Effect of temperature on spinal cord regeneration in the weakly electric fish, Apteronotus leptorhynchus
- First Online:
- Cite this article as:
- Sîrbulescu, R.F. & Zupanc, G.K.H. J Comp Physiol A (2010) 196: 359. doi:10.1007/s00359-010-0521-9
- 111 Views
Temperature manipulation has been shown to significantly affect recovery after spinal cord injury in various mammalian model systems. Little has been known thus far about the impact of temperature on structural and functional recovery after central nervous system lesions in regeneration-competent, poikilotherm organisms. In the present study, we addressed this aspect using an established model of adult spinal cord regeneration, the weakly electric teleost fish Apteronotusleptorhynchus. We observed an overall beneficial effect of increased temperature on both structural and behavioral recovery after amputation of the caudal spinal cord. Fish kept at 30°C recovered the amplitude of the electric organ discharge at more than twice the rate observed in fish kept at 22°C, within the first 20 days post-injury. This improved recovery was supported by increased cell proliferation and decreased apoptosis levels in fish kept at 30°C. The high temperature appeared to have a direct inhibitory effect on apoptosis and to lead to a compression of the duration of the wave of post-lesion apoptosis. The latter effect was presumably induced through the acceleration of the metabolic rate, a phenomenon also supported by the observation that re-growth of the tail was significantly increased in fish kept at 30°C.
KeywordsApoptosisCell proliferationElectric organ dischargeGliosisSpinal cord injury
Bovine serum albumin
Central nervous system
Electric organ discharge
Glial fibrillary acidic protein
The central nervous system (CNS) of teleost fish is capable of a remarkable degree of regeneration following injury (for reviews, see Zupanc 2001, 2008, 2009; Becker and Becker 2008). This contrasts with the largely irreversible tissue damage observed in mammals after CNS lesions (Profyris et al. 2004; Johansson 2007; Bramlett and Dietrich 2007). In the latter, spinal cord injury is initially characterized by hemorrhage and necrosis, while at later stages the hemorrhagic fronts enlarge, leading to edema and local ischemia as blood vessels become clotted (Profyris et al. 2004; Bramlett and Dietrich 2007). One therapeutic approach to alleviate the immediate effects of CNS injury in mammals involves the application of experimentally induced hypothermia (Martinez-Arizala and Green 1992; Dietrich et al. 2009; Marion and Bullock 2009). Various physiological effects of cooling have been observed, including a reduction in polymorphonuclear leukocyte invasion (Chatzipanteli et al. 2000), reduced edema (Yu et al. 1999), reduced histopathological damage, and improved locomotor function (Yu et al. 2000), or reduced vascular dysfunction (Wei et al. 2009). However, the use of this therapeutic measure remains controversial, as other studies have failed to find significant improvements after induced hypothermia (for reviews, see Hayashi 2009; MacLellan et al. 2009; Marion and Bullock 2009).
Very few studies have addressed the effect of temperature manipulation in regeneration-competent organisms. In line with results from studies in mammals, in goldfish low temperature had a protective effect on severed axons of Mauthner neurons. At 10°C, the axons survived for over 5 months, but at 30°C they did so for only 1 month (Blundon et al. 1990). Interestingly, in lampreys, recovery after spinal cord injury was significantly better at higher temperatures, with animals exhibiting less dysfunctional motor patterns (Cohen et al. 1999). Also, lampreys kept at 22–24°C showed less serotonin depletion close to the spinal cord lesion site as compared to lampreys kept at 11–13°C (Cohen et al. 2005).
In the present study, we aimed to explore the effect of temperature manipulation in a well-established model of spinal cord regeneration, the brown ghost knife fish, Apteronotus leptorhynchus. This teleostean species is capable of regenerating the caudal part of its tail, including spinal cord tissue, after amputation (Sîrbulescu et al. 2009). Moreover, because the electric organ is formed by enlarged axonal terminals of spinal cord motoneurons (Bennett 1971), functional recovery after spinal cord injury can be monitored by recording the amplitude of the electric organ discharge (EOD; Sîrbulescu et al. 2009). By combining this behavioral assay with neuroanatomical analysis, we have examined the effect of temperature on both structural and functional regeneration after spinal cord injury in A. leptorhynchus.
Materials and methods
Fish were purchased from a tropical fish importer and kept in individual aquaria at a water conductivity of approximately 200 μS/cm and a pH of approximately 6.8. A total of 22 adult fish were used in this study. They were selected such that they exhibited similar sizes (mean total length 15.1 ± 1.5 cm SD, mean body mass 7.3 ± 2 g SD). Seven fish were males and 5 females, as revealed by post-mortem gonadal inspection, while 10 fish, used for behavioral analysis only, were not sexed. In the 12 fish analyzed, the gonadosomatic index, determined as fresh weight of gonads divided by body mass, was on average 0.0027 (±0.0007 SD) in the males and 0.0449 (±0.0165 SD) in the females. All experiments were performed in accordance with the relevant German law, the Deutsches Tierschutzgesetz, of 1998. All efforts were made to reduce the number of animals used, and to minimize animal suffering.
Behavioral recovery and EOD recording
The fish (n = 10) were transferred into individual aquaria and kept at either low (approximately 22°C) or high (approximately 30°C) temperature for 3 days prior to the start of the experiment. The EOD was then recorded for 7 days to establish baseline levels of EOD amplitude and frequency. On day 8, the fish were anesthetized with urethane (Sigma, Taufkirchen, Germany), and after application of 2% lidocaine (Euro OTC Pharma, Hannover, Germany) as a local anesthetic, 1 cm of the tail, measured rostrally from the white caudal ring, was amputated. After recovery from the anesthesia, the fish were returned to their home aquarium. Over the following 38 days, the EOD was recorded daily. To minimize the impact upon the amplitude of the recorded signal, the conductivity of the water in each aquarium was kept within a standard deviation of ±2 μS/cm over the entire recording interval. The EOD of the fish was differentially recorded and digitized as described (Zupanc et al. 2006a). Briefly, recordings were made using pairs of stainless steel electrodes mounted on the inside of an opaque plastic tube. The fish was restrained in this tube by closing the ends with netting. The signal was AC amplified 30× (low-pass filter: none, high-pass filter: 200 Hz) by a CED 1902 amplifier (Cambridge Electronic Design, Cambridge, England). The analog signal was converted into a digital one at a sampling rate of 50 kHz using CED Micro1401 mkII analog-to-digital converters (Cambridge Electronic Design), and analyzed using the software Spike 2 Version 5.00 (Cambridge Electronic Design). Recordings were performed daily for 5–10 min. From each recording, a 10-s sequence of maximal amplitude was selected and analyzed using Matlab (version 7.3; The MathWorks, Natick, MA, USA). The waveform was band-filtered around the fundamental frequency using Fast Fourier Transformation, and the maximal amplitude of the resulting signal was determined.
Under general anesthesia with urethane (Sigma) and local anesthesia with 2% lidocaine (Euro OTC Pharma), 1 cm of the tail was amputated (n = 12 fish). After an 18-h or 10-day survival period (n = 6 fish per survival time), the animals were deeply anesthetized using ethyl 3-aminobenzoate methanesulfonate (MS-222; Sigma) and intracardially perfused. The caudalmost 0.5 cm of the tail was removed, post-fixed in 2% paraformaldehyde for 2 h, and cryoprotected as described (Zupanc et al. 2005). Transverse sections were cut serially at a thickness of 16 μm and mounted onto SuperFrost Plus Gold slides (Menzel-Gläser, Braunschweig, Germany).
For immunohistochemical detection of antigens, sections were dried in a desiccator for 90 min at room temperature (RT) and rehydrated through three changes of 0.1 M Tris-buffered saline (TBS; pH 7.5). To permeabilize the tissue and block unspecific binding sites, the sections were treated for 1 h at RT with blocking solution [TBS containing 1% bovine serum albumin (BSA; fraction 5, pH 5.2; AppliChem, Darmstadt, Germany), 1% teleostean gelatin (Sigma), 3% normal sheep serum (Sigma), and 0.3% Triton X-100 (AppliChem)]. The sections were incubated overnight at 4°C with one of the following primary antibodies: monoclonal rabbit anti-active caspase-3 antibody (1:100; BD Pharmingen, Heidelberg, Germany; Cat. No. 559565), monoclonal mouse anti-GFAP antibody (1:50; Sigma; clone G-5-A; Cat. No. G3893), or polyclonal rabbit anti-phosphorylated histone H3 (Ser 10) antibody (1:200; Biomol, Hamburg, Germany; Cat. No. 06-570). Unbound primary antibody was removed by three rinses for 5 min each in TBS. Sections were further incubated for 30 min at RT in blocking solution, as described above, except that normal goat serum (PAN Biotech, Aidenbach, Germany) was used instead of normal sheep serum. Antigenic sites were visualized by incubating the sections at RT for 90 min with the following corresponding secondary antibodies: Alexa Fluor 488-conjugated goat anti-rabbit IgG (1:200; Molecular Probes/Invitrogen, Karlsruhe, Germany; Cat. No. A-11008), Alexa Fluor 488-conjugated goat anti-mouse IgG (1:200; Molecular Probes/Invitrogen; Cat. No. A-11029), or Cy3-conjugated goat anti-rabbit IgG (1:1,000; Dianova, Hamburg, Germany; Cat. No. 111-165-003), respectively. After three washes for 10 min each with TBS, the sections were counterstained by incubation with 2 μg/ml of 4′,6-diamidino-2-phenylindoledihydrochloride (DAPI; Sigma) in PBS for 3 min at RT. Finally, the sections were washed three times for 3 min each in TBS and embedded in polyvinyl alcohol containing n-propyl gallate.
Specificity of the rabbit anti-active caspase-3 primary antibody in A. leptorhynchus was demonstrated previously both by pre-adsorption of the respective primary antibody with the specific antigen and by omission of the primary antibody during immunohistochemical processing (Sîrbulescu et al. 2009). Specificity of the rabbit anti-phosphorylated histone H3 was tested by omission of the primary antibody during the labeling procedure (Rajendran et al. 2007).
Microscopy and analysis
Sections were examined under a Zeiss Axioskop epifluorescence microscope (Carl Zeiss, Göttingen, Germany). For quantitative analysis of active caspase-3-positive and phosphorylated histone H3-positive cells, every complete section was analyzed (46–53 sections per fish), and all immunopositive cells were counted. Confocal images were taken using a Zeiss LSM 510 META laser-scanning microscope equipped with Argon and Helium–Neon lasers. Optical sections were taken with a pinhole opening of 1 Airy and at a resolution of 1,024 × 1,024 pixels, using LSM 5 (version 3.2; Carl Zeiss) software.
To quantify GFAP immunoreactivity, confocal images of complete sections located at consecutive points along the caudalmost 4 mm of the spinal cord were taken using identical settings for pinhole opening, laser intensity, and intensity gain. The outer perimeter of the spinal cord and the central canal perimeter were outlined as borders of the region of interest using LSM 5 software. The overlayed images were then exported as grayscale bitmap files and imported into Matlab. Self-developed software was used to quantify pixels with a brightness above 10% of the maximum within the region of interest. The number of such ‘labeled’ pixels divided by the total number of pixels in the spinal cord area was used as a measure of GFAP immunoreactivity within each analyzed section. Between 21 and 32 sections were analyzed per fish.
Photographs of regenerated tails
Photographs of the regenerated tails of fish at 30 days post-amputation were taken at a resolution of 2,592 × 1,944 pixels, using a Canon PowerShot S5 IS digital camera. The lens (focal length of 6.0–72.0 mm) was equipped with a macro function.
Normal distribution of all data sets was verified using the Kolmogorov–Smirnov test. Significant group differences were identified using two-tailed independent-samples Student’s t test. Results are reported as means ± standard error of the mean.
Effect of temperature on EOD recovery
In A. leptorhynchus, the electric organ is formed by modified axonal terminals of large motoneurons located in the spinal cord (Bennett 1971). This property enabled us to use changes in the amplitude of the EOD to monitor functional recovery during spinal cord regeneration. Previously, we showed that the degree of decrease in EOD amplitude after amputation can be used as a reliable indicator of the extent of spinal cord ablation (Sîrbulescu et al. 2009).
Effect of temperature on apoptosis, cell proliferation, and gliosis
Quantitative analysis in transverse sections from the caudalmost 4 mm of spinal cord showed that the number of apoptotic cells was high in fish kept at 22°C (10 ± 3 cells per section), while in fish kept at 30°C the number of such cells was significantly lower, averaging only 2 (±0.2) cells per section (Fig. 2a″). Cell proliferation showed the opposite trend, with relatively low levels in fish kept at 22°C (1.2 ± 0.2 cells per section), and significantly higher values in fish kept at 30°C (2.0 ± 0.2 cells per section; Fig. 2b″). GFAP immunoreactivity (Fig. 2c–c″) covered on average 60% (±1.7%) of the spinal cord section area in fish kept at 22°C, and an average of 74% (±2%) of the spinal cord section area in fish kept at 30°C. The difference between the two groups was significant (p < 0.01; Fig. 2c″).
Does high temperature inhibit apoptosis?
The reduced number of apoptotic cells in fish kept at 30°C, relative to fish kept at 22°C, at 10 days post-amputation could be explained by an inhibitory effect of high temperature on apoptosis. Alternatively, it is possible that the higher temperature accelerates tissue regeneration, so that fish kept at 30°C return, after an initial increase in the rate of apoptosis, faster back to baseline levels than fish kept at 22°C.
Effect of temperature on re-growth of caudal spinal cord
The present study explores the effect of ambient temperature on structural and functional spinal cord recovery in a regeneration-competent, poikilotherm model organism. Moderate temperature manipulations were employed, with values not exceeding 3–4°C below or above the temperature at which the fish are normally kept. We observed a distinct beneficial effect of mild hyperthermia on functional regeneration after amputation of the caudal spinal cord. Fish kept at 30°C recovered the EOD amplitude at a significantly higher rate than fish kept at 22°C. The improved functional recovery observed in fish kept under warm-water conditions was paralleled, and probably caused, by a reduced duration of the post-lesion wave of apoptosis and by an increased rate of cell proliferation. Increased gliosis levels observed in animals kept at 30°C did not appear to be detrimental to the recovery process. In addition, fish kept under high-temperature conditions exhibited a significantly accelerated re-growth of the tail at 30 days post-amputation, while this process appeared to be delayed in fish kept at 22°C.
Effect of temperature on functional recovery after spinal cord injury
In the present study, we found that mildly hyperthermic conditions were associated with a significant improvement in behavioral recovery after spinal cord injury. This result is in line with findings made in other regeneration-competent organisms. A study in lampreys (Cohen et al. 1999) revealed that the proportion of animals with dysfunctional motor patterns was as high as 74% at 12°C. By contrast, only 17% of the animals that recovered at 22°C showed motor pattern anomalies during swimming. Interestingly, this difference occurred despite the fact that lampreys are naturally adapted to cold-water conditions.
By contrast, a number of studies in mammals have focused mostly on the potential beneficial effects of hypothermia, which has been shown to improve locomotor recovery after spinal cord injury (for reviews, see Inamasu et al. 2003; Dietrich et al. 2009; Marion and Bullock 2009). Nevertheless, various other studies have reported no benefits, or even detrimental effects, associated with hypothermia after CNS injuries. In rats, systemic hypothermia failed to improve recovery of motor function after spinal cord lesion (Dimar et al. 2000; Westergren et al. 2000). After fluid percussion brain injury in rats, hypothermic treatment reduced vascular dysfunction, but failed to protect against persistent learning and memory impairments (Wei et al. 2009). An in vitro study on the re-sealing of cut axons in isolated strips of white matter from guinea pig spinal cord found that low temperature (25°C) severely hindered the repair of damaged membranes (Shi and Pryor 2000).
Despite some controversy, in general, the results of mammalian studies suggest a beneficial effect of hypothermia. By contrast, the effects of hyperthermia after CNS injury are usually found to be detrimental to the recovery process, worsening both histopathological and behavioral outcomes (for review, see Dietrich and Bramlett 2007). This effect of temperature in mammals contrasts with the observations made in the present study in A. leptorhynchus, potentially indicating fundamental differences in the recovery process between homeotherm mammalian model systems and regeneration-competent organisms, all of which are poikilotherms.
Effect of temperature on cellular and molecular processes underlying recovery after spinal cord lesion
To gain insight into possible mechanisms by which temperature influences behavioral recovery in A. leptorhynchus, we addressed three phenomena which play a major role in the regeneration process after spinal cord injury: apoptosis, cell proliferation, and gliosis.
We observed an overall accelerating effect of hyperthermia on apoptotic cell death. At 18 h post-amputation, when the rate of apoptosis peaks, the number of caspase-3+ cells was elevated in both fish kept at 22°C and fish kept at 30°C, with a higher mean number of immunopositive cells in the latter group. By 10 days post-amputation, however, the number of caspase-3+ cells in fish kept at 30°C had decreased markedly, reaching levels observed in a previous study at 10–20 days post-lesion in fish kept at 25–26°C (Sîrbulescu et al. 2009). By contrast, in fish kept at 22°C, apoptotic cell levels were still increased at 10 days post-amputation, approximating values observed at temperatures of 25–26°C at a post-lesion survival time of 5–10 days (Sîrbulescu et al. 2009). These results suggest that hyperthermic conditions accelerate the rate by which cells undergo apoptosis.
Notably, the number of caspase-3+ cells observed in fish kept at 30°C, at 18 h post-amputation, was several times lower than the numbers found at any time point when apoptosis peaks (15–48 h post-amputation) in fish kept at 25–26°C (Sîrbulescu et al. 2009). Thus, a certain inhibitory effect of temperature on apoptosis cannot be excluded in our experiments. Such a modulatory effect could be mediated by the induction of heat shock proteins, most of which have in general an anti-apoptotic function (for reviews, see Arya et al. 2007; Giffard et al. 2008). Interestingly, heat shock protein 70 was found to be significantly up-regulated after brain injury in A. leptorhynchus (Zupanc et al. 2006b). This protein can inhibit the apoptotic cascade at multiple levels—by preventing the release of cytochrome c from mitochondria, the formation of the apoptosome, and the activation of initiator caspases (Arya et al. 2007). High levels of heat shock proteins already present after injury could be further elevated by hyperthermic conditions, leading to improved protection against apoptotic cell death.
We observed an overall increase in the number of mitotic cells after injury under mildly hyperthermic conditions. Fish kept at 30°C showed an almost twofold increase in the number of phosphorylated histone H3+ cells, as compared to fish kept at 22°C. This result is in agreement with the hypothesis that high temperature has an overall stimulating effect on cellular metabolism in poikilotherm organisms. In juvenile turtles, higher numerical densities of proliferating cells were found in the brain and spinal cord of warm-acclimated animals (Radmilovich et al. 2003). Similarly, much higher levels of cell proliferation were observed in the retina of tench (Tinca tinca) kept at 20°C, as compared to animals kept at 6°C (Velasco et al. 2001). In goldfish, the proliferation rate of primary cells in culture increased proportionally with increasing temperature, up to 35°C (Kondo and Watabe 2004). Although cells of homeotherm organisms have a much narrower optimal temperature range, a similar effect was observed in cells of mammalian origin. Low temperature (32°C) led to decreased cell proliferation in CHO and NS0 cell lines (Marchant et al. 2008), as well as in hybridoma C2E7 cells (Chong et al. 2008). However, hyperthermia above 40–41°C can lead to cell cycle arrest and even induce apoptosis in mammalian cells (Park et al. 2008).
Glial scar formation due to enhanced proliferation of astrocytes after injury is considered one of the major impediments for CNS regeneration in mammals (for reviews, see Fitch and Silver 2008; Rolls et al. 2009). In the present study, we found that mild hyperthermia leads to a relatively small, but significant increase in gliosis. Fish kept at 30°C showed an increase of approximately 20% in GFAP immunoreactivity, as compared to fish kept at 22°C. In contrast to mammals, elevated levels of gliosis do not appear to be detrimental to the process of regeneration in fish. In A. leptorhynchus, regeneration of brain tissue after an experimental lesion takes place despite the development of a dense network of GFAP- and vimentin-positive astrocytic processes after injury (Clint and Zupanc 2001, 2002). Similarly, pronounced GFAP immunoreactivity was observed in the regenerated spinal cord of a related species, Apteronotus albifrons (Anderson et al. 1984). It has been hypothesized that in the CNS of A. leptorhynchus, as in other regeneration-competent organisms, astrocytes and radial glia have a supportive function, aiding in the migration of newly generated cells and possibly performing a neurotrophic role (Clint and Zupanc 2001, 2002; Sîrbulescu et al. 2009; Zupanc 2009). Furthermore, there is evidence that the local CNS microenvironment in these organisms lacks the inhibitory factors present in mammals (for reviews, see Stuermer et al. 1992; Becker and Becker 2007). It is interesting to note that the observed increase in GFAP immunoreactivity at high temperature may be due to a higher proliferation rate of GFAP-positive radial glia, which have been shown to function as progenitor cells in fish (Reimer et al. 2009).
Taken together, the results of this study demonstrate that temperature manipulation can be employed to improve the rate of structural and functional regeneration after spinal cord injury in a regeneration-competent poikilotherm model organism. We found that mild hyperthermia can significantly accelerate the regeneration process, leading to faster recovery of behavioral function. Moreover, we did not observe any detrimental effect of this degree of hyperthermia.
One aspect that affords further investigation is the effect of temperature modulation on immune function. In mammals, the main negative aspects associated with hypothermia are its inhibitory effects on the activation of the immune system and on blood coagulation (Marion and Bullock 2009). In regeneration-competent organisms, it is likely that mild hyperthermia might help circumvent these issues, while still not triggering an enhanced inflammatory response. Advantages include a general acceleration of the metabolic rate, so that the overall duration of detrimental processes occurring immediately after injury is shortened; the enhancement of cell proliferation and cell differentiation; and the activation of heat shock proteins which subsequently can exert a complex protective function in affected cells.
Importantly, mild hyperthermia appears to affect poikilotherms and homeotherms in a fundamentally opposite manner. In mammals, fever-like moderate hyperthermia can lead to markedly worsened outcomes after CNS injury, both in histopathological and in behavioral measures. Main reasons for this appear to be the induction of inflammatory cascades and the increase in neuronal excitotoxicity (for review, see Dietrich and Bramlett 2007). Why the same detrimental effects do not manifest in regeneration-competent organisms remains to be investigated. It is likely that such comparative studies may reveal novel protective factors and mechanisms that allow poikilotherms to benefit from an enhanced metabolic rate, while avoiding the associated negative side effects. Such factors could prove useful for the development of future therapeutic strategies applicable to mammals.
This study received financial support from the Ernst A.-C. Lange-Stiftung and from Jacobs University Bremen.