Fibroblast growth factor homologous factors control neuronal excitability through modulation of voltage-gated sodium channels

Abstract

Neurons integrate and encode complex synaptic inputs into action potential outputs through a process termed "intrinsic excitability." Here, we report the essential contribution of fibroblast growth factor homologous factors (FHFs), a family of voltage-gated sodium channel binding proteins, to this process. Fhf1-/-Fhf4-/- mice suffer from severe ataxia and other neurological deficits. In mouse cerebellar slice recordings, WT granule neurons can be induced to fire action potentials repetitively (approximately 60 Hz), whereas Fhf1-/-Fhf4-/- neurons often fire only once and at an elevated voltage spike threshold. Sodium channels in Fhf1-/-Fhf4-/- granule neurons inactivate at more negative membrane potential, inactivate more rapidly, and are slower to recover from the inactivated state. Altered sodium channel physiology is sufficient to explain excitation deficits, as tested in a granule cell computer model. These findings offer a physiological mechanism underlying human spinocerebellar ataxia induced by Fhf4 mutation and suggest a broad role for FHFs in the control of excitability throughout the CNS.

Figure 1. Generation of Fhf1 null allele and Fhf1^−/− mice

A) Maps of endogenous Fhf1 gene locus, targeting vector, and the product of successful gene targeting. NEO, G418-resistance cassette driven by phosphoglycerate kinase gene promoter; B, BamHI; P, PstI; R, EcoRI; S, SpeI; X, XhoI; Xb, XbaI. B) Southern blot analysis of mouse embryonic stem (ES) cell G418-resistant clones. Only ES clones demonstrating gene targeting assayed with both 3′ and 5′ probes were selected for mouse chimera production. C) Southern blot analysis of progeny from intercross between Fhf1^+/− mice.

Figure 2. Motor coordination assays

A) Ledge test. Mice were placed on a high, narrow ledge and timed until fall up to a maximum of 60 seconds. Significance of time differences in mutant mouse groups were analyzed by ANOVA for significance (p < .05). *, Fhf4^−/− mice significantly worse than wild type; **, Fhf1^−/−Fhf4^−/− mice performed significantly worse than wild-type or Fhf4^−/−. B–E) Gait test. Paws of mice were painted (forepaws red, hind paws green) and their footprints recorded while walking through a 7 cm diameter horizontal cylinder. Representative footprints from Fhf1^−/− (B), Fhf4^−/− (C), and Fhf1^−/−Fhf4^−/− (D) mice are shown, along with analysis of hind paw misplacement along the forward axis (E). *, Fhf4^−/− mice significantly impaired in comparison to Fhf1^−/−; **, Fhf1^−/−Fhf4^−/− mice significantly worse than Fhf4^−/−.

Figure 3. Cerebellar histology and granule neuron integrity in Fhf1^−/−Fhf4^−/− mice

(A, B) Histology of wild type (A) and Fhf1^−/−Fhf4^−/− (B) cerebellar cortex cryosections stained with cresyl violet, eosin Y, and Luxol fast blue. m, molecular layer; g, internal granule layer; p, Purkinje cell soma; f, fiber layer. (C – H) Confocal immunofluorescence detection of Nav1.6 in cerebellar granule cells in situ. Cryosections of wild-type (C–E) and (F–H) granule layer were probed for axon initial segment marker ankyrin G (C, F) and Nav1.6 (D, G); merged images (E, H). (I, J) Phase contrast micrographs of P10 cerebellar neurons cultured for 21 days. Abundant small round-soma granule neurons (e.g. arrows) are discernable in wild-type (I) and Fhf1^−/−Fhf4^−/− (J) cultures. (K – N) Immunofluorescence analysis of granule neurons cultured for 30 days. Wild type (K, M) and Fhf1^−/−Fhf4^−/− (L, N) cells were stained with either anti-glutamate transporter EAAC1 (green) and anti-pan-sodium channel alpha (red) (K, L) or anti-pan-sodium channel alpha (green) and ankyrinG (red) (M, N) and counterstained with DAPI (blue). Surface and vesicular EAAC1 (K, L) visualizes cell soma and processes. Arrows point to axon initial segments. Overlapping ankyrinG and sodium channel staining (M, N) appears yellow.

Figure 4. FHF proteins localize to the axon initial segment and associate with sodium channels

A) FHF1/sodium channel complexes in cerebellum. Triton X-100 extracts of mouse cerebellum were immunoprecipitated with pan-sodium channel alpha subunit monoclonal antibody, rabbit anti-FHF1, or rabbit anti-FRS2β. For FHF1 immunoprecipitation, the antibodies were preincubated with the immunizing FHF1 12-mer peptide, a non-specific 12-mer peptide, or no peptide. Captured immunoprecipitates were electrophoresed on a 7.5% polyacrylamide-SDS gel, blot transferred, and probed with pan-sodium channel monoclonal antibody. Migration of molecular weight standards is indicated (arrows), as is the position of detected sodium channel (arrowhead). The specificity of FHF1/channel coprecipitation is established by loss of pulldown when anti-FHF1 is preincubated with immunizing peptide, but not with an alternative peptide, and by virtual absence of precipitation using the unrelated anti-FRS2β antibody raised and purified by identical procedures. B–G) FHF1-GFP fusion proteins localize to cerebellar granule neuron axon initial segment. Wild type mouse granule cells cultured for 18 days were transfected with expression constructs for GFP fused to mouse FHF1a (B–D) or human FHF1b (E–G). After 24 hrs, cells were fixed and incubated with anti-GFP (mouse IgG2a) and anti-pan-sodium channel alpha (mouse IgG1), and bound antibodies were visualized by indirect immunofluorescence with biotinylated anti-mIgG2a + Streptavidin-ALEXA488 (B, E) and ALEXA594-conjugated anti-mIgG1 (C, F). Merged images (D, G) include DAPI nuclear stain. H–M) Native FHF localizes to axon initial segment. Rat hippocampal neurons in primary culture for 12 days were fixed and incubated with rabbit polyclonal anti-FHF2 antibodies (I, L) together with either anti-ankyrinG (mIgG1) (H) or anti-pan-sodium channel alpha (mIgG1) (K). Bound antibodies were visualized with biotinylated anti-rabbit IgG + streptavidin-ALEXA 488 (I, L) and anti-mouse IgG-ALEXA 594 (H, K). Merged images (J,M) include DAPI stain. Arrows denote axon initial segment.

Figure 5. Intrinsic excitability of wild-type and Fhf mutant granule neurons from brain slice preparations

A) Maximal spike frequency for wild-type and mutant cells. For each patched granule neuron, current injection command sweeps of 800 millisecond duration and of different amplitudes were applied, action potentials recorded, and the maximum number spikes in any sweep determined. n, number of cells recorded in genotype. *, Fhf4^−/− and Fhf1^−/−Fhf4^−/− less excitable than wild type (p < .01 by ANOVA). B) Distribution of spikes during 800 msec current injection. For sweep giving highest spike frequency in each cell, the number of spikes in each of the first five 100 msec intervals was counted. *, number of spikes during interval for mutant cells less than for wild type cells (p < .05 by ANOVA). C) Current-to-spike relationship among wild-type neurons. For each cell, smallest injected current to generate any spikes was normalized to 1. ——, cells displaying graded current-to-spike profile; - - - - -, cells displaying steep current-to-spike profile. D–G) Membrane voltage recordings of representative cells for each genotype during current injection: (D) wild type, (E) Fhf1^−/−, (F) Fhf4^−/−, (G) Fhf1^−/−Fhf4^−/−. Repetitive firing is impaired in the Fhf4^−/− neuron, and the Fhf1^−/−Fhf4^−/− neuron generated only one initial spike. Also note elevated voltage spike thresholds and afterhyperpolarizations in (F) and (G). H, I) Current traces of wild type (H) and Fhf1^−/−Fhf4^−/− (I) cells following voltage-clamp depolarization from −80 to 0 mV (inset). Both cells generate typical fast inward inactivating current and both inactivating and noninactivating outward currents.

Figure 6. Voltage-dependent activation and inactivation of sodium channels in wild-type and Fhf1^−/−Fhf4^−/− cultured granule neurons

Recordings were in solutions only allowing for gated sodium currents. A) Sodium channel activation in a wild-type neuron. Superimposed current traces are from several voltage depolarization step commands from −85 mV holding potential (left inset). Peak conductance as a percentage of maximal conductance is plotted versus V_m (right inset). B) Sodium channel activation in a Fhf1^−/−Fhf4^−/− neuron. Voltage dependence of activation is similar to that of the wild-type cell (A). C) Sodium channel inactivation in a wild-type neuron. Superimposed current traces are from a several voltage commands (left inset) consisting of preconditioning voltage step (125 msec) followed by depolarization to −15 mV (only time frame bracketing the −15 mV depolarization is shown). Peak available conductance as a percentage of maximal conductance is plotted versus V_m (right inset). D) Sodium channel inactivation in a Fhf1^−/−Fhf4^−/− neuron. Voltage dependence of inactivation is shifted leftward more than 10 mV in comparison to transition in the wild-type cell (C). E) Voltage dependence of activation and inactivation for all recorded cells. The graph shows the average voltage dependence of channel activation in wild-type (closed circles, n = 9) and Fhf1^−/−Fhf4^−/− (open circles, n = 11) cells, together with average voltage dependence of channel inactivation in wild-type (closed diamonds, n = 8) and Fhf1^−/−Fhf4^−/− (open diamonds, n = 9) cells. The 13 to 14 mV negative shift for voltage-dependent channel inactivation in Fhf1^−/−Fhf4^−/− cells is highly significant (P < 0.00003). G) Persistent sodium currents. Current traces for a representative wild-type (solid line) and Fhf1^−/−Fhf4^−/− (dappled line) cell following depolarization to −35 mV illustrate current persisting after 25 msec, which is reduced in the mutant.

Figure 7. Rate constants for sodium channel inactivation and recovery in wild-type and Fhf1^−/−Fhf4^−/− cultured granule neurons

Recordings were in solutions only allowing for gated sodium currents. A) Rates of inactivation. Wild type (n = 6) and Fhf1^−/−Fhf4^−/− (n = 10) granule cells were depolarized to different voltages, and the decay from peak inward current was used to calculate the time constant (τ_inactivation) associated with channel fast inactivation. At either −15 mV or −5 mV, τ_inactivation in Fhf1^−/−Fhf4^−/− cells was significantly shorter than in wild type cells (P < 0.001). B) Superimposed and scaled current traces from wild-type (solid line) and Fhf1^−/−Fhf4^−/− (dashed line) cells following depolarization to −5 mV illustrates the faster current decay in the mutant cell. Traces were minimally offset along x-axis to visualize the similar rapid onset of sodium current. C–E) Recovery of sodium channels from inactivation. Superimposed current traces recorded from a wild-type neuron (C) in response to voltage commands consisting of two 10 msec depolarizations separated by variable recovery phases at −85 mV. The second peak sodium current associated with several recovery times is indicated (arrowheads). Arrow denotes resurgent sodium current accompanying repolarization. Superimposed current traces recorded from a Fhf1^−/−Fhf4^−/− neuron (D) show slower rate of recovery. (E) Percent maximal sodium channel recovery versus recovery time for wild-type (solid squares, n = 7) and Fhf1^−/−Fhf4^−/− (open squares, n = 6) cells.

Figure 8. Computer modeling of intrinsic excitability in wild-type and Fhf1^−/−Fhf4^−/− granule neurons

The kinetics of state-to-state sodium channel transitions was modeled to approximate the recorded physiological behavior of channels in wild-type and mutant neurons. A) Sodium currents generated in a voltage clamp simulation from the holding potential of −80 mV to various test potentials, as indicated. Note lesser amplitude and faster inactivation kinetics in the Fhf1^−/−Fhf4^−/− mutant. The decrease of persistent sodium current in the Fhf1^−/−Fhf4^−/− mutant (dotted line) model in comparison to wild-type (solid line) model is shown in the inset. B) Voltage-dependent activation, inactivation, and recovery rate of sodium channels in the wild-type and Fhf1^−/−Fhf4^−/−mutant models. The voltage dependence of activation is minimally affected by mutation, while steady-state inactivation shows a 10mV leftward shift in the Fhf1^−/−Fhf4^−/− mutant simulation. The rate of recovery from inactivation for sodium channels in the Fhf1^−/−Fhf4^−/− mutant model is slowed, analogous to channel properties in recorded cells. At 5 milliseconds, only 50% of channels in the mutant model have recovered, while recovery is 90% in the wild-type model. C) Current-clamp simulation. The wild-type granule cell model shows repetitive firing upon a depolarizing current injection of 10 pA. For the Fhf1^−/−Fhf4^−/− model, sodium channel density was doubled to allow similar peak sodium conductance, as seen in some recorded neurons. Even with this compensation, firing is abolished in the mutant model at 10 pA current injection, and just a single early spike is elicited with more injected current (70 pA). V_threshold is also higher in the Fhf1^−/−Fhf4^−/− model.

Scheme 1

Kinetic scheme of the sodium channel used in the present simulations. The effect of this scheme on granule cell sodium currents and excitability are shown in Figure 8.