Inorganic polyphosphate regulates neuronal excitability through modulation of voltage-gated channels

Background Inorganic polyphosphate (polyP) is a highly charged polyanion capable of interacting with a number of molecular targets. This signaling molecule is released into the extracellular matrix by central astrocytes and by peripheral platelets during inflammation. While the release of polyP is associated with both induction of blood coagulation and astrocyte extracellular signaling, the role of secreted polyP in regulation of neuronal activity remains undefined. Here we test the hypothesis that polyP is an important participant in neuronal signaling. Specifically, we investigate the ability of neurons to release polyP and to induce neuronal firing, and clarify the underlying molecular mechanisms of this process by studying the action of polyP on voltage gated channels. Results Using patch clamp techniques, and primary hippocampal and dorsal root ganglion cell cultures, we demonstrate that polyP directly influences neuronal activity, inducing action potential generation in both PNS and CNS neurons. Mechanistically, this is accomplished by shifting the voltage sensitivity of NaV channel activation toward the neuronal resting membrane potential, the block KV channels, and the activation of CaV channels. Next, using calcium imaging we found that polyP stimulates an increase in neuronal network activity and induces calcium influx in glial cells. Using in situ DAPI localization and live imaging, we demonstrate that polyP is naturally present in synaptic regions and is released from the neurons upon depolarization. Finally, using a biochemical assay we demonstrate that polyP is present in synaptosomes and can be released upon their membrane depolarization by the addition of potassium chloride. Conclusions We conclude that polyP release leads to increased excitability of the neuronal membrane through the modulation of voltage gated ion channels. Together, our data establishes that polyP could function as excitatory neuromodulator in both the PNS and CNS.


Background
Inorganic polyphosphate (polyP) is a bioactive polymer of 10 to several 100 orthophosphates linked together by phosphoanhydride bonds. Recent studies have implicated polyP as a regulatory molecule among non-excitable cells [1][2][3][4]. Notably, polyP has been shown to play a signaling role in astrocytes where its release facilitates intercellular communication [5] and in stimulated platelets, which release significant amounts of polyP triggering inflammation and plasma clotting [6,7]. Compounds released from stimulated platelets during injury and inflammation are involved in pain signaling mechanisms, and these compounds can directly induce firing of sensory neurons [8].
Here we investigated whether polyP influences neuronal activity. We show that polyP dramatically affects the activity of excitable cells through the modulation of distinct ion channels. Specifically, application of polyP resulted in the rapid generation of high frequency action potentials in both DRG and hippocampal neurons. The neuronal response is due to a hyperpolarizing shift in Na V channel activation, potentiation of Ca V channels, and block of K V channels. As polyP is a neuroactive compound, we propose that its systemic release by platelets contributes to inflammatory pain. Furthermore, we show here that polyP is present in synaptic vesicles using both biochemical and live imaging techniques, and is released from and reloaded into the vesicles by neuronal depolarization. This suggests the possibility that polyP may function as a neurotransmitter under non-pathological conditions. We conclude that polyP functions as a peripheral neuroactive compound and as an endogenous neuromodulator of the CNS.

PolyP as a neuroactive compound
Current-clamp recordings of cultured hippocampal and DRG neurons determined that polyP can functionally stimulate excitable cells. Puff application of medium chain length polyP molecules (60 orthophosphate groups in length on average, see Methods for concentration details) elicited a dramatic increase in AP firing rates from both the CNS ( Figure 1A) and PNS ( Figure 1B) neurons. Additionally, some hippocampal neurons (n = 8 of 24) had a multi-phasic response to polyP: AP firing frequency increased initially, but ceased as the membrane resting potential depolarized, then AP firing reappeared as the neurons repolarized ( Figure 1C). The polyP concentration was chosen based on the reported average polyP concentrations observed in CNS tissue [9,10].

PolyP application hyperpolarizes the voltage sensitivity of Na V channel activation
Voltage-clamp recordings of cultured hippocampal neurons were conducted to identify the ion channel constituents of the AP whose activity is altered by polyP application. Tetrodotoxin (TTX)-sensitive voltage-gated sodium (Na V ) channels are highly expressed in most neurons, driving AP generation with the inward flux of a large rapidly activating and rapidly inactivating sodium current. Holding at −86 mV, polyP evoked current spiking activity (Figure 2A, n = 4) completely blocked by TTX (n = 3). Using recording conditions tailored to study Na V channel currents (I NaV , see Methods), a step protocol revealed that I NaV activation significantly shifted to hyperpolarized voltages in the presence of polyP ( Figure 2B, C and Table 1). A difference of more than 20 mV in the voltage of half activation (V h ) was calculated from Boltzmann fits of the normalized cumulative data ( Figure 2C) before and after polyP application.

PolyP significantly inhibits I KV and activates I CaV
Enhanced excitability of CA1 pyramidal neurons through G protein coupled receptor (GPCR)-evoked inhibition of K V channels is well documented [11][12][13][14]. With TTX in the bath, a ramp protocol revealed that polyP addition blocked I KV by 36.3% ± 5.4 (n = 12; Figure 3A & B). However, 500 μM suramin failed to prevent the polyPinduced block of I KV (34.5% ± 6.3, n = 6), suggesting polyP directly blocks I KV . To ensure the outward currents are attributable to I KV , recordings were conducted with cesium-gluconate internal solutions. Under these conditions, I outward was 15.8 pA/pF ± 5.2 and increased by Figure 1 Neuronal response to polyP. Both CNS and PNS neurons responded to the application of polyP with a vehement burst of activity. A) Representative response of a hippocampal neuron to direct application of polyP (n = 16 of 24), in which rapid, persistent firing is observed. B) Representative response of a DRG neuron to polyP. Small nociceptive DRG neurons (<35 microns) responded with rapid firing (n = 24). C) Some neurons responded to polyP with a multiphasic response: rapid firing followed by protracted depolarization, during which time firing did not occur, that was restored upon recovery of the membrane potential (8 of 24).

PolyP stimulates neurons and glia in hippocampal co-cultures
To confirm our electrophysiology data and extend our analysis to glia cells, we assessed the response of   dissociated hippocampal neuronal/glia co-cultures to the addition of polyP using high spatial and temporal resolution calcium imaging. Our analysis software separates the slower responses attributable to depolarization of the underlying astrocyte layer from neuronal action potentials, producing activity "fingerprints" of >300 neurons simultaneously (see Methods for experimental details). Figure 4A shows the response of the neurons to polyP, where most neurons responded immediately with a burst of activity and some but not all had a secondary burst a few minutes later ( Figure 4A), consistent with the patterns observed in Figure 1B. In Figure 4B we show the underlying glial activity patterns corresponding to the neuronal pattern seen in the experiment in Figure 4A. PolyP evokes glial calcium signals during polyP addition in synchrony with neuronal activity, potentially driven by glutamate release from the actively firing neurons. In astrocytes, polyP binding to P2Y 1 receptors stimulates Gq-coupled calcium release from internal stores [5]; here we show that suramin prevents the initial burst of glial activity ( Figure 4C). However, a few minutes after polyP addition glial activity increased, potentially in response to neuronal activity.

DAPI stains polyP in distinct punta along axon-like projections
PolyP is a gliotransmitter, released by astrocytes to signal neighboring astrocytes [5]. Here we demonstrate that polyP is also released and taken up by neuronal synapses. PolyP can be imaged in live cells using DAPI stain and collecting emission wavelengths at 560 nm [1,[20][21][22]. In our hippocampal cultures polyP is found in distinct puncta along axon-like projections and in the cell body perinuclear space and cytoplasm ( Figure 5A). In contrast, at 456 nm emission, DAPI stains nuclear DNA and dendritic nucleic acids ( Figure 5B). Overlap of the two emission signals clearly shows the differential compartmentalization of polyP versus nucleic acids ( Figure 5C). Photoconductive stimulation can induce neuronal cultures grown on silicon wafers to fire in a high frequency, physiological manner [23]. Before stimulation, multiple polyP positive puncta are visible along axon-like projections ( Figure 5E). Immediately following stimulation a subpopulation of the puncta disappear ( Figure 5F), suggesting that the polyP is released concurrent with neuronal activity. As not all puncta vanish (~16% +/−5 of total puncta remain, n = 6 experiments), polyP may be in vesicles and other compartments outside of the readily released pool. The remaining puncta provide corroborative evidence that there was no shift in focal plane during the imaging. Interestingly, puncta reappeared within 30 seconds of the firing event ( Figure 5G), suggesting that the polyP is rapidly replenished. Together, these data suggest polyP is can be released and replenished into neuronal puncta, similar to other neurotransmitters. As DAPI fails to label polyP in fixed cells, co-staining with cellular/synaptic markers was not possible. However, polyP isolated from synaptosomes can be labeled with DAPI ( Figure 5D lane 5). Compared to synthetic polyP of varying lengths ( Figure 5D lanes 1-3), synaptically isolated polyP ran parallel to long polyP (~130 polymer chain length). An in vitro enzymatic assay (see methods) indicated that the synaptic vesicle fraction contained~35 pmol/mg of protein. Phosphatase treatment completely hydrolyzed synaptic polyP ( Figure 5D, lane 6 and 7). Importantly, synaptosomal release of polyP can be triggered by KCl-evoked membrane depolarization as evidenced by the reduction of the polyP band density ( Figure 5D, lane 4). This supports the possibility that similar release of polyP can occur in response to membrane depolarization in the intact neurons.

Discussion
PolyP stimulates neuronal activity by modulating several key components of the action potential. Our results show that polyP blocks K V channels, and potentiates Na V and Ca V channel activity. K V channels contribute to the maintenance of the neuronal resting membrane potential. When K V 7 in particular, is inhibited, depolarization follows [24], and spontaneous CA1 pyramidal neuron firing is induced [14], presumably as the threshold for Na V channel activation is reached. PolyP further promotes AP generation by shifting Na V channel voltage sensitivity so that less depolarization is required to trigger Na V channel activation. PolyP also magnifies the entry of calcium through Ca V channels, potentially amplifying downstream intracellular signaling pathways. The concerted effects of polyP on neuronal ion channels may promote normal AP generation, while excessive release of polyP may contribute to hyperexcitability disorders such as epilepsy.
In astrocytes, polyP activation of the P2Y1 receptor Gq signaling pathway underlies glial communication [5]. The polyP effect we report on neuronal ion channels differs significantly from the well characterized effect of P2Y GPCR activity on these same channels [25]. In particular, nucleotides inhibit I CaV through the activation of P2Y1 receptors [26], while polyP potentiates I CaV. Nucleotides inhibit the TTX-sensitive I NaV [27], as does GPCR activation of PKA/PKC [28]; polyP significantly hyperpolarizes activation of I NaV . P2Y1 receptors activate hippocampal neuron calcium-dependent K + channels [29], however, polyP did not elicit this activity. Nucleotides inhibit I KV through P2Y1 receptors [24], an effect mimicked by polyP. In our calcium imaging experiments, suramin did prevent polyP-evoked signaling in glia. However, suramin failed to prevent neuronal polyP I KV inhibition, suggesting the GPCR pathway is not involved. Rather, we suggest that polyP effects neuronal ion channel gating properties by altering the local surface charge. Long chain polyP is high in negative charges and may increase the density of fixed negative charges surrounding the channels [30]. As a result polyP would create an environment promoting Na V and Ca V channel activation. We predict that the K V channels are directly blocked by polyP. The ability of polyP to simultaneously modulate the activity of a range of channels further supports the interpretation that its activity is linked to the global electrostatic change of the channels environment rather than its binding to specific sites. A parallel can be made to a recent study which proposes that polyP acts as a chaperone, simultaneously interacting with a number of proteins [31]. In contrast, polyP modulation of TRPM8 channels is attributed to direct ionic interactions [3,32].
We found that polyP accumulates in synaptic vesiclelike structures under physiological conditions. Upon depolarization, and in response to a burst of neuronal activity, polyP is released into the extracellular milieu where it will stimulate further neuronal and glial communication. Astrocytes are also a source of polyP [5]. Abramov's research group recently reported that less that 3% of neurons responded to 50 μM polyP ( [5]; calcium imaging experiments). For our electrophysiology experiments, we chose to apply polyP using a puffer pipette. The technique mimics synaptic release more closely; however the targeted neurons see a higher concentration of the applied agonist before it diffuses into the bath. 100% of neurons tested responded to polyP.
PolyP regulates many critical cellular processes including mitochondrial ion transport and respiratory chain activity [33,1], and is a constituent of the normal physiological environment. However, pathological conditions can lead to a dramatic increase in polyP concentrations. During inflammation polyP is released from activated platelets [34,6], stimulating blood coagulation [7]. Given its activity on sensory DRG neurons, polyP is expected to elicit a pain response. Supporting this hypothesis, peripheral sensory TRP channel activity is dependent on polyP. Specifically, normal activity of TRPM8 requires the presence of polyP as a structural component of the channel [3] while TRPA1 channels rely on polyP as a cofactor regulating their activity [35]. Centrally, epileptic seizure activity could be expected with a dramatic increase in polyP. It will be interesting to determine if polyP concentrations rise under conditions of inflammation, stroke, seizure, or ischemia, contributing to global neuronal changes.

Conclusions
We conclude that inorganic polyphosphate is a previously unrecognized modulator of neuronal Na V , Ca V , and K V channels, potentially playing an important role in normal neuronal synaptic transmission as well as pain and other neuropathological conditions.

Neuronal cultures
Hippocampal co-cultures (containing both neurons and glia) were prepared as previously described [36] and used for the electrophysiology, network activity and polyP imaging experiments. Briefly, hippocampi were dissected from P0 (newborn) Sprague Dawley rat pups, and neurons and glia were dissociated and plated together on pretreated silicon wafers [23]. Cultures were maintained in Eagle's Basal Media (BME, GIBCO-Invitrogen) supplemented with B-27, penicillin, streptomycin and L-glutamine. Cultures are maintained with 4% Fetal Bovine Serum (FBS) for the first week to aid in establishment of the cultures; FBS is subsequently reduced to restrict growth of the astrocytes. Functional synapses form after a few days in culture, and robust spontaneous network activity could be observed after approximately 2 weeks. Dorsal root ganglion (DRG) neurons were cultured as previously described [37].

DAPI staining of neuronal cultures
2 weeks old hippocampal neuronal cultures were stained with DAPI (4′,6-diamidino-2-phenylindole) at a final concentration of 0.1 μg/ml in extracellular bath solution (135 mM NaCl, 2 mM MgCl 2 , 3 mM CaCl 2 , 5 mM KCl, 5 mM HEPES, and 10 mM Glucose, pH 7.4 with TEA-OH and 320 mOSM with D-sorbitol) for 5 minutes. Cultures were rinsed briefly and imaged in the 560 nm (emission) range to detect polyP, and at 456 nm to detect nucleic acids. For the polyP channel, color images of the emission band were taken with a Nikon D3 camera, and then the specific wavelength further refined with color band filtering of the RGB image. Separated images were then recombined for the static image presentation ( Figure 5A-C) using Photoshop.
Live imaging of release of polyP from synapses ( Figure 5E-G) was performed on the same cultures, using the same DAPI-polyP filter set, without the post color band processing.

Synaptosome and synaptic vesicle isolation
Synaptosomes and synaptic vesicles were isolated from cortical tissue of one adult Sprague Dawley rat. A synaptic vesicles isolation kit (Sigma, SV0100) was used to prepare an enriched synaptic vesicle fraction. The synaptosomal pellet was resuspended in lysis buffer and incubated on ice for 45 minutes, followed by centrifugation at 20,000 g for 20 minutes. Synaptic vesicles were collected as a pellet by final centrifugation of resultant supernatant at 70,000 g for 45 minutes and resuspended in storage buffer. Isolated vesicles were frozen and further processed for polyP extraction and concentration determination.

PolyP extraction and concentration determination
PolyP was extracted from synaptic vesicle preparation using chloroform-methanol extraction protocol. 100 μL of synaptic vesicles in storage buffer were incubated with 60 μL of methanol for 20 minutes followed by the addition of 120 μL of chloroform. PolyP was collected from the aqueous (upper) phase of the resultant mixture. Collected sample was concentrated to 40 μL using a Speed Vac and run on a PAGE gel (15%). The running buffer (TBE) contained (mM) 90 Tris, 90 borate, 2 EDTA, pH 8.3. Gels were run at 100 V for 1 hour and incubated in a fixative solution containing 25% methanol and 5% glycerol at pH = 8.0. PolyP was visualized using negative staining with DAPI as previously described [38]. Briefly, gels in fixative solution were incubated with 2 μg/ml of DAPI for 30 min and then exposed to 365 nm light using a UV transilluminator for 20-30 min to induce photobleaching. DAPI bound to polyP produced strong photobleaching which resulted in the appearance of dark bands on the gel. Samples were treated with calf alkaline phosphatase (10 U/ml; Sigma).
PolyP was colorimetrically quantified as the amount of orthophosphate residues (Pi) released upon sample treatment with recombinant exopolyphosphatase (PPX; from yeast Saccharomyces cerevisiae). Sample was mixed with 100 mM Tris-HCl, pH 7.5, 50 mM ammonium acetate, 5 mM magnesium acetate, and PPX in excess to ensure complete polyP hydrolysis. 50 μl of the processed sample was then mixed with 500 μl of 2.5% solution of ammonium molybdate prepared in 5 N sulfuric acid and 50 μl Fiske-Subbarow reagent (Sigma), with the total volume adjusted to 1 ml with water. The optical density of the solution at 650 nm was compared with Pi standards (Sigma) to determine the sample quantity of Pi [21].

Electrophysiology
Current and voltage clamp recordings were obtained from primary culture hippocampal and DRG neurons. Small (nociceptive) DRG neurons were used for electrophysiological recording after more than 24 hours in culture to allow adequate extension of processes. Hippocampal neurons were cultured for 2 weeks to allow extensive synaptic networks to develop before recordings were made, with the exception of I NaV records. I NaV recordings were obtained from hippocampal neurons cultured 3 days and treated for 5 minutes with PBS versene (Invitrogen) to mitigate space clamp issues arising from extensive neuronal processes. For I NaV recordings the external media Whole-cell patch clamp recordings from DRGs and hippocampal neurons were made using a Multiclamp 700B amplifier, a Digidata 1440A digital-to-analog converter, and pClamp10 software (Molecular Devices, Palo Alto, CA). Current-clamp recordings were performed by switching to current-clamp mode after a stable whole cell configuration was formed in voltage clamp mode. A membrane potential of at least of −40 mV was obtained from cells used in the statistical analysis. In a few cases, depolarizing pulses of 10 to 20 pA, lasting about 1-2 s, caused local increases in membrane potential to −30 mV. The experiments were performed at room temperature (21-23°C). During voltage-clamp recordings, capacitance and series resistance were compensated by 75-85%. Ramp protocol: from a holding potential (HP) of −70 mV, voltage ramps from −100 to +40 mV (400 ms in duration, with a 40 ms delay at −100 and +40 mV) were applied every 2 seconds. Data were low-pass filtered at 2 kHz and digitized at 10 kHz.
Step protocol: from a HP of −100 mV, 30 ms voltage steps in 5 mV increments from −70 to +20 mV were applied every 2 seconds. Data were low-pass filtered at 6 kHz and digitized at 50 kHz. Data are presented with liquid junction potentials (LJP) corrected for.
Data analysis of I NaV : Peak I NaV from the step protocol was plotted as a function of the test potential (LJP corrected) to generate current-voltage relations (I-V). For the mean I-V ( Figure 3C), paired data were normalized to the I peak before polyP application and averaged. Individual (Table 1) and mean I-V relations ( Figure 3C) were fit with the Boltzmann equation, where I peak = (V -E rev ) G (1/1 + exp (V a,1/2 -V)/S)), and E rev is the reversal potential, V a,1/2 is the half-activation potential, G is the maximum slope conductance, and S is the slope factor that is inversely proportional to the effective gating charge.

Data analysis of I KV and I CaV
Currents recorded during ramp protocols were normalized by cell capacitance measurements for the comparison of current densities.

Statistical analysis
A paired Student's t-test was used to determine if significant differences resulted from polyP application. Results are reported in the figures.
Calcium imaging and recording of network activity 18-21 DIV neuronal cultures for the network activity analysis were prepared as described above. To capture neuronal and glial activity, cultures were first loaded with the calcium dye Fluo4-AM by incubating the silicon wafer in 500 μl of recording solution and 2 μg of Fluo-4 at 37°C for 30 minutes prior to optical recording. Wide field images of neuronal activity were made at 15 frames per second using an Olympus BX60 microscope with a 10× water immersion objective and a Watec WAT-120 N camera. Individual neuronal cell bodies in the images were first identified from an averaged frame spanning the time course of the recording, based on their morphology. Neurons were numbered and their activity traces extracted by calculating the mean grey value in a region of interest (ROI) encompassing the cell body. Their identity as neurons was confirmed by the criterion that action potentials are detected, as determined by a sharp (less than 15 ms) rise in internal calcium. We then removed the depolarization patterns of the neurons from the underlying glia calcium signal, and segregated the field of view into a grid array. Changes in these surrounding regions are taken as glial depolarizations, again confirmed by their gradual increase and decrease in comparison to the more rapid speed of a neuronal firing event.
Once the neuronal and glial movies were split, each was processed by custom written routines in Matlab. For the neuronal analysis, activity fingerprints were made that display the neuron index number vertically, and the point in time that a firing event occurred horizontally. For the glial depolarizations, each grid reference point was treated like a cell, and displayed in a similar manner to the neurons. This provided a general correspondence of the physical region of the glial activity with the reference point of the neurons. In all experiments light intensity was kept to the minimum in order to avoid photobleaching. 3 separate rounds of hippocampal cultures were used for imaging experiments.

Use of animals
Newborn rat pups (not gender identified) for primary neuronal cultures were obtained from Sprague Dawley dams, purchased from Charles River Laboratories International, Inc. All experimental protocols were approved by the University of Calgary Conjoint Faculties Research Ethics Board, and performed under protocol #M5083.

Statistics
No parametric statistical analysis of imaging was performed for this manuscript as all observations were binary events (e.g. if bursting occurred or not, if synaptic staining disappeared or not). n values represent the number of times an experiment was repeated and the phenomena was observed.