A Functional Precision Medicine Pipeline Combines Comparative Transcriptomics and Tumor Organoid Modeling to Identify Bespoke Treatment Strategies for Glioblastoma

Abstract

Li Fraumeni syndrome (LFS) is a hereditary cancer predisposition syndrome caused by germline mutations in TP53. TP53 is the most common mutated gene in human cancer, occurring in 30-50% of glioblastomas (GBM). Here, we highlight a precision medicine platform to identify potential targets for a GBM patient with LFS. We used a comparative transcriptomics approach to identify genes that are uniquely overexpressed in the LFS GBM patient relative to a cancer compendium of 12,747 tumor RNA sequencing data sets, including 200 GBMs. STAT1 and STAT2 were identified as being significantly overexpressed in the LFS patient, indicating ruxolitinib, a Janus kinase 1 and 2 inhibitors, as a potential therapy. The LFS patient had the highest level of STAT1 and STAT2 expression in an institutional high-grade glioma cohort of 45 patients, further supporting the cancer compendium results. To empirically validate the comparative transcriptomics pipeline, we used a combination of adherent and organoid cell culture techniques, including ex vivo patient-derived organoids (PDOs) from four patient-derived cell lines, including the LFS patient. STAT1 and STAT2 expression levels in the four patient-derived cells correlated with levels identified in the respective parent tumors. In both adherent and organoid cultures, cells from the LFS patient were among the most sensitive to ruxolitinib compared to patient-derived cells with lower STAT1 and STAT2 expression levels. A spheroid-based drug screening assay (3D-PREDICT) was performed and used to identify further therapeutic targets. Two targeted therapies were selected for the patient of interest and resulted in radiographic disease stability. This manuscript supports the use of comparative transcriptomics to identify personalized therapeutic targets in a functional precision medicine platform for malignant brain tumors.

Keywords: Li Fraumeni; glioblastoma; organoid; precision medicine; transcriptomics.

Conflict of interest statement

A.M.S., T.M.D. and C.R.T.V. are employees of KIYATEC. All other authors declare no conflict of interest.

Figures

Figure 1

LFS patient was identified as a suitable candidate for comparative transcriptomics. (A) Treatment timeline with relevant MRI scans for the patient of interest (Patient 1) is shown. Where the primary tumor is referred to as 1a, and all additional recurrences are labeled 1b-g in order of occurrence. Histopathology findings of the index tumor identified the mass as a high-grade glioma with diffuse (B), ependymoma-like (C), and pleomorphic xanthrocytoma-like (D) area. Mitotic activity (arrow in B), necrosis (* in C), and pseudorosettes (arrow in C) are present at a magnification of 200×.

Figure 2

Patient 1 index tumor was extensively characterized using RNA and DNA sequencing. (A) TumorMap of Patient 1 relative to the pan-cancer polyA compendium of solid tumors. Spearman correlation was used to identify the top six most correlated samples, which are highlighted with red markers and correlated to tumor type of either ependymoma (purple, 4 tumors), DIPG astrocytoma (yellow, 1 tumor), and glioblastoma (green, 1 tumor). (B) Oncoplot depicting somatic mutations present in the UAMS GBM compendium. Mutations were detected using targeted exome sequencing. Barplot at the top shows total number of mutations per sample. Barplot at the right shows number of affected samples. The stacked barplot at the bottom identifies the proportion of each type of single nucleotide variant present. Sex, tumor grade, MGMT methylation status, and EGFR mutation status (assayed via PCR) are also shown below. All 8 genes displayed were significantly mutated in these samples (p < 0.1).

Figure 3

Comparative transcriptomic analysis of Li Fraumeni glioblastoma elucidates novel tumor characteristics. (A) Comparative transcriptomic ex vivo functional validation pipeline is described. (B) Heatmap identifying STAT1 and STAT2 expression from 45 grade III and grade VI astrocytoma patients from the UAMS compendium. Where * identifies patient samples selected for future validation studies and the red * indicates the patient of interest tumor. Patient 1 was compared to the UAMS comparative transcriptomics compendium and identified as having significantly overexpressed STAT1 (C) and STAT2 (D). Where red line indicates patient of interest, and each dot represents a patient tumor sample. Patient 1 was compared to the UCSC pan-cancer tumor compendium consisting of 12,747 tumors, and STAT1 (E) and STAT2 (F) expression levels were identified. Patient of interest RNAseq data were compared to the UCSC glioblastoma compendium consisting of 200 tumors, and STAT1 (G) and STAT2 (H) expression levels were identified.

Figure 4

Selected patient-derived cell cultures were characterized prior to functional validation. (A) Table identifying tumor location, sex, age, mutation status, and predicted drug targets for the four selected patient cell lines. (B) Heatmap identifying STAT1 and STAT2 expression for each of the four patient cell lines with combined value shown. TumorMap analysis of selected patient cell lines; Patient 1 (C), Patient 17 (D), Patient 23 (E), and Patient 25 (F) with red markers highlighting the top six most correlated cell lines. Colors are correlated with tumor types, as shown next to each TumorMap.

Figure 5

Predicted sensitivity to ruxolitinib was functionally validated. (A) Representative clonogenic assay images for 0 and 100 μM ruxolitinib treatment. (B) Graph of percent survival of colonies after treatment with ruxolitinib at each of the assayed drug concentrations. Where a colony was only scored if it consisted of at least 25 cells. At least three biological replicates were performed for each cell line, and a significant difference in percent survival between cell lines was determined for 100 μM ruxolitinib treatment using a one-way ANOVA, where ** is p ≤ 0.01 and *** is p ≤ 0.001. (C) Treatment scheme for the patient-derived organoid screening method used. (D) Dose-response graphs of the PDO panel. Percent viability was plotted over the log of ruxolitinib concentration to generate EC50 values for each cell line (shown in (E)). The red dotted line indicates 50% viability. Each patient-derived organoid drug screen was performed in at least three biological replicates.

Figure 6

Patient 1 was found to benefit from a combination of JAK/STAT and mTOR inhibition. (A) 3D-PREDICT response panels for tumors 1f and 1g. Yellow indicates drug response, teal indicates moderate drug response, and dark blue indicates no drug response. (B) Graphs of longitudinal log2(TPM+1) gene expression analysis of JAK1, STAT1, STAT2, and mTOR of all LFS patient tumors 1a-1g (color-coded in key above graphs). Gene expression was recorded relative to the UAMS GBM compendium and was plotted as either low, median, or high expression. Dose-response graphs of index (1a) and recurrent (1f and 1g) patient-derived organoids were generated. Percent viability was plotted over the log of ruxolitinib (C) or everolimus (D) concentration to generate EC50 values for each tumor. The red dotted line indicates 50% viability. Each patient-derived organoid drug screen was performed in at least three biological replicates. (E) Heatmap showing EC50 for all tumors after treatment with either ruxolitinib or everolimus. (F) MRI of Patient 1 after resection of tumor 1g. MRI scans were repeated at 2 (G), 3 (H), and 4+ (I) months of concurrent treatment with ruxolitinib and everolimus and were confirmed by radiology to show stable disease in each scan.