LI Shun, GAO Yi-Ming2,, XU Yue, HU Jia-Qi, TANG Wen-Xu, SUN Xiao-Quan,XUE Lei2,, WANG Wan-Chun6)
(1)Jiangxi University of Traditional Chinese Medicine, Nanchang330004,China;2)State Key Laboratory of Medical Neurobiology and MOE Frontiers Center for Brain Science, Fudan University, Shanghai200433,China;3)Department of Physiology and Neurobiology, School of Life Sciences, Fudan University, Shanghai200438,China;4)Research Institute of Intelligent Complex Systems, Fudan University, Shanghai200433,China;5)Center for Rehabilitation Medicine, Department of Pain Management, Zhejiang Provincial People"s Hospital,Affiliated People"s Hospital, Hangzhou Medical College, Hangzhou310014,China;6)Department of Traditional Chinese Medicine Surgery, Affiliated Hospital of Jiangxi University of Chinese Medicine, Nanchang330006,China)
Abstract Objective Quercetin, a flavonol compound widely distributed in fruits, vegetables, and medicinal plants, has been suggested to act as a neuroprotective agent. In the present study, we investigated the presynaptic effect of quercetin on synaptic transmission and plasticity. Methods Using whole-cell patch clamp and capacitance measurement technique, we recorded miniature excitatory postsynaptic currents (mEPSC), presynaptic calcium influx, vesicle release and recycling, and the replenishment of readily releasable pool (RRP). Additionally, we stimulated the axon with 5-200 Hz and recorded short-term depression (STD) in the postsynaptic neuron. Results We found that 100 μmol/L quercetin in the extracellular solution did not affect the mEPSC amplitude and frequency, indicating a presynaptic mechanism modulating synaptic transmission. At the presynaptic nerve terminals,100 μmol/L quercetin did not induce notable changes in calcium influx or vesicle release, but significantly inhibited clathrindependent slow endocytosis following exocytosis. The inhibition of endocytosis led to a slowdown of vesicle mobilization during stimulation, a reduction in readily releasable pool replenishment after stimulation, and enhancement of short-term depression during high-frequency repetitive stimulation in the postsynaptic principal neurons. Conclusion These results provide new insights into quercetin-modulated neuronal signaling and suggest a protective effect that prevents excessive excitatory synaptic transmission in brain circuits.
Key words quercetin, synaptic transmission, vesicle exocytosis, vesicle endocytosis, capacitance measurement
Synapses are the essential building blocks of the central nervous system (CNS) and responsible for transmitting information between neurons[1]. The highly efficient process of presynaptic vesicle recycling plays a crucial role in maintaining precise synaptic transmission and plasma membrane homeostasis[2]. When the action potential (AP) arrives at the nerve terminal and activates the voltagedependent calcium channels, the calcium influx triggers the neurotransmitter-containing vesicles to fuse with the presynaptic membrane, leading to exocytosis[3]. In the mammalian CNS, neurons can fire in the range of 10 to several hundred Hertz.However, typical presynaptic terminals only contain several hundred synaptic vesicles per active zone(AZ)[4]. Therefore, efficient endocytosis following exocytosis is critical to prevent exhaustion of vesicles and to help maintain the efficiency of synaptic transmission[5].
Quercetin is a flavonol compound present in traditional Chinese medicine but also widely found in various fruits, vegetables, and medicinal plants[6].Recently, cumulative evidence has shown that quercetin can penetrate the blood-brain barrier and has multiple physiological effects, such as in antioxidative stress, the anti-inflammatory response, and promotion of nerve regeneration[7-9]. Quercetin may also have therapeutic effects on ischemia and hypoxiainduced brain injury and Alzheimer’s disease,suggesting great potential in the treatment of neurodegenerative diseases[10-11].
Previous studies have reported that quercetin can modulate neuronal excitability. Renet al.[12]showed that quercetin can reduce the firing rate of pyramidal neurons in the prefrontal cortex. Toyotaet al.[13]observed similar phenomena in which acute administration of quercetin inhibited the firing rate in trigeminal ganglion cells. In rat retinal ganglion cells,quercetin can reduce the frequency of miniature excitatory postsynaptic currents (mEPSCs) and increase the frequency of miniature inhibitory postsynaptic currents (mIPSCs) in both ON and OFF type cells to suppress the excitotoxicity caused by glaucoma[14]. In addition, quercetin has been shown to enhance the paired-pulse ratio (PPR) and long-term potentiation (LTP) in the rat hippocampus, relieving chronic lead exposure-induced impairment of synaptic plasticity[15]. On the other hand, the quercetinmodulated presynaptic mechanisms remain largely unclear. Luet al.[16]showed that quercetin inhibits Ntype and P/Q-type calcium channels and reduces excessive glutamate release in the synaptic boutons of the cerebral cortex, exerting a neuroprotective effect.Another study reported an opposite conclusion in the rat hippocampal CA1 neurons that quercetin promoted presynaptic glutamate release by reducing the postsynaptic PPR, leading to a dose-dependent increase in the excitatory postsynaptic current (EPSC)amplitude[17]. The uncertainty surrounding quercetinmodulated presynaptic mechanisms requires direct electrophysiological recordings at the presynaptic nerve terminals.
In the present study, we examined how quercetin modulates synaptic transmission at a giant glutamatergic synapse, the calyx of Held, located at the medial nucleus of the trapezoid body (MNTB) in the brainstem[18]. The large presynaptic nerve terminal allows for direct recordings of presynaptic calcium current (ICa) and calcium-triggered vesicle exoendocytosisviacapacitance measurements at a high temporal resolution[19-20]. Our findings show that quercetin inhibits vesicle endocytosis in a calciumindependent manner, which leads to slowdown of vesicle mobilization and the readily releasable pool(RRP) replenishment. Furthermore, quercetin enhances short-term depression (STD) in a frequencydependent manner and dramatically reduces the EPSC amplitude upon high-frequency stimulation. Our results provide new insights into quercetin-modulated neuronal signaling and short-term plasticityviainhibition of presynaptic vesicle endocytosis, which leads the reduced synaptic transmission and enhanced STD.
1.1 Slice preparation
Sprague-Dawley rats (8-10 d old, p8-p10) of either sex were decapitated and tissue blocks containing the MNTB placed in a low-Ca2+artificial cerebrospinal fluid (ACSF) solution (125 mmol/L NaCl, 25 mmol/L NaHCO3, 3 mmol/L myo-inositol,2 mmol/L Na-pyruvate, 2.5 mmol/L KCl, 1.25 mmol/L NaH2PO4, 0.4 mmol/L ascorbic acid, 25 mmol/L glucose, 3 mmol/L MgCl2, and 0.05 mmol/L CaCl2).Brain slices (~200 μm thick) were sectioned using a vibratome (VT 1200s, Leica, Germany) and incubated in normal ACSF with 2 mmol/L CaCl2at 37°C for 30-40 min before experiments. Quercetin was diluted in dimethylsulfoxide (DMSO) at a concentration of 10 mmol/L and stored frozen (-20°C). For use, the stock solution was vortexed to ensure complete mixing of components and diluted into the extracellular solution to a final working concentration of 10-250 μmol/L,with a DMSO concentration of 0.1% maintained in all solutions. All electrophysiological recordings were made at room temperature (22-24°C).
All of the methods were carried out in accordance with approved guidelines, and all animal experimental protocols were approved by the Animal Care and Use Committee of Fudan University and Jiangxi University of Chinese Medicine.
1.2 Electrophysiology
The voltage-clamp recordings of AMPA receptormediated mEPSCs and EPSCs were obtained using an EPC-10 amplifier (HEKA, Lambrecht, Germany).The pipette (2-3 MΩ) solution contained 125 mmol/L K-gluconate, 20 mmol/L KCl, 4 mmol/L Mg-ATP,10 mmol/L Na2-phosphocreatine, 0.3 mmol/L GTP,10 mmol/L HEPES, and 0.5 mmol/L EGTA (pH 7.2,adjusted with KOH). The extracellular solution was similar to the ACSF except with 2 mmol/L CaCl2.The series resistance (<10 MΩ) was compensated by 95%(lag 10 μs) throughout the experiment. The mEPSCs were recorded at a holding membrane potential of-80 mV with 0.5 μmol/L tetrodotoxin (TTX).Bicuculline (10 μmol/L), D-APV (10 μmol/L), and strychnine (10 μmol/L) were included in the extracellular solution to inhibit the GABAA receptor,NMDA receptor, and glycine receptor-mediated currents. For EPSC recordings, a bipolar electrode was placed near the midline of the MNTB.Stimulation pulses (0.1 ms, 2-20 V with AM2100,A-M systems, USA) at 5-200 Hz were delivered to evoke APs at the presynaptic nerve terminal, which induced an AMPA receptor-mediated EPSC at the principal neuron of the MNTB.
The presynaptic membrane capacitance measurements were obtained using an EPC-10 amplifier with lock-in software. The presynaptic pipette (3.5-5 MΩ) solution contained 125 mmol/L Cs-gluconate, 20 mmol/L CsCl, 4 mmol/L Mg-ATP,10 mmol/L Na2-phosphocreatine, 0.3 mmol/L GTP,10 mmol/L HEPES, and 0.05 mmol/L BAPTA(pH 7.2, adjusted with CsOH). The series resistance(<10 MΩ) was compensated by 65% (lag 10 μs). Slow endocytosis was induced by a 20-ms depolarization pulse from -80 mV to +10 mV (depol20ms), and rapid endocytosis was induced by 10 depol20msat 10 Hz(depol20ms×10). To measure the vesicle mobilization during stimulation, 10 pulses of depol20msat 1 Hz were applied.
1.3 Data analysis
The Mini Analysis Program (version 6.07,Synaptosoft, USA) was used to analyze the mEPSCs.Capacitance jumps were measured 250 ms after depolarization to avoid artifacts. The initial rate of endocytosis (Rateendo) was measured within 2 s after depolarization with depol20msor depol20ms×10. The residual capacitance 15 s after depol20ms(ΔCm15s) or 30 s after depol20ms×10(ΔCm30s) was measured to represent endocytosis recovery. All data were expressed as mean±SE.Pvalues were determined by Student’st-test, one-way ANOVA with Bonferroni’s multiple comparisons test, or Kruskal-Wallis test, andP<0.05 was considered significant.
2.1 Quercetin does not affect the sensitivity of postsynaptic AMPA receptors
To investigate how quercetin affects synaptic transmission in the CNS, we examined the mEPSCs of the principal neurons at the rat calyx of Held synapses in the brain stem. The postsynaptic neurons were held at a resting potential of -80 mV in the whole-cell configuration with 0.5 μmol/L TTX in the extracellular solution. Bicuculline (10 μmol/L),D-APV (10 μmol/L), and strychnine (10 μmol/L)were also included. In control p8-p10 rats, the mean mEPSC amplitude and frequency were (30.7±2.6) pA and (0.8±0.2) Hz (722 events,n=5; Figure 1a-c). The mean 10%-90% rise time and 20%-80% decay time were (0.34±0.02) ms and (1.58±0.05) ms, respectively(Figure 1d). These results were similar to our previous studies at calyx-type synapses[21-22]. In the quercetintreated groups, we applied 10 μmol/L quercetin to the extracellular solution at least 30 min before recordings were made. The mEPSC amplitude and frequency were not significantly different from the control group (amplitude: (33.2±2.3) pA,P=0.84;frequency: (0.8±0.3) Hz,P=0.99; 884 events,n=6;Figure 1a-c). We also found no significant differences in the 10%-90% rise time and 20%-80% decay time(rise time: (0.36±0.03) ms,P=0.53; decay time: (1.65±0.10) ms,P=0.24;n=6; Figure 1d). Previous studies have tested quercetin concentrations up to 200-300 μmol/L[16,23]. Therefore, we also investigated 100 and 250 μmol/L extracellular quercetin. However,neither mEPSC amplitude and frequency nor 10%-90% rise time and 20%-80% decay time showed significant differences among the control and two high-concentration quercetin-treated groups (Figure 1c-d). We also plotted the cumulative probability curve of the mEPSCs in the control and quercetintreated groups and found no significant differences among all four groups (Kruskal-Wallis test,P=0.31;Figure 1e). Therefore, we conclude that quercetin does not affect the spontaneous mEPSCs, suggesting that quercetin does not alter the sensitivity of the postsynaptic AMPA receptors of principal neurons at calyces.
Fig. 1 Quercetin does not affect the sensitivity of postsynaptic AMPA receptors
Fig. 2 Quercetin inhibits presynaptic vesicle endocytosis at the calyx-type synapses
Fig. 3 Quercetin slows down vesicle mobilization and RRP replenishment
Fig. 4 Quercetin enhances short-term depression under intense stimulation
2.2 Quercetin inhibits presynaptic vesicle endocytosis at the calyx of Held synapses
Having shown that quercetin has no effect on the postsynaptic mEPSCs, we further investigated whether quercetin affects synaptic transmission from the presynaptic site. At the calyx-type synapse, the giant presynaptic nerve terminal provides a unique opportunity to study the presynaptic key parameters modulating synaptic transmission, including the calcium influx, vesicle exocytosis, and endocytosis[20].
At the presynaptic nerve terminal, APs depolarize the presynaptic membrane and induce calcium influx to trigger neurotransmitter release[2].Endocytosis, which retrieves the membrane released by its immediately preceding exocytosis, is also of vital importance in maintaining the efficiency of synaptic transmission[24]. Here, we measured theICa and vesicle exo-endocytosis at the presynaptic nerve terminal of the calyx-type synapse with direct capacitance recordings. As we have already shown that quercetin does not affect the mEPSCs in a concentration range of 10-250 μmol/L (Figure 1), we chose 100 μmol/L extracellular quercetin, a relatively high concentration, for all of the studies in the presynaptic nerve terminal.
Our previous studies showed that the kinetics of vesicle exo-endocytosis critically depend on stimulation intensities[25-26]. Specifically, depol20mscan deplete the RRP and induce clathrin-dependent,dynamin-dependent slow endocytosis[2], whereas depol20ms×10can induce exocytosis to a greater extent and a rapid dynamin-dependent, but clathrinindependent, form of endocytosis[27]. In control rats,depol20msinduced aICa and capacitance jump reflecting exocytosis (ΔCm) of (2.1±0.1) nA and(511±27) fF (n=6; Figure 2a-c), respectively, which were similar to our recent study[20]. The immediately subsequent capacitance decay could be fitted monoexponentially with a time constant (τ) of (16.5±0.1) s (n=6), showing slow endocytosis. The Rateendomeasured 1-2 s after depol20mswas (42±4) fF/s (n=6;Figure 2d). In the quercetin-treated group, depol20msinduced anICa and exocytosis of (2.0±0.2) nA and(573±37) fF (n=6), respectively, which were not significantly different from controls (ICa:P=0.75;ΔCm:P=0.24; Figure 2a-c). However, endocytosis was inhibited in the quercetin-treated group. Rateendowas slightly but not significantly reduced ((37±5) fF/s,P=0.56;n=6), and the capacitance decay did not return to baseline within 30 s, which is difficult to fit monoexponentially (Figure 2a, e). The remaining capacitance jump 15 s after depol20mswas (56±4)%(ΔCm15s%;n=6; Figure 2e) of the maximum ΔCm induced by depol20ms, which was significantly higher than in controls ((40±2)%,n=6;P=0.01), indicating that quercetin inhibits slow endocytosis.
Next, we applied intensive stimulation of depol20ms×10to induce the clathrin-independent but dynamin-dependent rapid endocytosis at calyces. In controls, depol20ms×10induced a much larger calcium influx (measured by calcium charge, QICa; (316±19) pC;n=9), leading to a higher capacitance jump((1 461±97) fF,n=9; Figure 2f-h). The capacitance decay reflecting endocytosis could be fit biexponentially with rapid and slow time constants of(2.0±0.1) s and (20.6±0.1) s, respectively. The Rateendorepresenting rapid endocytosis was (238±24) fF/s(n=9). In the presence of 100 μmol/L quercetin in the extracellular solution, depol20ms×10induced a similar calcium influx and capacitance jump as in controls(QICa: (326±27) pC,P=0.78; ΔCm: (1 526±101) fF,P=0.66;n=9; Figure 2f-h). The rapid endocytosis was also slightly but not significantly reduced (Rateendo:(222±13) fF/s,P=0.61;n=9; Figure 2i). However, the residual capacitance jump measured 30 s after depol20ms×10(ΔCm30s%) was significantly higher than in controls (Control: (16±3)%; Quercetin: (32±3)%;P=0.002;n=9; Figure 2j), indicating inhibition of slow endocytosis. A previous study demonstrated that calcium initiates all forms of endocytosis[28]. In our study, we found that quercetin inhibited slow endocytosis while not affecting theICa and exocytosis. Therefore, our results may indicate that quercetin inhibits presynaptic vesicle endocytosis,especially slow endocytosis, in a calcium-independent manner.
2.3 Quercetin slows vesicle mobilization and RRP replenishment
Recruiting vesicles to the RRP after exocytosis is critical for RRP replenishment. Inhibition of endocytosis may slow replenishment down. To test this possibility, we applied 10 pulses of depol20msat 1 Hz at the presynaptic nerve terminal of calyces and examined the RRP recovery during the stimulation.The 1 s between pulses, which was much longer than for 10 Hz stimulation, could help us measure the endocytosis rate more accurately during stimulation.The first depol20msdepleted the RRP[29], and the capacitance jump evoked by the 2ndto the 10thdepolarizing pulse would reflect vesicle mobilization to the RRP. In the presence of 100 μmol/L quercetin in the extracellular solution, we found that the RRP replenishment rate was significantly reduced starting from the 2nddepol20ms(Figure 3a, b). A previous study showed that calcium/calmodulin sped up both the endocytosis rate and vesicle recruitment to the RRP[28,30]. However, our results demonstrated that the reduced capacitance jump in the presence of quercetin was not caused by changes in calcium influx,suggesting that quercetin may directly inhibit RRP replenishment by inhibiting endocytosis.
To further consolidate our conclusion, we measured the RRP replenishment at various times (Δt=0.1-30 s) after depol20msand plotted the RRP replenishment curve. In controls, the RRP replenishment could be fitted double exponentially with a rapid time constant of 0.2 s and slow time constant of 7.1 s, consistent with previous reports(Figure 3c, upper; Figure 3d)[25,30]. In the presence of 100 μmol/L quercetin in the extracellular solution, the RRP recovery was much slower than in controls, with a rapid time constant of 0.2 s and a slow time constant of 13.1 s (Figure 3c, lower; Figure 3d). The slowing of RRP replenishment further confirmed that quercetin reduces RRP replenishment by inhibiting endocytosis.
2.4 Quercetin enhances short-term depression
The quercetin-induced inhibition of endocytosis and replenishment of RRP may lead to reduced postsynaptic responses. We investigated this possibility with fiber stimulation at the midline of the trapezoid body to evoke APs at the calyx-type synapses and recorded the EPSCs from the postsynaptic principal neuron[31-32]. At calyces, basal neuronal firing can be increased to 600-800 Hz upon stimulation[18,33]. Thus, we evoked 40 APs at a range of physiological stimulation frequencies at calyces, 5 to 200 Hz[34]. During the stimulation, the EPSC amplitude rapidly declined to a steady-state level(Figure 4a-d). We calculated the steady-state depression ratio (EPSCss/EPSC1) by averaging the EPSCs from the 31stto the 40thEPSC. As shown in Figure 4e, the steady-state depression ratio decreased as the stimulation frequency increased. In the presence of 100 μmol/L quercetin in the extracellular solution, the depression ratio was not significantly different from that of the control group at a low stimulation frequency of 5 Hz (Figure 4a, e).However, at higher stimulation frequencies (50-200 Hz), quercetin significantly enhanced synaptic depression (50 Hz: ctrl (4.5±1.1)% (n=7) and quercetin (1.4±0.5)% (n=6),P=0.03; 100 Hz: ctrl(3.4±0.5)% (n=6) and quercetin (1.4±0.5)% (n=6),P=0.02; 200 Hz: ctrl (3.0±0.6)% (n=6) and quercetin(1.0±0.4)% (n=4),P=0.049; Figure 4e). These results demonstrate that inhibition of vesicle endocytosis and RRP replenishment led to reduced EPSCs during highfrequency stimulation, with the steady-state depression being more pronounced at higher stimulation frequencies. In addition, we also calculated the PPR from the first two EPSC events of STD recordings both in the control and quercetintreated groups. We found PPR changed after quercetin treatment, further confirming that quercetin was involved in the presynaptic modulation of synaptic transmission.
Efficient synaptic transmission in the CNS critically relies on presynaptic vesicle exo- and endocytosis. In the present study, we report that quercetin, a flavonol compound commonly found in fruits, vegetables, and medicinal plants, can inhibit vesicle endocytosis at rat calyx-type synapses located in the brainstem. Inhibition of endocytosis leads to slowed vesicle mobilization during stimulation, a reduction in RRP replenishment after stimulation, and enhancement of STD during high-frequency repetitive stimulation, a form of short-term plasticity in the postsynaptic principal neurons.
Accumulated studies have demonstrated that quercetin can modulate neuronal excitabilityviadifferent underlying mechanisms in the CNS. In acutely isolated rat hippocampal neurons, quercetin has been reported to reduce voltage-dependent sodium currents and delay depolarization to exert neuroprotective effects[35]. Recently, TWIK-related potassium channel-1 (TREK-1) was shown to be a new target of quercetin action, which can control neuronal excitability by keeping the membrane potential below the depolarization threshold[36]. Renet al.[12]showed that quercetin increased TREK-1 current by activating PKC and reduced the excitability of pyramidal neurons in the prefrontal cortex, thereby alleviating manic behavior in mice. In addition,quercetin has been observed to modulate other channel activities in the CNS, including calcium channels[16], BK channels (large-conductance Ca2+-regulated potassium channels)[37], and TRPV1 channels[38]. The quercetin-modulated neuronal activities may further lead to modulation of the efficiency of synaptic transmission and plasticity in different brain regions. However, many studies of quercetin-modulated synaptic transmission are still controversial[16-17]due to a lack of direct evidence from the presynaptic side. The possible mechanisms for presynaptic function mainly come from the mEPSC/mIPSC frequency recordings or PPR evaluations.
APs depolarize the presynaptic membrane,inducing calcium influx, which triggers neurotransmitter release[39-40]. Endocytosis is crucial to maintaining the efficiency of synaptic transmission by retrieving the membrane released by its immediately preceding exocytosis[41-42]. Our previous studies have shown that the kinetics of vesicle exoendocytosis critically depend on stimulation intensities[26,28]. Two major forms of vesicle endocytosis were investigated in the present study.Depol20mscan deplete the RRP and induce clathrindependent, dynamin-dependent slow endocytosis[2,43-44], whereas depol20ms×10can induce a rapid dynamin-dependent, but clathrin-independent, form of endocytosis[27-28,43,45]. We found that administration of quercetin in the extracellular solution did not affect theICa and exocytosis, but slightly reduced the rapid endocytosis and significantly inhibited the slow endocytosis (Figure 2). The slowdown of endocytosis further inhibited vesicle mobilization and RRP replenishment, which may provide a protective effect that prevents excessive excitatory synaptic transmission.
We found that quercetin enhanced the synaptic depression during high-frequency repetitive stimulation because of slowed endocytosis (Figure 4).A previous study reported that, in synapsin I/II double knock-out (DKO) mice, the kinetics of vesicle exoendocytosis are not altered. However, the STD was dramatically enhanced in DKO mice during highfrequency stimulation[31]. Synapsin proteins may help boost the release probability during repetitive stimulation. Whether quercetin can also inhibit the synapsin function and contribute to the enhanced STD remains unclear.
Our previous study showed that calmodulin is the calcium sensor of endocytosis, and calcium/calmodulin facilitates RRP replenishment, which is critical in maintaining release during repetitive firing[28]. However, in the present study, we found that quercetin inhibits vesicle endocytosis and RRP replenishment while not affecting the calcium influx and exocytosis. This observation is somewhat surprising because it is not consistent with the calcium/calmodulin-dependent mechanism.Interestingly, a previous study also reported that the RRP replenishment time course was slowed by more intense stimulation, which presumably should induce more calcium influx and activate calmodulin to a higher level[25]. We also confirmed this phenomenon with similar protocols[30]. In addition, calcineurin,which is activated by calcium/calmodulin, has been shown to control the speed of both rapid and slow endocytosis by dephosphorylating endocytic proteins[46]. Whether calcineurin is also involved in the calcium/calmodulin-independent mechanism needs to be further clarified. Thus, it would be of great interest to explore a possible mechanism other than the calcium/calmodulin-dependent mechanism in future studies, which may help elucidate the generation of postsynaptic short-term plasticity upon intense stimulation.
In conclusion, we investigated the quercetinmodulated presynaptic mechanisms at a glutamatergic central synapse. By uncovering the precise kinetics in synaptic transmission and plasticity, our research may provide new insights for clinical application of quercetin.