Transmembrane voltage potential of somatic cells controls oncogene-mediated tumorigenesis at long-range
In Xenopus, CLIC1 is expressed throughout development, with progressive increase in expression from mid-blastula transition to the tadpole stages; spatially, CLIC1 transcripts are present primarily in the ectoderm and ectodermally-originated organs [94]. Interestingly, CLIC1 expression is missing from the gut region where we frequently observe ITLS on the ectoderm. Thus, CLIC1 proteins are located at a distance from oncogene-induced foci, showing that they can act like the overexpressed potassium channels – in long-range hyperpolarization. By overexpressing wildtype Xenopus CLIC1 and raising extracellular chloride levels, we are able to achieve higher suppression of oncogenic transformation of cells located at a distance. Moreover, this suppression can be blocked – despite the presence of high extracellular chloride levels – by introducing a dominant negative chloride channel mutant away from oncogene-expressing cells (Fig. 4). Interestingly, another study from our lab has documented the endogenous presence of glycine-gated channels: though not present in melanocytes their depolarization confers highly proliferative and metastatic phenotype in melanocytes in a long-range, serotonergic signaling pathway [60, 61].
How do resting potential changes such as hyperpolarization impact on the transcriptional and epigenetic pathways of cancer? Recent studies have identified several mechanisms that transduce bioelectric signals at the cell membrane into biochemical responses (reviewed in [63, 104, 171]. Using a pharmacological suppression and molecular loss of function (Fig. 5A, B), we identified possible roles for gap junction communication (to be reported in a forthcoming study) and HDAC1 inhibition as signaling elements required for the long-range ITLS suppression. The introduction of a dominant negative HDAC1 mRNA [102] in a long-range manner was able to significantly increase neoplastic conversions of oncogene-expressing distant cells. The results are surprising relative to the well-known function of HDAC inhibitors as having antitumor activities, including reduced proliferation [172], increased differentiation/apoptosis [173], and that a number of them are on the clinical development pipeline for anti-cancer therapeutics [174-177]. In fact, variants of well-known HDAC inhibitors, such as vorinostat, (Merck) and romidepsin (Celgene) have already been developed and approved by the Food and Drug Administration for treating cutaneous T-cell lymphoma [178, 179]. However, consistent with our data, HDAC1 knock-down promotes early tumorigenesis in oncogene expressing cells [180], and genomic instability in a given cell can have a system-wide effect mediated by the immune system and other factors [181, 182]. Overall, this is consistent with our proposal that what is disrupted in cancer is a specific and highly modulated pattern of resting potentials - it is likely that the desired signaling cannot be achieved by universally increasing “anti-cancer” gene products.
Depending on other factors that vary widely in different systems (such as ion concentrations, butyrate availability, and spatial arrangement of tumor and tissue), disbalance of voltage-dependent signaling in either direction could promote or suppress tumorigenesis.
We sought to understand how Vmem changes and acetylation state of chromatins could control tumor growth in a non-cell autonomous manner. Given the link between Vmem-guided butyrate transport and HDAC activities the control cell behavior during regeneration [62] and local tumorigenesis [43], we tested the effect of targeting butyrate-producing bacteria on ITLS formation. Both novobiocin, an antibiotic against gram-positive bacteria, and a cocktail of antibiotics (gentamicin, metrodinazole, vancomycin, and clindamycin, data not shown) that have been verified to reduce levels of butyrate, significantly increased the rate of ITLS formation (Fig. 6). Our data are consistent with, and provide a novel mechanism explaining, several studies that investigated the effects of antibiotics on the gut microbiota, including butyrate-producing bacteria species, and their subsequent implications in increased cancer risks and incidences [183-186].
Our analysis was limited in two main ways: the lack of a technology for monitoring location of butyrate in vivo, and the inability to isolate and culture butyrate producing bacteria (which are highly anaerobic and do not grow under standard culture conditions [187]). Future work using GFP-expressing strains of appropriately engineered bacterial model species will further refine our understanding of the role of native microbiota on cancer progression and bioelectric signaling.
Collectively, the data presented here suggest a possible model (Fig. 7). Pharmacologically/molecular-genetically hyperpolarized cells have membrane that is negative on the inside, which creates a positively-charged cell surface environment. While bacteria – whether gram-positive or gram-negative – readily attach to a positively-charged surface, their viability is greatly diminished due to strong electrostatic interaction [188, 189]. As a result, distribution of butyrate producing-bacteria is skewed away from hyperpolarized cells in favor of oncogene-expressing cells, which are known to be depolarized pre-neoplastic transformation [43]. Availability of butyrate to these prospective tumor sites leads to HDAC inhibition and increased hyperacetylation of histones, which promote cell cycle arrest, leading to reduced rates of proliferation, apoptosis, and differentiation of cancer cells lines [105-107].
While the spatial dynamics of butyrate and bacteria remain to be investigated, our data reveal a novel long-range pathway regulating oncogene-mediated tumorigenesis at a distance. We do not yet know the maximum extent, in a large organism, of such long-range signaling by hyperpolarization. Future work must also examine the roles of remote endogenous bioelectric states in mammalian cancer models; if conserved, these data suggest a number of detection and treatment modalities focused on Vmem modulation. Because such strategies could be designed using ion channel drugs [190, 191] already approved for human use (as anti-epileptic agents), exciting opportunities for biomedicine may be presented by investigations of bioelectric signaling in cancer at a level beyond that of single cells.
Xenopus laevis embryos were collected and fertilized in vitro according to standard protocols [192], in 0.1X Modified Marc's Ringers (MMR; pH 7.8) with 0.1% Gentamicin. Xenopus embryos were housed at 14-18°C and staged according to Nieuwkoop and Faber [193]. All experimental procedures involving the use of animals for experimental purposes were approved by the Institutional Animal Care and Use Committees (IACUC) and Tufts University Department of Lab Animal Medicine (DLAM) under the protocol number M2011-70.
Fertilized Xenopus embryos were transferred into mesh-bottomed dishes with 3% Ficoll and injected with capped, synthetic mRNAs (made using the Ambion Message Machine kit) dissolved in water at the stages indicated. 2 hours post injection, embryos were transferred into 0.75× MMR for 45 minutes before they were washed and cultured in 0.1X MMR until desired stage was reached. Constructs used included: FUCCI (fluorescent ubiquitination-based cell cycle indicator) pair: mKO2-Cdt1 and mAG-Geminin [71, 72]; Gli1 [68], Xrel3 [194], KRASG12D [69], and Kv1.5 [86]; CLIC1-2A-GFP3 and CLIC1C24S-tdTomato; DN-HDAC1 [102].
Embryos were exposed in 0.1× MMR for the stages indicated to: Anthracene-9-Carboxylic Acid (ACA) 67 μM; indanyloxyacetic acid (IAA-94), 55 μM; 5-Nitro-2-(3-phenylpropylamino) benzoic acid (NPPB, Tocaris Bioscience, Bristol, UK), 8.5 nM; nalidixic acid 43 μM; novobiocin sodium salt 78 μM; gentamicin, 20 μM; metrodinazole, 20 μM; vancomycin hydrochloride (Santacruz Biotechnology, Texas, USA) 20 μM; Choline chloride, 70mM and clindamycin hydrochloride, 20 μM. All compounds were obtained from Sigma-Aldrich, St. Louis, MO unless otherwise noted.
Spatial analysis of leukocyte presence was performed by immunohistochemistry in paraffin sections, using an anti-XL2 antibody [195]. Briefly, embryos were fixed overnight in MEMFA [192], embedded in paraffin and sectioned at 5 μm using a Leica microtome. After deparaffinizing and rehydrating, tissue sections were permeabilized in phosphate buffered saline (PBS) + 0.1X Triton X-100 for 30 minutes, blocked with 10% goat serum in PBS + 0.1% tween-20 for 1 hour, and incubated at 4°C overnight with anti-XL2 primary antibody. Sections were then washed six times with PBST (30 minutes each at room temperature) and incubated with Alexa-Fluor-555-conjugated secondary antibody at 1:1000 in PBST + 10% goat serum overnight at 4°C. After five 30-minute washes in PBST, sections were mounted on a slide and photographed using the TRITC filter set on an Olympus BX61 spinning-disk confocal microscope with Hamamatsu ORCA digital CCD camera.
Immunochemical detection of tissue hypoxia was performed by immunoperoxidase reaction in paraffin sections, using the Hypoxy™-1 Plus Kit (Hypoxyprobe, Inc, MA, USA). St.34 embryos with KRASG12D ITLS were incubated in 300 µM of pimonidazol HCl (Hypoxyprobe™-1). Paraffin sections were prepared for immunostaining as described above. Tissue sections were then incubated using 1:50 dilution of FITC-MAB1, which binds to protein adducts of pimonidazole in hypoxic regions, and counterstained with 1:50 dilution of horseradish peroxidase conjugated anti-FITC secondary.
pH measurements were calculated from BCECF fluorescent signals as described previously in [196]. Briefly, embryos were incubated in a 5 µM BCECF, AM (2',7'-Bis-(2-Carboxyethyl)-5-(and-6)-Carboxyfluorescein, Acetoxymethyl Ester (Life Technologies, NY, USA) solution. Excess dye was washed out, and embryos were anesthetized with MS222. To image BCECF, a dual excitation dye, filters were EX 450/20, D 460, EM 535/30 (the isobestic point) and EX 500/20, D 515, EM 535/30. To calibrate BCECF in our system, a prototonophore, CCCP, and NaOH were used to drive pH values to 4 (limiting minimum) and 9 (limiting maximum), respectively, by observing gradual decrease and increase of fluorescence. A plot profile of pixel intensities was then taken from ITLS and control regions, and conversions to pH values were made as described in [196]
Sections taken through ITLS and control regions were stained with Hoechst Blue (Life Technologies, NY, USA), imaged using the DAPI filter set on an Olympus BX61 spinning-disk confocal microscope with Hamamatsu ORCA digital CCD camera, and analyzed by using ImageJ. The threshold and analyze particle tools in ImageJ were used to interactively outline each nucleus and determine its size and shape.
All statistical analyses were performed using GraphPad InStat v. 3.10 (GraphPad Software, La Jolla, CA, USA). Data were expressed as the mean unless otherwise noted. Error bars represent standard error. The differences between treatment groups were analyzed using Student's t-test, Chi-squared test or One-way ANOVA, (tukey's post hoc comparisons), and the null hypothesis was rejected at the 0.05 level.
This paper is dedicated to Bjorn Nordenstrom, a pioneer of long-range bioelectrical signaling in cancer. We thank Leonard Zon for the KRAS mutant construct, Florian Lang for the Kv1.5 construct, Ariel Ruiz Altaba for the Gli1 construct, John Wallingford for the Xrel3 construct, Atsushi Miyawaki for the FUCCI plasmid pair (mKO2-zCdt1 and mAG-zGeminin), and Saburo Nagata for the XL-2 antibody. We thank Jeff Wolf for H&E analysis of ITLS sections. We thank Amber Currier, Erin Switzer, and Amanda Allen for general lab assistance and frog husbandry, and Joan Lemire and Jean-Francois Pare for molecular biology assistance. We are grateful to Douglas Brash, Robert Gillies, and the members of the Levin lab for many useful discussions on these topics. M.L. is grateful for support of the G. Harold and Leila Y. Mathers Charitable Foundation, NIH R01-AR055993, and of DARPA (subaward W911NF-09-1-0125).