Cannabinoid-induced increase of quantal size and enhanced neuromuscular transmission

The cannabinoid WIN increases presynaptic vesicle size and transmitter release

WIN (10 µM) and AM630 (10 µM) caused no change in the area or density of postsynaptic nicotinic acetylcholine receptors (AChR) that might explain the increase in mEPP amplitude (Fig. 2a–e; Supplementary Fig. ^S1). WIN also did not enhance mEPP amplitudes by prolonging the actions of acetylcholine in the synaptic cleft (Supplementary Fig. ^S2), a well-described effect of cholinesterase inhibitors. By exclusion, these results suggest that WIN acts presynaptically to increase quantal size. Since the amount of acetylcholine released from each synaptic vesicle cannot be measured directly, we instead tested the effect of drugs that block the filling of synaptic vesicles. Vesamicol (4 µM) inhibits the vesicular acetylcholine transporter, while bafilomycin (0.1 µM) blocks the H⁺-ATPase that is needed to drive the acetylcholine transporter^24,25. Neither of these drugs on their own affected mEPP amplitudes in our recordings (Supplementary Fig. ^S3a and see refs^24,25), but both inhibitors blocked the WIN-induced increase in mEPP and EPP amplitudes (Fig. 2f, Supplementary Fig. ^S3). Thus, the cannabinoid-induced increase in quantal amplitude appears to depend upon the active transport of acetylcholine into synaptic vesicles. Consistent with enhanced vesicle filling, electron microscopy revealed an increase in synaptic vesicle diameter after WIN treatment²⁶(Fig. 2g–i; 43.9 nm vs 47.4 nm; P < 0.01). Assuming spherical geometry, this equates to a 48% increase in vesicle volume (Fig. 2j; see Methods and Supplementary Table ¹). Together, our results strongly suggest that WIN acutely enhances neuromuscular transmission by increasing the loading of synaptic vesicles within the presynaptic nerve terminal.

Endplate potential recordings

The diaphragm with the phrenic nerve attached was quickly dissected, pinned on a Sylgard-coated dish and bathed in oxygenated Ringer’s solution containing physiological calcium levels (in mM: NaCl-136, KCl-5, NaH₂PO₄-1, NaHCO₃-12.8, MgCl₂-1, CaCl₂ and glucose-10, pH adjusted to 7.3). Spontaneous miniature endplate potential (mEPP) and nerve-evoked endplate potential (EPP) recordings were made at room temperature as previously described^28,40. Contraction was blocked using the muscle sodium channel blocker, µ-conotoxin GIIIb (1 μM μCTX, Peptide Institute, Japan). EPP and mEPP recordings were started 1–2 hrs after the diaphragm was placed in the bath solution. Spontaneous mEPP amplitudes were recorded for 1–3 min and were normalized to a resting potential of −80 mV. A train of 40 stimuli (1 Hz) was used for EPP analyses (average amplitude; 20 stimuli for the vesamicol and bafilomycin experiments). EPP amplitudes were normalized to −80 mV and then corrected for non-linear summation⁴¹. Quantal content was calculated by dividing the normalized and corrected EPP amplitudes by the normalized mEPP amplitude for each muscle fiber. Recordings were made from 12–18 muscle fibers from each muscle to derive mean mEPP and EPP values for each sample. Only fibers with a stable resting membrane potential (RMP) ≤ −60 mV were analyzed. To calculate the 90–10% decay times before and after pyridostigmine treatment the average decay time for all sample traces (irrespective of peak amplitude) for n = 3–4 mice was compared.

Drug treatments

In all experiments, drug stock solutions were diluted into the physiological saline solution bathing the preparation. Drugs were obtained from Sigma (MO, USA). Stock solutions for R(+)-WIN 55,212 (WIN), AM251, AM630, L-+/-vesamicol, bafilomycin, anandamide and URB597 were first dissolved in DMSO, stored at −20 °C, and added to the bath 1–2 hrs before the start of recordings. This drug exposure time allowed for dissection of the tissue and set up of the ex-vivo preparation prior to start our recordings. It also ensured enough time for the drugs in the bath to diffuse through the ex-vivo diaphragm muscle preparations. Total DMSO in the Ringers solution was maintained at <0.01%. Control preparations received the same concentration of DMSO vehicle. To test the effect on presynaptic vesicle filling, control recordings were obtained first and then L-+/-vesamicol or bafilomycin were added to the bath. After incubation for approximately 30 min the cannabinoid agonist WIN was also added to the bath before recordings were recommenced a further 60–90 min later.

For the in vivo experiments, compound muscle action potentials (CMAP) were recorded from myasthenic mice (see below). The mice then received an intraperitoneal injection of R(+)-WIN 55,212 (5 mg/kg; diluted in 0.9% saline) and a second set of CMAP recordings were made 3 hrs later. At the dose employed mice showed some signs of hyperactivity but did not display behavior suggestive of intoxication (such as nasal secretions or muscle fasciculation). After the phrenic nerve-diaphragm preparation was dissected it was thoroughly washed for 1.5 hrs in Ringer’s solution to remove traces of the injected WIN prior to baseline EPP recordings. The acute effects of WIN were then tested by adding WIN to the bath solution for 1.5 hrs.

Confocal microscopy, NMJ morphometry and FRET

Immunostaining, imaging (Zeiss LSM510 Meta confocal microscope) and endplate morphometry was performed as described previously^40,42. For comparison of AChR staining intensities, sections were processed together and examined during the same imaging acquisition using identical imaging settings. Confocal Fluorescence Resonance Energy Transfer (FRET) was used to compare the extent to which AChRs were tightly packed within endplate AChR clusters (<10 nm AChR-AChR spacing) as described previously⁴³. FRET efficiency was calculated from the increase of the fluorescence intensity of the donor after the acceptor fluorophor was selectively photobleached.

Electron microscopy and synaptic vesicle measurements

Diaphragms prepared as described above were divided into two hemidiaphragms. The right hemidiaphragm was bathed in WIN (10 µM) for ~2 hrs while the left hemidiaphragm was incubated in control solution (without WIN). Samples were then immersion fixed in 4% paraformaldehyde +2.5% glutaraldehyde in 0.05 M cacodylate buffer (containing 2% sucrose and 0.15 mM CaCl2, pH 7.4) for 10 min (modified from)⁴⁴. Diaphragms were then sliced into approx. 1–2 mm strips (perpendicular to the main nerve) and, fixed for a further 45 min followed by heavy metal staining⁴⁵. Finally, each strip of diaphragm was flat embedded in EPON resin and polymerized overnight at 60 °C. An ultramicrotome (Ultracut7, Leica, Germany) was used to produce sections for LM and EM observations. Semi-thin sections (500 nm) across the entire length of the diaphragm were produced, stained with Toluidine blue and observed under a light microscope: this step aided in trimming the diaphragm down to only the areas located within 2–3 mm of the main nerve (where most NMJs are located). Ultra-thin sections (70 nm) were then produced from the trimmed area, collected on 200 mesh copper grids, stained with Uranyl acetate and Lead citrate for 10 min each before being observed with a TEM, operating at 120 kV (JEM 1400, JEOL, Japan).

Measurements of vesicle diameter were performed on digital micrographs by an operator who was blind to the treatment group and following the protocol outlined in Karunanithi et al.²⁶. Briefly, the outer diameters of synaptic vesicles were measured for at least 200 vesicle profiles, which was sufficient to obtain a characteristic distribution⁴⁶. Only clearly distinguished vesicles that were circular in shape, had a grey core and a uniform membrane were used for the analysis. At least 15 NMJs were sampled from each of four hemidiaphragms/mice. A greater number of micrograph images were analysed from WIN-treated muscles than from control muscles (Control: n = 597 vesicles; WIN: n = 2098 vesicles). Vesicle diameters were corrected using the method of Froesch so as to take into consideration EM section sampling^26,47. Volumes were calculated assuming spherical geometry (Volume = 4/3*π*radius³) based upon the inner diameter of the vesicles (outer diameter minus twice the vesicle membrane thickness (8.9 ± 0.19 nm; n = 103 vesicles).

Mouse model of anti-MuSK myasthenia gravis

The passive IgG transfer model of anti-MuSK myasthenia gravis was described previously^40,48,49. Briefly, mice received daily injections of IgG from anti-MuSK-positive patient 4 (45 mg/day i.p., patient suffering grade 4B weakness – Myasthenia Gravis Foundation of America; IgG batch AM4.4). Passive transfer of IgG from this patient was previously reported to cause overt myasthenic weakness after 12 daily injections⁵⁰.

Electromyography

Compound muscle action potentials (CMAP) were recorded from the gastrocnemius muscle during repetitive stimulation of the sciatic nerve, as previously described⁴⁰. Briefly, the mouse was first anesthetized with 2–3% isoflurane/oxygen. Two ~3 mm custom-made single monopolar recording electrodes were glued to the surface of the skin: one over the dorsal aspect of the gastrocnemius muscle and the second electrode at the ankle of the same hind limb. Electrolyte gel (VIASYS Healthcare, Madison, USA) was applied directly at the electrode sites. Stimulation of the sciatic nerve was accomplished via a 4 mm incision in the sciatic notch and by placing the nerve on custom-made silver hook electrodes (0.6 mm diameter). Ten stimuli were delivered at 3 impulses/sec and three such trains recorded from each muscle were averaged.

This work was supported by the Brain Foundation and Australian Myasthenic Association in New South Wales (to SWR), and Australian Research Council (DP0988227 to DAP). We thank Drs Louise Cole (Advanced Microscopy Facility, Bosch Institute) and Shanker Karunanithi for technical advice and assistance, Joshua A. Way, who performed some mEPP and EPP recordings, the molecular medicine laboratory at Concord Hospital, the apheresis team, Sr Beth Newman, Louise Wienholt and patients from around Australia who contributed their plasma to this research, and Prof Mark Connor for critical comments on the manuscript.