclear accumulation and more strongly induced DR5 (Fig. 3K and L). Conversely, T86D mutant was defective in nuclear accumulation and DR5 induction (Fig. 3K and L), resistant to CDK4/6 inhibition (Supplementary Fig. S4G), and unable to restore DR5 induction in p73-KO cells (Supplementary Fig. S4H). Together, these results demonstrate that CDK4/6 inhibition suppresses Thr86 phosphorylation and leads to nuclear translocation of p73, which activates DR5 transcription.

We tested if CDK4/6i sensitize colorectal cancer cells to the DR5 ligand TRAIL and other anticancer agents. CDK4/6 KD or inhibition markedly potentiated HCT116 and DLD1 cells to TRAIL, as shown by a lower IC₅₀ and enhanced growth suppression, apoptosis induction, and caspase activation (Supplementary Fig. S5A–S5H). Similar effects of CDK4/6i were observed in breast cancer cells (Supplementary Fig. S5I–S5K), but not in NCM356 cells (Supplementary Fig. S5L). CDK4/6i also sensitized colorectal cancer cells to commonly used anticancer drugs, including 5-FU, oxaliplatin, CPT-11, cetuximab, and cisplatin, with enhanced growth suppression, apoptosis induction, and caspase activation (Supplementary Fig. S6A–S6F).

DR5 KO by homologous recombination in HCT116 cells abolished the potentiating effects of CDK4/6i on growth suppression, apoptosis induction, caspase activation, and inhibition of colony formation by TRAIL (Fig. 4A–D; Supplementary Fig. S7A–S7D) and by 5-FU (Fig. 4E–H; Supplementary Fig. S7E). DR5 KO in DLD1 and RKO cells by CRISPR/Cas9 confirmed the requirement of DR5 for the potentiating effects of CDK4/6i (Supplementary Fig. S7F–S7I). Furthermore, perturbation of DR5 downstream apoptosis regulators, including FADD, caspase-8, and Bid, suppressed the potentiating effects of CDK4/6i on TRAIL and 5-FU (Supplementary Fig. S8A–S8F). In contrast, DR5 KO did not affect the inhibition of RB1 phosphorylation by CDK4/6i (Supplementary Fig. S8G), and RB1 KD did not change TRAIL sensitivity (Supplementary Fig. S8H). Together, these findings demonstrate a critical role of DR5-dependent apoptosis in mediating the anti–colorectal cancer activity of CDK4/6i.

We then analyzed palbociclib-induced therapeutic sensitization in vivo. WT and DR5-KO HCT116 xenograft tumors growing in athymic nude mice were treated with palbociclib alone or in combination with TRAIL or 5-FU. At relatively low doses used, palbociclib or TRAIL alone had little effect on WT HCT116 xenografts (Fig. 5A), despite DR5 induction and inhibition of p73 Thr86 phosphorylation by palbociclib (Fig. 5B). Combining palbociclib and TRAIL more strongly inhibited tumor growth than palbociclib alone (P < 0.01, ANOVA with Fisher LSD posthoc test; Fig. 5A) and enhanced TUNEL and active caspase-3 and 8 immunostaining (P < 0.001, two tailed Student t test; Fig. 5C; Supplementary Fig. S9A and S9B). Similarly, combining palbociclib and 5-FU at nontoxic doses significantly improved antitumor efficacy (Fig. 5D), and enhanced DR5 induction (Fig. 5E) and tumor cell apoptosis (Fig. 5F; Supplementary Fig. S9C and S9D). Importantly the antitumor and apoptotic effects of palbociclib/TRAIL and palbociclib/5-FU combinations were abolished in DR5-KO tumors (Fig. 5A–F; Supplementary Fig. S9A–S9D). We also generated stable p73-KO HCT116 cells with reconstituted WT or T86D mutant p73 (Supplementary Fig. S10A). p73 KO phenocopied DR5 KO in suppressing the antitumor and apoptotic effects of palbociclib combined with TRAIL (Supplementary Fig. S10B–S10E). Reintroducing WT p73, but not T86D mutant, restored these effects in p73-KO tumors (Supplementary Fig. S10B–S10E).

We further analyzed the antitumor effects of CDK4/6i on 2 PDX models established from treatment-naïve primary colorectal cancer tumors, including PDX1 (T4N0M1; KRAS-G13D; NRAS-G12D; MMR-proficient) and PDX2 (T2N0; MMR-proficient; ref. 29). Palbociclib combined with TRAIL, but not either agent alone, markedly suppressed the growth of PDX1 and PDX2 tumors without affecting animal weight (Fig. 6A and B; Supplementary Fig. S11A). PDX growth inhibition was associated with enhanced tumor cell loss, DR5 induction, p73 Thr86 dephosphorylation, apoptosis induction, and proliferation inhibition (Fig. 6C–E; Supplementary Fig. S11B and S11C). Combining palbociclib and 5-FU also strongly suppressed PDX1 and PDX2 growth and enhanced DR5 induction, apoptosis, and proliferation inhibition (Fig. 6F–J; Supplementary Fig. S11D–S11F). These results suggest that the in vivo antitumor effects of CDK4/6i are mediated by p73 and DR5 via enhanced apoptosis.

The induction of DR5 by CDK4/6i prompted us to test if the immunogenic effects of CDK4/6i are mediated by ICD, which is characterized by cell-surface exposure of the ER chaperone CRT and DC maturation (15, 16). Treating HCT116 cells with palbociclib or ribociclib markedly increased cell-surface CRT, which was blocked in DR5-KO and p73-KO cells (Fig. 7A and B). To test if CDK4/6i-induced dying colorectal cancer cells can be recognized and phagocytized by DCs, we labelled DCs differentiated from peripheral blood mononuclear cells of healthy donors with Far Red (red), and palbociclib- or ribociclib-treated HCT116 cells with carboxyfluorescein succinimidyl ester (CFSE; green). Upon coculturing labelled DCs and HCT116 cells, we detected by fluorescence microscopy and flow cytometry a significant increase in double positive fused cells following palbociclib or ribociclib treatment, which was suppressed by DR5 or p73 KO (Fig. 7C–E). These results suggest a key role of p73-dependent DR5 induction in mediating the immunogenic effects of CDK4/6i.

We then used immuno-competent CT26-BALB/cJ syngeneic tumor model to determine if CDK4/6i enhance anti–PD-1 immunotherapy through DR5-dependent ICD. BALB/cJ mice bearing WT or isogenic DR5-KO (generated by CRISPR/Cas9) CT26 tumors were treated with anti-mouse PD-1 antibody at a previously described dose (200 μg/dose i.p.) and schedule (3× every 2 days followed by 2× weekly; Fig. 8A; ref. 37), with or without combination with palbociclib at a reduced dose (100 mg/kg) and frequency (3 days on, 4 days off; Fig. 8A). Although either agent alone had little or no efficacy, the combination group showed striking therapeutic efficacy with complete inhibition of WT CT26 tumor growth in 4 out of 8 mice at the endpoint on day 28 (Fig. 8B; Supplementary Fig. S12A), along with significant extension of animal survival (Fig. 8C). The combination was well tolerated without affecting body weight (Supplementary Fig. S12B). In contrast, the therapeutic effect was almost completely lost in DR5-KO CT26 tumors (Fig. 8B and C; Supplementary Fig. S12A).

The combination treatment significantly increased infiltration of CD3⁺ and CD8⁺ T cells, and decreased infiltration of CD25⁺/FoxP3⁺ immunosuppressive regulatory T cells (Tregs) in WT tumors (Fig. 8D–G; Supplementary Fig. S12C–S12F). The treatment also increased GzmB and IFNγ production by tumor-infiltrating CD8⁺ T cells (Fig. 8H and I), and expanded infiltrating CD11c⁺ DCs expressing increased CD86 and MHCII (Fig. 8J–L). Remarkably, all of the observed changes were suppressed in DR5-KO CT26 tumors (Fig. 8D–L), indicating a critical role of DR5-dependent cell death in mediating the antitumor immune response to this combination. Furthermore, depleting CD8⁺ cytotoxic T cells (Fig. 8A) abolished the effects of combination treatment on WT tumors (Fig. 8M), confirming the role of CD8⁺ T cells in tumor suppression.

We identified p73-dependent and RB1-independent induction of DR5 as a novel antitumor mechanism of CDK4/6i in colorectal cancer cells. CDK4/6 inhibition suppresses p73 Thr86 phosphorylation, which leads to p73 nuclear translocation and increased DR5 promoter binding and transcription. Thr86 of p73 is highly conserved among different species. Thr86 phosphorylation is mediated by CDK1/2 in lung cancer cells (35) but by CDK4/6 in colorectal cancer cells. A noncanonical function of CDKs seems to regulate p73 Thr86 phosphorylation in cancer cells (38). DR5 induction by CDK4/6i is largely p53-independent, reflecting the distinct functions of p73 and p53 in regulating common target genes. However, p53 seems to impact DR5 induction (Fig. 2B), likely via its effects on p73 (39). Interestingly, our recent study showed that CDK4/6 inhibition in mice protects against radiation-induced intestinal injury by abolishing p53 and PUMA induction (40). The distinct functional roles and complex interplay of p53, p73, and other p53 family members in coordinating response to CDK4/6i in cancer and normal cells remain to be further elucidated. Some of our mechanistic experiments are limited by a single cell line and need to be repeated using other cell lines.

Our results suggest that rational combinations of CDK4/6i with other anticancer agents can markedly improve therapeutic efficacy through DR5-dependent ICD. The striking therapeutic effects of CDK4/6i combined with TRAIL via DR5 induction provide a compelling rationale for developing the TRAIL-CDK4/6i combination for treating advanced and therapy-resistant colorectal cancers. DR5 induction mediates the enhanced apoptotic and antitumor effects of combining 5-FU and CDK4/6i in colorectal cancer cells. DR5 induction may lower the apoptotic threshold to sensitize colorectal cancer cells to different stimuli, including TRAIL, chemotherapy, or activated immune cells. The enhanced DR5 induction in colorectal cancer cells treated by palbociclib combined with 5-FU may be explained by simultaneous activation of DR5 transcription by p73, p53, NF-κB, CHOP, and/or other transcription factors. DR5 induction may be a useful indicator for therapeutic efficacy and effective drug combinations.

Recent studies have demonstrated multiple effects of CDK4/6i in antitumor immunity, such as activating innate immunity by inducing endogenous retroviral elements, intracellular double-stranded RNA, and IFN response (6), upregulating PD-L1 by dephosphorylating the E3 ubiquitin ligase SPOP (7), and stimulating T-cell activation by derepressing NFAT family proteins and targets (41). As DR5 functions as a nodule point of different apoptotic pathways, including cytokine-induced apoptosis in immune response, these immunogenic effects of CDK4/6i may converge and culminate in DR5 induction required for robust killing of tumor cells. DR5-mediated tumor cell death may function as a key link between innate and adaptive immune responses through increased phagocytosis by DCs.

The clinical benefit of immune checkpoint inhibitors (ICI) is limited to 10% to 15% of colorectal cancers that are MMR-deficient (42). A burning issue is to convert immunologically “cold” tumors into “hot” tumors (43). CDK4/6 inhibition represents a promising strategy for targeting immune escape mediated by aberrant CDK signaling in cancer cells, especially for those that are insensitive to ICIs. DR5 induction may be useful for predicting effective combinations and responsiveness to ICI combinations.

Collectively, our results reveal a novel, on-target mechanism of CDK4/6i through p73-dependent DR5 induction, which is crucial for the apoptotic, chemo-sensitizing, and immunogenic effects of CDK4/6i in colorectal cancer cells.

Y. Huang reports grants from NCI during the conduct of the study. J. Yu reports grants from NIH during the conduct of the study and grants from NIH outside the submitted work. L. Zhang reports grants from NIH during the conduct of the study. No disclosures were reported by the other authors.

J. Tong: Conceptualization, data curation, formal analysis, validation, investigation, visualization, methodology, writing–original draft, writing–review and editing. X. Tan: Data curation, investigation, methodology. X. Song: Data curation, investigation. M. Gao: Data curation, validation, investigation. D. Risnik: Data curation, investigation, methodology. S. Hao: Data curation, investigation. K. Ermine: Data curation, investigation. P. Wang: Resources, investigation. H. Li: Resources, investigation. Y. Huang: Writing–original draft, writing–review and editing. J. Yu: Resources, formal analysis, supervision, funding acquisition, writing–original draft, project administration, writing–review and editing. L. Zhang: Conceptualization, resources, data curation, software, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing.

This work was supported by U.S. NIH grants (grant nos. R01CA203028, R01CA217141, R01CA236271, R01CA247231, and R01CA248112 to L. Zhang; grant nos. U19AI068021 and R01CA215481 to J. Yu; grant no. R01CA260357 to Y. Huang; grant no. T32GM133332 to K. Ermine). This project used Animal Facility, Cytometry Facility, and Tissue and Research Pathology Services supported by P30CA047904.

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.