Abstract
Diffuse gliomas, particularly glioblastomas, are incurable brain tumours1. They are characterized by networks of interconnected brain tumour cells that communicate via Ca2+ transients2,3,4,5,6. However, the networks’ architecture and communication strategy and how these influence tumour biology remain unknown. Here we describe how glioblastoma cell networks include a small, plastic population of highly active glioblastoma cells that display rhythmic Ca2+ oscillations and are particularly connected to others. Their autonomous periodic Ca2+ transients preceded Ca2+ transients of other network-connected cells, activating the frequency-dependent MAPK and NF-κB pathways. Mathematical network analysis revealed that glioblastoma network topology follows scale-free and small-world properties, with periodic tumour cells frequently located in network hubs. This network design enabled resistance against random damage but was vulnerable to losing its key hubs. Targeting of autonomous rhythmic activity by selective physical ablation of periodic tumour cells or by genetic or pharmacological interference with the potassium channel KCa3.1 (also known as IK1, SK4 or KCNN4) strongly compromised global network communication. This led to a marked reduction of tumour cell viability within the entire network, reduced tumour growth in mice and extended animal survival. The dependency of glioblastoma networks on periodic Ca2+ activity generates a vulnerability7 that can be exploited for the development of novel therapies, such as with KCa3.1-inhibiting drugs.
This is a preview of subscription content, access via your institution
Access options
Access Nature and 54 other Nature Portfolio journals
Get Nature+, our best-value online-access subscription
$29.99 / 30 days
cancel any time
Subscribe to this journal
Receive 51 print issues and online access
$199.00 per year
only $3.90 per issue
Buy this article
- Purchase on Springer Link
- Instant access to full article PDF
Prices may be subject to local taxes which are calculated during checkout
Similar content being viewed by others
Data availability
Bulk RNA-seq data of connected and unconnected glioma cells have been deposited in the Sequence Read Archive (SRA) database under accession number PRJNA554870. Bulk RNA-seq data of cells treated with senicapoc and control have been deposited in the Gene Expression Omnibus under accession number GSE215365. Source data are provided with this paper.
Code availability
The code supporting the current study has been deposited in Zenodo at https://doi.org/10.5281/zenodo.7242228.
References
Weller, M. et al. Glioma. Nat. Rev. Dis. Primers 1, 15017 (2015).
Osswald, M. et al. Brain tumour cells interconnect to a functional and resistant network. Nature 528, 93–98 (2015).
Venkataramani, V. et al. Glutamatergic synaptic input to glioma cells drives brain tumour progression. Nature 573, 532–538 (2019).
Venkatesh, H. S. et al. Electrical and synaptic integration of glioma into neural circuits. Nature 573, 539–545 (2019).
Gritsenko, P. G. et al. p120-catenin-dependent collective brain infiltration by glioma cell networks. Nat. Cell Biol. 22, 97–107 (2020).
Winkler, F. & Wick, W. Harmful networks in the brain and beyond. Science 359, 1100–1101 (2018).
Albert, R., Jeong, H. & Barabasi, A. L. Error and attack tolerance of complex networks. Nature 406, 378–382 (2000).
Weil, S. et al. Tumor microtubes convey resistance to surgical lesions and chemotherapy in gliomas. Neuro Oncol. 19, 1316–1326 (2017).
Lee, J. et al. Tumor stem cells derived from glioblastomas cultured in bFGF and EGF more closely mirror the phenotype and genotype of primary tumors than do serum-cultured cell lines. Cancer Cell 9, 391–403 (2006).
Watts, D. J. & Strogatz, S. H. Collective dynamics of ‘small-world’ networks. Nature 393, 440–442 (1998).
Barabasi, A. L. & Albert, R. Emergence of scaling in random networks. Science 286, 509–512 (1999).
Milo, R. et al. Network motifs: simple building blocks of complex networks. Science 298, 824–827 (2002).
Dupont, G., Combettes, L., Bird, G. S. & Putney, J. W. Calcium oscillations. Cold Spring Harb. Perspect. Biol. 3, a004226 (2011).
Chen, Y. J. et al. The potassium channel KCa3.1 constitutes a pharmacological target for neuroinflammation associated with ischemia/reperfusion stroke. J. Cereb. Blood Flow Metab. 36, 2146–2161 (2016).
D’Alessandro, G. et al. KCa3.1 channels are involved in the infiltrative behavior of glioblastoma in vivo. Cell Death Dis. 4, e773 (2013).
Turner, K. L., Honasoge, A., Robert, S. M., McFerrin, M. M. & Sontheimer, H. A proinvasive role for the Ca2+-activated K+ channel KCa3.1 in malignant glioma. Glia 62, 971–981 (2014).
Ruggieri, P. et al. The inhibition of KCa3.1 channels activity reduces cell motility in glioblastoma derived cancer stem cells. PLoS ONE 7, e47825 (2012).
D’Alessandro, G. et al. KCa3.1 channel inhibition sensitizes malignant gliomas to temozolomide treatment. Oncotarget 7, 30781–30796 (2016).
Wang, H. Y. et al. A three ion channel genes-based signature predicts prognosis of primary glioblastoma patients and reveals a chemotherapy sensitive subtype. Oncotarget 7, 74895–74903 (2016).
Smedler, E. & Uhlen, P. Frequency decoding of calcium oscillations. Biochim. Biophys. Acta 1840, 964–969 (2014).
Parekh, A. B. Decoding cytosolic Ca2+ oscillations. Trends Biochem. Sci. 36, 78–87 (2011).
Wacquier, B., Voorsluijs, V., Combettes, L. & Dupont, G. Coding and decoding of oscillatory Ca2+ signals. Semin. Cell Dev. Biol. 94, 11–19 (2019).
Taniguchi, K. & Karin, M. NF-κB, inflammation, immunity and cancer: coming of age. Nat. Rev. Immunol. 18, 309–324 (2018).
Vogelstein, B. et al. Cancer genome landscapes. Science 339, 1546–1558 (2013).
Dolmetsch, R. E., Xu, K. & Lewis, R. S. Calcium oscillations increase the efficiency and specificity of gene expression. Nature 392, 933–936 (1998).
Kupzig, S., Walker, S. A. & Cullen, P. J. The frequencies of calcium oscillations are optimized for efficient calcium-mediated activation of Ras and the ERK/MAPK cascade. Proc. Natl Acad. Sci. USA 102, 7577–7582 (2005).
Eshete, F. & Fields, R. D. Spike frequency decoding and autonomous activation of Ca2+-calmodulin-dependent protein kinase II in dorsal root ganglion neurons. J. Neurosci. 21, 6694–6705 (2001).
Tompa, P., Töth-Boconádi, R. & Friedrich, P. Frequency decoding of fast calcium oscillations by calpain. Cell Calcium 29, 161–170 (2001).
Jin, L. W. et al. Repurposing the KCa3.1 inhibitor senicapoc for Alzheimer’s disease. Ann. Clin. Transl. Neurol. 6, 723–738 (2019).
Staal, R. G. W. et al. Inhibition of the potassium channel KCa3.1 by senicapoc reverses tactile allodynia in rats with peripheral nerve injury. Eur. J. Pharmacol. 795, 1–7 (2017).
Ataga, K. I. et al. Improvements in haemolysis and indicators of erythrocyte survival do not correlate with acute vaso-occlusive crises in patients with sickle cell disease: a phase III randomized, placebo-controlled, double-blind study of the Gardos channel blocker senicapoc (ICA-17043). Br. J. Haematol. 153, 92–104 (2011).
Maezawa, I., Jenkins, D. P., Jin, B. E. & Wulff, H. Microglial KCa3.1 channels as a potential therapeutic target for Alzheimer’s disease. Int. J. Alzheimers Dis. 2012, 868972 (2012).
Kaushal, V., Koeberle, P. D., Wang, Y. & Schlichter, L. C. The Ca2+-activated K+ channel KCNN4/KCa3.1 contributes to microglia activation and nitric oxide-dependent neurodegeneration. J. Neurosci. 27, 234–244 (2007).
Weisbrod, D. et al. SK4 Ca2+ activated K+ channel is a critical player in cardiac pacemaker derived from human embryonic stem cells. Proc. Natl Acad. Sci. USA 110, E1685–E1694 (2013).
Blankenship, A. G. & Feller, M. B. Mechanisms underlying spontaneous patterned activity in developing neural circuits. Nat. Rev. Neurosci. 11, 18–29 (2010).
Venkataramani, V. et al. Glioblastoma hijacks neuronal mechanisms for brain invasion. Cell 185, 2899–2917.e2831 (2022).
Gerdes, J. et al. Cell cycle analysis of a cell proliferation-associated human nuclear antigen defined by the monoclonal antibody Ki-67. J. Immunol. 133, 1710–1715 (1984).
Neftel, C. et al. An integrative model of cellular states, plasticity, and genetics for glioblastoma. Cell 178, 835–849.e821 (2019).
Acknowledgements
We thank K. Frey for technical assistance and M. Fischer for acquiring the MRI images; O. Bracko for the training in cranial window preparation and in vivo multi-photon imaging; A. Schulz and M. Vogel for RNA-sequencing services; and M. Brom and D. Krunic for their help with confocal microscopy. This work was supported by the Deutsche Forschungsgemeinschaft (SFB 1389, UNITE Glioblastoma, project ID 404521405, to W.W., F.W., E.J., T. Kessler, M.O.B., M.R., F.S., and M.S., and Emmy Noether programme DFG to M.O.B., BR 6153/1-1), the Heinrich F. C. Behr Foundation to D.H. and the German Research Foundation (VE1373/2-1), Else Kröner-Fresenius-Stiftung (2020-EKEA.135), the Stiftung SET to F.W. P.K. and the University of Heidelberg (Physician Scientist-Programm and Krebs- und Scharlachstiftung) to V.V., and the Bundesministerium für Bildung und Forschung within the framework of the e:Med research and funding concept to M.A.K. (01ZX1913D).
Author information
Authors and Affiliations
Contributions
D.H. designed, conducted and analysed experiments and contributed to all aspects of the study, in particular confocal and in vivo multi-photon imaging of glioma network activity, proliferation, and invasion, immunofluorescence, immunochemistry, cranial window implantation, tumour implantation, quantification and analysis of the data, data interpretation, writing the code, TCGA data and RNA-expression data analysis. D.H. initially discovered the intrinsically rhythmic cells. D.H. and F.W. wrote the manuscript with the input of all co-authors. D.C.H. performed sample preparation and transcriptional analyses. D.C.H., V.V. and E.J. provided conceptual and methodological input and data interpretation. V.V. and S.K.T. performed electrophysiological recordings under the supervision of T. Kuner. S.H. and A.J. conducted brain organoid experiments under the supervision of P.K. D.D.A and S. Weil performed cranial window implantation and tumour injections and provided conceptual input. L.H. and T. Kessler performed bioinformatic analysis of RNA-expression data. T. Kessler provided conceptual input. A.K. provided the KCa3.1-knockout constructs. P.S. and A.H. provided staining of human paraffin sections under the supervision of F.S. M.O.B. provided MRI and subsequent analysis. M.A.K. and M.R. provided conceptual input. J.M.M., Y.Y. and E.R. performed tumour injections. S. Wendler and C.L. performed cell culture work, C.L. and C.M. performed immunostaining. K.F. and O.G. provided the Twitch-3A vector. M.O. provided in vivo Ca2+ imaging data and conceptual input. G.S. performed in vivo Ca2+ imaging. M.S. provided conceptual input for network analyses. W.W. provided conceptual input, performed data interpretation and supervised RNA-expression data analysis. F.W. conceptualized and supervised all aspects of the study and performed data interpretation.
Corresponding author
Ethics declarations
Competing interests
E.J., M.O., W.W. and F.W. are inventors on patent no. WO2017020982A1 titled ‘Agents for Use in the Treatment of Glioma’. F.W. reports a research collaboration with DC Europa Limited, GlaxoSmithKline, Genentech and Boehringer. F.W. is co-founder of DC Europa Limited. M.R. reports a research grant from Novocure and honoraria, consulting fees and lecture fees from Novocure and Alexxion (now AstraZeneca). G.S. is an employee of Carl Zeiss Microscopy GmbH. F.S. received honoraria from Bayer and Illumina. W.W. is an inventor and patent holder on ‘Peptides for use in Treating or Diagnosing IDH1R132H Positive Cancers’ (patent no. EP2800580B1) and ‘Cancer Therapy with an Oncolytic Virus Combined with a Checkpoint Inhibitor’ (US11027013B2). He has consulted for Apogenix, AstraZeneca, Bayer, Enterome, Medac, MSD and Roche/Genentech, with honoraria paid to the Medical Faculty at the University of Heidelberg.
Peer review
Peer review information
Nature thanks Adam Cohen, Heike Wulff and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data figures and tables
Extended Data Fig. 1 Glioblastoma network communication depends on TMs and gap junctions and drives tumor cell proliferation.
a,b, Time-lapse images of Ca2+ transients travelling between two glioblastoma cells along a TM in the mouse brain in vivo; S24 line (a) and BG5 line (b); scale bars are 25 µm (a) and 15 μm (b); region of a is indicated in Fig. 1a, region of b is indicated in c by dotted box; arrows denote Ca2+ peaks in cell somata; arrowheads denote intercellular Ca2+ transient traveling through TMs. c, Network plot of coactive cell pairs derived from Ca2+ recordings of BG5 tumor cells; arrowheads indicate the direction of the transients. d, Synchronized Ca2+ transients of TM-connected tumor cells. Red: Ca2+ transients detectable; black: no Ca2+ peaks detectable. e, same as (a), but with a novel in vitro tumor cell network growth assay; S24 line; scale bar is 25 µm; region is indicated in Fig. 1c. f—h, same as b—d, but with a novel in vitro tumor cell network growth assay; scale bar in d is 15 μm. i, Number of coactive cells versus number of anatomical connections per cell; S24 line in vitro, n = 1602 cells from 4 recordings and 2 biologically independent experiments; linear regression. j, Fraction of cells with at least one coactive cell out of all active cells (here and below cells with ≥ 4 Ca2+ peaks). k,l, Normalized fraction of coactive cell pairs in in vivo (k) and in vitro (l) data and its corresponding null control; Kruskal-Wallis test, Dunn’s test for in vivo data and one-way ANOVA, Dunnett’s test for in vitro data; j—l, n = 23 recordings from 3 mice (S24 in vivo), n = 22 recordings from 9 biologically independent experiments (S24 in vitro), n = 40 recordings from 4 mice (BG5 in vivo), and n = 10 recordings from 5 biologically independent experiments (BG5 in vitro). m, Representative traces of Ca2+ transients of cells stained with Rhod-2AM and imaged with multiphoton microscopy in vivo and confocal microscopy in vitro (see Methods for more information); note that Ca2+ peak morphology is very similar in vivo and in vitro when the same Ca2+ sensor and similar recording technologies are used. n, Intercellular coactivity in Ca2+ recordings of control versus 100 µM Carbenoxolone (Cbx, gap junction inhibition), 100 µM MFA (gap junction inhibition), 1 µM Latrunculin B (LatrB, reduction of TM growth), and 100 µM Suramin (purinergic receptor inhibition) treatment in vitro; intercellular coactivity was calculated by dividing the number of coactive cells of a recording by the number of coactive cells of its corresponding null control data; note that the inhibition of extracellular transfer of ATP by Suramin increased the intercellular coactivity, possibly by reducing extracellular noise, and thereby revealing the pure, TM-mediated and highly synchronized Ca2+-mediated communication of the tumor cells; Kruskal-Wallis test, uncorrected Dunn’s test. o, Normalized global Ca2+ activity in control versus 100 µM Carbenoxolone (Cbx), 100 µM MFA, 1 µM Latrunculin B (LatrB), and 100 µM Suramin treatment in vitro; global calcium activity (number of Ca2+ peaks per time and cell) was normalized by dividing the number of peaks per cell and time by the mean of the corresponding control recording; one-way ANOVA, Dunnett’s test. n, o, n = 22 recordings (S24; control), n = 25 recordings (S24; Cbx), n = 6 recordings (S24; MFA), n = 5 recordings (S24; LatrB), n = 8 recordings (S24; Sur), n = 10 recordings (BG5; control), n = 4 recordings (BG5; Cbx, MFA, and Sur), n = 7 recordings (BG5; LatrB) from ≥ 2 biologically independent experiments, respectively. p, Normalized global Ca2+ activity in control versus SOCE inhibitor BTP2 treatment in vitro; n = 22 recordings from 9 biologically independent experiments (S24, control), n = 4 recordings from 2 biologically independent experiments (S24, 2 µM, 5 µM, and 10 µM; BG5, 2 µM, 5 µM, and 10 µM); one-way ANOVA, Dunnett’s test. q, Fraction of EdU-positive cells, measured as number of EdU-positive cells divided by DAPI-positive cells in control versus BTP2 treatment in vitro; n = 26 recordings (S24; Control and BTP2 2 µM), n = 23 recordings (S24; BTP2 5 µM), n = 29 recordings (S24; BTP2 10 µM), n = 27 recordings (BG5; Control and BTP2 2 µM), and n = 25 recordings (BG5; BTP2 5 µM and 10 µM) from 2 biologically independent experiments; one-way ANOVA, Dunnett’s test. r, Fraction of dead cells, measured as number of PI-positive cells divided by DAPI-positive cells in control versus BTP2 treatment in vitro; n = 18 recordings (S24; all groups), n = 24 recordings (BG5; Control and BTP2 2 µM), n = 17 recordings (S24; BTP2 5 µM), n = 22 recordings (S24; BTP2 10 µM) from 2 biologically independent experiments; Kruskal-Wallis test, Dunn’s test. s, Global Ca2+ activity in control versus 10 µM BAPTA-AM treatment in vitro; n = 11 recordings (S24, control), n = 4 recordings (S24, BAPTA-AM), n = 5 recordings (BG5, control), and n = 4 recordings (BG5, BAPTA-AM) from 2 biologically independent experiments; Brown-Forsythe and Welch ANOVA test, Dunnett’s test. t, Fraction of EdU-positive cells measured as the number of EdU-positive cells divided by DAPI-positive cells in control versus BAPTA-AM treatment in vitro; n = 30 recordings (S24, control and BAPTA-AM; BG5, control) and n = 29 recordings (BG5, BAPTA-AM) from 3 biologically independent experiments; Kruskal-Wallis test, Dunn’s test. u, Fraction of dead cells, measured as number of PI-positive cells divided by DAPI-positive cells in control versus BAPTA-AM treatment in vitro; n = 18 recordings (S24, control), n = 18 recordings (S24, BAPTA-AM) from 2 biologically independent experiments, n = 18 recordings (BG5, control), and n = 18 recordings (BG5, BAPTA-AM) from 2 biologically independent experiments; one-way ANOVA, Dunnett’s test. v, Fraction of EdU-positive cells and w, fraction after of dead cells, measured as number of PI-positive cells divided by DAPI-positive cells after control versus MFA treatment (gap junction inhibition) in vitro; n = 18 recordings per group from 2 biologically independent experiments, respectively; one-way ANOVA, Dunnett’s test. x, Fraction of activated caspase-3-positive tumor cells in tumor-bearing human cerebral organoids after 14 days of control versus 100 µM MFA treatment; n = 6 regions in 3 cerebral organoids per group; S24 line; two-sided Mann-Whitney test. Error bars show s.e.m. ns, not significant (P ≥ 0.05).
Extended Data Fig. 2 Characteristics and targeting of periodically active cells.
a, Ca2+ traces from a representative recording in vitro; traces of periodic cells are indicated by thick red lines; BG5 line. b,c, Representative recording of S24 (b) and BG5 (c) tumor cells in vitro; arrow indicates typical periodic cell with a high number of anatomical (TM-mediated) cell-cell connections. d, Number of TM-mediated cell-cell connections of periodic cells and non-periodic cells in vitro. b—d, n = 120 periodic cells and n = 1482 non-periodic cells from 4 recordings from 3 biologically independent experiments (S24); n = 40 periodic cells and n = 1364 non-periodic cells from 3 recordings from 2 biologically independent experiments (BG5); Kruskal-Wallis test, Dunn’s test. e, Histogram of the standard deviation (std) of the peak-peak-intervals of all active cells; red bars indicate cells with periodic Ca2+ activity; S24 line in vivo; n = 2198 cells from 24 recordings in 3 mice. f, Histograms of the frequency of Ca2+ peaks of all active cells; red bars indicate cells with periodic Ca2+ activity; n = 2198 cells (S24 in vivo) from 24 recordings in 3 mice. g, Fraction of periodic cells of all cells in empirical (control) and scrambled data versus 100 µM Carbenoxolone (Cbx; gap junction inhibition), 100 µM MFA (gap junction inhibition), and 1 µM Latrunculin B (LatrB; reduction of TM growth) treatment in vitro; n = 12 recordings (S24; control), n = 10 recordings (S24; Cbx), n = 4 recordings (S24; MFA), n = 5 recordings (S24; LatrB), n = 19 recordings (BG5; control), n = 4 recordings (BG5; Cbx, MFA), n = 7 recordings (BG5; LatrB) from n ≥ 2 independent experiments, respectively; Kruskal-Wallis test, Dunn’s test. h, Ca2+ traces from a representative recording after gap junction inhibition with 100 µM MFA in vitro; traces of periodic cells are indicated by thick red lines; BG5 line. i, Representative Ca2+ transients of periodic cells (dark red, upper lines) and anatomically connected non-periodic cells (lighter red and black) in vitro; S24 line. j, Number of Ca2+ peaks of all cells as a function of the degree of separation to the closest periodic cell, which corresponds to the number of cell-cell connections between the cell and the closest periodic cell in vitro; BG5 line; n = 40 cells (0), n = 297 cells (1), n = 320 cells (2), n = 304 cells (3), n = 196 cells (4), n = 98 cells (5), n = 48 cells (6), n = 31 cells (7), n = 14 cells (8), n = 17 cells (9), n = 11 cells (10), n = 6 cells (11), n = 3 cells (12), n = 20 cells (inf.) from 3 recordings from 3 independent experiments; Kruskal-Wallis test, Dunn’s test; **P < 0.01; ***P < 0.001; ****P < 0.0001. k, Directionality of the Ca2+ signals between communicating cells for periodic cells and non-periodic cells in vitro; n = 555 periodic cells and n = 2719 non-periodic cells from 22 recordings from 9 biologically independent experiments (S24) and n = 189 periodic cells and n = 3291 non-periodic cells from 10 recording from 5 biologically independent experiments (BG5); Kruskal-Wallis test, Dunn’s test. l, Fold change of global Ca2+ activity, m, fold change of fraction of periodic cells of all active cells, and n, fold change of frequency of Ca2+ oscillations of periodic cells after control (n = 49 and n = 22) versus BTP2 (n = 4 and n = 4), Cbx (n = 9 and n = 4), Gap19 (n = 4 and n = 3), TTX (n = 4 and n = 3), Verapamil (n = 4 and n = 3), Mibefradil (n = 4 and n = 3), U73122 (n = 2 and n = 3), EGTA (n = 2 and n = 3), Thapsigargin (n = 2 and n = 3), TRAM-34 (n = 5 and n = 5), Senicapoc (n = 12 and n = 11), Suramin (n = 8 and n = 4), ATP (n = 4 and n = 3), EGF (n = 4 and n = 4), FGF (n = 3 and n = 4), Latrunculin B (n = 5 and n = 7), and Y-27632 (ROCKI) (n = 4 and n = 4) treatment in vitro; see the Supplementary Discussion for mechanistic insights and Methods for respective concentrations; U73122, EGTA, Thapsigargin were not included in (n), as there was no Ca2+ activity in the presence of these drugs and therefore no periodic cells were detected; error bars show s.d.; ‘n’ is provided for S24 and BG5, respectively from n ≥ 2 independent experiments per group; one-way ANOVA, Uncorrected Fisher’s LSD (l) and Brown-Forsythe and Welch ANOVA, Dunnett’s test (m,n). o,p, Fraction of periodic cells of all cells before and after laser ablation; after laser ablation of random cells the number of periodic cells increased (only significant in o) and after laser ablation of all periodic cells, the number of periodic cells was not zero as expected, because new periodic cells appeared in the recordings, that were, however, not able to compensate the loss of the previous periodic cells (Fig. 3h, i, Extended Data Fig. 4f–h); S24 line (o) and BG5 line (p); n = 6 recordings (BG5 control and random), n = 7 (S24 control; BG5 periodic), and n = 8 (S24 random and periodic) from 2 biologically independent experiments; Mixed-effects analysis, Holm-Šídák test. g, j—p, Data are presented as mean values ± s.e.m. ns, not significant (P ≥ 0.05).
Extended Data Fig. 3 Tumor cell networks have scale-free and small-world properties, and their functional network hubs frequently display periodic calcium oscillations in vivo and in vitro.
a—c, Mean probability distribution of coactive cells per cell from in vitro and in vivo Ca2+ imaging plotted in a log-log scale with linear regression fit in red and the fitted random Poisson distribution in green; the grey box indicates highly connected network hubs; Pearson correlation. d, ‹k› (mean number of connections per cell) and σ (standard deviation of number of connections per cell) of all recorded networks from each condition. For a random network with Poisson degree distribution the standard deviation of the degrees follows σ = ‹k›1⁄2 shown as a red dashed line on the figure. For each network σ is larger than the value expected for a random network with the same ‹k›, which leads to the manifestation of scale-free properties in the networks7. e—h, Network parameters λ (mean shortest path length) and σ (clustering coefficient) calculated for recorded networks and their corresponding Erdős-Rényi random networks of equal number of nodes and edges; error bars show s.e.m.; paired two-sided t-test. a—h, n = 24 recordings from 3 mice (S24 in vivo), n = 22 recordings from 9 biologically independent experiments (S24 in vitro), n = 40 recordings from 4 mice (BG5 in vivo), and n = 10 recordings from 5 biologically independent experiments (BG5 in vitro). i—n, Network plot of cross-correlation coefficients larger than cut-off derived from the same Ca2+ recordings of tumor cells; BG5 line in vitro (i,l); S24 line in vivo (j,m); BG5 line in vivo (k,n); scale bars are 100 μm (i) and 50 μm (j,k). o—q, Ca2+ transients of indicated TM-connected tumor cells; dark red transients originated from a periodic cell. ns, not significant (P ≥ 0.05).
Extended Data Fig. 4 Ablation of periodically active cells reduces network integrity.
a, Fraction of periodic cells of all active cells and all network hubs in experimental data in vivo and all cells in scrambled data; n = 24 recordings from 3 mice (S24) and n = 40 recordings from 4 mice (BG5); error bars show s.e.m.; Mixed-effects analysis, Dunnett’s test. b, Venn diagram of active cells, communicating cells, periodic cells, and functional network hubs; periodic cells are solely defined by their low peak-peak-interval variability (bright red) and convey their effect on the network by often being highly connected network hubs (dark red). c, Number of ablated periodic cells and randomly chosen active cells per FOV in vitro (comparable in both groups); n = 6 recordings (BG5 control and random), n = 7 (S24 control; BG5 periodic), and n = 8 (S24 random and periodic) from 2 biologically independent experiments; two-sided t-test. d, Fraction of activated caspase-3-positive cells (i.e., cells undergoing apoptosis) in ablated and non-ablated cells; all ablated cells were apoptotic; n = 6 recordings per group from 2 biologically independent experiments; two-sided t-test. e−g, Global Ca2+ activity and number of communicating cells per FOV of Ca2+ recordings in vitro before and 1 h after cell-specific laser ablation of either no cells (control), all periodic cells, or the corresponding number of randomly chosen active cells; BG5 line, n = 6 recordings (control and random ablation) and n = 7 recordings (ablation of periodic cells) in 2 biologically independent experiments (e); S24 line, n = 7 recordings (control) and n = 8 recordings (random and periodic) in 2 biologically independent experiments (f,g); one-way ANOVA, Dunnett’s test. h, Fraction of dead cells, measured as number of PI-positive cells (dead cells) divided by DAPI-positive cells (nuclei), 24 h after the cell-specific laser ablation of either periodic cells (periodic), the corresponding number randomly chosen active cells (random), or no cells (control); S24 line; n = 6 recordings per group from 2 biologically independent experiments; Kruskal-Wallis test, Dunn’s test. c—h, Error bars show s.d. ns, not significant (P ≥ 0.05).
Extended Data Fig. 5 KCa3.1 inhibition reduces tumor cell proliferation in vitro and in human cerebral brain organoids.
a, Fraction of EdU-positive cells after control versus 1 µM TRAM-34 and 1 µM Senicapoc treatment in vitro; n = 18 FOVs from 2 biologically independent experiments for each group; one-way ANOVA, Dunnett’s test. b, Fraction of dead cells, measured as number of PI-positive cells divided by DAPI-positive cells in control versus TRAM-34 and Senicapoc treatment in vitro; n = 18 FOVs from 2 biologically independent experiments for each group; one-way ANOVA, Dunnett’s test. c, Global Ca2+ activity and d, fraction of periodic cells of all active cells in control versus TRAM-34 (10 µM) treatment in a human cerebral organoid glioma model; patient-derived human glioblastoma cell lines S24 and P3; n = 7 recordings (S24, control) from 4 organoids, n = 4 recordings (S24, TRAM-34) from 2 organoids, and n = 3 recordings (P3, control and TRAM-34) from 2 organoids, respectively; one-way ANOVA, Dunnett’s test. e, Representative immunofluorescence images, f, tumor cell density and g, fraction of Ki67-positive tumor cells in tumor-bearing human cerebral organoids after 14 days of control versus 1 µM TRAM-34 or 1 µM Senicapoc treatment; n = 12 regions in 6 human cerebral organoids per group; S24 line; one-way ANOVA, Dunnett’s test (f) and Kruskal-Wallis test, Dunn’s test (g). h, and i, Fraction of activated caspase-3-positive tumor cells (h) and neurons (i) in tumor-bearing human cerebral organoids after 14 days of no treatment (control), DMSO control, TRAM-34 and Senicapoc treatment; n = 6 regions in 3 organoids per group; S24 line; one-way ANOVA, Dunnett’s test (h) and Kruskal-Wallis test, Dunn’s test (i). Error bars show s.e.m. ns, not significant (P ≥ 0.05).
Extended Data Fig. 6 Genetic perturbation of KCa3.1 reduces intrinsic rhythmicity, global Ca2+ activity, and tumor cell proliferation.
a, KCa3.1 knockout (KO; 2 different sgRNAs) versus knockout control (non-targeting sgRNA); S24 line. For both KOs no KCa3.1 signal was detected. Therefore, the qPCR cycle limit +1 was used as the hypothetical timepoint of detection for quantifying the relative KO of KCa3.1. b, KCa3.1 knockdown (KD) versus knockdown control (non-targeting shRNA); S24 line. c, KCa3.1 knockdown (KD) versus knockdown control (non-targeting shRNA); P3 line. a1–3,b1, and c1, Relative expression of KCa3.1 as determined via qPCR; n = 2 technical replicates (S24 KO control, S24 KCa3.1 KO-1 & −2), n = 5 technical replicates from 2 biologically independent experiments (S24 KD control, S24 KCa3.1 KD), n = 3 technical replicates (P3 KD control, P3 KCa3.1 KD). a4 b2, and c2, Fractions of periodic cells of all active cells in vitro. a5,b3, and c3, Frequencies of Ca2+ oscillations of periodic cells in vitro. Note that the difference between the groups in b3 reaches statistical significance but is minor, and thus of questionable biological relevance. a6,b4, and c4, Global Ca2+ activity in vitro. a7, Ca2+ traces of each cell from representative recordings; dark red: periodic cells. a8,b5, and c5, Fraction of EdU-positive cells in vitro. a9,b6, and c6, Fractions of dead cells in vitro. a4–6,b2–4, and c2–4, n = 5 recordings (S24 KO control, S24 KCa3.1 KO-1 & −2), n = 6 recordings (S24 KD control and S24 KCa3.1 KD), n = 5 recordings (P3 KD control and P3 KCa3.1 KD) from 2 biologically independent experiments per group, respectively. a8-9,b5–6, and c5–6, n = 27 FOVs (S24 KO control (EdU) and S24 KCa3.1 KO-1 & −2 (EdU)), n = 18 FOVs (S24 KO control (dead cells) and S24 KCa3.1 KO-1 & −2 (dead cells)), n = 45 FOVs (S24 KD control (EdU) and S24 KCa3.1 KD (EdU)), n = 18 FOVs (S24 KD control (dead cells) and S24 KCa3.1 KD (dead cells)), n = 27 FOVs (P3 KD control and P3 KCa3.1 KD) from 2 biologically independent experiments per group, respectively. a4–6,a8, One-way ANOVA, Dunnett’s test. a9, Kruskal-Wallis test, Dunn’s test. b1–6,c1–2, and c4, two-sided t-test. c3, and c5–6, two-sided Mann-Whitney test. d, Representative immunofluorescence images, e, tumor cell density f, fraction of Ki67-positive tumor cells, and g, fraction of activated caspase-3-positive tumor cells of tumor-bearing cerebral human organoids after implantation of S24 KO control and S24 KCa3.1 KO-2 line; n = 12 regions in 6 human cerebral organoids per group (e,f) and n = 6 regions in 3 human cerebral organoids per group (g); two-sided t-test (e,g) and two-sided Mann-Whitney test (f). Error bars show s.e.m. ns, not significant (P ≥ 0.05).
Extended Data Fig. 7 KCa3.1 drives periodic Ca2+ oscillations and their proliferation-stimulating effects in glioblastoma cell networks.
a—d, After gap junction inhibition via MFA, pharmacological inhibition, and genetic perturbation of KCa3.1 strongly reduce the fraction of cells displaying periodic Ca2+ activity but do not affect the remaining sporadic non-periodic Ca2+ activity, demonstrating a specific effect on periodic Ca2+ activity that – due to gap junction inhibition – is unlikely to be mediated by nonspecific effects on global Ca2+ communication. a,b, Fraction of cells displaying periodic (a) and non-periodic (b) Ca2+ activity of all cells after treatment with 100 µM MFA and with both MFA and 1 µM Senicapoc in vitro; S24 line; n = 10 recordings (Control), n = 11 recordings (MFA), and n = 13 recordings (MFA+Senicapoc) from 2 biologically independent experiments, respectively; One-way ANOVA, Dunnett’s test. c,d, Fraction of cells displaying periodic (c) and non-periodic (d) Ca2+ activity after treatment with MFA and after both genetic knockout of KCa3.1 and treatment with MFA in vitro; n = 5 recordings (S24 KO control line, no treatment; “Control”), n = 6 recordings (S24 KO control line, MFA treatment; “MFA”), and n = 6 recordings (S24 KCa3.1 KO-2 line, MFA treatment; “MFA+KCa3.1 KO”) from 2 biologically independent experiments, respectively; One-way ANOVA, Dunnett’s test. e—h, Intercellular coactivity is not reduced in Ca2+ recordings of control versus 1 µM TRAM-34 and 1 µM Senicapoc treatment (e), S24 KO control versus S24 KCa3.1 KO-1 and KO-2 (f), S24 KD control versus S24 KCa3.1 KD (g), and P3 control versus P3 KCa3.1 KD (h) in vitro; intercellular coactivity was calculated by dividing the number of coactive cells of a recording by the number of coactive cells of its corresponding null control data, representing the overall degree to which active cells are communicating their Ca2+ activity to other cells; increased intercellular coactivity after KCa3.1 inhibitor treatment might be because the strongly reduced number of periodic cells leads to less interference of different signals and therefore an even stronger synchronization between cell pairs. e, n = 11 recordings (S24, control treatment), n = 6 recordings (S24, TRAM-34), n = 15 recordings (S24, Senicapoc), n = 10 recordings (BG5, control treatment), n = 7 recordings (BG5, TRAM-34), and n = 11 recordings (BG5, Senicapoc) from 2 biologically independent experiments; Kruskal-Wallis test, Dunn’s test. f—h, n = 5 recordings (S24 KO control, S24 KCa3.1 KO-2), n = 4 recordings (S24 KCa3.1 KO-1), n = 6 recordings (S24 KD control), n = 5 recordings (S24 KCa3.1 KD), and n = 5 recordings (P3 KD control, P3 KCa3.1 KD) from 2 biologically independent experiments per group, respectively. f, One-way ANOVA, Dunnett’s test; g, two-sided t-test; h, two-sided Mann-Whitney test. i, Representative images of cells in adherent and spheroid conditions; in spheroid conditions tumor cells do not form networks (j) and do not display periodic activity (l), resulting in a much lower KCa3.1 expression (k), and therefore also do not show any Ca2+ activity (m). j, Number of TMs per cell in adherent versus spheroid conditions; n = 45 cells from 3 recordings from 3 biologically independent experiments (adherent) and n = 1087 cells from 5 recordings from 2 biologically independent experiments (spheroid); two-sided Mann-Whitney test. k, Relative expression of KCa3.1 in adherent versus spheroid conditions as determined via qPCR; n = 2 technical replicates; two-sided t-test. l, Fraction of periodic cells and m, global Ca2+ activity in adherent versus spheroid conditions; n = 11 recordings from 3 biologically independent experiments (adherent) and n = 5 recordings from 2 biologically independent experiments (spheroid); S24 line; two-sided t-test. n—q, AlamarBlue proliferation assay demonstrates that specific KCa3.1 inhibition with 1 µM TRAM-34 and 1 µM Senicapoc and genetic knockout of KCa3.1 reduces proliferation in adherent conditions (n,p) but not in spheroid conditions (o,q); S24 line; n = 12 measurements per group from 2 biologically independent experiments; one-way ANOVA, Dunnett’s test. r, Fraction of wild-type (WT) cells of all cells in Ca2+ recordings (Ca2+) displayed in Fig. 4c–f and Extended Data Fig. 7u,v and in recordings of the EdU proliferation assay (EdU) displayed in Fig. 4g–i and Extended Data Fig. 7w after coculturing S24 wild-type cells with S24 KCa3.1 KO-2 cells; n = 9 recordings (Ca2+) and n = 40 recordings (EdU) in 2 biological independent experiments, respectively. s, Ca2+ traces from representative recordings of S24 wild-type cells (WT), S24 KCa3.1 KO-2 cells and of a co-culture of 10% S24 wild-type cells with S24 KCa3.1 KO-2 cells (10% WT); traces of wild-type cells are depicted in red, traces of KCa3.1 knockout cells are depicted in green, and traces of cells displaying periodic Ca2+ activity are depicted thicker and darker; adding wild-type cells rescues the effect of the KCa3.1 knockout on global Ca2+ activity. t, Rarely detected periodic activity in knockout cells (green) after cocultivation with wild-type cells is due to close cupelling with periodically active wild-type cells (red); representative Ca2+ traces from s. u—w, Same data as shown in Fig. 4c–i, which originates from joint experiments with multiple experimental groups, and is shown here in one instead of two graphs to allow statistical comparison between the subpopulations (SP) of the co-culture and the control conditions (WT and KCa3.1 KO). Black p-values indicate these comparisons and grey p-values indicate comparisons that are also depicted in the respective main figure (Fig. 4c–i). P-values can differ here due to multiple testing: u, Fraction of periodic cells of all cells, v, global Ca2+ activity, and w, fraction of EdU-positive cells in recordings of S24 wild-type cells (WT), of S24 KCa3.1 KO-2 cells, of a co-culture of 10% S24 wild-type cells with S24 KCa3.1 KO-2 cells (10% WT) and of the subpopulations (SP) of S24 wild-type cells and S24 KCa3.1 KO-2 cells in the respective co-culture. u,v, n = 7 recordings (WT), n = 8 recordings (KCa3.1 KO), and n = 9 recordings (Co-culture) in 2 biological independent experiments; one-way ANOVA, Dunnett’s test (u) and Kruskal-Wallis test, Dunn’s test (v). w, n = 20 recordings in 2 biological independent experiments for all groups respectively; Kruskal-Wallis test, Dunn’s test. x, Fraction of wild-type (WT) cells and y, fraction of EdU-positive cells in recordings of S24 wild-type cells (100%), S24 KCa3.1 KO-2 cells (0%) and of a co-culture of 3, 5 and 10% S24 wild-type cells with S24 KCa3.1 KO-2 cells; y, for 0, 3, 5, an 10% WT cells the fraction of proliferating KO cells is shown on the left y-axis, for comparison the fraction of proliferating cells in WT-monoculture (100%) is shown on the right y-axis. While the co-culture of 5% WT cells still significantly increases the proliferation of the KCa3.1 KO, the co-culture of 3% of WT cells does not, placing the lower limit of WT cells to rescue the effects of the KCa3.1 KO somewhere between 3–5%. Kruskal-Wallis test, Dunn’s test. x,y, n = 20 recordings (0%, 10%, 100%), n = 6 recordings (3%), and n = 9 recordings (5%), in 2 biological independent experiments for all groups. Error bars show s.e.m. ns, not significant (P ≥ 0.05).
Extended Data Fig. 8 Tumor biological effects of KCa3.1 overexpression, and further in vitro characterizations.
a, KCa3.1 overexpression (OE) versus control; S24 line. b, KCa3.1 overexpression (OE) versus control; P3 line. a1 and b1, Relative expression of KCa3.1 as determined via qPCR; n = 3 technical replicates. a2, Normalized KCa3.1 immunofluorescence values of all cells (see Methods for more details). a3, Whole-cell patch-clamp electrophysiology of S24 cells overexpressing KCa3.1; arrows indicate glioblastoma cells connected via gap junctions filled via the patch pipette containing Alexa 594 dye; on the right side, I-V curves of voltage ramps between −105 mV to 55 mV are shown before (grey trace) and after wash-in of 1 µM TRAM-34 (red trace). TRAM-34 sensitive currents were detected in all S24 KCa3.1 overexpression cells (n = 4 cells), but in none of the control S24 cells (n = 3 cells). b2, Fraction of periodic cells of all active cells in vitro. a4, and b3, Frequency of Ca2+ oscillation of periodic cells in vitro. a5 and b4, Global Ca2+ activity in vitro. a6, Ca2+ traces of each cell from representative recordings; dark red: periodic cells. a7 and b5, Fraction of EdU-positive cells in vitro. a8, and b6, Fraction of dead cells in vitro. a2, n = 1810 cells (S24 OE control) and n = 1532 cells (S24 KCa3.1 OE) from 6 FOVs, respectively; a4–5, and b2–4, n = 10 recordings (S24 OE control), n = 11 recordings (S24 KCa3.1 OE), n = 6 recordings (P3 OE control), n = 7 recordings (P3 KCa3.1 OE) from 2 biologically independent experiments per group, respectively. a7–8, and b5–6, n = 27 FOVs from 2 biologically independent experiments per group, respectively. a1, a4–8, b1–2, and b4, two-sided t-test. a2,b3, and b5–6, two-sided Mann-Whitney test. c, Representative Nestin and KCa3.1 double-immunostaining after Ca2+ recordings of the same region in vitro; arrows indicate cells that displayed periodic Ca2+ oscillations in the preceding Ca2+ recording; S24 line. d, Normalized KCa3.1 immunofluorescence values after calcium recordings in periodic cells and non-periodic cells (see Methods for more details); periodic cells show a 51.1% ± 11.1% (95% CI) for S24 and 43.2% ± 17.0% (95% CI) for BG5 higher fluorescence intensity than non-periodic cells. c,d, n = 3177 S24 non-periodic cells, n = 75 S24 periodic cells, n = 3108 BG5 non-periodic cells, and n = 29 BG5 periodic cells from in n = 5 recordings per cell line from n = 3 biologically independent experiments; Kruskal-Wallis test, Dunn’s test. e, fraction of KCa3.1high tumor cells growing in mice; n = 27 regions in 4 mice (S24 line) and n = 21 regions in 3 mice (BG5 line). f, Fraction of tumor cells with the respective number of TMs per tumor cell in KCa3.1high and KCa3.1low cells; immunostainings of tumors growing in mice; error bars show s.d.; S24 line; n = 12 regions in 3 mice; Kruskal-Wallis test, Dunn’s test. g, Representative KCa3.1 and EdU immunostainings of S24 line; arrows indicate the KCa3.1high cells; dashed circles indicate the radius of 100 µm around the KCa3.1high cells; n = 18 recordings from 2 biologically independent experiments. h–k, Fraction of EdU-positive cells with a distance < 100 µm versus > 100 µm to the closest KCa3.1high cell normalized to the fraction of EdU-positive cells of each recording; h, S24 line control conditions, n = 106 ROIs (<100 µm) and n = 30 (>100 µm) in 2 biologically independent experiments, i, BG5 line control conditions, n = 36 ROIs (<100 µm) and n = 12 (>100 µm) in 2 biologically independent experiments, j, S24 line BAPTA-AM treatment, n = 42 ROIs (< 100 µm) and n = 11 (>100 µm) in 2 biologically independent experiments, and k, BG5 line BAPTA-AM treatment, n = 31 ROIs (<100 µm) and n = 12 (>100 µm) in 2 biologically independent experiments; two-sided Mann-Whitney test. Error bars show s.e.m. ns, not significant (P ≥ 0.05).
Extended Data Fig. 9 KCa3.1 expression is associated with distinct molecular features in glioblastoma.
a, Double-immunofluorescence of Nestin and KCa3.1 in a human glioblastoma sample of a representative tumor network region and invasion zone; arrow indicates KCa3.1high tumor cell; n = 30 ROIs (tumor network) and n = 30 ROIs (invasion zone) in 3 samples from different patients. b, Ingenuity Pathway Analysis (IPA) was performed using the differential gene expression of KCa3.1high versus KCa3.1low cells in a scRNA-seq dataset of 28 glioblastoma by Neftel et al.38. and the differential gene expression of the top versus the bottom quartile of KCa3.1 expressing TCGA-GBM glioblastoma (n = 36 patients (top) and 37 patients (bottom)); scatter plot and simple linear regression with f-test of all activation z-scores of the upstream regulator analysis by IPA reveals accordance between both datasets; n = 621 upstream regulators. c—e, Ingenuity Pathway Analysis (IPA) using the differential gene expression of KCa3.1high versus KCa3.1low cells in a scRNA-seq dataset of 28 glioblastoma by Neftel et al.38. c, Top differentially regulated molecular and cellular functions; the respective activation z-score is color-coded as indicated (higher values (orange) represent an upregulation in KCa3.1high cells and lower values (blue) a downregulation); right-tailed Fisher’s exact test. d,e, Box plot showing the activation z-scores of all differentially regulated molecular and cellular functions associated with cell death and survival (d) and growth of protrusions (e); n = 18 (death) and n = 3 (survival); two-sided t-test; error bars show s.e.m. f,h, Ingenuity Pathway Analysis (IPA) was performed using the differential gene expression of the top versus the bottom quartile of KCa3.1 expressing TCGA glioblastoma (n = 36 patients (top) and 37 patients (bottom)). f, Top differentially regulated molecular and cellular functions; the respective activation z-score is color-coded as indicated; right-tailed Fisher’s exact test. h, Activation z-scores of the top activated signaling pathways. g,i, Using the differential gene expression of KCa3.1high versus KCa3.1low cells in the scRNA-seq dataset of 28 glioblastoma by Neftel et al.38. and the differential gene expression of the top versus the bottom quartile of KCa3.1 expressing TCGA-GBM glioblastoma (n = 36 patients (top) and 37 patients (bottom)) a gene set expression analysis (GSEA) of the GO-Terms “cell proliferation”, “cell-cell signaling”, “pro-survival”, “MAPK cascade”, “NIK/NF-KAPPAB signaling”, “calcium-dependent cysteine-type endopeptidase activity”, and “calmodulin-dependent protein kinase activity” was performed. j, Representative phospho-CaMKII (Thr286) and k, MAP2 immunofluorescence images and immunofluorescence values; arrows indicate cells that displayed periodic Ca2+ oscillations in the preceding Ca2+ recording; Calpain translates high-frequency Ca2+ transients into decomposition of its sensitive substrate, MAP228. The MAP2 fluorescence intensity therefore correlates invers with the Calpain activity; S24 line; n = 3069 cells (non-periodic, CaMKII), n = 114 (periodic, CaMKII), n = 3279 cells (non-periodic, MAP2), and n = 102 (periodic, MAP2) from 8 recordings from 2 biologically independent experiments, respectively; two-sided Mann-Whitney test. l,m, Ingenuity Pathway Analysis (IPA) was performed using the differential gene expression after control versus KCa3.1 inhibition in vitro using Senicapoc; right-tailed Fisher’s exact test. l, Activation z-score of top downregulated signaling pathways. m, Top differentially regulated molecular and cellular functions; the respective activation z-score is color-coded as indicated. n, Using the differential gene expression after control versus KCa3.1 inhibition in vitro using Senicapoc a gene set expression analysis (GSEA) of the GO-Terms “cell proliferation”, “cell-cell signaling”, “pro-survival”, “MAPK cascade”, “NIK/NF-KAPPAB signaling”, “calcium-dependent cysteine-type endopeptidase activity”, and “calmodulin-dependent protein kinase activity” was performed. o, Representative phospho-NF-κB and p, phospho-p44/42 MAPK immunofluorescence images and normalized mean immunofluorescence values of all cells in the FOV in vitro, S24 KCa3.1 overexpression (KCa3.1 OE) versus S24 OE control (Control); n = 20 FOVs (KCa3.1 OE) and n = 20 FOVs (Control) from 2 biologically independent experiments, respectively; two-sided t-test. q,r, Coculturing 10% wild-type cells with KCa3.1 knockout cells rescues the suppressive effect of the KCa3.1 KO on NF-κB and p44/42 MAPK activity. Representative phospho-NF-κB (q) and phospho-p44/42 MAPK (r) immunofluorescence images and normalized mean immunofluorescence values of all cells in the FOV in vitro; data is shown for S24 wild-type cells (WT), for S24 KCa3.1 KO-2 cells, for a co-culture of 10% S24 wild-type cells with S24 KCa3.1 KO-2 cells (10% WT) and for the subpopulation (SP) of S24 wild-type cells and S24 KCa3.1 KO-2 cells in the respective co-culture; NF-κB: n = 18 recordings from 2 biologically independent experiments for all conditions, respectively, one-way ANOVA, Dunnett’s test; p44/42 MAPK: n = 39 FOVs (WT & 10% WT) and n = 20 FOVs (KCa3.1 KO) from 2 biologically independent experiments, respectively, Kruskal-Wallis test, Dunn’s test. s, Fraction of wild-type (WT) cells of all cells in recordings of the immunocytochemistry of the NF-κB- and MAPK-pathways displayed in q and r after coculturing S24 wild-type cells with S24 KCa3.1 KO-2 cells; n = 18 recordings (NFκB) and n = 39 recordings (MAPK) in 2 biological independent experiments, respectively. Error bars show s.e.m.
Extended Data Fig. 10 Relevance of KCa3.1 functionality for brain tumor progression.
a, Fold change of intercellular coactivity in Ca2+ recordings from control and either TRAM-34 (i.p., 120 mg/kg) or Senicapoc (i.p., 50 mg/kg) treated mice; intercellular coactivity was calculated by dividing the number of coactive cells of a recording by the number of coactive cells of its corresponding null control data; KCa3.1 inhibition did not hinder cell-to-cell signaling as such, ruling out that the reduced global Ca2+ activity (Fig. 6b) is solely due to nonspecifically altered Ca2+ dynamics; n = 8 and 11 recordings in 4 mice for control and TRAM-34 treatments, respectively in S24, n = 14 and 6 recordings in 4 mice for control and Senicapoc treatments, respectively in S24, n = 7 recordings in 4 and 3 mice for control and TRAM-34 treatments in BG5, respectively; Kruskal-Wallis test, Dunn’s test. b, Representative time series of tumor cells (green) and brain microvessels (red) under control conditions versus Senicapoc treatment; arrows indicate somata of tumor cells; scale bars are 25 µm. c, Invasion speed of tumor cells; BG5 line. b,c, n = 49 cells under control conditions in 4 mice and n = 71 cells under Senicapoc treatment in 4 mice; Mann Whitney test. d, Representative immunofluorescence staining of Nestin (tumor cells) and IBA1 (microglia) in patient-derived human glioblastoma xenografts (S24 line) in mouse brains after TRAM-34 treatment (2x 120 mg/kg per day) versus control; arrowheads show activated microglia and arrows show surveilling microglia according to morphological criteria (see Methods for more details). e, Number of microglia and f, fraction of activated microglia in patient-derived human glioblastoma xenografts in mouse brains after TRAM-34 treatment (2x 120 mg/kg per day) and Senicapoc treatment (2x 50 mg/kg per day) versus control. d—f, n = 31 FOVs in 5 mice for control and n = 38 FOVs in 3 mice for TRAM-34; n = 48 FOVs in 4 mice for control and n = 61 FOVs in 4 mice for Senicapoc; Kruskal-Wallis test, Dunn’s test. g, Representative immunofluorescence stainings for Nestin and Ki67 and h, fraction of Ki67-positive tumor cells in patient-derived human glioblastoma xenografts (S24 line) in mouse brains after TRAM-34 treatment (2x 120 mg/kg per day) versus control; n = 36 FOVs in 5 mice for control and n = 28 FOVs in 3 mice for TRAM-34; two-sided Mann-Whitney test. i, Representative immunofluorescence staining of Nestin and EdU and j, fraction of EdU-positive tumor cells in patient-derived human glioblastoma xenografts (BG5 line) in mouse brains after Senicapoc treatment (2x 50 mg/kg per day) versus vehicle control and in vivo EdU incorporation; n = 39 FOVs in 4 mice for control and n = 64 FOVs in 4 mice for Senicapoc; two-sided Mann-Whitney test. k, Change in tumor cell density on day 14 versus day 0 under control conditions versus TRAM-34 treatment (2x 120 mg/kg per day); n = 14 regions in 4 mice (control) and n = 21 regions in 4 mice (TRAM-34); Kruskal-Wallis test, Dunn’s test. l, Number of tumor cells and m, representative tile-scan images of tumors (green) and brain vessels (red) on day 14 versus day 0 under control versus Senicapoc treatment (2x 50 mg/kg per day); BG5 line; n = 5 mice for control conditions and n = 4 mice for Senicapoc treatment; two-sided t-test. n, KCa3.1 expression values (log2(FPKM + 1)) in glioblastoma (n = 230 patients), IDH-mutant and 1p/19q intact lower grade glioma (n = 241 patients), and IDH-mutant and 1p/19q codeleted oligodendroglioma (n = 176 patients) from TCGA datasets; Kruskal-Wallis test, Dunn’s test. o, Kaplan-Meier survival plot of glioblastoma patients of the mesenchymal gene expression subtype with the highest versus lowest third of KCa3.1 expression; the significant difference between groups indicates that the prognostic effect of KCa3.1 is not simply due to its correlation with mesenchymal tumors, which have been associated with a worse prognosis70; log-rank test. p, Normalized expression of KCa3.1 in different subtypes of glioblastoma; n = 49 patients (mesenchymal), n = 31 patients (proneural), n = 26 patients (neural), and n = 39 patients (classical). q, Schematic illustration of main findings. Error bars show s.e.m. ns, not significant (P ≥ 0.05).
Supplementary information
Supplementary Methods
This file contains the entire Methods section for the manuscript as well as the references for the extended data figures.
Supplementary Methods
Ca2+ communication in vivo and in vitro and mechanistic insights into critical compounds used in the manuscript.
Supplementary Video 1
Ca2+ imaging in vivo and in the newly developed in vitro model. Awake in vivo multiphoton Ca2+ imaging of gliomas growing in the brain under a chronic cranial window in mice. Glioblastoma cells were transduced with the GCaMP6s Ca2+ sensor using lentiviral vectors. In vitro Ca2+ imaging was performed in the monolayer assay where tumour cells were stained with the Rhod-2AM Ca2+ sensor and imaged using confocal microscopy. Recordings demonstrate the dynamics as well as the robust and frequent nature of Ca2+ transients in the in vivo and in vitro model, indicative of intercellular Ca2+ communication in these tumours.
Supplementary Video 2
Periodic glioblastoma cells in vivo. Awake in vivo multiphoton Ca2+ imaging of glioblastoma cells growing in the brain under a chronic cranial window in mice. Tumour cells were transduced with the Twitch3 Ca2+ sensor using lentiviral vectors. The representative recording demonstrates autonomous rhythmic Ca2+ activity of a periodic cell that triggers Ca2+ activity in regionally connected tumour cells.
Supplementary Video 3
Periodic glioblastoma cells in the in vitro monolayer assay. Ca2+ imaging in the in vitro monolayer assay of tumour cells stained with the Rhod-2AM Ca2+ sensor using confocal microscopy. The representative recordings demonstrate autonomous rhythmic Ca2+ activity of a periodic cell. Gap junction inhibition with MFA blocks the transmission of the autonomous activity to regionally connected tumour cells but does not affect the autonomous Ca2+ activity of periodic cells itself.
Supplementary Video 4
Ca2+ imaging mixing experiment. Ca2+ imaging in the in vitro monolayer assay of tumour cells stained with the Rhod-2AM Ca2+ sensor (displayed in red) using confocal microscopy. Recording of S24 wild-type cell monoculture shows vivid global Ca2+ activity. Recording of S24 KCa3.1 knockout cell monoculture shows strongly reduced global Ca2+ activity and no intrinsically rhythmic activity. Mixing 10% S24 wild-type cells (GFP negative) with 90% S24 KCa3.1 knockout cells (GFP positive, green) fully recovers the effect of the KCa3.1 knockout on global Ca2+ activity and the S24 wild-type cells show an unusually high fraction of periodic cells.
Supplementary Video 5
Glioma Ca2+ imaging before and after TRAM-34 treatment in vivo. Awake in vivo multiphoton Ca2+ imaging of GCaMP6s-labeled glioblastoma cells before and 1h after single TRAM-34 treatment. TRAM-34 treatment markedly reduced the tumour autonomous rhythmic Ca2+ activity and thereby the overall Ca2+ signaling in the tumour.
Source data
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Hausmann, D., Hoffmann, D.C., Venkataramani, V. et al. Autonomous rhythmic activity in glioma networks drives brain tumour growth. Nature 613, 179–186 (2023). https://doi.org/10.1038/s41586-022-05520-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/s41586-022-05520-4
This article is cited by
-
PerSurge (NOA-30) phase II trial of perampanel treatment around surgery in patients with progressive glioblastoma
BMC Cancer (2024)
-
KCNA1 promotes the growth and invasion of glioblastoma cells through ferroptosis inhibition via upregulating SLC7A11
Cancer Cell International (2024)
-
A phase Ib/II randomized, open-label drug repurposing trial of glutamate signaling inhibitors in combination with chemoradiotherapy in patients with newly diagnosed glioblastoma: the GLUGLIO trial protocol
BMC Cancer (2024)
-
How cancer hijacks the nervous system to grow and spread
Nature (2024)
-
Pan-cancer ion transport signature reveals functional regulators of glioblastoma aggression
The EMBO Journal (2024)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.
