Loss of intracellular FGF14 (iFGF14) increases excitability of mature hippocampal pyramidal neurons

The intracellular fibroblast growth factor subfamily, Fgf11–14 in mice and FGF11–14 in humans, generates four iFGF proteins, iFGF11–14, with similar “core” sequences and distinct N-termini (Goldfarb, 2005; Pablo and Pitt, 2017; Ornitz and Itoh, 2022). Alternative exon usage and splicing (Munoz-Sanjuan et al., 2000) generate further iFGF protein diversity. Also referred to as fibroblast growth factor homologous factors, FHF1–4, the iFGFs share sequence and structural homology with the canonical FGFs but lack the signal sequence for secretion (Smallwood et al., 1996; Hartung et al., 1997; Munoz-Sanjuan et al., 2000; Olsen et al., 2003; Goldfarb, 2005; Pablo and Pitt, 2017; Ornitz and Itoh, 2022). The iFGFs have been shown to bind to the C-termini of voltage-gated Na⁺ (Nav) channel pore-forming (α) subunits (Liu et al., 2001, 2003; Wittmack et al., 2004; Lou et al., 2005; Goetz et al., 2009; Wang et al., 2012) and to modulate the time- and voltage-dependent properties of heterologously expressed Nav channels (Liu et al., 2001, 2003; Wittmack et al., 2004; Lou et al., 2005; Rush et al., 2006; Laezza et al., 2007, 2009; Dover et al., 2010; Wang et al., 2011a), as well as of native Nav currents in cardiac (Wang et al., 2011b, 2017; Park et al., 2016; Abrams et al., 2020; Chakouri et al., 2022; Angsutaraux et al., 2023; Fischer et al., 2024, Preprint) and neuronal (Goldfarb et al., 2007; Dover et al., 2010; Goldfarb, 2012; Venkatesan et al., 2014; Yan et al., 2014; Bosch et al., 2015; Puranam et al., 2015; Alshammari et al., 2016; Yang et al., 2017; Effraim et al., 2019, 2022) cells.

Unlike canonical FGFs, the iFGFs are broadly and robustly expressed in the adult peripheral and central nervous systems (Smallwood et al., 1996; Hartung et al., 1997; Wang et al., 2000; Ornitz and Itoh, 2022), and mutations in FGF12 and FGF14 have been linked to inherited disorders of neuronal excitability (van Swieten et al., 2003; Dalski et al., 2005; Brusse et al., 2006; Misceo et al., 2009; Choquet et al., 2015; Al-Mehmadi et al., 2016; Siekierska et al., 2016). FGF14, for example, was identified as the locus of mutations in spinocerebellar ataxia type 27, SCA27 (now SCA27A), a rare autosomal dominant neurological disorder that presents with profound and progressive motor and cognitive impairment (van Swieten et al., 2003; Dalski et al., 2005; Brusse et al., 2006; Misceo et al., 2009; Choquet et al., 2015). More recently, GAA repeat expansions in FGF14 were identified in patients with late-onset ataxias, now referred to as SCA27B (Pellerin et al., 2023; Rafehi et al., 2023a, 2023b; Mohren et al., 2024, Preprint).

Mice (Fgf14^−/−) harboring a targeted disruption of Fgf14 display profound motor and cognitive deficits (Wang et al., 2002; Xiao et al., 2007; Wozniak et al., 2007), that is, phenotypes similar to those seen in individuals with SCA27A (Brusse et al., 2006; Misceo et al., 2009). We previously demonstrated that spontaneous, high-frequency repetitive firing is markedly attenuated in Fgf14^−/− cerebellar Purkinje neurons (Shakkottai et al., 2009; Bosch et al., 2015), the sole output neurons of the cerebellar cortex (Billard et al., 1993; Gauck and Jaeger, 2000; Ito, 2001). Similar deficits in balance and motor coordination and spontaneous firing were observed with acute, in vivo shRNA-mediated knockdown of Fgf14 in adult Purkinje neurons (Bosch et al., 2015). Loss of iFGF14, however, did not measurably affect the distribution of Nav α subunit or Ankyrin G labeling at the axon initial segments (AIS) of mature cerebellar Purkinje neurons. Voltage-clamp experiments revealed that the loss of iFGF14 results in a hyperpolarizing shift in the voltage dependence of inactivation of the transient component of the Nav current (Bosch et al., 2015). We also showed that membrane hyperpolarization can “rescue” repetitive firing in adult Purkinje neurons with acute, Fgf14-shRNA-mediated knockdown of Fgf14 and that viral-mediated expression of iFGF14 in adult Fgf14^−/− Purkinje neurons rescued spontaneous firing and improved motor coordination and balance (Bosch et al., 2015).

Loss of iFGF14 is also linked to cognitive dysfunction. Individuals with SCA27A, for example, have progressive deficits in learning and memory (Brusse et al., 2006; Misceo et al., 2009), and compared with wild type (WT) littermates, Fgf14^−/− animals perform poorly in the Morris water maze (Wozniak et al., 2007), suggesting impaired hippocampal functioning. The experiments here were undertaken to test the hypothesis that iFGF14 also modulates Nav current inactivation to regulate the intrinsic excitability of hippocampal pyramidal neurons, the main output neurons of the hippocampus. In contrast to expectations, we found that evoked repetitive firing rates are higher (not lower) in Fgf14^−/− than in WT, hippocampal CA1 pyramidal neurons. In addition, whole-cell voltage-clamp recordings demonstrate that the loss of iFGF14 does not measurably alter the voltage dependence of Nav current inactivation in CA1 pyramidal neurons. In addition and in contrast with previously published findings on dissociated hippocampal neurons (Pablo et al., 2016), the loss of iFGF14 does not disrupt the localization of Nav1.6 channels at the AIS of mature hippocampal pyramidal neurons in vivo.

All experiments involving animals were performed in accordance with the guidelines published in the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and all protocols were approved by the Washington University Institutional Animal Care and Use Committee (IACUC). Animals harboring a targeted disruption of the Fgf14 locus (Fgf14^−/−), described in Wang et al. (2002), were generated by crossing Fgf14^+/− heterozygotes congenic in the C57BL/6J background; genotypes were confirmed by PCR. Adult male and female WT and Fgf14^−/− animals of varying ages (4–10 wk) were used, as indicated in descriptions of the individual experiments.

For electrophysiological recordings, acute (350 μM) slices were prepared from (4- to 8-wk old) male and female WT and Fgf14^−/− animals: horizontal sections were prepared for recordings from hippocampal CA1 pyramidal neurons and coronal slices were prepared for recordings from pyramidal neurons in layer 5 of the primary visual cortex using standard procedures (Davie et al., 2006); one slice was prepared from each animal (N) and recordings were obtained from multiple neurons (n) in each slice. Briefly, animals were anesthetized with 1.25% Avertin and perfused transcardially with an ice-cold cutting solution containing (in mM): 240 sucrose, 2.5 KCl, 1.25 NaH₂PO₄, 25 NaHCO₃, 0.5 CaCl₂, and 7 MgCl₂ saturated with 95% O₂/5% CO₂. After perfusion, the brains were rapidly removed and coronal or horizontal slices were cut in an ice-cold cutting solution on a Leica VT1000 S microtome (Leica Microsystems, Inc) and incubated in a holding chamber in artificial control cerebrospinal fluid (control ACSF) containing (in mM): 125 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 25 NaHCO₃, 2 CaCl₂, 1 MgCl₂, and 25 dextrose (pH 7.2; ∼310 mOsmol l⁻¹), saturated with 95% O₂/5% CO₂. Horizontal slices were incubated at 33°C for 25 min, followed by at least 35 min at room temperature prior to electrophysiological recordings, and coronal slices were incubated for at least 1 h at room temperature prior to electrophysiological recordings.

Whole-cell current-clamp recordings were obtained at 33 ± 1°C from individual visually identified hippocampal CA1 (or layer 5 visual cortical) pyramidal neurons (n) in acute slices (N), prepared from adult (6–8 wk) WT and Fgf14^−/− animals, as described above, using differential interference contrast optics with infrared illumination. Experiments were controlled and data were collected using a Multiclamp 700B patch clamp amplifier interfaced with a Digidata 1332 acquisition system and the pCLAMP 10 software (Molecular Devices) to a Dell computer. For current-clamp recordings, slices were perfused continuously with warmed (33 ± 1°C) control ACSF, saturated with 95% O₂/5% CO₂. Recording pipettes contained (in mM): 120 KMeSO₄, 20 KCl, 10 HEPES, 0.2 EGTA, 8 NaCl, 4 Mg-ATP, 0.3 Tris-GTP, and 14 phosphocreatine (pH 7.25; ∼300 mOsmol l⁻¹); pipette resistances were 2–4 MΩ. Tip potentials were zeroed before membrane–pipette seals were formed. Pipette capacitances were compensated using the pCLAMP software. Signals were acquired at 50–100 kHz and filtered at 10 kHz prior to digitization and storage. Initial resting membrane potentials (V_r) were between −60 and −80 mV. In each cell, single action potentials and action potential trains were elicited from V_r in response to brief (2.5 ms) and prolonged (500 ms), respectively, depolarizing current injections of variable amplitudes.

Voltage-clamp recordings were obtained at 33 ± 1°C from hippocampal CA1 pyramidal neurons in acute slices prepared from young adult (4–6 wk) WT and Fgf14^−/− animals, as described above. For voltage-clamp experiments, recording pipettes (2–4 MΩ) contained (in mM): 110 CsCl, 20 TEA-Cl, 1 CaCl₂, 2 MgCl₂, 10 EGTA, 4 Mg-ATP, 0.3 Tris-GTP, and 10 HEPES (pH 7.25; ∼310 mOsmol/liter). After formation of a gigaseal, the control ACSF superfusing the slice was switched to a low Na⁺ ACSF which contained (in mM): 25 NaCl, 100 TEA-Cl, 2.5 KCl, 1.25 NaH₂PO₄, 25 NaHCO₃, 2 CaCl₂, 1 MgCl₂, 0.1 CdCl₂, and 25 dextrose, (pH 7.25; ∼310 mOsmol/liter), also saturated with 95% O₂/5% CO₂. Series resistances were compensated ≥80% and the voltage errors resulting from the uncompensated resistances were always ≤4 mV and were not corrected. Voltage-clamp protocols were tested and optimized using the strategy outlined in Milescu et al. (2010) to minimize space clamp errors and enable reliable measurement of Nav currents in WT and Fgf14^−/− hippocampal CA1 pyramidal neurons.

Whole-cell current- and voltage-clamp data were analyzed using ClampFit (Molecular Devices), Microsoft Excel, Mini Analysis (v. 6.0, Synaptosoft, Inc.), and Prism (GraphPad Software, Inc.). In all experiments, input resistances (R_in) were determined from the change in membrane potential produced by a 20-pA hyperpolarizing current injection from V_r. The current threshold for action potential generation in each cell was defined as the minimal current injection, applied (for 5 ms) from V_r, required to evoke a single action potential. Action potential durations at half-maximal amplitude (APD₅₀) were determined from measurements of the widths of action potentials when the membrane voltage returned halfway to V_r (from the peak).

Results obtained in the analyses of current-clamp (waveforms of individual action and repetitive firing properties) and voltage-clamp recordings from individual hippocampal CA1 (or visual cortical) pyramidal neurons (n) are displayed, and averaged/normalized data are presented as means ± SEM. Unless otherwise indicated, statistical analyses were performed using a Student’s (unpaired) t test, and P values are presented in the text in Table 1, in the figures, and/or in the figure legends.

Using previously described methods (Bosch et al., 2015), adult (8–10 wk) WT and Fgf14^−/− mice were anesthetized with 1 ml/kg (intraperitoneal injection) ketamine/xylazine and perfused transcardially with 0.9% NaCl, followed by freshly prepared, ice-cold 1% formaldehyde in 0.1 M phosphate-buffered saline (PBS) at pH 7.4. Brains were removed and post-fixed (1% formaldehyde in 0.1 M PBS) for 1 h at 4°C. Following washing with PBS, brains were cryoprotected in 30% sucrose in PBS overnight at 4°C, subsequently embedded in Tissue Tek OCT (Sakurai Finetek), and frozen. Horizontal (40 μm) cryostat sections of the hippocampus (or cortex) were cut, collected in PBS at room temperature (22–23°C), and stored at 4°C until processed for immunostaining.

Free-floating sections were rinsed twice in PBS, followed by 30 min in PBS with 0.1% Triton X-100 (vol/vol). Sections were then incubated in blocking buffer (PBS plus 10% goat serum) for 1 h, followed by overnight incubation at 4°C with primary antibodies, diluted (1:1,000) in PBS with 0.1% Triton X-100 and 0.1% bovine serum albumin. The primary antibodies used were purified mouse monoclonal Neuromab antibodies obtained from Antibodies, Inc.: anti-iFGF14 (mAb-α-iFGF14; clone N56/21, IgG1), anti-Ankyrin G (mAb-α-Ankyrin G; clone N106/36, IgG2a), and anti-Nav1.6 (mAb-α-Nav1.6; clone K87A/10, IgG1). Following washing in PBS, sections were incubated with goat anti-mouse secondary antibodies conjugated to Alexa Fluor 488, Alexa Fluor 594, or Alexa Fluor 647 (1:400; Life Technologies) in PBS for 1 h at room temperature. Sections were washed in PBS, mounted on positively charged slides, coated with a drop of Vectashield Hardset (Vector Laboratories), and coverslipped; slides were stored at 4°C.

Images were captured using an Olympus Fluoview-500 confocal microscope with a 60× oil immersion objective. Laser intensity, gain, and pinhole size were kept constant across experiments, and data were acquired and analyzed blinded to the experimental group. Sequential acquisition of data from multiple channels was used and z-stacks were collected in 0.5-μm steps and converted into maximum intensity z-projections using NIH ImageJ (Schneider et al., 2012). Line-scan graphs were also prepared using ImageJ, and integrated intensity plots were generated in MATLAB using previously described methods (Grubb and Burrone, 2010; Bosch et al., 2015). With ImageJ, three-channel TIFF files were analyzed from line profiles drawn along the axon initial segment (AIS) of individual C1 hippocampal pyramidal neurons (n), beginning at the soma and using the α-Ankyrin G staining to indicate (mark) the AIS. Fluorescence intensity plot profiles were obtained from each channel and data were imported into Excel for normalization and averaging. The start and end of each AIS were determined by the proximal and distal AIS points, respectively, at which the normalized and smoothed α-Ankyrin G fluorescence intensity declined to 0.33 of the maximal value (in the same AIS). Fluorescence intensities obtained from these line profiles were smoothed using a 40-point sliding mean (Grubb and Burrone, 2010; Bosch et al., 2015). Integrated fluorescence intensities along each AIS were calculated as the area under the curve from the beginning to the end of the AIS. Mean ± SEM values for α-Ankyrin G and α-Nav1.6 immunofluorescence labeling intensities are plotted as a function of distance along the AIS of WT and Fgf14^−/− CA1 pyramidal neurons, and peak integrated intensities in individual AIS are provided. In addition, α-Ankyrin G and α-Nav1.6 labeling intensities measured along each AIS were normalized to the maximal and minimal values in the same AIS; mean ± SEM normalized values along the AIS are also presented. Statistical analyses were conducted using a Student’s (unpaired) t-test, and P values are presented in Fig. 6.

For RNA isolation, hippocampal CA1, primary visual cortical, or cerebellar tissue pieces were dissected from anesthetized (1.5% avertin) adult (8–10 wk) female and male Fgf14^−/− (N = 6) and WT (N = 6) animals and immediately flash-frozen (individually) in liquid nitrogen. Total RNA was isolated from each tissue sample using Trizol and DNase-treated using the RNeasy Tissue Mini Kit (Qiagen) according to the manufacturer’s instructions. The expression levels of the transcripts encoding iFGF11–14 and Navβ1–4 were determined by SYBR green quantitative RT-PCR using sequence-specific primers (Table S1). Data were analyzed using the threshold cycle (C_T) relative quantification (2^−∆Ct) method (Livak and Schmittgen, 2001; Marionneau et al., 2008) using phosphoglycerate kinase (Pgk1) or hypoxanthine guanine phosphoribosyl transferase I (Hprt) as the endogenous control.

Lysates were prepared from frozen hippocampal CA1 and primary visual cortical tissue from adult (8–10 wk) WT (N = 4) and Fgf14^−/− (N = 4) animals in 20 mM HEPES + 150 mM NaCl buffer with 0.5% CHAPS and a protease inhibitor tablet (Roche) using established methods (Brunet et al., 2004; Norris and Nerbonne, 2010). Immunoprecipitations (IPs) from these lysates were performed with the (Neuromab) antibody mAb-α-iFGF14 conjugated to protein G Dyna beads (Life Technologies) using previously described methods (Bosch et al., 2016). The proteins eluted from the beads following the IPs were fractionated on SDS-PAGE (12%) gels, transferred to polyvinylidene fluoride (PVDF) membranes (Biorad), and probed for iFGF14 expression using a rabbit polyclonal α-iFGF14 antiserum (Rb-α-iFGF14; 1:1,000; custom antibody made by Covance Antibody Services), validated as described previously (Bosch et al., 2016). The lysates prepared from frozen hippocampal CA1, primary visual cortical, and cerebellar tissues were also fractionated, transferred, and probed for iFGF12 and iFGF13 expression using validated (Angsutararux et al., 2023; Nerbonne, unpublished) rabbit polyclonal anti-iFGF12 (Rb-α-iFGF12, 1:500; catalog # HPA071557; Sigma-Aldrich) and anti-FGF13 (Rb-α-iFGF13, 1:1,000; generous gift of Geoffrey Pitt, Weill Medical College, Cornell University, New York, NY, USA) antibodies. Additional lysates were prepared from frozen hippocampal and cerebellar tissues from adult (8–10 wk) WT (N = 4) and Fgf14^−/− (N = 4) animals, fractionated, transferred, and probed for Navβ1, Navβ2, Navβ3, and Navβ4 expression using validated (Wong et al., 2005; Bosch et al., 2016) rabbit polyclonal anti-Navβ1, anti-Navβ2, anti-Navβ3 or anti-Navβ3 (Rb-α-Navβ1, 1:1,000; Rb-α-Navβ2, 1:1,000; Rb-α-Navβ3, 1:1,000; or RB-α-Navβ4, 1:1,000; generous gift of N. Nukina, Riken Science Institute, Wako, Japan) antibodies. To ensure equal loading of all of the lanes of each of the gels in these experiments, membranes were also probed with either a mouse monoclonal antibody targeting glyceraldehyde-3-phosphate dehydrogenase (mAb-α-GAPDH; 1:8,000; catalog # MA5-15738; Thermo Fisher Scientific [Invitrogen]) or a mouse monoclonal antibody targeting α-tubulin (mAb-α-Tubulin; 1:10,000; catalog # ab7291; Abcam). The loading control western blot data are presented alongside the iFGF12/13 and Navβ1/2/3/4 blots, and full western blots are provided in the source data figures.

Fig. S1, S2, and S3 present data demonstrating the robust expression of the Fgf14 transcript and the iFGF14 protein in WT adult mouse primary visual cortex and the loss of both (Fgf14/iFGF14) in Fgf14^−/− cortex (Fig. S1), the expression of the Fgf12 (Fgf12B) and Fgf13 (Fgf13A) transcripts and the iFGF12 and iFGF13 proteins in layer 5 of the primary visual cortex of adult WT and Fgf14^−/− animals (Fig. S2), and the functional consequences of the loss of Fgf14/iFGF14 on the evoked repetitive firing properties of primary visual cortical layer 5 pyramidal neurons (Fig. S3). Quantitative RT-PCR data presented in Fig. S4 demonstrate that expression of the Scn1a, Scn2a, and Scn8a transcripts are similar in the CA1 region of adult WT and Fgf14^−/− hippocampus, whereas expression of Scn2a and Scn8a (but not Scn1a) are lower in Fgf14^−/− than in WT, cerebellum. Table S1 lists primers for quantitative real-time PCR.

In addition to expression in the cerebellum, previous in situ hybridization studies identified Fgf14/FGF14 transcripts in adult mouse and human hippocampus and cerebral cortex, as well as in the amygdala, striatum, and thalamus (Wang et al., 2000, 2002). In the initial experiments, we directly examined Fgf14 transcript and iFGF14 protein expression in the CA1 region of the hippocampus of adult WT mice. As illustrated in Fig. 1, quantitative RT-PCR analyses confirmed the expression of Fgf14 transcripts in the CA1 region of the hippocampus (Fig. 1 A) of WT animals and the complete loss of Fgf14 in Fgf14^−/− hippocampus (Fig. 1 A). Further experiments revealed, consistent with previous reports (Wang et al., 2002), that Fgf14B is the predominant splice variant in WT hippocampus and Fgf14A is barely detectable (Fig. 1 A). Western blot analyses, using a validated polyclonal α-iFGF14 antiserum (Bosch et al., 2015, 2016), confirmed robust expression of the iFGF14 protein in the CA1 region of the hippocampus of adult WT mice and the complete loss of the iFGF14 protein in CA1 hippocampal tissue from Fgf14^−/− animals (Fig. 1 B). Also consistent with previous findings (Wang et al., 2002), Fgf14B (Fig. S1 B) and the iFGF14 protein (Fig. S1 B) are abundant in the adult mouse cortex.

Additional RT-PCR and Western blot analyses revealed robust expression of Fgf12 (specifically Fgf12B) and Fgf13 (specifically Fgf13A) transcripts and of the iFGF12 and iFGF13 proteins in the CA1 region of the hippocampus (Fig. 2) of adult WT mice. In addition, and in spite of the complete loss of Fgf14 transcripts (Fig. 1 A) and the iFGF14 protein (Fig. 1 B) in the CA1 region of adult Fgf14^−/− hippocampus, no remodeling (upregulation or downregulation) of the Fgf12/Fgf13 transcripts (Fig. 2 A) and/or the iFGF12/iFGF13 proteins (Fig. 2 B) is evident. Similar results were obtained when WT and Fgf14^−/− cortices were compared (Fig. S2).

Immunohistochemical experiments with a monoclonal α-iFGF14 antibody (Xiao et al., 2013; Bosch et al., 2015) revealed α-iFGF14 labeling at the AIS of WT hippocampal CA1 pyramidal neurons (Fig. 1 C) identified using an α-Ankyrin G antibody (Bosch et al., 2015). Similar patterns of α-FGF14 and α-Ankyrin G labeling were observed in pyramidal neurons in WT adult mouse primary visual cortex (Fig. S1 C). Confirming robust expression of iFGF14 in adult mouse pyramidal neurons, subsequent experiments were focused on testing the hypothesis that, similar to our previous findings in studies of cerebellar Purkinje neurons (Bosch et al., 2015), loss of iFGF14 attenuates the intrinsic excitability of mature hippocampal pyramidal neurons.

Previous studies have demonstrated that the loss of iFGF14 is linked to cognitive dysfunction. Individuals with SCA27A, for example, have progressive deficits in learning and memory (Brusse et al., 2006; Misceo et al., 2009) and Fgf14^−/− animals perform poorly in the Morris water maze (Wozniak et al., 2007), suggesting impaired hippocampal functioning. To determine directly if the loss of iFGF14 impairs the excitability of adult CA1 hippocampal pyramidal neurons, whole-cell current-clamp recordings were obtained (as described in Materials and methods) from visually identified CA1 pyramidal neurons in acute slices prepared from adult (6–8 wk) WT and Fgf14^−/− mice. As illustrated in the representative recordings shown in Fig. 3 A (left panel), WT CA1 pyramidal neurons fired repetitively in response to prolonged (500 ms) depolarizing current injections, and the rates of repetitive firing increased as a function of the amplitude of the injected current. Similar to WT cells, prolonged depolarizing current injections also elicited repetitive firing in Fgf14^−/− CA1 pyramidal neurons (Fig. 3 A, right panel), and firing rates also tracked the amplitudes of the injected currents. As is also evident in the representative records shown in Fig. 3 A, repetitive firing rates were higher in the Fgf14^−/−, compared with the WT, CA1 pyramidal neuron for all injected current amplitudes. Similar results were obtained in recordings from additional WT and Fgf14^−/− CA1 pyramidal neurons, and the mean ± SEM input–output curves (numbers of spikes evoked plotted versus injected current amplitudes) in WT (n = 13; N = 3) and Fgf14^−/− (n = 14; N = 3) hippocampal CA1 pyramidal neurons (Fig. 3 B) are distinct (P < 0.001; two-way ANOVA). Evoked repetitive firing rates are also higher in Fgf14^−/−, than in WT, cortical pyramidal neurons (Fig. S3).

Although spike frequency adaptation is apparent in the representative recordings from both WT and Fgf14^−/− hippocampal CA1 pyramidal neurons (Fig. 3 A), quantification of the ratio of the average firing frequency at the end (last 100 ms) and at the beginning (first 100 ms) for current injections that evoked the same mean number of action potentials (400 pA in Fgf14^−/− cells and 460 pA in WT cells, see: Fig. 2 B) revealed similar mean ± SEM firing rate ratios of 0.44 ± 0.07 (n = 14; N = 3) and 0.52 ± 0.05 (n = 13; N = 3) in Fgf14^−/− and WT neurons, respectively. Loss of iFGF14, therefore, results in markedly increased repetitive firing rates but does not measurably affect spike frequency adaptation in mature hippocampal CA1 pyramidal neurons (Fig. 3).

Additional experiments and analyses were completed to quantify the effects of the loss of iFGF14 on the waveforms of individual action potentials in hippocampal CA1 pyramidal neurons. As illustrated in Fig. 4 A, the waveforms of single action potentials evoked in Fgf14^−/− and WT CA1 pyramidal neurons in response to brief (2.5 ms) depolarizing current injections (1.5 nA) are similar (Fig. 4 A). Analyses of the voltage records revealed, however, that there are differences in the voltage thresholds (V_thr) for action potential generation (Fig. 4 B) and in action potential durations (APD₅₀) at 50% repolarization (Fig. 4 C) in Fgf14^−/− and WT CA1 pyramidal neurons. The mean ± SEM V_thr for action potential generation determined in Fgf14^−/− CA1 pyramidal neurons, for example, is depolarized (P = 0.015) relative to the value in WT CA1 cells (Fig. 4 B and Table 1). In addition, the mean ± SEM APD₅₀ measured is smaller (P < 0.001) in Fgf14^−/−, than in WT, CA1 pyramidal neurons (Fig. 4 C and Table 1).

Although the membrane voltage (V_m) trajectories during individual action potentials are similar in WT and Fgf14^−/− CA1 pyramidal neurons (Fig. 4 D, top traces), there are marked differences in the first (dV/dt) (Fig. 4 D, middle traces) and second (dV²/dt) (Fig. 4 D, bottom traces) derivative plots of action potential trajectories versus time. In the representative WT CA1 pyramidal neuron presented in Fig. 4 D, for example, there are two distinct components in the rising phase of the dV/dt plot (Fig. 4 D, black middle trace), and these two inflection points, which correspond to action potentials generated in the axon and the cell soma (Palmer and Stuart, 2006; Meeks and Mennerick, 2007), are clearly resolved in the dV²/dt plot (Fig. 4 D, black lower trace). Although also evident in the dV/dt (Fig. 4 D, red middle trace) and dV²/dt (Fig. 4 D, red lower trace) plots generated from the action potential recorded from a representative Fgf14^−/− CA1 pyramidal neuron (Fig. 4 A), the time difference between the two inflection points is smaller in the Fgf14^−/− than in the WT cells (Fig. 4 D). Similar results, obtained in recordings from additional WT and Fgf14^−/− CA1 pyramidal neurons, are plotted in Fig. 4 E. As illustrated, the mean ± SEM time interval between the two inflection points is much smaller (P = 0.006) in Fgf14^−/− than in WT hippocampal CA1 pyramidal neurons (Fig. 4 E and Table 1). Phase plots (Fig. 4 F) also revealed that peak dV/dt values are higher (P = 0.007) in Fgf14^−/−, compared with WT, CA1 pyramidal neurons (Fig. 4 G and Table 1).

Previous studies in cerebellar Purkinje neurons have shown that the loss of iFGF14 results in a marked hyperpolarizing shift in the voltage-dependence of steady-state (closed state) inactivation of the transient component of the Nav current, I_NaT (Bosch et al., 2015). To explore the contribution of iFGF14-mediated effects on the voltage-dependent properties of I_NaT in hippocampal CA1 pyramidal neurons to the observed differences in the waveforms of individual action potentials (Fig. 4) and the repetitive firing rates (Fig. 3) of Fgf14^−/−, compared with WT cells, whole-cell Nav currents were recorded from cells in acute slices prepared from young adult (4–6 wk) animals. Voltage-clamp experiments were performed using a previously described experimental strategy (Milescu et al., 2010; Bosch et al., 2015), optimized to minimize space clamp errors and to allow reliable measurements of the voltage dependences of the inactivation of I_NaT (see Materials and methods). To determine the voltage dependences of closed state inactivation of I_NaT in WT and Fgf14^−/− CA1 pyramidal neurons, the peak of I_NaT amplitudes, evoked at 0 mV from various conditioning voltages (Fig. 5 A), was measured and normalized to the amplitude of the current evoked from the most hyperpolarized membrane potential (−130 mV) in the same cell. Mean (±SEM) normalized peak I_NaT amplitudes were then plotted as a function of the conditioning voltage and fitted (Fig. 5 B). The closed state inactivation data for I_NaT in Fgf14^−/− (n = 8; N = 2) and WT (n = 6; N = 2) CA1 pyramidal neurons were well described by single Boltzmann functions with similar mean ± SEM V_1/2 values of −68.2 ± 0.6 and −69.0 ± 1.0 mV in WT and Fgf14^−/−, respectively, CA1 pyramidal neurons. The slope factors, 9.0 ± 0.6 and 13.0 ± 1.0, determined in WT and Fgf14^−/−, CA1 neurons, respectively, in contrast, are (P < 0.001) different. Specifically, the slope is steeper in WT compared with Fgf14^−/−, CA1 pyramidal neurons.

It is important to note that the protocol used here to determine the voltage-dependences of closed-state inactivation (Fig. 5) of I_NaT in adult CA1 pyramidal neurons in situ required a voltage-clamp paradigm in which brief (3 ms) voltage steps were used to recover Nav channels from inactivation so that currents through these channels could be evoked (and measured) during the subsequent voltage step (see Materials and methods). Using this paradigm, the evoked/measured Nav currents arise predominantly from channels (near the patch electrode) localized to the somatic and perisomatic spaces, which may not extend through the entire AIS and likely do not, therefore, reflect the full complement of Nav channels present that could/would be recovered with longer (and more negative) voltage steps.

Immunohistochemical analyses of the AIS of mature cerebellar Purkinje neurons in situ revealed no measurable effects of the loss of iFGF14 on α-Ankyrin G and/or α-Nav α subunit labeling intensity or distribution (Bosch et al., 2015). It has also previously been reported, however, that the acute knockdown of Fgf14 in dissociated hippocampal neurons in culture disrupts the AIS localization of Nav channels (Pablo et al., 2016). To determine if these seemingly disparate observations reflected differences in neuronal cell types or experimental protocols, we completed experiments to determine directly if iFGF14 affects the AIS localization/distribution of Ankyrin G and/or Nav1.6 in mature CA1 pyramidal neurons in situ. In these experiments, hippocampal sections prepared from adult WT and Fgf14^−/− mice were stained with α-Ankyrin G and α-Nav1.6 specific antibodies and labeling intensities were determined from analyses of confocal line scans along the AIS of individual WT and Fgf14^−/− CA1 hippocampal pyramidal neurons (see Materials and methods).

As illustrated in Fig. 6 B, these immunohistochemical experiments revealed robust α-Ankyrin G and α-Nav1.6 immunostaining at the AIS of adult WT and Fgf14^−/− hippocampal CA1 pyramidal neurons. Quantification of AIS images such as those shown in Fig. 6 B reveals that the mean ± SEM intensity of α-Ankyrin G (Fig. 6 C, top panel) labeling was similar along the AIS of WT and Fgf14^−/− neurons, whereas the mean ± SEM α-Nav1.6 (Fig. 6 C, lower panel) labeling intensity was somewhat higher along the AIS of Fgf14^−/− compared with WT, CA1 hippocampal pyramidal neurons. In addition, analysis of the α-Nav1.6 (Fig. 6 D, lower panel) immunolabeling intensities, measured along individual AIS, reveals higher integrated α-Nav1.6 (P < 0.001) labeling intensities along the AIS of Fgf14^−/− compared with WT, hippocampal CA1 pyramidal neurons, whereas integrated α-Ankyrin G labeling intensities along the AIS of Fgf14^−/− and WT hippocampal CA1 pyramidal neurons are similar. As the expression of Scn8a, which encodes Nav1.6, is not measurably affected by the loss of Fgf14 (Fig. S4 A), the increase in α-Nav1.6 labeling must reflect enhanced AIS localization of Nav1.6 and/or an increase in total Nav1.6 protein expression in Fgf14^−/− compared with WT, CA1 pyramidal neurons. In addition, further analyses revealed that the mean ± SEM normalized distributions of both α-Ankyrin G (Fig. 6 E, top panel) and α-Nav1.6 (Fig. 6 E, bottom panel) immunofluorescence labeling intensities along the AIS of Fgf14^−/− and WT hippocampal CA1 pyramidal neurons are indistinguishable. The higher integrated intensity of α-Nav1.6 labeling along the AIS, however, likely contributes to the increased excitability (see: Fig. 3) and higher peak dV/dt (see: Fig. 4 G) observed in Fgf14^−/− compared with WT, CA1 hippocampal pyramidal neurons (see Discussion).

The results presented above demonstrate that the effects of the targeted deletion of Fgf14 on the excitability of hippocampal CA1 pyramidal neurons are distinct from previous findings in Fgf14^−/− cerebellar Purkinje neurons (Shakkottai et al., 2009; Bosch et al., 2015). Repetitive firing rates are higher (Fig. 3) in adult hippocampal CA1 pyramidal neurons lacking Fgf14, whereas repetitive firing rates were lower in adult Fgf14^−/− than in WT, cerebellar Purkinje neurons (Shakkottai et al., 2009; Bosch et al., 2015). These differences reflect distinct effects of Fgf14 deletion on Nav currents: the loss of iFGF14 in cerebellar Purkinje neurons results in a hyperpolarizing shift in the voltage dependence of Nav current inactivation (Bosch et al., 2015), whereas the loss of iFGF14 in hippocampal CA1 pyramidal neurons affect the voltage dependence of Nav current activation. As Nav1.6 is the predominant Na α subunit responsible for the initiation of action potentials at the AIS of hippocampal CA1 pyramidal (Royeck et al., 2008) and cerebellar Purkinje (Levin et al., 2006) neurons, the distinct functional effects of the loss of Fgf14 likely reflect differences in one or more of the other molecular components, that is, other iFGFs and/or Navβ subunits of native Nav1.6 channel complexes in these two cell types. Quantitative RT-PCR analysis revealed that the expression levels of the Fgf11, Fgf12A, and Fgf12B transcripts are higher in the WT cerebellum (Fig. 7 A) than in WT hippocampus (Fig. 7 B), whereas Fgf13A expression is higher in WT hippocampus (Fig. 7 B) than in WT cerebellum (Fig. 7 A). In addition and in contrast with the hippocampal CA1 data (Fig. 7 B), the expression levels of both Fgf12A and Fgf12B are much lower (P < 0.0001) in Fgf14^−/− compared with WT, cerebellum (Fig. 7 A). Western blot analyses, however, revealed that in contrast with the lower expression levels of the Fgf12A and Fgf12B transcripts in Fgf14^−/− cerebellum (Fig. 7 A), iFGF12 protein expression is similar in WT and Fgf14^−/− cerebellum (Fig. 7 C), findings similar to those obtained in the hippocampus (Fig. 2 B). Consistent with the lower expression levels of Fgf13 transcripts (Fig. 7 A), iFGF13 protein expression is also lower and more variable in the cerebellum (Fig. 7 D) compared with the hippocampus (Fig. 2 C) (see Discussion).

In addition to the iFGFs, Nav channel accessory β subunits are also components of native Nav channel macromolecular complexes (O’Malley and Isom, 2015; Bosch et al., 2016). There are four Navβ subunits, Navβ1–Navβ4, all of which are single transmembrane, immunoglobulin domain-containing proteins that bind to Nav α subunits and modulate the trafficking and/or the biophysical properties of assembled Nav channels, including Nav1.6-encoded channels (Zybura et al., 2021). Quantitative RT-PCR analyses revealed robust expression of the Scn1B, Scn2B, and Scn4B transcripts in WT and Fgf14^−/− cerebellum, whereas Scn3B is barely detectable (Fig. 8 A) and all four transcripts, i.e., Scn1B, Scn2B, Scn3B, and Scn4B, are detected in the hippocampus (Fig. 8 B). These analyses also revealed that the expression levels of the Scn1B and Scn2B transcripts are lower in Fgf14^−/− than in WT, cerebellum, whereas the expression levels of the Scn1B, Scn2B, Scn3B, and Scn4B transcripts are not significantly different in the CA1 region of the hippocampus of WT and Fgf14^−/− mice. Western blot analyses revealed that the Navβ3 protein (Fig. 8 C3), like the Scn3b transcript (Fig. 8 A), is barely detectable in WT and Fgf14^−/− cerebellum whereas Navβ1 (Fig. 8 C1), Navβ2 (Fig. 8 C2), and Navβ4 (Fig. 8 C4) are robustly expressed and at similar levels in WT and Fgf14^−/− cerebellum. In contrast and in spite of the low expression level of the Scn3b transcript (Fig. 8 B), the Navβ3 protein is readily detected in the hippocampus (Fig. 8 D3), whereas the Navβ1 (Fig. 8 D1) and Navβ4 (Fig. 8 D4) proteins are less abundant and more variable in the hippocampus than in the cerebellum (Fig. 8, C1 and C4) (see Discussion).

In contrast with previous findings in adult cerebellar Purkinje neurons (Bosch et al., 2015), the results here demonstrate that excitability and repetitive firing rates are increased, not decreased, in adult Fgf14^−/− compared with WT, hippocampal CA1 pyramidal neurons. Also distinct from cerebellar Purkinje neurons (Bosch et al., 2015), the loss of iFGF14 does not measurably affect the voltage dependence of closed-state inactivation of I_NaT in adult hippocampal CA1 pyramidal neurons. Similar to previous findings demonstrating preserved AIS localization of Nav α subunits in Fgf14^−/− cerebellar Purkinje neurons (Bosch et al., 2015), however, the loss of iFGF14 does not disrupt the localization or measurably alter the distribution of Nav1.6 channels along the AIS of mature hippocampal CA1 pyramidal neurons. The mean integrated intensity of α-Nav1.6 immunolabeling, however, was significantly higher along the AIS of Fgf14^−/− compared with WT, hippocampal CA1 pyramidal neurons, consistent with the increased excitability (Figs. 3 and 4) of these cells.

It has previously been reported that the acute knockdown of Fgf14 in dissociated neonatal (rat) hippocampal pyramidal neurons resulted in reduced Nav current densities and reduced AIS localization of Nav α subunits in spite of preserved AIS architecture, as revealed by the expression/distribution of Ankyrin G (Pablo et al., 2016). The results of the immunohistochemical analyses conducted here, however, revealed no significant differences in the normalized distribution of α-Nav1.6 α subunit (or Ankyrin G) labeling of the AIS of Fgf14^−/− compared with WT, adult (mouse) hippocampal CA1 pyramidal neurons and, in addition, that the mean integrated intensity of α-Nav1.6 immunolabeling was higher along the AIS of Fgf14^−/− compared with WT, hippocampal CA1 pyramidal neurons. It is possible that these conflicting results reflect iFGF14-dependent differences in the ages of the neurons in the previous (Pablo et al., 2016), versus the present, study and/or the fact that some remodeling has occurred in Fgf14^−/− adult hippocampal CA1 pyramidal neurons studied here, as these cells lacked iFGF14 throughout development. It is relevant to note here, however, that we have previously shown that the electrophysiological properties of adult Fgf14^−/− cerebellar Purkinje neurons are recapitulated with acute shRNA-mediated knockdown of Fgf14 in adult neurons and, in addition, that virus-mediated expression of iFGF14 restores spontaneous and evoked repetitive firing in adult Fgf14^−/− cerebellar Purkinje neurons and improves motor coordination and balance in adult Fgf14^−/− animals (Bosch et al., 2015). There may, however, be effects of the developmental and/or long-term loss of iFGF14 in hippocampal CA1 pyramidal neurons that are not observed in cerebellar Purkinje neurons. It is also possible that these disparate results reflect a species-specific difference in the functioning of iFGF14 as mouse hippocampal pyramidal neurons were studied here, and Pablo et al. (2016) reported the results of experiments conducted on rat hippocampal neurons. A role for cell dissociation in the loss of Nav1.6 AIS localization in the previous studies also seems possible. Additional experiments designed to compare directly AIS architecture and the impact of the loss of iFGF14 on the AIS localization of Nav channels in mouse and rat hippocampal neurons in dissociated cell culture and in adult mouse and rat hippocampal neurons in situ will be needed to distinguish among these various possibilities.

The observations here that evoked repetitive firing rates are higher in Fgf14^−/− compared with WT, adult hippocampal CA1 pyramidal neurons and that the voltage-dependence of Nav current closed state inactivation in CA1 neurons is not affected by the loss of Fgf14 contrast markedly with our previous findings showing that spontaneous repetitive firing rates were significantly reduced in Fgf14^−/− compared with WT, cerebellar Purkinje neurons (Shakkottai et al., 2009; Bosch et al., 2015), and this effect was mediated by a hyperpolarizing shift in the voltage dependence of Nav current inactivation with the loss of iFGF14 (Bosch et al., 2015). The motivation for undertaking studies to define the role of iFGF14 in hippocampal pyramidal neurons was that cognitive deficits, like those observed in individuals with SCA27A (Brusse et al., 2006; Misceo et al., 2009), are also evident in Fgf14^−/− animals, suggesting a role for changes in excitability in neurons in additional brain regions, such as the hippocampus, involved in learning and memory (Wang et al., 2002; Xiao et al., 2007; Wozniak et al., 2007). Also, we selected hippocampal pyramidal neurons here for comparison with cerebellar Purkinje neurons because pyramidal cells are functionally similar in that they are like Purkinje cells (Billard et al., 1993; Gauck and Jaeger, 2000; Ito, 2001) and projection neurons. These cell types are, however, distinct in other ways in that cerebellar Purkinje neurons, which express the neurotransmitter GABA (gamma-aminobutyric acid), are inhibitory (Ito et al., 1964; Ito, 2001), whereas hippocampal pyramidal neurons, which express the neurotransmitter glutamate, are excitatory (Storm-Mathisen, 1981; Jones, 1986). It seems unlikely, however, that the distinct functional effects observed with the loss of iFGF14 in hippocampal pyramidal neurons, compared with cerebellar Purkinje neurons, simply reflect these differences in neurotransmitter phenotypes as loss of Fgf14 in cerebellar granule neurons, which are also glutamate-expressing excitatory central neurons (Eccles et al., 1966; Ito, 2001), resulted in a marked decrease in evoked repetitive firing rates, attributed to a hyperpolarizing shift in the voltage-dependence of Nav current inactivation (Goldfarb et al., 2007), i.e., cellular phenotypes that are remarkably similar to those observed in cerebellar Purkinje neurons (Shakkottai et al., 2009; Bosch et al., 2015).

Studies in heterologous cells have demonstrated diverse effects of iFGF co-expression on Nav current amplitudes, as well as on the voltage dependences of Nav current inactivation (Liu et al., 2001, 2003; Wittmack et al., 2004; Lou et al., 2005; Laezza et al., 2009; Wang et al., 2011b). Comparison of the results obtained in these various studies suggest that the diverse functional effects observed reflect the specific Nav α subunit and the specific iFGF protein that are co-expressed, as well as the cellular expression environment, that is, the effects of an individual iFGF on the expression levels and/or properties of Nav currents encoded by a single Nav α subunit have been shown to be different in different cell types (Wittmack et al., 2004; Lou et al., 2005; Laezza et al., 2007, 2009; Wang et al., 2011b). Cell-type-specific differences in the expression of the Nav1.1 and Nav 1.2 α subunits in cerebellar Purkinje (Xiao et al., 2013) and hippocampal CA1 pyramidal (Yamagata et al., 2023) neurons, respectively, could, therefore, contribute to the distinct functional effects of the loss of iFGF14 in hippocampal CA1 pyramidal neurons reported here, compared with previous findings in cerebellar Purkinje neurons (Shakkottai et al., 2009; Bosch et al., 2015). Nav1.6, however, is the predominant Na α subunit responsible for the initiation of action potentials at the AIS of central neurons, including hippocampal pyramidal neurons and cerebellar Purkinje (and granule) neurons (Levin et al., 2006; Royeck et al., 2008; Hu et al., 2009; Katz et al., 2018; Leterrier, 2018; Zybura et al., 2021). Given the critical role of Nav1.6-encoded AIS channels and that we examined the effects of iFGF14 and, more specifically, iFGF14B, it seems reasonable to conclude that the distinct functional effects of the targeted deletion of Fgf14 observed reflect the different cellular expression environments, that is, the various accessory proteins and/or post-translational modifications of the components of functional Nav channel complexes in adult mouse cerebellar Purkinje versus hippocampal CA1 pyramidal neurons.

The molecular analyses here reveal differences in the expression levels of the Fgf11, Fgf12A, Fgf12B, and Fgf13A transcripts in the WT cerebellum and hippocampus, as well as selective effects of the targeted deletion of Fgf14 on Fgf12A and Fgf12B expression in the cerebellum. iFGF12 protein expression levels, however, were similar in WT and Fgf14^−/− cerebellum. In contrast, but consistent with the lower expression levels of Fgf13 transcripts, iFGF13 protein expression was lower and more variable in the cerebellum, compared with the hippocampus, suggesting a possible role for iFGF13 in mediating the differential effects of the loss of iFGF14. Differences in the expression of the Scn1B–Scn4B transcripts and the Navβ1–Navβ4 proteins were also evident in WT cerebellum and hippocampus. The Scn3b transcript and the Navβ3 protein, for example, were barely detectable in the cerebellum, whereas the Navβ3 protein was readily detected in the hippocampus. In addition, the Navβ1 and Navβ4 proteins are less abundant and more variable in the hippocampus than in the cerebellum. Differences in the relative expression of the Navβ1, Navβ3, and Navβ4 proteins could certainly contribute to the differential effects of the loss of iFGF14 in adult hippocampal pyramidal compared with adult cerebellar Purkinje neurons. Additional experiments aimed at defining the functional roles of each of the other iFGF proteins and of the Navβ1–Navβ4 subunits (O’Malley and Isom, 2015; Ransdell et al., 2017), as well, perhaps, as other Nav channel accessory subunits, such as calmodulin (Ben-Johny et al., 2015), in controlling the voltage-dependent properties of Nav1.6-encoded currents and/or in modulating the effects of iFGF14B on the voltage-dependent properties of Nav1.6-encoded currents will be of considerable interest.

The results presented here demonstrate a distinct physiological role for iFGF14 in the regulation of the intrinsic excitability of adult hippocampal CA1 pyramidal neurons (Figs. 3 and 4). It is unclear, however, whether iFGF14 should be considered an obligatory accessory subunit of Nav channels in hippocampal CA1 pyramidal neurons and/or if iFGF14-Nav α subunit interactions are dynamically regulated through interactions with other iFGF and/or Navβ subunit proteins or, possibly, via posttranslational modifications of the pore-forming Nav1.6 protein and/or of one or more Nav channel accessory subunits (Shao et al., 2009; Onwuli and Beltran-Alvarez, 2016; Pei et al., 2018; Zybura et al., 2021; Marosi et al., 2022). In addition, it seems reasonable to suggest that, if there are intracellular second messenger pathways that regulate or modulate the interaction(s) between iFGF14 and the components of native neuronal Nav1.6-encoded channels, these could also affect the membrane expression/density of Nav channels and, therefore, impact the excitability and the functioning of hippocampal CA1 pyramidal neurons. Delineation of cell signaling pathways that modulate (attenuate or enhance) iFGF14-Nav channel interactions through kinase-mediated and other intracellular second messenger pathways could also provide new strategies to target iFGF-Nav channel interactions and indirectly impact neuronal excitability, potentially in a cell type–specific manner.

Data are available in the published article and in the accompanying Online supplemental material and source materials.

Olaf S. Andersen served as editor.

The authors thank Dr. Geoffrey Pitt for providing the polyclonal α-iFGF13 antibody and Drs. Nobuyuki Nukina and Haruko Miyazaki for providing the polyclonal anti-Navβ1-4 antibodies used in the western blot experiments presented.

The authors acknowledge the financial support provided by the National Institute of Neurological Disorders and Stroke of the National Institutes of Health (National Institutes of Health [NIH]) (Grant #R01 NS065761 to D.M. Ornitz and J.M. Nerbonne; Individual Postdoctoral Fellowship Awards #F32 NS090765 to J.L. Ransdell and #F32 NS065582 to Y. Carrasquillo). M.K. Bosch was supported by NIH Institutional Training Grants (T32 GM007200 and T32 HL007275). This work was supported by the Hope Center Neuroimaging and Vector Cores at Washington University Medical School. The monoclonal antibodies α-iFGF14, α-Ankyrin G, and α-Nav1.6 were developed by the University of California at Davis/NIH NeuroMab facility, supported by NIH grant U24NS050606, and maintained by the University of California at Davis, and were purchased from Antibodies, Incorporated (Davis CA).

Author contributions: J.L. Ransdell: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Validation, Visualization, Writing - original draft, Writing - review & editing, Y. Carrasquillo: Conceptualization, Formal analysis, Funding acquisition, Investigation, Writing - review & editing, M.K. Bosch: Investigation, Validation, Visualization, R.L. Mellor: Investigation, D.M. Ornitz: Conceptualization, Writing - review & editing, J.M. Nerbonne: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing - review & editing.