Age-related increase in the excitability of mouse layer V pyramidal neurons in the primary motor cortex is accompanied by an increased persistent inward current

Abstract

Sarcopenia, or pathological age-related loss of muscle strength and mass, contributes to physical function impairment in older adults. While current understanding of sarcopenia is centered mostly on neuromuscular mechanisms, mounting evidence supports that deficits at the level of the primary motor cortex (PMC) play a significant role. Despite the importance of the PMC to initiate movement, understanding of how age affects the excitability of layer V pyramidal neurons (LVPNs) of the PMC is limited. To address this, we used the whole-cell patch clamp technique to measure the excitability of LVPNs of the PMC in young, late adulthood, and old mice. Old LVPNs had increased firing frequency and membrane input resistance, but no differences in action potential kinetics versus young and late adulthood mice. Since changes in the persistent inward current (PIC) are known to contribute to changes in motor neuron excitability, we measured LVPN PICs as a putative contributor to LVPN excitability. The PIC amplitude was increased in old LVPN via increases in Na⁺ and Ca²⁺ PICs, in addition to being active across a wider voltage range. Given that LVPN function is integral to initiation of voluntary muscle contraction, altered LVPN excitability likely contributes to age-related impairment of physical function.

Introduction

Aging is intricately linked to progressive decline across physiological systems and is a major risk factor for chronic disease. Among manifestations of aging, sarcopenia or pathological age-related skeletal muscle failure stands out as a significant concern with far-reaching implications for public health [1]. Sarcopenia contributes not only to difficulties in daily activities, increased fall risks, and decreased quality of life but is also a major predictor of all-cause mortality [2], with current treatment options being limited to exercise and nutritional interventions [3]. The current consensus agrees that sarcopenia is a disorder involving muscle weakness, with a growing recognition of its multifactorial nature and increased emphasis on the potential neural mechanisms [4] driving this age-related pathology. With an increasing age of the world’s population, there is an urgent need for an improved understanding of the neural mechanisms of sarcopenia and the development of improved treatments.

Over the past decade, the focus of sarcopenia research has significantly shifted to incorporate a greater focus on exploring potential neural mechanisms, but these investigations have primarily focused on peripheral nervous system pathologies at the level of lower motor neurons and the neuromuscular junction [4–6]. Yet, normal muscle contractions and motor control depend on the orchestrated signaling between the central and peripheral nervous systems and skeletal muscles [7, 8]. Importantly, appropriate motor control is complex and requires areas of the central nervous system like the primary motor cortex, which play a critical role in motor skill acquisition, motor pattern saliency, movement inhibition, and motor sensory feedback [9]. Additionally, the primary motor cortex, in particular pyramidal neurons in layer V (LVPNs), can drive voluntary motor sequences that are then transmitted to lower motor neurons through corticospinal tracts and subsequently to skeletal muscle via neuromuscular junctions [10, 11].

Despite the essential role of the central nervous system in motor control, in particular the primary motor cortex, understanding of how aging impacts cortical motor circuits and how deficits in this area may contribute to sarcopenia remains incomplete [4]. A few studies, utilizing indirect assessments such as transcranial magnetic stimulation (TMS), suggest that excitability of the primary motor cortex and corticospinal tracts may be altered, particularly in weak older adults (OAs) [12–15]. However, the results from these studies show that age-related changes at the level of the primary motor cortex are nuanced, with some studies finding limb-specific changes in cortical excitability [16] and other studies reporting changes in cortical excitability without considering age-related changes in motor unit connectivity [15]. Thus, while assessments like TMS provide a valuable read-out of the excitability of the primary motor cortex in aging, which is useful, they do not directly inform on the function and mechanisms driving potential excitability changes of aging LVPNs.

Accordingly, we used whole-cell patch-clamp recordings to investigate the age-related changes in single-cell excitability of LVPNs from slices of the mouse primary motor cortex. First, we sought to determine the presence of age-related changes in the intrinsic membrane excitability of LVPNs. Second, we also aimed to shed light on putative intrinsic mechanisms of neural decline during aging that might contribute to loss of physical function in OAs. To do so, we measured age-related changes in the persistent inward current (PIC) of LVPNs across the lifespan including young, late adulthood, and old mice (Fig. 1A). Decrements in the PIC at the level of spinal motor neurons have been associated with decreased spinal excitability and motor function in aging [17]. Thus, we hypothesized that age-related changes in the PIC could be associated with changes in LVPNs excitability.

Fig. 1.

Experimental design to investigate the excitability of LVPNs across the mouse lifespan. A For experiments measuring LVPN excitability using a whole-cell current clamp, 3 mouse age groups (50% female) were selected: young (2 months; N¹ = 8 mice, n² = 51), late adulthood (12 months; N = 6, n = 37), and old (23 months; N = 6, n = 38). For experiments measuring the persistent inward current (PIC) using a whole-cell voltage clamp, 2 mouse age groups were selected (50% female): young (2 months; N = 4, n = 24) and old (23 months; N = 4, n = 24). B Whole-cell patch clamp experiments were performed on tissue harvested from the brains of each age group. We identified the primary motor cortex, generated coronal slices (300 µm thick), and identified LVPNs. ¹N, mouse sample size; ²n, LVPN measurements

Methods

Animals

All experiments performed were approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Missouri–Columbia (protocol # 39,905). C57BL/6N (Charles River Laboratories) mice were used for all experiments. Whole-cell LVPN current-clamp experiments were performed in mice at three ages (“N” = mouse sample size; “n” = LVPN measurements): young (2 months: N = 8, n = 51); late adulthood (12 months: N = 6, n = 37); and old (23 months: N = 6, n = 38). Whole-cell LVPN voltage-clamp experiments to measure the PIC were performed in a separate group of mice at two ages: young (2 months: N = 4, n = 24) and old (23 months: N = 4, n = 24). Whole-cell LVPN voltage-clamp experiments to measure the sodium and calcium components of the PIC were performed in a separate group of mice at two ages: young (3 months: N = 4, n = 24) and old (24 months: N = 4, n = 24). Each age group had equal female and male distribution. The age range of 2 to 24 months was chosen to cover the majority of the lifespan in C57BL/6N mice and approximates roughly 20 to 70 years of age in humans [18].

Brain slice preparation

Mice were deeply anesthetized via isoflurane inhalation (open-drop exposure in a tightly sealed bell-jar chamber: 95% isoflurane, approximately 1 cc/500 cc jar volume). After the absence of the toe-pinch reflex, mice were quickly decapitated, and the brain was quickly removed. Given the reduced slice viability and known increased difficulty in patching the neurons of old mice [19], we used a modified version of the NMDG-recovery method. This method has been shown to improve slice viability and giga-ohm seal formation [19, 20]. Additionally, multiple studies have used this or a modified form of the NMDG-recovery method to improve slice viability in aging animals [21–23]. Brains were then submerged in the artificial cerebral spinal fluid (ACSF) NMDG-ACSF [19] containing (in mM) the following: 92 NMDG, 2.5 KCl, 20 HEPES, 30 NaHCO₃, 1.25 NaH₂PO₄, 25 D-glucose, 0.5 CaCl₂*2H₂O (added 500 μL from 1M stock solution), 10 MgCl₂*2H₂O (added 10 mL from 1 M stock solution), 2 Thiourea, 5 Na-ascorbate, 3 Na-pyruvate, and equilibrated to pH = 7.3–7.4 with 95% O₂/5% CO₂ (Millipore Sigma, Burlington, MA, USA; Fisher Scientific, Hampton, NH, USA). Brains were then blocked in the coronal plane by removing the cerebellum, glued with cyanoacrylate glue (Loctite SuperGlue), and then coronal brain slices were made (300-μm thick) from rostral to caudal using an Electron Microscopy Sciences 7000 vibratome (frequency 70Hz, amplitude 1.50 mm, advance speed 0.10 mm/s). Slicing was performed under 10 min while brains were fully submerged in ice-cold bubbled NMDG-ACSF. The primary motor cortex was isolated by cutting slices between − 1 and + 2.30 mm from the bregma [24, 25]. Slices were then transferred to a holding chamber containing pre-warmed NMDG-ACSF at 34–35 °C and incubated for no more than 30 min. Slices were then transferred to a holding chamber with regular ACSF, containing (in mM): 125 NaCL, 2.5 KCl, 5 HEPES, 25 NaHCO₃, 1.25 NaH₂PO_4, 12.5 D-glucose, 2 CaCl₂*2H₂O (added 2 mL from 1 M stock solution), 1 MgCl₂*2H₂O (added 1 mL from 1 M stock solution). Slices were incubated for at least 45 min to 1 h at room temperature prior to performing any recordings. Upon transfer to the recording chamber, slices were perfused with regular ACSF at a rate of 2 mL/min. The primary motor cortex (Fig. 1B) was identified depending on the slice distance from the bregma (atlas.brain-map.org). We anatomically identified layer V motor neurons of the primary motor cortex based on their large soma size of about 20 μm [26, 27] and the presence of large apical dendrites [28, 29] compared to neurons from layers 2/3 or layer 6.

Electrophysiology

Slices were transferred to a recording chamber of an upright Nikon FN1 eclipse microscope (Nikon Instruments, Melville, NY, USA) and visualized with 60 × magnification under IR-DIC optics with a Nikon FN1 eclipse camera (Nikon Instruments, Melville, NY, USA) and perfused with regular-ACSF at room temperature. All current-clamp and voltage-clamp experiments were performed on the Axopatch 200B amplifier using a Digidata 1550B (Molecular Devices, San Jose, CA, USA). Data were sampled at 10 kHz and low-pass filtered at 5 kHz. Pipettes for current-clamp recordings were pulled in a 2-stage protocol on a P-1000 puller (Sutter Instruments, Novato, CA, USA) and had a tip resistance between 4 and 4.5 MΩ. Patch pipettes were filled with intracellular solution containing (in mM): 135 K-gluconate, 4 KCl, 2 MgCl₂, 10 HEPES, 0.3 EGTA, 4 Na₂ATP, 0.3 Na₂GTP, pH = 7.3 adjusted with KOH. We did not correct voltage recordings for the liquid junction potential (calculated potential = 10 mV). We purchased thin-wall glass for patch pipette electrodes from Harvard Apparatus (1.5OD × 1.17ID × 100L mm: Holliston, MA, USA). After entering the whole-cell configuration, neurons were voltage-clamped at − 65 mV and allowed to stabilize for at least 3 min. Specifically for current-clamp experiments, while we did not directly compensate for series resistance (R_s), we performed the following measures of quality control during recordings: (1) neurons that had an R_s greater than 20 MΩ or that changed by at least 20% were always discarded; (2) neurons were also discarded if they began sealing up (the capacitive transient current in the whole-cell configuration becomes noticeably narrower, thereby increasing R_s) and we were unable to reverse this with slight negative pressure; (3) pipette tip resistances was always in the same range as described above, and recordings began immediately after 3 min of stabilization (as describe above).

To generate the frequency-current relationship (F-I) of motor neurons and to calculate active and passive properties, whole-cell current-clamp recordings consisted of incremental 500 ms 50 pA current steps from − 150 to 300 pA. The duration of each pre-step before a current injection was applied was 100 ms, and the duration of the post-step after a current injection was applied was 900 ms. The total duration of each step (pre-step + current injection step + post-step) was 1500 ms. Therefore, the total amount of time elapsed between the end of a current injection step (500 ms) and the beginning of another current injection in the subsequent step was 1000 ms. Easy Electrophysiology software (www.easyelectrophysiology.com London, UK) was used offline to analyze all recordings. The following passive properties were calculated. RMP was measured from 0 pA of input current. Input resistance (R_in) was calculated from the first current step at − 150 pA as (Average V_m of pre-step baseline – Average Vm of step decay)/ − 150 pA). The membrane time constant (Tau) was calculated at the − 150 pA step using a mono-exponential fit to membrane-voltage, where the fit duration was 50 ms starting at 2 ms after current injection. Membrane capacitance (C_m) was calculated as Tau/ R_in. The voltage sag ratio was calculated at the − 150 pA step as the ratio of the sag hump (difference between peak V_m and steady state V_m) and the maximum downward deflection from baseline. Values for active properties were determined by averaging the values for all action potentials (AP) generated at the 50 pA step (Fig. 3A). AP amplitude was calculated as the ΔVm between AP peak and AP threshold voltage. AP half-width was calculated as the duration of AP at half of the maximal amplitude. AP threshold was calculated as the Vm measured 0.5 ms before the onset of the AP. Medium-duration afterhyperpolarization (mAHP) was calculated as the voltage difference between the fast AHP peak and minimum baseline V_m after the fAHP.

Fig. 3.

Passive properties of LVPNs are altered with age. A To quantify passive properties, a -150 pA current step lasting 500 ms was used across the 3 age groups. B Resting membrane potential (was not significantly different across the age groups (p > 0.05, one-way nested ANOVA). C Membrane input resistance (R_IN) was significantly higher for old (23 months) LVPNs when compared to young (2 months) LVPNs but not different when compared to late adulthood LVPNs (12 months) and (p < 0.0001 and p > 0.05, one-way nested ANOVA). Late adulthood LVPNs had a higher R_IN than young LVPNs (p < 0.001, one-way nested ANOVA). D The membrane time constant (Tau) was significantly higher for old LVPNs compared to young and late adulthood LVPNs (p < 0.001 and p < 0.05, one-way nested ANOVA). Tau for late adulthood LVPNs and young LVPNs was not different (p > 0.05, one-way nested ANOVA). E The sag ratio of young LVPNs was bigger than that of late adulthood LVPNs (p < 0.01, one-way nested ANOVA) but not significantly different when compared to old LVPNs (p > 0.05, one-way nested ANOVA). F Membrane capacitance of LVPNs was not different across age groups (p > 0.05, one-way nested ANOVA). The sag ratio of old LVPNs was not significantly different from late adulthood LVPNs (p < 0.05, one-way nested ANOVA). Error bars represent the standard deviation (SD) from the mean. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

To generate slow 4-s triangular current bi-ramps, we increased the current at 125 pA/s in the ascending portion of the ramp and decreased it at 125 pA/s in the descending portion of the ramp with a peak ramp current of 250 pA (Fig. 5A) [24, 30–34]. Slow current ramps are used to uncover the presence of prolonged membrane depolarization that can persist even after the input current is decreased [35]. Thus, slow triangular current ramps are utilized for the detection of sustained repetitive firing that is driven by the persistent inward current (PIC) [30, 33–38] and have been used to identify the presence of a PIC in other brain areas like the medulla [33]. During the ascending portion of a slow current ramp, input current triggered repetitive firing in LVPNs (Fig. 5A). If firing continued even when the input current began to decrease in the descending ramp, we considered this to suggest the presence of an LVPN PIC (Fig. 5A) [30, 35, 37]. From this protocol, we calculated the following parameters. We measured the on and off currents similarly to previous studies, with the on-current calculated as the minimum current (rheobase) to elicit the first action potential, and the off-current calculated as the current necessary to elicit the last action potential (Fig. 5A) [33, 37–39].

Fig. 5.

Smaller rheobase and prolonged sustained repetitive firing in old LVPNs. A To quantify sustained repetitive firing properties, a slow triangular 4 s current ramp was used, with 125 pA/s on the ascending ramp, and 125 nA/s on the descending ramp, with a peak current of 250 pA. Representative traces are shown for LVPNs from each group. Qualitatively, old (23 months) LVPNs have a smaller rheobase on the ascending ramp, higher frequency on the ascending and descending ramps, and longer firing duration in the descending ramp. B The on-current (rheobase) for repetitive firing was smaller in old LVPNs when compared to late adulthood (12 months) and young (2 months) LVPNs (p < 0.01 and p < 0.01, one-way nested ANOVA). The on-current was not significantly different between the young and late adulthood group (p > 0.05, one-way nested ANOVA). C The off-current (current at which repetitive firing terminated) was smaller in old LVPNs when compared to late adulthood and young LVPNs (p < 0.05 and p < 0.05, one-way nested ANOVA). The off-current was not significantly different between the young and late adulthood groups (p > 0.05, one-way nested ANOVA). D Old LVPNs had a higher action potential (AP) frequency on the ascending portion of the current ramp when compared to young LVPNs but not when compared to late adulthood LVPNs (p < 0.01 and p > 0.05, one-way nested ANOVA). The firing frequency on the ascending portion of the ramp was not significantly different between young and late adulthood LVPNs (p > 0.05, one-way nested ANOVA). E Old LVPNs had a higher action potential (AP) frequency on the descending portion of the current ramp when compared to the young LVPNs but not compared to the late adulthood descending portion of the ramp was not significantly different between young and late adulthood LVPNs (p > 0.05, one-way nested ANOVA). Error bars represent standard deviation (SD) from the mean. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

During voltage-clamp experiments, to ensure adequate cancelation of the pipette capacitance, patch pipettes were wrapped with a thin layer of parafilm to at least 200 μm of the pipette tip [40]. Upon breaking into a cell, a 10 mV pulse was applied in the voltage-clamp configuration from a holding potential of − 65 mV and we ensured that stable transient currents were observed for at least 1 min. If the cell began to re-seal (the transient currents became smaller), we applied gentle negative pressure to ensure stable transient currents. Series resistance (R_s) was estimated by dividing the 10 mV pulse by the observed transient current amplitude [40]. The average uncompensated R_s for LVPNs from old animals was 13.5 MΩ ± 2.6 and 15.4 MΩ ± 2.3 for young animals. There was no significant difference between the R_s of old and young animals (Student’s T-test, p = 0.9790). The uncompensated R_s value was then dialed into the Axopatch 200B amplifier as the calculated R_s. R_s was compensated by 80–90% for all cells. We calculated, dialed in, and measured R_s as described in Gray and Santin (2023) [40]. For all voltage-clamp recordings, R_s did not change by more than 20% between the beginning and end of voltage-clamp protocols.

After R_s compensation, we measured PICs. To do so, we used a triangular ramp protocol of 10 mV/s, with a holding potential of − 100 mV on both sides of the ramp and a peak of − 20 mV [32, 33, 38]. Preliminary studies revealed the activation of voltage-dependent K + currents during the ramp. Thus, for all PIC recordings, all recordings were made in 20 mM of tetraethylammonium chloride to block these currents [33] (TEA; purchased from TCI chemicals). To do so, we made a TEA-ACSF, which was the same as regular-ACSF with the exception of a 20 mM equimolar substitution of TEA for NaCl (NaCl was reduced to 105 mM from 125 mM). After recording of PICs, we performed a subtraction of the leak current [33, 41]. This allowed us to better visualize and calculate the different PIC parameters. To leak subtract, during the ascending portion of the ramp protocol, we drew a line along the linear phase of the rising current (Fig. 6A). From this line, we calculated the R_in and used the pCLAMP software (Molecular Devices, San Jose, CA, USA) to subtract the linear phase from the total raw PIC current trace. We then calculated the following PIC parameters (Fig. 6B). The PIC peak current was calculated as the maximal inward current during the ramp protocol compared to the baseline [33, 37, 38]. The PIC peak voltage was calculated as the voltage at which the peak PIC amplitude occurred. To calculate The PIC onset voltage, we drew a line along the baseline of the recorded current corresponding to the ascending portion of the voltage ramp [33]. To calculate the onset voltage of the PIC we used similar approaches as in previous studies measuring the PIC [33, 34, 37, 39]. Briefly, we drew a line along the baseline current, and then measured the point at which the current was no longer parallel to this line, but became tangential to this line, with the corresponding voltage on the voltage ramp defining the onset voltage (Fig. 6B). To calculate the PIC offset voltage, we measured the point at which the current was no longer tangential to the line drawn along the baseline, but became parallel with it, with the corresponding voltage on the ramp defining the offset voltage (Fig. 6B) [37]. To calculate the PIC total voltage range, we measured the absolute magnitude between the onset and offset voltage of the PIC. To calculate PIC hysteresis, we compared the difference between the offset and onset voltage of the PIC for each age group. To isolate and measure the different components of the PIC, we sequentially applied pharmacological blockers in each LVPN, after which we performed a series of offline subtractions similar to Quinlan et al. [38]. To measure the sodium component of the PIC, we first measured PICs in TEA (Total-PIC) as previously described. We then blocked the sodium PIC with 500 nM tetrodotoxin [37, 38] (TTX; HelloBio, Princeton, New Jersey, USA) and measured the remaining PIC in TTX (PIC in TTX). To quantify the different parameters of the sodium PIC, we performed an offline subtraction of the PIC in TTX from the total-PIC using the pCLAMP software. To isolate the calcium component of the PIC, after removing the sodium PIC in TTX, we blocked the calcium PIC with 25 μM nimodipine (Alfa Aesar, Tewksbury, MA, USA) [33, 37] and measured the PIC in nimodipine (PIC in nimodipine). To quantify the different parameters of the calcium PIC, we performed an offline subtraction of the nimodipine-PIC from the TTX-PIC using the pCLAMP software. To quantify the different parameters of the resistant PIC (the PIC that remained in the presence of both TTX and nimodipine), we used only the nimodipine-PIC current traces.

Fig. 6.

Old LVPNs have larger persistent inward currents with altered voltage-sensitivity. To quantify the PIC, triangular voltage ramps (10 mV/s) were run using the whole-cell voltage-clamp configuration. A Representative trace of non-leak subtracted PICs. PICs were measured in the presence of 20 mM tetraethylammonium to eliminate I_K contamination. PIC was leak subtracted offline by calculating the input resistance of the leak (linear current during ascending ramp) and subtracting this value from the total non-leak subtracted trace. B Representative traces of the persistent inward current across young (2 months) and old (23 months) LVPNs. Qualitatively, old LVPNs have a bigger peak PIC that is achieved at more hyperpolarized voltages (refer to pink and gray dotted lines); the PIC activates and deactivates closer to the resting membrane potential (refer to pink and gray dotted lines). Old LVPNs have a bigger peak PIC (C), smaller PIC onset voltage (D), smaller PIC peak voltage (E), smaller PIC offset voltage (F), bigger total PIC voltage range (G), and bigger changes in hysteresis with a more hyperpolarized deactivation (when compared to activation voltage) when compared to young LVPNs (p < 0.05, p < 0.05, p < 0.01, p < 0.001, p < 0.01, p < 0.05, nested T-test). Error bars represent the standard deviation (SD) from the mean. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Statistical analysis

Statistical analyses and figure generation were performed using GraphPad Prism 10 (GraphPad Version 10.1.1 (270) Software, San Diego, CA, USA) and “R studio” (version 2023.09.0 + 463). A mixed effect two-way ANOVA (that accounts for interactions between multiple patch-clamp recordings from within the same animal) was used to analyze data sets that had 2 simple main effects. We used nested one-way ANOVAs for data sets that had 1 simple main effect. Nested ANOVAs were followed by Tukey’s Honest Significance Difference test to quantify multiple comparisons. Group data are presented as means ± standard deviation (SD). Groups were considered significantly different when p < 0.05. For figures showing data, each circle corresponds to data collected from a single LVPN. Additionally, each data point corresponds to an outcome measure from a single patch-clamp protocol run in a single LVPN and does not represent the average of multiple protocol runs. The total number of cells is represented by the letter “n,” total number of animals by the letter “N” (refer to Methods 2.1 and Fig. 1). For current-clamp experiments, based on pilot data collected, we performed a power analysis where a sample size of at least n = 30 per group gave us 97% power to detect a 53% difference in action potential firing frequency at 50 pA (firing frequency at 50 pA used as a primary outcome; 2 months old: 10.5 Hz ± 7.3; 23 months old: 18.2 Hz ± 8.2). For voltage-clamp experiments, we performed a power analysis where a sample of size of at least n = 24 per group gave us 90% power to detect an 18.7% difference in peak PIC current (peak PIC current used as a primary outcome; 2 months old: 23 pA ± 2.4; 23 months old: 27.8 Hz ± 7.6).

Results

Excitability is increased in old LVPNs

First, we wanted to determine if age had an effect on the excitability of LVPNs. To do so, we injected incremental current steps and then calculated the firing frequency (Fig. 2A).

Fig. 2.

Excitability is increased in old LVPNs. A To quantify frequency-currents relationships of LVPNs, whole-cell current clamp recordings consisted of incremental 500 ms 50 pA current steps from − 150 to 300 pA. Representative traces of 50 pA and 300 pA current-clamp traces are shown for each age group. Qualitatively, the same stimulus elicits higher firing frequency in old LVPNs. B Bottom left. Frequency-current relationship curves for the 3 age groups. Action potential frequencies for current steps from − 150 to 0 pA are not shown because in all cases, neurons from each age group did not spike at those current steps. A mixed effect two-way ANOVA (p > 0.05, age x current-step interaction; p < 0.0001, age or current-step simple main effect) revealed that at every current step at or above 50 pA (multiple comparisons are denoted by letters), old (23 months) LVPNs had higher action potential frequencies than late adulthood (12 months) and young (2 months) LVPNs, but late adulthood and young LVPNs were not significantly different (Tukey’s HSD). Bottom right. At 50 pA, old LVPNs had higher action potential frequencies than late adulthood and young LVPNs (p < 0.05 and p < 0.01, mixed effect two-way ANOVA and Tukey’s HSD). Young and late adulthood LVPNs were not different from each other (p > 0.05). At 300 pA old LVPNs has higher firing frequencies than late adulthood or young (p < 0.001 and p < 0.001, mixed effect two-way ANOVA and Tukey’s HSD). Young and late adulthood were not significantly different (p > 0.05). Error bars represent the standard deviation (SD) from the mean. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

A mixed effect two-way ANOVA revealed no significant interaction between mouse age and current-step (p = 0.3690). However, both main effects (age and current-step) had a significant effect on firing frequency (p < 0.001 and p < 0.001 respectively). We found that − 150 to 0 pA produced no action potentials and therefore a firing frequency of 0 Hz, with no significant difference between firing frequencies (Table 1). We found that at every current injection step at or after 50 pA, old LVPNs had a significantly higher firing frequency compared with the late adulthood and young LVPNs (Fig. 2B and Table 1). Specifically, at 50 pA (Fig. 2B and Table 1), the old LVPNs had 52% increases in firing frequency than late adulthood (17.3 Hz ± 9.7 and 11.4 Hz ± 7.8, p = 0.0344) and 75% increase when compared to young LVPNs (17.3 0 ± 9.7 Hz and 9.9 Hz ± 10.2, p = 0.0078). Young and late adulthood LVPNs were not significantly different from each other. At 300pA (Fig. 2B and Table 1), old LVPNs had a 34% increase in firing frequency compared to late adulthood (40.7 Hz ± 11 and 30.4 Hz ± 12, p = 0.0007) and 33% increase when compared to young LVPNs (40.7 Hz ± 11 and 30.6 Hz ± 11.5, p = 0.0006). However young and late adulthood LVPNs were not significantly different from each other. These results show that when given the same current stimulus, the firing frequency of old LVPNs is higher when compared to young and late adulthood LVPNs, but generally does not change between young and late adulthood animals.

Table 1.

Quantification of active properties, passive properties, action potential properties, repetitive firing during slow current ramps, and PIC properties collected from LVPNs of young (2 months), late adulthood (12 months), and old (23 months) animals

Mean and SD				Multiple comparisons
	A = 2 month (n = 51)	B = 12 months (n = 37)	C = 23 months (n = 38)	2 months vs 12 months	2 months vs 23 months	12 months vs 23 months
Passive properties (nested one-way ANOVA)
Rin (MΩ) ANOVA p-value: < 0.0001 ANOVA F-value: 18.31	147.8 ± 26.6	179.2 ± 38.1	192.5 ± 44	***	****	ns (p = 0.4323)
RMP (mV) ANOVA p-value: 0.5542 ANOVA F-value: 1.934	− 60.6 ± 1.6	− 62.89 ± 1.5	− 59.98 ± 1.4	ns (p = 0.1811)	ns (p = 0.2341)	ns (p = 0.1529)
τc (ms) ANOVA p-value: 0.0061 ANOVA F-value: 7.669	26 ± 7.3	23.3 ± 8.8	30.7 ± 9.1	ns (p = 0.4461)	*	***
Capacitance Cm (pF) ANOVA p-value: 0.0712 ANOVA F-value: 1.939	181.9 ± 63.3	158.4 ± 23.5	165.6 ± 51.4	ns (p = 0.1398)	ns (p = 0.2216)	ns (p = 0.8496)
Sag ratio ANOVA p-value: 0.0012 ANOVA F-value: 7.272	0.2 ± 0.1	0.1 ± 0.1	0.1 ± 0.1	**	ns (p = 0.1003)	ns (p = 0.6611)
Active properties (mixed effect two-way ANOVA)
Mixed effect two-way ANOVA Interaction current x Age (p-value: 0.369; F-value:1)/ Main effect: current (p-value: < 0.0001; F-value: 57.14) Main effect: age (p-value: < 0.0001; F-value 80.66)
Firing frequency at -150 pA (Hz)	0 ± 0	0 ± 0	0 ± 0	ns (p = 0.000)	ns (p = 0.000)	ns (p = 0.000)
Firing frequency at -100 pA (Hz)	0 ± 0	0 ± 0	0 ± 0	ns (p = 0.000)	ns (p = 0.000)	ns (p = 0.000)
Firing frequency at -50 pA (Hz)	0 ± 0	0 ± 0	0 ± 0	ns (p = 0.000)	ns (p = 0.000)	ns (p = 0.000)
Firing frequency at 0 pA (Hz)	0 ± 0	0 ± 0	0 ± 0	ns (p = 0.000)	ns (p = 0.000)	ns (p = 0.000)
Firing frequency at 50 pA (Hz)	9.9 ± 10.2	11.4 ± 7.8	17.3 ± 9.7	ns (p = 0.2211)	**	*
Firing frequency at 100 pA (Hz)	18.4 ± 8.9	20 ± 7.6	27.5 ± 8.5	ns (p = 0.2462)	****	***
Firing frequency at 150 pA (Hz)	22.8 ± 8.2	26 ± 7.2	33.1 ± 7.5	ns (p = 0.0811)	****	***
Firing frequency at 200 pA (Hz)	26.2 ± 9	29.3 ± 8.5	36.4 ± 8.7	ns (p = 0.4312)	****	**
Firing frequency at 250 pA (Hz)	28.8 ± 9.7	30.5 ± 9.9	38.4 ± 9.3	ns (p = 0.9771)	****	**
Firing frequency at 300 pA (Hz)	30.6 ± 11.5	30.4 ± 12	40.7 ± 11	ns (p = 0.1937)	***	***
AP properties (one-way ANOVA)
AP amplitude (mV) ANOVA p-value: 0.4119 ANOVA F-value: 1.890	80.5 ± 14.8	74.3 ± 10.1	77.1 ± 11.8	ns (p = 0.2212)	ns (p = 0.8821)	ns (p = 0.9012)
AP threshold (mV) ANOVA p-value: 0.1182 ANOVA F-value: 1.794	− 39.1 ± 4.4	− 39.9 ± 5	− 37.5 ± 5.9	ns (p = 0.2119)	ns (p = 0.4321)	ns (p = 0.1619)
AP halfwidth (ms) ANOVA p-value: 0.3451 ANOVA F-value: 1.415	1.9 ± 1.3	2.0 ± 0.7	2.4 ± 1.5	ns (p = 0.2274)	ns (p = 0.0911)	ns (p = 0.4544)
mAHP (mV) ANOVA p-value: 0.0918 ANOVA F-value: 2.258	− 9.4 ± 4.5	− 7.79 ± 3.3	− 7.8 ± 3.3	ns (p = 0.4611)	ns (p = 0.0945)	ns (p = 0.8170)
Current-ramp (one-way ANOVA)
On-current (pA) ANOVA p-value: 0.018 ANOVA F-value: 6.741	61.2 ± 32	64.2 ± 44.5	37.3 ± 16.4	ns (p = 0.7171)	**	**
Off-current (pA) ANOVA p-value: 0.039 ANOVA F-value: 4.630	123.8 ± 84.9	132.7 ± 90.8	79.3 ± 40.5	ns (p = 0.4539)	*	*
Firing frequency ascending ramp (Hz) ANOVA p-value: 0.0012 ANOVA F-value: 5.215	15.1 ± 5.8	16.6 ± 5	19.1 ± 5.6	ns (p = 0.8902)	**	ns (p = 0.8123)
Firing frequency descending ramp (Hz) ANOVA p-value: 0.0077 ANOVA F-value: 5.613	15.7 ± 9.1	18 ± 11.3	23.3 ± 10.3	ns (p = 0.5552)	**	ns (p = 0.1311)
PIC properties (unpaired T-test)
	2 months (n = 24)	23 months (n = 24)
PIC peak current (pA) T-value: 2.500	− 1402 ± 446	− 1722 ± 443.2		*
PIC peak voltage (mV) T-value: 2.958	− 23 ± 2.3	− 27.8 ± 7.5		**
PIC onset voltage (mV) T-value: 2.496	− 65.4 ± 5.9	− 69.7 ± 6		*
PIC offset voltage (mV) T-value: 3.587	− 66.5 ± 7	− 73.9 ± 7.3		***
PIC total voltage change (mV) T-value: 3.357	131.9 ± 11.8	143.6 ± 12.3		**
PIC hysteresis (mV)	Mean onset voltage =  − 65.3 Mean offset voltage =  − 66.5 ns (p = 0.7421)	Mean onset voltage =  − 69.6 Mean offset voltage =  − 73.9 p = 0.028*		* = only 23 month old LVPNs

All data reported as group data means ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

The passive properties of LVPNs are altered with age

Next, we characterized the passive properties of LVPNs across the three age groups (Fig. 3A). Resting membrane potential (RMP) did not change across age groups (Fig. 3B and Table 1). Input resistance (R_in) (Fig. 3C and Table 1) increased by 30% in old LVPNs when compared to young LVPNs (192.5 MΩ ± 44 and 147.8 MΩ ± 26.6, p < 0.0001), but not different when compared to late adulthood LVPNs (192.5 MΩ ± 44 and 179.2 MΩ ± 38.1, p = 0.4323). Late adulthood LVPNs had a 21% increase in R_in when compared to young LVPNs (179.2 MΩ ± 38.1 and 147.8 ± 26.6 MΩ, p = 0.0008). We found that the membrane time constant (τ) (Fig. 3D and Table 1) increased significantly in old LVPNs when compared to late adulthood (30.7 ms ± 9.1 and 23.3 ms ± 8.8, p = 0.0002) and young LVPNs (30.7 ms ± 9.1 and 26 ms ± 7.3, p = 0.0412), consistent with an increase in R_in. Young and late adulthood LVPNs were not significantly different from each other. We calculated the sag ratio (Fig. 3E and Table 1), a measure of I_h, for each age group (refer to Methods 2.3). Briefly, I_h (hyperpolarization-activated current) is a voltage-dependent current, carried by the HCN channel that provides depolarizing drive at hyperpolarized membrane voltages, therefore enhancing membrane excitability [42], with a bigger ratio corresponding to a greater I_h. We found that only young and late adulthood LVPNs were significantly different from each other (0.2 ± 0.1 and 0.1 ± 0.1, p = 0.0082), while old LVPNs were not different from young LVPNs or late adulthood group. Finally, we found that capacitance (C_m) (Fig. 3F and Table 1), an indirect measure of cell surface area, was not significantly different between age groups. Overall, these results show that age has a significant effect on some of the passive properties of LVPNs.

Action potential properties of LVPNs do not change with age

We then assessed whether aging influenced the action potential properties of LVPNs. To do so, we quantified AP amplitude, AP halfwidth, AP threshold, and medium-duration afterhyperpolarization (mAHP) (Fig. 4A and Table 1). We found no significant differences for any of the AP properties measured across all age groups (Fig. 4B–E and Table 1).

Fig. 4.

Action potential properties of LVPNS generally do not change with age. A To quantify action potential properties, a 50 pA step lasting 500 ms was used and then the average values for all properties were quantified. All measures quantified are depicted in this representative action potential trace. B–E There were no significant differences across age groups for any of the AP properties (p < 0.05, one-way nested ANOVA). Error bars represent the standard deviation (SD) from the mean

Smaller rheobase and prolonged sustained repetitive firing in old LVPNs

AP properties are in part driven by sodium and potassium currents with relatively short time courses occurring in the range of milliseconds [43, 44]. Since we observed no differences in AP kinetics between young and old LVPNs, we decided to explore slower currents known to contribute to the increased excitability and action potential frequency we observed during our step protocols (Fig. 2B). Thus, we hypothesized that PICs could be a contributing factor to this phenotype. PICs are voltage-dependent inward cation currents that are long-lasting [45] and remain active even after depolarizing stimulus begins to decrease [46]. By addressing differences in ascending and descending F–I relationships and in the onset and offset of firing during slow current ramp steps, these measures provide insight into the presence and function of PICs [16, 36–39, 47] (5A). Therefore, we measured multiple parameters to address hysteresis (Fig. 5A) in the ascending and descending portions of these slow-current ramps (Fig. 5B–E and Table 1).

On the ascending portion of the current ramp, we found that old LVPNs required less current to initiate repetitive firing (smaller on-current) (Fig. 5B) when compared to young LVPNs (37.3 pA ± 16.4 and 61.2 pA ± 32, p = 0.0071) and late adulthood LVPNs (37.2 pA ± 16.4 and 64.2 pA ± 44.5, p = 0.0081). The on-currents between late adulthood and young LVPNs were not different. Next, during the descending portion of the current ramp, old LVPNs continued to fire at smaller currents (off-current: a smaller off-current corresponds to more prolonged sustained repetitive firing) (Fig. 5C) when compared to young LVPNs (79.3 pA ± 40.5 and 123.8 pA ± 84.9, p = 0.0441) and late adulthood LVPNs (79.3 pA ± 40.5 and 132.7pA ± 90.8, p = 0.0381). The off-currents between late adulthood and young LVPNs were not different. Additionally, we found that during the ascending portion of the current ramp, the firing frequency (Fig. 5D) of old LVPNs was higher compared to young LVPNs (19.1 Hz ± 5.6 and 15.1 Hz ± 5.8, p = 0.0011) but not different when compared to late adulthood LVPNs. During the ascending portion of the current ramp, young LVPNs and late adulthood LVPNs firing frequencies were not different from each other. We also found that during the descending portion of the current ramp, the firing frequency (Fig. 5E) of old LVPNs was higher compared to young LVPNs (23.3 Hz ± 10.3 and 15.7 Hz ± 9.1, p = 0.0079) but not different when compared to late adulthood LVPNs. Young LVPN and late adulthood LVPN firing frequencies were not different from each other during the descending portion of the current ramp. Overall, these results show that old LVPNs have a smaller rheobase to initiate repetitive firing and have sustained repetitive firing at a smaller input current compared with LVPNs from young and late adulthood animals. These results suggest enhanced PICs in older LVPNs.

Old LVPNs have larger PICs with altered voltage-sensitivity

Next, we wanted to directly measure the PIC in the whole-cell voltage-clamp configuration to gauge the extent of its contribution to age-related changes in LVPN repetitive firing and excitability (Fig. 6 and Table 1). Given the strong difference between young and old LVPNs from our previous experiments (Fig. 5), we used standard voltage clamp protocols (refer to Methods 2.3) [32, 33, 37–39] to measure PICs in LVPNs in young and old mice. Figure 6A shows an example of a raw PIC trace that was then isolated offline through leak subtraction (Fig. 6A and refer to methods). Figure 6B shows 2 PIC traces from old and young LVPNs: in all cases, currents were measured in 20 mM TEA-ACSF to eliminate voltage-dependent K⁺ currents that would contaminate the PIC [33, 38]. We found the PIC inward peak current was 22% larger in old LVPNs than in young LVPNs (Fig. 6C) (-1722 pA ± 443.2 and -1402 pA ± 446, p = 0.0212). The voltages at which the peak PIC occurred (Fig. 6D) were also more hyperpolarized in old LVPNs than in young LVPNs (the voltage was closer to the beginning of the ascending ramp) (− 27.8 mV ± 7.5 and − 23 mV ± 2.3, p = 0.0016). Both the onset voltage (Fig. 6E) (the voltage at which the PIC activates) (− 69.7 mV ± 6 and − 65.4 mV ± 5.9, p = 0.0411) and the offset voltage (Fig. 6F) (the voltage at which the PIC deactivates) (− 73.9 mV ± 7.3 and − 66.5 mV ± 7, p = 0.0002) were significantly more hyperpolarized in old LVPNs than in young MNs. Additionally, we calculated the PIC total voltage range (Fig. 6G) between the onset and offset of the PIC, with old LVPNs having a larger absolute voltage range when compared to young LVPNs (143.6 mV ± 12.3 and 131.9 mV ± 11.8, p = 0.0019). We also compared the onset/offset hysteresis of the PIC (Fig. 6H) for both young and old LVPNs. We found that young LVPNs had no significant difference in the PIC onset vs offset voltages (65.3 mV ± 5.9 and − 66.5 mV ± 6.9, p = 0.7421) while old LVPNs had a PIC offset that was more hyperpolarized than its PIC onset (− 73.9 mV ± 7.3 and − 69.6 mV ± 6.9, p = 0.028). These results show that PICs in old LVPNs are larger, activate at more hyperpolarized voltages, deactivate at more hyperpolarized voltages, and have a longer voltage trajectory when compared to young LVPNs.

The sodium and calcium PICs of old LVPNs are also larger and have altered voltage sensitivity

The PIC is made up of both a sodium and calcium component [37, 38, 46]. Since we found that the total PIC was increased in old LVPNs, we aimed to determine the contribution of each individual PIC component to this age-related increase: the Toal-PIC, the Na⁺-PIC, the Ca²⁺-PIC, and the resistant-PIC (Fig. 7F). To do so, we repeated the previous measurements of the PIC (Fig. 6) in a separate cohort of young and old mice.

Fig. 7.

The peak amplitude of the total PIC, sodium PIC, calcium PIC, and resistant PIC of old LVPNs is larger. The distinct PIC components were isolated pharmacologically. The peak amplitude of the total PIC (A), the sodium PIC (B), the calcium PIC (C), and the resistant PIC are all increased in old LVPNs (p < 0.0001) (D). E To quantify the relative contribution of each PIC component to the total PIC peak, we calculated the ratio of each PIC component to the total PIC. A mixed effect two-way ANOVA (p > 0.05, age x PIC component interaction; p < 0.0001, age or PIC-component simple main effect) revealed that every PIC component had an increased contribution to the total PIC in old LVPNs (p < 0.05, Tukey’s HSD). F Representative traces of the different components of the PIC that were isolated pharmacologically (refer to methods 2.3): total-PIC (purple trace), sodium PIC (blue trace), calcium PIC (gray trace), and the resistant PIC (orange trace)

We found that the total-PIC inward peak current was 43% larger in old LVPNs than in young LVPNs (Fig. 7A) (− 2577 pA ± 423.4 and − 1724 pA ± 331.1, p < 0.0001). The voltages at which the peak total-PIC occurred (Table 2) were also more hyperpolarized in LVPNs than in young MNs (− 27.5 mV ± 6.2 and − 23.8 mV ± 3.3, p = 0.0266). Both the onset voltage (Table 2) (− 92 mV ± 10.1 and − 75.8 mV ± 11.6, p < 0.0001) and the offset voltage (Table 2) (− 90.7 mV ± 7.3 and − 79.5 mV ± 8, p < 0.0001) were significantly more hyperpolarized in old LVPNs than in young LVPNs. Additionally, we calculated the Total-PIC total voltage range (Table 2) between the onset and offset of the Total-PIC, with old LVPNs having a larger absolute voltage range when compared to young LVPNs (182.7 mV ± 16.4 and 155.3 mV ± 18.5 p < 0.0001). These results confirm the previous age-related increase in the peak current and voltage sensitivity we described in Fig. 6.

Table 2.

Quantification of PIC components: Total-PIC, Na⁺-PIC, Ca²⁺-PIC, and resistant-PIC in LVPNs of young (3 months) and old (24 months) animals. All data reported as group data means ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Total PIC (nested T-test)
	3 months (n = 24)	24 months (n = 24)	Multiple comparisons
PIC peak current (pA) T-value: 4.378	− 1724 ± 331.1	− 2577 ± 423.4	****
PIC peak voltage (mV) T-value: 2.648	− 23.8 ± 3.3	− 27.5 ± 6.2	*
PIC onset voltage (mV) T-value: 5.207	− 75.8 ± 11.6	− 92 ± 10.1	****
PIC offset voltage (mV) T-value: 5.108	− 79.5 ± 8	− 90.7 ± 7.3	****
PIC total voltage change (mV) T-value: 5.485	155.3 ± 18.5	182.7 ± 16.4	****
Na+PIC (nested T-test)
	3 months (n = 24)	24 months (n = 24)	Multiple comparisons
PIC peak current (pA) T-value: 6.131	− 1157 ± 446.7	− 1957 ± 466.6	****
PIC peak voltage (mV) T-value: 1.598	− 29.3 ± 6	− 31.8 ± 5	ns (p = 0.3219)
PIC onset voltage (mV) T-value: 8.063	− 76.8 ± 10.9	− 96.6 ± 5.2	****
PIC offset voltage (mV) T-value: 9.812	− 70.1 ± 8.5	− 93.3 ± 8.1	****
PIC total voltage change (mV) T-value: 10.62	147.0 ± 16.5	189.9 ± 11.2	****
Ca2+PIC (nested T-test)
	3 months (n = 24)	24 months (n = 24)	Multiple comparisons
PIC peak current (pA) T-value: 5.792	− 461 ± 232.4	− 877.3 ± 270	****
PIC peak voltage (mV) T-value: 1.544	− 23.6 ± 2.5	− 22.5 ± 2.4	ns (p = 0.0711)
PIC onset voltage (mV) T-value: 5.539	− 64.7 ± 18.3	− 88.3 ± 10.2	****
PIC offset voltage (mV) T-value: 7.082	− 63.8 ± 11.4	− 83.1 ± 7.2	****
PIC total voltage change (mV) T-value: 6.738	128.5 ± 27.6	171.4 ± 15	****
Resistant PIC (nested T-test)
	3 months (n = 24)	24 months (n = 24)	Multiple comparisons
PIC peak current (pA) T-value: 6.126	− 172.4 ± 122.5	− 446.4 ± 185.4	****
PIC peak voltage (mV) T-value: 1.282	− 20.8 ± 1.2	− 21.4 ± 1.9	ns (p = 0.3166)
PIC onset voltage (mV) T-value: 0.8933	− 76.3 ± 10.8	− 79.8 ± 16.3	ns (p = 0.4223)
PIC offset voltage (mV) T-value: 1.743	− 75.9 ± 7.1	− 79.6 ± 7.9	ns (p = 0.0912)
PIC total voltage change (mV) T-value: 1.354	152.2 ± 13.2	159.5 ± 21.6	ns (p = 0.3887)
PIC component contribution to Total PIC peak ( mixed effect two-way ANOVA)
Mixed effect two-way ANOVA Interaction PIC component x age (p-value: 0.8167; F-value: 0.035)/ Main effect: PIC component (p-value: < 0.0001; F-value: 388) Main effect: age (p-value: < 0.0001; F-value 21.70)	3 months (n = 24)	24 months (n = 24)	Multiple comparisons
Na+ PIC/total PIC	0.7	0.8	*
Ca2+ PIC/total PIC	0.3	0.4	*
Resistant PIC/total PIC	0.1	0.2	*

We found that the Na⁺-PIC inward peak current was also 69% larger in old LVPNs than in young LVPNs (Fig. 7B) (− 1957 pA ± 466.6 and − 1157 pA ± 446.7, p < 0.0001). The voltages at which the peak Na⁺-PIC occurred (Table 2) were not significantly different between old LVPNs and young LVPNs. Both the onset voltage (Table 2) (− 96.6 mV ± 5.2 and − 76.8 mV ± 10.9, p < 0.0001) and the offset voltage (Table 2) (− 93.3 mV ± 8.1 and − 70.1 mV ± 8.5, p < 0.0001) were significantly more hyperpolarized in old LVPNs than in young LVPNs. Additionally, we calculated the Na⁺-PIC total voltage range (Table 2) between the onset and offset of the Na⁺-PIC, with old LVPNs having a larger absolute voltage range when compared to young LVPNs (189.9 mV ± 11.2 and 147 mV ± 16.5, p < 0.0001).

We found that the Ca²⁺-PIC inward peak current was 90% larger in old LVPNs than in young LVPNs (Fig. 7C) (− 877.3 pA ± 270 and − 461 pA ± 232.4, p < 0.0001). The voltages at which the peak Ca²⁺-PIC occurred (Table 2) were not significantly different between old LVPNs and young. Both the onset voltage (Table 2) (− 88.3 mV ± 10.2 and − 64.7 mV ± 19.3, p < 0.0001) and the offset voltage (Table 2) (− 83.1 mV ± 7.2 and − 63.8 mV ± 11.4, p < 0.0001) were significantly more hyperpolarized in old LVPNs than in young LVPNs. Additionally, we calculated the Ca²⁺-PIC total voltage range (Table 2) between the onset and offset of the Ca²⁺-PIC, with old LVPNs having a larger absolute voltage range when compared to young LVPNs (171.4 mV ± 15 and 128.5 mV ± 27.6, p < 0.0001).

Finally, we measured the resistant-PIC: the component of the PIC insensitive to both TTX and nimodipine [48]. We found that the resistant-PIC inward peak current was 159% larger in old LVPNs than in young LVPNs (Fig. 7D) (− 446.4 pA ± 185.4 and − 172.4 pA ± 122.5, p < 0.0001). We found no significant differences between old and young LVPNs for the voltages at which the peak resistant-PIC occurred, for the onset voltage, for the offset voltage, or for the total voltage range between the onset and offset (Table 2).

Overall, these results show that old LVPNs have larger PICs with altered voltage sensitivity, while the sodium and calcium PIC components are also larger and also have altered voltage sensitivity. However, the resistant-PIC showed only an age-dependent change in peak current. Additionally, we found that each of these PIC components changed their overall contribution to the total PIC peak (ratio of a PIC component to the total PIC peak) in old LVPNs. We performed a mixed effect two-way ANOVA to quantify the effect of age and PIC component on relative contribution to the total PIC amplitude. The two-way ANOVA revealed no significant interaction between mouse age and PIC component (p = 0.8167). However, both simple main effects, age and PIC component, had a significant effect on the relative contribution to total PIC amplitude (p < 0.0001 and p < 0.0001 respectively). We found that the contributions of the Na₊-PIC (0.8 and 0.7, p = 0.0319), Ca²⁺-PIC (0.4 and 0.3, p = 0.0226), and resistant-PIC (0.2 and 0.1, p = 0.0236) to the total-PIC peak increased significantly in old LVPNs (Fig. 7E).

Discussion

In this study, we investigated parameters of LVPN excitability across the lifespan of mice. Overall, we found that LVPNs, a neuron population critical in voluntary motor command initiation [49, 50], become more excitable in old mice. This is accompanied by two key electrophysiological changes: (1) greater membrane resistances and (2) larger PICs with a wider range of activation. Thus, we propose that increases in these parameters may represent factors associated with age-related and cell-type-specific increases in excitability.

Sarcopenia poses a significant health challenge with limited treatment options, emphasizing the need for an improved understanding of potential mechanisms [51]. The majority of research investigating mechanisms of age-related motor dysfunction and sarcopenia has focused on changes that occur at the levels of skeletal muscle and connectivity of the neuromuscular system [5], but sarcopenia is multifactorial and likely involves changes in the central nervous system including the primary motor cortex [4]. To quantify these changes, prior evidence relying on useful methods like TMS and motor evoked potentials [52] (MEPs: recordings of summated electrical potentials from muscles activated via corticospinal tracts or cortical stimulation) has found that the excitability of the primary motor cortex is itself multifactorial and nuanced. For example, both older age and weakness (in the context of aging) are associated with MEP amplitude reduction in upper limb muscles [13, 16]. Interestingly, if factors like muscle group and physical activity are also considered, differential effects on recorded MEPs are noted [16]. Despite these nuances, the excitability of the aged primary motor cortex is nevertheless susceptible to change during aging, and, when considered in the context of age-related weakness, excitability of the primary motor cortex has consistently been shown to be reduced [4, 13].

Thus, given these established, age-related changes in cortical excitability [15, 53], we aimed to characterize excitability, at the single-cell resolution, in LVPNs using whole-cell patch clamp recordings across the mouse lifespan. Our results show that LVPNs of old mice produce trains of action potentials at a higher frequency than those of young or late adulthood animals (Fig. 2B), which suggests that old LVPNs have an increased excitability. This increased excitability could be attributed to changes in the passive properties of old LVPNs such as the higher R_in, known to contribute to the increased excitability of aged neurons [54, 55]. We also did not observe differences in membrane capacitance among age groups suggesting no obvious age-related changes in cell morphology. However, we did observe that young LVPNs have a larger sag ratio than late adulthood and old LVPNs, which seemed to normalize back to the level of young animals in old age. It remains unclear how sag (a measure of I_h) may be contributing to the excitability changes we observed given that previous studies report that I_h has the potential to either decrease [56] or increase [57] excitability in neurons.

After confirming that excitability was increased in old LVPNs, we were interested in understanding the putative mechanisms of these age-related changes. We showed no age-related differences in action potential (AP) kinetics (Fig. 4), supporting that ionic currents involved in shaping AP kinetics like sodium [58] and potassium currents [59] are not necessarily responsible for the increased excitability and increased repetitive firing of old LVPNs. Thus, we directed our focus to other putative mechanisms that could be influencing changes in the repetitive firing of LVPNs. Repetitive firing in neurons can be shaped by multiple current types [42] and is a major functional characteristic of spinal motor neurons that allows motor neurons to control and maintain steady muscle contractions [60] but is also found in different neuron types within the brain such as the medulla [33], hippocampus [61], and entorhinal cortex [62]. A current critical for sustaining the repetitive firing of motor neurons is the PIC, [46, 60, 63], a depolarizing current carried by voltage-sensitive sodium and calcium channels that are located in its majority on the dendritic tree and amplifies synaptic inputs by more than fivefold [60, 64]. Importantly, PICs are voltage-dependent currents that inactivate slowly/are persistent [64, 65] and are thus able to sustain motor neuron repetitive firing crucial to maintaining sustained muscle contractility, even after voluntary drive is reduced below the level required to initially activate muscle firing [60]. In OAs, decreases in the PIC can trigger reduction in spinal motor neuron firing rates [63] and reduction in sustained muscle force production [17]. Conversely, increases in the PIC are conducive to increased repetitive firing and neuronal excitability [30, 37, 38, 66]. Thus, we hypothesized that the increased excitability of old LVPNs could be associated with changes in the PIC.

To ascertain the possible presence of the PIC in LVPNs, we first assessed differences in ascending and descending F-I relationships during slow current ramp steps (Fig. 5A). During these ramp protocols, we found that during the ascending ramp, old LVPNs have a higher firing frequency, and are activated at smaller depolarizing currents (lower rheobase), confirming the increased excitability phenotype using currents injection steps. During the descending ramp, both young and old LVPNs demonstrated persistent firing even when the input current began to decrease, with old CMMs having a higher firing frequency and a longer maintenance of persistent firing. This persistent repetitive firing during the descending portion of the ramp suggests the presence of a persistent current in both young and old LVPNs and that changes in this current during the ascending and descending ramps may contribute to the prolonged persistent repetitive firing in old LVPNs.

Given these findings in current clamp, we used slow voltage ramps to measure PICs in LVPNs and demonstrate that mouse LVPN PICs possess 3 main qualities proper to a “classic” PIC that undergo age-dependent changes. The first major feature of a PIC is its slow deactivation. In agreement with previous studies in spinal motor neurons [30, 32, 38, 39, 46] and other brain areas [33, 61, 62], our results show that as we depolarized the membrane in the ascending ramp, LVPNs exhibited a strong inward current that deviates from the baseline current and which remains persistently active even when the membrane voltage begins to repolarize in the descending ramp, suggesting the presence of PICs in both young and old LVPNs (Fig. 6B). Importantly, the amplitude of the PIC is larger in old LVPNs, consistent with current clamp data showing that the firing rate is higher on the descending part of the ramp in old animals. A second major feature that defines the PIC is its activation onset and offset hysteresis. We found that in the LVPNs of old mice, PICs possessed a widened total activation range (Fig. 6G), suggesting an age-related change in PIC hysteresis. This was confirmed by old LVPNs possessing a change in hysteresis while young LVPNs did not: the PIC offset voltage occurs at a more hyperpolarized voltage than its onset (Fig. 6H). In mammalian spinal motor neurons, for example, the PIC deactivation current occurs at more hyperpolarized voltages than the activation voltage [37, 46]. However, it is important to remark that while the hysteresis proper to spinal motor neurons has been extensively studied [46], this hysteresis modality in not shared by all neuronal populations. Other brain areas like the medulla for example possess multiple neuronal populations within the same region that present with different forms of hysteresis [33], suggesting that PIC hysteresis may be a cell-type specific (as evidenced by our data in young LVPNs; see Fig. 6H) or even an injury- and disease-type specific attribute. Finally, the PIC we observed is carried by TTX-sensitive Na⁺ channels and nimodipine-sensitive Ca²⁺ channels, as observed in other cell types [37, 38, 45, 46]. We once again measured the PIC (Total-PIC) in an additional cohort of young and old mice LVPNs (Fig. 7 and Table 2) and found that our results from Fig. 6 repeated. In this cohort of additional animals, we found that the individual Na⁺ and Ca²⁺ components of the PIC also possessed increased peak amplitudes and widened activation ranges in old LVPNs (Fig. 7 and Table 2). Interestingly, chronic spinal rats possess larger PICs with sodium and calcium PICs driving their increased sensitivity to activation and delayed deactivation [30], while Quinlan et al. [38] reported age-dependent increases in the total-, sodium-, and calcium-PIC of ALS spinal motor neurons during early development [38]. These studies suggest that factors such as injury or neurodegeneration can drive changes in the PIC and its components. In the context of these previous studies, we show here that aging is another factor that has the potential to be a major driver of changes to the PIC and its sodium and calcium components, and that these changes may contribute to the age-related changes in PIC hysteresis and excitability we report. Additionally, we found that a resistant PIC component that we were unable to eliminate pharmacologically (Fig. 7F), also had an increased peak amplitude in old LVPNs. While previously reported in spinal and medulla neurons [33, 48], its role remains unclear, and its resistant nature suggests its coexistence with both sodium and calcium components. Thus, age-dependent changes in these 3 important properties of the PIC, make it plausible that the PIC may represent a putative mechanism of increased excitability and prolonged sustained repetitive firing in LVPNs, as has been previously reported in spinal motor neurons following spinal cord injury [30, 46, 67], spinal motor neurons in ALS [38, 68, 69], CA1 hippocampal neurons in Alzheimers disease [61], entorhinal neurons in fragile-x syndrome [62], and CA1 hippocampal and layer V neurons in epilepsy [62, 70]. However, it is important to note, that space-clamp issues during whole-cell voltage clamp recordings, especially in large neurons like LVPN that possess large dendrites, prevents clamping of areas of the membrane distal from the recording electrode [71]. Thus, space-clamp has the potential to occlude the true contribution of the PIC to LVPN excitability.

The data we present here highlights whether increased excitability in old LVPNs is a loss of function during aging or, alternatively, a compensatory response to other age-related neurological changes. Hyperexcitability of old LVPNs could be occurring in response to loss of excitability at the level of spinal motor neurons as a compensatory mechanism for maintaining functional output from a hypoexcitable neuromuscular system. For example, our finding of increased excitability of LVPNs contrasts prior findings of age-related decreases of spinal motor neuron [72] and motor unit firing rates [73, 74] which could suggest that changes in cortical excitability are compensatory in nature. Moreover, if increased LVPN excitability is compensatory, increasing LVPN activity would be expected to be correlated with an increase (or maintenance) of motor activity in older animals and humans. A recent study measuring changes in oxygenated hemoglobin (HbO2) found that when compared to young individuals, fast walking required a greater increase of HbO2 in the primary motor cortex of OAs [75]. This suggests that compensatory increases in primary motor cortex excitability may impose a greater metabolic burden to sustain motor behaviors given the energetically demanding nature of neural activity. Alternatively, hyperexcitability of old LVPNs could also be occurring in response to loss of excitability at the level of the primary motor cortex. At first glance, our finding that LVPNs of the primary motor cortex have increased excitability, seemingly contradicts previous findings that the overall output of the primary motor cortex is hypoexcitable [4]. In mice specifically, quantification of age-related changes in MN number has mostly focused on the spinal cord, with studies arguing for [76] and against [77, 78] age-related changes in spinal MN number, and no clear consensus on age-related changes in brain MN number. However, in humans, evidence points to age-related loss of spinal MNs [79], with no real loss of cortical neurons [80, 81], but rather these undergo morphological changes, such as axonal degeneration [80, 82, 83] and demyelination [82] which could disrupt the faithful transmission of cortical commands to the periphery. This could lead to compensatory changes [84] such as the increased excitability of LVPNs that we report here.

Age-related hyperexcitability may also represent a maladaptive consequence of aging and could prove to be detrimental in the coordination of complex movements. Walking for example, requires synchronized activation and deactivation of agonistic and antagonistic muscle groups in rapidly repeating patterns [85] that are in turn driven by specific neuronal populations [86]. This is particularly relevant in OAs, where walking speed and gait coordination may be impaired and slow gait speed is included as diagnostic criteria for sarcopenia [87–89]. Therefore, we propose that while age-related hyperexcitability of LVPNs could be compensatory in nature, it holds the potential to disrupt the production of normal motor behaviors. Additionally, while increased excitability or “hyperexcitability” is a common phenomenon in the aged brain, other prior studies have highlighted its potential detrimental effects. In the somatosensory cortex, age-related increases in cortical receptive fields of the somatosensory cortex have been associated with hyperexcitable cortical neurons, with hyperexcitability in those neurons driven by decreased synaptic inhibition and decreases in AP threshold [54]. In senescence accelerated mice (SAM), a mouse model of accelerated cellular aging and senescence that undergoes changes such as glial inflammation and oxidative stress [90, 91], amygdala neurons present with decreased inhibitory inputs, which triggers neuronal hyperexcitability and contributes to declines in learning and memory [92]. Hyperexcitability of the hippocampus has also been reported in aged mice, specifically driven in CA3 neurons by Kv channel dysfunction, which may also contribute to memory impairment in aging [93]. In fact, using anti-epileptic drugs to decrease hippocampal hyperexcitability in Alzheimer’s disease actually improves memory performance [94]. Interestingly, hyperexcitability has also been associated with age-related decline in sleep patterns of aged mice; specifically, age-related decline of KCN channels in hypothalamic neurons that drive their hyperexcitability increases patters of fragmented sleep [95]. The results like these suggest that hyperexcitability of distinct cortical circuits may represent an emerging phenotype that could potentially represent a loss of function, leading to maladaptive patterns of behavior. Thus, given our results and these lines of evidence, it remains unclear what the functional purpose or long-term effects of LVPN hyperexcitability are on motor control and sarcopenia.

Conclusions

In this study, we present evidence that the LVPNs of old mice possess increased excitability. We propose this phenotype may arise as a result of increases in membrane R_in and enhanced function of ionic currents such as the Na⁺ and Ca²⁺ PICs. Therefore, our results introduce two intriguing possibilities, whereby increases in LVPN excitability may be compensatory for age-related loss of neuromuscular function or directly contribute to its functional decline.

Author contribution

JAV, JMS, and WDA conceived and designed the research; JAV performed the experiments; JAV analyzed the data; JAV, JMS, and WDA interpreted the results of the experiments; JAV prepared the figures; JAV drafted the manuscript; JAV, JMS, NB, and WDA edited and revised the manuscript; JAV, NB, JMS, and WDA approved the final version of the manuscript.

Funding

This project was funded by the following:

JMS: R01NS114514 and ‘Mizzou Forward’ startup.

WDA: Funding from NIA/NIH (1R01AG067758 and R01AG078129).

Data availability

Data can be available upon reasonable request to the corresponding authors.

Declarations

Competing interests

The authors declare no competing interests.