L0 (grade IV glioblastoma from a 43-year-old man) and S7 (grade II glioma from a 54-year-old woman with EGFR amplification) cell lines were isolated and maintained as previously described (26, 33–35). ARL13B-deficient S7 cells were generated using CRISPR/Cas9 as previously described (32). L0 and S7 cells were grown as floating spheres and maintained in NeuroCult NS-A proliferation medium and 10% proliferation supplement (Cat# 05750 and #05753; STEMCELL Technologies, Vancouver, Canada), 1% penicillin–streptomycin (Cat# 15140122; ThermoFisher, Waltham, MA, USA), 20 ng/ml human epidermal growth factor (hEGF) (Cat# 78006; STEMCELL Technologies, Vancouver, Canada), and 10 ng/ml basic fibroblast growth factor (bFGF) (Cat# 78003; STEMCELL Technologies, Vancouver, Canada). For S7 cells, the media were supplemented with 2 μg/ml heparin (Cat# 07980; STEMCELL Technologies, Vancouver, Canada). All cells were grown in a humidified incubator at 37°C with 5% CO₂. When cells reached confluency, or spheres reached approximately 150 μm in diameter, they were enzymatically dissociated by digestion with Accumax (Cat# AM-105; Innovative Cell Technologies, San Diego, CA, USA) for 10 min at 37°C. For human cells grown on glass coverslips, NeuroCult NS-A Proliferation medium was supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Cat# SH30070.03HI; Cytiva, Marlborough, MA, USA).

Primary neural cultures were similar to previously described (32, 36). Briefly, acutely microdissected C57/BL6 mouse cortices from postnatal day 0 to 2 pups were dissected into Gey’s Balanced Salt Solution (Cat# G9779; Sigma-Aldrich, St. Louis, MO, USA) at ~37°C under oxygenation for ~20 min. Dissociated cells were triturated with pipettes of decreasing bore size, pelleted by centrifugation at 1,500 rpm for 3–5 min, and resuspended and plated in glial medium containing DMEM (Cat# SH3002201; Cytiva HyClone, Marlborough, MA, USA), FBS (Cat# 50-753-2981; Gemini BioProducts, West Sacramento, CA, USA), insulin (Cat# 15500; Sigma-Aldrich, St. Louis, MO, USA), Glutamax (Cat# 35050061; Gibco, Waltham, MA, USA), and penicillin–streptomycin (Cat# 15140122; Gibco, Waltham, MA, USA). Cells were plated at a density of 80,000 cells/coverslip on 12-mm glass coverslips coated with 0.1 mg/ml poly-d-lysine followed by 5 μg/ml laminin in minimal essential medium. After approximately 2 h, cells were supplemented with 2 ml neuronal media containing Neurobasal A (Cat# 10888022; Gibco) supplemented with B27 (Cat# A3582801; Gibco), GlutaMAX (35050061; Gibco), kynurenic acid (Sigma-Aldrich, Cat# K3375), and glial cell line-derived neurotrophic factor (GDNF; Cat# SRP3200; Sigma-Aldrich). Every 4 days, half of the media was replaced with fresh neuronal media as described above but lacking kynurenic acid and GDNF. On DIV12, coverslips were transferred into TTFields dishes and fixed after 24 h or 3 days after treatment as described below.

For adherent and spheres, 5 × 10⁴ cells were seeded in 2 ml growth media with or without 10% FBS, respectively. Adherent cells, spheres, or biopsies were placed in TTFields ceramic dishes, each dish approximately the size of a single well of a 6-well plate, and mounted into Inovitro™ base plates (Novocure Ltd., Haifa, Israel). The base plates were connected to a power generator which delivered TTFields at a frequency of 200 kHz at a target intensity of 1.62 V/cm (37). During TTFields treatment, cells were maintained in an incubator (ESCO Technologies, Horsham, PA, USA) with the ambient temperature set to 18°C with 5% CO₂ and a target temperature of 37°C inside each ceramic dish. Treatment duration is as indicated but ranged from 1 to 72 h for a single treatment. To prevent media evaporation during TTFields application, parafilm was placed over each TTFields ceramic dish. In between repeated exposures or for recovery experiments, cells were dissociated and transferred back to a regular incubator. Control samples were grown at 37°C in 5% CO₂ in 6-well plates. In some experiments, we pretreated cells before TTFields with either vehicle or specified drugs. Unless otherwise stated, data in each experiment were pooled from at least 4 dishes per condition and per timepoint.

For timelapse imaging combined with TTFields, we plated 50,000 cells in S7/L0 media supplemented with 5% FBS into 35 mm glass bottom culture dishes (Cat# 81158; Ibidi, Gräfelfing, Germany) which were maintained at 37°C in 5% CO₂. Twenty-four hours before imaging at about 70% confluency, cells were transfected with 500 ng total cDNA/dish of pDest-Arl13b:GFP (gift from T Caspary) and pCMV-myc/mCherry:hOFD1 (Vectorbuilder.com, vector ID: VB201119-1128fyp) using Lipofectamine 3000 (Cat# L3000015; Life Technologies, Carlsbad, CA, USA). A TTFields-delivering ceramic insert was placed into the culture dish and connected to a generator that delivered TTFields at a frequency of 200 kHz at an approximate intensity of 1.2 V/cm and target temperature of 37°C inside each dish. Imaging was conducted on an inverted Zeiss AxioObserver D1 microscope using a Zeiss 40×/0.95 plan Apochromat air objective. The microscope stage was equipped with a Tokai Hit stage incubation system that maintained a humid environment and ambient temperature of 22°C–23°C and 5% CO₂. Baseline images were captured every minute, whereas after TTFields onset, images were collected every 5 min with exposure times ranging in duration from 400 to 750 ms (EGFP) and 300–400 ms (Cy3) per image. Image acquisition and processing were performed using Zeiss ZEN software.

In accordance with our institutional IRB protocol (# 201902489), we collected several fresh, surgically resected tumor biopsies that were subsequently pathologically confirmed. Within 1 h of the resection, biopsies were taken to the laboratory and dissected into several pieces using a sterile scalpel blade. Tissues were immediately fixed and/or transferred into 2 ml of S7 media for culture at 37°C in 5% CO₂ or transferred into TTFields dishes for a 24-h exposure as described above. Following TTFields, control and treated samples were fixed and prepared as described above.

For cell proliferation assay, cells (2.5–5 × 10⁴) were seeded in 2 ml of growth media per well in 6-well plates or in 1 ml of growth media per well in 24-well plates for indicated duration. Cells were then treated with various drugs including chloroquine (CQ) (Cat# C6628; Sigma) (20 µM diluted in sterile water), TMZ (Cat# T2577; Sigma) (0.3 to 100 µM diluted in DMSO), 1,2-bis(o-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA) (Cat# B1205; Sigma) (1 µM diluted in DMSO), or ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid (EGTA) (Cat# RES3010E; Sigma) (0.6 mM diluted in DMSO). After indicated treatment durations, cells were enzymatically dissociated and replaced in ×1 phosphate-buffered saline (PBS). Total cell counts were collected using a Bio-Rad TC20 automated cell counter. Bar graphs show the mean (+/− SEM) and were analyzed statistically using analysis of variance (ANOVA).

For immunocytochemical (ICC) and immunohistochemical (IHC) analyses, samples were fixed at indicated timepoints with 4% paraformaldehyde in 0.1 M phosphate buffer (4% PFA) for 30 min or ice-cold methanol for 15 min (ICC) to 1 h (IHC) and washed with 1× PBS. Spheres or biopsies were cryoprotected in 30% sucrose in PBS followed by a 1:1:30% sucrose and optimal cutting temperature compound (OCT) (Cat# 4585; Fisher Healthcare, Loughborough, UK), frozen in OCT over liquid N₂ and cryosection at 16 µm. Samples were stained for the indicated primary antibodies ( Table S1 ). Samples were incubated in blocking solution containing 5% normal donkey serum (NDS) (Cat# NC9624464; Jackson Immunoresearch, West Grove, PA, USA) and 0.2% Triton-X 100 in 1× PBS for 1 h and then incubated in primary antibodies with 2.5% NDS and 0.1% Triton-X 100 in 1× PBS either for 2 h at room temperature (RT) or overnight at 4°C. For samples stained sequentially with mouse antibodies against gamma- and acetylated alpha-tubulin, samples were blocked with donkey anti-mouse IgG Fab fragments (20 μg/ml; Cat# 715-007-003; Jackson Immunoresearch, West Grove, PA, USA) as previously described (26). Appropriate FITC-, Cy3-, or Cy5-conjugated secondary antibodies (1:1,000; Jackson ImmunoResearch) in 2.5% NDS with 1× PBS were applied for 1–2 h at RT, and coverslips were mounted onto Superfrost™ Plus-coated glass slides (Cat# 12-550-15; Fisher Scientific, Waltham, MA, USA) in Prolong Gold antifade media containing DAPI (Cat# P36935; ThermoFisher). Stained coverslips were examined under epifluorescence using an inverted Zeiss AxioObserver D1 microscope using a Zeiss 40×/0.95 plan Apochromat air objective or a Zeiss 63×/1.4 plan Apochromat oil objective. Images were captured and analyzed using Zeiss ZEN software.

We exposed two patient-derived glioma cell lines, L0 (a grade IV glioblastoma) and S7 (a grade II glioma) that grow primary cilia (26, 29, 31, 32, 38) to TTFields. We used Novocure’s Inovitro™ system to deliver low-intensity (1–4 V/cm), 200 kHz alternating electric fields to cultured cells which presumably mimic the type of fields delivered by the Optune^® device in patients, similar to recent studies (16, 39). Generally, glioma cells were grown adherent (in serum) on coverslips or as free-floating spheres (without serum) for 3 days in vitro (DIV), then performed a single exposure to TTFields for up to 1 day or 3 continuous days at which point we analyzed cells immediately (“acute”) ( Figure 1A ). For repeated TTFields exposures, we dissociated cells after 3 days of continuous TTFields and repeated the cycle 2 more times ( Figure 1A ). For recovery, after the last day of single or repeated TTFields exposures, we dissociated spheres and cultured cells adherently (with serum) on coverslips and examined cells after 4–5 days.

After a single 3-day exposure to TTFields, we immunostained cells for ARL13B and orodigital facial syndrome 1 (OFD1), a protein that concentrates around the basal body (40–42). In control S7 cells, ARL13B+ cilia were readily identifiable extending from OFD1⁺ basal bodies ( Figures 1B , b1 ). After TTFields, the presence of ARL13B⁺ cilia was largely undetected. Cilia that remained were typically elongated ( Figures 1C , c1 ), or appeared detached from ( Figures 1D , d1 ) or dissolved ( Figures 1D , d2 ) around the basal bodies. Most TTFields-treated cilia displayed reduced intensity of OFD1 around the basal body compared with control ( Figures 1 b1, d1, d2 and Figure S1 ). We observed a similar phenomenon in L0 cells ( Figure 1E and Figure S1 ). The appearance ( Figure 1E ) and percent of ARL13B⁺ cilia ( Figure 1F ) in L0 cells were significantly reduced. The effects of TTFields on glioma cilia can be seen within 24 h posttreatment. In both L0 and S7 cell lines, there was a significant loss of ARL13B⁺ cilia after TTFields exposure ( Figures S2A–F ). To confirm that TTFields are affecting the cilium and not just ARL13B localization along the ciliary membrane, we performed triple immunostaining to label a different component of cilia axoneme, acetylated-alpha tubulin (aaTUB) along with gamma-tubulin (gTUB) a microtubule component of the basal body/centriole, and ARL13B. In both L0 and S7 control cells, we found cilia that colocalized aaTUB⁺ and ARL13B⁺ extended from gTUB⁺ basal bodies ( Figure 1G ). However, after TTFields exposure, ARL13B puncta clustered around gTUB⁺ basal body/centrioles without obvious aaTUB⁺ axoneme extending from gTUB⁺ puncta ( Figure 1G ). Quantification of cilia with both aaTUB⁺ and ARL13B⁺ cilia revealed significantly reduced frequencies after TTFields exposure in the two cell lines ( Figure 1H ), indicating that TTFields disrupt the integrity of the entire organelle.

The above observations suggest TTFields effects on the cilia of glioma initiate within hours. Indeed, TTFields have been shown to disrupt glioma cell membrane permeability within the first hour of treatment (16). Thus, we examined the cilia axoneme and membrane 1 and 6 h after TTFields using antibodies against aaTub, gTUB, ARL13B, and inositol polyphosphate-5-phosphatase E (INPP5e). INPP5e localizes to the ciliary membrane where it interacts with ARL13B (22, 43). We found that after 6 h, cilia appeared longer than controls in both cell lines. The elongated ARL13B⁺ cilia displayed underlying colocalization with aaTUB⁺, suggesting that TTFields may stimulate a transient lengthening of the entire organelle within hours ( Figures S3A, C ). We also observed some anomalies in the ciliary distribution of ARL13B and INPP5e staining. ARL13B and INPP5e seemed to distribute evenly along the ciliary axoneme in control, but after 60 min and 6 h exposures to TTFields, the staining pattern appear clustered or polarized toward the proximal and distal tips of the cilium ( Figures S3A–C ). In S7 cells, we also observed unusual clusters of INPP5e surrounding the basal body after TTFields exposure ( Figure S3B ). Thus, in addition to a ciliary lengthening that precedes the loss of cilia, TTFields affect properties of the cilia membrane and surrounding base within hours of exposure in vitro.

The significant depletion of primary cilia by TTFields led us to ask if this effect was permanent. That is, would the frequency of ciliated glioma cells remain low if treatment is stopped and cells are allowed to recover? After single or repeated TTFields exposure, we plated cells in serum for 4 days, fixed, and immunostained for ARL13B alone or combined with pericentriolar material 1 (PCM1), another protein that clusters around the basal body and centrioles in glioma cells (29), or OFD1. In S7 cells, ARL13B+ cilia were detectable but appeared longer than control ( Figures 2A, B ). Quantification of lengths of ARL13B⁺ cilia demonstrated a significant increase in the TTFields-exposed groups ( Figures 2D, E ). A similar increase in length in L0 was observed after recovery from TTFields ( Figures 2C, F ). However, in either S7 or L0 cells, there was no change in cilia frequency after recovery from TTFields ( Figures 2G–I ). These data indicate that frequencies of ciliated glioma cells are restored after TTFields but are affected in a way that leads to elongation.

Considering the robust depletion of glioma cilia within 24 h of exposure to TTFields, we next asked whether cilia of normal primary neural cell types are similarly affected by TTFields. To test this, we cultured dissociated mouse embryonic cortices on glass coverslips for 11 days in vitro (DIV). At 11DIV, we assigned coverslips as control or TTFields (24 h or 3 days exposure). One advantage of this type of culture is that we can examine the effects of TTFields on cells that differentiate into various subtypes including astrocytes and neurons, marked by glial fibrillary acidic protein (GFAP) and neuronal nuclei (NeuN) expression, respectively.

We first examined astrocyte cilia through a combined immunostaining for GFAP, ARL13B, and pericentrin (Pcnt, a protein concentrated around the cilia basal body) ( Figures 3A, D ). After 24 h of TTFields, we did not observe significant differences in the frequency ( Figure 3B ) or length ( Figure 3C ) of astrocyte cilia. However, after 3 days of TTFields, there were significantly fewer ciliated GFAP⁺ cells ( Figure 3E ), though the lengths of these cilia were comparable with control ( Figure 3F ). These data suggest that, at least at acute timepoints after TTFields, astrocyte cilia appear more resistant to TTFields than glioma cells.

Next, we examined neuronal cilia by triple immunostaining for NeuN, type 3 adenylyl cyclase (AC3), an enzyme enriched in most neuronal cilia in the cortex (44–46), and Pcnt (47, 48) ( Figures 3G, J ). After 24 h of TTFields, we did not observe any significant changes on the frequency of neurons with AC3⁺ cilia ( Figure 3H ), but the lengths of AC3⁺ cilia were significantly reduced ( Figure 3I ). After 3 days of TTFields, both the frequency and length of AC3⁺ cilia on NeuN+ cells were significantly reduced compared with control ( Figures 3K, L , respectively). The extent of the reduced ciliary frequency was ~40% of control neurons, compared with about a 90% decrease in L0 cells after similar duration. Thus, neurons, especially after 24 h of TTFields, appear more resistant though not completely spared from the effects of TTFields.

The neural cultures also contained populations of multiciliated cell types (presumably cells that differentiated into ependymal cells) and proliferating cells. However, the cells bearing tufts of cilia, detected by a combination of ARL13B and Pcnt, appeared comparable after 24 h (data not shown) and 3 days of continuous TTFields ( Figure 3M ). A fraction of the cells in the culture were still Ki67⁺, suggesting they were still active in the cell cycle ( Figure 3N ). However, among the Ki67⁺ cells ( Figure 3O ), we did not observe a significant change in the percentage of ciliated cells ( Figure 3P ). Thus, TTFields seems to spare the ability for multiciliated cells and ciliated cycling cells to form or maintain their cilia. Overall, these data suggest a differential sensitivity to TTFields between normal neural cell types and glioma cells.

What is the mechanism through which TTFields promotes rapid cilia loss in glioma cells? A number of factors and pathways promote cilia disassembly (49). Examples include calcium shock/influx (50), or autophagy activation (51), processes that have been shown to rapidly increase after TTFields onset (14, 17). We examined whether buffering extra/intracellular Ca²⁺ by pretreating cells with 600 mM EGTA or 1 µM BAPTA increases cilia frequency during TTFields; however, we did not observe any prevention of cilia loss (data not shown). We then examined if the autophagy pathway changes at cilia were involved, in part because the reduced OFD1 expression we observed around the basal bodies/centrioles after TTFields ( Figure S1 ), is a potential indicator of autophagy activation (42, 52, 53). OFD1 is also an autophagy receptor that controls the early phases of the autophagy cascade and autophagosome biogenesis (54). In addition, we observed microtubule-associated proteins 1A/1B light chain 3B (LC3B) and phospho-AMPK (pAMPK) recruitment to basal bodies after single and repeated TTFields application ( Figure S4 ) consistent with reports that autophagy proteins are recruited to primary cilia (51, 55). Thus, we pretreated S7 and L0 cells 30 min before TTFields induction with vehicle or the autophagy inhibitor chloroquine (CQ) (20 µM), fixed cells after 6 or 24 h, and analyzed the frequency and length of cilia. A concentration of 20 µM CQ was selected because it inhibited autophagy pathway activation in response to TTFields in U87 and other glioma cell lines (14). We found that S7 glioma cells pretreated with CQ 30 min before TTFields led to a significant increase in the percent of ciliated cells after 24 h compared with control ( Figures 4A–D, M ). The TTFields-induced increase in cilia length in S7 cells was also reduced by CQ at 24 h ( Figure 4O ). In L0 cells, we observed significantly more ciliated cells after 6 h ( Figures 4E–H, N ) and 24 h ( Figures 4I–L, N ) pretreatment with CQ and exposed to TTFields. In addition, cilia length of CQ-treated L0 cells at 6 h was significantly reduced compared with vehicle after 6 h TTFields ( Figure 4P ), suggesting autophagy activation may be underlying the observed elongation after TTFields. It is noteworthy that in both S7 and L0 cells, 6 h of TTFields was sufficient to observe significant cilia elongation. Furthermore, because CQ did not fully restore the frequency of cilia to control suggests that either CQ may not inhibit autophagy in all cells, or that the activation of autophagy may represent one of the factors resulting in TTFields-induced cilia depletion.

To more directly examine how ciliated cells respond to TTFields, we transfected L0 and S7 cells with two cDNA constructs encoding ARL13B:GFP and OFD1:mCherry allowing us to track isolated cells displaying ARL13B:GFP⁺ cilia with OFD1:mCherry⁺ clusters around the basal body overnight ( Figure S5A ). We live-imaged cells up to 24 h after transfection using Novocure’s Inovitro LIVE imaging system. Notably within several hours after TTFields onset, we observed ciliated L0 cells that appeared to die ( Figure S5B, C ), with similar observations in S7 cells during TTFields ( Figure S5D ). These data indicate TTFields may have a direct impact on the survival and proliferation of ciliated glioma cells or cells that are derived from ciliated glioma cells.

The survival benefit promoted by TTFields in patients occurs during TMZ maintenance therapy (3, 4). TMZ has recently been reported to increase the frequency and length of ARL13B+ cilia patient-derived glioblastoma cells (30). Thus, we asked if the effects of TTFields on cilia would be similar in the presence of TMZ chemotherapy. The effects of TMZ on glioma ciliogenesis have not been extensively analyzed with respect to different cell lines and different concentrations and durations of exposure. Thus, we first examined how different durations and concentrations of TMZ, doses lower than those typically used to kill cells in in vitro assays, affect the frequency and length of ARL13B+ glioma cilia in our cell lines.

In L0 and S7 cells, 10 µM TMZ appeared to elongate primary cilia in both cell lines 24 h after treatment ( Figures 5A, B ). In S7 cells, we observed a dose-dependent effect with 0.3, 3, and 10 µM TMZ sufficient to increase cilia length 24 h after exposure ( Figure 5C ). In L0 cells, we found that 10 µM TMZ significantly increased the length of ARL13B⁺ cilia compared with lower TMZ concentrations and vehicle ( Figure 5D ). Three days after exposure to TMZ (a duration chosen because we performed 3 days of continuous TTFields), we found that in S7 cells, 10 and 50 µM TMZ significantly reduced the length of cilia ( Figures 5E, G ) but significantly increased the frequency of ciliated cells ( Figure 5H ). However, 3 days after TMZ exposure in L0 cells did not affect cilia length after 10 or 50 µM ( Figures 5F, I ) but significantly increased the frequency of ciliated L0 cells ( Figure 5J ) as in S7 cells. These results support results of recent studies (30) and suggest TMZ is capable of stimulating the elongation of ARL13B⁺ cilia, at least acutely, and increasing the frequency of ciliated glioma cells.

Since TMZ generally stimulates ciliogenesis, and TTFields inhibit it, we examined glioma cilia with a combination of these treatments in adherent cells and spheres ( Figure 6 ). First, we pretreated adherent S7 and L0 cells with concentrations of TMZ that stimulated ciliogenesis about 30 min before a 3-day exposure to TTFields. We found that in both adherent S7 ( Figures 6A –D ) and L0 ( Figures 6E –H ) cells, ciliogenesis was not observed in the presence of TMZ plus TTFields ( Figures 6D, H ). We also examined if TTFields had the same effect in gliomaspheres. We cultured S7 and L0 spheres for 3 days, and then treated them for 3 days with vehicle, 50 µM TMZ, TTFields, or TTFields plus 50 µM TMZ (added 30 min before onset) ( Figures 6I –L ). We then collected and fixed spheres, and sectioned and immunostained them for ARL13B. For each sphere, we normalized the number of cilia to the area of the sphere. In S7 cells, we found that TMZ alone increased the frequency of cilia ( Figure 6M ), consistent with what we observed in adherent cells ( Figure 5E ). As expected, TTFields significantly reduced the frequency of cilia in the spheres but TMZ-induced increase did not occur in the presence of TTFields ( Figure 6M ). Interestingly, we did not observe a change in the length of cilia across groups ( Figure 6N ). Unlike S7 cells, TMZ alone did not increase the frequency of cilia in L0 spheres ( Figure 6O ), and TMZ plus TTFields cotreatment increased cilia lengths compared to vehicle plus TTFields ( Figure 6P ). However, like S7 cells, TTFields reduced the frequency of cilia in the L0 spheres which was not enhanced by the cotreatment of TMZ plus TTFields ( Figure 6O ). Together, these results indicate that TTFields disrupt the prociliogenic effects of TMZ.

Previously, we found that deleting key ciliogenesis genes (e.g., PCM1, KIF3A) enhanced sensitivity of glioma cells to TMZ (29). Similarly, glioma cells expressing ARL13B shRNA, which depleted cilia, were more sensitized to TMZ in vitro and in vivo (30). Using our S7 glioma transgenic cell line depleted in ARL13B and cilia using CRISPR/Cas9 (32), we examined how these cells proliferated in response to TMZ, TTFields, or TMZ and TTFields cotreatment. After 4 days of growing S7 parental or ARL13B KO spheres, we treated them with vehicle or 50 or 100 µM TMZ in the absence or with 3 days of TTFields ( Figures 7A , B ). In parental S7 cells, proliferation was reduced in 100 µM TMZ group ( Figure 7A ). However, in ARL13B KO cells, proliferation was significantly reduced at 50 and 100 µM TMZ groups ( Figure 7A ), consistent with results of previous studies that ARL13B⁺ cilia are associated with resistance to TMZ. However, when we cotreated S7 parental or ARL13B KO cells with TMZ and TTFields, there was no added toxicity in either group ( Figure 7B ). This suggests that in cells with cilia ablation via treatment with TTFields or thru genetic means, the cotreatment of TMZ and TTFields may not lead to an acute additive toxicity.

Surprised that 3 days of TMZ and TTFields cotreatment in S7 parental cells did not show additive toxicity and trended towards increased proliferation ( Figure 7B ), we wondered if potential cotoxicity relates to treatment sequence or the treatment effect could be delayed. For example, would stimulating ciliogenesis with TMZ render more cells susceptible to TTFields, and/or would TTFields suppression of ciliogenesis sensitize more glioma cells to TMZ? To test this, we administered TMZ before (PRE) ( Figures 7C, D ), after (POST) ( Figures 7E, F ), or during (WITH) ( Figures 7G, H ) a 24-h window of TTFields application. In the PRE-experiment, we found that the acute numbers of TMZ + TTFields-treated cells were similar to vehicle (Veh) + TTFields ( Figure 7C ). However, the fold expansion of TMZ + TTFields-treated cells 7 days after treatment was significantly reduced compared with Veh + TTFields ( Figure 7D ). In the POST-experiment, we also did not observe an acute reduction of TTFields + TMZ-treated cells compared with control ( Figure 7E ) but did observe a significant reduction of TTFields-TMZ-treated cells 7 days after treatment ( Figure 7F ). Surprisingly, in the “WITH” experiment, we observed an acute increase in TMZ + TTFields-treated cells compared with Veh + TTFields ( Figure 7G ), and a significant increase in the expansion of TMZ + TTFields cotreated cells 7 days after treatment ( Figure 7H ). Thus, TMZ given pre- or post-TTFields slows tumor cell recurrence, but coadministration enhanced tumor cell recurrence. We next asked if either of the treatment paradigms that slowed tumor cell recurrence correlated with the presence of ARL13B⁺ cilia. Since we observed a significant effect of giving TMZ after TTFields on parental cells, we examined fold expansions of S7 ARL13b KO cells 6 days after TTFields + TMZ treatment ( Figure 7I ). However, there was no significant difference in the expansion between TTFields + TMZ-treated and TTFields + Veh-treated ARL13B KO cells ( Figure 7I ). These results suggest that the relative timing of TMZ exposure to TTFields impacts subsequent tumor cell expansion in vitro that also correlates with the presence of ARL13B⁺ cilia.

While we currently do not have the technical capability to test TTFields in an intracranial, xenograft model, we asked if the effects of TTFields on cilia we observed in adherent or spheres of glioma cells are detectable within the patient tumor microenvironment. To test this, we divided fresh biopsy samples into 3 groups: immediate/acute fixation, 24-h control, or 24 h of TTFields. We then fixed and immunostained cryosections of biopsies ( Figure 8A ). First, we examined a subependymoma (a grade 1 glioma), a tumor type reported to possess cilia (56). We found that ARL13B⁺ cilia extending from OFD1⁺ basal bodies were readily detectable in control ( Figure 8B ) whereas we only clearly observed OFD1⁺ basal bodies in the TTFields-treated biopsy ( Figure 8C ). We also received newly diagnosed GBM biopsies from a 34-year-old man and a 66-year-old man which we treated similarly except for the immunostained basal bodies/centrioles with gTUB and cilia with ARL13B antibodies. We observed gTUB⁺ basal bodies with ARL13B⁺ cilia in acutely fixed tissue and in overnight control ( Figures 8D, E, G ), whereas the cilia appeared blunted or generally reduced in TTFields-treated tissue ( Figures 8F, G ). Similar observations were made in a GBM biopsy of a 66-year-old man exposed to 24 h of TTFields ( Figure S6 ). These data are consistent with our cultured adherent cells and gliomaspheres, raising the possibility that TTFields disrupts ARL13B⁺-ciliated tumor cells within the tumor.