5-HT7 receptors have been implicated in the control of locomotion. Here we use 5-HT7 receptor knockout mice to rigorously test whether 5-HT acts at the 5-HT7 receptor to control locomotor-like activity in the neonatal mouse spinal cord in vitro and voluntary locomotion in adult mice. We found that 5-HT applied onto in vitro spinal cords of 5-HT7+/+ mice produced locomotor-like activity that was disrupted and subsequently blocked by the 5-HT7 receptor antagonist SB-269970. In spinal cords isolated from 5-HT7−/− mice, 5-HT produced either uncoordinated rhythmic activity or resulted in synchronous discharges of the ventral roots. SB-269970 had no effect on 5-HT-induced rhythmic activity in the 5-HT7−/− mice. In adult in vivo experiments, SB-269970 applied directly to the spinal cord consistently disrupted locomotion and produced prolonged-extension of the hindlimbs in 5-HT7+/+ but not 5-HT7−/− mice. Disrupted EMG activity produced by SB-269970 in vivo was similar to the uncoordinated rhythmic activity produced by the drug in vitro. Moreover, 5-HT7−/− mice displayed greater maximal extension at the hip and ankle joints than 5-HT7+/+ animals during voluntary locomotion. These results suggest that spinal 5-HT7 receptors are required for the production and coordination of 5-HT-induced locomotor-like activity in the neonatal mouse and are important for the coordination of voluntary locomotion in adult mice. We conclude that spinal 5-HT7 receptors are critical for alternating activity during locomotion.
Strong evidence suggests that 5-hydroxytryptamne (5-HT) is a neurotransmitter involved in a descending command system sufficient for production of locomotion (Jordan and Schmidt 2002; Liu and Jordan 2005; Schmidt and Jordan 2000). Furthermore, there is increasing interest in the possibility that recovery of locomotion after injury can be facilitated by 5-HT agonists or transplantation of 5-HT neurons into the spinal cord. (Antri et al. 2002, 2005; Feraboli-Lohnherr et al. 1999; Gimenez y Ribotta et al. 2000; Kim et al. 2001; Landry et al. 2006; Majczynski et al. 2005; Slawinska et al. 2000). The specific receptors responsible for the locomotor effects of 5-HT have not been determined.
Involvement of 5-HT7 receptors in the control of locomotion induced by 5-HT has been suggested by Hochman and co-workers, who used clozapine as an antagonist (Cina and Hochman 1998; Hochman et al. 2001). This notion has been supported by experiments using the specific 5-HT7 antagonist SB-269970 in both rat and mouse isolated spinal cord preparations (Landry et al. 2006; Madriaga et al. 2004; Pearlstein et al. 2005). The in vitro locomotor rhythm is slowed, and coordination among antagonist flexor-extensor nerve pairs and the two sides of the spinal cord is disrupted by SB-269970 (Liu and Jordan 2005; Madriaga et al. 2004; Pearlstein et al. 2005), suggesting that neurons of the central pattern generator (CPG) for locomotion possess 5-HT7 receptors. Stimulation of a discrete population of 5-HT neurons in the parapyramidal region (PPR) of the medulla elicits locomotor-like activity in the neonatal rat isolated brain stem-spinal cord preparation (Liu and Jordan 2005), and this activity can be disrupted or blocked by 5-HT7 receptor antagonists. Locomotion evoked in the decerebrate cat by stimulation of the mesencephalic locomotor region (MLR) can also be blocked by the 5-HT7 antagonist clozapine (Schmidt and Jordan 2000).
Because of the possibility that antagonists may act at more than one receptor, further evidence besides that obtained in pharmacological experiments is required to firmly establish a role for 5-HT7 receptors in the control of locomotion. Mice with a targeted deletion of 5-HT7 receptors (Hedlund et al. 2003; Roberts et al. 2004) have been used to show that 5-HT7 receptors are among those needed for the induction of locomotor-like activity by systemic administration of the 5-HT agonist 8-hydroxy-2-(di-N-propylamino)-tetralin (8-OH-DPAT) in adult spinal mice (Landry et al. 2006). Here we use these mice and their wild-type littermates to examine the role of 5-HT7 receptors in 5-HT-induced locomotion in vitro and voluntary locomotion in vivo. We reasoned that if 5-HT7 receptors normally participate in the production of 5-HT-induced locomotor rhythm, then the ability of 5-HT to induce locomotor-like activity in vitro should be disrupted in 5-HT7−/− mice. Moreover, if 5-HT7 receptors normally play a role in the production of locomotion in adult animals, then it is expected that voluntary locomotion would be disrupted by a 5-HT7 receptor antagonist in adult wild-type mice (5-HT7+/+) but not in 5-HT7−/− mice. Preliminary accounts of some of these results have been reported (Jordan et al. 2008; Liu et al. 2006).
Mice lacking 5-HT7 receptors were created by targeted disruption of the 5-HT7 receptor gene, previously described by Hedlund et al. (2003). Crossings of heterozygous 5-HT7± mice produced homozygous wild-type (5-HT7+/+), homozygous knockout mice (5-HT7−/−) and heterozygous (5-HT7±) mice at the expected Mendelian frequency. Genotyping consisted of detecting the neomycin resistance gene and a part of exon II of the receptor gene using PCR.
Exon II primers: F: GAC AAA GTG TGC TTG ATC AGC CAG G; R: ATG CAG CTA CAG GAG GTG CCA CAG; PCR product: 400 bp; Neo primers: F: CTT GGG TGG AGA GGC TAT TC; R: AGG TGA GAT GAC AGG AGA TC; PCR product: 280 bp; control primers: mouse beta casein (MBC): F: GAT GTG CTC CAG GCT AAA GTT CAC; R: AGA AAC GGA ATG TTG TGG AGT GGC; PCR product: 500 bp.
Animals and procedures, in vitro experiments
The experiments were performed on neonatal C57B/6J 5-HT7−/− and 5-HT7+/+ mice (0–5 day old). All experiments were performed in accordance with the Canadian Council on Animal Care guidelines and were approved by the University of Manitoba Animal Protocol Committee and by the University of Alberta animal welfare committee. Following anesthesia with halothane, the animal was immediately decapitated, eviscerated, and removed to a silicone elastomer (Sylgard)-coated recording chamber. A laminectomy was performed, and the spinal cord was isolated. The spinal cord was transected below T1 and was superfused with artificial cerebrospinal fluid (ACSF, concentration in mM: 128 NaCI, 3.0 KCI, 0.5 NaH2PO4, 1.5 CaCl2, 1.0 MgSO4, 21 NaHCO3, and 30 glucose) and oxygenated with 95%-5% CO2 at room temperature. The preparation was pinned ventral side up to the Sylgard surface and allowed to stabilize for 30 min before recording. Temperature was maintained at 25°C during experiments.
Animals and procedures, in vivo experiments
CD1/ICR (n = 4) and C57B/6J (n = 9) mice were used in these experiments. Thirteen adult wild-type and six littermates of C57B/6J 5-HT7−/− (Hedlund et al. 2003) were used to determine if 5-HT7 receptors are essential for the action of SB-269970. Anesthesia was induced with Forane (Isoflurane, Baxter, Toronto, Ontario, Canada) as previously described (Akay et al. 2006, 2008). Briefly, prior to surgery, the opioid receptor agonist Buprenex (buprenorphine hydrochloride, Reckitt Benckiser Healthcare., Hull, UK, 0.15 mg/kg) was injected subcutaneously. A laminectomy was performed at the T12 and T13 vertebrae to expose the L2–L4 segments of the spinal cord. The dura mater was removed with a fine spring scissor. Additionally little openings were also made on the pia to ease the drug diffusion into the spinal cord. The 5-HT7 antagonist SB-269970 (1–10 mM) or vehicle (saline) was then applied for 30 min in a bath created with petroleum jelly (Vaseline) around the opening in the dura. At the end of the 30-min period, the drug or saline solution and the Vaseline was quickly removed, the wound was closed and the anesthesia was discontinued. The mouse was placed on a transparent walkway, as previously described (Akay et al. 2006, 2008; Pearson et al. 2005). The mice normally commenced voluntary locomotion approximately 1 or 2 min after recovery from the anesthesia. Due to the application of the opioid receptor agonist buprenorphine, the mice walked undisturbed without any signs of pain from the surgery. “Precontrol,” “drug application,” and “postcontrol” recordings were captured. The effect of the applied drug lasted ≤1.5 h, dependent on the dose. After the recordings were made, the animals were anesthetized again and sodium pentobarbital (0.96 mg/ml) 0.2 ml was injected intraperitoneally for euthanasia.
Electrophyisiological recordings, in vitro experiments
Glass suction electrodes were applied to record the lumbar 2 (L2) and lumbar 5 (L5) ventral discharges of both sides as monitors of locomotor-like activity (Cowley and Schmidt 1997; Kjaerulff and Kiehn 1996). Ventral root recordings were amplified, filtered (30 Hz to 2 kHz), digitized, stored, and analyzed using software developed at the Winnipeg Spinal Cord Research Centre (see http://www.scrc.umanitoba.ca/doc/ for details). Locomotor-like activity consisted of alternating activity recorded between both the left and right L2 and L5 ventral root pairs as previously described for the in vitro isolated mouse spinal cord (Madriaga 2004; Nishimaru et al. 2000; Whelan et al. 2000). Locomotor-like activity was evoked by bath application of 5-HT (20–50 μM). SB-269970 (5–20 μM) was used to block 5-HT7 receptors, while ketanserin (15–20 μM) and spiperone (10–15 μM were used to block 5-HT2 receptors. Drugs were obtained from Sigma-Aldrich.
Electrophyisiological recordings, in vivo experiments
The fabrication and implantation of electromyographic (EMG) electrodes was as previously described (Akay et al. 2006, 2008; Pearson et al. 2005). For implantation, the mice were anesthetized with isoflurane, and the hindlimbs and the dorsum of the neck were shaved. Small incisions were made in the skin over the dorsal neck area, tibialis anterior (TA, ankle flexor) and gastrocnemius (GS, ankle extensor) muscles, and bipolar electrodes were drawn under the skin from the neck incision to the leg incisions. The electrodes were inserted into the muscles as previously described (Pearson et al. 2005). The incisions on the hindlimbs were closed, and a headpiece with the electrodes sutured to it was stitched to the skin near the neck incision. Buprenex was injected subcutaneously for analgesia, and the mice were left in their cages for recovery from surgery ≥2 days before performing the experiments.
Kinematic recordings during voluntary locomotion in adult mice
Custom-made three-dimensional reflective markers (2 mm diam) were glued onto the shaved skin at the level of the iliac crest, hip, knee, ankle, paw, and tip of the fourth digit (toe) of the left hindlimbs as well as on the wrist of the left foreleg. Knee position was calculated by triangulation from the position of hip and ankle joint markers, using the measured lengths of the femur and tibia. All the joint-angles (hip, knee, and ankle) were then determined by using that point (triangulated ankle position). For the video recordings during free walking, the mice where placed into a custom-made Plexiglas walkway (90 cm long, 5 cm wide, and 13 cm high). After recovery from anesthesia, the mice walked freely back and forth in the walkway. The center of the walkway was viewed with a high speed camera (Photron Fastcam) with a capture rate of 250 frame/s. Video data were stored in computer memory and were analyzed using Peak Motus 8.2 motion-analysis software (ViconPeak, Denver, CO). Kinematic parameters of the stepping movements, such as swing and stance durations and joint angles, were measured from data files created by the Peak Motus System. The coordination between the hindlimbs was determined from video images captured by a mirror placed underneath the walkway set at ∼45° from vertical. These images allowed measurement of the time of swing onset of each hindlimb, defined as time of the onset of forward movement of the paw.
Measurements and statistical analysis
Step cycle frequency and duration were determined by measuring right or left L2 ventral root activity (in vitro experiments) or TA and GS EMG after filtering and rectification (in vivo experiments). A 5-min episode of data with stable bursting was used for analysis after drug administration. EMG recordings were taken immediately after removal of the anesthetic and ≤90 min after removal from anesthesia. The statistical values are indicated as the means ± SD; n refers to the number of experiments. Student's t-test was used to compare the differences between the mean values, and P values <0.05 were considered significant. Circular statistics was used to determine the coupling strength (Φ; also called phase value) between flexor-extensor and right-left ventral roots or left and right TA and GS EMG recordings (Kjaerulff and Kiehn 1996; Kriellaars et al. 1994; Liu and Jordan 2005). The phase value was displayed graphically as data points on a polar plot (Zaporozhets et al. 2006; Zar 1974). A phase value of 0.5 indicates strict alternation between the onsets of bursts of activity in the chosen roots or muscles. The mean value (r), reflecting the concentration of phase values, was plotted as the length of the vector, which ranged from 0 to 1. If r was greater than the critical Rayleigh's value as indicated with a dotted circle then the relationship was considered phase-related. In our experiments, cross-correlation analysis was also used to determine the rhythmic stability between the flexor-extensor and right-left ventral root electroneurographic recordings or ipsilateral TA/GS and right-left TA or GS EMG recordings, measured as the minimum negative value at time 0 subtracted from the maximum first positive value of the cross-correlogram, or peak-to-trough correlation coefficient (PTCC) (Madriaga et al. 2004). The mean PTCC was averaged of all individual values observed in a given experimental condition. The maximum PTCC is defined as −2. In video recordings of the in vivo experiments, the cycle periods were measured from the onset of swing (defined as the start of forward movement of the toe) to the onset of the following swing phase.
Properties of locomotor-like activity produced in vitro by 5-HT in wild-type mice
We characterized the effects of 5-HT applied to the isolated spinal cord of wild-type neonatal C57B/6J mice to provide baseline data for comparison with their knockout littermates. This strain has not been used in any previous in vitro study on the action of 5-HT. Bath application of 5-HT (20–50 μM)-induced locomotor-like activity in 53.8% (7 of 13) 5-HT7+/+ spinal cord preparations. The mean cycle rate was 0.12 ± 0.09 Hz, and the burst duration was 5.07 ± 2.82 s. The initial exposure to 5-HT always resulted in an increase in tonic discharges in all four roots, followed by rhythmic alternate activity (n = 7). In two of these preparations, uncoordinated rhythmic activity was observed in the first 10 min after 5-HT application followed by alternate locomotor activity. A representative example of locomotor-like activity recorded in the ventral roots and evoked by 5-HT is presented in Fig. 1A. Figure 1B shows a raster display of the four ventral roots from 15 cycles during which the alternation between flexor (L2) and extensor (L5) and right and left ventral roots was evident. Locomotor-like activity recorded from seven preparations was pooled in polar plots (Fig. 1C). The mean value of the phase (Φ) between right and left side (RL2–LL2) was 0.47 ± 0.08; coupling strength was highly significant (r = 0.87, P < 0.001). The mean value of the phase (Φ) between flexor and extensor ventral roots (RL2–RL5) was 0.46 ± 0.07, and the coupling strength was also significant (r = 0.89, P < 0.001). Cross-correlation analysis revealed stable flexor-extensor (Fig. 1D) and right-left alternation (E). The PTCCs between flexor-extensor and right-left ventral roots in this example were −1.46 and −1.50, respectively. The mean PTCCs between flexor-extensor and right-left ventral roots for seven preparations were −1.25 ± 0.26 and −1.44 ± 0.31, respectively. Data from both qualitative and quantitative analyses suggest that 5-HT could induce well-coordinated locomotor-like activity in 5-HT7+/+ mice. An increase in tonic activity of ventral roots was observed in the remaining six preparations. In mice with initial failure of induction of locomotion, increasing the 5-HT concentration to 60–100 μM (n = 3) was ineffective, indicating that the concentration of 5-HT (20–50 μM) in our experiments is appropriate for inducing locomotor-like activity and increasing the dose does not improve the occurrence of locomotor-like activity.
5-HT-induced locomotor-like activity was disrupted by 5-HT7 receptor blockade
To test the importance of 5-HT7 receptors in the production of 5-HT-induced locomotion, SB-269970, a specific 5-HT7 receptor antagonist, was applied to the bath after locomotor-like activity was induced by 5-HT (n = 5). Figure 2A illustrates an example of locomotor-like activity evoked by 5-HT (30 μM). Φ and r values between flexor and extensor ventral roots were 0.48 and 0.99 (P < 0.001) and 0.49 and 0.98 (P < 0.001) between right and left side. Application of SB-269970 (15 μM) first disrupted the locomotor rhythm at 5 min (Fig. 2B) as indicated by a change in the Φ and r values between flexor and extensor ventral roots to 0.36 and 0.44 (P > 0.05). Φ and r value between the right and left side were also altered (Φ = 0.28 and r = 0.30, P > 0.05), indicating disruption of locomotor-like activity characterized by the loss of alternation between ipsilateral flexor and extensor ventral roots and between ventral roots on the two sides. Locomotor-like activity was completely blocked after 10 min (Fig. 2C). The blockade of locomotor-like activity by SB-269970 was consistently observed in all five tested preparations. The results indicate that 5-HT7 receptors play a critical role in 5-HT-induced locomotor-like activity. This finding is consistent with previous experiments in rats (Hochman et al. 2001; Liu and Jordan 2005; Pearlstein et al. 2005) and another strain of mice (Madriaga et al. 2004), thus providing the background for examining the 5-HT-induced effects in the spinal cords of 5-HT7−/− mice.
5-HT produced uncoordinated rhythmic activity in vitro in 5-HT7−/− mice
5-HT (20–50 μM) was applied to 5-HT7−/− mouse in vitro spinal cord preparations (n = 10). In most of these preparations (7/10), 5-HT produced nonlocomotor-like uncoordinated rhythmic bursting (Fig. 3). Well-coordinated locomotor-like activity was evoked in only one preparation in which Φ and r values were 0.52 and 0.90 (P < 0.001) for flexor and extensor ventral roots and 0.47 and 0.83 (P < 0.001) for right and left ventral roots and was blocked by ketanserin (20 μM), a 5-HT2A receptors antagonist. However, it is noteworthy that the cycle rate in this one case was markedly slower (0.06–0.07 Hz) than that seen in wild-type animals, where the mean cycle rate was 0.12 Hz. Increased tonic activity without any rhythmic activity was produced by 5-HT in the remaining two preparations.
A representative example of the predominant response to 5-HT (uncoordinated rhythmic activity) is shown in Fig. 3A. Figure 3B is a raster display of the four ventral roots from the same preparation in which uncoordinated discharges in the four roots is evident. Polar plots in Fig. 3C show the coupling strength between flexor-extensor and right-left ventral roots; in all but one case, the poor coupling between flexor-extensor and right-left ventral roots was appreciated. The mean flexor-extensor Φ from 10 cases was 0.53 ± 0.32; the coupling strength was not significant (r = 0.46, P > 0.05). The mean right-left Φ was 0.43 ± 0.24, and the coupling strength was also not significant (r = 0.37, P > 0.05). PTCCs between flexor-extensor (Fig. 3D) and right-left (E) ventral roots were −0.55 and −0.50, respectively, for the example shown in Fig. 3A. Ipsilateral synchronous discharges were evoked in 2 of the 10 preparations (example in Fig. 3F). The mean PTCCs between flexor-extensor and right-left ventral roots in the 10 preparations were −0.49 ± 0.23 and −0.56 ± 0.13. These are much weaker relationships than observed in the wild-type animals (compare with Fig. 1) where the mean PTCCs were −1.25 and −1.44. In three preparations with uncoordinated rhythmic activity, increasing the 5-HT concentration to 60–100 μM did not significantly change the discharge pattern. Together, these results indicate that 5-HT produced uncoordinated rhythmic activity, rather than well-coordinated locomotion, in most 5-HT7−/− animals.
To test whether SB-269970 has effects on 5-HT-induced rhythmic activity in 5-HT7−/− mice, SB-269970 (15–30 μM) was applied after uncoordinated rhythmic activity was evoked by 5-HT in four preparations. Figure 4A is an example of 5-HT-induced uncoordinated rhythmic activity. As shown in the polar plots, Φ and r between flexor-extensor and right-left ventral roots were Φ = 0.38, r = 0.54 (P > 0.05) and Φ = 0.15, r = 0.60 (P > 0.05). Ten minutes after adding SB-269970 (15 μM), Φ and r between flexor-extensor and right-left ventral roots were Φ = 0.31, r = 0.41 (P > 0.05) and Φ = 0.18, r = 0.19 (P > 0.05). There were no significant differences in the ventral root discharges prior to and after SB-269970 (Fig. 4B). These results suggest that the action of SB-269970 on 5-HT-induced activity in vitro requires functional 5-HT7 receptors.
Previous experiments demonstrated that 5-HT2 receptors play a role in 5-HT-induced locomotor-like activity in mouse preparations (Madriaga et al. 2004). To test the possible involvement of 5-HT2 receptor in 5-H-induced uncoordinated rhythm in the 5-HT7−/− mice, ketanserin, a 5-HT2A antagonist, was used. As shown in Fig. 4C, ketanserin (15 μM) effectively blocked uncoordinated discharges in all preparations tested (n = 4). The involvement of 5-HT2A receptors was further confirmed using a second 5-HT2A receptor antagonist, spiperone. Spiperone (10–15 μM, n = 4) consistently abolished this bursting (data not shown). Blockage by both ketanserin and spiperone is taken as an indication of 5-HT2A receptor involvement (Glennon et al. 2002). It is possible that 5-HT2A receptors play an increasing role in the production of locomotion in the absence of 5-HT7 receptors.
5-HT7 receptor blockade disrupts voluntary locomotion in adult wild-type mice
To test the hypothesis that 5-HT7 receptors are involved in the control of voluntary locomotion in adult mice, we used adult 5-HT7+/+ mice and directly applied the 5-HT7 receptor antagonist, SB-269970 (1 or 10 mM), to the exposed spinal cord at the level of the L2–L4 segments. Previous experiments using topical application of drugs suggest that these doses are within the expected effective range (Brumley et al. 2007; Chau et al. 2002). The most pronounced effect of SB-269970 was prolongation of extension in the hindlimbs that severely impaired locomotion. This occurred in both CD1/ICR (4/4) and C57B/6J (8/9) strains of mice. Figure 5A shows an example of video frames of one step cycle prior to and after SB-269970 administration in a C57B/6J mouse. The effects of SB-269970 could be associated with prolonged extension of the limb beyond the point of successful weight support, and sustained extension at the ankle joint with failure of plantar placement of the hindpaw. Complete recovery from the effects of SB-269970 occurred on average after 51.75 ± 25.33 min. All of the preceding results were obtained with doses of 10 mM SB-269970. In three cases, 1 mM SB-269970 was used, but only very transient effects on locomotion were observed.
Kinematic and EMG analysis was performed to determine the changes in joint angles and muscle activation induced by SB-269970 in wild-type animals. Increased hip extension and persistent extension of the ankle joint occurred in the example shown in Fig. 5B. The stick figures represent one step cycle (during the horizontal bar below the EMG traces) for the control (left) and SB-269970 (right) conditions. Ankle, knee, and hip joint angles are shown for five step cycles (swing phase indicated by gray vertical bars) in the control condition and four step cycles after SB-269970 administration. Flexor (TA) and extensor (GS) EMGs are shown for these same step cycles. Increased hip extension was observed in 8/9 cases, and the changes in the ankle joint angle were also observed in these same eight animals. The joint angle changes were accompanied by disturbed coordination between ipsilateral ankle flexor (TA) and extensor (GS) muscles (Fig. 5B). Repetitive extensor EMG activity continued during the prolonged extension, and the ankle extensor EMG never persisted throughout the period of ankle extension. We conclude that the extension was not due to prolonged extensor EMG activity, but was associated with a disrupted rhythm in the flexor EMG that failed to produce a successful swing phase of the locomotor cycle.
Polar plots and correlation analysis were used to more clearly demonstrate the altered coordination produced by SB-269970 in adult animals. Figure 6A shows an example of EMG recording prior to SB-269970 application, and alternation between flexor-extensor and right-left side is illustrated in polar plots (middle) and histograms of the correlation-coefficient (right). After the drug application, flexor-extensor and right-left alternation were disturbed, with frequent co-contractions of ipsilateral flexors and extensors (Fig. 6B). Recovery occurred within 60 min after drug application as shown in Fig. 6C. Figure 7 summarizes EMG polar plots in four wild-type animals in the control and SB-269970-treated conditions. The mean Φ from four wild-type animals between flexor-extensor and right-left EMGs was 0.57 ± 0.16 and 0.47 ± 0.07; coupling strength between flexor-extensor (r = 0.82, P < 0.001) and right-left (r = 0.85, P < 0.001) EMGs was highly significant. Application of SB-269970 disturbed or decreased flexor-extensor and right-left alternation in all cases. The mean Φ between flexor-extensor and right-left EMGs was 0.63 ± 0.31 and 0.52 ± 0.13; and the coupling strength value between flexor-extensor (r = 0.58, P > 0.05) and right-left (r = 0.53, P > 0.05) was not significant. In one of these animals, the ipsilateral flexor and extensor coupling became synchronous (see also Fig. 6B). These disturbances in EMG coordination are similar to the effects of SB-269970 on ventral root recordings in vitro (Fig. 3). Forelimb walking was not affected by SB-269970 during these experiments and no indications of systemic actions of the drug were observed. All animals recovered fully from the drug treatment.
Voluntary locomotion in adult mice lacking 5-HT7 receptors
Adult 5-HT7−/− mice are capable of voluntary locomotion (Fig. 8 A, control), as previously reported (Hedlund et al. 2003). This is consistent with the presence of multiple activating pathways for the control of locomotion, such that apparently normal locomotion in the absence of 5-HT7 receptors can be accounted for by these other pathways. In contrast to the wild-type animals (see Fig. 5), the 5-HT7 antagonist SB-269970 did not produce prolonged extension (Fig. 8A, SB-269970) in 4/6 5-HT7−/− mice. Two examples of stick figures and joint angle measurements from 5-HT7−/− mice prior to (control) and after SB-269970 administration are illustrated in Fig. 8, B and C. The stick figures correspond to the steps indicated by horizontal bars below the joint angle traces, and ankle, knee, and hip joint angles are shown for several step cycles. In these two examples, no appreciable difference from the control condition was produced by SB-269970. In the remaining two 5-HT7−/− animals, SB-269970 resulted in prolongation of hindlimb extension qualitatively similar to that observed in the wild-type animals.
To compare the prolonged extension produced by SB-269970 in 5-HT7+/+ and 5-HT7−/− mice, we measured the time to recovery from the prolonged extension after the drug application. The mean time to recovery from the SB-269970-induced prolonged extension in wild-type animals was 51.75 ± 25.33 min, and the mean time to recovery from after SB-269970 in 5-HT7−/− animals was 18 ± 17.22 min. This difference was highly significant (P = 0.012). This analysis confirms our observation that blockage of 5-HT7 receptors leads to prolonged extension in the hindlimbs of wild-type animals and suggests that 5-HT7 receptors are required for this effect.
Prolonged hindlimb extension was far more dramatic in the 5-HT7+/+ mice as compared with 5-HT7−/− mice. As shown in Fig. 5B, SB-269970 produced large changes in hip and ankle joint angles in 5-HT7+/+ mice. We therefore conducted a detailed analysis of the changes produced by SB-269970 in wild-type and knockout animals for these joints. The results of this analysis are summarized in Fig. 9. SB-269970 significantly increased the mean maximal hip joint angle by 12.7° in eight wild-type mice as compared with 6.4° increase in 5-HT7−/− mice (Fig. 9A). In knockout animals (control condition), the maximal hip joint angle was 4.2° (P < 0.001) greater than in wild-type animals. Similar results were obtained with maximal ankle joint extension (Fig. 9B). In 5-HT7+/+ mice, SB-269970 significantly increased ankle extension by almost 13.8° as compared with 7.3° increase in 5-HT7−/− mice. The control ankle extension in knockout animals was 9.6° (P < 0.001) greater than in wild-type animals. Thus neurons with 5-HT7 receptors most likely participate in the regulation of both ankle and hip joint angle to prolong extension and delay the onset of flexion. The addition of SB-269970 results in a further increase in the maximal joint angles in both wild-type and knockout animals, possibly due to effects of the drug that are not dependent on 5-HT7 receptors.
5-HT7 receptors are required for 5-HT-induced locomotor-like activity in vitro
We have shown that locomotor-like activity could be evoked by bath application of 5-HT in the majority of neonatal C57B/6J 5-HT7+/+ mice (7/13, 58.3%), and this locomotor activity was first disrupted and then blocked by a specific 5-HT7 receptor antagonist, SB-269970. We also showed that 5-HT produced only uncoordinated rhythmic activity characterized by the absence of alternation between ipsilateral flexor-extensor and right-left ventral roots in 5-HT7−/− mice, and this rhythmic activity was not affected by SB-269970. These results clearly demonstrate that functional activation of 5-HT7 receptors plays a critical role in 5-HT-induced locomotor-like activity in the neonatal mouse spinal cord. Furthermore, the absence of effect of SB-269970 on 5-HT-induced activity in 5-HT7−/− mice indicates that this antagonist requires 5-HT7 receptors for its action in vitro. This finding strengthens our previous report (Liu and Jordan 2005) using SB-269970 to show that 5-HT7 receptor activation is required for brain stem locomotor activity evoked by stimulation of 5-HT neurons in the parapyramidal region of the medulla in neonatal rats. Previous work in this lab demonstrated that intrathecal administration of clozapine, a nonspecific 5-HT7 receptor antagonist, abolished locomotion induced by brain stem stimulation in adult decerebrate cats (Schmidt and Jordan 2000). Thus our results provide further evidence for the hypothesis that 5-HT7 receptors play an important role in 5-HT-induced locomotor-like activity in mammals (Liu and Jordan 2005; Madriaga et al. 2004; Schmidt and Jordan 2000).
The 5-HT-induced uncoordinated rhythmic activity or synchronous discharges of the ventral roots in 5-HT7−/− mice was blocked by the 5-HT2A receptor antagonist ketanserin, indicating that activation of 5-HT2A receptors are likely involved in rhythmic activity in 5-HT7−/− mice. A nonspecific 5-HT2 agonist, α-methyl-5-HT, could elicit coordinated locomotor-like activity in the isolated spinal cord of neonatal mice (Madriaga et al. 2004; Pearson et al. 2003), an effect which was disrupted by ketanserin, consistent with our finding that the locomotor network may be activated by 5-HT2 receptor agonists. In one case, well-coordinated locomotion with extremely slow step frequency induced in 5-HT7−/− mice was observed, providing clues that compensation might occur at the postnatal stage and other 5-HT receptor subtypes (e.g., 5-HT2) may take on the role of rhythmogenesis for 5-HT-induced locomotion.
The action of 5-HT on the locomotor central pattern generator is altered in 5-HT7−/− mice such that coordination between ipsilateral flexor and extensor muscle groups and between the left and right sides is lost. This suggests that the neurons that normally are activated via 5-HT7−/− receptors participate in the coordination of locomotor output. An explanation for the uncoordinated locomotion produced by 5-HT7−/− mice might be that 5-HT7 receptors normally regulate the excitability of coordinating interneurons. In the absence of activation of sufficient numbers of these cells in 5-HT7−/− mice or in the presence of SB-269970, rhythmic activity ensues, but coordination among the neurons producing excitation of motoneurons is lacking. This is consistent with the suggestion that the locomotor pattern generator consists of two layers, one the rhythm generator layer and another pattern formation layer (Lafreniere-Roula and McCrea 2005). In accordance with this hypothesis, the 5-HT7 receptor positive neurons are most likely to be included in the pattern formation layer.
Certainly one of the possible mechanisms causing co-contractions of flexors and extensors following SB-269970 application is that inhibitory interneurons responsible for reciprocal inhibition during walking possess 5-HT7 receptors, and their excitability might be suppressed after the drug application. During locomotion alternate contraction of antagonist muscles at the same joint is realized partly by reciprocal inhibition through excitation of inhibitory interneurons (Jordan 1983; Orsal et al. 1986; Shefchyk and Jordan 1985). Blocking the action of inhibitory interneurons produces synchronous discharge of ventral roots in the isolated neonatal rat spinal cord (Cowley and Schmidt 1995). Ia inhibitory interneurons may be involved in this inhibition (Pratt and Jordan 1987), but they appear to not be required for reciprocal inhibition during fictive locomotion (Gosgnach et al. 2006). Thus one possibility is that the inhibitory interneurons responsible for alternating activity during locomotion are normally controlled by 5-HT via 5-HT7 receptors.
There is one report describing 5-HT7 receptor distribution within the spinal cord (Doly et al. 2005), showing that they are confined largely to the dorsal horn, with a low level of labeling of motoneurons in lamina X of the rat lumbar cord. Our results are not consistent with an action of 5-HT on motoneurons because the major effect of 5-HT7 receptor blockage or deletion is loss of alternating rhythmic activity, not a decrease in motoneuron output. 5-HT7 receptor-positive cells in the locomotor area of the ventral horn were not described by Doly and co-workers (2005). Using another antibody, we described 5-HT7 receptor positive locomotor neurons in the ventral horn spinal cord of adult rats labeled using an antibody to c-fos after a treadmill locomotor task (Jordan and Schmidt 2002).
5-HT7 receptor activation is required for normal voluntary locomotion in adult wild-type mice
Our data show that application of SB-269970 to the exposed spinal cord (L2–L4) resulted in a pronounced effect on locomotion in eight of nine 5-HT7+/+ mice, including prolongation of extension, synchronous activation of antagonist muscles, disorganized rhythmic activity in the ankle flexor muscle TA, and increased maximal extension of the hip and ankle joints. Such disrupted locomotion was observed in both CD1/ICR and C57B/6J strains of mice.
There was a striking similarity between the disrupted ventral root coordination seen in vitro and the uncoordinated EMG activity observed in the adult mice. For example, disrupted rhythmic activity and ipsilateral synchronous discharges were both seen in the in vitro and adult mice. Although the mechanisms underlying these similarities cannot be addressed in the present study, it may be that there is a loss of 5-HT control of inhibitory coordinating interneurons, as suggested in the preceding text to explain our in vitro results.
The increased maximal extension of the hip and ankle joints observed in 5-HT7−/− mice, as well as the increased maximal extension observed in 5-HT7+/+ mice after blockage of 5-HT7 receptors by SB-269970, suggests that some mechanism involved in the onset of flexion at the end of the stance phase of locomotion is likely controlled by neurons with 5-HT7 receptors. In both wild-type and knockout animals, there is a further change in the maximal extension at the hip and ankle joints produced by SB-269970 that cannot be attributed to an action at the 5-HT7 receptor. We interpret this further change as due to nonspecific actions of SB-269970 (Thomas and Hagan 2004). A possible explanation for the increase in maximal extension in 5-HT7−/− mice is that neurons responsible for the onset of the swing phase of locomotion might possess 5-HT7 receptors, and failure to activate them adequately would lead to the prolonged extension observed in these experiments. It is well known that one of the functions of sensory feedback during walking is to control the timing of the transition from stance to swing (Grillner and Rossignol 1978; Pang and Yang 2000; Pearson 1995; Whelan et al. 1995). Our results suggest that the presence of 5-HT7 receptors might be a marker for the neurons of the locomotor central pattern generator involved in the onset of the flexion phase of locomotion. This is consistent with our finding that during prolonged extension the control of TA, an ankle flexor, is disturbed (Fig. 5).
Another observation derived from this study was that adult 5-HT7−/− mice, despite the presence of increased maximal extension at the hip and ankle joints, do not show major deficiencies in locomotion. This is consistent with a previous description of locomotion in these mice (Hedlund et al. 2003) and with the observation that 5-HT does not appear to be necessary for locomotion in adult cats (Rossignol 2006; Schmidt and Jordan 2000; Steeves et al. 1980). In mice lacking Pet-1, the majority of brain stem 5-HT neurons fail to differentiate, giving rise to a transient decrease in exploratory locomotion (Hendricks et al. 2003). Recent experiments using Lmx1b conditional knockout mice resulted in depletion of all 5-HT neurons, but no overt locomotor deficits were observed (Zhao et al. 2006). The authors concluded, “the central serotonergic system is not required for normal locomotor activity,” consistent with the widely accepted notion that there are multiple descending locomotor command systems capable of eliciting locomotion, such that “none of the descending pathways plays an indispensable role in the basic generation of locomotion” (Rossignol 2006). It is clear from our results that the 5-HT system is one of the descending pathways normally involved in the control of locomotion, and a portion of the action of this pathway occurs via 5-HT7 receptors in rodents. This is shown by our finding that blockage of this receptor has profound effects on locomotion in adult animals that are naïve to alterations in the 5-HT system. Consistent with our findings, depletion of 5-HT with the 5-HT synthesis inhibitor ρ-chlorophenylalanine in neonatal rats resulted in decreased velocity of locomotion and prolonged extension of the hindlimbs (Myoga et al. 1995).
In conclusion, we have used 5-HT7−/− mice to confirm the hypothesis that 5-HT induces locomotor-like activity in the isolated neonatal mouse spinal cord by acting at the 5-HT7 receptor. Our results provide for the first time essential evidence that a 5-HT descending system influences voluntary locomotion and that 5-HT7 receptors are involved in the production of coordinated locomotion in adult wild-type mice.
This work was supported by the Canadian Institutes of Health Research (CIHR) grants to L. M. Jordan and K. G. Pearson and a National Science and Engineering Research Council (NSERC) Canadian Graduate Scholarship to J. Liu.
The authors thank J. McVagh and C. Gibbs for technical assistance, M. Setterbom for preparing the figures, and P. Gardiner, D. McCrea, P. Whelan, and B. Schmidt for comments on the manuscript.
Present addresses: J. Liu, University of British Columbia, Brain Research Center, F270 - 2211 Wesbrook Mall, Vancouver, BC, V6T 2B5, Canada; T. Akay, Columbia University Medical Center, Dept. of Biochemistry & Molecular Biophysics, 701 West 168th Street, HHSC 1013, New York, NY 10032.
- Copyright © 2009 the American Physiological Society
- Akay 2006.↵
- Akay 2008.↵
- Antri 2005.↵
- Antri 2002.↵
- Brumley 2007.↵
- Chau 2002.↵
- Cina 1998.↵
- Cowley 1995.↵
- Cowley 1997.↵
- Doly 2005.↵
- Feraboli-Lohnherr 1999.↵
- Gimenez y Ribotta 2000.↵
- Glennon 2002.↵
- Gosgnach 2006.↵
- Grillner 1978.↵
- Hedlund 2003.↵
- Hendricks 2003.↵
- Hochman, 2001.↵
- Jordan 1983.↵
- Jordan 2008.↵
- Jordan 2002.↵
- Kim 2001.↵
- Kjaerulff 1996.↵
- Kriellaars 1994.↵
- Lafreniere-Roula 2005.↵
- Landry 2006.↵
- Liu 2006.↵
- Liu 2005.↵
- Madriaga 2004.↵
- Majczynski 2005.↵
- Myoga 1995.↵
- Nishimaru 2000.↵
- Orlovsky 1999.
- Orsal 1986.↵
- Pang 2000.↵
- Pearlstein 2005.↵
- Pearson 1995.↵
- Pearson 2005.↵
- Pearson 2003.↵
- Pratt 1987.↵
- Roberts 2004.↵
- Rossignol 2006.↵
- Schmidt 2000.↵
- Shefchyk 1985.↵
- Slawinska 2000.↵
- Steeves 1980.↵
- Thomas 2004.↵
- Whelan 2000.↵
- Whelan 1995.↵
- Zhao 2006.↵
- Zar 1974.↵
- Zaporozhets 2006.↵