Biophysical Properties of Human Medulloblastoma Cells
- First Online:
- Cite this article as:
- Ernest, N.J., Logsdon, N.J., McFerrin, M.B. et al. J Membrane Biol (2010) 237: 59. doi:10.1007/s00232-010-9306-x
- 130 Views
Medulloblastoma is a pediatric high-grade cerebellar malignancy derived from neuronal precursors. Although electrophysiologic characteristics of cerebellar granule neurons at all stages of cell development have been well described, such characterization has not been reported for medulloblastoma. In this study we attempt to characterize important electrophysiologic features of medulloblastoma that may distinguish it from the surrounding cerebellum. Using patient-derived cell lines and tumor tissues, we show that medulloblastoma cells have no inward Na+ current or transient K+ current involved in action potential generation and propagation, typically seen in granule neurons. Expression and function of calcium-activated, large-conductance K+ channels are diminished in medulloblastoma, judged by electrophysiology and Western analysis. The resting membrane potential of medulloblastoma cells in culture is quite depolarized compared to granule neurons. Interestingly, medulloblastoma cells express small, fast-inactivating calcium currents consistent with T-type calcium channels, but these channels are activated only from hyperpolarized potentials, which are unlikely to occur. Additionally, a background acid-sensitive K+ current is present with features characteristic of TASK1 or TASK3 channels, such as inhibition by ruthenium red. Western analysis confirms expression of TASK1 and TASK3. In describing the electrophysiologic characteristics of medulloblastoma, one can see features that resemble other high-grade malignancies as opposed to normal cerebellar granule neurons. This supports the notion that the malignant phenotype of medulloblastoma is characterized by unique changes in ion channel expression
KeywordsMedulloblastomaMembrane biophysicsCalcium-activated K+ channelModulation of neuronal ion channel
Medulloblastoma is a highly malignant posterior fossa tumor found primarily in children. It is the most common malignant brain tumor diagnosed in pediatrics. The putative cell of origin of medulloblastoma is the cerebellar granule neuron precursor (Hatton et al. 2008; Schuller et al. 2008). Detailed information is known regarding the development, migration and maturation of cerebellar granule neurons. However, despite their neuronal origin, little has been described regarding the neuronal properties of medulloblastoma cells; the similarities and differences between these tumor cells and their neighboring normal counterparts are still being investigated. More specifically, the ion channel properties of this aggressive and highly metastatic cancer have not been defined. Membrane ion channel activity plays an important role in managing cell volume, influencing migration potential, progression through the cell cycle and responding to environmental signals in order to maintain homeostasis (Bloch et al. 2007; Han et al. 2007, 2008; Kunzelmann 2005; Mu et al. 2003; Szabo et al. 2004; Wang 2004; Zhang et al. 2009). Often, a malignant cell will have a channel expression pattern distinct from the normal tissue from which it is thought to arise and from neighboring uninvolved tissue. Therefore, channel activity is extremely relevant to the malignant potential of tumor cells. We have characterized the electrophysiologic properties of malignant medulloblastoma cells in order to reveal important information regarding the nature of medulloblastoma activity in the brain. Specifically, we find that medulloblastoma cells do not have channels important in signal generation (Na+) and propagation (transient K+) but retain K+ channels implicit in basic cell biology. Also, small, fast-inactivating calcium currents are evident in some cells. We discuss the relevance of these current alterations to the malignant character of medulloblastoma cells.
Materials and Methods
Cerebellar granule neurons were cultured as described (Han et al. 2007) with a few modifications. Briefly, P2-6 Sprague-Dawley rats were decapitated and the cerebella were dissected in ice-cold Hank’s balanced salt solution with glucose. Following removal of the meninges, the tissue was placed in an O2-saturated papain and DNAse solution (Worthington Biochemical, Lakewood, NJ) for 20 min at 37°C. The tissue was triturated, centrifuged and resuspended in Earle’s balanced salt solution. The cell suspension was layered over a sterile fetal bovine serum cushion and centrifuged for 10 min at 100×g with no brake. The recovered cell pellet was triturated with a polished glass Pasteur pipette. Cells were maintained in NeuroBasal medium supplemented with B-27 (GIBCO, Grand Island, NY), 1 μM ara-C, 20 mM KCl and 1× penicillin–streptomycin. Cells were counted and seeded onto glass coverslips or culture dishes coated with 25 μg/ml poly-l-lysine at a density of 105 cells/cm2. The next day and every third day afterward, medium was changed. Fungizone (Invitrogen, Carlsbad, CA) was included in the medium for the first 4–6 days of culture.
Medulloblastoma experiments were performed on DAOY and D283 (American Type Culture Collection, Manassas, VA), two commercially available medulloblastoma cell lines, and a cell line, D425, obtained from Duke University, which was established from a primary medulloblastoma (kindly provided by Dr. D. Bigner, Duke University, Durham, NC) (Di et al. 2005; Adamson et al. 2010). D283 and DAOY medulloblastoma cell cultures were maintained in MEMα supplemented with 2 mM l-glutamine (Mediatech, Manassas, VA) and 10% characterized fetal bovine serum (HyClone, Logan, UT). D425 cells were maintained in Improved MEM Zinc Option + 10% FBS + 5 ml 1 M HEPES + 14.7 ml 7.5% sodium bicarbonate (GIBCO). Cells were kept at 37°C in a 95% O2–5% CO2 humidified atmosphere.
Whole-cell voltage-clamp recordings on cultured cells were obtained as previously described (Hamill et al. 1981). Patch pipettes were made from thin-walled (outer diameter 1.5 mm, inner diameter 1.12 mm) borosilicate glass (TW150F-4; WPI, Sarasota, FL) and had resistances of 3–6 MΩ. Current recordings were obtained with an Axopatch 200A amplifier (Axon Instruments, Foster City, CA). Current signals were low-pass-filtered at 2 kHz and digitized online at 10–20 kHz using a Digidata 1322A digitizing board (Axon Instruments). Data acquisition and storage were conducted with the use of pClamp 8.2 (Axon Instruments). Resting membrane potentials, cell capacitances and series resistances were measured directly from the amplifier, with the upper limit for series resistance being 12 MΩ and series resistance compensation adjusted to 80% to reduce voltage errors.
Standard KCl pipette solution contained (in mM) 145 KCl, 1 MgCl2, 10 EGTA and 10 HEPES Na+ salt, with the pH adjusted to 7.3 with Tris base. CaCl2 (0.2 mM) was added to the pipette solution just before recording, resulting in a free calcium concentration of 1.9 nM using an available equation based on the calculations previously described (Portzehl et al. 1964). Calcium concentration was very low in order to prevent interference of calcium activation of channels that are also voltage-activated. Cells were continuously superfused at room temperature with a saline solution containing (in mM) 125 NaCl, 5.0 KCl, 1.0 CaCl2, 10.5 glucose and 32.5 HEPES acid. The pH of this solution was adjusted to 6.4 or 7.4 with NaOH, and the osmolarity of this solution was confirmed at 290–310 mOsm using a micro-osmometer (Fiske Associates, Norwood, MA).
A cesium chloride pipette solution was used to facilitate isolation of Ca2+ current, containing (in mM) 140 CsCl, 1 MgCl2, 0.2 CaCl2, 10 EGTA and 10 HEPES at pH 7.2, yielding a calculated free calcium concentration of approximately 1.9 nM. The external solution contained (in mM) 125 NaCl, 5 KCl, 2.0 CaCl2, 10.5 glucose and 32.5 HEPES at pH 7.4. Note, the external Ca2+ concentration was increased from 1 to 2 mM to increase the concentration of charge carrier and enhance the driving force.
Channel inhibitors were added directly to bath solutions or media from stock solutions. Stock solutions were created by dissolving iberiotoxin (Ibtx; AnaSpec, San Jose, CA), ruthenium red and mibefradil at 1,000× final concentration in ddH2O. Nimodipine stock solution was made by dissolving drug at 5,000× final concentration in DMSO. All drugs were obtained from Sigma-Aldrich (St. Louis, MO) unless otherwise noted.
Current Density Calculations
To calculate BK current densities, whole-cell currents elicited over a range of potentials from −160 to +160 mV in the presence of Ibtx were subtracted from control currents and the current–voltage relationship was plotted. Currents at +140 mV were normalized using whole-cell capacitance to estimate cell size and expressed as picoamperes per picofarad (pA/pF). Current densities of pH-sensitive currents were calculated by plotting current–voltage relationships elicited from cells stepped to a range of depolarizing potentials (−70 to +90) from a prepulse potential of −110 mV. The protocol was done in the presence of Ibtx to exclude BK currents, and recordings evoked at pH 6.4 were subtracted from those obtained at pH 7.4 to give the pH-sensitive portion of the current. Currents obtained at +60 mV were normalized to whole-cell capacitance as for the BK analysis.
Cell pellets collected from 10-cm culture dishes were rinsed in PBS and lysed in RIPA (50 mM Tris–HCl [pH 8.0], 150 mM NaCl, 1% IGEPAL detergent, 0.1% SDS, 0.5% Na+ deoxycholate) containing 1× final concentration protease and phosphatase inhibitors (Sigma-Aldrich). Tissues (10–30 mg) were homogenized on ice in lysis buffer (100 mM Tris–HCl [pH 7.4], 1% SDS) containing protease and phosphatase inhibitors. Samples were sonicated briefly, incubated at 4°C with occasional agitation for 1 h, then clarified by centrifugation at 14,000×g for 30 min at 4°C. Total protein concentrations of the supernatants were measured using the modified detergent-compatible Lowry assay (Bio-Rad Laboratories, Hercules, CA). Twenty micrograms of total protein per lane were separated on 10% SDS-PAGE gels and transferred to PVDF membranes (Millipore, Billerica, MA). Membranes were blocked in TBS-T (20 mM Tris–HCl [pH 7.6], 137 mM NaCl, 0.1% Tween-20) containing 5% nonfat dry milk for 1 h at room temperature. Primary antibodies (TASK1 and TASK3, Sigma-Aldrich; BK, UC Davis NeuroMab Facility, Davis, CA) were applied overnight at 4°C in TBS-T + 5% milk. After extensive washing, membranes were incubated with the appropriate secondary antibody (goat anti-mouse or goat anti-rabbit-HRP conjugate; Santa Cruz Biotechnology, Santa Cruz, CA) at room temperature for 1 h. Detection was accomplished with ECL plus detection reagent (GE Biosciences, Piscataway, NJ), and images were collected on a Kodak (Rochester, NY) Image Station 4000MM. Membranes were stripped for 1 h at 50°C in strip buffer (62.5 mM Tris–HCl [pH 6.7], 2.0 % SDS, 100 mM β-mercaptoethanol), rinsed extensively in TBS-T and reprobed with actin antibody (Sigma-Aldrich) as a loading control.
Cultured Cerebellar Granule Neurons Demonstrate Classic Na+ and K+ Currents
This same voltage-step protocol also resulted in the activation of a large whole-cell outward current that appeared to decay in two phases, while current evoked following a depolarizing prepulse of −50 mV consisted mainly of a delayed outward current. Subtracting currents evoked from these two protocols allowed us to isolate a transient outward current with a reversal potential of approximately −80.2 mV. Similarly, a plot of the current–voltage relationship for the delayed outward current revealed a reversal potential of approximately −35.2 mV (Fig. 1b, c). These currents very closely resembled the transient K+ (IA) and late K+ (IK) currents that have been previously described in various preparations of excitable cells, including CGNs.
Cultured Medulloblastoma Cells Lack Na+ and Transient K+ Currents
Whole-cell voltage-clamp recordings were similarly obtained from cultured medulloblastoma cells from three separate cell lines. Notably, the entrance potential for each of the three cell lines (D283, −27.5 mV; D425, −13.4 mV; DAOY, −18.1 mV) was significantly depolarized compared to cultured CGNs (−52.7 mV). For this reason, the holding potential used for the following experiments was adjusted to −40 mV.
While this protocol did result in the activation of an outward current, there was no evidence of a rapidly inactivating, or transient, outward current in any of the medulloblastoma cell lines. In order to determine if the hyperpolarizing prepulse had any effect on the outward current, we also recorded whole-cell currents in the absence of a prepulse. Subtracting the whole-cell currents evoked from these two protocols also failed to reveal a transient K+ current, as this procedure did in CGNs. Rather, the subtracted current was slowly activating, outwardly rectifying with properties characteristic of delayed-rectifier currents (Fig. 2c, d). The late outward current evoked in the absence of a hyperpolarizing prepulse was similar to that observed in cultured neurons.
Expression of Large-Conductance Ca2+-Activated K+ (BK) Channels Is Decreased in Medulloblastoma Cells
In these studies, the glioblastoma cell line D54-MG was used as a positive control since gliomas are known to have abundant expression of BK (Ransom and Sontheimer 2001). To determine if loss of BK expression in medulloblastoma cells was a phenomenon of cell culture, we examined the protein expression of BK in whole-tissue extracts from normal human cerebellum and primary medulloblastoma tissue. Similar to what was observed in cell culture, each of the four primary medulloblastoma tumors expressed reduced levels of BK compared to normal human brain (Fig. 3d).
Since BK channels are activated by Ca2+ as well as voltage, the presence of voltage-activated calcium currents was investigated. Recordings using cesium chloride to isolate Ca2+ currents demonstrated small, mibefradil-sensitive, fast-inactivating transient currents at each voltage step ≥−40 mV, with a peak current observed typically from −30 to −10 mV (Fig. 3e). These current transients were not detected when cells were recorded from a holding potential of −50 mV, consistent with T-type calcium channel inactivation. These T-type channels were identified in about 68% of D283 cells (11 of 16). No L-type Ca2+ channels were detected in similar experiments using nimodipine.
Functional Expression of pH-Sensitive K+ Channels is Enhanced in Cultured Medulloblastoma Cells
Studying the biophysical properties of primary medulloblastoma is challenging because the tumor is diagnosed so infrequently (approximately 500 cases per year in the United States) and the tumor identity may not be known until after the surgery is performed, making fresh tissue acquisition difficult. However, a handful of cell lines are available, each derived from primary disease. One of the few electrophysiologic reports on medulloblastoma used one such cell line, MHH-MED-3, to demonstrate the similarities between immature CGNs and medulloblastoma in order to argue for granule cell precursors as the medulloblastoma cell of origin (Codina et al. 2000). We investigated human cell lines of medulloblastoma compared to cultured rat CGNs to demonstrate differences between the malignant cells and native neurons with which tumor cells are surrounded. We discovered a number of differences in whole-cell recordings from three medulloblastoma cell lines and CGNs.
Medulloblastoma cultured cells are much more depolarized at rest than CGNs, with entrance potentials of −27.5 to −13.4 mV compared to −52.7 mV for CGNs. This is similar to findings in tumors derived from nonexcitable cells such as cultured glioma cells (Kunzelmann 2005; Pardo et al. 2005; Sontheimer 2008). A more depolarized resting potential appears necessary for cell proliferation, since oscillating ion fluxes are necessary for proceeding through the cell cycle, and is a common feature of dividing cells. Channels do exist, however, in medulloblastoma cell lines that are largely inactive at these resting potentials, such as T-type Ca2+ channels. T-type Ca2+ channels reportedly play a role in tumor cell proliferation (Panner and Wurster 2006). Similar to our findings with BK channels, not all D283 cells have these Ca2+ channels; we have not established whether the cells that have BK currents are the same population that has T-type Ca2+ channels. Interestingly, splice variants exist for both types of channels in gliomas, suggesting that each makes a unique contribution to malignant pathology (Latour et al. 2004; Liu et al. 2002). The significance of channels that are inactive in these cells at rest and how this plays out within the environment of the tumor are not known.
Medulloblastoma is considered an embryonal tumor, meaning that it derives from fetal tissue. Therefore, it is important not to assume that mature granule neurons would be the proper cells to compare to this malignant cell type in our studies. However, cerebellar granule cells develop the three predominant current components very early in development, though their proportions may change as the organism matures (Mathie et al. 2003). These currents include an inward Na+ current and two K+ currents, a transient one and a noninactivating one (Radden et al. 1994). Our CGNs did indeed demonstrate the three expected currents. Medulloblastoma cells display only one of these three typical currents. Specifically, as seen when recordings of D283 cells were made in the presence of the voltage-gated Na+ channel inhibitor TTX, the transient inward Na+ current responsible for CGN action potentials is absent. This corroborates the results reported previously in the MHH-MED-3 cell line (Codina et al. 2000). One study of immature granule cells in rats describes small inward and outward currents that increase in amplitude after weeks in culture (Hockberger et al. 1987). These are eventually robust enough to generate action potentials. Our medulloblastoma plots most closely resemble immature granule neurons early in vitro, with no significant inward currents. This is not surprising since medulloblastomas are thought to originate from immature granule neurons. However, it exposes one of the unique features of malignant cells: They do not take on a mature phenotype with time in culture. Understanding the character of malignant cells in the context of normal tissue in which it resides does not inform how a tumor became established but can be important as therapies look at differences that can be exploited in malignant cells while preserving normal brain.
Several K+ channels have been identified in healthy CGNs. Rapidly activating and inactivating channels that result in transient K+ currents that follow depolarization are abundant in the brain (Mathie et al. 2003). There is no inactivating current displayed in any of the medulloblastoma cell lines. There appears to be a loss of Na+ and transient K+ channels important for generating and propagating action potentials. However, they have retained K+ channels necessary for cell homeostasis and perhaps gained K+ channel function that is supportive of cancer.
It has been suggested that K+ channels may play a role in tumorigenesis, affecting cell proliferation (Kunzelmann 2005), apoptosis (Patel and Lazdunski 2004; Szabo et al. 2004) and even metastases (Sontheimer 2008). Reports of K+ channel regulation influencing malignant growth are found in studies of lung cancer, breast cancer, prostate cancer, pancreatic cancer, glioblastoma multiforme and acute myelogenous leukemia, to name a few (Han et al. 2002; Pardo et al. 2005). However, not all K+ channels promote tumors in all cells. For example, BK expression is increased in glioblastoma, correlating with increasing tumor grade (Sontheimer and Waxman 1993). However, it has no effect on adhesion properties of hepatocellular carcinoma cells (Zhou et al. 2003) or proliferation in colon cancer cells (Han et al. 2002). Furthermore, there is evidence to suggest that BK channels have anticancer effects in osteosarcoma (Cambien et al. 2008) and ovarian cancer (Han et al. 2008). In our studies, BK was expressed more in normal neurons and brain than in medulloblastoma by current density evaluation and protein expression. The significance of loss of functional BK channels in medulloblastoma is not yet clear but may suggest that BK is protective to normal neurons such that decreased expression may be significant to malignant transformation. More investigation of this property of BK loss is warranted.
We have identified an acid-sensitive current in our cells and hypothesized that TASK channels may be responsible for that current. Consistent with this assumption, currents were outwardly rectifying and activated by a step to pH 6.4, with both TASK1 and TASK3 present by Western blot. TASK1 and TASK3 RNAs are abundant in normal cerebellum, while TASK2 and TASK4 are not (Mathie et al. 2003). TASK3 has recently been implicated in growth and survival of other cells, including a variety of tumor types (Mu et al. 2003; Patel and Lazdunski 2004). We demonstrate higher protein expression by Western of TASK1 and TASK3 in some medulloblastoma primary tumor tissue and cultured cells than in their respective nonmalignant controls. In primary tumor tissue, each of the four samples showed markedly enhanced expression of either TASK1 or TASK3 but not both.
The biology of solid tumors typically involves outward proliferation and invasion along the tumor borders, while the tumor’s core fails to build a sound vascular structure and becomes necrotic. One could consider that acid-sensitive channels would be important for malignant cells to recognize this inhospitable environment of low oxygen tension and acidification in order to survive. However, the current density information presented reveals more functional pH-sensitive channels in CGNs than medulloblastoma cell lines. Perhaps the tumor cells have TASK channels available that can be called upon, in vivo, as needed for just that situation but outside of the three-dimensional context of a tumor fail to utilize them. Future investigation of the electrophysiology of TASK activity in fresh primary tumor would help to clarify this matter.
We anticipated that comparison of medulloblastoma cells to normal neurons would reveal minor but important differences in ion channel activity. Instead, we find that medulloblastoma cells display characteristics found in other high-grade tumors of nonexcitable tissue types. This tumor can be added to the growing list of tumors with aberrant channel expression, which is likely necessary for establishing and maintaining malignancy.
This work was supported in parts by NIH-RO1NS31234.