Single-cell profiling of peripheral blood mononuclear cells from patients treated with oncolytic adenovirus TILT-123 reveals baseline immune status as a predictor of therapy outcomes

Abstract

Oncolytic adenovirus Ad5/3-E2F-d24-hTNFa-IRES-hIL2 (TILT-123, igrelimogene litadenorepvec) shows promise as a therapeutic agent capable of causing tumor regression and activating host immunity. A phase I clinical study TUNIMO (NCT04695327) assessed its safety as monotherapy in patients with various solid tumors. Through single-cell profiling of peripheral blood, we identified distinct immunological features distinguishing responders from non-responders. Specifically, at baseline, responders demonstrated enhanced cytotoxic markers and stronger immune cell communication networks. Moreover, higher baseline CD16+ monocytes correlated with improved survival, while elevated regulatory T cells predicted poor response. T and B cell evaluation revealed contrasting patterns: responders showed higher numbers of T cells with predicted specificity to both adenovirus and tumor antigens, while elevated total memory B cells, regardless of specificity, predicted poor survival. Several T and B cell receptor segments matched those previously reported in other viral infections, suggesting possible cross-reactive immune responses. These findings emphasize that comprehensive biomarker analysis of peripheral blood should include not only cell frequencies but also transcriptional changes and distinct patterns of cellular and humoral immunity.

Fig. 1. Cell clustering and cell dynamics in patients underwent TILT-123 therapy.

A UMAP plot of combined baseline and day 64 PBMCs collected from patients before (BL) and after (D64) they were administered with TILT-123. CM CD4 T = Central Memory CD4 T cells, CM CD8 T = Central Memory CD8 T cells, EM CD8 T = Effector Memory CD8 T cells, Treg = T regulatory cells, cDC = conventional dendritic cells, pDC = plasmacytoid dendritic cells. B Dot plot showing expression of common markers used for clustering annotation. C Cell type frequencies in individual patients at baseline (BL) and after the treatment (D64) grouped based on the treatment response (Responders n = 5, Non-Responders n = 4). Data are shown as violin plots displaying distribution density with median indicated by the central line. Values are expressed as percentage of total cells per sample. D Cell type frequencies in individual patients at baseline (BL) and after the treatment (D64) grouped based on survival length (Longer survivors n = 4, Shorter survivors n = 5). Patients with OS > 12 months were considered Longer survivors. Data are shown as violin plots with median values indicated. Differences between groups (Responders vs. Non-Responders, Longer vs. Shorter survivors) were assessed using unpaired t-test or Wilcoxon test depending on the data distribution. For comparing baseline and day 64 timepoints within groups, paired t-test or Wilcoxon signed-rank test were used. p values < 0.05 were considered significant.

Fig. 2. Transcriptional differences between Responders and Non-Responders at baseline (BL).

A Transcriptional changes in Tregs comparing Responders to Non-Responders. B Transcriptional changes in CD16+ Monocytes comparing long survivors (OS > 12 months) to short survivors. C Volcano plot showing differential gene expression in PBMCs between Responders and Non-Responders. Gray dots indicate genes with p ≥ 0.05, with colored genes showing strongest expression differences. D Heatmap showing top upregulated genes per cluster in Responders at baseline. E KEGG pathway enrichment analysis comparing Responders to Non-Responders in PBMCs. F Cell-cell interaction counts and strength at baseline calculated using CellChat (v1.5.0). G Differential signaling patterns in Responders at baseline, with more active signaling cells shown in red and less active in blue. H Pathway fold changes between Responders and Non-Responders, with significant changes (p < 0.05) shown in red and trending differences (p < 0.1) in gray. I Ligand-receptor pair analysis showing differential interactions between cell populations, with dot size indicating interaction strength and color indicating statistical significance. Differential expression analysis was performed using Wilcoxon test. Cell-cell interaction significance was assessed using permutation tests implemented in CellChat. p values < 0.05 were considered significant.

Fig. 3. Cellular and transcriptional differences between Responders and Non-Responders after TILT-123 administration (day 64).

A Cell type frequencies comparing patients who received lower (Cohorts 1–3, Table 1) versus higher (Cohorts 4–6, Table 1) virus doses. Data are shown as violin plots displaying distribution density with median indicated by the central line. Values are expressed as percentage of total cells per sample. B Correlation between cell percentage fold change at day 64 and overall survival time. C Correlation between cell percentage fold change at day 64 and tumor size change at day 64 (Table 1). D Kaplan-Meier analysis of overall survival based on NK cell fold change at day 64. E Kaplan-Meier analysis of overall survival based on memory T cell fold change at day 64. F Kaplan-Meier analysis of overall survival based on anti-Ad5/3 neutralizing antibody titers at day 64. G Volcano plot showing differential gene expression in Responders comparing day 64 to baseline. Gray dots indicate genes with p ≥ 0.05, with colored genes showing strongest expression differences. H Volcano plot showing differential gene expression in Non-Responders comparing day 64 to baseline. Cell frequencies were compared using unpaired t-test or Wilcoxon test depending on data distribution. For comparing baseline and day 64 timepoints within groups, paired t-test or Wilcoxon signed-rank test were used. Correlations were assessed using Pearson or Spearman correlation based on data normality, with p < 0.1 shown in gray and p < 0.05 in red. Survival analyses were performed using log-rank (Mantel–Cox) test. p values < 0.05 were considered significant.

Fig. 4. B cell receptor analysis and diversity evaluation.

A UMAP plot showing B cell subsets based on immunoglobulin isotype expression. B Percentage of cells expressing different immunoglobulin isotypes (IGHA, IGHM, IGHG, IGHD) in Responders and Non-Responders at baseline and day 64. Data are shown as violin plots displaying distribution density with median indicated by the central line. Values are expressed as percentage of total B cells per sample. C Correlation between percentage of immunoglobulin-expressing cells and overall survival at baseline and day 64. Results with p < 0.1 are shown with gray and statistical significance (p < 0.05) is shown with red. D Correlation between memory B cell numbers at baseline and overall survival. E Kaplan-Meier analysis of overall survival based on baseline anti-Ad5/3 antibodies. F Memory B cell percentage at baseline and day 64. Data are shown as violin plots displaying distribution density with median indicated by the central line. Values are expressed as percentage of total B cells per sample. G Kaplan-Meier analysis of overall survival based on memory B cell fold change at day 64. H BCR clonality analysis showing percentage of unique clonotypes and Shannon diversity index in response groups at baseline and day 64. Data are shown as violin plots displaying distribution density with median indicated by the central line. Values are expressed as percentage of total B cells per sample. I Analysis of immunoglobulin gene segments showing differential expression between response groups. Results with p < 0.1 are shown with grey and statistical significance (p < 0.05) is shown with red. J Distribution of significantly differentially expressed immunoglobulin segments across cell types. Cell frequencies and clonality metrics were compared using t-test or Wilcoxon test depending on data distribution. For comparing baseline and day 64 timepoints within groups, paired t-test or Wilcoxon signed-rank test were used. Correlations were assessed using Pearson or Spearman correlation based on data normality, with p < 0.1 shown in grey and p < 0.05 in red. Survival analyses were performed using log-rank (Mantel–Cox) test. p values < 0.05 were considered significant.

Fig. 5. T cell receptor analysis and predicted antigen specificity.

A TCR clonality analysis showing percentage of unique clonotypes and Shannon diversity index in response groups at baseline and day 64. Data are shown as violin plots displaying distribution density with median indicated by the central line. Values are expressed as percentage of total T cells per sample. B UMAP plot showing distribution of TCR clones across T cell subsets. Clone size is defined as follows: Single—X = 1, Small—1 < X ≤ 5, Medium—5 < X ≤ 10, Large—10 < X ≤ 50, Hyperexpanded—X > 50. C T cell clonality correlation with overall survival at baseline and day 64. D Analysis of TCR segments showing differential expression between response groups. E Percentage of predicted anti-Ad5 specific T cells in Responders and Non-Responders at baseline and day 64. F Percentage of predicted anti-tumor specific T cells in Responders and Non-Responders at baseline and day 64. G Distribution of predicted anti-Ad5 TCRs across T cell subsets. H Distribution of predicted anti-tumor TCRs across T cell subsets. I Correlation between percentage of specific T cells at baseline and overall survival. Cell frequencies and clonality metrics were compared using t-test or Wilcoxon test depending on data distribution. For comparing baseline and day 64 timepoints within groups, paired t-test or Wilcoxon signed-rank test were used. Correlations were assessed using Pearson or Spearman correlation based on data normality. p values < 0.05 were considered significant.