Trabectedin decreases myeloid resistance to improve the efficacy of anti-PD1 immunotherapy and delay glioma malignant progression

Abstract

Low-grade gliomas (LGGs) are the most common brain tumors diagnosed in adolescents and young adults. LGGs demonstrate a longer latency and are associated with a lower relative risk compared to high-grade gliomas (HGGs), despite their potential to malignantly transform into fatal HGGs. We previously showed that malignant progression of glioma is associated with an accumulation of immunosuppressive myeloid cells that impair anti-tumor T cell function. Trabectedin is an FDA-approved chemotherapy used for the treatment of soft tissue sarcoma that has also been shown to selectively deplete monocytes and macrophages in those cancer types, although its role in modulating the brain tumor microenvironment (TME) has not yet been investigated. Here, we employed a multimodal approach, including mass cytometry, cytokine assay, and bulk RNA sequencing, to investigate the impact of trabectedin on the glioma immune TME in an immunocompetent, transgenic RCAS-tva murine glioma model that recapitulates spontaneous malignant progression. We report that trabectedin significantly decreased bone marrow-derived myeloid cells, alleviated immunosuppressive myeloid phenotypes, and diminished T cell exhaustion while increasing T cell infiltration in the glioma TME. Finally, trabectedin improved the efficacy of immune checkpoint blockade and significantly increased survival in the RCAS-tva glioma progression model.

Graphical abstract

The FDA-approved drug trabectedin remodels the glioma immune microenvironment by depleting immunosuppressive myeloid cells and enhancing T cell infiltration, both locally and systemically. In the immunocompetent RCAS-tva murine glioma progression model, trabectedin improves the efficacy of anti-PD1 immune checkpoint blockade and significantly extends survival.

Introduction

Approximately 22.9% of primary malignant brain and other central nervous system (CNS) tumors are gliomas, which exhibit wide heterogeneity in genetic, environmental, and prognostic features influencing pathogenesis and disease severity.^1,2 High-grade gliomas (HGGs) may arise de novo as primary tumors or transform from low-grade gliomas (LGGs) in a process called malignant transformation.³ LGGs, defined by the World Health Organization (WHO) as grades I and II, are the most common brain tumors in pediatric patients and account for 30% of all cases.⁴ While grade I gliomas are generally associated with favorable survival, grade II LGGs are problematic due to their tendency to recur and their potential to transform into more aggressive, lethal HGGs, which is associated with poor survival.^5,6 For example, Jaeckle et al. reported that in a cohort of 99 patients with grade II gliomas, 66 later exhibited malignant transformation, suggesting that a “watchful waiting” strategy is insufficient for clinical management.³ Despite extensive research comparing pediatric and adult LGGs, the adolescent and young adult (AYA) population, defined as patients between 15 and 39 years old, remains underrepresented and vulnerable, with 65,399 primary brain and other CNS tumor diagnoses reported between 2017 and 2021, comprising 14% of all such cases.^7,8 Current treatment typically involves maximal safe resection followed by chemotherapy (temozolomide), radiation, or both, but these approaches are associated with substantial side effects, including neurocognitive defects, endocrine disruption, and secondary malignancy, which are particularly harmful during neurological development, making pediatric and AYA patients especially susceptible.^8,9 Unlike adult LGGs, clinicians may opt to treat LGGs in younger patients with surgery alone to avoid the long-term impacts of chemotherapy and radiation, yet there is variability and uncertainty regarding how these treatments impact AYA patients.¹⁰ Moreover, temozolomide has been implicated in driving malignant transformation of LGGs through hypermutation, and immunotherapies, particularly immune checkpoint blockade (ICB), which show promise in other solid tumors, have been largely unsuccessful in CNS tumors due to factors such as myeloid cell-mediated immunosuppression.^2,11 Collectively, these limitations underscore that current therapies for LGGs in the AYA population are insufficient and highlight the urgent need for novel therapeutic approaches.

Throughout glioma progression, bone marrow-derived myeloid cells (BMDMs) traffic to and accumulate in the tumor microenvironment (TME) and exert potent immunosuppressive, pro-tumor activity.¹² These BMDMs often comprise 30%–50% of the tumor mass in glioblastoma (GBM) and severely impact the ability of the host immune system to combat the tumor.¹³ In HGGs, BMDMs differentiate into tumor-associated macrophages (TAMs), myeloid-derived suppressor cells (MDSCs), and monocytes that often express checkpoint receptors, downregulate antigen presentation and costimulatory molecules, and release immunosuppressive mediators, such as IL-10, arginase-1, and TGF-β, into the TME and periphery.¹³ Together with the highly proliferative and heterogeneous nature of GBM cells, these TAMs create an immunosuppressive TME that favors glioma progression and resistance to established immunotherapies, such as anti-PD1 ICB.² This combination of a highly heterogeneous, proliferative, and immunosuppressive nature of GBM cells, coupled with a myriad of immunosuppressive TAMs and MDSCs, creates an environment ideal for T cell dysfunction. T cell activity in GBM is often characterized by exhaustion, excessive Treg infiltration, and reduced proliferation, resulting in poor responses to immunotherapies.^14,15 Recent studies have demonstrated that tumor-associated myeloid cells impact T cell dysfunction in glioma.^16,17 Thus, depleting and/or modulating TAMs and monocytes may be an attractive strategy for reducing tumor burden and improving the efficacy of immunotherapies in patients with glioma.

Trabectedin is an FDA-approved chemotherapy currently used in the treatment of soft tissue sarcomas. Trabectedin kills cancer cells by binding to the minor groove of DNA and interfering with transcription-coupled nucleotide excision repair processes to generate DNA breaks.¹⁸ In addition to its chemotherapeutic effects, trabectedin selectively induces apoptosis in macrophages and monocytes via caspase-8 activation downstream of TRAIL-R1/2 engagement.¹⁹ Further, trabectedin inhibits transcription of inflammatory and angiogenic mediators, such as IL-6, VEGF, and CCL2, while simultaneously enabling transcription of anti-tumor cytokines, including TNF-α.²⁰ Importantly, trabectedin has been shown to increase T cell infiltration and activation in pancreatic, ovarian, and sarcoma cancer models.^19,21,22 Thus, trabectedin offers an attractive approach to combatting glioma progression because of its cytotoxic effect on cancer cells, induction of apoptosis in macrophages and monocytes, and interference with the transcription of inflammatory chemokines and angiogenic factors.

The use of trabectedin as an anti-tumor agent in glioma has yet to be investigated, and its immunomodulatory effects in this setting are not fully understood. Combination therapy with trabectedin and anti-PD1 has demonstrated preclinical success in ovarian and sarcoma cancer models but has yet to be investigated in any CNS tumor setting.^22,23,24 Myeloid-modulating therapies, including CSF1R inhibitors, have shown minimal success in treating GBM in clinical trials, and alternative strategies to modulate TAMs are necessary.²⁵ Currently, no data reporting trabectedin passing the blood-brain barrier (BBB) have been published. In humans, trabectedin is an FDA-approved drug with an acceptable toxicity profile, consisting of clinically manageable adverse effects, including neutropenia, hepatotoxicity, and thrombocytopenia.²⁶ Here, we utilize a murine glioma malignant transformation model to demonstrate the utility of trabectedin in reducing myeloid-based immunosuppression, increasing T cell infiltration, and enhancing the efficacy of anti-PD1 ICB toward delaying malignant progression.

Results

Trabectedin modulates the myeloid and lymphocyte populations of both the glioma TME and the systemic immune system

We sought to evaluate whether trabectedin is capable of altering immune cell populations in the glioma TME in vivo. A single retro-orbital injection of trabectedin was administered to tumor-bearing nestin tv-a (NTV-a) x Xfm mice at week 4 of glioma progression (Figure 1A), and immune cells were isolated from the tumors of mice treated with trabectedin or vehicle 4 days post-treatment 1 (4DPT1) and immunophenotyped via mass cytometry (CyTOF) (Figure 1B). Trabectedin significantly decreased the proportion of BMDMs (p = 0.03) and neutrophils (p = 0.03) compared to vehicle in the TME (Figure 1C). Further, trabectedin decreased populations of monocytes (p = 0.05) and peripherally derived macrophages (p = 0.06) compared to vehicle. Additionally, the proportion of microglia was significantly increased in the trabectedin-treated group (p = 0.02) (Figure 1C). Interestingly, the frequency of CD3+ T cells was significantly increased in the trabectedin-treated group (p = 0.001) (Figure 1D). Within the CD3+ T cell compartment, CD4+ helper T cells (TH) were significantly increased (p = 0.01), while CD8+ cytotoxic T cells (CTL) showed a modest increase (p = 0.17) in the trabectedin-treated group (Figure 1D). No significant changes were observed in regulatory T cells (Tregs, p = 0.84) or natural killer (NK) cells (p = 0.82) between the treatment groups (Figure 1D). Overall, trabectedin decreased the proportion of BMDMs, including macrophages and monocytes, while simultaneously increasing T cell infiltration in the glioma TME.

Figure 1.

Trabectedin modulates the intratumoral myeloid and lymphocyte proportions in the glioma tumor microenvironment in vivo

(A) Schematic depicting the immunocompetent RCAS-tva murine glioma progression model. Pups with a genetic background of heterogeneous tumor suppressor gene loss are injected intracranially with DF-1 cells carrying RCAS-PDGFB and RCAS-Cre on days 0–2 after birth (P0–P2) to induce spontaneous gliomagenesis, resulting in glioma progression from low-grade glioma (LGG) to high-grade glioma (HGG) over the course of 5–7 weeks. (B) Experimental schematic demonstrating a singular dose of trabectedin (0.15 mg/kg) injected retro-orbitally at week 4 of glioma progression and assessment of immune cells 4 days post-treatment via mass cytometry (CyTOF). (C) CyTOF analysis of the myeloid and (D) lymphocyte (CD11b-) populations in the tumor microenvironment in trabectedin- or vehicle-treated mice 4 days post-treatment (n = 4–5 per group). Significance was determined using a paired t test with Welch’s correction (p < 0.0005∗∗∗∗, p < 0.005 ∗∗∗, p < 0.005∗∗, p < 0.05∗). Error bars show SD.

To evaluate the systemic immune effects of trabectedin in the RCAS-tva glioma model, we performed CyTOF on immune cells isolated from the bone marrow and spleens 4DPT1 of mice treated with trabectedin or vehicle (Figure 1B). We observed significant decreases in splenic CD11b+ myeloid cells (p = 0.02), macrophages (p = 0.01), monocytes (p = 0.006), and monocytic dendritic cells (mDCs, p = 0.04) in trabectedin-treated mice compared to vehicle (Figure 2A). Furthermore, trabectedin significantly increased the proportions of CD3+ T cells (p < 0.0001), CD4+ TH (p = 0.002), Tregs (p = 0.03), and CD8+ CTL (p = 0.004), while significantly decreasing the proportion of B cells (p = 0.008) (Figure 2B). In the bone marrow of trabectedin-treated mice, we observed a general increase in CD11b+ myeloid cells (p = 0.05), driven by a significant increase in neutrophils (p = 0.005) (Figure 2C). A reduction in MDSC-like cells was observed (p = 0.15) (Figure 2C). Additionally, neutrophils were increased (p = 0.005) in the bone marrow of trabectedin-treated mice (Figure 2C). Lastly, trabectedin significantly increased the proportion of CD3+ T cells (p = 0.006) in the bone marrow, specifically the CD8+ CTL population (p = 0.02), although CD4+ TH cells were also increased (p = 0.05) (Figure 2D). Interestingly, a significant decrease in B cells (p = 0.04) was also observed in the bone marrow of the trabectedin-treated group (Figure 2D). Overall, trabectedin decreased splenic myeloid cells, including monocytes and macrophages, while increasing the expansion of both CD4+ and CD8+ T cells in the spleen and bone marrow.

Figure 2.

Trabectedin modulates the systemic myeloid and lymphocyte proportions in the spleen and bone marrow of glioma-bearing mice

Spleens (A and B) and bone marrow (C and D) from glioma-bearing mice treated with a single dose of trabectedin or vehicle at week 4 of glioma progression were analyzed via mass cytometry (CyTOF) 4 days post-treatment (n = 4–5 per group). Significance was determined using a paired t test with Welch’s correction (p < 0.0005∗∗∗∗, p < 0.005 ∗∗∗, p < 0.005∗∗, p < 0.05∗). Error bars depict the SD.

Trabectedin modulates the serum levels of circulating cytokines in glioma-bearing mice

To further investigate whether trabectedin is exerting its myeloid-modulating impact systemically, in the glioma TME, or both, we analyzed the serum of trabectedin-treated glioma-bearing mice. Trabectedin or vehicle was delivered retro-orbitally at weeks 3.5 and 4.5 to glioma-bearing RCAS-tva mice, and serum was harvested 3DPT2. Serum from vehicle- or trabectedin-treated mice was analyzed via Luminex Multiplex Soluble Cytokine Assay, and of the 42 cytokines measured, 30 were detectable. The mean serum concentration for each treatment group are summarized in Figure 3A. Serum from trabectedin-treated mice displayed a decrease in the monocyte chemoattractant CCL2 (p = 0.06) compared to vehicle (Figure 3B). Additionally, trabectedin increased serum levels of the cytokines IL-2 (p = 0.29) and IL-15 (p = 0.11) (Figure 3B). In summary, trabectedin modulates the levels of circulating cytokines, specifically those involved in monocyte chemotaxis and effector cell function.

Figure 3.

Trabectedin modulates circulating cytokines in the serum and impacts gene expression pathways in tumors of glioma-bearing mice

Glioma-bearing mice were treated with 2 doses of trabectedin or vehicle at weeks 3.5 and 4.5 of glioma progression. Serum and RNA from tumors were harvested 3 days after the second treatment for Luminex cytokine assay and bulk-RNA-seq, respectively. (A) Heatmap summarizing the mean serum concentration of circulating cytokines and chemokines (n = 4 per group). The average serum concentration in pg/mL for trabectedin- or vehicle-treated mice is displayed in the corresponding cell. (B) Quantification of serum concentrations of the cytokines CCL2, IL-15, and IL-2. Significance was determined using a paired t test. Error bars show the SD. (C–F) Gene set enrichment analysis (GSEA) of bulk RNA-seq data comparing trabectedin-treated tumors to vehicle-treated tumors across 3 gene databases (n = 3–4 per group). A positive NES indicates upregulation in trabectedin-treated tumors compared to vehicle. GSEA of the CancerHallmark, GOBP, and Reactome databases indicates significant upregulation of gene pathways related to interferon-γ, interferon-α, cell cycle checkpoints, homology-directed repair, and DNA double-strand break repair in trabectedin-treated tumors. Alternatively, GSEA demonstrates significant downregulation of gene pathways associated with TGF-β signaling, myeloid cell development, and myeloid cell homeostasis in trabectedin-treated tumors compared to vehicle.

Trabectedin impacts gene expression pathways within the glioma TME

To further evaluate the impact of trabectedin on the glioma TME, we performed bulk RNA sequencing (RNA-seq) of the tumor. Trabectedin or vehicle was delivered retro-orbitally at weeks 3.5 and 4.5 to glioma-bearing RCAS-tva mice. Tumors were harvested 3DPT2, and RNA was isolated for RNA-seq. To investigate gene expression differences between the vehicle- and trabectedin-treated groups, gene set enrichment analysis (GSEA) was performed across 3 databases (CancerHallmark, Gene Ontology Biological Processes [GOBP], and Reactome). Interestingly, the IFN-γ and IFN-α response pathways were significantly enriched in the trabectedin-treated group (Figure 3C). Additionally, the TGF-β signaling pathway was significantly downregulated in trabectedin-treated tumors compared to vehicle (Figure 3C). Furthermore, trabectedin-treated tumors significantly downregulated the myeloid cell development and homeostasis pathways (Figures 3D and 3E), suggesting trabectedin’s myeloid-modulating role within the TME. Lastly, trabectedin significantly upregulated pathways associated with cell cycle checkpoints, homology-directed repair, and double-strand break repair (Figure 3F). We observed that 69% of differentially expressed pathways in the GOBP database were broadly associated with DNA damage/repair, mitosis, and the cell cycle (data not shown). These results suggest that trabectedin, in addition to its immune modulation, may be exerting a chemotherapeutic effect in the TME. To confirm the ability of trabectedin to act as a chemotherapy in glioma, as this has not yet been reported, we performed a WST-1 cell viability assay on four different glioma cell lines with various concentrations of trabectedin over 72 h. Trabectedin was effective at killing glioma cells at very low concentrations in vitro in the U251, GL261, U87, and CT2A cell lines, with relative IC50 values of 0.262, 0.721, 1.03, and 0.398, respectively (Figure S2). Overall, trabectedin was capable of upregulating pathways related to anti-tumor immune responses including IFN-γ and IFN-α, downregulating the immunosuppressive TGF-β pathway, and increasing DNA damage-related pathways within the tumor.

Trabectedin improves the efficacy of anti-PD1 ICB in the RCAS-tva glioma progression model

Trabectedin reduced myeloid cells and improved T cell infiltration into the glioma TME and systemically (Figures 1 and 2). Next, we investigated whether this immune modulation by trabectedin improves the efficacy of anti-PD1 ICB in the RCAS-tva glioma progression model.

To determine the optimal timing of anti-PD1 ICB treatment, we conducted flow cytometry on the splenic myeloid cells of trabectedin- or vehicle-treated mice 1DPT1, 3DPT2, and 7DPT2. We observed the most robust decrease in splenic myeloid cells at 3DPT2 and thus adopted this time point to begin anti-PD1 treatment 72 h following trabectedin treatment (Figure S3). Four groups of mice were treated with either vehicle, trabectedin (2 doses, 1 week apart, starting at week 4), anti-PD-1 ICB (3 doses, 3 days apart, starting at week 3.5), or a combination of trabectedin and anti-PD-1 ICB (Figure 4A). Trabectedin monotherapy showed a significant survival benefit compared to vehicle (p = 0.004), whereas anti-PD1 monotherapy did not show a significant survival benefit compared to vehicle (p = 0.05) (Figure 4B). Interestingly, trabectedin monotherapy did not show a significant survival benefit compared to anti-PD1 monotherapy (p = 0.96). However, combination therapy with trabectedin and anti-PD1 ICB showed a significant survival benefit compared to vehicle (p < 0.0001), anti-PD1 monotherapy (p = 0.009), and trabectedin monotherapy (p = 0.01) (Figure 4B). To investigate apoptosis occurring within the tumor in response to trabectedin or trabectedin in combination with anti-PD1, we performed immunohistochemical (IHC) staining of cleaved caspase-3 (c-cas3) 3DPT2, using a tonsil as a positive control (Figure 4C). Minimal positive staining for c-cas3 was observed in the vehicle- or trabectedin-treated tumors; however, a significant increase in c-cas3 staining was observed in the combination group (Figure 4D). Together, these data indicate that trabectedin improves the efficacy of anti-PD1 ICB and results in increased tumor cell death within the glioma TME.

Figure 4.

Combination trabectedin and anti-PD1 ICB prolongs survival of glioma-bearing mice

(A) Schematic depicting the treatment regimen for the Ntv-a x XFM-luc RCAS murine glioma progression model. Four groups of mice were treated with vehicle, trabectedin (2 doses, 1 week apart, starting at week 4), anti-PD-1 immune checkpoint blockade (ICB; 3 doses, 3 days apart, starting at week 3.5), or a combination of trabectedin and anti-PD-1 ICB. (B) Kaplan-Meier survival analysis of the four treatment groups. Vehicle (n = 12), anti-PD1 (n = 8), trabectedin (n = 13), and trabectedin + anti-PD1 (n = 12) had median survivals of 39, 43, 43, and 53 days, respectively. Significance was determined using the log-rank (Mantel-Cox) test (p < 0.0001∗∗∗∗,p < 0.0002 ∗∗∗, p < 0.0021∗∗, p < 0.0332∗). (C) Representative images of cleaved caspase-3 (c-cas3) immunohistochemical (IHC) staining of tumors isolated from glioma-bearing mice 3 days post treatment 2 (3DPT2) with vehicle, trabectedin, or trabectedin + anti-PD1 combination. (D) Quantification of c-cas3 IHC staining. Positive pixels were calculated by sampling 5 equal-sized regions within each tumor core and comparing the number of positive pixels per region (n = 2–5 mice per treatment group). Significance was determined using a paired t test with Welch’s correction (p < 0.0005∗∗∗∗, p < 0.005 ∗∗∗, p < 0.005∗∗, p < 0.05∗). Error bars show SD.

Trabectedin reduces myeloid cell immunosuppression and T cell exhaustion in the glioma TME to improve anti-PD1 ICB

We next investigated the immunological impacts of trabectedin in combination with anti-PD1 in the glioma TME via CyTOF. Glioma-bearing RCAS-tva mice were treated with vehicle, trabectedin, anti-PD1 ICB, or combination therapy as described in Figure 4A, and immune cells were isolated from the tumor for CyTOF 3DPT2. Anti-PD1 ICB monotherapy significantly increased the proportion of BMDMs (p = 0.03), peripherally derived macrophages (p = 0.01), and Tregs (p = 0.04) compared to vehicle (Figure 5A). Combination therapy alleviated this effect and significantly reduced BMDMs (p = 0.007), macrophages (p = 0.009), and Tregs (p = 0.02) compared to anti-PD1 monotherapy (Figure 5A). These differences were associated with an increase in MHCII expression on CD11b+ myeloid cells in the trabectedin, anti-PD1 ICB, and combination groups compared to vehicle (Figure 5B). To investigate how trabectedin monotherapy and combination therapy with anti-PD1 ICB affect T cell function, we generated CyTOF uniform manifold approximation and projection (UMAPs) from CD3+ T cells for each treatment group and analyzed the expression of functional markers on different T cell subsets (Figure 5C). The expression of PD1, both an activation and exhaustion marker depending on the context, was significantly decreased on CD4+ TH and CD8+ CTL in the trabectedin group compared to vehicle (p = 0.04, p = 0.002) (Figure 5D). Further, trabectedin significantly decreased the exhaustion marker Lag3 on both CD4+ and CD8+ T cells compared to vehicle (p = 0.004, p = 0.005). Anti-PD1 monotherapy significantly increased the expression of PD1, Lag3, and Tim3 on CD4+ T cells compared to trabectedin (p = 0.02, p = 0.01, p = 0.01, respectively) (Figure 5D). Although anti-PD1 monotherapy increased the expression of PD1, Tim3, and Lag3 on T cells compared to trabectedin monotherapy, combination therapy trended toward lower expression of these markers than anti-PD1 ICB alone, suggesting that trabectedin is able to partially alleviate anti-PD1 ICB’s induction of checkpoint markers (Figure 5D). In summary, trabectedin was able to mitigate some of the suboptimal immunological effects of anti-PD1 ICB monotherapy by decreasing BMDMs, macrophages, Tregs, and T cell exhaustion markers.

Figure 5.

The combination of trabectedin and anti-PD1 ICB modulates the glioma tumor microenvironment in vivo

(A–D) Immune cells were isolated from murine glioma tumors treated with trabectedin (n = 3), anti-PD1 immune checkpoint blockade (ICB) (n = 3), combination therapy (n = 4), or vehicle (n = 4) 3 days post-treatment 2 (3DPT2) and profiled by CyTOF. (A) CyTOF analysis demonstrates that the combination of trabectedin and anti-PD1 ICB significantly reduced bone marrow-derived myeloid cells (BMDMs, CD45high CD11b+), macrophages (CD45high CD11b+ F4/80+), and regulatory T cells (CD45+ CD3+ CD4+ CD25high Foxp3+) compared to anti-PD1 monotherapy. (B) Representative CyTOF gating on CD45+ immune cells indicates that trabectedin monotherapy and combination treatment increase MHCII+ CD11b+ myeloid cells. (C) Dimensional reduction of CD3+ T cells by treatment group illustrates differences in Tregs, central memory T cells (TCM), and exhaustion markers PD1, Lag3, and Tim3. (D) Quantification of exhaustion markers on CD4+ and CD8+ T cells by CyTOF. Significance was determined using a paired t test with Welch’s correction (p < 0.0005∗∗∗∗, p < 0.005 ∗∗∗, p < 0.005∗∗, p < 0.05∗). Error bars show SD.

Trabectedin impacts systemic myeloid cell function and T cell memory globally

Next, we investigated how trabectedin impacts functional myeloid and T cell phenotypes in the spleen and bone marrow, alone and in combination with anti-PD1 ICB. Glioma-bearing RCAS-tva mice were treated with vehicle, trabectedin, anti-PD1 ICB, or combination therapy as described in Figure 4A, and immune cells were isolated from the spleens and bone marrow for CyTOF staining 3DPT2. Trabectedin monotherapy significantly decreased Trem2 expression on splenic CD11b+ myeloid cells compared to vehicle (p = 0.04) (Figures 6A and 6B). Trem2 expression on myeloid cells is associated with an immunosuppressive TAM phenotype.²⁷ Furthermore, trabectedin alone and in combination with anti-PD1 ICB significantly increased the expression of MHCII on CD11b+ splenic myeloid cells compared to vehicle (p = 0.004, p = 0.02) (Figures 6A and 6B).

Figure 6.

The combination of trabectedin and anti-PD1 ICB systemically modulates immune cells

(A–D) Immune cells were isolated from murine glioma tumors treated with trabectedin (n = 3), anti-PD1 immune checkpoint blockade (ICB) (n = 3), combination therapy (n = 4), or vehicle (n = 4) 3 days post-treatment 2 (3DPT2) and profiled by CyTOF. (A and B) CyTOF quantification and representative gating of splenic CD11b+ myeloid cells demonstrate that trabectedin significantly reduces Trem2 and increases MHCII expression compared to Vehicle. (C) CyTOF analysis of bone marrow immune cells indicates variable expression of PD-L1, CCR2, and CX3CR1 on macrophages and CD69 on CD335+ natural killer (NK) cells across treatment groups. (D) Quantification of CD8+ CD62L + CD44+ central memory T cells in spleen, BM, and brain TME across treatment groups. (E) Representative CyTOF gating of splenic memory CD8+ T cells. Significance was determined using a paired t test with Welch’s correction (p < 0.0005∗∗∗∗, p < 0.005 ∗∗∗, p < 0.005∗∗, p < 0.05∗). Error bars show SD.

We observed significant decreases in PD-L1+ MHCII- immunosuppressive macrophages in the trabectedin-treated group compared to vehicle (p = 0.02) and anti-PD1 ICB monotherapy (p = 0.003) in the bone marrow (Figure 6C). Additionally, combination therapy reduced the proportion of PD-L1+ MHCII- macrophages in the bone marrow compared to vehicle (p = 0.06) and anti-PD1 ICB monotherapy (p = 0.04) (Figure 6C). Next, we assessed the expression of homing receptors CCR2 and CX3CR1, which are implicated in myeloid cell trafficking and chemotaxis to the glioma TME.^28,29 Trabectedin monotherapy and combination therapy with anti-PD1 ICB significantly decreased the expression of CX3CR1, the receptor for CX3CL1, on bone marrow macrophages compared to vehicle (p = 0.0004, p = 0.002) (Figure 6C). Further, trabectedin significantly reduced the expression of CCR2, the receptor for the potent monocyte chemokine CCL2, on bone marrow macrophages compared to vehicle (p = 0.04) (Figure 6C). We also observed a significant increase in the NK activation marker CD69 in the combination group compared to trabectedin monotherapy (p = 0.02). Lastly, we observed a significant increase in CD8+ central memory T cells (TCM, CD8+ CD62L + CD44+) in the spleen (p = 0.02) and bone marrow (p = 0.02), as well as a modest increase in the TME (p = 0.08), of trabectedin-treated mice compared to vehicle (Figures 6D and 6E). Taken together, these results indicate that trabectedin is capable of functional modulation of myeloid and effector cells systemically, rather than solely depletion, and functions well immunologically in combination with anti-PD1 ICB to reduce immunosuppression.

Discussion

Trabectedin is an FDA-approved drug for the treatment of soft tissue sarcoma, with ongoing clinical trials in various solid tumors.¹⁸ Here, we conducted the first study to investigate the use of trabectedin in any CNS tumor and describe its immunological and therapeutic effects in glioma in vivo, both within the TME and systemically. During glioma malignant transformation from LGG to HGG, BMDMs traffic from the bone marrow to the TME and differentiate into TAMs and MDSCs that exert pro-tumor, immunosuppressive functionality.³⁰ Trabectedin, in addition to its chemotherapeutic effects, has been shown to deplete monocytes and macrophages via cell-specific TRAIL-R1/2 signaling, interfere with pro-tumor gene transcription in myeloid cells, and increase effector cell functions in sarcoma, pancreatic, and ovarian cancer models.²⁰ In the present study, we determined that trabectedin was able to reduce BMDMs and increase T cell infiltration in the glioma TME, reduce myeloid cell immunosuppression, enhance effector cells systemically, and significantly improve survival as a monotherapy and in combination with anti-PD1 ICB in the immunocompetent RCAS-tva glioma progression murine model.

Approximately 30%–50% of a GBM tumor mass is comprised of myeloid cells, including infiltrating BMDMs consisting of monocytes, macrophages, MDSCs, and neutrophils, as well as resident microglia.¹³ As glioma tumors progress, microglia become outnumbered by these BMDMs, which become educated toward an immunosuppressive phenotype in the TME.^12,13 Furthermore, glioma patients exhibit a rise in immunosuppressive myeloid cells peripherally, which can contribute to disease progression.³¹ Here, we demonstrated the ability of trabectedin to reduce the myeloid cell burden in the glioma TME as well as systemically in the spleen. Decreases specifically observed in peripherally derived macrophages and monocytes in the TME and spleen align with various other studies of trabectedin in other solid tumors.²⁰ Trabectedin exerts its cytotoxic effect against myeloid cells by binding TRAIL-R1 or TRAIL-R2 on monocytes and macrophages to trigger caspase-8-dependent apoptosis. T cells and neutrophils are spared due to their expression of the decoy receptor TRAIL-R3, although less is known about trabectedin’s effect on other immune cell populations.¹⁹ Furthermore, NK cells showed no change in viability when exposed to trabectedin in vitro.³² Here, we observed that trabectedin decreases B cell proportions in the TME, spleen, and bone marrow of treated mice. Interestingly, resident microglia populations in the TME increased following trabectedin treatment. Matysiak et al. reported that microglia express high levels of the decoy receptor TRAIL-R3, potentially protecting these cells from trabectedin-induced TRAIL-triggered death.³³ Lastly, we observed an increase in myeloid cell proportions in the bone marrow, consisting mostly of neutrophils. It is possible that the trabectedin-induced immunological imbalance skewed toward lymphocyte expansion in the spleen and bone marrow triggered a compensatory expansion of neutrophils. Taken together, these novel results indicate that trabectedin can reduce myeloid populations in brain tumors locally and systemically, creating an environment more primed for T cell-based immunotherapy.

Recent studies in sarcoma and breast cancer models demonstrate the ability of trabectedin to modulate myeloid cell phenotypes, rather than solely deplete monocytes and macrophages.^32,34 Swartz et al. reported that trabectedin decreases the proportion of CD206^high M2-like and increases MHCII^high M1-like splenic macrophages in a breast cancer model.³² In the present study, trabectedin increased MHCII and decreased Trem2 expression on splenic myeloid cells, further supporting a myeloid-modulating impact of the drug beyond myeloid-depletion. In cancer, TREM2 is a major regulator of the myeloid immune response, and its expression on TAMs is associated with immunosuppression and resistance to anti-PD1 therapy.^27,35 Furthermore, we observed a significant decrease in MHCII- PD-L1+ macrophages in the bone marrow following trabectedin treatment. Bone marrow-resident macrophages with high PD-L1 expression could potentially interact with circulating, memory, or developing T cells to suppress their activation or cytotoxicity. Additionally, trabectedin significantly decreased the expression of homing receptors CCR2 and CX3CR1 on macrophages in the bone marrow. CCR2 is a receptor for the monocyte chemoattractant CCL2, and the CCL2/CCR2 axis is heavily associated with the trafficking to and accumulation of MDSCs and Tregs in the glioma TME.^28,29 Additionally, we observed a decrease in serum CCL2 levels following treatment with trabectedin. CX3CR1, a receptor predominantly expressed by monocytes and macrophages, binds to the chemoattractant CX3CL1 and facilitates the movement of monocytes out of the bone marrow and into the periphery.³⁶ Overall, these data suggest that trabectedin systemically modulates myeloid cell phenotypes by decreasing immunosuppressive molecules and reducing the trafficking of BMDMs to the glioma TME.

T cells in glioma often display a dysfunctional phenotype, consisting of exhaustion, reduced proliferation, and poor anti-tumor activity in the TME.^16,37 Furthermore, immunosuppressive CD4+ Tregs are often enriched in the glioma TME and contribute to a potent pro-tumor environment.³⁸ The immunosuppression of effector T cells by TAMs and MDSCs is often a key resistance mechanism to T cell-based immunotherapies.³⁹ Here, we demonstrated the ability of trabectedin to increase T cell abundance in the TME and expand T cell populations in the spleen and bone marrow, likely by reducing local and systemic myeloid-based immunosuppression. We further observed increases in serum IL-2 and IL-15 levels in trabectedin-treated mice. The cytokines IL-2 and IL-15 are both positively associated with T cell proliferation, differentiation, and cytotoxic effector function.⁴⁰ Additionally, we observed a significant decrease in exhaustion markers, Lag3 and PD1, on both CD4+ and CD8+ T cells in the glioma TME following trabectedin treatment. Moreover, trabectedin increased the proportion of TCM in the brain TME, spleen, and bone marrow. Studies have demonstrated that TCM produce higher levels of cytokines, exhibit higher cytotoxic activity, and demonstrate a better ability to eradicate murine tumors compared to effector T cells.⁴¹ Given these promising immunological results, we hypothesized that trabectedin would increase the efficacy of anti-PD1 ICB. Trabectedin and anti-PD1 ICB combination therapy significantly improved survival in the immunocompetent RCAS-tva glioma progression model compared to either monotherapy or vehicle. Although all of the tumors eventually progressed to high-grade regardless of treatment in this study, we were able to successfully delay tumor progression and extend survival, which offers promise in a setting with excessive therapeutic resistance. Furthermore, the combination therapy significantly reduced the increased infiltration of BMDMs, macrophages, and Tregs elicited by anti-PD1 ICB monotherapy. It is probable that priming of the systemic and TME immune cells with trabectedin temporarily alleviates myeloid resistance, allowing anti-PD1 ICB to exert its intended function.

Clinical trials investigating the use of myeloid-depleting drugs, namely CSF1R inhibitors, were largely unsuccessful in showing a significant survival benefit in glioma patients.^42,43 Drugs targeting TAMs and MDSCs for depletion or polarization offer promise for delaying the malignant progression of glioma and improving responses to ICB, but they face numerous barriers. For example, the CSF1R-inhibitor PLZ3397 was able to deplete TAMs but was unsuccessful in recruiting T cells to the TME as a monotherapy in a model of mesothelioma.⁴⁴ Furthermore, the discovery of novel myeloid-modulating drugs is necessary, but the average length of time from drug discovery to the clinic is approximately 9 years.⁴⁵ Alternatively, trabectedin is an FDA-approved drug that is generally well tolerated in patients with appropriate clinical considerations, and pretreatment with dexamethasone reduces hepatotoxicity.⁴⁶ Here, we propose utilizing trabectedin as a primer for ICB therapy, rather than as a sole treatment option, to reduce myeloid resistance and improve systemic T cell activity. Further studies are necessary to evaluate trabectedin as a suitable clinical option for glioma patients, including its BBB penetrability. We provide some evidence for BBB penetrability by trabectedin, with the RNA-seq data suggesting a chemotherapeutic effect within the TME. Additionally, a more in-depth analysis of the infiltrating T cells in the TME by single cell RNA-seq will inform the ability of the combination therapy to elicit a tumor-specific T cell response. Overall, our study offers novel insight into the use of trabectedin in the CNS tumor setting and describes expansive immunological results that could potentially be propagated into the clinic with further studies.

Materials and methods

Mice

The RCAS-NTV-a murine model has been described in detail previously.⁴⁷ Ntv-a mice are genetically engineered to express the tv-a receptor for the RCAS plasmid under the control of the nestin promotor. The nestin promotor is overexpressed in neural progenitor cells. Ntv-a (Ink4a^+/+ Arf ^+/+) mice were crossed with Xfm (Ink4a ^−/− Arf ^−/− PTEN ^fl/fl) mice to generate immunocompetent NTV-a x Xfm (Ink4a^+/− Arf ^± PTEN^+/fl) pups. The heterozygous pups were injected intracranially at day 0–2 into the brain parenchyma with 1 μL of a combination of 50,000 DF-1 carrier cells transfected with RCAS-PDGFB and 50,000 DF-1 carrier cells transfected with RCAS-Cre using a 10 μL Hamilton syringe. Tumor progression was validated by histological analysis and confirmed by a board-certified neuropathologist at Nationwide Children’s Hospital. Mice were treated with a maximum of two doses of 0.15 mg/kg of trabectedin (MCE, #HY-50936S, Monmouth Junction, NJ) delivered 7 days apart retro-orbitally under anesthesia. Mice were monitored for adverse effects of trabectedin, including injection site inflammation and/or necrosis, by the veterinary staff of the Animal Resources Core (ARC) at Nationwide Children’s Hospital (NCH) and were euthanized if deemed necessary. Mice treated with anti-PD1 ICB (BioXCell, #BEO146, clone RMP1-14, Lebanon, NH) were injected intraperitoneally with 200 μg 3 days apart for a total of 3 doses. Equal ratios of male and female mice were employed in each experiment, although no sex-specific differences were observed. The Institutional Animal Care and Use Committee (IACUC) of NCH approved all mouse experiments under Protocol AR19-00146. All mice were housed at the ARC facilities at NCH in pathogen-free rooms with requirements for PPE and maintenance of consistent temperature, humidity, and light/dark cycles. All cages were provided with chow and organized in racks with an independent sterile water supply, enrichment, and HEPA-filtered airflow.

Survival study

Gliomagenesis was induced in NTV-a x Xfm pups at 0–2 days, as described. At week 3.5, mice were randomized into 4 treatment groups of approximately 8–12 animals each: vehicle, trabectedin, anti-PD1 ICB, or a combination of trabectedin and anti-PD1 ICB. Mice were injected with trabectedin or vehicle at week 3.5 of glioma progression, as described. Three days following the first injection, anti-PD1 or PBS was administered, as described. Anti-PD1 or PBS was then administered every 3 days for a total of 3 doses. A second dose of trabectedin was injected at week 4.5 of glioma progression. After the full regimen was administered for each treatment group, mice were monitored twice daily and humanely euthanized at the onset of phenotypic decline, with symptoms including lethargy, hyperactivity, seizures, hydrocephalus, abnormal behavior, and/or weight loss.

Mass cytometry

To determine immune cell phenotypes following treatment, cells from the spleen, bone marrow, and brain TME were isolated and stained with a comprehensive panel of antibodies to assess innate and adaptive immune markers (Table S1). Spleens were excised, passed through a 70 μM cell strainer, and washed with PBS. Bone marrow was harvested according to an established protocol.⁴⁸ For TME investigation, the right hemisphere of the brain was excised, passed through a 70 μM cell strainer in PBS containing 1% fetal bovine serum (FBS) and 0.1 mg/mL DNase, and then layered on top of 25% Percoll. The tube was centrifuged at 500 ×g for 20 min with no brake. The supernatant containing myelin and Percoll was aspirated, and the remaining immune pellet was washed twice with PBS. The CyTOF staining and acquisition protocol was performed as described by Canella et al.⁴⁹ Briefly, 1–2 million cells per sample were incubated with Fc-blocking agent (1:100, BD, Frankin Lakes, NJ), stained with the surface antibody cocktail, and incubated at 4°C for 30 min. Cells were then washed twice and fixed with 1.5% paraformaldehyde (PFA). For intracellular staining, cells were permeabilized by resuspension in ice-cold MeOH and incubated for 20 min at −20°C. Cells were washed three times and incubated in the intracellular antibody cocktail for 50 min. Cells were washed twice and fixed in 1.5% PFA with 1:4000 Ir (500 μM) intercalator for 10 min and stored at −80°C until acquisition. Samples were acquired on a Helios mass cytometer (Standard BioTools, San Francisco, CA). Single cells⁵⁰ were gated and analyzed using Omiq (Boston, MA). The gating strategies for the brain TME, spleen, and bone marrow are outlined in Figure S1. Briefly, immune cells in the brain TME were defined as follows: BMDM (CD45^high CD11b+), microglia (CD45^int CD11b+), peripherally derived macrophages (CD45^high CD11b+ F4/80+), mDCs (CD45^high CD11b+ CD11c+ MHCII^high), neutrophils (CD45^high CD11b+ Ly6C- Ly6G+), monocytes (CD45^high CD11b+ Ly6C + Ly6G-), MDSC-like (CD45^high CD11b+ Ly6C + Ly6G+), Tregs (CD3+ CD4+ CD25^high foxp3+), and NK cells (CD11b- CD3- CD335+ CD49b+).

Flow cytometry

Immune cells were isolated from spleens as described in the CyTOF methods. Briefly, cells were washed twice with fluorescence-activated cell sorting (FACS) buffer (PBS +1% FBS). One million cells per sample were counted and stained with Zombie NIR viability stain (1:2500, Biolegend 423105, San Diego, CA) in the dark for 20 min on ice. Cells were washed once with FACS buffer and resuspended in FACS buffer containing murine Fc blocking reagent (BD 553142, 1:100, BD) and incubated in the dark on ice for 20 min. Cells were centrifuged at 300 g and then resuspended in 100 μL of a cocktail of titrated antibodies: CD45-PE (Biolegend 103106, 1:200), CD11b-BV421 (Biolegend 101235, 1:600), F4/80-APC (Biolegend 123116, 1:200), Ly6C-BV510 (Biolegend 128033, 1:200), Ly6G-BV650 (Biolegend 127641, 1:100), and CD115-PECy7 (Biolegend 135523, 1:200). Cells were incubated in the cocktail for 45 min on ice in the dark, washed 2× with FACS buffer, and fixed in 2% PFA. Samples were acquired on a BD Fortessa X20. Data were analyzed using Flowjo and gated on viable singlets. Fluorescence minus one (FMO) controls were for gating.

Luminex

To compare the levels of circulating cytokines in the serum of vehicle- or trabectedin-treated mice, we utilized a Luminex Multiplex assay. Whole blood was harvested from mice 3DPT2 via cardiac puncture and collected in EDTA-coated tubes. The blood was centrifuged at 4,000 RPM for 5 min. The top layer of serum was collected and immediately frozen at −20°C. The Mouse Cytokine 44-Plex Discovery Assay was run by Eve Technologies Corporation (Alberta, Canada). For analysis, the blank was subtracted, and analytes below the range of detection were omitted from analysis. Analytes above the range of the highest standard were replaced with the highest detectable standard in the assay.

Bulk-RNA-seq

Brain tumor lysates were generated from glioma-bearing mice 3DPT2 with trabectedin or vehicle (n = 3 per group) by excising the right hemisphere, disassociating it, and passing it through a 40 μM cell strainer in PBS containing 1% FBS and 0.1 mg/mL DNase, followed by two washes with PBS. RNA was isolated from the tumor lysates and purified using the RNeasy Plus Mini Kit (Qiagen 74134) and the RNA Cleanup and Concentration Kits (23600, Norgen, Ontario, Canada). RNA was submitted to the Genomics Services Lab at NCH for bulk RNA-seq. Strand-specific RNA-seq libraries were prepared using the NEBNext Ultra II Directional RNA Library Prep Kit for Illumina, following the manufacturer’s protocol. A 1,000 ng aliquot of total RNA was rRNA-depleted using NEB’s Human/Mouse/Rat RNAse-H based Depletion kit (New England BioLabs, Ipswich, MA). Following rRNA removal, mRNA was fragmented and used for first- and second-strand cDNA synthesis with random hexamer primers. ds cDNA fragments underwent end-repair and A-tailing and were ligated to dual-unique adapters (Integrated DNA Technologies). Adaptor-ligated cDNA was amplified by limited-cycle PCR. Library quality was analyzed using the Agilent Tapestation High-Sensitivity D-1000 ScreenTape (Agilent Biotechnologies, Santa Clara, CA) and quantified using Quibit. Libraries were subsequently sequenced with 2 × 150 bp read lengths on the Illumina NovaSeq6000 using v1.5 chemistry, generating approximately 60–80 million paired-end reads per sample. Low-quality reads (q < 10) and adapter sequences were eliminated from raw reads using bbduk. Each sample was aligned to the GRCm38.p6 assembly of the Mus musculus reference from NCBI using the RNA-Seq aligner STAR. Features were identified from the GFF file provided with the assembly from Gencode (V28). Feature coverage counts were calculated using featureCounts, and raw counts were normalized using log2-quantile normalization. GSEA was performed using GSEA software v4.3.3 with MsigDB Mouse Collections, including CancerHallmark, GOBP, and Reactome.

WST-1 assay

ATCC glioma cell lines GL261, U87, CT2A, and U251 were utilized to assess the cytotoxicity of trabectedin against glioma cells in vitro. GL261, U87, and U251 were purchased from ATCC, while CT2A was kindly provided by the Cassidy lab at NCH. Cell lines were cultured in DMEM (Thermo Fisher Scientific #12491015, Waltham, MA) supplemented with 10% FBS and 1% penicillin/streptomycin (P/S). For the WST-1 experiment, 2,000 cells per well were plated in a 96-well round-bottom plate in 100 μL of DMEM containing 10% FBS and 1% P/S, with increasing concentrations of trabectedin ranging from 0 nM (untreated) to 5 nM. After 72 h, WST-1 reagent (Sigma Aldrich #05015944001, St. Louis, MO) was added to each well and incubated for 1 h according to the manufacturer’s protocol. Luminescence was measured at 450 nm (target) and 630 nm (reference) using a microplate reader. The blank was subtracted from each measurement, and cell viability was calculated relative to untreated cells. Each concentration was assayed in triplicate, and the experiment was repeated for a total of 3 times. Relative IC50 values were generated using nonlinear regression in GraphPad Prism 10.

Immunohistochemistry

Whole brains were harvested from mice 3DPT2 and fixed in 10% buffered formalin phosphate for 24 h. Tissue processing and IHC staining of the fixed brain were performed by the Histopathology Core at NCH. Briefly, tissues were processed into paraffin-embedded blocks, and 5 μm sections were created. Heat-mediated antigen retrieval was done using EDTA (pH9). Endogenous peroxidase activity was blocked with 3% hydrogen peroxide. Slides were washed in TBST and stained with c-cas3 (R&D Systems MAB835, Minneapolis, MN) at a 1:100 dilution for 2 h at room temperature. Signal Stain Boost IHC detection reagent (HRP, rabbit) was used for 30 min at room temperature, followed by visualization with Vector ImmPACT DAB. Slides were counterstained with hematoxylin, dehydrated through graded ethanol to xylene, and coverslipped. The number of positive pixels was calculated by sampling 5 equal-sized regions within each tumor core and comparing the positive pixel counts per region (n = 2–5 mice per treatment group) using Aperio ImageScope (Leica Biosystems, Deer Park, IL). Tonsil tissue was utilized as a positive control for c-cas3 staining.

Statistics

Statistical analyses and graphs were generated using GraphPad Prism 10. All results reported in the graphs are expressed as the mean ± standard deviation (SD). Two-tailed unpaired Student’s t test were utilized to calculate statistical significance between two experimental groups. Significance is indicated as ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.005, and ∗∗∗∗p < 0.001. Overall survival was measured as the time from gliomagenesis induction to the humane endpoint and shown using Kaplan-Meier curve. Survival analyses were conducted using the Log Rank (Mantel-Cox) test. Sample sizes and statistical tests are described in each figure caption.

Data and code availability

Upon request, the corresponding author (rajappp@ccf.org) will grant access to all data for the scientific community.

Acknowledgments

This study was supported by the U54-CA232561 and R01NS127984 NIH grants. We are grateful to Haley Wrightnour, Miranda Menke, Erin Grove, and all other members of the Animal Resources Core (ARC) at Nationwide Children’s Hospital (NCH) for their help in safely injecting trabectedin and diligently caring for the mice. We thank Diana L. Thomas of the Department of Pathology and Laboratory Medicine at NCH for assisting in tumor confirmation and evaluation in the murine model. We also thank to former Rajappa lab members Sarabeth Anderson and Zoe Tapp for their feedback during the beginning of this study. We acknowledge Marvel Tranquille for their support in analyzing RNA-seq data and Justin Lyberger for their assistance in CyTOF data acquisition. Summary illustrations were created with https://biorender.com. Graphs were generated using GraphPad Prism 10.

Author contributions

M.N., E.R.M., P.R., T.P.C., and D.A.L. contributed to the conceptualization and supervision of the study. Funding acquisition and resources were secured by P.R., T.P.C., E.R.M., and G.B. Data curation was completed by M.N. M.N. conducted the methodology, investigation, and formal analysis of Luminex assays, CyTOF, in vitro experiments, flow cytometry, and murine survival studies. B.B. and O.E. conducted the formal analysis of bulk-RNA-seq data. A.G. and M.C. assisted in the methodology of trabectedin administration, and G.B. assisted with CyTOF. M.N. created the visualization for all figures and wrote the original draft. M.N., E.R.M., A.C., B.B., O.E., G.B., M.C., A.G., D.A.L., T.P.C., and P.R. reviewed and edited the manuscript.

Declaration of interests

The authors report there are no competing interests to declare.