Additional Supporting Information may be found in the online version of this article.

ANA_22036_sm_SupFig1.tif11208KSupplementary Fig 1: Rationale of the present study and procedure for derivation of tissue. (A) Cartoon showing a model for brain tumor initiation and spread, including a hypothetical point of origin (red), the GBM center (green), and the GBM periphery (orange). For general illustration, a GBM sample (patient 002) was processed for p53 immunolabeling (DAB, brown). Note that accumulation of the p53 gene product is neither seen in every GBM sample, nor a mandatory feature of every GBM cell. In the presented case, however, the immune reaction exposes malignant cells. (B) Macroscopic appearance of the surgery site. From each patient, two tissue specimens were obtained representing routinely resected GBM tissue during standard neurosurgery (GBM center, c), and, after completion of standard neurosurgery a second, experimental biopsy was obtained from tissue surrounding the resection cavity (GBM periphery, p), respectively. (C) MRI analysis before (T1-weighted images) and after surgery (T2-weighted images) of the patient shown in (B). Note that asterisks in (B) and (C) highlight the respective biopsy sites.
ANA_22036_sm_SupFig2.tif2192KSupplementary Figure 2: Comparative analysis of multiple biopsies. In some patients (#'s: 023, 046, 066, 078 and 106), neurosurgical procedures allowed the separation and the individual procurement of multiple, tissue samples from the same tumor region (i.e. in situ located > 1cm apart from each other). These samples were propagated and characterized in analogy to the routine methods applied in this study. Histology confirmed - for all of these samples, INF characteristics in biopsies from the resection margin, and GBM-typical characteristics in tumor center biopsies (in analogy to Fig 1 and Supplementary Table 1). Side-by-side analysis revealed some degree of heterogeneity in every patient investigated, however, in every case, residual cells could clearly be distinguished from GBM center cells. Moreover, data on the kinetics of cell expansion (A), the degree of self-renewal (B), and the profile of molecular markers (C) revealed ‘intra-individual’ profiles of residual vs. center GBM cells that closely resembled the findings presented in Figures 1 and 2 for multiple patients. Asterisks indicate statistically significant (p<0.05) findings.
ANA_22036_sm_SupFig3.tif2208KSupplementary Figure 3: In vitro drug and irradiation challenge. (A) Pairs of adherent passage-7 cells were investigated in 96-well plates. These included in vitro application of irradiation (5 Gy), temozolomide (500?μM), lomustine (CCNU; 380?μM), and combinations thereof. Metabolic activity of cells was recorded five days after application using the alamarBlueR assay. Representative cell morphologies at the time of recording are presented as insets (differential interference contrast; Co=control/untreated; TMZ=temozolomide). Data showed that in 16/25 measurements, GBM periphery cells responded dissimilar (modalities indicated in red) from the corresponding center cells. Note: the rationale for conducting the in vitro drug and irradiation challenge was to gather additional evidence for functional dissimilarities between center and residual GBM cells.
ANA_22036_sm_SupFig4.tif2021KSupplementary Figure 4: Distribution of tissue and cells to the experimental paradigms. Flow chart describing the stepwise protocol applied in this study.
ANA_22036_sm_SupFig5.tif6406KSupplementary Figure 5: Immediate in vitro analysis. As a first step of comparative analysis immediate evaluation of single cell suspensions (18-paired samples; 001-013; 023; 025; 026; 029; 030) was conducted in defined non-adhesive culture conditions (see Supplementary Methods above). (A) Under these conditions, center biopsy cells (17/18 samples) formed spherical structures, while specimens from the tumor periphery appeared less organized. The formation of spherical structures was observed in only 4/18 periphery samples. (B) Viable cell counting at the end of the observation period demonstrated similar ratios of surviving cells in center vs. periphery biopsy-derived cultures. However, case-specific comparison revealed in 8/18 measurements that cell numbers varied more than 50% between the respective pairs of cultures (inset: C=center; P=periphery; mean value ± standard error). These findings, in addition to the distinct growth patterns (i.e. the ability to form spherical structures) served as an initial indication for intra-individual and regional diversity of GBM cells. For a more standardized comparative evaluation and to estimate the content of self-renewing and multipotent cells in paired biopsies, however, we additionally applied a standardized neurosphere assay (see Fig 2). Scale bar in (A), 200 μm
ANA_22036_sm_SupTab1.pdf56KSupplementary Table 1: Patient data. List of patients and tissue specimens investigated. Representative fragments of each biopsy sample used for in vitro analysis underwent parallel histological analysis. Classification of tissue was performed applying standard H&E-, and immunohistochemistry-profiling with antibodies against GFAP, MAP2c, and Ki67. Specifically, tissue showing vascular proliferation and necrosis, a presence of pleomorphic glial tumor cells, and an abundance of mitotic/proliferative activity was classified as GBM. Tissue exhibiting an increase in cellular density with abnormal grouping of pleomorphic glial cells and occasional mitotic figures was classified as glioma infiltration zone (INF). According to this histological classification, 30/36 (83%) of the GBM center biopsy samples showed exclusive morphological GBM signs. 32/38 (84%) periphery biopsy samples exhibited exclusive morphological signs of an INF, and in 11 of a total of 74 obtained biopsies (15%) mixed GBM and INF characteristics were detected. Exclusive INF features were never observed in center biopsy samples; however, exclusive GBM characteristics could be demonstrated in 1/38 (3%) periphery biopsy samples. Clinical data include the prognostic groups according the RTOG RPA classificationSuppl7-9, the primary treatment schemes, and the clinical courses of each patient investigated.
ANA_22036_sm_SupTab2.pdf24KSupplementary Table 2: Primers. List of primers used for quantitative RT-PCR studies. Note that for Musashi-1, CD133, uPAR, and hGAPDH commercially available, pre-designed sets of primers were in part used (Invitrogen-ID is indicated for purchase information).
ANA_22036_sm_SupApp.doc64KSupplementary data

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.