Unless otherwise stated, all chemicals were purchased from Sigma Aldrich (Taufkirchen, Germany).

The cell lines 1321N1 and U-251MG were obtained from Sigma Aldrich (Taufkirchen, Germany), and the cell lines U87, T98G, and LN229 from the American Type Culture Collection (ATCC; Manassas, United States). MZ54, G55T2, and MZ18 were initially obtained from Donat Kögel (Frankfurt, Germany), and the lines LN405 and U-343MG from the German Collection of Microorganisms and Cell Cultures (DMSZ; Braunschweig, Germany). Table 1 provides a detailed description of all cell lines used. Cells were cultured at 37°C, 5% CO₂, and 95% air in DMEM (Dulbecco’s Modified Eagle Medium containing 4.5 g glucose/mL) supplemented with 2 mM Glutamax, 1% Penicillin/Streptomycin (all from Gibco Life Technologies, now Thermo Fisher Scientific, Darmstadt, Germany) and 10% Fetal Bovine Serum (FBS; Biochrom GmbH, Berlin, Germany), hereafter referred to as “culture medium.” This culture medium was used for all experiments, including the formation of spheroids. Confluence was checked by phase contrast microscopy.

Single cells were harvested in their exponential growth phase at 75%–85% confluence using Accutase (Gibco), and the cell number was determined using an automatic cell counter (TC10™ automatic cell counter; Bio-Rad Laboratories GmbH, Feldkirchen, Germany). Spheroids were developed in 96-well ultra-low attachment plates (Corning^® 96-well Black/Clear Round Bottom Ultra-Low Attachment Spheroid Microplate, with Lid, Sterile, Corning Incorporated, Corning, NY, United States). A specific number of cells (see details for each experiment) were pipetted into each well in 150 µL of culture medium. The plates were centrifuged at 241 × g for 5 min, transferred to a cell incubator, and incubated at 37°C, 5% CO₂ and 95% air. (Note: The seeding of cells is day zero). Every 3–4 days, 50% of the medium was removed, and fresh medium was added. 2D cultures were performed in 96-well plates (µClear; Greiner BioONE, Frickenhausen, Germany) under the same culture conditions.

Staining with propidium iodide (PI) was performed by adding 0.1 μg/mL PI to the cells at the beginning of the experiments. We used a concentration well below the commonly used concentration of 0.5 μg/mL (or even 10 μg/mL) to avoid toxicity by PI (Krämer et al., 2016), continuously monitoring necrosis without producing artifacts by staining. Imaging was performed with a BZ-X810 Microscope (Keyence, Neu-Isenburg, Germany), using a TexasRed-Filter (OP-87765, Keyence, EX 560/40 nm, DM 585 nm, BA 630/75 nm) for PI fluorescence staining. Phase contrast and PI images were taken at ×20 magnification.

The radii of the spheroids were measured using the BZ-X800 Analyzer Software (Keyence). Spheroids’ total volume and the inner core of dead cells (stained by propidium iodide) within the spheroids were calculated based on the radii (r) according to the following formula: V = 4 3 π r 3

The BZ-X800 Analyzers Stitching Tool generated a single image for measuring spheroids larger than a field of view. (Note: In rare cases, when spheroids deviated significantly from a spherical appearance, these structures were omitted from analysis.) At this point, it should be noted that our oversimplified method of determining spheroids’ volume might need to be revised to determine the response of spheroid volume to drugs. However, our method proved sufficient to show the relation between volume and cell number and volume and incubation time.

The size of individual cells from different cultures was determined from cells harvested in their exponential growth phase at 75%–85% confluence. Trypan blue staining was employed to exclude dead cells from the analysis. The cells were applied to Dual-Chamber Cell Counting Slides (Bio-Rad) to perform microscopy using a Keyence microscope. Cell radii were measured by the BZ-H4XI/Image Cytometer Module (Keyence), which allows setting a lower bound for the exclusion of debris and an upper bound for the exclusion of artifacts. In addition, it allows the separation of doublets. Size determinations were also controlled visually for possible doublets and other artifacts.

From each well containing a spheroid in 150 µL culture medium, 50 µL of the medium was removed and replaced with 50 µL of fresh medium containing 600 µM Temozolomide (TMZ) in DSMO, resulting in 200 µM TMZ per Well. Note: We used a concentration of 200 µM because we evaluated this concentration as appropriate in previous experiments with 2D cultures (Oppermann et al., 2019). However, one should consider that this concentration is ∼3–4 times higher than the peak plasma concentration obtained in a patient receiving the standard dose of 150–200 mg/m²/day (Ortiz et al., 2021). As a control, 50 µL of medium containing only DMSO was used. Single dose irradiation (8 Gy) was performed 2 h after TMZ was added to the cells using a 200 kV X-ray machine (Xstrahl, Ratingen, Germany) at a dose rate of 1.43 Gy/min (150 kV, 10 mA).

The viability of cells within a spheroid was visualized using PI staining. The CellTiter-Glo 3D Cell Viability Assay (Promega, Mannheim, Germany) was used according to the manufacturer’s recommendations for quantitively determining spheroids' viability. In brief, an equal volume of CTG-3D reagent was added to each well and then mixed by pipetting and shaking for 5 min. After incubation at room temperature for 25 min, luminescence was determined using a SpectraMax M5 (Molecular Devices, Biberach, Germany). Cell viability in 2D culture was determined using the CellTiter-Glo Cell Viability Assay (Promega) according to the manufacturer’s recommendations.

The statistical analysis was performed with SPSS [IBM, Armonk, United States; version: 28.0.0.0 (190)]. For multiple comparisons, a one-way ANOVA with a Games-Howell or Bonferroni post hoc test was performed after the test for equality of variance (Levene test). The results were considered statistically significant at a value of p < 0.05. Graphical representations were created using OriginPro (2021b; OriginLab Corporation, Northampton, MA, United States) and CorelDraw Graphics Suite 2020 (Corel Corporation, Ottawa, Canada).

The following strategy was used to analyze whether treatment had a statistically different effect on cells in 2D culture than on cells in 3D culture: Cells or spheroids were treated with TMZ, radiation, or a combination of both treatment modalities (all n = 6). The effect of the treatment was compared with untreated control cells or spheroids (defined as 100% viability). After determining the mean value for each treatment group, each value (n = 6 for each modality) was normalized to this mean value. These normalized data were used for a one-factorial ANOVA after a Kolmogorov-Smirnov test, followed by a Levene test to decide whether a Games-Howell or Bonferroni post hoc correction should be applied to determine significance.

The correct spelling and grammar were checked using DeepL (DeepL SE, Cologne, Germany) and Grammarly (Grammarly Inc., San Francisco, United States).

First, we investigated the spheroid size as a function of the initial number of cells. Therefore, 250, 500, 1,000, 2000, 5,000, and 15,000 cells were used to form spheroids. The size was measured after 24, 48, 72, 96, 120, 144, and 168 h. The results of the experiment are shown in Figures 1, 2. In addition, pictures from spheroids can be seen in Supplementary Material 1. As can be seen in Figure 1, the size of the spheroids appears to be a linear function of the initial number of cells in this double logarithmic presentation. A linear fit using the original non-logarithmic data confirmed a linear relation in most cases (see Table 2; all adjusted R ² values for the linear fit, including intercept and slope are presented in Supplementary Table 1). However, some lines, such as U87, G55T2 and T98G deviate from a linear relation with increasing incubation time. These lines also respond to increased incubation time with an increased slope (pL Spheroid/cell number) of the linear regression curves whereas the other lines show a tendency towards decreased slopes. More important, as shown in Figure 2, which plots spheroid volume as a function of cultivation time, the spheroids of lines LN405, MZ18, and MZ54 do not grow, regardless of the initial number of cells used. The data in Figure 2 also show that growth is generally better with a lower initial number of cells. While spheroid growth of U-251MG cells can only be observed from 250 or 500 cells, spheroids of U87 cells still show some growth, even from 15,000 cells. Although spheroid growth differs between the lines, the best growth can generally be observed at an initial cell number between 250 and 500.

As microscopy of cells in 2D culture indicated that cells of different origins differ in volume and size distribution, we wondered whether the volume of the single cells may influence the formation and growth of spheroids. We especially considered, that a cell line that shows a high variety of cell volumes within a culture may form spheroids that have a different growth behavior compared to a cell line with cells that do not show a high variety of different cell volumes. Therefore, we measured cell volume and size distribution of cells after their detachment from culture flasks before they were used to produce spheroids. The different sizes determined and their distribution are presented in Figure 3, along with example images used to analyze cell volumes. In general, we found no apparent correlation between the individual cell size of a line and the resulting spheroid volume. However, there was a tendency for lines with smaller cells and a smaller range of volumes to be better suited for spheroids when higher numbers of cells are desired.

As demonstrated by others, spheroids consist of an outer layer of living and proliferating cells, an intermediate layer of quiescent cells, and an inner core of dying or dead cells [e.g.,: (Sutherland et al., 1971; Onozato et al., 2017)]. To investigate whether the different lines employed in our study form spheroids with comparable cores of dead cells surrounded by living cells, we performed a microscopic analysis after staining dead cells with propidium iodide (PI). Therefore, we added PI at a 0.1 μg/mL concentration before seeding the cells for spheroid formation (2,500 and 5,000 cells; N = 5 each). The spheroids were grown for 48, 96, and 168 h and received fresh medium (50%) after 96 h. Representative images from microscopy are shown in Figure 4. As can be seen, almost all spheroids show an inner core of cells stained by PI and an outer rim of cells without staining. Staining with PI differed between the different cells; for example, cells from lines LN229 and U87 exhibited low staining at all time points, whereas a strong signal was seen in spheroids from lines such as LN405 and 1321N1. Remarkably, in some lines, staining with PI becomes more intense with increasing incubation time, indicating increasing cell death in the inner core (for example, U-251MG and U-343MG). We also tried to determine the ratio between living cells (unstained) and dead cells (stained by PI) in a spheroid. This analysis pointed towards the notion that the number of living cells in a spheroid is usually less than 10% (see Supplementary Table 2). However, as determining the border between stained and unstained cells is difficult there is a high uncertainty in this estimation.

Finally, to determine whether spheroids respond differently to treatment than 2D cultures of the same cell line, we performed an experiment in which we treated both types of cultures with TMZ and irradiation, using cell lines G55T2, U87, and U-343MG. For this study, we created spheroids and adherent cultures using either 500 or 1,000 cells per culture well. The initial numbers were derived from Figure 2, which revealed that these lines exhibited good spheroid growth at these initial densities. After a growth period of 48 h, cells received 50 µL of fresh medium containing TMZ (200 µM final concentration) or a vehicle control without the drug. Two hours later, the cultures were irradiated (8 Gy) or sham-irradiated (0 Gy). The cells were grown for a further 5 days. During this time, selected wells were analyzed by microscopy (Figure 5). In addition, we determined the size of the spheroids of Figure 5 (data presented in Figure 6). As can be seen in Figure 6, spheroids from all lines show strong size differences at day 4 and 5. However, it should be noted that in some cases size determination was difficult because some spheroids responded to treatment with blurred edges (e.g., U87 spheroids from 1,000 cells ad day 4; Figure 5). Finally, the viability of all cultures was determined by measuring ATP in lysates after 168 h of growth (Day 5/120 h after treatment). The result of the analysis is presented in Figure 7. For the comparison, we tested whether an effect of a treatment in 2D culture is statistically different from the effect of the same treatment in 3D culture and indicated this by asterisks (e.g., *** above the column “2D culture of G55T2 cells” starting with 1,000 cells (lower left panel) indicates that the effect of irradiation alone (Rx) is significantly different (p > 0.0005) between 2D and 3D culture (in this case, the effect is less pronounced in 2D culture); data is also available as Supplementary Material 2 (for details of the calculation see Materials and Methods; “Statistical analysis and graphical representation”). As can be seen, cells of line U87 provide the same result regardless of the model used. The same is true for G55T2 cells when seeded at an initial density of 500 cells. In contrast, at a density of 1,000 cells, significantly stronger effects were observed in 2D culture compared to 3D culture after treatment with irradiation (Rx) or a combination of irradiation with TMZ (RCT). This different behavior could indicate that the spheroids respond more strongly to irradiation at a higher cell density.

However, it should be noted that the observed qualitative effect is the same in both models, demonstrating that irradiation has a more substantial effect on the cells than TMZ alone, and the combination of both treatment modalities has a more substantial effect than either modality alone, regardless of the initial number of cells. An even more significant difference between the responses of cells in 2D and 3D cultures can be observed for U-343MG. Here, in 2D culture, TMZ alone is more effective compared to Rx and RCT, which is not observed in 3D culture. This effect is independent of the initial number of cells used for the experiment.