Abnormal Neuronal Excitability and Reduced Parvalbumin Expression in Shank3-Deficient Parvalbumin Neurons of the Thalamic Reticular Nucleus
- Yubo Lai1,2,*,
- Lulu Lü2,*,
- Runfeng Sun2,*,
- Junhui Lai2,
- Yang Wang2,
- Silin Chen2,
- Shan Li1,
- Qinglian Liu3, and
- Lei Zhou2
- 1South China University of Technology, Guangzhou 510006, China
- 2Institute of Molecular Physiology, Shenzhen Bay Laboratory, Shenzhen 518132, China
- 3Department of Physiology and Biophysics, School of Medicine, Virginia Commonwealth University, Richmond, Virginia 23298
-
Author contributions: S.L., Q.L., and L.Z. designed research; Y.L., L.L., R.S., J.L., Y.W., S.C., and L.Z. performed research; Y.L., L.L., R.S., J.L., Y.W., S.C., and L.Z. analyzed data; L.Z. wrote the paper.
Abstract
Fast-spiking, nonadaptive inhibitory neurons in the thalamic reticular nucleus (TRN) critically gate the reciprocal communication between the thalamus and the cortex. Parvalbumin (PV) neurons express high levels of PV, the sole role of which appears to be calcium buffering. The significance of the PV protein—and its related high calcium-buffering capacity—under pathological conditions, especially in various neuropsychiatric disorders, is underappreciated. Deficiency of SHANK3, an important neuronal protein containing ankyrin, SH3, and PDZ, three canonical domains for protein recognition, causes behavioral changes relevant to autism spectrum disorders (ASDs). Here we report TRN PV neurons of Shank3−/− (exon 4–22 deletion) mice of either sex exhibit pronounced increases in burst firing occurrence, decreased tonic firing frequency, and faster dendritic calcium transient decay. We pinpointed reduced PV expression as the culprit and used the added buffer approach to confirm the decrease in calcium-buffering capacity in mutant neurons. Conversely, supplementing Shank3−/− PV neurons with extra EGTA reverses the abnormal action potential (AP) firing. In addition, the PV neurons from HCN2−/− mice exhibit consistent changes in neuronal excitability, PV expression, and calcium signaling. Together with the study of dopaminergic (DA) neurons in the ventral tegmental area (VTA), these results uncover reduced PV expression, calcium-buffering capacity, and altered neuronal excitability in Shank3−/− and HCN2−/− mice. This pathway, downstream of Shank3 deficiency and HCN channelopathy, may form an important pathological basis not only for ASD but also other neuropsychiatric disorders.
Significance Statement
SHANK3 is a scaffolding protein that is highly enriched in the postsynaptic density of synapses. Mutations and deletions of the SHANK3 gene are directly connected to Phelan–McDermid syndrome and autism spectrum disorders. However, the links between genetic alterations and abnormalities at the cellular, network, and behavioral levels remain unclear. This study uncovered abnormal physiological changes in inhibitory neurons in the thalamus. A definitive link at the cellular level is established between the Shank3 protein deficiency and the pathological basis of related neuropsychiatric disorders.
Introduction
The thalamus plays a central role in the process and relay of sensory information as well as in regulating sleep, consciousness, motor control, and memory (Steriade et al., 1993; Sherman and Guillery, 2001, 2002; Wimmer et al., 2015). TRN neurons are clustered within a shell-like structure, encircling the core of the thalamus where the thalamocortical (TC) neurons reside (Pinault, 2004). TRN neurons are mostly GABAergic, and their classification involves merits of transcriptome, function, and morphology (Li et al., 2020; Martinez-Garcia et al., 2020). TRN neurons receive excitatory inputs from TC neurons and corticothalamic neurons in the cortex but they only innervate TC neurons within the thalamus (Crandall et al., 2015; Hou et al., 2016; Hadinger et al., 2023). Thus, in the cortico-thalamo-cortical (CTC) loop, TRN functions as the gatekeeper and controls the information flow as well as oscillations within the CTC loop (Fogerson and Huguenard, 2016; Clemente-Perez et al., 2017; Crunelli et al., 2018; Shepherd and Yamawaki, 2021).
Abnormalities in the thalamus have been linked to various neuropsychiatric disorders, including autism spectrum disorders (ASDs; Nair et al., 2013, 2015). Autistic patients show anatomical changes in the thalamus, including an expanded surface area in the right posterior region and increased functional connectivity between the thalamus and the precentral gyrus (Schuetze et al., 2016; Karavallil Achuthan et al., 2023). Using a strain of Shank3 knock-out mice (exon 13–16 deletion) as a rodent model for ASD, we investigated changes in neuronal excitability of TC neurons in the thalamus (Zhu et al., 2018). ShankΔ13–16 TC neurons exhibit decreased resting membrane potential (RMP), increased input resistance (Rin), imbalance in excitatory/inhibitory synaptic transmission, and downregulation of HCN channel expression and function. Excitingly, a recent study reports that restoring HCN2 function in Shank3−/− TC neurons is able to correct sensory hypersensitivity and sleep fragmentation in mutant mice (Guo et al., 2024). In addition, TRN dysfunction has been found in various neuropsychiatric diseases and contributes to sensory abnormalities, attention deficits, and sleep disorder (Krol et al., 2018). Abnormalities in the physiology of Shank3−/− TRN neurons remain to be clarified.
TRN PV neurons exhibit features shared by PV neurons from other regions, including high frequency of AP firing in tonic mode, almost none calcium-dependent adaptation, and minimal membrane sag upon hyperpolarizing stimulations (Hu et al., 2014; Clemente-Perez et al., 2017; Perumal and Sah, 2021). PV is a calcium-binding protein of ∼110 aa (Permyakov and Uversky, 2022). Unlike calmodulin, which undergoes significant structural changes upon calcium binding and is involved in various signaling pathways, the PV protein adopts a globular structure and mainly functions as a buffer for calcium ions.
It is often overlooked that ∼99% of the total calcium ions within the cytoplasm are bound to various calcium-binding proteins, including PV, calbindin, and calretinin (Brinley, 1978; Matthews and Dietrich, 2015; Schwaller, 2020). The calcium-buffering system critically determine the spatiotemporal profile of calcium signaling: it dampens sudden increases in intracellular calcium and prolongs the decrease in calcium concentration by releasing calcium back into the cytosol during the decay phase (Fakler and Adelman, 2008). PV has been reported to play an important role in maintaining the normal calcium levels, especially under stress and disease conditions. Downregulation of PV expression and pathological damages to PV neurons have been reported in ASD and various neurodegenerative diseases (Marin, 2012; Steullet et al., 2018; Filice et al., 2020; Ruden et al., 2021).
Here we propose that the PV protein in TRN PV neurons is not merely a molecular marker but rather plays an important role in regulating the spatiotemporal profiles of intracellular calcium dynamics. Using the Shank3Δ4–22 as the animal model, we explored the downstream events caused by Shank3 deficiency in TRN PV neurons, including HCN channelopathy, reduced PV protein expression and dampened calcium-buffering capacity, and correlated the results with the observations in VTA DA neurons.
Materials and Methods
Ethical approval
All animal experiments were approved by the IACUC Committee of Shenzhen Bay Laboratory (SZBL, #D-2020-0021). The number of animals and procedures involving pain were minimized in compliance with regulations of SZBL. Animals were subjected to regular veterinary care on a daily basis. The Center for Research Animal Resources was available 24 h for any case in which veterinary care or consultation was required. Mice were housed in a standard 12 h light/dark cycle, with ad libitum access to food and water. C57BL/6J mice were used in the current study. The Shank3 knock-out mice were purchased from The Jackson Laboratory (B6.Cg-Shank3tm2.1Bux/J, Stock No: 017688; jax.org; Drapeau et al., 2018). The brain-specific HCN2 knock-out mice were generated by crossing the Nestin-Cre mice [B6.Cg-Tg(Nes-Cre)1Kln/J; Stock No: 003771; Jackson Laboratory] with Floxed-HCN2 mice (exons 3–4 floxed, kindly provided by Dr. Peter McNaughton from King's College, London; Emery et al., 2011).
Acute brain slice preparation and whole-cell patch-clamp recording
Coronal brain slices were prepared from mice 3–6 weeks old of either sex. In brief, mice were deeply anesthetized by intraperitoneal injection of 1.25% tribromoethanol (Avertin; 0.2 ml/10 g body weight) and then transcardially perfused through the ascending aorta with preoxygenated (95% O2 and 5% CO2) and ice-cold sucrose-based artificial cerebrospinal fluid (sucrose-ACSF). Brain slices of 300 μm thickness were prepared using a vibratome (VT1200S, Leica) at 4°C. Slices were incubated first at 34°C for 30 min in standard ACSF and subsequently kept at room temperature for at least 1 h prior to recording.
During recording, brain slices were perfused continuously in the recording chamber with prewarmed (34°C) carbonated standard ACSF at a rate of 2–3 ml/min. Visually identified healthy neurons were recorded with glass pipettes (4–8 MΩ) filled with the internal solution. Voltage and current signals recorded from neuron soma were amplified by MultiClamp 700B (Molecular Devices), digitized by an A/D converter (NI USB-6341, National Instruments), and acquired by the WinWCP program (developed by Dr. John Dempster, University of Strathclyde). All electrophysiological data is recorded with the amplifier filter set at 10 kHz and digitized at 20 kHz. RMP and Rin data are recorded immediately after membrane rupture. Action potential and other signals were recorded from neurons that have remained stable for at least 5 min after membrane rupture and the formation of the whole-cell recording mode. Data were off-line converted to the Axon file format and analyzed by the program of Clampfit (Molecular Devices). Neurons with an RMP more negative than −50 mV and the peak of action potential (AP) passing 0 mV were selected for recording. Rin was measured at RMP based on the change in membrane potential in response to a 10 pA current injection (500 ms). APs were elicited with depolarizing or hyperpolarizing current pulses according to the description in the text. AP and other electrophysiology parameters were analyzed by Clampfit (Fig. S1A).
Two-photon calcium imaging
An Olympus two-photon microscope (FVMPE-RS) with custom IR-DIC was used for calcium imaging in neuron soma and dendrites. Femtosecond pulsed laser (Spectra-Physics) at 810 nm was used for the excitation of calcium dyes (Fluo-4 and Fluo-5F) and AF594, a morphology tracer. Fluorescence signals were collected by a 25× water immersion objective lens for multiphoton recording (ND 1.05, WD 2.0 mm, 400–1,600 nm applicable, Olympus) and recorded through the green channel (495–540 nm emission filter) and the red channel (575–645 nm filter) by two GaAsP PMTs (RNDD3G; RNDD4G). The TTL signal generated by the WinWCP program was used to trigger the recording of time-lapse images at the frequency of 25 frames/second. To minimize photobleaching, the laser power was typically set in the range of 3.5–5%. To approach optimal intensity of each frame, Galvan instead of Resonant scanner was used and the exposure time was in the range between 10 and 20 μs/pixel. The size of ROI was adjusted so that the frame rate can reach 25 Hz. The fluorescence intensity of each ROI was monitored on-line during each experiment and analyzed off-line using Clampfit. Background fluorescence was subtracted by manual selection of an area away from the electrode, soma, and dendrites. Optical and electrical recordings were carried out simultaneously. Burst firing of TRN neurons was triggered by three or four current injections (500 pA/10 ms duration/1 s interval) at the RMP of −70 mV. No averaging and filtering were applied to the fluorescence signal during data analysis.
Ratiometric calcium imaging
Fura-8, a high-affinity ratiometric calcium indicator, was alternatively excited by two LED light sources, 365 and 410 nm. Fluorescence signals were collected by a 40× water immersion lens (CFI Apo NIR 40× W, Nikon), passed through the filter cube, and detected by a back-illuminated sCMOS camera (Dhyana 400BSI by Tucsen). The exposure time and frame rate were set at 20 ms and 40 Hz, respectively. TTL signals generated by the WinWCP program were used to synchronize the electrophysiology amplifier (AM2400, A-M Systems), the LED light sources, and the sCMOS camera. For each recording episode, a total of 390 images were recorded at 40 Hz (195 data points at 20 Hz). Burst firing of TRN neurons was triggered by three current injections (500 pA/10 ms duration/10 ms interval) at the RMP of −70 mV.
The procedure for calculating the free calcium centration was based on published studies (Zhou and Neher, 1993; Maravall et al., 2000; Delvendahl et al., 2015; Helmchen and Tank, 2015). The endogenous calcium buffer power was estimated by the “added buffer” approach where exogenous calcium buffers at different capacities compete with endogenous buffers. This method relies on the time-dependent diffusion of calcium dyes added in the pipette solution into the cytosol, which occurs in the timescale of minutes. One sweep of calcium transients was recorded at each time point and collectively the changes in the amplitude and the decay time constant reflect the increase in total calcium buffer capacity. The analysis involves the calculation of fluorescence intensity (Fi) and the effective binding constant (Keff), as outlined in Equations 1 and 2.
Step 1: intracellular fluorescence intensity Fi
Fi can be calculated using the following formula:<mml:math display="block"><mml:mi>F</mml:mi><mml:mi>i</mml:mi><mml:mo>=</mml:mo><mml:mi>F</mml:mi><mml:mn>1</mml:mn><mml:mo>+</mml:mo><mml:mi>α</mml:mi><mml:mi>F</mml:mi><mml:mn>2</mml:mn><mml:mo>,</mml:mo><mml:mspace
width="2em"></mml:mspace><mml:mo stretchy="false">(</mml:mo><mml:mn>1</mml:mn><mml:mo stretchy="false">)</mml:mo></mml:math>Fi=F1+αF2,(1)
where F1 and F2 represent the fluorescence intensity measured at 365 and 410 nm, respectively, and α represents the isocoefficient so that Fi is independent of calcium concentration. The background fluorescence intensity was subtracted from all measurements. Fi serves as an indicator independent of calcium concentration and reflects the concentration of Fura-8 inside the neuron. After
the rupture of the membrane seal and the formation of the whole-cell patch-clamp recording mode, the Fura-8 gradually diffuses
into the cell and the value of Fi gradually increases.
Step 2: effective binding constant Keff of Fura-8
<mml:math display="block"><mml:msub><mml:mi>K</mml:mi><mml:mrow><mml:mi mathvariant="normal">eff</mml:mi></mml:mrow></mml:msub><mml:mo>=</mml:mo><mml:msub><mml:mi>K</mml:mi><mml:mi>D</mml:mi></mml:msub><mml:mstyle
displaystyle="true" scriptlevel="0"><mml:mrow><mml:mfrac><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mo movablelimits="true"
form="prefix">max</mml:mo></mml:msub><mml:mo>+</mml:mo><mml:mi>α</mml:mi></mml:mrow><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mo
movablelimits="true" form="prefix">min</mml:mo></mml:msub><mml:mo>+</mml:mo><mml:mi>α</mml:mi></mml:mrow></mml:mfrac></mml:mrow><mml:mo>,</mml:mo><mml:mspace
width="2em"></mml:mspace><mml:mo stretchy="false">(</mml:mo><mml:mn>2</mml:mn><mml:mo stretchy="false">)</mml:mo></mml:mstyle></mml:math>Keff=KDRmax+αRmin+α,(2)
where KD represents the dissociation constant of Fura-8 (260 nM), Rmax and Rmin are the maximum and minimal ratios of intensity measured using an electrode solution containing 10 mM CaCl2 and 10 mM EGTA, respectively.
Step 3: intracellular free calcium ion concentration
<mml:math display="block"><mml:mo stretchy="false">[</mml:mo><mml:mrow><mml:mrow><mml:mi mathvariant="normal">C</mml:mi></mml:mrow><mml:msup><mml:mrow><mml:mi
mathvariant="normal">a</mml:mi></mml:mrow><mml:mrow><mml:mn>2</mml:mn><mml:mo>+</mml:mo></mml:mrow></mml:msup></mml:mrow><mml:mo
stretchy="false">]</mml:mo><mml:mo>=</mml:mo><mml:msub><mml:mi>K</mml:mi><mml:mrow><mml:mrow><mml:mi mathvariant="normal">eff</mml:mi></mml:mrow></mml:mrow></mml:msub><mml:mstyle
displaystyle="true" scriptlevel="0"><mml:mrow><mml:mfrac><mml:mrow><mml:mi>R</mml:mi><mml:mo>−</mml:mo><mml:msub><mml:mi>R</mml:mi><mml:mrow><mml:mo
movablelimits="true" form="prefix">min</mml:mo></mml:mrow></mml:msub></mml:mrow><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mrow><mml:mo
movablelimits="true" form="prefix">max</mml:mo></mml:mrow></mml:msub><mml:mo>−</mml:mo><mml:mi>R</mml:mi></mml:mrow></mml:mfrac></mml:mrow><mml:mo>,</mml:mo><mml:mspace
width="2em"></mml:mspace><mml:mo stretchy="false">(</mml:mo><mml:mn>3</mml:mn><mml:mo stretchy="false">)</mml:mo></mml:mstyle></mml:math>[Ca2+]=KeffR−RminRmax−R,(3)
where R represents the ratio of fluorescence intensities excited by 365 and 410 nm light. The resting calcium ion concentration [Ca2+]rest and the peak calcium ion concentration [Ca2+]peak after induced AP firing can be obtained from Equation 3.
Step 4: calcium buffer capacity KB
After the rupture of the membrane seal, the concentration of Fura-8 as well as the corresponding calcium buffer capacity KB in the cytosol gradually increases. [B]t represents the concentration of the calcium dye inside the neuron and can be calculated by the following equation:<mml:math display="block"><mml:mo stretchy="false">[</mml:mo><mml:mi>B</mml:mi><mml:msub><mml:mo stretchy="false">]</mml:mo><mml:mi>t</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:mo
stretchy="false">[</mml:mo><mml:mi>B</mml:mi><mml:msub><mml:mo stretchy="false">]</mml:mo><mml:mi>p</mml:mi></mml:msub><mml:mo>⋅</mml:mo><mml:msub><mml:mi>F</mml:mi><mml:mi>i</mml:mi></mml:msub><mml:mo>/</mml:mo><mml:msub><mml:mi>F</mml:mi><mml:mrow><mml:mi>i</mml:mi><mml:mi
mathvariant="normal">∞</mml:mi></mml:mrow></mml:msub><mml:mo>,</mml:mo><mml:mspace width="2em"></mml:mspace><mml:mo stretchy="false">(</mml:mo><mml:mn>4</mml:mn><mml:mo
stretchy="false">)</mml:mo></mml:math>[B]t=[B]p⋅Fi/Fi∞,(4)
where [B]p is the total concentration of Fura-8 in the electrode solution (100 µM for TRN PV neurons or 50 µM for VTA DA neurons). The
Fi value at a given time can be calculated from Equation 1. Fi∞ represents the Fi value when Fura-8 reaches equilibrium inside the neuron (typically after 20 min) and can also be calculated from Equation 1.
KB can be calculated by the following:<mml:math display="block"><mml:msub><mml:mi>K</mml:mi><mml:mi>B</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"
scriptlevel="0"><mml:mrow><mml:mfrac><mml:mrow><mml:mo stretchy="false">[</mml:mo><mml:mi>B</mml:mi><mml:msub><mml:mo stretchy="false">]</mml:mo><mml:mi>t</mml:mi></mml:msub><mml:mo>⋅</mml:mo><mml:msub><mml:mi>K</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:mo
stretchy="false">(</mml:mo><mml:mrow><mml:msub><mml:mrow><mml:mo stretchy="false">[</mml:mo><mml:mrow><mml:mrow><mml:mi mathvariant="normal">C</mml:mi></mml:mrow><mml:msup><mml:mrow><mml:mi
mathvariant="normal">a</mml:mi></mml:mrow><mml:mrow><mml:mn>2</mml:mn><mml:mo>+</mml:mo></mml:mrow></mml:msup></mml:mrow><mml:mo
stretchy="false">]</mml:mo></mml:mrow><mml:mrow><mml:mrow><mml:mi mathvariant="normal">rest</mml:mi></mml:mrow></mml:mrow></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi>K</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow><mml:mo
stretchy="false">)</mml:mo><mml:mo stretchy="false">(</mml:mo><mml:mrow><mml:msub><mml:mrow><mml:mo stretchy="false">[</mml:mo><mml:mrow><mml:mrow><mml:mi
mathvariant="normal">C</mml:mi></mml:mrow><mml:msup><mml:mrow><mml:mi mathvariant="normal">a</mml:mi></mml:mrow><mml:mrow><mml:mn>2</mml:mn><mml:mo>+</mml:mo></mml:mrow></mml:msup></mml:mrow><mml:mo
stretchy="false">]</mml:mo></mml:mrow><mml:mrow><mml:mrow><mml:mi mathvariant="normal">peak</mml:mi></mml:mrow></mml:mrow></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi>K</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow><mml:mo
stretchy="false">)</mml:mo></mml:mrow></mml:mfrac></mml:mrow><mml:mo>,</mml:mo><mml:mspace width="2em"></mml:mspace><mml:mo
stretchy="false">(</mml:mo><mml:mn>5</mml:mn><mml:mo stretchy="false">)</mml:mo></mml:mstyle></mml:math>KB=[B]t⋅KD([Ca2+]rest+KD)([Ca2+]peak+KD),(5)
where KD is the dissociation constant of Fura-8 (260 nM). [Ca2+]rest and [Ca2+]peak can be obtained from Equation 3.
Step 5: decay time constant of calcium transients
After rupturing the cell membrane, the exogenous calcium buffer power (calcium-buffering capacity) KB increases as Fura-8 gradually enters the neuron. Increased calcium buffer power tends to decrease the amplitude of the calcium
transients and increase the time constant for the decay phase. The decay time constant τ of the calcium transients can be obtained using Clampfit software. The relationship between KB and τ at different time points can be fitted into a linear equation:<mml:math display="block"><mml:mi>τ</mml:mi><mml:mo>=</mml:mo><mml:mstyle displaystyle="true" scriptlevel="0"><mml:mrow><mml:mfrac><mml:mrow><mml:mn>1</mml:mn><mml:mo>+</mml:mo><mml:msub><mml:mi>K</mml:mi><mml:mi>B</mml:mi></mml:msub><mml:mo>+</mml:mo><mml:msub><mml:mi>K</mml:mi><mml:mi>E</mml:mi></mml:msub></mml:mrow><mml:mi>γ</mml:mi></mml:mfrac></mml:mrow><mml:mo>,</mml:mo><mml:mspace
width="2em"></mml:mspace><mml:mo stretchy="false">(</mml:mo><mml:mn>6</mml:mn><mml:mo stretchy="false">)</mml:mo></mml:mstyle></mml:math>τ=1+KB+KEγ,(6)
where τ is obtained from Clampfit software, KB is calculated using Equation 5, and the negative x-axis intercept provides an estimate of the endogenous buffer capacity KE. Following this procedure, we determined the endogenous calcium buffer (KE) capacity of neurons from WT, Shank3−/−, and HCN2−/− mice. Notably, KE values can also be derived from the A−1-KB plots. Negative KE values can be attributed to the potential dialysis of endogenous calcium buffer. In addition, the low temporal resolution
of 20 Hz means that image-based recording more accurately captures the decay kinetics than the amplitude of the calcium transients.
Therefore, only the KE values derived from the t-KB plots were used in the final analysis.
Three types of calcium dyes are used, Fluo-4 (2P, Kd 345 nM, 100 μM), Fluo-5F (2P, Kd 2.3 μM, 250 μM), and Fura-8 (ratiometric
dye, Kd 260 nM, 100 μM), and their buffer capacity changes as a function of peak calcium concentration (based on Eq. 5; Fig. S1B). Each trace of calcium transients contains 500 (two-photon imaging, 25 Hz, 20 s) or 390 (ratiometric imaging, 20 Hz, 19.5 s)
images. The standard exponential function for one term from Clampfit (ver 10.7) was used to fit the data window between the
peak value and the last data point:<mml:math display="block"><mml:mi mathvariant="normal">Δ</mml:mi><mml:mi>F</mml:mi><mml:mo stretchy="false">(</mml:mo><mml:mi>t</mml:mi><mml:mo
stretchy="false">)</mml:mo><mml:mo>=</mml:mo><mml:mi>A</mml:mi><mml:mo>⋅</mml:mo><mml:msup><mml:mi>e</mml:mi><mml:mrow><mml:mo>−</mml:mo><mml:mi>t</mml:mi><mml:mo>/</mml:mo><mml:mi>τ</mml:mi></mml:mrow></mml:msup><mml:mo>+</mml:mo><mml:mi>C</mml:mi><mml:mo>.</mml:mo><mml:mspace
width="2em"></mml:mspace><mml:mo stretchy="false">(</mml:mo><mml:mn>7</mml:mn><mml:mo stretchy="false">)</mml:mo></mml:math>ΔF(t)=A⋅e−t/τ+C.(7)
Notably, the decay of calcium transients involves a number of factors, including calcium sources, endogenous buffer proteins
and exogenous calcium dye, and the clearance system, so that curve fitting with a single-exponential function is a simplified
treatment (Higley and Sabatini, 2012).
Solutions used (all in mM)
Transcardial perfusion and brain slice preparation solution (in mM): 220 sucrose; 2.5 KCl; 1.25 NaH2PO4; 6 MgSO4; 0.5 CaCl2; 25 NaHCO3; 20 glucose; 0.4 ascorbic acid; 2 Na pyruvate.
Brain slice incubation and recording solution (in mM): 125 NaCl; 2.5 KCl; 1.25 NaH2PO4; 2 MgSO4; 2 CaCl2; 25 NaHCO3; 20 glucose; 0.4 ascorbic acid; 2 Na pyruvate.
Standard electrode solution (in mM): 130 K gluconate; 5 KCl; 2 MgSO4; 5 creatine phosphate; 1 ATP; 0.3 GTP; 10 HEPES; 0.5 EGTA.
Electrode solution for calcium imaging (in mM): 130 K gluconate; 5 KCl; 2 MgSO4; 5 creatine phosphate; 1 ATP; 0.3 GTP; 10 HEPES.
Reagents used
For two-photon calcium imaging, two calcium indicators were used: Fluo-5F (pentapotassium salt) and Fluo-4 (pentapotassium salt). To trace neuronal soma and dendrites, a red fluorescent dye Alexa Fluor 594 (commonly abbreviated as AF594) was used.
Immunofluorescence staining
Mice under anesthesia were transcardially perfused with ∼20 ml of ice-cold PBS solution, followed by 20 ml of precooled 4% paraformaldehyde. Then the brain tissue was dissected out and immersed in 4% paraformaldehyde for overnight fixing, followed by soaking with 30% sucrose for cryoprotection and then frozen in OCT medium.
Brain slices of 25 µm were prepared using a Leica cryostat and then washed in 0.1% PBST for three times and 0.3% PBST for 20 min. After blocking with PBST and 10% goat serum for 1 h, brain slices were incubated with primary antibodies (rabbit anti-PV antibody; 1:1,000 dilution; Cell Signaling, #80561) at 4°C overnight. After rinsing three times, brain slices were incubated with secondary antibodies in the dark (Alexa Fluo488 conjugated goat anti-rabbit, ab150077, Abcam, 1:1,000) at room temperature for 1 h. After extensive rinsing, brain slices were mounted with sealing agent DAPI-Aqueous (ab104139, Abcam). Parameters of the confocal microscope, especially the laser power and detector gain, were preadjusted to minimize photobleaching. The same set of parameters was used in the data collection for parallelly processed WT and mutant samples. After raw image to TIFF format (16 bit), ImageJ was used for data analysis. For each brain slice, the TRN region was manually selected (Freehand selections) and cell bodies were registered with the PV signal (green channel) as the guidance. The size parameter was set as 0.01–3 inch2 (Analyze Particles, under the menu command Analyze). Then the ROI outputs were used to measure the PV signal (green channel) of the selected cell bodies. For each brain slice, we obtain the total number of PV+ neurons and the average PV signal intensity. Final statistics were based on results of all brain slices used in data analysis (Fig. S6A–E). Images from the mouse brain atlas, including the in situ hybridization data of PV, were shown for reference purposes (Fig. S6F).
Statistics
All statistical tests were performed using the program of OriginPro. Data are presented as mean ± SEM. Statistical significance was assessed with unpaired Student’s t test. P < 0.05 was considered statistically significant, and significance is indicated as follows: ***p < 0.001; **p < 0.01; *p < 0.05. p values are added to the figure legends.
Results
Abnormal neuronal excitability and AP firing in Shank3−/− and HCN2−/− PV neurons from TRN
Previous studies on engineered human neurons, cultured mouse hippocampal neurons, and mouse thalamocortical neurons identified HCN channelopathy as the primary downstream event caused by Shank3 deficiency (Yi et al., 2016; Zhu et al., 2018). We started from an electrophysiological examination of TRN PV neurons from WT, Shank3−/−, and HCN2−/− mice. TRN and TC neurons exhibit two distinct action potential (AP) firing modes: the tonic mode, in which neurons show sustained and nonadaptive firing, and the burst mode, involving sequential activation of HCN and T-type voltage-gated calcium (CaVT) channels, followed by calcium-activated potassium and chloride channels (McCormick and Pape, 1990; von Krosigk et al., 1993; Cueni et al., 2008; Astori et al., 2011; Errington et al., 2012; Ha et al., 2016). Burst firing can be initiated by a cluster of inhibitory inputs (hyperpolarization), CaVT deinactivation, HCN channel activation, membrane depolarization, CaVT activation and formation of low-thresh spike (LTS), and a cluster of AP firings (Na spikes) riding on top of LTS.
Our recordings revealed a clear distinction exists between two groups of neurons in AP firing and other basic characteristics (Fig. 1, Fig. S2). PV neurons exhibit high-frequency, nonadapting AP firing in tonic mode, no membrane sag in response to hyperpolarizing current injections, typical burst firing at hyperpolarizing membrane potentials, and a short AP half-width (∼1 ms). In contrast, the other group of neurons, presumably SOM+ neurons, exhibited lower frequency of AP firing, a lack of burst firing, clear calcium-dependent adaption (accommodation), a wider AP half-width (∼1.5 ms), a more depolarized RMP, and a higher Rin. Notably, a recent study reported that most inhibitory neurons in TRN are PV positive, and they can be further separated into calbindin (CB) neurons in the center and somatostatin (SST) neurons near the edge of TRN (Martinez-Garcia et al., 2020). The profile of AP firing and the location of the neurons we recorded are consistent with the PV+ and CB+ neurons in the TRN center. In this study, we focused on PV neurons in TRN and excluded SOM neurons from data analysis.
Difference between two types of inhibitory neurons in TRN, presumably PV and SOM types, in AP firing elicited by depolarizing and hyperpolarizing current injections. A, Membrane potential traces of PV and SOM neurons in response to 50, 100, and 150 pA current injection. AP firing by PV neurons exhibits two distinct modes, burst mode and tonic mode. In contrast, SOM neurons only show tonic mode. In addition, in tonic firing mode, PV neurons exhibit little adaptation whereas SOM neurons show strong frequency adaptation, which is presumably calcium dependent. B, Membrane potential traces of PV and SOM neurons in response to hyperpolarizing current injections. The membrane potential was aproximately −100 mV before rebound. Rebound AP firing by PV neurons exhibits two distinct modes, burst mode and tonic mode (nonadaptive). In contrast, SOM neurons only show tonic mode (adaptive). C, Compared with PV neurons, SOM neurons exhibit higher Rin and more depolarized RMP. p value: 0.0033 (RMP); 0.0024 (Rin). D, Statistics of key properties of AP firing. Animal numbers: 6 (PV), 5 (SOM). Total numbers of neurons used in the analysis: 10 (PV), 10 (SOM). p value: <0.0001 (AP number); 0.037 (AP half-width); 0.011 (AP half-amplitude); 0.0003 (tonic frequency); 0.097 (1st latency).
In HCN2−/− TRN PV neurons, we observed a negatively shifted RMP, from −71.8 ± 0.4 mV (WT, n = 15) to −79.3 ± 1.4 mV (n = 15), as well as an increase in Rin (Fig. 2A). These results align with the expected contributions by HCN channels to RMP and Rin. Similarly, Shank3−/− TRN neurons exhibited a more negative RMP (−73.7 ± 0.7 mV, n = 15). Since HCN channels are known to be distributed in the dendrites of TRN neurons, the negatively shifted RMP and increased Rin suggest that dendritic HCN channels can still make substantial contributions to somatic excitability, and in addition, Shank3 deficiency likely reduces HCN channel functionality (Abbas et al., 2006; Ying et al., 2007).
Increased episodes of burst firing and reduced frequency of tonic firing in TRN neurons from Shank3−/− and HCN2−/− mice. A, Left, Bright-field image of a typical brain slice containing TRN, ventral posteromedial nucleus (VPM), and ventral posterolateral nucleus (VPL). Right, RMP and Rin of WT (−71.8 ± 0.4 mV, n = 15; 388.3 ± 23.2 MΩ, n = 15), Shank3−/− (−73.7 ± 0.7 mV, n = 15; 458.9 ± 45.7 MΩ, n = 15), and HCN2−/− (−79.3 ± 1.4 mV, n = 15; 516.6 ± 35.6 MΩ, n = 14) TRN neurons. Animal numbers: 3 (WT), 3 (Shank3−/−), 6 (HCN2−/−). Total numbers of neurons used in the analysis: 15 (WT), 15 (Shank3−/−), 15 (HCN2−/−). p value: RMP, 0.0021 (WT-Shank3−/−), <0.0001 (WT-HCN2−/−); Rin, <0.0001 (WT-Shank3−/−), 0.00467 (WT-HCN2−/−). B, AP firing of TRN neurons from WT, Shank3−/− and HCN2−/− mice elicited by depolarizing current injections (50, 100, 150 pA). To prevent spontaneous firing of AP and also for comparison purposes, the RMP was adjusted to approximately −70 mV. C, Compared with WT TRN neurons, Shank3−/− and HCN2−/− neurons exhibit increased episodes of burst firing. Animal numbers: 6 (WT), 4 (Shank3−/−), 5 (HCN2−/−). Total numbers of neurons used in the analysis: 10 (WT), 10 (Shank3−/−), 10 (HCN2−/−). p value: 0.0055 (WT-Shank3−/−), 0.0215 (WT-HCN2−/−). D, Shank3−/− and HCN2−/− TRN neurons exhibit decreased frequency of tonic firing. Animal numbers: 6 (WT), 4 (Shank3−/−), 5 (HCN2−/−). Total numbers of neurons used in the analysis: 10 (WT), 10 (Shank3−/−), 10 (HCN2−/−). p value: 0.0051 (WT-Shank3−/−), <0.0001 (WT-HCN2−/−).
Next, we examined AP firing in TRN PV neurons, with the membrane potential adjusted to a level approximat −70 mV and stimulated by a series of depolarizing current injections. WT TRN neurons tend to fire one or two episodes of burst firing and then transition to tonic firing (Fig. 2B). In contrast, Shank3−/− TRN neurons fire significantly more episodes of burst firing before transitioning to tonic firing. In response to +100 pA current injection, Shank3−/− neurons generated an average of three episodes of burst firing, in contrast to only one in WT neurons. Notably, HCN2−/− TRN neurons exhibited a similar increase in burst firing, albeit to a lesser extent than Shank3−/− neurons (Fig. 2C).
For tonic firing, we observed opposite trends of changes, with mutant TRN PV neurons showing reduced frequency of AP firing (Fig. 2D). The frequency of tonic firing increases in response to more depolarizing current injections. WT TRN neurons can reach a firing rate over 80 Hz with +400 pA current injection (Fig. S3). In contrast, the rate of tonic firing was significantly reduced in Shank3−/− neurons, reaching 75 Hz at +400 pA current injection, with a significantly increased half-amplitude. Notably, HCN2−/− TRN neurons exhibited an even more dramatic decrease in AP firing rate, down to the level of 58 Hz at +400 pA current injection. If counting both Na spikes in burst firing and APs during tonic firing, Shank3−/− neurons showed a slight decrease whereas the decrease in HCN2−/− neurons was more robust.
To evaluate the rebound burst firing, negative current injections were delivered to TRN PV neurons to reach the hyperpolarizing potential approximately −100 mV (Fig. 3A). Shank3−/− neurons took a longer time to generate the first Na spike (latency of 1st rebound Na spike), accompanied by a long time interval between the first Na spike and AHP (Fig. 3B). For Na spikes in burst firing, Shank3−/− and HCN2−/− TRN neurons exhibit a reduced frequency. HCN2−/− TRN neurons exhibit a reduced number of Na spikes with a shorter half-width, whereas Shank3−/− TRN neurons show a slight increase in the number of Na spikes (Fig. 3C). In addition, we used a different protocol to elicit one or two burst firing by injecting minimum currents into neurons with negative membrane potentials aproximately −70 mV (Fig. S4). Compared with WT neurons, Shank3−/− and HCN2−/− neurons exhibit shorter interval between 1st Na spike and AHP (1 burst firing) and longer interval between two burst firings.
Rebound burst firing of TRN neurons from WT, Shank3−/− and HCN2−/− mice. A, AP firing in response to a series of increasing hyperpolarizing current injections. The traces with the hyperpolarizing membrane potential before membrane potential rebound approximately −100 mV were selected for analysis. Right, Zoom views over the initial burst firing after membrane potential rebound. Animal numbers: 2 (WT), 2 (Shank3−/−), 2 (HCN2−/−). Total numbers of neurons used in the analysis: 10 (WT), 10 (Shank3−/−), 10 (HCN2−/−). B, Summary of burst firing after membrane potential rebound. From left to right, latency of the 1st burst firing, time interval between 1st Na spike and AHP, time interval between the 1st Na spikes of two sequential burst firings as marked above. p value: WT-Shank3−/−, 0.032, 0.005, 0.359; WT-HCN2−/−, 0.466, 0.225, 0.287. C, Summary of Na spikes in the 1st rebound burst firing. From left to right, frequency of Na spikes, number of Na spikes, half-width of Na spikes. p value: WT-Shank3−/−, 0.006, 0.026, 0.065; WT-HCN2−/−, 0.009, 0.006, 0.005.
Faster decay of dendritic calcium transients and reduced PV expression in Shank3−/− TRN neurons
A key feature of the PV+ neurons is their high PV content in the cytosol. Consequently, the calcium-buffering capacity and the spatial-temporal profile of calcium signaling in PV neurons ought to be unique. Through the whole-cell recording pipette, we loaded TRN neurons with the calcium dye and examined somatic and dendritic calcium transients using two-photon calcium imaging (Fig. S5A). For TRN neuron recording, we chose Fluo-5F, a low-affinity dye with a KD of ∼2.3 μM because its binding capacity under physiological range of [Ca2+]i is more stable than those of high affinity dyes (Fig. S1B). Three short pulses of current injection were delivered at the frequency of 1 Hz to trigger AP firing (Fig. S5B,D,E). We recorded calcium transients within the soma and at four spots along the dendritic trunk: 25, 50, 75, or 100 μm away from the soma (Fig. S5C). At the soma, no clear differences were observed between WT and Shank3−/− neurons. However, dendritic calcium transients exhibited a much faster decay (Fig. S5D).
In a parallel study of VTA dopaminergic neurons, we made similar observations, except that calcium transients decay faster in both the soma and the dendrites (unpublished results). To explore the corresponding molecular basis, we utilized a systematic proteomics approach to analyze brain tissue from the VTA area of WT and Shank3−/− mice. The PV protein was found to be at the top of the list of proteins showing differential expression. Reduced PV expression results in decreased calcium buffer capacity, leading to increased amplitude and faster decay of the calcium transients (Eisner et al., 2023).
To confirm the above results obtained with VTA DA neurons, we performed immunostaining for PV proteins in fixed brain slices from WT and Shank3−/− mice (Fig. 4A). Brain slices were collected from paired littermates and processed in parallel during the staining and imaging steps. Quantification of the immunostaining results from the TRN area revealed a significant decrease in PV fluorescence signal in Shank3−/− TRN neurons (Fig. 4B–D; Fig. S6). Notably, the basal level of PV in TRN neurons is much higher than in VTA DA neurons. In addition, Shank3−/− PV+ neurons in the cortex also exhibited a significant decrease in PV expression (Gogolla et al., 2009; Filice et al., 2016; Orefice et al., 2019; Shih et al., 2023).
Reduced PV expression in Shank3−/− TRN PV neurons. A, Immunostaining of PV (green) and DAPI (blue) in coronal slices of WT and Shank3−/− mouse brain. Scale bar, 1 mm. B, PV expression in the TRN area of WT and Shank3−/− mice. Scale bar, 50 μm. C, Zoomed-in view over the boxes marked above showing weaker PV staining in Shank3−/− TRN neurons. Scale bar, 20 μm. D, Statistics of immunostaining results. Animal numbers: 3 (WT), 3 (Shank3−/−). Total number of analyzed brain slices: 10 (WT), 12 (Shank3−/−). p value: 0.153, 0.0007.
Reduced calcium buffer capacity in Shank3−/− TRN neurons
It is perplexing that calcium transients in the soma of Shank3−/− TRN neurons decay normally, yet immunostaining revealed a significant decrease in PV expression in mutant cells. To investigate this discrepancy further, we used the added buffer approach to quantify the capacity of the calcium-buffering system. After the formation of whole-cell recording mode, the calcium dye added in the recording pipette slowly diffuses into the neuron. Consequently, the calcium buffer capacity inside the neuron gradually increases, reflected by the gradual changes in the amplitude and the decay time constant of the calcium transients. KE, which represents endogenous calcium binding capacity, can be extracted through linear regression.
We established a dual wavelength excitation system and used Fura-8, an improved version of the widely used ratiometric calcium dye Fura-2 (Fig. 5, Fig. S7). Fura-8 was alternately excited by 365 and 410 nm light (Fig. 5A,B). After membrane break-in, calcium transients were recorded every 2 min. The corresponding amplitudes and decay time constants were used to derive the value of KE (Fig. 5C–G). Our results revealed a more significant decrease in KE, from 138.8 of WT to 32. 7 of Shank3−/− TRN PV neurons (Fig. 5H). Thus, the reduced calcium buffer capacity in Shank3−/− neurons is consistent with the immunostaining results which revealed reduced PV expression in mutant neurons.
Reduced calcium-buffering capacity in Shank3−/− TRN neurons. A, Radiometric calcium imaging results of WT and Shank3−/− TRN neurons. Fura-8 at 100 mM was added to the pipette solution. Top, Current injection (500 pA × 10 ms, 3 stimulations at 50 Hz). Bottom, Membrane potential recordings. Zoomed views are shown in Figure S7. B, Fluorescence intensities with alternative excitation of Fura-8 by 410 and 365 nm LED lights and the radiometric trace. C, Calcium transients (F365/F410) as a function of time after membrane break-in. D, Fura-8 concentration (Right) and fluorescence intensity (Fluo, left) as a function of time after membrane break-in and the formation of whole-cell recording mode. E, KB as a function of time after membrane break-in. F, Cross-plot of the time constant of calcium transient decay versus KB. G, Cross-plot of the amplitude of calcium transient decay versus KB. Results in C–G are from the same WT TRN neuron. H, Statistics of KE of WT and Shank3−/− neurons. Animal numbers: 6 (WT), 5 (Shank3−/−). Number of neurons in the analysis: 10 (WT), 10 (Shank3−/−). p value: <0.0001.
To explore the potential link between decreased PV expression and abnormal AP firing in Shank3−/− TRN PV neurons, we added an additional 2 mM EGTA to the whole-cell recording pipette, with the aim to compensate the decreased calcium-buffering capacity in mutant cells (Johnson et al., 1999; Caillard et al., 2000; Muller et al., 2007). Just as expected, the number of burst firing was significantly reduced in mutant neurons. With a 100 pA current injection, the number of bust firing was reduced from ∼3.0 to ∼1.3 in Shank3−/− TRN PV neurons with extra EGTA. In addition, for the mutant neurons with EGTA, the frequency of tonic firing increased significantly, with the half-amplitude and other characteristics more-or-less closer to the control levels (Figs. S8, S9). Conversely, in the absence of EGTA in pipette solution, the number of burst firing increases compared with the recordings collected with normal EGTA (0.5 mM) added (Fig. S10). Taken together, these results are consistent with the notion that reduced PV expression has connections with the abnormal AP firing, including the increased burst firing and decreased tonic firing, in Shank3−/− TRN neurons.
TRN neurons from HCN2−/− mice show comparable changes in PV expression and calcium signaling to Shank3−/− TRN neurons and further evidence of VTA DA neurons
The above results prompted us to further investigate the relationship between Shank3 deficiency, altered calcium signaling, and cell excitability in TRN neurons from HCN2−/− mice. We examined the cell excitability of Shank3−/− and HCN2−/− TRN neurons (Fig. S11A). The decay of calcium transients in HCN2−/− TRN neurons, in both the soma and four spots along the dendrite, proceeds at much faster rates than in WT neurons (Fig. S11B–F). To determine if the faster decay of calcium transients is due to a smaller calcium-buffering capacity in HCN2−/− TRN neurons, we examined PV expression by immunostaining. Indeed, PV expression in HCN2−/− neurons was significantly decreased (Fig. S6). Taken together, the results with TRN neurons from Shank3−/− and HCN2−/− mice uncovers PV downregulation, reduced calcium-buffering capacity, and abnormal calcium signaling as possible downstream events.
To further evaluate the molecular events downstream of Shank3 deficiency, we examined VTA DA neurons from Shank3−/− and HCN2−/− mice. To elicit AP firing, current injections at the interval of +50 pA were applied to DA neurons (Fig. 6A). Remarkably, Shank3−/− and HCN2−/− DA neurons exhibited similar changes in neuronal excitability, including decreases in AP firing, a shorter AP half-width, and a lower half-amplitude of AP (Fig. 6B). HCN channels are known to contribute to the maintenance of RMP and Rin and the generation of membrane sag after hyperpolarization stimulations. In both Shank3−/− and HCN2−/− DA neurons, we observed negative shifts in RMP and reduced sag ratio, with the changes in HCN−/− DA neurons being more severe (Fig. 6C,D). These observations in VTA DA neurons are consistent with the notion that Shank3 deficiency causes the downregulation of HCN channel expression and function, thereby altering neuronal excitability. Notably, sag is reduced but not completely abolished in HCN2−/− VTA neurons, which may reflect contributions of other HCN subtypes, most likely HCN1 and HCN4.
Shank3−/− and HCN2−/− VTA DA neurons exhibit similar changes in neuronal excitability, reduced PV expression, and faster decay of calcium transients. A, AP firing of VTA DA neurons from WT, Shank3−/−, and HCN2−/− mice evoked by current injection (150 pA). B, Averaged RMP and Rin of WT, Shank3−/−, and HCN2−/− DA neurons. Animal numbers: 4 (WT), 3 (Shank3−/−), 3 (HCN2−/−). Neuron numbers: 12 (WT), 14-9-14 (Shank3−/−), 7 (HCN2−/−). p value: number of AP, WT-Shank3−/− 0.002, WT-HCN2−/− 0.041; half-width, WT-Shank3−/− 0.024, WT-HCN2−/− 0.022; half-amplitude, WT-Shank3−/− 0.008. C, Reduced AP firing number, half-width, and half-amplitude in Shank3−/− and HCN2−/− DA neurons. Animal numbers: 7 (WT), 10 (Shank3−/−), 5 (HCN2−/−). Neuron numbers: 14 (WT), 24 (Shank3−/−), 10 (HCN2−/−). p value: WT-Shank3−/−, RMP 0.020, Rin 0.901; WT-HCN2−/−, RMP 0.006; Rin 0.560. D, Membrane potential sag evoked by a −200 pA current injection. Left, Representative traces of membrane potential. Right, Averaged results. Animal numbers: 4 (WT), 3 (Shank3−/−), 3 (HCN2−/−). Neuron numbers: 11 (WT), 12 (Shank3−/−), 13 (HCN2−/−). p value at −200 pA: WT-Shank3−/−, 0.022; WT-HCN2−/−, 0.0009.
The above results prompted us to investigate further the relationship between Shank3 deficiency, altered calcium signaling, and cell excitability. We repeated the immunostaining of PV in HCN2−/− mice (Fig. 7A). Indeed, a significant decrease in PV expression was observed in the HCN2−/−VTA area, which was quite similar to the Shank3−/− VTA staining results (Fig. 7B). Next, we examined the profile of calcium signaling in VTA DA neurons, in both the soma and the dendritic trunk, using a two-photon calcium imaging system with Fluo-4 as the calcium indicator (Fig. 7C). The decrease in the time constant of calcium transient decay in the soma of HCN2−/− DA neurons was more dramatic, from 6.9 ± 7.0 s to 3.1 ± 5.3 s (Fig. 7D). Furthermore, along the dendrites, the time constant for the decay phase was comparable between HCN2−/− and Shank3−/− DA neurons but was significantly smaller than that of WT neurons (Fig. 7E,F). Consistently, the amplitudes of the calcium transients in the dendrites of HCN2−/− and Shank3−/− DA neurons were similar, but both were higher than those of WT neurons.
Reduced PV expression and faster decay of calcium transients in HCN2−/− VTA DA neurons. A, Immunostaining results of PV (green) and TH (red) in WT and HCN2−/− VTA areas. B, Statistics of immunostaining results. Animals used: 3 (WT), 3 (HCN2−/−). p value, 0.002 (neuron number), 0.003 (TH), <0.0001 (PV). C, Two-photon fluorescence image of a VTA DA neuron filled with AF594 (red channel, 10 mM) and Fluo-4 (200 mM). D, Averaged results of time constant (left) and amplitude (right) of calcium transients recorded from the soma. Calcium transients were evoked by a train of 4 current pulses (2 nA, 2 ms, 2 Hz). Animal numbers, 2 (WT), 3 (HCN2−/−). Neuron numbers: 7 (WT), 9 (HCN2−/−). p value: 0.0006 (tau), 0.203-0.153-0.215-0.234 (peaks 1–4). E, Decay time constant of calcium transients recorded from three dendritic locations (50, 100, 150 μm from soma). p values: 0.013 (50 μm), 0.005 (100 μm), 0.003 (150 μm). p values (50–100–150 μm): peak 1, 0.009 0.003 0.016; peak 2, 0.026 0.001 0.007; peak 3, 0.012 0.002 0.005; peak 4, 0.0172 < 0.0001 0.001.
Discussion
Well-balanced chemical and electrical communications form the basis of higher-order of neurological functions. Undoubtedly, fast-spiking (FS) inhibitory neurons play a key regulatory role in every neuronal circuitry and contribute to various neuropsychiatric functions. However, the high speed of action potential firing by FS interneurons comes at the cost of high metabolic expenditure and an excessive, and often cytotoxic, influx of calcium ions into the cytosol. In numerous neuropsychiatric diseases, PV+ neurons are more vulnerable to pathological stresses. Intriguingly, a key molecular marker of FS interneurons is the PV protein, a small calcium-binding protein with no other known significant roles besides calcium binding. The expression of PV is dynamic, and its downregulation has been reported in autism, Alzheimer's disease, etc. (Wohr et al., 2015; Filice et al., 2016; Lauber et al., 2018; Terstege et al., 2025). The exact connections among neuronal excitability, calcium signaling, and the expression and function of PV in these neuropsychiatric disorders have not been established (Filice et al., 2020).
TRN PV neurons critically regulate the excitability of TC neurons, and the reciprocal contacts and excitation between TC and TRN neurons form the basis of oscillatory brain activities, such as sleep, attention, and motor control. In a previous study, we addressed TC neurons from Shank3−/− mice and uncovered HCN channelopathy and associated alterations in neuronal excitability. In this study, we tackled fast spiking, PV+ neurons in the TRN from the Shank3−/− and HCN2−/− mice. Our results show that Shank3−/− TRN neurons exhibit increased burst firing episodes but oppositely decreased frequency of tonic firing.
The display of burst and tonic firings, two distinct AP firing modes, is a unique feature shared by both excitatory TC and inhibitory TRN neurons and carries important physiological significance in processes including sleep and attention (Crunelli et al., 2018). By producing postsynaptic responses distinctively different from tonic firing, bursting firing of TRN neuron provides strong, concentrated, and long-lasting inhibitory control over the thalamus–cortex communication (Desai and Varela, 2021). Bursting firing of TRN neurons has been reported to activate both GABAA and GABAB receptors whereas tonic firing only activates GABAA receptor (Kim et al., 1997; Kim and McCormick, 1998). Thus, our discovery that Shank3−/− TRN neurons are prone to burst firing sheds light on the related pathophysiological changes within the thalamus–cortex circuitry as well as the relevant behavioral phenotypes.
As PV is a key component of the calcium-buffering system, we used two-photon calcium imaging to examine activity-dependent calcium transients. In Shank3−/− TRN PV neurons, dendritic calcium transients decay faster, consistent with the expectation of reduced calcium-buffering capacity. Immunostaining results confirmed reduced PV expression in Shank3−/− neurons. To quantify the impact on calcium-buffering capacity, we used the added buffer approach and found a significant decrease in KE, a quantitative measure of cytosolic calcium-buffering capacity. These results demonstrate the association among Shank3 deficiency, abnormal neuronal excitability, reduced PV expression, and the related decrease in calcium-buffering capacity. Remarkably, supplementing Shank3−/− neurons with extra EGTA, to compensate for the reduced PV level, resulted in a robust decrease in episodes of burst firing and an increase in tonic firing frequency and thus reversed the abnormal excitability in mutant neurons. This result further supports the critical role played by PV and its involved calcium buffer system in the physiology of TRN PV neurons. It is worth mentioning that other than bursting firing number and tonic firing frequency, the effects of EGTA on other aspects of AP firing of Shank3−/− neurons are limited. This could be attributed to the intrinsic complexity of PV as a calcium buffer compared with EGTA, or compensatory changes during the development of Shank3−/− mice, including altered expression of certain channels and receptors.
Notably, the calcium signaling underlying the burst firing or LTS in TRN PV neuron is intrinsically different from the calcium signal triggered by isolated action potentials. Given the AP stimulation protocol and the temporal resolution of calcium imaging (two-photon/Fluo-5F, 25 Hz; sCMOS/Fura-8, 20 Hz) used in the current study, a potential impact of PV deficiency on the fast component of the calcium transient, which occurs in the time window of ∼100 ms after stimulation, cannot be excluded. A significant prolongation of calcium transients was observed with the calcium transients recorded from the calyx of Held, ∼30 ms for WT and 70 ms for PV−/− neurons, indicating PV actually helps curtail the calcium transients (Muller et al., 2007). Additionally, in cerebellar inhibitory interneurons, calcium transients follow a biexponential time course, and the absence of PV significantly prolongs the decay of first/fast component but shortens the second/slow component (Collin et al., 2005). These observations are in-line with the high PV concentration and the related large calcium-buffering power in these inhibitory neurons. Simulation of calcium transients and calibrated immunohistochemistry determined the concentration of PV in cerebellar basket cells in the range of 150 (Collin et al., 2005) and 563 μM (Eggermann and Jonas, 2011). On the other hand, PV is known to exist in three forms, magnesium bound (∼73%), calcium bound (∼22%), and the free forms (∼5%; Eggermann and Jonas, 2011). Therefore, high concentrations of PV in PV+ neurons result in a substantial amount of PV staying in its free form in the resting state. The free form of PV, with its fast binding kinetics (Kon, 1 × 108 M−1s−1) and high affinity (Kd, 5–9 nM), make PV function like BAPTA or effectively as an absorbing system to reduce the amplitude and speed up the initial (fast) decay of calcium transients (Eberhard and Erne, 1994; Schmidt et al., 2003; Collin et al., 2005; Eggermann and Jonas, 2011).
However, under the condition of longer calcium transients such as LTS or more profound calcium influx, with calcium clearance mainly mediated by transporters, PV's role as a calcium source during the decay phase become more significant. Thus, calcium release from PV contributes to the slowdown of the secondary phase (slow) of calcium transients (Lee et al., 2000; Collin et al., 2005; Franconville et al., 2011; Higley and Sabatini, 2012). These studies are consistent with our observations of faster calcium transient decay and decreased PV expression in Shank3−/− neurons.
Still, it is evident that the whole-cell recording configuration used in the current study can be affected by gradual dialysis and wash-off of the intracellular system and loss of fast, mobile calcium buffer proteins such as calbindin. The washout of endogenous calcium buffer is understandably more effective in the soma than dendrites, which potentially explain the comparable somatic calcium transients between WT and Shank3−/− TRN neurons as shown in Figure S5D. To this end, the configuration of perforated patch-clamp recording or dye loading method without disturbance of intracellular environment are promising approaches (Delvendahl et al., 2015; Hess et al., 2021).
Published studies by Südhof's group, based on engineered human neurons and cultured hippocampal neurons from newborn mice, and by our group based on mouse TC neurons have reported HCN channelopathy in Shank3-deficient neurons (Yi et al., 2016; Zhu et al., 2018). Remarkably, a recent study targeted the HCN channelopathy in thalamocortical circuity in Shank3−/− mice and successfully corrected certain autistic behaviors (Guo et al., 2024). In TRN neurons from Shank3−/− and HCN2−/− mice, we observed comparable alterations in neuronal excitability and calcium signaling. In addition, VTA DA neurons from Shank3−/− mice and HCN2−/− mice exhibit similar changes in neuronal excitability, particularly RMP and the membrane sag—a hallmark of contribution to membrane excitability by HCN channels, as well as faster decay in calcium transients. An important question arises: what is the cause-and-effect relationship between HCN channelopathy and the downregulation of PV expression in Shank3−/− neurons?
Multiple lines of evidence suggest that the PV reduction occurs downstream of HCN channelopathy. In Shank3−/− VTA DA neurons, the changes in AP and calcium transients, in both the soma and dendrites, mirror those changes observed in Shank3−/− neurons. In the case of TRN PV neurons, Shank3−/− and HCN2−/− neurons exhibit certain similarities in AP firing. The nonsignificant change in somatic calcium transients in Shank3−/− TRN PV neurons is most likely due to the absence of somatic HCN channel expression. However, the altered profile of dendritic calcium transients in HCN2−/− TRN GABA neurons, specifically the increased amplitude and faster decay, is consistent with the dendritic localization of HCN channels observed in these neurons and again mirrors the changes in Shank3−/− neurons (Abbas et al., 2006; Ying et al., 2007). These similarities between HCN2−/− and Shank3−/− neurons provide further evidence that HCN channelopathy may play a causal role in PV downregulation. Notably, Shank3 interacts with and stabilizes the membrane expression of a number of surface receptors and channels other than HCN channels, and they may also contribute to the altered excitability in Shank3−/− neurons.
Thus, this study expands the current understanding of Shank3 neuropathology to fast-spiking PV neurons and supports the HCN channelopathy, abnormal neuronal excitability, downregulation of PV expression, and abnormal calcium signaling as possible downstream events. Importantly, this study highlights the critical role of the PV protein in PV neurons, as well as the associated intracellular calcium-buffering system in ASD. These insights will facilitate the development of novel therapeutic strategies by targeting elements within the calcium-buffering system and the HCN channels in related neuropsychiatric disorders.
Footnotes
- Received June 19, 2025.
- Revision received December 29, 2025.
- Accepted January 28, 2026.
-
We acknowledge the discussion and technical support from other members of the lab. L.Z. is supported by funding from the National Natural Science Foundation of China (#32171150), Shenzhen Bay Laboratory (Startup Fund and Major Program of SZBL), Pearl River Talents Program of Guangdong Province, and Shenzhen Peacock Project. Q.L. is supported by funding from Virginia Commonwealth University (CTSA) and National Institutes of Health (R01GM098592).
-
↵*Y.L., L.L., and R.S. contributed equally to this work.
-
The authors declare no competing financial interests.
-
This paper contains supplemental material available at: https://doi.org/10.1523/JNEUROSCI.1202-25.2026
- Correspondence should be addressed to Lei Zhou at zhoulei{at}szbl.ac.cn.
- Copyright © 2026 the authors
