Patient- and xenograft-derived organoids recapitulate pediatric brain tumor features and patient treatments

Abstract

Brain tumors are the leading cause of cancer-related death in children. Experimental in vitro models that faithfully capture the hallmarks and tumor heterogeneity of pediatric brain cancers are limited and hard to establish. We present a protocol that enables efficient generation, expansion, and biobanking of pediatric brain cancer organoids. Utilizing our protocol, we have established patient-derived organoids (PDOs) from ependymomas, medulloblastomas, low-grade glial tumors, and patient-derived xenograft organoids (PDXOs) from medulloblastoma xenografts. PDOs and PDXOs recapitulate histological features, DNA methylation profiles, and intratumor heterogeneity of the tumors from which they were derived. We also showed that PDOs can be xenografted. Most interestingly, when subjected to the same routinely applied therapeutic regimens, PDOs respond similarly to the patients. Taken together, our study highlights the potential of PDOs and PDXOs for research and translational applications for personalized medicine.

Keywords: brain tumors; organoids; patient-derived; pediatric cancer; translational applications.

Figure 1. In vitro culture of patient‐derived organoids (PDOs) and maintenance of tumorigenic potential in vivo

1. A

   Axial T2‐weighted (upper panels) and sagittal T1‐weighted MPRAGE (lower panels) MRI images relative to the indicated cases.

2. B

   Schematic representation of primary tumor samples management workflow.

3. C

   Schematic representation of primary tumor samples management for generation of PDOs as tumor piece.

4. D–F

   Brightfield images of EPN‐ (D), MB‐ (E), and LGG‐ (F) derived PDOs as tumor pieces at different timepoints.

5. G, H

   Confocal images of DAPI staining and immunofluorescence of human nuclear antigen and Ki67 of sagittal brain sections of immunodeficient mice engrafted with EPN‐ (G) and MB‐derived PDOs (H).

Data information: The white square marks the region shown at higher magnification in (G″, H″). Scale bar 200 μm (D–F), 500 μm (G′–H′), 100 μm (G″–H″). Source data are available online for this figure.

Figure EV1. In vitro culture of patient‐derived organoids (PDOs)

1. A

   Schematic representation of primary tumor samples management for generation of PDOs as tumor single cells spheroid.

2. B–D

   Brightfield images of tumor single cells spheroids EPN‐ (B), MB‐ (C), and LGG‐ (D) derived PDOs at different timepoints.

3. E, F

   Confocal images of DAPI staining and immunofluorescence of human nuclear antigen and Ki67 of sagittal brain sections of immunodeficient mice engrafted with MB‐derived PDOs.

4. G, H

   Copy number variation profiles comparison between primary parental tumor and 3 different MB‐PDOs (H) and EPN‐PDOs (G).

Data information: X axis: chromosomes; Y axis: Log₂ copy number ratio. Scale bar 200 μm (B–D), 100 μm (E, F). DNA methylation (CNV) experiments (G, H) were performed once per primary tumor/matching PDOs.

Figure 2. Maintenance of genomic and genetic profiles in PDOs

1. A–C

   Copy number variation profiles comparison between primary parental tumor and EPN‐ (A), MB‐ (B), and LGG‐ (C) derived PDOs.

2. D–H

   Venn diagram of primary tumor/PDOs at different timepoints (D'–H') and relevant shared or new variants (D''–H'') for EPN‐ (D, E) and MB‐ (F–H) derived PDOs.

Data information: X axis: chromosomes; Y axis: Log2 copy number ratio. TMB, tumor mutational burden; MSI, microsatellite instability. DNA methylation (CNV) (A‐C) and TrueSight Oncology (D‐H) experiments were performed once per primary tumor/matching PDOs. Source data are available online for this figure.

Figure EV2. Data from RNA‐based assay showed maintenance of genomic features of fusion‐positive tumors and their corresponding PDOs

1. Tumor #16, LGG with FGFR1‐TACC1 fusion compared to derived PDOs at day 28: the assay detects the same fusion.

2. Tumor #19, KIAA1549‐BRAF fusion was detected in pilocytic astrocytoma and in derived PDOs at day 28. Both PDOs shared the same breakpoints compared to primary parental tumors.

Data information: Experiments were performed once per primary tumor/matching PDOs.

Figure 3. Maintenance of cellular heterogeneity in PDOs

1. A–C

   Confocal images of immunofluorescence of Ki67, SOX2, OLIG2, Nestin, IBA1, CD3, GFAP, B3‐tubulin, CD34 of EPN‐ (A), MB‐ (B) and LGG‐ (C) derived PDOs.

2. D–F

   Quantification in EPN‐ (D), MB‐ (E), and LGG‐ (F) derived PDOs of Ki67⁺, SOX2⁺, OLIG2⁺, and IBA1⁺. Cells are shown as percentage of specific marker⁺ cells/DAPI.

Data information: Data are presented as mean ± s.e.m.; each dot represents a ROI/image. For each marker, n = 5–12 ROI/image of primary tumor was considered. For each marker, n = 2–3 PDOs (biological replicates) were considered; for each PDO, n = 3–4 ROI/image was used. Quantification experiments were performed once per primary tumor/matching PDOs. Kolmogorov–Smirnov test for data with non‐normal distribution; ***P ≤ 0.001, **P ≤ 0.01, *P ≤ 0.05 (D–F). Exact P values are reported in figure. The white arrows highlight specific cells. Scale bar 50 μm (A–C). Source data are available online for this figure.

Figure EV3. Maintenance of cellular heterogeneity in PDOs and comparison with already published medium

1. A–C

   Confocal images of immunofluorescence of Ki67, SOX2, OLIG2, Nestin, IBA1, CD3, GFAP, B3‐tubulin, CD34 of EPN‐ (A′–A″), MB‐ (B) and LGG‐ (C) derived PDOs.

2. D

   Brightfield images of LGG‐derived PDOs as tumor pieces at different timepoints in PDOs medium and cultured according to (Abdullah et al, 2022) (D′), confocal images of immunofluorescence of Ki67, SOX2, OLIG2, Nestin, IBA1, CD3, GFAP, B3‐tubulin (D″) and quantification in PDOs of Ki67⁺, SOX2⁺, OLIG2⁺ and IBA1⁺ cells (D″′).

3. E, F

   Confocal images of immunofluorescence of YAP1 (E′, F′) and p75 NGFR (E″, F″) of MB‐derived PDOs.

4. G, H

   Confocal images of immunofluorescence of synaptophysin of LGG‐derived PDOs. Quantifications are shown as percentage of specific marker⁺ cells/DAPI (D″′).

Data information: Data are presented as mean ± s.e.m.; each dot represents a ROI/image. For each marker, n = 5–7 ROI/image of primary tumor was considered. For each marker, n = 2–3 PDOs (biological replicates) were considered; for each PDO, n = 3–4 ROI/image was used. Quantification experiments were performed once per primary tumor/matching PDOs. Kruskal–Wallis test with Dunn's post hoc correction; **P ≤ 0.01, *P ≤ 0.05. Adjusted and exact P values are reported in figure. The white arrows highlight specific cells. Scale bar 50 μm (A–C, D″, E–H), 200 μm (D′).

Figure EV4. Maintenance of morphological features and cellular heterogeneity in PDOs and PDOs‐derived tumors

1. A–C

   Morphological features (A′, B′, C′) and immunohistochemical expression of lineage markers GFAP (A″), H3K27me3 (A″′), synaptophysin (B″), and OLIG2 (C″) of 2 EPN, 2 MB and 2 LGG paired parental tumors/PDOs samples.

2. D, F

   Confocal images of immunofluorescence of human nuclear antigen and GFAP of sagittal brain sections of immunodeficient mice engrafted with EPN‐ (D) and MB‐ (F) derived PDOs.

3. E, G

   Confocal images of immunofluorescence of human nuclear antigen and OLIG2 of sagittal brain sections of immunodeficient mice engrafted with EPN‐ (E) and MB‐ (G) derived PDOs.

4. H

   Confocal images of immunofluorescence of human nuclear antigen and SOX9 of sagittal brain sections of immunodeficient mice engrafted with MB‐derived PDOs.

Data information: Scale bar 100 μm (A, C, D–H), 50 μm (B).

Figure 4. scRNA‐seq data analysis of primary tumor #15 and matching PDOs samples describes G3 MB‐specific intratumoral heterogeneity, recapitulated in the PDOs model

1. UMAP dimensionality reduction plot showing the cluster distribution of cells obtained from tumor and PDOs samples.

2. UMAP plot showing the different independent clusters obtained by integrating the malignant cells from “Primary tumor #15,” “PDO Day 28,” and “PDO Day 61” datasets.

3. Expression dotplot representing the key markers identified for each cluster and belonging to cellular and/or functional categories.

4. FeaturePlot showing the expression levels of key markers in each cell.

5. Stacked barplot representing the relative proportion (expressed in %) of the tumor and PDO cells across the different subclusters.

Data information: scRNA‐seq experiment was performed once per primary tumor/matching PDOs. Source data are available online for this figure.

Figure EV5. scRNA‐seq data analysis of primary tumor #2 and matching PDOs samples describes EPN PFA‐specific intratumoral heterogeneity, recapitulated in the PDOs model

1. UMAP dimensionality reduction plot showing the cluster distribution of cells obtained from tumor and PDO samples.

2. UMAP plot showing the different independent clusters obtained by integrating the malignant cells from “Primary tumor #2,” “PDO Day 14,” and “PDO Day 28” datasets.

3. Expression dotplot representing the key markers identified for each cluster and belonging to cellular and/or functional categories.

4. FeaturePlot showing the expression levels of key markers in each cell.

5. Stacked barplot representing the relative proportion (expressed in %) of the tumor and PDO cells across the different subclusters.

Data information: scRNA‐seq experiment was performed once per primary tumor/matching PDOs.

Figure 5. In vitro culture of patient‐derived xenograft organoids (PDXOs) and maintenance of genomic aberrations

1. A

   List of MB PDX‐derived tumor samples with information about MB subtypes, patients (gender M: male, F: female; age in months), model name, and methods of processing (spheroid, piece).

2. B

   Schematic representation of PDX‐derived tumor samples management workflow.

3. C

   Schematic representation of PDX‐derived tumor samples management for generation of PDXOs as tumor single cells spheroid.

4. D, E

   Brightfield images of tumor single cells spheroid PDXOs from SHH and G3 MB‐derived PDXs at different timepoints.

5. F

   Schematic representation of PDX‐derived tumor samples management for generation of PDXOs as tumor piece.

6. G, H

   Brightfield images of tumor piece PDXOs from SHH and G3 MB‐derived PDXs at different timepoints.

7. I

   Schematic representation of PDX‐derived tumor samples management for generation of PDXOs as tumor piece in suspension.

8. J, K

   Brightfield images of tumor piece PDXOs in suspension from SHH and G3 MB‐derived PDXs at different timepoints.

9. L, M

   Copy number variation profiles comparison between PDX‐derived tumor and PDXOs from SHH (L) and G3 MB‐derived PDXs (M). CNVs profiles of PDXOs as tumor piece kept in multiwell are shown in (L′–M′), kept in suspension are shown in (L″–M″).

Data information: X axis: chromosomes; Y axis: Log2 copy number ratio. Scale bar 200 μm in (D, E, G, H, J, K). DNA methylation experiments (CNV) (L, M) were performed once per primary tumor/matching PDXOs. Source data are available online for this figure.

Figure 6. ST EPN‐derived PDOs respond to SIOP Ependymoma II, Stratum 3 protocol as the correspondent patient

1. Summary of PDOs generation and treatment according to SIOP Ependymoma II, Stratum 3 protocol (Massimino et al, ; Leblond et al, 2022) for both patient and in vitro adaptation.

2. PDOs live cells analysis with brightfield and fluorescence images (B′) and quantification of area, integrated intensity, and integrated intensity/area after the treatment (B″).

3. Confocal images of immunofluorescence of Ki67 of treated ST EPN‐PDOs.

4. Confocal images of immunofluorescence of cleaved caspase‐3 of treated ST EPN‐PDOs.

5. T2‐weighted MRI coronal images of the patient pre‐surgery, post‐surgery and post‐chemotherapy. Morphological features (E′, E″) and immunohistochemical expression of Ki67 (E″′) of post‐initial surgery and post‐chemotherapy residual tumor. The red asterisk marks the tumor mass.

Data information: Data are presented as mean ± s.e.m.; each dot represents a PDOs. For each treatment condition, n = 8–9 PDOs (biological replicates) were considered. Unpaired t‐test with Welch's correction or Kolmogorov–Smirnov test; *P ≤ 0.05. Exact P values are reported in figure. Scale bar 1 mm in (B′); 100 μm and 20 μm (higher magnification) in (C, D); 50 μm (E″, E″′). CBDCA, carboplatin; CYCLO, cyclophosphamide; MTX, methotrexate; VCR, vincristine. SIOP Ependymoma II, Stratum 3 protocol treatment experiment was performed once in ST EPN PDOs. Source data are available online for this figure.

Figure 7. G3 MB‐derived PDOs respond to high‐risk medulloblastoma protocol as the correspondent patient

1. A

   Summary of PDOs generation and treatment according to high‐risk medulloblastoma protocol (Gandola et al, ; Massimino et al, 2013) for both patient and in vitro adaptation.

2. B

   PDOs live cells analysis with brightfield and fluorescence images (B′) and quantification of area, integrated intensity, and integrated intensity/area after the treatment (B″).

3. C

   Confocal images of immunofluorescence of Ki67 of treated G3 MB‐PDOs.

4. D

   Confocal images of immunofluorescence of cleaved caspase‐3 of treated G3 MB‐PDOs.

5. E, F

   Confocal images of DAPI staining and immunofluorescence of human nuclear antigen and Ki67 of sagittal brain sections of immunodeficient mice engrafted with treatment control G3 MB‐PDOs (E) and treated G3 MB‐PDOs (F). The white square marks the region shown at higher magnification in (E″, F″).

6. G

   T2‐weighted MRI coronal images of the patient pre‐surgery (G′), post‐surgery (G″), post‐chemotherapy (G″′) and post‐radiotherapy (G″″). The red asterisk marks the tumor mass.

Data information: Data are presented as mean ± s.e.m.; each dot represents a PDOs. For each treatment condition, n = 6–10 PDOs (biological replicates) were considered. Ordinary one‐way ANOVA or Kruskal–Wallis test with Dunn's post hoc test; ****P ≤ 0.0001, ***P ≤ 0.001, **P ≤ 0.01, *P ≤ 0.05. Adjusted P values are reported in figure. Scale bar 1 mm in (B′); 100 μm and 20 μm (higher magnification) in (C, D); 500 μm (E′–F′), 100 μm (E″–F″). CBDCA, carboplatin; CCNU, lomustine; CPC, circulating progenitor cells; CYCLO, cyclophosphamide; G‐CSF, granulocyte colony‐stimulating factor; HD, high dose; MTX, methotrexate; VCR, vincristine; VP16, etoposide. High‐risk medulloblastoma protocol treatment experiment was performed once in G3 MB PDOs. Source data are available online for this figure.