Age-related changes in excitatory and inhibitory intra-cortical circuits in auditory cortex of C57Bl/6 mice

Abstract

A common impairment in aging is age-related hearing loss (presbycusis), which manifests as impaired spectrotemporal processing. Aging is accompanied by alteration in normal inhibitory (GABA) neurotransmission, and changes in excitatory (NMDA and AMPA) synapses in the auditory cortex (ACtx). However, the circuits affected by these synaptic changes remain unknown. Mice of the C57Bl/6J strain show premature age-related hearing loss and changes in functional responses in ACtx. We thus investigated how auditory cortical microcircuits change with age by comparing young (~ 6 weeks) and aged (>1 year old) C57Bl/6J mice. We performed laser scanning photostimulation (LSPS) combined with whole-cell patch clamp recordings from Layer (L) 2/3 cells in primary auditory cortex (A1) of young adult and aged C57Bl/6J mice. We found that L2/3 cells in aged C57Bl/6 mice display functional hypoconnectivity of both excitatory and inhibitory circuits. Compared to cells from young C57Bl/6 mice, cells from aged C57Bl/6J mice have fewer excitatory connections with weaker connection strength. Whereas young adult and aged C57Bl/6J mice have similar amounts of inhibitory connections, the strength of local inhibition is weaker in the aged group. We confirmed these results by recording miniature excitatory (mEPSCs) and inhibitory synaptic currents (mIPSCs). Our results suggest a specific reduction in excitatory and inhibitory intralaminar cortical circuits in aged C57Bl/6J mice compared with young adult animals. We speculate that these unbalanced changes in cortical circuits contribute to the functional manifestations of age-related hearing loss.

Introduction

Sensory and cognitive declines with aging are profound and likely involve changes in cortical processing (Nusbaum, 1999; Scialfa, 2002; Jayakody et al., 2018). One common impairment is age-related hearing loss, also known as presbycusis (Brant and Fozard, 1990; Pearson et al., 1995; Matthews et al., 1997; Cruickshanks et al., 1998). Age-related high-frequency hearing loss has been associated with changes in tonotopic organization and deterioration of temporal processing in the inferior colliculus (IC) and auditory cortex (ACtx) (Willott, 1986; Willott et al., 1993; Mendelson and Ricketts, 2001; Lee et al., 2002; Mendelson and Lui, 2004; Gourevitch and Edeline, 2011; Engle and Recanzone, 2013; Trujillo et al., 2013; Brewton et al., 2016; Recanzone, 2018). These functional changes are likely due to altered functional connections between neurons.

The interplay of glutamatergic excitation and GABAergic inhibition contributes to normal brain function. Disruption of inhibition is thought to be responsible for age-related sensory impairment (Chao and Knight, 1997; Kok, 1999; Caspary et al., 2008; Stanley et al., 2012; Rozycka and Liguz-Lecznar, 2017; Recanzone, 2018). Indeed, aging decreases the density of inhibitory synapses in the prefrontal cortex (Peters et al., 2008). It is also associated with decreased levels of glutamate decarboxylase (GAD) and vesicular GABA transporter (VGAT), hypofunction of NMDA receptors, and changes in ionic conductances in different sensory cortices (De Luca et al., 1990; Chaudhry et al., 1998; Milbrandt et al., 2000; Shi et al., 2004; Stanley and Shetty, 2004; Ling et al., 2005; Burianova et al., 2009; Stanley et al., 2012; Gold and Bajo, 2014; Liguz-Lecznar et al., 2015; Liao et al., 2016; Kumar et al., 2019). Interestingly, administering GABA or GABA agonists facilitates visual function in vision-impaired aged non-human primates (Leventhal et al., 2003). Similarly, GABA-associated changes are observed in the auditory pathway in aged animals. Aging alters GABA receptor composition, the levels of GAD and calcium binding proteins (CBPs) in the IC, thalamus, cochlear nucleus, and ACtx in aging rats, non-human primates (Milbrandt et al., 1996; Caspary et al., 1999; Ling et al., 2005; Ouda and Syka, 2012; Caspary et al., 2013; Richardson et al., 2013; Recanzone, 2018; Richardson et al., 2020). Aging mice of both the C57Bl/6 as well CBA strain also show decreased in parvalbumin and perineuronal nets around parvalbumin neurons suggesting changes in inhibition (Brewton et al., 2016; Rogalla and Hildebrandt, 2020).

These results suggest that the primary functional deficit in aging may lie in a hypofunction of inhibitory circuits. At the same time, however, synaptic excitation also changes with aging, since the density of dendritic spines is decreased in aged brains and the structure of NMDA receptor complex changes with aging (Magnusson et al., 2010; Dickstein et al., 2013; Hickmott and Dinse, 2013; Fetoni et al., 2015). These results suggest that both excitatory and inhibitory circuits may be altered in aged brains. Changes in functional microcircuits could contribute to the observed changes in sound processing, e.g., increased convergence or weakening of connections across the tonotopic axis could contribute to changes in bandwidth and temporal processing (Willott et al., 1993; Mendelson and Ricketts, 2001; Lee et al., 2002; Mendelson and Lui, 2004; Engle and Recanzone, 2013; Brewton et al., 2016; Recanzone, 2018; Richardson et al., 2020), as well as population encoding in aged CBA mice (Shilling-Scrivo et al., 2021).

Reduced peripheral input can lead to alterations of intra-cortical microcircuits. Acoustic trauma or congenital hearing loss result in altered excitatory and inhibitory synaptic connections, including decreased inhibitory synaptic strength (Kotak et al., 2005; Kotak et al., 2008; Takesian et al., 2012). Mice of the C57Bl/6J strain show age-related hearing loss with significant high-frequency hearing loss present at 4 months of age and are a commonly used animal model of presbycusis (Henry and Chole, 1980; Willott, 1986; Li and Borg, 1991; Willott et al., 1993; O’Neill et al., 1997; Lee et al., 2002; Francis et al., 2003; Prosen et al., 2003; Ison et al., 2007; Martin del Campo et al., 2012; Engle and Recanzone, 2013; Trujillo et al., 2013; Brewton et al., 2016; Rogalla and Hildebrandt, 2020; Rumschlag et al., 2021; Rumschlag and Razak, 2021; Bishop et al., 2022). In particular, mice of the C57Bl/6 strain have been shown to have abnormal functional responses to sounds in the ACtx (Rumschlag and Razak, 2021; Bishop et al., 2022), including broader frequency tuning, reduced dynamic range, increased response duration, as well as a higher firing rates (Bishop et al., 2022). Abnormal tuning bandwidth has also been observed in aging rats (Turner et al., 2005).

Given the age-related changes in sound-evoked responses in C57Bl/6 mice, we thus speculated that these changes could reflect altered excitatory and inhibitory microcircuits. We thus investigated the changes in microcircuits of A1 in aging C57Bl/6J mice. We used laser scanning photostimulation (LSPS) and combined it with whole-cell patch clamp recordings of L2/3 cells. Our results show that L2/3 cells in aged A1 receive inputs from fewer presynaptic locations within L2/3. At the same time, the spatial laminar pattern of inhibitory connections is unchanged, but the strength of inhibitory connections is weaker. These changes are paralleled by a decrease in the frequency of mEPSCs and decreased amplitude of both mEPSCs and mIPSCs. Calculating the excitation/Inhibition ratio (E/I ratio) showed an imbalance towards inhibition in aged mice. Overall, our findings reveal a specific age-related laminar hypoconnectivity of intra-cortical excitatory and inhibitory circuits in A1. Our results thus suggest that not all circuits are changing equally and, therefore, therapeutic interventions must take these specific changes into account.

Methods

All procedures were approved by the University of Maryland Institutional Animal Care and Use Committee

Animals:

Male and female C57Bl/6J mice (Jackson Laboratory) were raised in 12-h light/12-h dark light cycle. C57Bl/6J mice from P370-P443 were used in the aging group (mean age P403). Young C57Bl/6J mice from P30 to P45 (mean age P40) were used as controls.

Slice Preparation:

Mice were deeply anesthetized with isoflurane (Halocarbon). A block of brain containing A1 and the medial geniculate nucleus (MGN) was removed, and slices (400 μm thick) were cut on a vibrating microtome in ice-cold artificial cerebrospinal fluid (ACSF) containing (in mM) 130 NaCl, 3 KCl, 1.25 KH₂PO₄, 20 NaHCO₃, 10 glucose, 1.3 MgSO₄, and 2.5 CaCl₂ (pH 7.35–7.4, in 95% O₂/5% CO₂). For A1 slices, the cutting angle is ~15 degrees from the horizontal plane to maintain the tonotopy (Zhao et al., 2009; Meng et al., 2015, 2017b; Meng et al., 2017a). Left hemisphere slices were incubated for 1 h in ACSF at 30°C and then kept at room temperature. For recording, slices were held in a chamber on a fixed-stage microscope (Olympus BX51) and superfused (2–4 ml/min) with high-Mg ACSF recording solution at room temperature to reduce spontaneous activity in the slice. The recording solution contained (in mM) 124 NaCl, 5 KCl, 1.23 NaH₂PO₄, 26 NaHCO₃, 10 glucose, 4 MgCl₂, and 4 CaCl₂. The location of the recording site in A1 was identified by landmarks (Cruikshank et al., 2002; Zhao et al., 2009; Meng et al., 2015).

Electrophysiology:

Whole-cell recordings were performed with a patch clamp amplifier (Multiclamp 700B, Molecular Devices) using pipettes with input resistance of 4–9 MΩ. Cells targeted for recording were located in an area of A1 overlying the rostral flexure of the hippocampus. Thus, we sampled cells in a restricted central area of A1. Targeted cells were located in layers 2/3 and showed pyramidal morphology. Data acquisition was performed by National Instruments AD boards and custom software (Ephus) (Suter et al., 2010), which is written in Matlab (Mathworks) and adapted to our setup. Voltages were corrected for an estimated junction potential of 10mV. Electrodes were filled with an internal solution containing (in mM): 115 cesium methanesulfonate (CsCH₃SO₃), 5 NaF, 10 EGTA, 10 HEPES, 15 CsCl, 3.5 MgATP, and 3 QX-314 (pH 7.25, 300 mOsm). Biocytin or neurobiotin (0.5%) was added to the electrode solution as needed. Series resistances were typically 20–25 MΩ.

Laser-Scanning Photostimulation:

LSPS was performed as described previously (Meng et al., 2015, 2017b; Meng et al., 2017a; Meng et al., 2019). Caged glutamate (0.8 mM N-(6-nitro-7-coumarylmethyl)-L-glutamate) (Muralidharan et al., 2016) was added to the high-divalent ACSF during recording. Laser stimulation (1 ms) was delivered through a 10x water immersion objective (Olympus). Laser power on the specimen was ~24 mW (23.4mW – 24.7mW) and was held constant between recordings. For each map, an array of up to 30 × 30 sites with 30-μm spacing was stimulated once at 1 Hz in a pseudorandom order. This stimulation paradigm evokes an action potential at the stimulation sites with similar spatial resolution (about 100 μm) over cells in all cortical layers (Meng et al., 2015, 2017b; Meng et al., 2017a). Putative monosynaptic excitatory postsynaptic currents (EPSCs) in GABAergic interneurons were classified by the post-stimulation latency of the evoked current. Evoked currents with latencies of less than 10 ms are likely to be the result of direct activation of glutamate receptors on the patched cell (Meng et al., 2015, 2017b; Meng et al., 2017a; Meng et al., 2019; Xue et al., 2022). Evoked currents with latencies between 10 ms and 50 ms were classified as monosynaptic evoked EPSCs. The first peak amplitude and the charge (the area of EPSC in the counting window) were quantified for each synaptic response. Recordings were performed at room temperature and in high-Mg²⁺ solution to reduce the probability of polysynaptic inputs. Cells that did not show any large (＞100 pA) direct responses were excluded from the analysis, as these could be astrocytes. Excitatory and inhibitory inputs were recorded at −70 mV and 0 mV respectively.

Statistics:

Data were analyzed by custom software written in MATLAB. Cortical layer boundaries were identified by features in the bright-field image as described previously (Meng et al., 2015, 2017b; Meng et al., 2017a; Xue et al., 2022). Input area was calculated as the area within each layer that gave rise to PSCs. Integration distance referred to the distance that covered 80% of evoked PSCs along the rostro-caudal direction. Mean charge was the average charge of PSCs from each stimulus spot. And the mean peak amplitude was calculated by the average amplitude of PSCs. The balance of excitation and inhibition was calculated by E/I ratio. The E/I ratio is based on the number of inputs and the average input strength (E_density/I_density and E_charge/I_charge). Results are plotted as means ± SD unless otherwise indicated. Populations are compared with a rank sum or Student’s t test (based on Lilliefors test for normality).

Results

To investigate the changes of intra-cortical circuits in A1 in aging C57Bl/6 mice, we performed LSPS, as described in our prior studies (Meng et al., 2015, 2017b; Meng et al., 2017a; Meng et al., 2019; Xue et al., 2022), in young adult (P30-P45) and in aged (older than P365) C57Bl/6 mice. Thalamocortical slices of A1 (Figure 1A) were cut and LSPS with caged glutamate was used to focally activate cortical neurons. Cell-attached patch recordings and whole-cell recordings were performed to test the photo-excitability of A1 neurons and the spatial connectivity of excitatory and inhibitory inputs to L2/3 neurons in A1 respectively (Figure 1A).

Figure 1. Photo-excitability of A1 Neurons Remains Unchanged During Aging.

A, Top, schematic of LSPS. Solid triangle represents recorded neuron. Red lines show the connections to recorded cell. Bottom, Infrared image of brain slice with patch pipette on L2/3 neuron. Stimulation grid is indicated by blue dots. Layer boundaries are indicated by white bars on the right. The scale bar at the bottom left is 200 μm. B-C, Number of action potentials (B) and effective stimulation distance (C) from cell-attached recordings of L2/3, L4 and L5/6 neurons. The numbers of evoked action potentials of neurons in all layers were similar (L2/3: p = 0.4049, L4: p=0.4129, L5/6: p=0.3755) and most spikes were evoked within 150 μm (L2/3: p = 0.4628, L4: p=0.9801, L5/6: p=0.5686).

Aging does not alter photo-excitability of A1 neurons.

To reliably compare the spatial connection pattern of cells in A1 using LSPS across age groups, we first confirmed that the spatial resolution of LSPS was similar across ages. We performed cell-attached patch recordings with LSPS in cells with pyramidal morphology from L2/3, L4 and L5/6 to test the ability of A1 neurons to fire action potentials in response to photoreleased glutamate. Short UV laser pulses (1 ms) were targeted to multiple stimulus locations to focally release glutamate and cause firing of action potentials. The grid of stimulation spots covered the entire A1, resulting in a high-resolution 2D photoactivation pattern for a given cell. We then counted numbers of evoked action potentials and measured the distance of effective stimulation sites. The numbers of evoked spikes were similar across ages (Figure 1B, L2/3: p=0.4049, median: 0.54 vs 0.50; L4: p=0.4129, median: 0.60 vs 0.60; L5/6: p=0.0755 median: 0.70 vs 0.67). Furthermore, the majority of action potentials were generated within 150 μm of the soma in both control and aged animals (Figure 1C, L2/3: p=0.4628 median: 80 vs 80; L4: p=0.9801 median: 40 vs 40; L5/6: p=0.5686 median: 160 vs 120;). These findings suggest that the sensitivity of cells to photoreleased glutamate and thus the spatial resolution of LSPS remains unchanged with age

L2/3 neurons in aged C57Bl/6J A1 show fewer and weaker intra-laminar excitatory connections.

We next investigated if intracortical circuits to L2/3 neurons change with aging. To visualize the connection pattern impinging on L2/3 neurons, we combined LSPS with whole-cell recordings from A1 L2/3 cells. Recorded cells had pyramidal morphology and were located at similar laminar positions (Figure 2A, p>0.05) and held at −70mV (E_GABA) to isolate excitatory synaptic currents. Laser pulses were targeted to ~900 distinct stimulus locations spanning all cortical layers around the recorded cell, and the resulting membrane currents were measured (Figure 2B). We observed large, short-latency (<10 ms) inward currents when the stimulation location was close to the recorded neuron due to direct activation of the cell body and the proximal dendrites. In contrast, post-synaptic currents caused by activation of presynaptic neurons were long-latency (>10ms) events (Meng et al., 2015, 2017b; Meng et al., 2017a; Meng et al., 2019; Xue et al., 2022).

Figure 2. Fewer and Weaker Intralaminar Excitatory Connections to A1 L2/3 Neurons in Aged Mouse.

A, Relative position of recorded neurons within L2/3. Plotted is the relative position within L2/3 with 10 referring to the border near L4 and 100 referring to the border with L1. Cells were sampled from the middle of L2/3 in young adult and aged animals (p=0.1113). B, Schematic of recording EPSCs. Whole-cell voltage clamp recordings at holding potentials of −70mV. If the presynaptic neuron synapses on the recorded neuron, an EPSC will be observed (black trace). Shown on the right are examplar patch-clamp recordings of direct response (red), EPSC (yellow), and IPSC (green), acquired at holding potentials of −70, −70, and 0 mV, respectively. Dashed blue line indicates time of photostimulation; solid blue line marks 8 ms post-stimulus, which is the minimal latency of synaptic responses. C, Left: Schematic illustration of how the connection probability map is calculated. Connection maps from recorded neurons are aligned to the soma location and averaged to create the connection probability map. Right: Maps of connection probability for excitatory connections in young adult (left) and aged (right) animals. Soma location is indicated by the white circle. Connection probability is encoded according to the pseudocolor scale. White horizontal lines indicate averaged laminar borders and are 100 μm long. D, Distributions of the area of excitatory input originating from L2/3 (top), L4 (middle), L5/6 (bottom) of young adult (black) or aged (red) animals; *, p<0.05. E, Fraction of excitatory input originating from L2/3 (top), L4 (middle), L5/6 (bottom) of young adult (black) or aged (red) animals; *, p<0.05. F, Distance of 80% of excitatory input to each L2/3 cell originating from L2/3 (top), L4 (middle), and L5/6 (bottom) of young adult (black) or aged (red) animals. Plotted is the laminar radius that covers 80% of inputs inside each layer. G, Distribution of numbers of excitatory input locations from rostral (solid) or caudal (dashed) side of each cell from L2/3 (top), L4 (middle), and L5/6 (bottom) of young adult (black) or aged (red) animals. H, Distributions of mean EPSC charge of inputs originating from L2/3 (top), L4 (middle), and L5/6 (bottom) of young adult (black) or aged (red) animals; **, p ＜ 0.01.

We mapped 75 L2/3 cells in A1 of C57Bl/6J mice (37 cells in 7 young adult animals and 38 cells in 7 aged animals). For each cell we identified stimulus locations that gave rise to an evoked EPSC and generated a binary input map. For each group, we aligned all maps to the soma position of the individual cells and averaged them (Figure 2C), resulting in a spatial connection probability map for excitatory inputs. These maps allowed us to identify cortical locations that over the population gave rise to inputs to L2/3 neurons. Qualitatively comparing the connection probability maps, we observed that in aged animals, fewer cortical locations seemed to provide inputs to L2/3 – indicating hypoconnectivity (Figure 2C). To quantify the changes, we calculated the laminar changes of the connection properties of each cell as in previous studies (Meng et al., 2015, 2017b; Meng et al., 2017a; Meng et al., 2019; Xue et al., 2022). We first identified layer boundaries in differential interference contrast (DIC) images. We next quantified the amount of convergence from each layer to L2/3. We calculated the total area within each layer from which EPSCs could be evoked. A comparison of input areas indicates that the amount of intralaminar excitatory input from within L2/3 is reduced in aging mice (Figure 2D). We observed a trend of increased input area from L4 and thus reasoned that aging might have affected the relative laminar balance of inputs. We thus computed the fraction of inputs originating from each layer and find that cells received relatively more inputs from L4 and L5/6 and fewer inputs from L2/3 (Figure 2E). These results indicate that L2/3 neurons in aged mice receive fewer inputs from within L2/3 and more inputs from thalamorecipient L4. Our areal measurement takes into account both changes in the input distribution within a layer, e.g., from L2 to L3, as well as changes in the orthogonal direction. Our thalamocortical slices preserve the macroscale rostro-caudally oriented tonotopic map in the slice plane. Therefore, the spatial extent of the input distribution along the rostro-caudal axis is a proxy for the integration along the tonotopic axis. To probe if the reduction of L2/3 inputs occurred along the tonotopic axis, we next calculated the distance that includes 80% of the evoked EPSCs and also calculated the fraction of EPSCs evoked towards the rostral or caudal side of the cell. We find that intralaminar integration distance is similar between the two groups (Figure 2F). Inputs on both the rostral and caudal side of the cells were reduced with aging, but the reduction of inputs was more prominent in the rostral side in aging animals (Figure 2G).

Functional circuit changes can occur through alteration of connection probabilities but also through changes of connection strength. Given that many synaptic proteins change with aging, we next tested if connection strength was altered. We measured the average size (transferred charge) of the evoked EPSCs and find that connections from within L2/3 showed weaker synaptic strength (Figure 2H). At the same time, the connection strength from deeper layers L4 and L5/6 was unchanged. Thus, these findings suggest fewer and weaker intralaminar connections from within L2/3 and an increased relative number of inputs from L4 in aged A1 and that this difference is most pronounced for inputs from the rostral (high-frequency) side.

L2/3 neurons in aged C57Bl/6J A1 show weaker and more focal intralaminar inhibitory connections.

We next investigated inhibitory circuits by holding cells at 0 mV and performing LSPS. We recorded evoked IPSCs and derived inhibitory input maps. Averaging these maps yielded inhibitory connection probability maps for both groups. In contrast to excitatory maps, the connection pattern of inhibitory circuits showed few qualitative differences (Figure 3A). Laminar analysis quantitatively supported these observations. L2/3 neurons from young and old mice received a similar amount of inhibitory input from each layer (Figure 3B, C). However, we found a reduction in the integration distance of L5/6 inputs (Figure 3D). Together, our observations indicate that L5/6 inputs are mostly originating from a focal area. Previous studies showed that the expression of rate-limiting enzymes, GAD65 and GAD67, which catalyze the conversion of glutamate to GABA, are decreased in aging in both IC and A1 (Milbrandt et al., 2000; Burianova et al., 2009), suggesting that strength of inhibition could be altered in aged mice. To test this hypothesis, we calculated the average charge of IPSCs. We found that IPSCs evoked from L2/3 neurons from aged mice had a smaller charge (Figure 3E). Thus, our findings show a selective weakening in intralaminar inhibitory inputs to L2/3 cells.

Figure 3. Weaker Intralaminar Inhibitory Connections to A1 L2/3 Neurons in Aged Mice.

A, Left, schematic of recording IPSCs. Whole-cell voltage clamp recordings at holding potentials of 0 mV. If the presynaptic neuron synapses onto the recorded neuron, an IPSC (black trace) will be observed. Right: Average maps of connection probability for inhibitory connections in young adult (left) and aged (right) animals. Maps are aligned to the soma location. Connection probability is encoded according to the pseudocolor scale. White horizontal lines indicate averaged laminar borders and are 100 μm long. B, Distributions of area of inhibitory input originating from L2/3 (top), L4(middle), and L5/6 (bottom) of young adult (black) or aged (red) animals. C, Distributions of fraction of total inhibitory input originating from L2/3 (top), L4(middle), and L5/6 (bottom) of young adult (black) or aged (red) animals D, Distance of 80% of inhibitory input to each L4 cell originating from L2/3 (top), L4 (middle), and L5/6 (bottom) of young adult (black) or aged (red) animals. Plotted is the laminar radius that covers 80% of inputs inside each layer. E, Distributions of mean IPSC charge of inputs originating from L2/3 (top), L4 (middle), and L5/6 (bottom) of adult (black) or aged (red) animals; ***, p ＜ 0.001.

mESPCs are reduced in frequency and amplitude in aged C57Bl/6J animals while mIPSCs show reduced amplitude

While LSPS can identify changes in meso-scale connectivity, it cannot measure very local connections due to the direct activation of the targeted neuron. We thus sought independent confirmations of our observations. We tested if spontaneous PSCs in L2/3 cells were altered in aged animals. We measured mEPSCs and mIPSCs by whole-cell recordings on L2/3 cells (N=21 cells from 4 adult animals; N=20 cells from 5 aged animals) in the presence of 500 nM tetrodotoxin (TTX) to block action potential propagation (Figure 4A). We then calculated the frequency, amplitude, and decay time of the mEPSC and mIPSCs in both groups. We found that the mEPSC rate decreases in aging animals while the mIPSC rate remained unchanged (Figure 4B). These results are consistent with the hypoconnectivity observed for excitatory connections in LSPS. In contrast, both mEPSCs and mIPSCs show reduced amplitude and decay time (Figure 4C, D). These changes in mPSC amplitude are consistent with the reduced amplitudes seen in LSPS recordings. Moreover, the reduced mPSC decay time constants suggest a change in the underlying receptor composition. Together, our results suggest that aging results in a loss of excitatory intralaminar L2/3 connections and in a weakening of both excitatory and inhibitory L2/3 connections to L2/3.

Figure 4. Spontaneous PSCs are altered in Aged mice.

A, Example traces showing mEPSCs and mIPSCs from adult (black) and aged (red) mice. B, Reduced mEPSC frequency in aged animals. Black bars for young adult group and red bars for aged group; *, p ＜ 0.05. C, Reduced mPSC amplitude in aged animals; *, p ＜ 0.05. D, Reduced decay time constant in aged animals; *, p ＜ 0.05 and **, p ＜ 0.01.

Decreased excitatory circuit similarity in aged C57Bl/6J mice

L2/3 cells in adult A1 show heterogeneity in their functional interlaminar circuits (Meng et al., 2017b). This heterogeneity emerges over development, coinciding with a decrease in functional activity correlations (Meng et al., 2019). While we observed a decrease in excitatory connections (Figure 2D), the spatial connection probability maps (Figure 2C) suggest a higher probability for excitatory connections originating from L4 just below the recorded neuron. This suggests that circuits from L4 to L2/3 may be less variable in old mice. To examine whether aging alters the circuit heterogeneity of L2/3 cells, we calculated the correlation between individual connection maps from all recorded cells (Figure 5A). We aligned the connection maps of recorded neurons within each group to their soma positions and compared the connection patterns between cells. This analysis showed that the correlation of excitatory circuits across all layers is decreased in old mice. However, the correlation between inhibitory circuits was similar between old and adult mice (Figure 5B). We next investigated if this effect varied across different layers. Connections from L2/3 and L5/6 showed a decreased similarity of excitatory connection patterns (Figure 5C). In contrast, inputs from L4 showed increased correlation between all connection maps, consistent with the locally increased spatial connection probability (Figure 2B). We did not observe any changes in inhibitory circuit similarity in any layer (Figure 5C). Together, these results indicate that aging changes the spatial functional connection pattern of excitatory inputs to L2/3 neurons, decreases the diversity of L4 inputs to L2/3, and increases the diversity of L2/3 and L5/6 connections.

Figure 5. Decreased similarity of excitatory circuits in aged mice.

A, Schematic illustration of how pairwise correlation between two maps is calculated. White circle represents recorded neuron. A yellow square represents a stimulus location that has monosynaptic connection to the recorded cell; a blue square represents a location with no connection to the recorded cell. For the pairwise correlation calculation, yellow and blue squares are assigned values of 1 and 0, respectively. B, Correlation of excitatory and inhibitory circuit patterns across all layers. Black bars represent adult animals and red bars represent aged animals. C, Correlation of circuit patterns within L2/3 (left), L4 (middle) and L5/6 (right). For all plots: *, p<0.05.

Imbalance towards inhibition in L2/3 of aged C57Bl/6J animals

The concomitant occurrence of synaptic excitation and inhibition maintains balance between excitation and inhibition. In all sensory cortices, excitation and inhibition wax and wane together when responding to sensory stimuli (Anderson et al., 2000; Swadlow, 2003; Wehr and Zador, 2003; Tan et al., 2004; Wilent and Contreras, 2005), and the balance of these two opposing forces is critical for proper cortical function (Sohal and Rubenstein, 2019). Given that the E/I ratio changes and is crucial during development (Hensch and Fagiolini, 2005; Chu and Anderson, 2015; Hu et al., 2017), we asked whether E/I balance also changes during the aging process. We calculated the E/I balance of the inputs from each layer based on the area and the charge of PSCs as in our prior studies (Meng et al., 2015). We found that E/I ratios based on density (Figure 6A) and on charge (Figure 6B) are both decreased in L2/3 in aged animals. These results indicate that the relative number of intralaminar excitatory inputs to L2/3 neurons decreased and that these inputs were relatively weaker. To gain an independent confirmation of these results, we next calculated the E/I balance based on the frequency and amplitude of mEPSCs and mIPSCs (Figure 6C). We find that the E/I balance of mPSC frequencies was reduced, consistent with a reduction in excitatory inputs. We observed no change in the relative amplitude. Together, our results show a change in the balance of functional inputs to L2/3 neurons towards inhibition.

Figure 6. Aged animals show E/I imbalance towards inhibition in intralaminar L2/3 inputs.

A, Distributions of E/I based on input numbers (density) from L2/3 (left), L4 (middle), and L5/6 (right) of young adult (black) or aged (red) animals; ***, p ＜ 0.01. Aged mice show lower E/I in L2/3 than young adult mice, evidenced by left shifts in the distribution. B, Distributions of E/I based on transferred charge from L2/3 (left), L4 (middle), and L5/6 (right) of young adult (black) or aged (red) animals; *, p ＜ 0.05. C, Distributions of E/I based on frequency and amplitude of mPSCs; *, p ＜ 0.05.

Discussion

Our results show a hypoconnectivity with weaker intralaminar connections and weaker inhibition within A1 L2/3 with aging in C57Bl/6J mice. The decreased E/I ratio indicates an imbalance towards inhibition, consistent with altered age-related functional responses in (Bishop et al., 2022).

We observed a loss of intralaminar excitatory inputs in L2/3 cells with age and a relative increase in L4 inputs. These results are consistent with the observation of reduced dendritic spine density in the aged brain (Dickstein et al., 2013; Hickmott and Dinse, 2013), and therefore suggest that these spines were innervated by L2/3 inputs. MGB cells in aged rats show hyperexcitability to reduced tonic inhibition (Richardson et al., 2013) possibly increasing thalamocortical drive. Assuming a homeostatic framework, the decreased intralaminar cortical connections in L2/3 could be a compensation mechanism for increased ascending drive from thalamocortical recipient layers or from higher order thalamic nuclei (Zhang and Bruno, 2019). Recent studies support the notion that L2/3 cells are involved in cross-modal plasticity and connect to higher-order cortical and subcortical areas (Lee and Whitt, 2015; Knöpfel et al., 2019; Zhang and Bruno, 2019; Wang et al., 2020). Thus, the changes in L2/3 circuits might affect cross-modal integration and plasticity (Peiffer et al., 2009; Glick and Sharma, 2017; Henschke et al., 2018).

We also observed a weakening of excitatory L2/3 connections. This is consistent with a loss of synaptic AMPA receptors and hypofunction of NMDA receptors as has been observed in the aged hippocampus and in other sensory areas (Gocel and Larson, 2013; Kumar et al., 2019). Taken together, our results suggest that the aged A1 shows reduced excitation, but that this reduction is selective to the supragranular connections.

Much focus has been placed on the changes in inhibition during aging (Chao and Knight, 1997; Kok, 1999; Caspary et al., 2008; Stanley et al., 2012; Rozycka and Liguz-Lecznar, 2017; Recanzone, 2018; Rogalla and Hildebrandt, 2020). In particular in mice of the C57Bl/6 strain, lower PV immunoreactivity in the auditory cortex was observed in layers 1–4 of old mice comparing with middle-aged animals (Martin del Campo et al., 2012). However, surprisingly, and unlike excitatory circuits, we did not observe any loss of functional inhibitory connections and only observed a weakening in inhibitory connections originating from L2/3. The unchanged connection probability of inhibition is consistent with the stable interneuron population after aging in rats and C57Bl/6 mice (Ouellet and de Villers-Sidani, 2014; Burianova et al., 2015; Rogalla and Hildebrandt, 2020). Meanwhile, no age-related changes in the overall number, total binding or affinity of GABA_A receptors have been observed in aged C57Bl/6 mice and Fischer 344 rats (Heusner and Bosmann, 1981; Komiskey and MacFarlan, 1983; Reeves and Schweizer, 1983; Komiskey, 1987). These results support our findings that the overall connection density of inhibitory circuits is unchanged. Furthermore, we observed a weaker local inhibition in aging mice consistent with decreased GAD level and structural changes on GABA_A receptors in rat A1 (Caspary et al., 1999; Ling et al., 2005; Caspary et al., 2013). The largest age-related GAD reduction was found in rat A1 layer 2 (Ling et al., 2005), which is consistent with the weaker connection strength in L2/3. Thus, we speculate that aging causes decreased release of GABA leading to a functionally weaker intralaminar inhibition. Such decrease in GABA release will functionally likely be most evident during conditions requiring sustained inhibitory tone such as suppression of neural activity due to noisy backgrounds (Shilling-Scrivo et al., 2021, 2022)

We found a reduction in the time constant of the mIPSC decay in aged animals suggesting a change in GABA_A receptor subunits. Indeed, the level of the α1 subunit of the GABA_A receptor decreases during aging along central auditory pathway including rat spiral ganglion neurons (SGNs), IC, MGB and auditory cortex (Caspary et al., 1999; Caspary et al., 2013; Tang et al., 2014; Ouda et al., 2015). The α1 subunit mediates relatively short-lasting IPSCs in mouse and rat (Goetz et al., 2007; Rudolph and Möhler, 2014); therefore, a decrease in the α1 subunit in A1 should lead to slower IPSCs, which is contrary to our results. However, the decrease in time constant we observed could suggest an increase in the relative amount of the α1 subunit and it could be that other subunits decrease even more than the α1 subunit. Moreover, the α1 subunit is not the only factor that affects decay of the IPSC; other factors that lower synaptic neurotransmitter concentration in the synaptic cleft could play a role (Nusser et al., 2001; Petrini et al., 2011; Labrakakis et al., 2014). Nevertheless, our results show a weakening of intralaminar IPSCs and a change in the mIPSC decay time constants are likely due to the reduction of multiple GABA receptor subunits as well as reduced GABA synthesis.

Aging C57Bl/6 A1 cells also show increased excitatory bandwidth and decreased dynamic range as well as increased response duration (Bishop et al., 2022), all suggestive of altered inhibition consistent with our findings. In particular, we found that the inhibitory input area originating from L4 is more concentrated in aged animals. It has been commonly accepted that inhibition is critical for shaping tuning properties (Sillito, 1977, 1979; Kyriazi et al., 1996; Wang et al., 2002). Since the integration distance preserved tonotopic organization, a less dispersive input area might contribute to an abnormal tuning bandwidth observed by prior studies (Turner et al., 2005; Bishop et al., 2022). Moreover, the relative increase of excitatory L4 inputs we observe might also contribute to altered tuning, as L4 neurons show broader tuning from L2/3 neurons (Bandyopadhyay et al., 2010; Winkowski and Kanold, 2013; Bowen et al., 2020).

Although changes of excitation and inhibition happen together during aging, our results show a decreased E/I balance. The lower ratio indicates that the excitation decreases more than inhibition to create an imbalance in the aged A1. A similar imbalance has been found in the parietal cortex with age-related cognitive impairment, but not in aged animals with normal cognition (Wong et al., 2006). Since intracortical inhibition shapes the functional properties of cortical neurons and influences the responses to sensory stimuli (Wang et al., 2000; Wang et al., 2002; Kurt et al., 2006; Razak and Fuzessery, 2009; Sun et al., 2010; Gaucher et al., 2013; Li et al., 2014; Zhou et al., 2014; Butman and Suga, 2019), such imbalance to inhibition could be an important substrate of functional age-related impairments.

Mice of the C57Bl/6J strain have accelerated hearing loss with significant high-frequency hearing loss after 4 months and have been used as a model of presbycusis (Henry and Chole, 1980; Willott, 1986; Li and Borg, 1991; Willott et al., 1993; O’Neill et al., 1997; Lee et al., 2002; Francis et al., 2003; Prosen et al., 2003; Ison et al., 2007; Martin del Campo et al., 2012; Engle and Recanzone, 2013; Trujillo et al., 2013; Brewton et al., 2016; Rogalla and Hildebrandt, 2020; Rumschlag et al., 2021; Rumschlag and Razak, 2021). Thus, the imbalance to inhibition could be at least partially due to this peripheral impairment potentially due to homeostatic mechanisms. However, reduced peripheral function would be expected to lead to reduced activation of A1 L4. While, MGB neurons in aged rats show hyperexcitability (Richardson et al., 2013) this might not fully compensate for decreased peripheral drive. Furthermore, hyperexcitability could also increase spontaneous activity reducing sound induced firing rate changes and possibly causing synaptic depression at the thalamocortical synapse. Together, sound induced activation of A1 L4 would be lowered and this decrease would be expected to lead to a reduction in L4→L2/3 inputs and a compensatory increase in L2/3→L2/3 or L5/6→L2/3 inputs. Indeed, cochlea lesioning in adults results in an increase in intracortical gain (Chambers et al., 2016) which is also present in aged C57Bl/6 mice albeit with reduced temporal fidelity (Rumschlag and Razak, 2021). Aged mice of the C57Bl/6 strain show age-related decrease in resting gamma power, a reduction in gamma-range synchrony to time varying stimuli, and an increase in noise evoked and induced gamma power (Rumschlag et al., 2021). Moreover, this reduction in temporal processing was dissociable from hearing loss suggesting an age-related component. Thus, the selective reduction in intralaminar excitatory inputs likely reflects age-related changes not primarily due to cochlear impairment in C57Bl/6 mice.

Taken together, our findings delineate age related changes in C57Bl/6 mouse primary auditory cortex. We demonstrate a loss and weakening of intralaminar excitatory circuits in aged animals, while local inhibition was weaker and more spatially concentrated. The simultaneous alterations of excitation and inhibition led to an imbalance towards inhibition. Our results suggest that the altered response properties to sound in aged C57Bl/6 mice are due to the changes of both excitatory and inhibitory connections.

Highlights.

• Age-related hearing loss (presbycusis) manifests as impaired spectrotemporal processing. Mice of the C57BL/6J strain show premature age-related hearing loss and changes in functional responses in auditory cortex. Investigating the underlying circuit changes, we found that compared to young mice L2/3 cells in aged C57Bl/6 mice have fewer excitatory connections with weaker connection strength. Whereas young adult and aged C57/Bl6 mice have similar amounts of inhibitory connections, the strength of local inhibition is weaker in aged mice. We speculate that unbalanced changes in cortical circuits contribute to the functional manifestations of presbycusis.