TIMEPOINT, a phase 1 study combining MTL-CEBPA with pembrolizumab, supports the immunomodulatory effect of MTL-CEBPA in solid tumors

Abstract

Many patients with cancer do not benefit from currently approved immune checkpoint inhibitors (ICIs), suggesting that additional immunomodulation of the immunosuppressive tumor microenvironment (TME) is required. MTL-CCAAT enhancer-binding protein alpha (CEBPA) specifically upregulates the expression of the master myeloid transcription factor, CEBPA, relieving myeloid-driven immunosuppression. Here, we report the safety, tolerability, pharmacokinetics, and efficacy of MTL-CEBPA in combination with pembrolizumab in patients with advanced solid tumors that typically show ICI resistance. Multimodal exploratory analyses of paired patient biopsies demonstrate biological changes associated with the combination treatment of MTL-CEBPA and pembrolizumab, including increased infiltration of T cell and antigen-presenting cells supporting conversion from an immune-desert toward a more immune-inflamed TME. Patients with disease stabilization demonstrate reductions in immunosuppressive myeloid cells post treatment. Collectively, these data support a role for MTL-CEBPA in reducing immunosuppression in the TME. This study was registered at ClinicalTrials.gov (NCT04105335).

Keywords: biomarker; cancer; gene activation; immunosuppression; immunotherapy; resistance; tumor microenvironment.

Graphical abstract

Figure 1

Best objective tumor response and pharmacokinetics of MTL-CEBPA in combination with pembrolizumab (A) Dosing schedule of MTL-CEBPA/pembrolizumab in the TIMEPOINT clinical trial. (B) Waterfall plot illustrating patient best objective tumor response across treatment groups in phase 1a and phase 1b. Note: an i prefix represents a response evaluated according to the iRECIST or irRECIST criteria. Patients with no post-baseline RECIST assessments are excluded from this figure. iRECIST, immune RECIST; irRECIST, immune-related RECIST; PD, progressive disease; PR, partial response; SD, stable disease; UPD, unconfirmed progressive disease. (C) Plasma CEBPA-51 concentration versus time profiles were collected over 7 days after the first dose of MTL-CEBPA (left) and over 3 days after the second dose of MTL-CEBPA (right). Cohort: 70 mg/m²N = 4, 98 mg/m²N = 3, 130 mg/m²N = 9. Black arrow denotes pembrolizumab dose.

Figure 2

MTL-CEBPA and pembrolizumab combination treatment is associated with changes across tumor types, consistent with immune inflammation and activation (A) Fold change of CEBPA RNA at C1 D2 normalized to pre-treatment values (C1 D1). Bar is at median with 95% confidence error bars, each point is one patient. Significance was determined by one sample Wilcoxon test with hypothetical median value of 1. (B) Fast gene set enrichment analysis (FGSEA) analysis of differentially expressed genes between pre- and post-treatment time points across all patients, testing Nanostring pathways of the IO Pan-cancer panel. Pathways in red reached significance (p adjust <0.05), with significance determined and adjusted for multiple testing by the fgsea algorithm. (C) Quantification of CD3⁺ CD8⁺ T cell numbers (left) and CD3⁺CD8+Ki67+GZMB+ T cell numbers (right) between pre- (screening) and post treatment (C2D16) of 23 paired biopsies as determined by IHC. p values are non-adjusted and are determined by paired two-sided Wilcoxon test. (D) Quantification of CD11b+CD68⁺CD64⁻CD206+CD163⁻ cells as a proportion of total macrophages (CD11b+CD68⁺ cells) (left) and stromal CD11b+CD68⁺CD64⁻CD206+CD163⁻ cell numbers (right) between pre- (screening) and post-treatment (C2D16) of 23 paired biopsies as determined by IHC. p values are non-adjusted and are determined by paired two-sided Wilcoxon test. (E) Quantification of parenchymal CD11b+CD15⁺CD14⁻HLA-DR⁻LOX1+ (PMN-MDSC) cells as a proportion of total myeloid cells (CD11b+ cells) (left), between pre- (screening) and post treatment (C2D16) of 23 paired biopsies as determined by IHC. Right, parenchymal PMN-MDSC cell numbers in patients with at least 1 PMN-MDSC per cell/mm² (N = 18). p values are non-adjusted and are determined by paired two-sided Wilcoxon test. The boxplots in (C), (D), and (E) show the data distribution, where the line denotes the median, the box edges show the interquartile range, and each dot is one patient. (F) Flow cytometric detection of PMN-MDSCs in the blood of TIMEPOINT patients. % change refers to change of PMN-MDSC as a proportion of total live cells (see gating strategy in Figure S2D) compared to C1 D1. Each point is one patient, line is at mean with SD error bars. Significance was determined by Wilcoxon test compared to expected value of 100.

Figure 3

MTL-CEBPA/pembrolizumab combination treatment is associated with immunomodulatory changes in the immune-desert TME, converting to immune-inflamed (A) Fold change in CD3⁺CD8⁺ T cells (left) and CD3⁺CD8⁺GZMB+ T cells (right) at C2D16 compared to screening, between patients with low or high levels of TIL at baseline using the median as threshold at screening. p value was determined by an unpaired two-sided Wilcoxon test. (B) Quantification of Immunosign, IS21 between pre- (screening) and post treatment (C2D16) between cold and hot tumors, left and right, respectively. p values are non-adjusted and were determined by paired two-sided Wilcoxon test. (C) IHC quantification of parenchymal CD3⁺CD8⁺Ki67+ T cells (left) and CD3⁺CD8⁺Ki67+GZMB+ T cells (right) between pre- (screening) and post treatment (C2D16) of 9 paired biopsies with an immune-desert TME (IS21 < 10). p value is determined by paired two-sided Wilcoxon test. (D) IHC quantification of CD11b+CD14⁺CD15⁻HLA-DR+ cells in the whole tumor (left) or stroma (right) between pre- (screening) and post treatment (C2D16) of 9 paired biopsies with an immune-desert TME (IS21 < 10). p value is determined by paired two-sided Wilcoxon test. The boxplots in (A)–(D) show the data distribution, where the line denotes the median, the box edges show the interquartile range, and each dot is one patient. (E) Differential gene expression calculated between pre- and post-treatment time points of 9 paired immune-desert patient tumor biopsies. p values on y axis refer to adjusted p values by DESeq2 algorithm. (F) FGSEA analysis of differentially expressed gene sets between pre- (screening) and post treatment (C2D16) of 9 paired biopsies with an immune-desert TME (IS21 < 10), using supplied Nanostring pathways of the IO Pan-cancer panel. Pathways in red reached significance (p adjust <0.05), adjusted by fgsea algorithm.

Figure 4

Characteristics linking to MTL-CEBPA mode of action are enriched in patients with stable disease after combination treatment (A) Fold change of CEBPA RNA at C1 D2 compared to pre-treatment, C1 D1 in patients with HCC from the OUTREACH clinical trial. Bar is at median with 95% confidence interval error bars, each point is one patient. Significance was tested with Kruskal-Wallis test with p value determined 0.1375. (B) Fold change of CEBPA RNA at C1 D2 normalized to pre-treatment, C1 D1 in iCCA patients from TIMEPOINT. Bar is at median with 95% confidence interval error bars, each point is one patient. Significance was not tested due to low patient numbers. For (A) and (B), clinical outcome refers to clinical response by RECIST or if unavailable, at C2 D22 determined by site. (C) Change in stromal CD11b+CD14⁺CD15⁻HLA-DR+ APC-like cells between pre- (screening) and post treatment (C2D16) calculated independently for patients with disease stabilization and progressive disease. Significance was determined by paired two-sided Wilcoxon test and p values are unadjusted. (D) Change in numbers of parenchymal CD8 T cells (CD3⁺CD8⁺) between pre- (screening) and post treatment (C2D16) in patients with disease stabilization and progressive disease independently calculated. Significance was determined by paired two-sided Wilcoxon test and p values are unadjusted. (E) Change in proportion of CD11b+CD68⁺CD64⁻CD206+CD163⁻ cells as a proportion of total macrophages (CD11b+CD68⁺ cells) between pre- (screening) and post treatment (C2D16) in patients with disease stabilization and progressive disease independently calculated. Significance was determined by paired two-sided Wilcoxon test and p values are unadjusted. (F) Change in proportion of parenchymal CD11b+CD14⁻CD15+HLA-DR-LOX1+ cells (PMN-MDSCs) as a percentage of total myeloid cells (CD11b+) between pre- (screening) and post treatment (C2D16) in patients with disease stabilization and progressive disease calculated independently. Significance was determined by paired two-sided Wilcoxon test and p values are unadjusted. The boxplots in (C)–(F) show the data distribution, where the line denotes the median, the box edges show the interquartile range, and each dot is one patient (N = 23). Clinical outcome of PD or non-PD refers to the clinical response determined by site at C2 D22.

Figure 5

Predictive transcriptomic- and proteomic-based signatures of clinical outcome (A) VST-RUVg visualization of CCGS gene expression (top) and pembrolizumab-related biomarkers (bottom) pre-treatment in 24 patient biopsies, split by progressive disease (red; PD) and non-progressive disease (blue; non-PD). (B) CCGS quantification at pre-treatment between patients with PD and non-PD post treatment. Significance was determined with two-sided Wilcoxon test. Each dot represents one patient (N = 24). (C) Spearman correlation of CCGS at pre-treatment in patient tumor biopsies with % tumor change from baseline at C2 D22 after two cycles of treatment, each point is one patient. (D) Kaplan-Meier plot of progression-free survival with CCGS score, splitting patients by 70th percentile of CCGS. Data are censored as indicated by crosses. Significance was determined by log-rank test. (E) Combined signature of OLINK analysis at pre-treatment between patients with PD and non-PD by site at C2 D22. Significance was determined with an unpaired two-tailed Wilcoxon test. Each dot represents one patient. (F) OLINK plasma analysis of circulating proteins using Inflammation panel of patient plasma at baseline (pre-treatment) in TIMEPOINT (top) as NPX values. Clinical outcome of PD or non-PD refers to the clinical response determined by site at C2 D22 for (A)–(F). Significance was determined with an unpaired two-tailed Wilcoxon. p values are non-adjusted. The boxplots in (B), (E), and (F) show the data distribution, where the center line denotes the median, the box edges show the interquartile range, and each point is one patient.