BGP-15 prevents the death of neurons in a mouse model of familial dysautonomia

Edited by Tomas G. M. Hokfelt, Karolinska Institutet, Stockholm, Sweden, and approved April 5, 2017 (received for review December 9, 2016)
April 24, 2017
114 (19) 5035-5040


Familial dysautonomia (FD) is a fatal genetic disorder that disrupts development of the peripheral nervous system (PNS) and causes progressive degeneration of the PNS and retina, ultimately leading to blindness. The underlying cellular mechanisms responsible for neuronal death in FD have been elusive. Using a mouse model that recapitulates impaired PNS development, we report here that developing peripheral neurons die in FD as the result of disruptions in mitochondrial and actin function. Cell death can be prevented in vivo by the small molecule BGP-15. Given that disrupted mitochondrial function appears to be a hallmark of neurodegenerative diseases, future studies on the efficacy of BGP-15 for mitigating neuronal loss in other disorders is warranted.


Hereditary sensory and autonomic neuropathy type III, or familial dysautonomia [FD; Online Mendelian Inheritance in Man (OMIM) 223900], affects the development and long-term viability of neurons in the peripheral nervous system (PNS) and retina. FD is caused by a point mutation in the gene IKBKAP/ELP1 that results in a tissue-specific reduction of the IKAP/ELP1 protein, a subunit of the Elongator complex. Hallmarks of the disease include vasomotor and cardiovascular instability and diminished pain and temperature sensation caused by reductions in sensory and autonomic neurons. It has been suggested but not demonstrated that mitochondrial function may be abnormal in FD. We previously generated an Ikbkap/Elp1 conditional-knockout mouse model that recapitulates the selective death of sensory (dorsal root ganglia) and autonomic neurons observed in FD. We now show that in these mice neuronal mitochondria have abnormal membrane potentials, produce elevated levels of reactive oxygen species, are fragmented, and do not aggregate normally at axonal branch points. The small hydroxylamine compound BGP-15 improved mitochondrial function, protecting neurons from dying in vitro and in vivo, and promoted cardiac innervation in vivo. Given that impairment of mitochondrial function is a common pathological component of neurodegenerative diseases such as amyotrophic lateral sclerosis and Alzheimer’s, Parkinson’s, and Huntington’s diseases, our findings identify a therapeutic approach that may have efficacy in multiple degenerative conditions.
The autonomic nervous system is essential for homeostasis, and its disruption in familial dysautonomia (FD) can have fatal consequences resulting from cardiovascular instability, respiratory dysfunction, and/or sudden death during sleep (13). In addition to developmental decreases in the number of sensory and autonomic neurons, FD patients undergo a progressive loss of peripheral neurons and retinal ganglion cells. The latter loss may ultimately lead to blindness (4, 5). More than 98% of FD cases result from a single base substitution (IVS20+6T > C) in the IKBKAP/ELP1 gene (3, 6). This mutation is carried by 1 in 27 to 1 in 32 Ashkenazi Jews (3, 6). The protein encoded by the IKBKAP/ELP1 gene, IKAP/ELP1, is a scaffolding protein for the six-subunit Elongator complex (ELP1–ELP6), which modifies tRNAs during translation (7). Why reduction in the IKAP/ELP1 protein results in neuronal death is unknown, and we have sought to elucidate the cellular and molecular mechanisms that cause progressive neurodegeneration in FD to help identify treatments that improve neuronal function and viability.
Accumulating evidence indicates that cells from FD patients and from mouse models with deletions in the Elongator subunits Ikbkap/Elp1 or Elp3 experience intracellular stress (4, 5, 810) resulting from the direct and/or indirect consequences of impaired translation (7, 11). The C terminus of IKAP/ELP1 has been shown to bind c-Jun N-terminal kinase (JNK) and to regulate JNK cytosolic stress signaling (12). Loss of Elp3 in mouse cortical neurons triggers endoplasmic reticulum stress (9), and Elp3 deletion in yeast causes disruptions in mitochondrial function (10). ELP3 has been demonstrated to localize in mitochondria in HeLa cells, mouse brain, and Toxoplasma (13, 14).
Recent studies suggest that the loss of Ikbkap/Elp1 can cause mitochondrial dysfunction. For example, investigations of retinal ganglion cells (RGCs) in mouse models of FD and in human patients indicate that metabolically active, temporal RGCs are more compromised in mutant (FD) retinae than are the less active nasal RGCs. This pattern is reminiscent of the pattern of RGC loss in optic neuropathies caused by disruption in mitochondrial genes (4, 5, 15). Furthermore, a recent study of FD patients demonstrates that they experience high rates of rhabdomyolysis (16). Finally, we previously have demonstrated that in the mouse model of FD (Wnt1-Cre;Ikbkap−/−) used in the present study, TrkA+ pain- and temperature-receptive neurons in the dorsal root ganglia (DRG) die in vivo as a result of p53, caspase-3–mediated apoptosis (8).
BGP-15 is a hydroxylamine derivative that has been shown to exert cyto- and neuroprotective effects in mammalian models of injury, stress, and disease (1730). These improvements in cellular function have been correlated with the activation of several intracellular pathways. For example, the heat-shock response is enhanced by increased levels of the molecular chaperone HSP72 (21, 24, 26, 28, 29), possibly mediated by Rac-1 signaling (31, 32). BGP-15 also decreases phospho-JNK (pJNK) and p38 stress signaling (21, 23) and increases AKT and IGFR1 protective signaling (23, 25). Additionally, in models of stress and injury, BGP-15 has been shown to decrease cell stress by restoring normal mitochondrial function. These studies have identified multiple mechanisms that may mediate BGP-15’s positive effects on mitochondria. These include reducing NAD+ depletion and overactive PARP, improving antioxidant status, decreasing reactive oxidant species (ROS), stabilizing mitochondrial membrane potential, increasing mitochondrial content, and blocking AIF translocation to the nucleus (1719, 22, 23, 2628, 30). Given the beneficial effects of BGP-15 in reducing intracellular stress via multiple potential pathways, including repairing mitochondrial function, the goals of this study were (i) to test the hypothesis that Ikbkap−/− neurons die because of impaired mitochondrial function, and, having demonstrated this causation, (ii) to investigate the therapeutic potential of BGP-15 in restoring mitochondrial function and neuronal survival in a mouse model of FD.


Mitochondrial Function Is Disrupted in Ikbkap−/− Neurons and Is Repaired by BGP-15.

DRG were dissected from and Wnt1-Cre;IkbkapLoxP/LoxP (hereafter, “mutant”) and littermate Ikbkap+/LoxP (hereafter, “control”) mice (8), dissociated, and cultured in the presence of NGF to select for the small-diameter, TrkA+ pain and temperature receptors. We first examined the inner mitochondrial membrane potential, because this membrane is essential for proper mitochondrial bioenergetics and cell survival (33). Based on the degree of accumulation of positively charged MitoTracker Red CMXRos, we found not only that mitochondria from mutant neurons were depolarized compared with those in their littermate controls, but also that their membrane potentials could be fully restored to control levels by BGP-15 (Fig. 1 A and B). To determine whether the reduced accumulation of MitoTracker Red in mutant neurons was caused simply by a reduction in mitochondrial number or size, we incubated neurons with both MitoTracker Red and MitoTracker Green. Because the latter accumulates independently of membrane potential and serves as a measure of mitochondrial mass, this staining allowed us to normalize the measurements of mitochondrial membrane potential to mitochondrial mass. The results of this analysis validated our initial findings that mitochondria of mutant neurons were depolarized but were fully restored to control levels in the presence of BGP-15 (Fig. S1). We next analyzed the mitochondrial morphology of mutant neurons, which might be indicative of dysfunctional mitochondrial dynamics (e.g., in fission, fusion, transport, and/or mitophagy) (34). CellLight Mitochondria-GFP was added to both BGP-15–treated and untreated embryonic control and mutant DRG neurons to visualize mitochondria (Fig. 1C). A blinded scoring method was used to determine the shape of mitochondrial networks (Fig. 1D). This analysis was followed by ImageJ’s particle analyzer function to measure the number, perimeter, and area of the mitochondrial particles as calculated from converted binary images of the Mitochondrial-GFP fluorescence (Fig. 1 C and EG). We found that mitochondria from mutant DRG neurons were severely fragmented compared with the mitochondria of neurons from their control littermates. BGP-15 improved mitochondrial integrity by significantly reducing fragmentation and partially restoring morphology to control levels. At a concentration of 50 µM, BGP-15 was toxic to cultured DRG neurons. A 30-µM concentration induced a slight improvement in the overall morphology of mutant neurons (see elongation score, Fig. 1D), but a decline in the morphology of control neurons. A 10-µM concentration of BGP-15 appeared to be optimal for improving mitochondrial function (Fig. 1 CG).
Fig. 1.
Mitochondrial membrane potential and morphology are disrupted in Ikbkap−/− neurons but are ameliorated by incubation with BGP-15. (A) Representative images of MitoTracker CMXRos retention (red) in mitochondria of Tuj-1+ (green) E17.5 TrkA+ DRG neurons show decreased MitoTracker intensity compared with control neurons. (B) Average intensity measurements of MitoTracker signal [region of interest (ROI), soma]. Data are presented as fold change where Ikbkap−/− intensities were normalized to PBS-treated control intensity (dashed line). BGP-15 (10 μM) restores MitoTracker intensity to control levels. n = 4 experiments from ∼100 TrkA+ neurons, isolated from a total of nine control and eight mutant embryos. (C) Representative images of Mitochondria-BacMam GFP (green), Tuj-1 (red), and DAPI+ (blue) E17.5 TrkA+ DRG neurons reveal fragmented Ikbkap−/− mitochondrial networks and the partial restoration of morphology with 10 and 30 μM BGP-15. The bottom row shows the mitochondria-GFP signal converted to binary images used for quantification. (D) The elongation score is a measure of mitochondrial morphology based on comparison with reference images (0 = complete fragmentation; 4 = elongated mitochondria). (EG) Binary images indicate fragmented Ikbkap−/− networks and improvement by BGP-15. Data are presented as fold change with all measurements normalized to PBS-treated control measurements (dashed line). Graphs show the average number (E), the average perimeter (F), and the average area [per unit area of the defined ROI (soma)] (G) of mitochondrial particles. n = 590 TrkA+ neurons from 14 control and 11 mutant embryos, collected over six experiments. (Scale bars, 10 μm.) Errors bars indicate SEM. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001.
Fig. S1.
Ratiometric analysis of the ratio of MitoTracker CMXRos to MitoTracker Green reveals that mitochondrial membrane potential is disrupted in Ikbkap−/− neurons but is fully restored by incubation with BGP-15. MitoTracker CMXRos, which accumulates within the mitochondrion based on the mitochondrial membrane potential, was normalized to MitoTracker Green, which is retained independently of membrane potential and is a measure of mitochondrial mass. The normalized data indicate that the collapsed membrane potential of Ikbkap−/− TrkA+ neurons is restored to the levels observed in littermate control neurons with 10 μM of BGP-15. n = 170 total neurons, ∼40 neurons of each genotype and treatment. Error bars indicate the SEM; ***P ≤ 0.001.

BGP-15 Reduces Elevated ROS in Ikbkap−/− Neurons.

The discovery of significant impairments in mitochondrial function of mutant neurons, in particular the collapsed mitochondrial membrane potential, prompted us to measure levels of ROS (35). BGP-15 has been shown to reduce ROS in stressed mammalian cells and rodent models (1719, 30). Using the probe CellROX Deep Red, which fluoresces only when oxidized by ROS, we discovered that ROS were significantly increased in mutant neurons. We observed a strong CellROX fluorescent signal in somas of mutant DRG neurons, a marked increase from the low levels of fluorescence observed in neurons from littermate control embryos (Fig. 2). Incubation with 10-µM BGP-15 reduced the ROS levels in mutant neurons to those seen in littermate control neurons (Fig. 2).
Fig. 2.
ROS are increased in Ikbkap−/− neurons but are reduced to control levels upon incubation with BGP-15. (A) Ikbkap−/− DRG neurons have increased ROS compared with E17.5 DRG neurons from control littermates, indicated by fluorescence of CellROX. TL, transmitted light. (Scale bar, 10 µm.) (B) Average intensity values of CellRox (ROI, soma). Data are presented as fold change with each set (a replicate of all conditions) normalized to the intensity of controls (dashed line). Incubation with 10 μM BGP-15 reduces ROS of Ikbkap−/− neurons to control levels. n = 6 sets over four experiments, representing 11 control and 11 mutant embryos and ∼40 cells of each genotype and treatment. Error bars indicate SEM; ***P ≤ 0.001.

BGP-15 Normalizes Impaired Actin Dynamics in Ikbkap−/− Neurons.

It has been postulated that neurons die in FD because of an inability to innervate correctly their targets from which they receive essential survival stimuli in the form of neurotrophins (36, 37). We show here that incubation in NGF is insufficient to rescue the small-diameter TrkA+, Ikbkap−/− neurons. However, the failure to extend or maintain axons to mediate normal transport during target innervation could also ultimately cause neuronal death. In support of this thesis, IKAP/ELP1 has been shown to be necessary for normal cytoskeletal dynamics (36, 38, 39) and for the retrograde transport of NGF (40), whereas ELP3, the acetyl-transferase subunit of the Elongator complex, has been proposed to regulate actin dynamics (13, 41). For a better understanding of why Ikbkap−/− neurons die, we investigated their behavior in vitro, examining their overall morphology, axon outgrowth, and growth cone dynamics. Using live time-lapse confocal imaging, we found that, as compared with control littermate neurons, mutant neurons had stunted axons marked by an atypical, highly branched morphology that included increased extensions off the primary axon and growth cones that extended profuse and abnormally elongated filopodia (Fig. S2 and Movies S1 and S2). Axonal extensions over identical time intervals indicated that mutant growth cones were less motile than controls, with little net gain in growth cone forward position compared with control axons (Fig. 3F and Movies S1 and S2).
Fig. 3.
BGP-15 restores normal filopodia morphology and number to Ikbkap−/− growth cones. (A and B) Ikbkap−/− neurons are stunted, extending less in culture. (A) Representative images of E15.5 control and Ikbkap−/− neurons with Tuj-1 (green) and phalloidin (red) labeling. Arrows mark characteristic branching pattern of Ikbkap−/− neurons. (Scale bar, 10 μm.) (B) Percent of axons that extend outside and within the ROI (field of view). BGP-15 (10 μM) is unable to restore axon length. n = 12 wells from three experiments. (C) Representative images of E15.5 growth cones stained with Tuj-1 (green) and phalloidin (red) show the irregular morphology of Ikbkap−/− growth cones; filopodia are more numerous and longer than those of controls (arrows) and often branch (*). (Scale bar, = 5 μm.) (D and E) Quantification of fixed growth cones indicating the average number (D) and length (E) of filopodia upon treatment. Both filopodia number and length are reduced to control levels with 10 μM BGP-15. n = 15–20 neurons of each genotype and treatment over three experiments. (F) Percent of both control and Ikbkap−/− growth cones upon treatment that are stationary or show a gain in position. n = 15–20 growth cones of each genotype and treatment over three experiments. (G) Tuj-1 (blue), phalloidin (red), and mitochondria-BacMam GFP (green) labeling confirms altered cytoskeletal morphology (microtubules, arrowhead) and mitochondria (arrows) in Ikbkap−/− axons, branches, and growth cones. (Scale bar, 5 μm.) Error bars indicate SEM. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001.
Fig. S2.
Ikbkap−/− axons are highly branched compared with axons of control littermates. (A and B) Transmitted light, 63× images of a control (A) and a Ikbkap−/− TrkA+ (B) DRG neuron. Ikbkap−/− neurons are more branched (arrowheads). Arrows indicate the abundant and filopodia-like projections of Ikbkap−/− growth cones. (Scale bar, 10 µm.) (C) Average number of axonal extensions within an axon length of 50 μm from the soma (includes any extension >2 μm). The addition of 10 μM BGP-15 does not significantly reduce the highly branched Ikbkap−/− axons. n = 15–20 randomly selected neurons of each genotype and condition that were imaged alternating between genotypes and conditions to account for lapsed time. Results were replicated over three experiments. Error bars indicate SEM; *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001.
To investigate the organization of the cytoskeleton, we immunolabeled actin and microtubules in fixed cultures of DRG neurons (Fig. 3 A and C). There were three major differences between Ikbkap−/− and control neurons: (i) mutant axons extended for shorter distances (Fig. 3B); (ii) their growth cones sent out an increased number of actin-filled filopodia (Fig. 3 CE); and (iii) their axonal branches were composed primarily of actin with only infrequent microtubule invasion (Fig. 3 A and G). Axonal branching is a key component of normal axon outgrowth during development (42). Microtubule invasion is a prerequisite for branch stabilization but is dependent on the local aggregation of mitochondria at branch points (42). To explore branch composition further, we analyzed the distribution of mitochondria at axonal branch points. In control DRG neurons, CellLight Mitochondria-GFP revealed an aggregation of mitochondria at branch points and branches that were composed of both actin and tubulin (Fig. 3G). In contrast, Ikbkap−/− axonal branches contained fewer microtubules. Mitochondria were also much less abundant at branch points and were more fragmented than in controls (Fig. 3G). To determine whether BGP-15 could correct the axon outgrowth impairments, we measured axon extension and branching along with filopodia number in control and mutant axons. Although BGP-15 did not promote axon extension or reduce axonal branching significantly (Fig. 3B and Fig. S2C), it was very effective in reducing both the number and length of the excessive actin-rich filopodial extensions from the growth cones (Fig. 3 CE). Together, these data indicate that, although BGP-15 did not rescue axonal microtubule dynamics, it significantly restored normal actin dynamics in Ikbkap−/− neurons.

BGP-15 Prevents the Death of Ikbkap−/− Sensory Neurons in Vitro and in Vivo.

Ikbkap−/− TrkA+ neurons begin dying within 24 h of being placed in culture, even in the presence of their preferred growth factor, NGF (Fig. 4A). Neurons from their control littermates can survive for weeks in the same conditions (Fig. 4A). Given the ameliorative effects of BGP-15 on mitochondrial activity and actin dynamics, DRG from Wnt1-Cre;Ikbkap−/− embryonic mice were removed, dissociated into single neurons, and cultured in the presence or absence of BGP-15. Although a 1-µM concentration of BGP-15 had little effect, and 50–100 µM was toxic to the cells, 10 µM BGP-15 significantly decreased the death of the TrkA+ small-diameter mutant neurons in vitro. The same concentration of BGP-15 on control neurons induced a small but significant increase in survival, perhaps resulting from a reduction in naturally occurring programmed cell death (Fig. 4A) (43). We next tested the efficacy of BGP-15 in vivo by injecting pregnant dams once daily from E12.5 to E16.5 with either PBS or 100 mg/kg of BGP-15 i.p. and counting the number of TrkA+ DRG neurons at E17.5. BGP-15 had a striking effect: It significantly reduced the loss of TrkA+ neurons in vivo (Fig. 4 C and D), restoring their numbers to normal levels (8).
Fig. 4.
Death of Ikbkap−/− DRG neurons is significantly decreased by BGP-15 both in vitro and in vivo. (A) Representative images of DAPI (blue) and Tuj-1+ (red) control and Ikbkap−/− E16.5 TrkA+ neurons. (Scale bar,100 μm.) (B) Average increase in TrkA+ neuronal survival upon the addition of 10 μM BGP-15, calculated as fold change in which counts of BGP-15–treated neurons were normalized to counts of PBS-treated neurons (dashed line). n = 5 experiments and DRG from a total of 10 control and 10 mutant embryos. (C) Representative images of TrkA+ neurons (green) at E17.5 after daily injection of 100 mg/kg BGP-15 or PBS (the dotted line indicates the DRG). (Scale bar, 50 μm.) (D) Average number of TrkA+ neurons per section of DRG of the upper lumbar region from E17.5 embryos. n = 6 control embryos treated with saline, 5 embryos treated with BGP-15, 4 mutant embryos treated with saline, and 4 mutant embryos treated with BGP-15 embryos, carried out over four separate experiments. Error bars indicate SEM; *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001.

BGP-15 Decreases Elevated pJNK in Ikbkap−/− DRG.

There is evidence that IKAP/ELP1 participates in JNK stress signaling (12), and pJNK levels have been demonstrated to be reduced by exposure to BGP-15 (21, 23). Therefore, we asked whether pJNK levels were increased in Ikbkap−/− DRG neurons, and, if so, whether they could be normalized by treatment with BGP-15 in vivo. Pregnant dams were injected with PBS or BGP-15 from E12.5 to E16.5, and embryos were fixed, sectioned, and stained with antibodies to pJNK. Compared with levels found in control DRG neurons, pJNK expression was significantly elevated in DRG of mutant embryos. pJNK returned to control levels following daily treatment with BGP-15 (Fig. S3).
Fig. S3.
pJNK is increased in Ikbkap−/− DRG of mutant embryos but is reduced with daily injection of BGP-15. (A) Representative cross-sections of DRG from E17.5 control and mutant embryos exposed in utero to either PBS or 100 mg/kg BGP-15 via daily i.p. injection into pregnant dams. DRG neurons of mutant embryos exhibit an increase in pJNK (green), but with BGP-15 treatment levels are reduced to those observed in control DRG. Tuj-1 (red) and DAPI (blue) labeling were used to identify individual neurons of the DRG. The top row represents tissue incubated without primary antibodies, used as a control to confirm the validity of the pJNK staining and to determine the level of background staining for quantification. (Scale bar, 10 μm.) (B and C) In utero exposure to BGP-15 significantly reduces pJNK in Ikbkap−/− DRG neurons of mutant embryos to levels found in DRG neurons of control embryos. (Scale bar, 10 µm.) (B) Average intensity measurements of pJNK, represented as the fold change calculated from the baseline intensity of PBS-treated control DRG (dashed line; the ROI is a 50-μm box in the center of the DRG; measurements were taken in all 16-μm sections of the DRG of the upper lumbar region). (C) Percent of pJNK+ neurons of the total neurons counted per field (three to five fields were counted in each treatment) after background was reduced (the background, derived from the intensity of the imaged tissue stained with only secondary antibodies, was determined for each experiment). n = 4 control + PBS embryos, 4 control + BGP-15 embryos, 4 mutant + PBS embryos, and 4 mutant + BGP-15 embryos, carried out over four experiments. Error bars indicate SEM; *P ≤ 0.05; ***P ≤ 0.001.

BGP-15 Improves Ikbkap−/− Cardiac Innervation in Vivo.

FD patients have impaired blood pressure regulation and suffer from sudden death resulting from cardiac arrhythmias and asystole (44). These symptoms coincide with a decrease in sympathetic innervation of the heart, particularly its ventral apex (37, 45). Based on our finding that BGP-15 could rectify actin dynamics in Ikbkap−/− neurons in vitro, we asked whether treatment with the compound could reverse the defect in cardiac innervation that has previously been reported in Ikbkap−/− embryos (37). BGP-15 was injected daily into pregnant dams from E12.5 to E16.5 (100 mg/kg, i.p.), and embryos were harvested at E17.5. As in previous studies, we found that hearts of mutant mice were significantly less well innervated than hearts from littermate controls. Mutants had significant reductions in axonal extension and branching (Fig. S4A). Daily injections of BGP-15 significantly improved cardiac innervation in mutant embryos, increasing both the number and ramification of branches in the heart and restoring innervation of the heart middle region to control levels. Drug treatment partially restored growth of the longest axons that innervate the inferior apical region of mutant hearts (Fig. S4).
Fig. S4.
Innervation is reduced in Ikbkap−/− hearts of mutant embryos but improves with daily injection of BGP-15. (A) Representative images of control and Ikbkap−/− hearts harvested from E17.5 embryos exposed in utero to either PBS or 100 mg/kg BGP-15 via daily i.p. injection into pregnant dams demonstrate that both the extent and complexity of Ikbkap−/− heart innervation is increased with BGP-15 treatment (arrowheads). Boxes mark the three ROIs [superior, middle, and inferior, separated by transition 1 and transition 2 (arrows)] used to quantify the degree and patterning of innervation. Dashed boxes demarcate the areas magnified in the Insets (magnification: approximately 2×). (BD) Quantification of Tuj-1 axons shows the average number of branch points in each ROI (P values show statistical significance compared with values of saline-treated mutant hearts) (B), the percent of ROIs with axons present (C), and the average number of axons extending across transition 1 and transition 2 (D). n = 10 control + saline, 12 control + BGP-15, 10 mutant + saline, and 11 mutant + BGP-15 hearts, imaged over three experiments. Error bars indicate SEM; *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001.


Why neurons die in both the developing and in the adult FD nervous system has not been resolved, but evidence is accumulating that neurons in both human FD patients and mouse models of FD experience intracellular stress. We report here that mitochondrial function is impaired in neurons lacking Ikbkap: Mitochondria are depolarized (Fig. 1 A and B and Fig. S1) and fragmented (Fig. 1 CG), and ROS levels are significantly increased (Fig. 2). The small molecule BGP-15 can restore the mitochondrial membrane potential (Fig. 1 A and B and Fig. S1), reduce the elevated ROS levels (Fig. 2), and reduce mitochondrial fragmentation (Fig. 1 CG). pJNK levels are also higher in Ikbkap−/− DRG neurons than in those of controls, and these elevated levels also were normalized by BGP-15 treatment (Fig. S3). Although BGP-15 exerts diverse actions on cells, our data suggest that, by restoring aspects of mitochondrial function, BGP-15 prevents the death of Ikbkap−/− neurons both in vitro and in vivo. Because these mitochondrial disruptions can trigger apoptosis (46) and loss of Ikbkap−/− TrkA+ DRG neurons is attributed to this method of programmed cell death (PCD) (8), we conclude that BGP-15 acts to prevent the apoptotic death of neurons that lack Ikap/Elp1.
Our data also reveal that axon extension and growth cone morphology and dynamics are perturbed in neurons that lack Ikbkap. The impaired actin-mediated functions can be corrected by BGP-15. Axons of Ikbkap−/− neurons are stunted and highly branched (Fig. S2), and growth cones extend abnormally long and abundant filopodia (Fig. 3 CE). Perhaps as a consequence of the excess filopodia, their axonal growth cones do not exhibit much, if any, total forward movement (Fig. 3F and Movies S1 and S2). Using the heart as a model to explore innervation in vivo, we found that cardiac innervation is reduced in the absence of Ikbkap, especially more inferiorly toward the heart apex (Fig. S4). BGP-15 treatment inhibited the formation of overabundant growth cone actin networks (Fig. 3 CE) and improved cardiac innervation (Fig. S4). Thus, improvements in cytoskeletal networks, especially of actin dynamics, may also enhance the survival of Ikbkap−/− neurons treated with BGP-15 in vitro and in vivo (Fig. 4).
These data implicate mitochondrial dysfunction as a major factor mediating the death of Ikbkap−/− neurons. First, we show that mitochondria of Ikbkap−/− neurons are fragmented and do not distribute correctly in the axon or coalesce at axonal branch points (Figs. 1 CG and 3G), suggesting impaired mitochondrial dynamics. Disrupted mitochondrial dynamics are detrimental to highly polarized sensory neurons, and such interruptions could contribute to apoptosis and later-onset neurodegeneration (34, 47). This feature is shared with neurons of patients who have Parkinson’s, Huntington’s, or Alzheimer’s disease (34). Second, we show that Ikbkap−/− neurons have a significant increase in ROS (Fig. 2), likely resulting from dysfunction in the mitochondrial respiratory chain, which can induce PCD (35). BGP-15 reduces ROS levels to normal in Ikbkap−/− neurons, in support of previous work demonstrating that BGP-15 can stabilize complexes of the respiratory chain (18, 28, 30), especially complexes I and III (28, 30). Third, our data indicate a widespread depolarization of mitochondrial networks (Fig. 1 A and B and Fig. S1). This depolarization may be caused in part by ROS-induced damage leading to a loss of charge across the inner mitochondrial membrane, a point of no return on the road to PCD (33, 35). BGP-15 appeared to prevent the collapse of the membrane potential of Ikbkap−/− neurons, thereby rescuing the cells. Whether mitochondrial impairment is the primary cause or simply a critical mediator of cell death is a question being investigated in a number of neurodegenerative diseases (47). Our data suggest that improving mitochondrial health may be beneficial to patients with such problems.
BGP-15–driven improvements in actin function (Fig. 3 and Fig. S4) could also have contributed to the increased survival of Ikbkap−/− neurons in vitro and in vivo. There is accumulating evidence that IKAP/ELP1 and the Elongator complex participate in the organization of the cytoskeleton (3639, 41, 4852). Specifically, Elongator has been demonstrated to be required for normal neuronal branching (36, 37, 41, 4851), organization of actin networks (38), and acetylation of α-tubulin (40, 50, 52). Microtubules also have been shown to be altered in growth cones from peripheral neurons lacking IKAP/ELP1 in chick (36). BGP-15 did not induce any significant improvements in Ikbkap−/− microtubule networks (i.e., it did not promote the growth of the Ikbkap−/− axons). Thus, although we cannot exclude the possibility that BGP-15 affects microtubule defects, we have no evidence for such benefits. Consequently, we focused on the compound’s obvious action on actin networks in growth cone filopodia, where we saw significant improvements (Fig. 3 CE). We also observed improved cardiac innervation (Fig. S4). It has been suggested that reduced target innervation in the absence of Ikbkap is the primary cause of death of sensory neurons in FD (37). Thus, the BGP-15–induced increase in cardiac innervation could result from the increase in neuronal cellular respiration and/ or actin-based growth cone function. This finding would suggest that reduced target innervation is the detrimental consequence of mitochondrial dysfunction and cytoskeletal disruption. These data are supported by the finding that mitochondria are required to initiate and maintain the formation of axonal branches in DRG neurons (42); the ability to extend and retract branches dynamically is integral to the navigation of axons toward their targets. Furthermore, restoration of actin network function could improve neuronal survival, because actin dynamics mediate vesicle transport and are required specifically for the TrkA/NGF endosomal retrograde survival signaling in DRG neurons (53) that is necessary to prevent the default apoptotic response (46). Previous studies have suggested a role of IKAP/ELP1 in neuronal transport (36, 54), and recent work has revealed that the speed of NGF retrograde transport is reduced in neurons lacking Ikbkap (40). Additionally, Rac-1, which interacts with actin to regulate growth cone dynamics (55), has been demonstrated in mammalian cells to be a target of BGP-15 treatment (31, 32) and potentially could contribute to the positive effects of BGP-15 on the actin cytoskeleton, because actin has been shown to activate signaling cascades that lead to apoptosis (56). Thus, improvements in cytoskeletal networks following BGP-15 treatment also could underlie its effects on Ikbkap−/− neurons.
Last, we show here that pJNK is increased in Ikbkap−/− DRG of mutant embryos (Fig. S3), although the mechanisms mediating this increase will require further study. Because IKAP possesses a JNK association site and has been demonstrated to potentiate JNK signaling in vitro (12), misregulated pJNK could be a direct response to the absence of IKAP. Alternatively, elevated pJNK could result from mitochondrial dysfunction, as suggested by elevated ROS levels (35), and/or from altered Rac and Cdc42 signaling, given the disrupted cytoskeletal networks of Ikbkap−/− neurons (56). BGP-15 reduced the elevated pJNK in Ikbkap−/− DRG neurons to control levels. The compound has been shown previously to reduce pJNK in many cell types by potentiating the heat-shock response (21), improving mitochondrial function (23), or modulating Rac-1 activity (31, 32).
Clinical data support impairment in mitochondrial function as a contributor to neuronal death in FD patients. For example, in the FD human (and mouse) retina, a progressive loss of RGCs occurs in the more metabolically active temporal half of the retina, reminiscent of the selective loss of temporal RGCs in Leber’s hereditary optic neuropathy, a consequence of a mutation in a mitochondrial gene (4, 5, 15, 57). Patients with FD also have been reported to be susceptible to developing rhabdomyolysis (16). These retina and muscle problems both suggest that mitochondrial dysfunction could play a significant role in FD.
IKBKAP/ELP1 is one of six subunits of the Elongator complex, and variants in three other Elongator subunits, ELP24, have also been associated with neurological disorders: ELP3 with amyotrophic lateral sclerosis (ALS) (49) and ELP2 and ELP4 with intellectual impairment and epilepsy (5860), demonstrating the importance of this complex in the nervous system. Furthermore, ALS is marked by mitochondrial dysfunction (61). Recent work has revealed that in the absence of Elp3 in mice, cortical neurogenesis is impaired because of the activation of the unfolded protein response, which is triggered by stress in the endoplasmic reticulum (9). That study and ours help elucidate the function of the Elongator complex in neurons and reveal that its absence triggers intracellular stress. Given the requirement for Elongator in translation, the question remains whether this stress is a direct or indirect consequence of impaired translation. Although the complete repertoire of BGP-15’s effects on neurons remains to be elucidated, our data demonstrate that it promotes mitochondrial health and cytoskeletal organization, both of which are essential for the function and survival of TrkA+ peripheral neurons that extend up to meter-long axons.
In summary, our experiments demonstrate the potential utility of BGP-15 in a neurological disease model. In addition, our data implicate disruption of mitochondrial function as a pathological mechanism contributing to the death of neurons in FD and place FD alongside more prevalent neurodegenerative disorders characterized by mitochondrial dysfunction (34, 47). BGP-15 has been shown to improve metabolic function in rodent models of several human degenerative diseases (18, 22, 23, 2628) and has been given to more than 400 patients in human clinical trials with no severe adverse drug-related events (62). The data reported here suggest that this drug, or compounds like it, may be effective in slowing or preventing the progressive loss of neurons in FD patients.

Materials and Methods

All research involving animals has been approved by the Institutional Animal Care and Use Committee at Montana State University. All procedures involving mice adhered to the NIH Guide for the Care and Use of Laboratory Animals (63). For additional descriptions of methods, please see SI Materials and Methods.

SI Materials and Methods

Primary Culture of Mouse DRG Neurons.

DRG were dissected from E15.5, E16.5, or E17.5 Wnt1-Cre;IkbkapLoxP/LoxP (mutant) and Ikbkap+/LoxP (control) embryos and were cultured in the presence of 25 ng/mL NGF to select for TrkA neurons. Embryos were first genotyped, and DRG from control and littermate mutant pups were dissected, pooled, and stored overnight at 4 °C in Hibernate-E medium (Thermo Fisher) supplemented with 2% B-27 (Gibco) and 2 mM l-glutamine (Gibco). Nunc LabTek II chambered coverglass eight-well slides (Fisher Scientific) were coated on the day of plating with 10 µg/mL poly-d-lysine (Millipore) for 30 min at room temperature) and with 20 µg/mL laminin (Thermo Fisher) for 2 h at 37 °C, 5% CO2. To dissociate neurons for plating, DRG were incubated with 0.05% trypsin + EDTA (Thermo Fisher) at 37 °C for 16 min. Trypsin was quenched with 10% FBS (Thermo Fisher) in Neurobasal medium (Thermo Fisher), and neurons were spun down at 188 × g for 2 min at 4 °C. The supernatant was decanted, and neurons were resuspended in Neurobasal medium supplemented with 2% B-27, 100 µg/mL penicillin/streptomycin (Mediatech), 2 mM l-glutamine, and 25 ng/mL NGF (Thermo Fisher) to select for TrkA+ neurons. Approximately 400 µL of supplemented Neurobasal medium per dissected embryo was added for the resuspension, resulting in an optimal concentration of cells for plating. Neurons were further dissociated with a series of fire-pulled Pasteur pipettes, and the cell solution was passed through a 40-µm Falcon Cell Strainer (Fisher Scientific) to eliminate any remaining clumps. Cells were counted with a hemocytometer and were plated at final concentration of 2.0 × 104 cells per well in a total volume of 300 µL. Cultures were incubated at 37 °C, 5% CO2, for up to 72 h.

Statistical Analysis.

Data are presented as the average ± SEM when a minimum of three experiments were conducted, representing a minimum of embryos from three separate litters. Statistical tests are noted in each experimental description. In some cases, data were preprocessed by normalization to account for experiment-to-experiment variability in plating, staining, or imaging; for that reason all genotype and treatment conditions were always represented in each experimental repetition. Any normalization of data is noted both in the figure legends and in the detailed description of methods below. The sample size, n, is also noted in the figure legends. When normalization of data was used (i.e., when an increase or decrease in neuronal survival was compared across experiments), n represents the number of experiments; for the in vivo studies, n represents the number of embryos considered. Statistical tests detailed in each specific method are found below, but in general, all one-way ANOVA tests assume independence of observations, normal distribution of data, and equality of variances. Student’s t tests are two-sided and assume normal distribution, random sampling, and continuous and independent observations. These tests use an alpha value of 0.05 and a 95% confidence level. The BootstRatio analysis was used to determine the statistical significance of data normalized to a control treatment, presented as fold change. The test assumes no underlying distribution of data and uses resampling methods, including bootstrap and permutation tests, of fold change ratios to determine significance. These data were considered significant at P ≤ 0.05.

Study Design.

For the in vitro studies (e.g., the in vitro neuronal survival and mitochondrial health assays), a minimum of three experiments were conducted, with each experiment consisting of data from littermate control and affected embryos. Studies were concluded after a maximum of five experiments were completed. For in vivo studies (e.g., the in vivo neuronal survival, pJNK, and heart innervation assays), pregnant dams were injected and killed until ∼5–10 embryos of each genotype and treatment were collected. Embryos were harvested from a minimum of three dams used over three separate experiments. An experiment represents a set of staining and imaging of tissue from all genotypes and treatments.
As a minimum, each experimental condition being tested (e.g., the addition of a specific concentration of BGP-15) was repeated and statistically analyzed over three experiments. When total neuronal counts of an entire well were evaluated, treatments were carried out over duplicate wells. Any duplicate wells (representing the same genotype and with neurons pulled from a common source of cells pooled from all littermate control or affected embryos) that differed in count by more than 250 neurons were deemed too variable (because of plating technique) and were excluded, and the experiment was repeated. For live-cell imaging, one experiment included measurements of 5–10 cells of each genotype and treatment condition; fixed-cell imaging included data from 20–50 cells of each genotype and treatment condition. All data points were considered. No outliers were excluded.
All experiments were carried out in a controlled laboratory environment. Tissue was collected from mice housed in our fully accredited Animal Resources Center, and observations were conducted using a Leica SP8 confocal microscope for organelle and cytoskeletal-level resolution. For data analysis, the investigator was blinded to the genotype and treatment condition of all collected images after randomization. Specific parameters of each study are detailed below.

Genetically Modified Mice.

The Ikbkap/Elp1 gene was conditionally deleted from the neural crest cell lineage using a Wnt1-Cre transgene as previously described (8). Briefly, we, crossed homozygous Ikbkap conditional-knockout (IC/C) mice to hemizygous Wnt1-Cre mice (Cre+) carrying one copy of the conditional Ikbkap allele (I+/C). For all analyses, Cre+; IC/C embryos were used as experimental mice, and Cre; I+/C littermates were used as controls. All genotyping was performed via PCR. Wnt1-Cre mice were purchased from The Jackson Laboratory (stock no. 003829). All procedures involving mice adhered to the NIH Guide for the Care and Use of Laboratory Animals.


After fixation, both adherent cells and sectioned tissue were blocked (1% glycine, 0.1% Triton, and 10% goat serum in 1× PBS) for 1 h at room temperature and then were incubated in primary antibody in blocking solution overnight at 4 °C and in secondary antibody for 1 h at room temperature. Primary antibodies included Tuj-1 (1:1,000; Covance), TrkA (a gift from Louis Reichardt of the University of California, San Francisco) (1:5,000), and anti-GFP (1:2,000; Thermo Fisher). Secondary antibodies and stains included Alexa 488 and 568 (1:2,000; Thermo Fisher), DAPI (1:3,000; Thermo Fisher), and Alexa Fluor 594 phalloidin (1:40; Thermo Fisher). Antibodies for whole-mount heart imaging and diaminobenzidine (DAB) immunohistochemistry included Tuj-1 (1:1,000) and the HRP-linked secondary antibody, goat anti–mouse-HRP (1:500; SouthernBiotech). The DAB-Plus reagent kit (Thermo Fisher) was used for development following the manufacturer’s protocol.

Mitochondrial Assays.

To measure the mitochondrial membrane potential, DRG cultures were incubated overnight with BGP-15 (10 µM) or PBS and were incubated the following morning (17–18 h later) with 200 nM MitoTracker Red CMXRos (Thermo Fisher) for 20 min at 37 °C, 5% CO2. Cultures then were fixed for 10 min with 4% paraformaldehyde at room temperature, stained with Tuj-1 and DAPI, and imaged using a 63× objective on a Leica TCS SP8 confocal microscope. Z-stacks were generated using a step size of 0.3 µm. For normalizing the accumulation of MitoTracker Red CMXRos to the mitochondrial mass, DRG cultures were incubated with 25 nM of both MitoTracker Red and MitoTracker Green for 20 min and were imaged immediately with a 20× objective using a 0.3-µm z-step. Because cultures were imaged live, images were taken alternating between genotypes and treatment conditions. All imaging parameters (laser power, pinhole size, PMT settings, z-step size, and resolution) were held constant for accurate comparison between control and Ikbkap−/− mitochondria in each experiment. Intensity measurements were taken at 488 and 568 excitations, and the Leica Application Suite Advanced Fluorescence software was used for taking intensity measurements of somas selected at random throughout the well. For the fixed MitoTracker Red measurements, data represent the average fold change (i.e., the increase or decrease in MitoTracker Red accumulation in Ikbkap−/− neurons compared with the accumulation in littermate control neurons) from four experiments. The BootstRatio analysis was used to calculate significance of the fold change across experiments. For the ratiometric analyses, MitoTracker Red intensity was normalized to MitoTracker Green intensity for each soma individually. The one-way ANOVA and Student’s t test were used to determine statistical significance.
To measure mitochondrial morphology, CellLight Mitochondria-GFP BacMam 2.0 (Thermo Fisher) was added according to the manufacturer’s protocol to DRG neurons at the time of plating, along with BGP-15 (10 and 30 µM) or PBS treatment. The length of mitochondria-GFP incubation was optimized at 45 h with an 8-µL addition of mitochondria-GFP per well (calculated using a delivery of 40 particles per cell). Neurons were fixed, stained, and imaged with a 63× objective at the level of the axon, because mitochondrial morphology was too convoluted in a complied z-stack. Morphology was calculated using both the Analyze Particles function of ImageJ, using binary images converted from the Mitochondria-GFP signal, and a method in which an elongation score was assigned to the overall mitochondria-GFP network based on blinded comparison with preselected reference images. A score of 4 denoted interconnected mitochondrial networks with elongated, tubule-like mitochondrial networks; a score of 0 indicated fragmented and disconnected mitochondrial particles. Mitochondria were scored after randomization of the images by an investigator blinded to the cell type and treatment condition. The significance of the elongation score was calculated using one-way ANOVA and Student’s t test; the BootstRatio analysis was used to determine the significance of the fold increase from data normalized each experiment relative to PBS-treated control DRG neurons.
To measure mitochondrial ROS, DRG were cultured overnight with BGP-15 (10 µM) or PBS to which CellRox Deep Red (Thermo Fisher) was added the following day (17–18 h later) to a final concentration of 5 µM, followed by further incubation for 30 min at 37 °C at 5% CO2. Cells were washed and fixed for 10 min with 10% formalin at room temperature and were imaged immediately (the probe must be imaged within 2 h according to the manufacturer’s protocol). Transmitted light was used to identify neurons, and intensity measurements of somas were calculated using the Leica Application Suite of CellROX Deep Red, which fluoresces only upon oxidation by ROS (emission ∼665 nm). Images were taken using a 63× objective, and z-stacks were generated using a step size of 0.3 µm. Representative images in Fig. 2 show a plane at the level of the axon from z-stack images. Only healthy, fully attached neurons with extended axons were selected and imaged, and images were taken alternating between genotypes and treatments to account for any reduced sensitivity of the probe over time. Each replicate (in which all genotypes and treatments were represented) was normalized to the average intensity measurements of PBS-treated control neurons. The significance of the fold changes between experiments was determined using the BootstRatio analysis.

In Vitro Cytoskeletal Morphology and Dynamics.

DRG neurons were dissected from E15.5 embryos and were cultured as described above in the presence of BGP-15 (10 µM) or PBS. For fixed imaging of the cytoskeleton, cultures were incubated with a solution of 4% paraformaldehyde and 0.2% glutaraldehyde for 10 min at room temperature, treated with 1 mg/mL sodium borohydride for 15 min, permeabilized with 0.1% Triton X-100 for 5 min, and labeled with Tuj-1 and Alexa Fluor 594 phalloidin. Quantification of axon extension was derived from overnight cultures that were fixed at the same time. All neurons in a well were counted, and the percent that extended outside each field of view versus those that were contained within the field of view was noted. Time-lapse imaging was executed at 63× magnification using transmitted light over 5 min, with images taken every 5 s. Any displacement of the lamellipodium over the 5 min was characterized as forward movement. To quantify growth cone morphology, the number of filopodia in a defined ROI (a 30-µm diameter circle centered over the leading edge of the growth cone) was counted, and their length was measured. To count the number of branches extending off the primary axon, quantification was limited to the first 50 µm of the axon just distal to the soma. All imaging was performed on a TCS SP8 confocal microscope with a 63× objective. Quantification was carried out using ImageJ, and statistical significance was determined from the one-way ANOVA and Student’s t test.

In Vitro Survival Assay.

Either 1× PBS (vehicle) or BGP-15 (1 or 10 µM) was added to E16.5 DRG neurons at the time of plating. Neurons were incubated for 40 h at 37 °C, 5% CO2 and were fixed with 4% paraformaldehyde for 10 min at room temperature. Cells were stained with DAPI and Tuj-1, and wells were imaged on a Leica TCS SP8 confocal microscope. The total number of neurons per well was determined in replicate wells. A total of 10 wells were tallied of each genotype and condition. Survival increase/decrease data represent the average fold change across five experiments, calculated from baseline PBS-treated control counts. Statistical significance was calculated using the BootstRatio analysis.

In Vivo Survival Assays.

Pregnant dams were injected with either 1× PBS (vehicle) or 100 mg/kg BGP-15 (N-Gene) from E12.5 to E16.5. Embryos were killed at E17.5, genotyped, fixed with 4% paraformaldehyde (2 h at 4 °C), and frozen in O.C.T. (Fisher Scientific) compound after a series of sucrose incubations at 4° (15% sucrose in 1× PBS overnight, 30% sucrose in 1× PBS overnight, followed by a 1:1 ratio of 30% sucrose to 1× PBS for 2 h). Tissue was cryosectioned transversely in 16-µm sections using a Leica CM 1950 cryostat and was labeled with anti-TrkA antibody. Slides were imaged as a set, representing all the genotypes and treatments (equaling one experiment), on a Leica TCS SP8 confocal microscope. The number of TrkA+ neurons was quantified as described in George et al. (8), and after the data were randomized counts were carried out by an investigator blinded to the genotype and treatment condition. Significance was determined using one-way ANOVA and Student’s t test.

Heart Innervation Analysis.

Pregnant dams were injected with either 1× PBS (vehicle) or 100 mg/kg BGP-15 from E12.5 to E16.5, and embryos were collected at E17.5. Hearts were dissected, genotyped, fixed for 2 h at 4 °C in 4% paraformaldehyde, and prepared for whole-mount imaging with DAB immunohistochemistry, which included dehydration in 50% and 80% methanol in 1× PBS (both for 30 min at room temperature) and overnight incubation in a 4:1:1 ratio of methanol:DMSO:H2O2 at 4 °C to quench endogenous peroxidase activity. Hearts were stored in 100% methanol at −20 °C. In preparation for staining, hearts were rehydrated and blocked overnight in 4% BSA and 1% Triton X-100 in 1× PBS at 4 °C. Hearts were incubated in primary antibody (Tuj-1 in BSA blocking solution) for 72 h at 4 °C, washed, and incubated overnight in secondary antibody (HRP goat anti-mouse in BSA blocking solution at 4 °C). Hearts were rinsed and developed using the DAB-Plus reagent kit (Thermo Fisher) following the manufacturer’s protocol. After development, hearts were postfixed in 4% paraformaldehyde for 20 min at 4 °C and were dehydrated with a series of incubations in methanol (20%, 50%, and 80% in 1× PBS, followed by 100% methanol, all for 10 min at room temperature). Tissue was cleared just before imaging with a 2:1 mixture of benzyl benzoate to benzyl alcohol (5 min at room temperature). Imaging was carried out with a 10× objective and a Zeiss Stemi SV-11 APO stereo microscope with ProgRes Mac Capture Pro software. For quantification, the pulmonary trunk and apex served as the upper and lower boundaries, respectively, in the creation of three ROIs of equal height (dubbed “superior,” “middle,” and “inferior” ROIs). Significance was calculated using the one-way ANOVA and Student’s t test.

Activation of JNK.

To determine the activation of JNK, pregnant dams were injected with either 1× PBS or 100 mg/kg BGP-15 from E12.5 to E16.5, and tissue was collected, fixed, embedded, and frozen as described previously. Transverse 16-µm sections of the upper lumbar region were stained with pJNK (1:200; Cell Signaling) and colabeled with Tuj-1 (1:1,000; Thermo Fisher) to identify neurons of the DRG. For heat-based antigen retrieval, citrate buffer (10 mM citric acid, 0.05% Tween 20, pH 6.0) was used: Slides were incubated in a water bath in coplin jars with citrate buffer at 90° for 30 min; then the coplin jars were removed from the heat and were cooled on the bench for a further 10 min. Slides were washed with 1× PBS, tissue was blocked for 1 h with normal goat serum, and primary antibodies were added for overnight incubation at 4 °C. Secondary antibodies included Alexa 488 and 568 (1:2,000; Thermo Fisher) and DAPI (1:3,000; Thermo Fisher). Tissue was imaged on a Leica SP8 confocal microscope with a 63× objective. The pJNK signal was quantified via (i) intensity, specifically whether intensity was increased or decreased in relation to the pJNK intensity of PBS-treated control tissue (represented as fold change; the ROI was a 50-μm box in the center of the DRG, and measurements were taken of all 16-μm sections of DRG of the upper lumbar region) and (ii) the percent of pJNK+ neurons in a field of view after elimination of background staining. Tuj-1 and DAPI were used to identify neurons, and all neurons in the field of view were counted. Background-level of staining of each experiment was determined from control tissue incubated without primary antibody but taken through all the steps of the antigen retrieval and secondary antibody incubation. Background was reduced using ImageJ’s color threshold function. The BootstRatio was used to determine the significance of any fold change from the baseline control intensity; ANOVA and Student’s t test were used for to determine the significance of the percent of pJNK+ neurons.


We thank Marta Chaverra for technical assistance. This work was funded by NIH Neurological Disorders and Strokes Grant R01NS086796 (to F.L.) and by grants from the Dysautonomia Foundation.

Supporting Information

Supporting Information (PDF)
Movie S1.
Control axons possess motile growth cones. Representative transmitted light, 63× time-lapse movie of an E15.5 control growth cone, imaged over 5 min with one frame every 5 s. Control growth cones show a gain in position indicated by movement of their lamellipodia (red arrowheads mark the leading edge of a lamellipodium). Black arrows denote filopodia. (Scale bar, 10 µm.)
Movie S2.
Ikbkap−/− axons possess more branched and less motile growth cones than axons in control littermates. Representative transmitted light, 63× time-lapse movie of an E15.5 Ikbkap−/− growth cone, imaged over 5 min with one frame every 5 s. Like the growth cones of control DRG neurons, mutant Ikbkap−/− growth cones have motile filopodia (black arrows), but the filopodia in the mutants are longer and more abundant than those in control neurons. Mutant growth cones also rarely show a gain in position, as indicated by the lack of movement of their lamellipodia (red arrowheads mark the leading edge of a lamellipodium). (Scale bar, 10 µm.)


FB Axelrod, A world without pain or tears. Clin Auton Res 16, 90–97 (2006).
L Norcliffe-Kaufmann, SA Slaugenhaupt, H Kaufmann, Familial dysautonomia: History, genotype, phenotype and translational research. Prog Neurobiol, June 15, 2016).
P Dietrich, I Dragatsis, Familial dysautonomia: Mechanisms and models. Genet Mol Biol 39, 497–514 (2016).
CE Mendoza-Santiesteban, et al., Clinical neuro-ophthalmic findings in familial dysautonomia. J Neuroophthalmol 32, 23–26 (2012).
CE Mendoza-Santiesteban, TR Hedges Iii, L Norcliffe-Kaufmann, F Axelrod, H Kaufmann, Selective retinal ganglion cell loss in familial dysautonomia. J Neurol 261, 702–709 (2014).
A Blumenfeld, et al., Precise genetic mapping and haplotype analysis of the familial dysautonomia gene on human chromosome 9q31. Am J Hum Genet 64, 1110–1118 (1999).
A Esberg, B Huang, MJO Johansson, AS Byström, Elevated levels of two tRNA species bypass the requirement for elongator complex in transcription and exocytosis. Mol Cell 24, 139–148 (2006).
L George, et al., Familial dysautonomia model reveals Ikbkap deletion causes apoptosis of Pax3+ progenitors and peripheral neurons. Proc Natl Acad Sci USA 110, 18698–18703 (2013).
S Laguesse, et al., A dynamic unfolded protein response contributes to the control of cortical neurogenesis. Dev Cell 35, 553–567 (2015).
M Tigano, et al., Elongator-dependent modification of cytoplasmic tRNALysUUU is required for mitochondrial function under stress conditions. Nucleic Acids Res 43, 8368–8380 (2015).
F Bauer, D Hermand, A coordinated codon-dependent regulation of translation by Elongator. Cell Cycle 11, 4524–4529 (2012).
C Holmberg, et al., A novel specific role for I kappa B kinase complex-associated protein in cytosolic stress signaling. J Biol Chem 277, 31918–31928 (2002).
D Barton, F Braet, J Marc, R Overall, J Gardiner, ELP3 localises to mitochondria and actin-rich domains at edges of HeLa cells. Neurosci Lett 455, 60–64 (2009).
KL Stilger, Jr WJ Sullivan, Elongator protein 3 (Elp3) lysine acetyltransferase is a tail-anchored mitochondrial protein in Toxoplasma gondii. J Biol Chem 288, 25318–25329 (2013).
Y Ueki, G Ramirez, E Salcedo, ME Stabio, F Lefcort, Loss of Ikbkap causes slow, progressive retinal degeneration in a mouse model of familial dysautonomia. eNeuro 3, 1–51 (2016).
J-A Palma, R Roda, L Norcliffe-Kaufmann, H Kaufmann, Increased frequency of rhabdomyolysis in familial dysautonomia. Muscle Nerve 52, 887–890 (2015).
E Szabados, P Literati-Nagy, B Farkas, B Sümegi, BGP-15, a nicotinic amidoxime derivate protecting heart from ischemia reperfusion injury through modulation of poly(ADP-ribose) polymerase. Biochem Pharmacol 59, 937–945 (2000).
R Halmosi, et al., Effect of poly(ADP-ribose) polymerase inhibitors on the ischemia-reperfusion-induced oxidative cell damage and mitochondrial metabolism in Langendorff heart perfusion system. Mol Pharmacol 59, 1497–1505 (2001).
I Racz, et al., BGP-15 - a novel poly(ADP-ribose) polymerase inhibitor - protects against nephrotoxicity of cisplatin without compromising its antitumor activity. Biochem Pharmacol 63, 1099–1111 (2002).
G Bárdos, et al., BGP-15, a hydroximic acid derivative, protects against cisplatin- or taxol-induced peripheral neuropathy in rats. Toxicol Appl Pharmacol 190, 9–16 (2003).
J Chung, et al., HSP72 protects against obesity-induced insulin resistance. Proc Natl Acad Sci USA 105, 1739–1744 (2008).
G Nagy, et al., BGP-15 inhibits caspase-independent programmed cell death in acetaminophen-induced liver injury. Toxicol Appl Pharmacol 243, 96–103 (2010).
Z Sarszegi, et al., BGP-15, a PARP-inhibitor, prevents imatinib-induced cardiotoxicity by activating Akt and suppressing JNK and p38 MAP kinases. Mol Cell Biochem 365, 129–137 (2012).
SM Gehrig, et al., Hsp72 preserves muscle function and slows progression of severe muscular dystrophy. Nature 484, 394–398 (2012).
G Sapra, et al., The small-molecule BGP-15 protects against heart failure and atrial fibrillation in mice. Nat Commun 5, 5705 (2014).
DC Henstridge, et al., Activating HSP72 in rodent skeletal muscle increases mitochondrial number and oxidative capacity and decreases insulin resistance. Diabetes 63, 1881–1894 (2014).
LL Wu, et al., Mitochondrial dysfunction in oocytes of obese mothers: Transmission to offspring and reversal by pharmacological endoplasmic reticulum stress inhibitors. Development 142, 681–691 (2015).
H Salah, et al., The chaperone co-inducer BGP-15 alleviates ventilation-induced diaphragm dysfunction. Sci Transl Med 8, 350ra103 (2016).
TL Kennedy, et al., BGP-15 improves aspects of the dystrophic pathology in mdx and dko mice with differing efficacies in heart and skeletal Muscle. Am J Pathol 186, 3246–3260 (2016).
K Sumegi, et al., BGP-15 protects against oxidative stress- or lipopolysaccharide-induced mitochondrial destabilization and reduces mitochondrial production of reactive oxygen species. PLoS One 12, e0169372 (2017).
I Gombos, et al., Membrane-lipid therapy in operation: The HSP co-inducer BGP-15 activates stress signal transduction pathways by remodeling plasma membrane rafts. PLoS One 6, e28818 (2011).
B Güngör, et al., Rac1 participates in thermally induced alterations of the cytoskeleton, cell morphology and lipid rafts, and regulates the expression of heat shock proteins in B16F10 melanoma cells. PLoS One 9, e89136 (2014).
G Kroemer, L Galluzzi, C Brenner, Mitochondrial membrane permeabilization in cell death. Physiol Rev 87, 99–163 (2007).
H Chen, DC Chan, Mitochondrial dynamics–fusion, fission, movement, and mitophagy–in neurodegenerative diseases. Hum Mol Genet 18, R169–R176 (2009).
S Marchi, et al., Mitochondria-ros crosstalk in the control of cell death and aging. J Signal Transduct 2012, 329635 (2012).
A Abashidze, V Gold, Y Anavi, H Greenspan, M Weil, Involvement of IKAP in peripheral target innervation and in specific JNK and NGF signaling in developing PNS neurons. PLoS One 9, e113428 (2014).
MZ Jackson, KA Gruner, C Qin, WG Tourtellotte, A neuron autonomous role for the familial dysautonomia gene ELP1 in sympathetic and sensory target tissue innervation. Development 141, 2452–2461 (2014).
LD Johansen, et al., IKAP localizes to membrane ruffles with filamin A and regulates actin cytoskeleton organization and cell migration. J Cell Sci 121, 854–864 (2008).
D Cheishvili, et al., IKAP/Elp1 involvement in cytoskeleton regulation and implication for familial dysautonomia. Hum Mol Genet 20, 1585–1594 (2011).
S Naftelberg, et al., Phosphatidylserine ameliorates neurodegenerative symptoms and enhances axonal transport in a mouse model of familial dysautonomia. PLoS Genet 12, e1006486 (2016).
S Tielens, et al., Elongator controls cortical interneuron migration by regulating actomyosin dynamics. Cell Res 26, 1131–1148 (2016).
M Spillane, A Ketschek, TT Merianda, JL Twiss, G Gallo, Mitochondria coordinate sites of axon branching through localized intra-axonal protein synthesis. Cell Reports 5, 1564–1575 (2013).
Y Yamaguchi, M Miura, Programmed cell death in neurodevelopment. Dev Cell 32, 478–490 (2015).
MS Carroll, AS Kenny, PP Patwari, J-M Ramirez, DE Weese-Mayer, Respiratory and cardiovascular indicators of autonomic nervous system dysregulation in familial dysautonomia. Pediatr Pulmonol 47, 682–691 (2012).
DS Goldstein, B Eldadah, Y Sharabi, FB Axelrod, Cardiac sympathetic hypo-innervation in familial dysautonomia. Clin Auton Res 18, 115–119 (2008).
S Elmore, Apoptosis: A review of programmed cell death. Toxicol Pathol 35, 495–516 (2007).
EA Schon, S Przedborski, Mitochondria: The next (neurode)generation. Neuron 70, 1033–1053 (2011).
BJ Hunnicutt, M Chaverra, L George, F Lefcort, IKAP/Elp1 is required in vivo for neurogenesis and neuronal survival, but not for neural crest migration. PLoS One 7, e32050 (2012).
CL Simpson, et al., Variants of the elongator protein 3 (ELP3) gene are associated with motor neuron degeneration. Hum Mol Genet 18, 472–481 (2009).
C Creppe, et al., Elongator controls the migration and differentiation of cortical neurons through acetylation of α-tubulin. Cell 136, 551–564 (2009).
N Singh, MT Lorbeck, A Zervos, J Zimmerman, F Elefant, The histone acetyltransferase Elp3 plays in active role in the control of synaptic bouton expansion and sleep in Drosophila. J Neurochem 115, 493–504 (2010).
JA Solinger, et al., The Caenorhabditis elegans Elongator complex regulates neuronal α-tubulin acetylation. PLoS Genet 6, e1000820 (2010).
AW Harrington, et al., Recruitment of actin modifiers to TrkA endosomes governs retrograde NGF signaling and survival. Cell 146, 421–434 (2011).
S Lefler, et al., Familial dysautonomia (FD) human embryonic stem cell derived PNS neurons reveal that synaptic vesicular and neuronal transport genes are directly or indirectly affected by IKBKAP downregulation. PLoS One 10, e0138807–e0138828 (2015).
TB Kuhn, MD Brown, JR Bamburg, Rac1-dependent actin filament organization in growth cones is necessary for beta1-integrin-mediated advance but not for growth on poly-D-lysine. J Neurobiol 37, 524–540 (1998).
MC Subauste, et al., Rho family proteins modulate rapid apoptosis induced by cytotoxic T lymphocytes and Fas. J Biol Chem 275, 9725–9733 (2000).
C Meyerson, G Van Stavern, C McClelland, Leber hereditary optic neuropathy: Current perspectives. Clin Ophthalmol 9, 1165–1176 (2015).
JS Cohen, et al., ELP2 is a novel gene implicated in neurodevelopmental disabilities. Am J Med Genet A 167, 1391–1395 (2015).
LJ Strug, et al., Centrotemporal sharp wave EEG trait in rolandic epilepsy maps to Elongator Protein Complex 4 (ELP4). Eur J Hum Genet 17, 1171–1181 (2009).
H Najmabadi, et al., Deep sequencing reveals 50 novel genes for recessive cognitive disorders. Nature 478, 57–63 (2011).
M Cozzolino, MT Carrì, Mitochondrial dysfunction in ALS. Prog Neurobiol 97, 54–66 (2012).
B Literáti-Nagy, et al., Improvement of insulin sensitivity by a novel drug, BGP-15, in insulin-resistant patients: A proof of concept randomized double-blind clinical trial. Horm Metab Res 41, 374–380 (2009).
; National Research Council Guide for the Care and Use of Laboratory Animals (National Academies, 8th Ed, Washington, DC, 2011).

Information & Authors


Published in

Go to Proceedings of the National Academy of Sciences
Go to Proceedings of the National Academy of Sciences
Proceedings of the National Academy of Sciences
Vol. 114 | No. 19
May 9, 2017
PubMed: 28439028


Submission history

Published online: April 24, 2017
Published in issue: May 9, 2017


  1. familial
  2. dysautonomia
  3. Ikbkap
  4. Elp1
  5. BGP-15


We thank Marta Chaverra for technical assistance. This work was funded by NIH Neurological Disorders and Strokes Grant R01NS086796 (to F.L.) and by grants from the Dysautonomia Foundation.


This article is a PNAS Direct Submission.



Sarah B. Ohlen
Department of Cell Biology and Neuroscience, Montana State University, Bozeman, MT 59717;
Magdalena L. Russell
Department of Cell Biology and Neuroscience, Montana State University, Bozeman, MT 59717;
Michael J. Brownstein
Dysautonomia Foundation, New York, NY 10018
Frances Lefcort1 [email protected]
Department of Cell Biology and Neuroscience, Montana State University, Bozeman, MT 59717;


To whom correspondence should be addressed. Email: [email protected].
Author contributions: S.B.O. and F.L. designed research; S.B.O. and M.L.R. performed research; M.J.B. and F.L. contributed new reagents/analytic tools; S.B.O. analyzed data; and S.B.O., M.J.B., and F.L. wrote the paper.

Competing Interests

Conflict of interest statement: A patent has been filed by Montana State University (MSU) and N-Gene, with F.L. and N-Gene named as inventors. F.L. has assigned the technology to MSU, and the technology has been licensed to N-Gene by MSU. F.L. has no interest or equity in N-Gene and was not involved in MSU's license negotiations. M.J.B. is on the Board of Directors of N-Gene, Inc. and has equity in the company.

Metrics & Citations


Note: The article usage is presented with a three- to four-day delay and will update daily once available. Due to ths delay, usage data will not appear immediately following publication. Citation information is sourced from Crossref Cited-by service.

Citation statements



If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.

Cited by


    View Options

    View options

    PDF format

    Download this article as a PDF file


    Get Access

    Login options

    Check if you have access through your login credentials or your institution to get full access on this article.

    Personal login Institutional Login

    Recommend to a librarian

    Recommend PNAS to a Librarian

    Purchase options

    Purchase this article to get full access to it.

    Single Article Purchase

    BGP-15 prevents the death of neurons in a mouse model of familial dysautonomia
    Proceedings of the National Academy of Sciences
    • Vol. 114
    • No. 19
    • pp. 4839-E3872







    Share article link

    Share on social media