Abstract
Spinal muscular atrophy (SMA) is a progressive motor neuron disease caused by a deficiency of survival motor neuron (SMN) protein. In this study, we evaluated the efficacy of intermittent transient hypothermia in a mouse model of SMA. SMA mice were exposed to ice for 50 s to achieve transient hypothermia (below 25°C) daily beginning on postnatal day 1. Neonatal SMA mice (Smn−/−SMN2+/−) who received daily transient hypothermia exhibited reduced motor neuron degeneration and muscle atrophy and preserved the architecture of neuromuscular junction when compared with untreated controls at day 8 post-treatment. Daily hypothermia also prolonged the lifespan, increased body weight and improved motor coordination in SMA mice. Quantitative polymerase chain reaction and western blot analyses showed that transient hypothermia led to an increase in SMN transcript and protein levels in the spinal cord and brain. In in vitro studies using an SMN knockdown motor neuron-like cell-line, transient hypothermia increased intracellular SMN protein expression and length of neurites, confirming the direct effect of hypothermia on motor neurons. These data indicate that the efficacy of intermittent transient hypothermia in improving outcome in an SMA mouse model may be mediated, in part, via an upregulation of SMN levels in the motor neurons.
Introduction
Spinal muscular atrophy (SMA) is an inherited motor neuron disease, characterized by progressive degeneration of spinal cord motor neurons, leading to muscular atrophy, paralysis and an attenuated lifespan. SMA exhibits an autosomal recessive pattern of inheritance with an incidence of approximately 1 in 6000–10 000 newborns (1). The majority of SMA patients present with a gene mutation at the telomeric copy of the survival of motor neuron gene (SMN1). A centromeric SMN gene (SMN2) is present in all SMA patients but is unable to compensate for the gene defect in SMN1, since the primary transcript of the SMN2 gene is defectively spliced (2,3). To date, no therapies have been found to be efficacious in SMA patients.
Therapeutic hypothermia remains part of the current standard treatment protocol for patients with cardiac arrest and those receiving cardiopulmonary resuscitation (4,5). Hypothermia induces multiple effects to inhibit the cascade of neuronal death, including an inhibition of cell apoptosis, diminution of oxygen consumption, reduction of neuronal excitotoxicity and elimination of inflammatory responses (6,7). Interestingly, in an International SMA registry study, the speed of SMA-related disease progression was found to be heterogeneous between different countries (8). In this study, the time to loss of ability to ambulate in type III SMA patients was delayed in patients living in countries with lower national average annual temperatures, suggestive of a possible protective effect of cold weather on SMA. Temperature might therefore represent a disease modifiable factor for SMA, and further investigation into the potential influence of hypothermia on the pathogenesis and outcome in SMA is warranted.
To evaluate the effect of hypothermia on SMA, we induced daily, transient hypothermia in a mouse model of SMA beginning on postnatal day 1. Daily hypothermia significantly reduced the extent of motor neuron cell death and muscle atrophy. Hypothermia-treated SMA mice also exhibited an increased lifespan and improved motor function. The results of in vitro studies suggest that these beneficial effects of hypothermia are likely the result of upregulation of SMN expression in motor neurons.
Results
Transient hypothermia treatment attenuates the loss of spinal cord motor neurons in SMA mice
Following exposure to crushed ice for 50 s, the core body temperature recorded from the intraperitoneal space in neonatal SMA mice decreased from 32.8 to 19.7°C and then returned to baseline within 5 min following warming (Fig. 1). No mortality was observed during or following the 50 s of transient hypothermia in neonatal SMA mice (Supplementary Material, Fig. S1). To evaluate the histological effects of transient hypothermia on SMA mice (Smn−/−SMN2+/−), spinal cord motor neuron cell counts were performed after 7 days of treatment using choline acetyltransferase (CHAT) immunostaining. SMA mice (n = 5) receiving daily transient hypothermia, beginning on postnatal day 1, showed a non-significant trend (P = 0.08) towards a higher motor neuron cell counts in the lumbar spinal cord when compared with untreated SMA mice (Fig. 2A). Hematoxylin and eosin (H&E) staining of spinal cord sections showed that untreated SMA mice had a significantly lower number of remaining motor neurons with large (>600 μm2) and midsize (400–600 μm2) cross-sectional areas when compared with heterozygous mice (Smn+/−SMN2+/−) (Fig. 2B). Transient hypothermia treatment in SMA mice increased the number of large-sized motor neurons when compared with untreated controls (P = 0.03; Fig. 2B).
Body temperature of neonatal SMA mice with transient hypothermia treatment. The core body temperature recorded from the intraperitoneal space of neonatal mice (postnatal day 2, n = 10) is shown.
The effects of daily transient hypothermia (HT) treatment on spinal cord motor neurons in SMA mice. (A) Immunostaining with the motor neuron marker CHAT (red), and counter-staining with the nuclear marker DAPI (blue) of motor neurons in the anterior (ventral) horn of lumbar spinal cord in untreated heterozygous, untreated SMA and HT-treated SMA mice on postnatal day 8. (B) H&E staining of neurons according to their cell size: 200–400, 400–600 and >600 μm2, respectively. (C) Results from retrograde tracing experiments using intra-gastrocnemius and intra-quadriceps injections of fluorogold. The number of fluorescent-labeled motor neurons is presented. Scale bars, 100 μm; n = 5 in each group (*P < 0.05; **P ≤ 0.01; Kruskal–Wallis test and Wilcoxon rank-sum test).
Retrograde tracing experiments using intramuscular (gastrocnemius and quadriceps) injections of a fluorogold neurotracing agent were employed to examine whether the remaining motor neurons contained axons that innervated the hind limb muscles. A positive fluorescent signal in the spinal anterior horn was considered indicative of motor neurons harboring axons that innervated the limb muscles. The number of fluorescence positive motor neurons in the lumbar spinal cord was significantly higher in hypothermia-treated SMA mice versus those from untreated SMA mice (P = 0.03; Fig. 2C), suggesting that transient hypothermia had not only reduced motor neuron degeneration but also preserved functional neuromuscular integrity.
Transient hypothermia prevents muscle atrophy and preserved the architecture at the neuromuscular junction in SMA mice
The effects of daily transient hypothermia (from day 1 to day 8) on muscle atrophy were also evaluated on SMA mice. Animals were sacrificed on day 8 and histological analysis of the cross-sectional area of individual myofibers in the quadriceps of SMA mice (n = 5) treated with hypothermia revealed that they were larger compared with untreated controls (P < 0.05) (Fig. 3A). Moreover, the thickness of both the intercostal muscles and diaphragm of hypothermia-treated SMA mice was found to be significantly larger when compared with those obtained from untreated SMA mice (P = 0.04 and P = 0.03, respectively) (Fig. 3B and C).
The effects of daily transient hypothermia treatment on quadriceps muscle, intercostal muscle, diaphragm and hamstring muscle in SMA mice. (A) The frequency distribution of muscle fiber size in quadriceps muscles using H&E stain in untreated heterozygous, untreated SMA and hypothermia (HT)-treated SMA mice on postnatal day 8. (B) H&E staining of the intercostal muscles for assessment of muscle thickness. (C) H&E staining of the diaphragm for assessment of muscle thickness. (D) Immunostaining with the axonal marker neurofilament H (green) and counter-staining with the NMJ marker α-bungarotoxin (red) in the hamstring muscles of SMA mice. Scale bars, 100 μm; n = 5 in each group. (*P < 0.05; **P ≤ 0.01; Kruskal–Wallis test and Wilcoxon rank-sum test).
The effect of hypothermia on the architecture of the neuromuscular junctions (NMJs) was also examined. Histochemical staining of the hamstring muscles for neurofilament H (axonal marker) and α-bungarotoxin (NMJ marker) showed significantly less denervated NMJs in the hypothermia-treated SMA mice compared with untreated SMA mice (P = 0.03) (Fig. 3D). However, the percentages of denervated NMJs in SMA mice with and without hypothermia treatment were both higher than that in heterozygous control mice (P < 0.05).
Since SMA mice are known to exhibit cardiac atrophy (9–11), we examined whether hypothermia treatment also impacted the cardiac musculature. Cardiac weight and thickness of cardiac ventricle and septum from untreated SMA mice were lower than heart tissue obtained from their heterozygous counterparts (P < 0.01). Treatment of SMA mice with transient hypothermia did not prevent cardiac atrophy that cardiac weights and thickness of ventricle and septum from SMA mice treated with hypothermia were not significantly different from those obtained from untreated controls (Supplementary Material, Fig. S2).
Transient hypothermia treatment prolongs lifespan and improves motor function of SMA mice
We next determined whether daily transient hypothermia treatment impacted lifespan, body weight and motor function in SMA mice. To test for lifespan and body weight, untreated SMA mice (n = 43), those receiving daily transient hypothermia from postnatal day 1 (n = 41) and those receiving transient hypothermia treatment every 3 days from postnatal day 1 (HT3, n = 40) were used for comparison. While both untreated SMA and HT3-SMA mice exhibited a median lifespan of 8 days, SMA mice receiving daily transient hypothermia exhibited a median lifespan of 11 days, representing a 37.5% increase relative to controls (P < 0.001) (Fig. 4A). Similarly, treatment of SMA mice with daily transient hypothermia induced a modest increase in body weight when compared with untreated SMA control mice and HT3 mice at postnatal days 6, 7 and 8 (P < 0.05) (Fig. 4B and C). No significant differences were observed in body weight gain between heterozygous mice with and without daily hypothermia treatment.
The effects of daily transient hypothermia treatment on the survival and health of SMA mice. (A) Lifespan, (B) body weight and (C) size of mice on postnatal day 8 in untreated SMA mice (n = 43), SMA mice receiving transient hypothermia treatment every 3 days beginning on postnatal day 1 (SMA/HT3, n = 40), SMA mice receiving daily transient hypothermia treatment (SMA/HT, n = 41), untreated heterozygous mice (n = 57), heterozygous mice receiving transient hypothermia treatment every 3 days beginning on postnatal day 1 (Hetero/HT3, n = 57) and heterozygous mice receiving daily transient hypothermia treatment (Hetero/HT, n = 63). Asterisk and arrowhead in (C) indicate daily hypothermia-treated SMA mice and untreated SMA mice, respectively. (A) P < 0.001, SMA/HT mice versus untreated SMA mice; log-rank test. (B) *P < 0.05 and **P ≤ 0.01, SMA/HT mice versus untreated SMA mice; #P < 0.05 and ##P ≤ 0.01, SMA/HT mice versus SMA/HT3 mice; one-way ANOVA with LSD post-hoc comparisons.
To test for the motor function, untreated SMA mice (n = 39), those receiving daily transient hypothermia from postnatal day 1 (n = 39), and those receiving transient hypothermia treatment every 3 days from postnatal day 1 (HT3, n = 32) were used for comparison. Assessment of motor function using the turnover test to evaluate righting reflex failed to reveal significant differences between groups on postnatal day 6 (Fig. 5A). However, on postnatal day 8, untreated SMA mice exhibited significantly longer latencies in turnover time when compared with heterozygous mice (P < 0.001). Treatment with daily transient hypothermia partially prevented this behavioral deficit (P < 0.001). The tube test analyzing hind limb strength performed on postnatal days 6 and 8 in untreated SMA mice revealed poorer performances than their age-matched heterozygous counterparts (P < 0.01) (Fig. 5B). Although treatment of SMA mice with daily transient hypothermia significantly improved hind limb strength on postnatal day 8 (P < 0.01), values did not return to those observed in heterozygous mice. Although tilting scores obtained from the negative geotaxis test did not differ between groups on postnatal day 6 (Fig. 5C), by postnatal day 8, the tilting scores of untreated SMA mice were significantly lower than their age-matched heterozygous counterparts (P < 0.01). Treatment of SMA mice with daily transient hypothermia significantly increased their tilting scores on postnatal day 8 (P < 0.01), but values did not reach those observed in heterozygous mice.
The effects of daily transient hypothermia treatment on motor function in SMA mice. (A) Turnover time, (B) tube test scores and (C) the tilting scores in untreated SMA mice (n = 39), SMA mice treated with transient hypothermia treatment every 3 days beginning on postnatal day 1 (SMA/HT3, n = 32), SMA mice treated with daily transient hypothermia (SMA/HT, n = 39), untreated heterozygous mice (n = 61), heterozygous mice receiving transient hypothermia treatment every 3 days beginning on postnatal day 1 (Hetero/HT3, n = 61) and heterozygous mice receiving daily transient hypothermia treatment (Hetero/HT, n = 76). *P < 0.05; **P ≤ 0.01; ***P ≤ 0.001; one-way ANOVA with LSD post-hoc comparisons.
Transient hypothermia increases tissue levels of SMN protein and transcripts in SMA mouse model
Tissue levels of SMN protein in the spinal cord and brain of SMA mice without treatment, those receiving transient hypothermia treatment daily, and those receiving transient hypothermia at 3 day intervals beginning on postnatal day 1 (n = 5/group) were assessed on postnatal day 8. Levels of SMN protein in SMA mice receiving daily transient hypothermia were significantly (2- to 3-fold) higher in both spinal cord and brain (Fig. 6A and B) when compared with untreated SMA controls (P < 0.01). However, the absolute levels of SMA protein in spinal cord remained significantly lower when compared with those obtained from heterozygous mice (P < 0.001). Although previous reports have suggested that the therapeutic mechanisms of hypothermia in the setting of cardiac arrest include, in part, an up-regulation of the anti-apoptotic factor Bcl-xl and concomitant down-regulation of the pro-apoptotic factor, Bax (6), no differences were observed between groups in tissue levels of Bcl-xl and Bax protein in the spinal cord or brain on postnatal day 8 in this study. Additionally, we assessed tissue levels of SMN protein in skeletal and cardiac muscles on postnatal day 8 following daily transient hypothermia treatment initiated on postnatal day 1. No differences were observed in SMN protein levels in SMA mice with and without hypothermia treatment (Supplementary Material, Fig. S3), and the levels were significantly lower than those in skeletal and cardiac muscle obtained from heterozygous mice (P < 0.01).
Analysis of protein and transcript expression in untreated SMA mice, SMA mice receiving transient hypothermia treatment every 3 days beginning on postnatal day 1 (HT3), SMA mice receiving daily transient hypothermia treatment (HT) and untreated heterozygous mice on postnatal day 8. (A) Western blots of homogenized spinal cords for SMN, Bax, Bcl-xl and β-actin. Analysis of the ratio of SMN, Bax and Bcl-xl protein levels to β-actin is shown. (B) Western blots of homogenized brain for SMN, Bax, Bcl-xl and β-actin. Analysis of the ratio of SMN, Bax and Bcl-xl protein levels to β-actin is shown. (C and D) Quantitative PCR of homogenized spinal cord (C) and brain (D) with primer pairs for full-length SMN, truncated SMN and GAPDH. Analysis of the ratio of full-length SMN levels to GAPDH and truncated SMN levels to GAPDH is shown. N = no treatment; n = 5 for each group. (E) Western blots of homogenized spinal cords and brain for p38 MAPK, HSP70, FGF21 and β-actin. Analysis of the ratio of p38 MAPK, HSP70 and FGF21 protein levels to β-actin is shown. n = 4 for each group (*P < 0.05; **P ≤ 0.01; Kruskal–Wallis test and Wilcoxon rank-sum test).
To clarify the molecular mechanisms underlying the observed increase in SMN protein levels following transient hypothermia treatment, quantitative real-time polymerase chain reaction (qRT-PCR) were performed on tissue lysates. qRT-PCR analysis of lysates from the spinal cord and brain of SMA mice receiving transient hypothermia revealed a higher level of full-length SMN gene transcript expression (P < 0.01) with similar truncated SMN expression when compared with tissue obtained from untreated SMA mice (Fig. 6C and D).
Tissue levels of two cold-inducible protein, phospho-p38 mitogen-activated protein kinase (p38 MAPK) and fibroblast growth factor 21 (FGF21) and a warm-inducible protein, heat shock protein 70 (HSP70) in the spinal cord and brain of SMA mice were also assessed on postnatal day 8. Levels of p38 MAPK protein in SMA mice receiving daily transient hypothermia were significantly (1.4- to 2-fold) higher in both spinal cord and brain (Fig. 6E) when compared with untreated SMA controls (P < 0.05). No differences were observed in p38 MAPK protein levels in muscle and heart tissues of SMA mice with and without hypothermia treatment (Supplementary Material, Fig. S3). In contrast, there were no differences between groups in tissue levels of HSP70 and FGF21 protein in the spinal cord, brain, muscle or heart on postnatal day 8 (Fig. 6E and Supplementary Material, Fig. S3).
In vitro studies: transient cold stimulation increases levels of SMN protein and length of neurites in an SMA-like motor neuron cell-line
To confirm that the motor neuron was a therapeutic target of transient hypothermia, we treated a motor neuron cell-line (NSC34) with daily transient cold stimulation. Transient cold exposure for 1 or 5 min increased the intracellular SMN protein levels by 1.3-fold (P = 0.04) and p38 MAPK protein levels by 1.7- to 2-fold (P = 0.04) in NSC34 (Fig. 7A). In contrast, cold stimulation did not significantly increase the SMN and p38 MAPK protein levels in a muscle cell-line (C2C12) (Fig. 7B). On the other hand, transient cold exposure did not change the intracellular HSP70 and FGF21 protein levels in the NSC34 cells, but increased the HSP70 protein levels (P = 0.04) and reduced FGF21 protein levels (P = 0.04) in the C2C12 cells (Fig. 7A and B).
The effects of transient cold exposure on protein expression and neurite length in a motor neuron cell-line and a muscle cell-line. (A and B) Western blots of a motor neuron cell-line (NSC34, A) and a muscle cell-line (C2C12, B) with or without daily 1-min or 5-min cold exposure for 7 days. Analysis of the ratio of SMN, p38 MAPK, HSP70 and FGF21 protein levels to β-actin is shown. (C) Western blots of an SMA-like motor neuron cell-line (NSC34-kd) for SMN and β-actin with different periods of Dox treatment. Analysis of the ratio of SMN protein levels to β-actin is shown. (D) Western blots for SMN, Bax, Bcl-xl and β-actin in the cultured motor neuron cell-lines, including naïve NSC34 without treatment (NSC34), NSC34-kd without treatment (NSC34-kd), NSC34-kd with Dox treatment (NSC34-kd/Dox), NSC34-kd with Dox treatment and daily 1-min cold exposure for 7 days (NSC34-kd/Dox/1 min), NSC34-kd with Dox treatment and daily 5-min cold exposure for 7 days (NSC34-kd/Dox/5 min). Analysis of the ratio of SMN, Bax and Bcl-xl protein levels to β-actin is shown. (E) Immunostaining for SMN (green) and β-III tubulin (red), and counter-stained with DAPI (blue) in NSC34-kd cells, NSC34-kd/Dox cells and NSC34-kd/Dox/5 min cells. Scale bars, 20 μm. (F) The length of neurites was measured and compared. n = 3 in each group (*P < 0.05; Kruskal–Wallis test and Wilcoxon rank-sum test).
We further treated an inducible SMA-like motor neuron cell-line (NSC34-kd) with daily transient cold stimulation. To verify the efficacy of the SMN knockdown, we first treated the NSC34-kd with Dox to induce SMN knock-down for varying periods from 1 to 7 days. Intracellular levels of SMN protein of the NSC34-kd were then measured. With Dox induction, the intracellular SMN protein levels were reduced in a time-dependent manner (Fig. 7C). After 7 days of Dox induction, SMN protein levels decreased to 17% of that in the original NSC34-kd cell line (P < 0.05), similar to the SMN expression levels noted in SMA patients (12). After confirming that we had induced SMN knock-down using 7-day Dox treatment, we then treated NSC34-kd (under Dox induction) with daily transient cold exposure for 1 or 5 min duration over 7 days. Transient cold exposure for 1 or 5 min increased the intracellular SMN protein levels by 2-fold (P = 0.05) in NSC34-kd with Dox induction (Fig. 7D). However, SMN protein levels remained significantly lower than those measured in NSC34-kd cells without Dox induction or in naïve NSC34 cells (P < 0.05). No significant differences were observed in levels of Bcl-xl and Bax protein between groups.
NSC34-kd cells, NSC34-kd cells with Dox treatment and NSC34-kd cells with Dox treatment and daily transient cold exposure for 5 min duration over 7 days (n = 3 in each group) were co-stained for β-III tubulin and SMN to measure the neurite length. Dox-treated NSC34-kd cells had significantly shorter length of neurites (P = 0.01) with lower intracellular SMN immunofluorescent signals than NSC34-kd cells without treatment (Fig. 7E and F). Transient cold exposure in NSC34-kd cells with Dox treatment increased the neurite length when compared with controls (P = 0.03; Fig. 7E and F), but values did not return to those observed in NSC34-kd cells without SMN knockdown.
Different intensity of transient hypothermia treatment on lifespan and motor function of SMA mice
We next determined the effects of different intensity of transient hypothermia on lifespan, body weight and motor function in SMA mice. We compared the lifespan and body weight of untreated SMA mice (n = 20) and those receiving transient hypothermia for 10 s (HT10, n = 18), 30 s (HT30, n = 14) and 50 s (HT50, n = 18) per day from postnatal day 1. The median lifespans of untreated and HT10 SMA mice were similar (8 days, P = 0.75). SMA mice receiving transient hypothermia for 30 and 50 s per day exhibited median lifespans of 9.5 and 10.5 days, respectively, which were both significantly higher than untreated control (P = 0.049 and P < 0.001) (Fig. 8A). In addition, treatment of SMA mice with transient hypothermia for 30 and 50 s per day induced a modest increase in body weight when compared with untreated SMA control mice and HT3-SMA mice at postnatal days 6 and 7 (P < 0.05) (Fig. 8B).
The effects of different intensity of daily transient hypothermia treatment on the survival and motor function of SMA mice. (A) Lifespan, and (B) body weight in untreated SMA mice (n = 20), SMA mice receiving hypothermia treatment for 10 s (SMA/HT10, n = 18), 30 s (SMA/HT30, n = 14) and 50 s (SMA/HT50, n = 18) per day beginning on postnatal day 1, untreated heterozygous mice (n = 27) and heterozygous mice receiving hypothermia treatment for 10 s (Hetero/HT10, n = 10), 30 s (Hetero/HT30, n = 18) and 50 s (Hetero/HT50, n = 31) per day beginning on postnatal day 1. (C) Turnover time, (D) tube test scores and (E) the tilting scores in untreated SMA mice (n = 11), SMA mice receiving hypothermia treatment for 10 s (n = 12), 30 s (n = 13) and 50 s (n = 16) per day beginning on postnatal day 1, untreated heterozygous mice (n = 20) and heterozygous mice receiving hypothermia treatment for 10 s (n = 10), 30 s (n = 22) and 50 s (n = 27) per day beginning on postnatal day 1. (A) P < 0.001, SMA/HT50 mice versus untreated SMA mice, SMA/HT10 mice versus SMA/HT30 and SMA/HT50 mice; P < 0.05, SMA/HT30 mice versus untreated SMA mice; log-rank test. (B) *P < 0.05, SMA/HT30 and SMA/HT50 mice versus untreated mice; #P < 0.05, SMA/HT30 mice versus SMA/H10 mice; †P < 0.05, SMA/HT5 P < 0.05, SMA/HT30 mice versus SMA/H10 mice 0 mice versus SMA/H10 mice. (C–E) *P < 0.05; **P ≤ 0.01; ***P ≤ 0.001; one-way ANOVA with LSD post-hoc comparisons.
To test for the motor function, untreated SMA mice (n = 11) and those receiving transient hypothermia for 10 s (n = 12), 30 s (n = 13) and 50 s (n = 16) per day from postnatal day 1 were used for comparison. Similarly, on postnatal day 8, assessment of motor function using the turnover test, tube test and negative geotaxis test all showed that the behavioral performances were worse in untreated and HT10 SMA mice when compared with heterozygous mice with or without hypothermia treatment (P < 0.001). Treatment with transient hypothermia for 30 or 50 s partially prevented this behavioral deficit (P < 0.05 and P < 0.01, respectively), but the performances did not reach those observed in heterozygous mice (P < 0.05).
Discussion
According to an International SMA registry study (8), the period from type III SMA diagnosis to the point at which patients lose their ability to ambulate varies markedly between different countries. This period is notably shorter for patients living in countries with higher national average annual temperatures (Ukraine, Hungary, Argentina and Serbia) versus those living in countries with lower temperatures (Switzerland, UK and Germany/Austria) (Supplementary Material, Fig. S4A) (13), suggestive of a negative correlation between the time from diagnosis to loss of ambulation and national average annual temperatures (Supplementary Material, Fig. S4B). These data suggest that temperature might be a disease modifiable factor and that hypothermia treatment may be beneficial for SMA. Here, we report that daily transient hypothermia treatment in an SMA mouse model reduced spinal motor neuron degeneration, increased the size of hind limb myofibers and muscle thickness in diaphragmatic and intercostal muscles and partially preserved the architecture of NMJ. These beneficial effects were associated with an improvement in the overall health of the mice as illustrated by increased body weight, improved motor function and a longer lifespan. We also demonstrate, for the first time, that daily transient hypothermia increased the levels of SMN transcripts and protein in spinal cord and brain of treated SMA mice.
Therapeutic hypothermia has been reported to improve neurological outcome in patients with anoxic brain injury and hypoxic ischemic neonatal encephalopathy primarily through multiple actions involving a blockade of neuronal death pathways (6,7). Among these, upregulation of the anti-apoptotic factor Bcl-xl, with concomitant downregulation of the pro- apoptotic factor Bax, is known to play a role in inhibiting neuronal apoptosis (6). Previous studies have shown that Bcl-xl over-expression or Bax knockout in SMA mice reduced motor neuron degeneration, preserved motor function and improved lifespan (14,15). Although we hypothesized that changes in Bcl-xl and Bax may underlie the beneficial effects of transient hypothermia in SMA mice, levels of Bcl-xl and Bax did not change in SMA mice receiving daily transient hypothermia. In addition, transient cold stimulation failed to alter levels of Bcl-xl and Bax in an SMA-like motor neuron cell-line. Although transient hypothermia treatment increased levels of SMN protein and transcripts in both spinal cord and brain, SMN protein levels remain unaltered in cardiac and hind limb muscles following hypothermia treatment and in a muscle cell-line following transient cold stimulation. Since SMA causes motor neuron degeneration in anterior horn of spinal cord, but not motor cortex of the brain, the higher SMN level in the brain tissue of hypothermia-treated SMA mice than untreated SMA mice is likely resulted from upregulation of SMN rather than only reduction in neuronal degeneration. Therefore, it remains possible that the beneficial effects of transient hypothermia in SMA mice were mediated, in part, via an augmentation of SMN in spinal cord, particularly in light of previous studies demonstrating that SMN expression levels are inversely correlated with disease severity of SMA (16,17).
Transient daily cold stimulation increased the intracellular SMN protein levels in a motor neuron cell-line, but not in a muscle cell-line. In order to model the in vivo conditions of SMA mice and further evaluate the possible cellular/molecular mechanisms underlying therapeutic benefit of transient hypothermia, we treated an SMA-like motor neuron cell line with daily transient cold stimulation. In motor neuronal culture cells associated with 80% SMN knockdown, hypothermia increased intracellular levels of SMN protein and the length of neurites. These findings confirm that the motor neurons are one of the primary targets responsive to transient hypothermia.
The protein levels of a cold-inducible protein, p38 MAPK (18), were increased in spinal cord and brain of SMA mice after hypothermia treatment and in a motor neuron cell-line after cold stimulation. Notably, a previous study has demonstrated that p38 MAPK activation increased SMN protein expression via stabilization of the SMN transcript (19). It remains possible that hypothermia upregulated p38 MAPK in spinal motor neurons, leading to increase in SMN protein expression via stabilization of the SMN transcript and subsequently partially prevented degeneration of motor neurons and muscles and preserved motor functions in SMA mice. The levels of p38 MAPK were not increased in muscle tissues of SMA mice and a muscle cell-line after cold stimulation. It may be the reason why the levels of SMN protein were not altered in muscle tissues and a muscle cell-line after cold exposure. Another cold-inducible protein, FGF21, and a rewarm associated protein, HSP70, are potentially neuroprotective (20–22). Previous study also showed that exogenous delivery of HSP70 benefited a mouse model of a motor neuron disease, amyotrophic lateral sclerosis (23). However, both levels of FGF21 and HSP70 were not altered in spinal cord, brain, muscle and heart tissues of SMA mice after transient cold stimulation. FGF21 and HSP70 were thus not likely play a major role on beneficial effects after hypothermia treatment in SMA mice, although we have noticed that the levels of HSP70 were increased after transient cold exposure in a muscle cell-line.
Using different intensity of cold stimulation, we demonstrated that there was dose-dependent benefit in survival and motor function in SMA mice. However, the duration of cold stimulation required at least 30 s with the lowest core body temperature of neonatal mice at about 25°C to get the beneficial effects. In humans, these extremely low body temperatures are known to cause fetal arrhythmia and multi-organ injury (24). According to current treatment guidelines for patients receiving therapeutic hypothermia for anoxic brain injury, the goal of hypothermia is to reduce core body temperature to 32–34°C (25). Therefore, the severe hypothermic treatment protocol used in this study cannot be directly translated to clinical practice. It remains possible, however, that long-term daily transient moderate hypothermia (35–36°C) may be beneficial in upregulating SMN expression and attenuating motor neuron degeneration in SMA patients. The time to loss of ability to ambulate in type III SMA patients was thus delayed in patients living in countries with lower national average annual temperatures in an International SMA registry study (8). However, the time of diagnosis in one country can be completely different than the other based on many situations such as doctor visits, family issues or even medical systems. Future large registration study with emphasis on the disease progression speed in SMA patients according to the average annual temperature is mandatory.
Hypothermia therapy remains an attractive treatment strategy for various neurological diseases. In addition to anoxic encephalopathy, for which therapeutic hypothermia has become standard of care (4,5,25), hypothermia treatment has also been shown to improve functional outcome and reduced histopathological damage in the experimental rodent models of traumatic brain injury and spinal cord injury, although its efficacy in humans remains controversial (26–28). Apart from a reduction of metabolic rate and anti-apoptotic effects, the basic mechanisms underlying hypothermic protection that have been suggested include anti-inflammatory effects, prevention of excitotoxicity, reduction of free radical production and suppression of epileptogenesis (6,7,26). Hypothermia has also been shown to improve survival and mobility in a Drosophila model of motor neuron disease, hereditary spastic paraplegia (23). It was suggested that hypothermia might be beneficial to synaptic function via a reduction in general neuronal activity or metabolic load (29). In this study, our results suggest that the beneficial effects of hypothermia in SMA mice are mediated, in part, via intracellular SMN augmentation. Although transient hypothermia may hold promise as a therapeutic approach for motor neuron diseases or other neurodegenerative diseases, recent studies have suggested that hypothermia may also contribute to the exacerbation of Alzheimer's disease, possibly resulting from enhanced formation of beta-amyloid plaques and tau protein hyperphosphorylation (30). The complex actions of therapeutic hypothermia in the central nervous system under different pathological conditions warrant further investigation.
Materials and Methods
Mice
A mouse model of SMA was produced via deletion of exon 7 of the Smn gene and knock-in of the human SMN2 (Smn−/−SMN2+/−) (31). We were able to generate a variant that presents with more severe disease symptomatology through back-crossing to obtain a more homogenous genetic background. This model of severe SMA harbors two copies of SMN2 transgene (Smn−/−SMN2+/−) and animals show an average lifespan of 8.7 days (32). Mouse genotypes were confirmed by PCR analysis as previously described (33). All procedures were approved by the Institutional Animal Care and Use Committee (Protocol # 20140324). Mice were under the care of the laboratory animal center of the National Taiwan University College of Medicine and were supplied with sterile water and rodent pellets ad libitum.
Hypothermia
For induction of hypothermia in the mouse model of SMA, neonatal mice were transiently exposed to crushed ice to lower the core body temperature. Since mortality rates of neonatal mice increased if the duration of ice exposure exceeded 60 s (Supplementary Material, Fig. S1), each neonatal mouse received transient hypothermia for the maximum of 50 s per day to achieve a lowered core body temperature. After removing mice from the crushed ice, animals were warmed using a heating lamp for 5 min. The core body temperature of neonatal mice was measured in the intraperitoneal space using a T-type thermocouple (5TC-TT-T-36-36; Omega Engineering, Stamford, CT) with an NI 9211 data acquisition system (National Instruments, Austin, TX).
Immunofluorescence study
The lumbar spinal cord and hind limb hamstring muscle were removed at postnatal day 8, placed in 4% paraformaldehyde at room temperature for 3 h, followed by incubation with 15% sucrose for 1 day and 30% sucrose at 4°C for another 3 days, and then rapidly frozen in liquid nitrogen-cooled isopentane. Frozen serial sections of spinal cord and muscles were cut to a thickness of 10 and 15 μm, respectively. Tissue sections were blocked with 5% serum. In addition, cultured cells were fixed with 4% paraformaldehyde at room temperature for 30 min and then blocked with 3% fetal bovine serum. Both tissue sections and cultured cells were incubated overnight at 4°C with the primary antibody, anti-CHAT (1:100; Millipore, Billerica, MA), SMN (1:100; BD Biosciences, San Diego, CA), anti-neurofilament 200 kDa (1:200; Millipore), α-bungarotoxin Alexa Fluor 555 conjugate (1:1000; Molecular Probes, Eugene, OR) or β-III tubulin (1:100; Cell Signaling, Beverly, MA). Samples were washed with TBS-T (10 mm Tris–HCl, pH 8.0, 150 mm NaCl and 0.1% Tween 20) and incubated for 2 h at room temperature with the fluorescence-conjugated secondary antibody (1:200; Molecular Probes, Eugene, OR). DAPI (4′,6-diamidino-2-phenylindole, Sigma Aldrich, St Louis, MO) was used for counter-staining the cell nucleus. Samples were examined using a fluorescence microscope (Zeiss AXIO Imager A1, Carl Zeiss, Gottingen, Germany) or laser-scanning confocal microscope (Leica TCS SP2 Spectral Confocal System, Leica Microsystems, Mannheim, Germany).
Hematoxylin and eosin staining
Frozen tissue sections from spinal cord (10 μm thickness), quadriceps muscles (5 μm), diaphragm (10 μm) and ribs (10 μm) were stained with H&E. Neurons with a size greater than 200 µm2 located in the anterior (ventral) horn of the spinal cord were analyzed. The number of neurons (only neurons showing nuclei) were counted and grouped according to their cell size: 200–400, 400–600 and >600 μm2, respectively. The size of myocytes from quadriceps was analyzed and the number of myocytes was counted according to their cross-sectional size. The thickness of diaphragmatic and intercostal muscles was also measured.
Retrograde tracing
For retrograde tracing in motor neurons (34), 4 μl of fluorogold (Fluorochrome, Denver, CO) was injected into bilateral gastrocnemius and quadriceps muscles of mice on postnatal day 5 using a 30-gauge needle at a rate of 1 μl/min. Mice were sacrificed 3 days post-injection and their spinal cord removed. After paraformaldehyde fixation, frozen serial 10-μm thick sections of whole lumbar spinal cord were cut and processed for histological analysis.
Western blotting
Western blot analysis was conducted as previously described (35). Briefly, mouse tissues were homogenized in ice-cold modified RIPA buffer (50 mm Tris–HCl, pH 7.4, 1% NP-40, 0.25% deoxycholic acid, 0.15 M NaCl, 1 mm EDTA, 1 mm PMSF/NaF/sodium orthovanadate and protease inhibitors cocktail). The supernatants were collected after centrifugation, separated on a 12% sodium dodecyl sulfate-polyacrylamide gel, and electro-transferred to a polyvinylidene fluoride membrane. The membranes were incubated in blocking solution for 1 h at room temperature, and then overnight at 4°C with the following primary antibodies: anti-SMN (1:500), anti-Bax (1:200, Santa Cruz Biotech, Santa Cruz, CA), anti-Bcl-xl (1:1000, Cell Signaling), anti-phostpo-p38 MAPK (1:1000, Cell Signaling), anti-HSP70 (1:1000, Abcam, Cambridge, UK), anti-FGF21 (1:1000, Abcam) and anti-β-actin (1:5000, Sigma Aldrich). The membranes were washed and incubated in blocking solution with the appropriate HRP-conjugated secondary antibodies for 2 h at room temperature (1:5000; GeneTex, Irvine, CA). The bands were visualized directly using a UV transilluminator (BioDoc-It® Imaging System, UVP Inc., Upland, CA). Semi-quantitative evaluation of the bands was performed by densitometric analysis using ImageJ, and the protein expression levels were normalized to β-actin.
Quantitative real-time PCR
Total RNA from mouse tissues were extracted with TRIZOL (Invitrogen, Carlsbad, CA) and subjected to reverse transcription using SuperScript III (Invitrogen) according to the Manufacturer's instructions. Quantitative real-time PCR was performed as previously described (36). Quantitative PCR reactions were run in triplicate for each sample on an IQ™5 Multicolor Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA). The SMN primer sets (full-length SMN and exon 7 truncated SMN) and an endogenous control (glyceraldehyde 3-phosphate dehydrogenase; GAPDH) primer set were used on each sample.
Motor function tests
Three behavioral tests were used to evaluate neurobehavioral motor function in SMA mice. In the turnover test, each mouse was placed on its back and the time required to right itself and place all four paws on the ground was recorded (cutoff time was 60 s) (32). Hind limb strength was determined using the tube test (37). Animals were suspended by their hind limbs on the lip of a 50-ml tube. The responses of the mice were scored as follows: 4 = normal hind limb separation; 3 = hind limbs close together; 2 = hind limbs close to each other and often touching; 1 = hind limbs almost always clasping and the tail raised; 0 = hind limbs constantly closed with the tail lowered. For the negative geotaxis test (32), each animal was placed on a 45° incline with its head pointing downward. The responses of the mice were scored as follows: 0 = subject slipped down immediately; 1 = subject held onto the incline with its head still pointing downward; 2 = subject turned around on the incline by 90°; 3 = subject turned around by 180° with its head pointing upward and 4 = subject turned around and climbed up the incline.
Cell culture
A mouse neuroblastoma-spinal cord hybrid cell-line (NSC34) expressing properties of motor neurons, Dox-inducible SMN knock-down NSC34 cells (NSC34-kd) and a mouse myoblast cell-line (C2C12) expressing properties of skeletal muscles were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum, 1 mm glutamine and 100 IU/ml penicillin and 100 mg/ml streptomycin, incubated at 37°C in a 5% CO2 humidified atmosphere (38–40). To reproduce daily transient hypothermia-like condition, culture medium of NSC34, NSC34-kd with or without Dox (Sigma Aldrich) treatment and C2C12 was replaced by 4°C phosphate-buffered saline for 1 or 5 min, and then changed back to 37°C culture medium once daily. After 7 days, one set of cells was detached by scraping and lysates were subjected for western blotting analysis as described above. Another set of cells was subjected for immunofluorescence study as described above.
Statistical analysis
Values are expressed as mean ± SEM. Statistical significance was analyzed using the Kruskal–Wallis test or a Wilcoxon rank-sum test, except for comparisons in body weight and motor function tests in which one-way analysis of variance (ANOVA) followed by LSD post-hoc comparisons was used. A Kaplan–Meier analysis was used for estimation of survival curves and a log-rank test was used to determine differences between the two survival curves. Two-tailed ‘P’ values of less than 0.05 were considered statistically significant. SPSS software was used for statistical analyses.
Funding
This work was supported by the National Science Council (grant numbers 102-2314-B002-065-MY3 and 104-2314-B-002-051-MY3).
Acknowledgements
We thank the Department of Medical Research, National Taiwan University Hospital, Taiwan, for excellent technical assistance and equipment support.
Conflict of Interest statement. The authors declare no conflict of interest.
References
