Scientific Papers

E3 ligase TRIM28 promotes anti-PD-1 resistance in non-small cell lung cancer by enhancing the recruitment of myeloid-derived suppressor cells | Journal of Experimental & Clinical Cancer Research


TRIM28 expression positively correlated with suppressive MDSC cell infiltration in NSCLC

Using data from the Cancer Genome Atlas (TCGA) (https://www.cancer.gov/ccg/research/genome-sequencing/tcga) databases, we examined TRIM28 expression across various cancer types and their corresponding non-tumor tissues. Our analysis revealed a significant elevation of TRIM28 expression in cancer samples compared to adjacent non-tumor tissues across multiple cancer types. Notably, TRIM28 mRNA levels were substantially higher in NSCLC tissues than in non-tumor tissues (Fig. 1A, Supplementary Table 1). Kaplan-Meier analysis indicated a significant negative correlation between high TRIM28 expression and the overall survival (OS) of NSCLC patients (Fig. 1B). To further investigate TRIM28’s role in NSCLC development, we conducted immunohistochemistry analysis on an NSCLC tissue microarray (Fig. 1C). Analysis of TRIM28 expression levels in conjunction with clinicopathological features revealed that high TRIM28 expression correlated with larger tumor size, lymph node metastasis, distant metastasis, and advanced pathological stage (Supplementary Table 2). Moreover, NSCLC patients with high TRIM28 expression levels exhibited a poorer prognosis (Fig. 1D). These findings collectively suggest that TRIM28 is upregulated in NSCLC tumor tissues and is inversely associated with the OS of NSCLC patients.

Fig. 1
figure 1

TRIM28 plays a mechanistic role in tumor progression by recruiting MDSCs into the tumor microenvironment. (A) TRIM28 expression levels in different tumor types from TCGA database were analyzed by TIMER2.0 (*p < 0.05, **p < 0.01, ***p < 0.001). (B) Survival analysis comparing the high and low expression of TRIM28 in lung adenocarcinoma according to TCGA dataset by using the website GEPIA 2 (http://gepia2.cancer-pku.cn/#survival). The high or low expression of TRIM28 were divided according to 50% of the total sample. (C) Immunohistochemical analysis of TRIM28 protein levels in NSCLC samples on tissue microarrays. Representative examples of TRIM28 expression in adjacent non-cancerous lung tissues, NSCLC tissues are shown. The scale bars represent 100 μm. (D) Overall survival analysis of patients with NSCLC stratified by the TRIM28 expression level in 90 samples. Kaplan-Meier survival analysis indicating a significant association between higher TRIM28 expression and poorer OS in NSCLC. (E) The correlations of TRIM28 expression and MDSCs infiltration in pan-cancers were analyzed by TIMER2.0. (F) Correlation of TRIM28 expression, tumor purity, and MDSCs infiltration in TCGA lung adenocarcinoma (LUAD) and lung squamous cell cancer (LUSC). The expression of TRIM28 positively correlates with MDSCs infiltration in NSCLC. (GH) Representative immunofluorescence staining of CD14 and TRIM28 in tissue from human lung adenocarcinoma and the correlation between TRIM28 and CD14 intensity. The expressions of TRIM28 and CD14 were measured with mean fluorescence intensities (MFIs) (in arbitrary units, a.u.), respectively. The pearson correlation between TRIM28 and CD14 expression (n = 90; p < 0.01, r = 0.567). Scale bars: 50 μm. (IJ) Cox regression analyses using data from TCGA indicated that high MDSC infiltration was significantly associated with poorer prognosis in NSCLC. Furthermore, elevated TRIM28 expression and a high proportion of MDSCs significantly correlated to poorer OS compared to their counterparts, strongly suggesting that TRIM28 influenced patient prognosis through an immune-related mechanism. Split infiltration percentage of patients: 50% (I). Split expression percentage of patients: 50% and split infiltration percentage of patients: 50% (J)

Aberrations in certain signaling pathways within tumor cells not only drive tumor cell proliferation and survival but also dictate the development of pro-tumorigenic microenvironments [27]. There is now substantial evidence indicating that oncogene expression is correlated with immune cell composition in human tumors [27]. Our findings prompted us to explore the role of TRIM28 in tumor progression and its impact on the tumor microenvironment. Consequently, we aimed to investigate the relationship between TRIM28 and immune cells through correlation analysis using the TIMER2.0 database. Interestingly, our results from TIMER2.0 analysis revealed a significant association between TRIM28 expression and MDSCs infiltration across various cancer types (Fig. 1E). This correlation was consistently observed in NSCLC (Fig. 1F). MDSCs in cancer patients were characterized by the expression of CD14, CD11b, CD33, and Arg1, along with low or absent expression of HLA-DR. When analyzing the tumor microenvironment of TRIM28-expressing NSCLC tumors, we observed an increase in CD14+ cells. Immunofluorescence staining demonstrated a positive correlation between the intensity of TRIM28 and CD14 expression in these tumors (Fig. 1G-H). Additionally, multivariate Cox analysis revealed an increased proportion of MDSCs was associated with a poor prognosis in NSCLC patients (Fig. 1I).

In addition, we analyzed the association between MDSC levels and OS in NSCLC patients, considering high or low TRIM28 expression using TIMER2.0. In the group of cancer patients with low TRIM28 expression, there was no significant difference in OS between those with low and high MDSC infiltration. However, in the group of cancer patients with high TRIM28 expression, higher levels of MDSCs were associated with worse patient survival, suggesting that elevated TRIM28 levels might enhance the immune-suppressive effect of MDSCs (Fig. 1J). These findings suggest an association between TRIM28 expression and immunosuppressive MDSCs in cancer. In summary, TRIM28 exhibits high expression in NSCLC tumor tissues and positively correlates with MDSC infiltration in the tumor microenvironment, potentially contributing to enhance tumor progression.

TRIM28 suppression sensitizes lung tumors to PD-1 blockade

Based on our observations that TRIM28 expression is strongly associated with MDSC infiltration in NSCLC, we hypothesized that TRIM28 promotes the chemotactic recruitment of MDSCs into the tumor microenvironment. To assess the effect of TRIM28 on MDSC infiltration in NSCLC, we conducted immunohistochemistry on MDSCs from syngeneic murine tumors. CMT-167 cells, with or without TRIM28 knockdown, were subcutaneously implanted into C57BL/6J mice. As expected, depletion of TRIM28 led to a profound reduction in MDSCs, whereas overexpression of TRIM28 increased MDSC infiltration in the tumor microenvironment (Supplementary Fig. 1). These results support our initial findings in patient tumor samples and emphasize the critical relationship between TRIM28 expression and MDSCs in cancer.

The effectiveness of checkpoint immunotherapy in NSCLC largely depends on the tumor microenvironment. Previous research has demonstrated that PD-1 blockade can transiently reduce tumor growth and metastasis in syngeneic tumor models by increasing CD8+ tumor-infiltrating lymphocytes (TILs) and decreasing exhausted CD8+TILs. However, increased MDSC infiltration is associated with treatment resistance to PD-1 blockade, reduced CD8+TILs, and increased exhausted CD8+TILs. Therefore, we investigated whether targeting TRIM28 in lung cancer could enhance the efficacy of anti-PD-1 therapy.

In syngeneic lung cancer models, we evaluated the impact of TRIM28 inhibition on the efficacy of anti-PD-1 therapy. As expected, the knockdown of TRIM28 alone decreased tumor growth. Both TRIM28 knockout and anti-PD-1 treatment alone reduced tumor size and weight compared to the control. Importantly, the combination of TRIM28 knockdown and anti-PD-1 significantly reduced tumor volume and weight, leading to a sustained reduction in tumor size throughout the treatment (Fig. 2A-B). In contrast, TRIM28 overexpression conferred resistance to anti-PD-1 therapy (Fig. 2C), further highlighting the significance of TRIM28 expression in maintaining sensitivity to anti-PD-1 therapy in NSCLC. To further investigate whether the therapeutic benefits of combined TRIM28 inhibition and anti-PD-1 extend to human NSCLC, we utilized a human lung adenocarcinoma cell line, H1299, in a humanized huHSC-NOG-EXL mouse model. Consistently, the combination of TRIM28 inhibition with anti-PD-1 resulted in sustained control of tumor progression (Fig. 2D-E). These findings suggest that TRIM28 promotes resistance to anti-PD-1 therapy in cancer by recruiting MDSCs into the tumor microenvironment.

Fig. 2
figure 2

TRIM28 inhibition enhances anti-PD-1 therapy in a syngeneic lung cancer model. (A) The schematic of tumor inoculation and treatment in mice. Western blot validated TRIM28 knockout or overexpression in CMT-167 cells. (B) C57BL/6J mice were subcutaneously injected with shControl or shTRIM28 CMT-167 cells and weekly with anti-PD-1 (200 µg/mouse) or isotype control (200 µg/mouse) by intraperitoneal injections starting on day 7 post tumor cell injection. Tumor growth curves and tumor weight were shown. n = 5 mice per treatment group. *p < 0.05; **p < 0.01. (C) C57BL/6J mice were subcutaneously injected with Control or TRIM28 CMT-167 cells and weekly with anti-PD-1 (200 µg/mouse) or isotype control (200 µg/mouse) starting on day 7 post-tumor cell injection. Tumor growth curves and tumor weight were shown. n = 5 mice per treatment group. *p < 0.05; **p < 0.01. (D) Experimental strategy. (E) Humanized huHSC-NOG-EXL mice were injected with shControl or shTRIM28 H1299 cells and treated with anti-PD-1(200 µg/mouse). Control animals were treated with isotype control. Tumor growth curves and tumor weight were shown. The treatment schema is as in (D). n = 5 mice per treatment group. Statistics were calculated using a one-way ANOVA post hoc Tukey test. *p < 0.05; **p < 0.01

TRIM28 is required for K63-linked ubiquitination of RIPK1

Structure-function evaluations have demonstrated that TRIM28’s tumor-promoting functions rely on its E3 ubiquitin ligase activity. To further investigate which protein TRIM28 interacts with and how this interaction affects the target protein’s physiological function, we conducted affinity capture-western assays, which confirmed that TRIM28 can interact with RIPK1 [28]. Subsequently, we conducted co-immunoprecipitation (Co-IP) analysis by co-transfecting Flag-RIPK1 and HA-TRIM28 into HEK293T cells. After immunoprecipitation with anti-Flag magnetic beads, we detected HA-TRIM28 and vice versa (Fig. 3A). Additionally, endogenous TRIM28 was found to bind endogenous RIPK1 in both CMT-167 and H1299 cells (Fig. 3A).

Fig. 3
figure 3

TRIM28 is required for K63-linked ubiquitination of RIPK1. (A) Ectopic HA-TRIM28 interacts with Flag-RIPK1 in 293T cells. Endogenous TRIM28 interacts with endogenous RIPK1 in CMT-167 and H1299 cells. (B) HEK293 cells were transfected with HA-TRIM28 and Flag-RIPK1 as indicated. Cell lysates were immunoprecipitated with anti-Flag and immunoblotted with anti-ubiquitin, anti-RIPK1, or anti-HA as indicated. (C and D) RIPK1 ubiquitination is increased upon overexpression of TRIM28-WT but not the ΔR mutant. 293T cells were transfected with HA-TRIM28 WT or ΔR mutant, and the cell lysates were subjected to immunoprecipitation using anti-Flag antibodies (C) or Ni-NTA pull-down under denaturing conditions (D), followed by immunoblotting with the indicated antibodies (E) RIPK1 ubiquitination was decreased upon TRIM28 depletion. 293T cells were co-transfected with the indicated plasmids or shRNAs, and Flag-RIPK1 was immunoprecipitated and analyzed by immunoblotting. (F) H1299 cells were transfected with control or TRIM28 shRNA, and endogenous RIPK1 was immunoprecipitated and analyzed for ubiquitination (G) TRIM28 modified RIPK1 by K63-linked ubiquitination. Cell lysates prepared in (C) were immunoprecipitated with anti-Flag and blotted with anti-ubiquitin K63 and anti-ubiquitin K48

Since TRIM28 possesses a RING finger domain, it has been considered an E3 ubiquitin ligase. Therefore, we investigated whether TRIM28 functions as an E3 ubiquitin ligase for RIPK1. We overexpressed Flag-RIPK1 with or without HA-TRIM28 and examined RIPK1 ubiquitination using immunoblot analysis with a ubiquitin antibody. As depicted in Fig. 3B, cells expressing TRIM28 exhibited induction of Flag-RIPK1 ubiquitination.

Next, HEK293 cells were transfected with plasmids encoding HA-TRIM28 or a mutant form, HA-TRIM28-∆RING, alongside Flag-RIPK1 and His-ubiquitin. Cells with HA-TRIM28 expression displayed increased ubiquitination of Flag-RIPK1 compared to control vector-transfected cells. Notably, RIPK1 ubiquitination was not observed in cells expressing the TRIM28-ΔRING mutant, indicating that TRIM28 potentiated RIPK1 ubiquitination through its E3 ligase activity (Fig. 3C-D). Consistently, the depletion of TRIM28 significantly inhibited RIPK1 ubiquitination (Fig. 3E). Moreover, TRIM28-depleted H1299 cells exhibited decreased ubiquitination of endogenous RIPK1 (Fig. 3F). Taken together, these findings indicate that TRIM28 is essential for RIPK1 ubiquitination, with RIPK1 being a substrate for TRIM28.

To determine the type of ubiquitin chain on RIPK1 modified by TRIM28, we performed an in vivo ubiquitination assay. Immunoblot analysis with a ubiquitin lysine (K) 63 antibody demonstrated that TRIM28 overexpression increased K63-linked ubiquitination of RIPK1. Conversely, TRIM28 overexpression did not affect the K48-linked ubiquitination of RIPK1, as determined by immunoblot analysis with a ubiquitin K48 antibody (Fig. 3G). These results suggest that TRIM28 primarily catalyzes the K63-linked ubiquitination of RIPK1. In summary, TRIM28 is an E3 ligase for K63-linked ubiquitination of RIPK1.

TRIM28 activates the NF-κB signaling pathway

It has been reported that TRIM28 plays a crucial role in activating the NF-κB signaling pathway. When TNF-α activates the NF-κB pathway, TRAF2 and RIPK1 are rapidly recruited to the membrane TNF-α receptor to form complex I, leading to K63-linked ubiquitination, which is a crucial step in NF-κB pathway activation. Previous studies have highlighted the importance of K63-linked polyubiquitination of RIPK1 in NF-κB signaling activation, a conserved regulator of immune and inflammatory responses [15]. Our results demonstrated that TRIM28’s E3 ligase activity is essential for mediating the K63-linked ubiquitination of RIPK1, suggesting that TRIM28 may play a significant role in driving NF-κB signaling pathway activation. Consistent with these findings, luciferase assays indicated that the activity of the NF-κB luciferase reporter gene increased in cells overexpressing wild-type TRIM28 but not the TRIM28-ΔRING mutant. Conversely, silencing TRIM28 reduced reporter activity, emphasizing the necessity of TRIM28’s E3 ligase activity for NF-κB activation (Fig. 4A).

Fig. 4
figure 4

TRIM28 activates NF-κB signaling. (A) HEK293 cells were co-transfected with the indicated plasmids along with pNF-κB-luc plasmids or the control-luciferase plasmid and subjected to a reporter assay. The luciferase assay showed that TRIM28, but not the ΔR mutant, induced the activation of NF-κB signaling. n.s., not significant; **p < 0.01. (B) p-IκBα, IκBα, p-IKKα/β, IKKα, IKKβ, p-p65, p65, TIRM28, and β-actin detected by western blot in TRIM28-knockdown and TRIM28-overexpressed CMT-167 cells. (C) p-IκBα, IκBα, p-IKKα/β, IKKα, IKKβ, p-p65, p65, TIRM28, and β-actin detected by western blot in TRIM28-knockdown and TRIM28-overexpressed H1299 cells. (D) Western blotting analysis of IκBα expression in the indicated cells treated with TNF-α (10ng/ml). β-actin is used as a loading control. (E) Assay of NF-κB luciferase reporter gene activity in TRIM28-overexpressing CMT-167 and H1299 cells transfected with vector or the IκBα dominant negative mutant (IκBα-mu). **p < 0.01. In (A) to (C), and (E), analyses were done in triplicate. Data represent mean ± SEM from each of three independent experiments. Statistics calculated using one-way ANOVA post hoc Tukey test for multi-group or two-tailed Student’s t-test for two-group comparisons

To further substantiate these results, we observed that TRIM28 overexpression substantially increased the phosphorylation of IκBα and IKKα/β while silencing TRIM28 significantly inhibited NF-κB signaling pathway activation in CMT-167 and H1299 cells (Fig. 4B-C, Supplementary Fig. 2). Additionally, we investigated whether TRIM28 facilitated sustained NF-κB activation by treating cells with TNF-α and assessing the protein levels of IκB. TRIM28-overexpressing cells exhibited longer-lasting reductions in IκB levels, while TRIM28-silenced cells exhibited shorter-lasting reductions (Fig. 4D). To further validate these findings, we blocked the NF-κB pathway by expressing an IκBα dominant-negative mutant (IκBα-mu) in TRIM28-overexpressing cells. As expected, the NF-κB pathway activation induced by TRIM28 overexpression was inhibited by IκBα-mu (Fig. 4E). Overall, these results illustrate that TRIM28 plays a pivotal role in NF-κB activation.

TRIM28 induces CXCL1 expression via the NF-κΒ pathway

Prior studies have shown that the NF-κB pathway activates CXCL1 transcription [29]. CXCL1 is a chemotactic factor that recruits MDSCs into the tumor microenvironment by binding to its receptor CXCR2 [30]. Therefore, we hypothesized that TRIM28-mediated NF-κB pathway activation might be responsible for enhancing CXCL1 expression. To explore this possibility, we analyzed nuclear phospho-p65 levels in lung cancer cell lines using western blot analysis. Phospho-p65, a key factor in the canonical NF-κB pathway, was decreased in TRIM28-depleted cell lines and increased in TRIM28-overexpressing cell lines (Fig. 5A-B, Supplementary Fig. 3). Subsequently, we examined phospho-p65 cellular localization in CMT-167 cells after modulating TRIM28 levels. Consistent with the western blot data, immunofluorescence results revealed a substantial increase in nuclear phospho-p65-positive staining in TRIM28-overexpressing cells compared to control cells, while TRIM28 knockdown attenuated nuclear phospho-p65 protein levels (Fig. 5C).

Fig. 5
figure 5

TRIM28 induces CXCL1 via the NF-κB pathway. (AB) Expression of p65 was determined by western blot in CMT-167 or H1299 cells overexpressing wild-type TRIM28 or TRIM28 knockdown with shRNA. The whole cell lysates and nuclear protein extracts were subjected to immunoblotting analysis. Data are representative of results obtained in three independent experiments. (C) The immunofluorescence assays of p-65 were performed in CMT-167 cells treated with the indicated plasmids. DAPI (blue) was used as a nuclear counterstain. The quantification of nuclear p-65 positive staining in at least 200 counted cells was presented as percentage ± SEM. Statistics calculated using one-way ANOVA post hoc Tukey for multiple comparisons. (DE) RT-qPCR determination of CXCL1 mRNA expression in TRIM28-knockdown and TRIM28-overexpressed CMT-167 or H1299 cells, treated with or without NF-κB inhibitor (BAY11-7085) at 10µM for 24 h. ELISA validation of levels of CXCL1 in cell culture supernatants from CMT-167 or H1299 cell culture. Data are representative of results obtained in three independent experiments. Statistics calculated using one-way ANOVA post hoc Tukey test for multi-group or two-tailed Student’s t-test for two-group comparisons. (FG) Representative IHC staining and quantification for CXCL1 in the indicated tumor models are shown. Additionally, ELISA was performed to validate the levels of CXCL1 in tumor lysates and sera from CMT-167 syngeneic tumor models, including subcutaneous tumors and blood samples collected from mice bearing CMT-167 control and CMT-167-shTRIM28 tumors. The scale bar represents 50 μm. (HI) Representative IHC staining and quantification for CXCL1 in the indicated tumor models are displayed. Additionally, ELISA was conducted to validate the levels of CXCL1 in tumor lysates and sera from CMT-167 syngeneic tumor models, including subcutaneous tumors and blood samples collected from mice bearing CMT-167 control and CMT-167-TRIM28 tumors. The scale bar represents 50 μm. Statistical analysis was performed using a two-tailed student’s t-test. **p < 0.01

Next, we assessed whether TRIM28 alters CXCL1 expression in murine and human lung cancer cell lines. RT-qPCR confirmed that TRIM28 knockout suppressed CXCL1 mRNA levels in CMT-167 and H1299 cells. Conversely, ectopic expression of TRIM28 in lung cancer cells had the opposite effect (Fig. 5D-E). To further validate that TRIM28-mediated NF-κB activation promotes CXCL1 and CCL2 expression, we treated TRIM28-overexpressing cells with the NF-κB inhibitor, BAY11-7085. CXCL1 expression significantly decreased after NF-κB inhibitor treatment compared to controls (Fig. 5D-E). ELISA assays consistently revealed that TRIM28 knockout significantly reduced CXCL1 secretion in cell culture supernatants of CMT-167 and H1299 cells (Fig. 5D-E). Furthermore, elevated TRIM28 levels caused a corresponding increase in CXCL1 expression. Activating the NF-κB signaling pathway could also promote CCL2 expression, contributing to MDSC accumulation in the tumor microenvironment [31]. Similarly, upregulated TRIM28 expression promoted CCL2 expression, while downregulated TRIM28 expression significantly reduced CCL2 expression (Supplementary Fig. 4).

To corroborate these findings in vivo, we examined CXCL1 levels in tumor tissue, tumor lysates, and serum from a syngeneic CMT-167 tumor model. TRIM28 overexpression in CMT-167 tumors led to an upregulation of CXCL1 levels in tumor tissue, tumor lysates, and serum (Fig. 5F-G). Conversely, inhibiting TRIM28 significantly decreased CXCL1 expression (Fig. 5H-I). In summary, these results suggest that TRIM28 in tumor cells induces the activation of the NF-κB pathway, which contributes to the expression of CXCL1.

TRIM28 recruitment and activation of MDSCs

Next, we examined the role of the TRIM28-CXCL1 axis in the chemotaxis of MDSCs, utilizing an in vitro MDSC migration assay. Notably, the deletion of TRIM28 significantly inhibited the migration of MDSCs towards the conditioned medium obtained from CMT-167 cells. Conversely, the introduction of TRIM28 into CMT-167 cells led to an increase in MDSC migration. Importantly, this effect was counteracted when an anti-CXCL1 neutralizing antibody was added to the conditioned medium (Fig. 6A). Furthermore, our analysis of TCGA data revealed a positive correlation between CXCL1 expression and the presence of MDSCs in NSCLC (Fig. 6B).

Fig. 6
figure 6

TRIM28 Promotes MDSCs recruitment in autochthonous tumors from KP mice. (A) Migration of MDSCs toward conditioned medium (CM) from CMT-167 cells were transfected with the indicated plasmids, treated with or without immunoglobulin G (IgG) control or CXCL1-neutralizing antibody. Data are representative of results obtained in three independent experiments. (B) Correlation of CXCL1 expression, tumor purity, and MDSCs infiltration in TCGA lung adenocarcinoma and lung squamous cell cancer. The expression of CXCL1 positively correlates with MDSCs infiltration in NSCLC. (C) Schematic representation of lentiviral TRIM28 or GFP overexpression in KP mice. KP mice were anesthetized with sodium pentobarbital followed by intranasal injection of Ad-Cre (1.5 × 106 pfu/mouse). (D) Representative images of HE staining in tumor-burdened lungs of KP mice were analyzed. (EF) Representative IHC (E) and IHC quantification (F) for immune cell markers (CD8, CD8+T cells; Gr-1, S100A8, S100A9, MDSC cells) in indicated in indicated tumor models. The scale bar represents 50 μm. Statistics calculated using a one-way ANOVA post hoc Tukey test. (G) Kaplan-Meier survival analysis of KP mice infected with Lenti-GFP-Cre or Lenti-TRIM28-Cre. Log-rank test. (HI) Flow cytometry gating strategy of immune cells. Quantification of flow-cytometry data for CD8+T, CD4+T, MDSCs as a percentage of leukocytes (CD45+) in tumor-burdened lungs from indicated tumor models. (J) Real-time qPCR and western blotting showed expression of representative MDSCs immunosuppressive gene in the indicated cell lines. Data are representative of results obtained in three independent experiments. (K) Quantification of the proliferation of CFSE-labelled CD8+T cells cocultured with MDSCs from tumor-burdened lungs of KP or KP-TRIM28 mice, analyzed by flow cytometry. Data are representative of results obtained in three independent experiments. Statistics calculated using one-way ANOVA post hoc Tukey test for multi-group or two-tailed Student’s t-test for two-group comparisons. **p < 0.01

To validate these findings in a more physiologically relevant setting, we conducted experiments in an autochthonous model of NSCLC, utilizing the KrasG12D; p53flox/flox (KP) lung adenocarcinoma system, which can be induced through lentiviral delivery of Cre recombinase alone or in combination with TRIM28 (Fig. 6C). Initially, we overexpressed TRIM28 in the KP adenocarcinoma model to investigate its impact on tumor development and MDSC levels. KP mice were infected with lenti-TRIM28-Cre or control lenti-GFP-Cre viruses. In mice infected with lenti-TRIM28-Cre (KP-TRIM28), histological analysis revealed significantly larger tumor sizes and areas in the lungs compared to control KP mice (Fig. 6D). Importantly, TRIM28 expression was substantially upregulated in the lungs of KP-TRIM28 mice (Fig. 6E). Furthermore, the enhanced expression of TRIM28 led to a significant reduction in the survival of KP-TRIM28 mice (Fig. 6G). To further investigate the association between TRIM28 expression and MDSC infiltration, we examined the percentages and numbers of MDSCs infiltrating the tumor-burdened lungs in KP-TRIM28 and KP mice.

Our observations indicated an increase in MDSCs and a decrease in CD8+T cell infiltration in the tumor-burdened lungs of KP-TRIM28 mice compared to KP mice (Fig. 6E-F). This was further supported by flow cytometry data, which showed a significant elevation in both the percentages and absolute numbers of MDSCs in the lungs of KP-TRIM28 mice, while the numbers of CD8+T cells and CD4+T cells were significantly reduced in comparison to KP mice (Fig. 6H-I).

It is widely acknowledged that antigen-presenting cells that capture tumor antigens and then migrate to lymphoid organs, where they activate tumor-specific naïve T cells, leading to their differentiation into effector T cells, play a crucial and central role in anti-tumor immunity. Thus, we assessed adaptive immune responses in the bronchial draining lymph nodes (dLNs). Flow cytometry analysis revealed a significant decrease in the numbers of CD4+, CD8+, CD4+IFNγ+ T cells, and CD8+IFNγ+ T cells in the bronchial dLNs of KP-TRIM28 mice when compared to KP mice. Additionally, there was an increase in the abundance of Treg cells in the dLNs of KP-TRIM28 mice compared to KP mice (Supplementary Fig. 5). These results further underscored the involvement of TRIM28 in reprogramming the tumor microenvironment towards immunosuppression.

To gain a deeper understanding of the different suppressive characteristics of MDSCs in the tumor-burdened lungs of KP-TRIM28 and KP mice, we examined the production and expression of factors associated with the immunosuppressive activity of MDSCs. We observed that MDSCs isolated from the lungs of KP-TRIM28 mice exhibited significantly higher levels of arginase-1 (Arg1), S100A9, and inducible nitric oxide synthase (iNOS) compared to MDSCs from the lungs of KP mice (Fig. 6J). Furthermore, we investigated whether TRIM28 influenced the suppressive function of MDSCs. MDSCs isolated from KP tumors were co-cultured with CD8+ T cells in vitro. Remarkably, MDSCs from KP-TRIM28 tumors displayed significantly stronger suppressive activity against the proliferation of CD8+T cells when compared to MDSCs from control tumors (Fig. 6K). These findings collectively demonstrated the crucial role of the TRIM28-CXCL1 axis in recruiting MDSCs into the lung cancer microenvironment.

RIPK1 inhibition sensitizes lung tumors to PD-1 blockade

Based on our above findings, we next analyzed the predictive role of TRIM28 in cancer patients undergoing anti-PD-1 therapy. Our analysis revealed that in patients with NSCLC, those with low TRIM28 expression experienced more favorable survival rates and longer survival times in comparison to patients with high TRIM28 expression (Fig. 7A). This relationship between TRIM28 expression and patient response was consistent in the GSE91061 melanoma cohort, where melanoma patients with low TRIM28 expression exhibited improved survival outcomes compared to those with high TRIM28 expression (Fig. 7B).

Fig. 7
figure 7

Targeting RIPK1 increases the sensitivity of lung tumors to anti-PD-1 therapy. (A) Kaplan-Meier curves predicting survival of LUSC patients receiving anti-PD-1 therapy based on net changes in TRIM28 mRNA levels in the LUSC-GSE93157-anti-PD-1 datasets. (B) Kaplan-Meier curves predicting survival of melanoma patients receiving anti-PD-1 therapy based on net changes in TRIM28 mRNA levels in the melanoma-GSE91061-anti-PD-1 datasets. (CD) C57BL/6J mice were subcutaneously injected with CMT-167 cells and treated with anti-PD-1, PK68 (RIPK1 inhibitor), PK68 plus anti-PD-1, or isotype control and vehicle. Tumor growth was monitored until the experimental endpoints. Data are shown as mean ± SEM. Tumor growth curves were shown. (EF) Representative images of IHC for CD8, Gr-1, and S100A8 + S100A9 in indicated mouse tumors (E) and IHC quantification (F). The scale bars represent 50 μm. Error bars indicate mean ± SEM. Statistics calculated using one-way ANOVA post hoc Tukey test for multi-group or two-tailed Student’s t-test for two-group comparisons. **p < 0.01. (G) High TRIM28 expression is positively correlated with MDSCs infiltration in multiple cohorts of cancer patients. TRIM28 promotes NF-κB activation by regulating K63-linked ubiquitination of RIPK1, leading to increased expression of the cytokine CXCL1, a chemoattractant for MDSCs via the CXCL1-CXCR2 axis. TRIM28-recruited MDSCs antagonize effector CD8+T cells in the tumor immune microenvironment of NSCLC, promoting anti-PD-1 resistance

PK68 is a potent and selective inhibitor of RIPK1 and also highlights its great potential for use in the treatment of cancer metastasis [32]. Given that tumor-associated TRIM28 induces CD8+T cell exhaustion through NF-κB signaling, we sought to assess RIPK1 as a potential therapeutic target in combination with immune checkpoint blockade. To explore this, we evaluated the impact of RIPK1 inhibition in a syngeneic mouse model. While the RIPK1 inhibitor PK68 exhibited a modest reduction in tumor growth when administered as a monotherapy, a significant reduction in tumor growth was observed only when PK68 was combined with anti-PD-1 treatment. This reduction in tumor growth became apparent within one week of treatment initiation and was sustained throughout the treatment course (Fig. 7C-D). Immunohistochemical analysis of tumor tissues revealed that treatment with the RIPK1 inhibitor alone enhanced the infiltration of CD8+T cells within CMT-167 tumors. However, the PK68/anti-PD-1 combination therapy resulted in a significantly greater infiltration of CD8 + T cells compared to monodrug treatment, a crucial characteristic associated with the success of ICB-based therapy. Furthermore, this combination therapy led to a notable reduction in the infiltration of MDSCs within the tumor microenvironment (Fig. 7E-F). Collectively, our study findings reveal a potential combined therapeutic strategy utilizing a RIPK1 inhibitor and PD-1 blockade to enhance anti-tumor efficacy.



Source link