Interneuron Functional Diversity in the Mouse Accessory Olfactory Bulb

Abstract

In the mouse accessory olfactory bulb (AOB), inhibitory interneurons play an essential role in gating behaviors elicited by sensory exposure to social odors. Several morphological classes have been described, but the full complement of interneurons remains incomplete. In order to develop a more comprehensive view of interneuron function in the AOB, we performed targeted patch clamp recordings from partially overlapping subsets of genetically labeled and morphologically defined interneuron types. Gad2 (GAD65), Calb2 (calretinin), and Cort (cortistatin)-cre mouse lines were used to drive selective expression of tdTomato in AOB interneurons. Gad2 and Calb2-labeled interneurons were found in the internal, external, and glomerular (GL) layers, whereas Cort-labeled interneurons were enriched within the lateral olfactory tract (LOT) and external cellular layer (ECL). We found that external granule cells (EGCs) from all genetically labeled subpopulations possessed intrinsic functional differences that allowed them to be readily distinguished from internal granule cells (IGCs). EGCs showed stronger voltage-gated Na⁺ and non-inactivating voltage-gated K⁺ currents, decreased I_H currents, and robust excitatory synaptic input. These specific intrinsic properties did not correspond to any genetically labeled type, suggesting that transcriptional heterogeneity among EGCs and IGCs is not correlated with expression of these particular marker genes. Intrinsic heterogeneity was also seen among AOB juxtaglomerular cells (JGCs), with a major subset of Calb2-labeled JGCs exhibiting spontaneous and depolarization-evoked plateau potentials. These data identify specific physiological features of AOB interneurons types that will assist in future studies of AOB function.

Keywords: accessory olfactory bulb; cell types; excitability; interneuron.

Figure 1.

Cre-mediated genetic labeling of interneurons in all 4 AOB sublaminae. A, Fluorescence micrograph of an adult male mouse AOB from a Gad2-tdTomato double transgenic mouse (Gad2-IRES-cre± x Rosa26-loxP-STOP-loxP-tdTomato±). Boxes labeled A1–A4 correspond to the magnified inset panels below. A↑: anterior. P↓: posterior. A1–A4, Magnified views of boxes from A centered on the GL, highlighting JGCs (A1), ECL, highlighting EGCs (A2), LOT, highlighting LOTCs (A3), and ICL, highlighting IGCs (A4). Arrowheads indicate the locations of selected tdTomato-positive interneuron cell bodies. B, Fluorescence micrograph of an adult male Calb2-tdTomato double transgenic mouse. B1–B4, Magnified views of boxes indicated in B. C, Fluorescence micrograph of an adult male Cort-tdTomato double transgenic mouse. C1–C4, Magnified views of boxes indicated in C. D, Quantification of the percentage of all tdTomato-positive cells in the GL, ECL, LOT, and ICL from Gad2-tdTomato (black bars), Calb2-tdTomato (gray bars), and Cort-tdTomato mice (white bars). E, Quantification of the cell density of tdTomato-positive cells by layer and genotype. F, Quantification of the normalized cell density of tdTomato-positive cells by layer and genotype (normalization was within genotype). Scale bars in A–C: 100 µm. Scale bars in magnified panels: 50 µm; *p < 0.05 and †0.05 < p < 0.1 (Student’s unpaired, two-tailed t test).

Figure 2.

Electrophysiological properties of AOB IGCs. A, Morphologic reconstruction of an IGC dendritic arbor. B1–D1, Responses of three representative IGCs (top) to current clamp challenges (bottom). Scale bars: 10 mV, 500 ms. Inset in B1 shows the blockade of the hyperpolarization-activated depolarizing I_H sag potential by ZD7288 (10 µM). B2–D2, Responses of the same three IGCs (top) to a series of command potential steps from –70 mV in voltage clamp (bottom). Hyperpolarizing responses were also recorded, but are not shown. Scale bars: 500 pA, 100 ms. E, Spike rate input-output curves for IGCs subjected to a series of step depolarizations in current clamp. Dashed green line indicates the mean ± SE (N = 52). F, Spike accommodation index for IGCs subjected to the same series of step depolarizations shown in C (N = 52). G, Maximal I_H sag potential for 52 IGCs. H, Input-output curves for peak inward voltage-gated Na⁺ currents (N = 51). I, Input-output curves for the ratio of peak voltage-gated Na⁺ to peak voltage-gated K⁺ currents (N = 52). J, Input-output curve for normalized voltage-gated K⁺ current inactivation (N = 52). K, Representative recordings of spontaneous synaptic currents (command potential –70 mV) before and during blockade of GABA_A receptors with gabazine (2.5 µM), AMPA and NMDA receptors (10 µM AP5, 1 µM NBQX), and Type I/II mGluRs (100 µM MCPG, N = 15). L, Blockade of net spontaneous synaptic currents during pharmacological blockade (holding potential –70 mV). Asterisks indicate p < 0.05 by multiple comparisons of mean ranks, Kruskal–Wallis test. M, sEPSC frequency (N = 52). N, sEPSC amplitude (N = 52).

Figure 3.

Multidimensional analysis of IGC physiologic properties. A, Multidimensional analysis of 26 intrinsic physiologic properties was used to investigate potential relationships between different genetically defined populations of IGCs, with mitral cells included for comparison. Each column represents the physiologic profile of a single cell. Data were normalized to the 95th percentile absolute value observed for each feature across 150 AOB neurons, including IGCs, EGCs, JGCs, and mitral cells. 42/58 recorded IGCs and 12/17 mitral cells contained information for all 26 parameters and were subjected to cluster analysis. Row labels refer to the intrinsic properties listed in Table 1, and markers below each column are color coded based on genetically defined and morphologically defined type (as in B, C). Clusters are separated by the vertical lines. B, C, Nonclassical multidimensional scaling analysis of the 54 cells shown in A. Each symbol represents a single cell. Relative symbol positions reflect similarity across all 26 measurements (dimensions). In B, colored symbols indicate the genetic type of the recorded neurons; in C, colored symbols indicate the morphologic type (same color scheme as A). Dashed lines identify each cluster from A.

Figure 4.

Electrophysiological properties of AOB EGCs. A, EGC morphologic reconstructions. B1–D1, Responses of three representative EGCs to current clamp challenges. Scale bars: 10 mV, 500 ms. B2–D2, Responses of the same 3 EGCs to a series of command potential steps from –70 mV in voltage clamp. Hyperpolarizing responses were also recorded, but are not shown here. Scale bars: 1 nA, 100 ms. E, Spike rate input-output curves for EGCs. Dashed purple line indicates the mean ± SE for EGCs, and dashed green for IGCs (from Fig. 2E; N for EGCs = 33, two-way ANOVA compared to IGCs: p = 1.48 × 10⁻⁵ main effect of cell type, p = 0.0051 interaction between cell type and stimulus intensity). F, Spike accommodation index for EGCs (N = 33, two-way ANOVA compared to IGCs: p = 0.83 main effect of cell type. p = 0.113 interaction between cell type and stimulus intensity). G, Maximal I_H sag potential (N = 33, compared to IGCs p = 1.48 × 10⁻⁹ by Wilcoxon rank-sum test). H, Input-output curves for peak inward voltage-gated Na⁺ currents (N = 29, two-way ANOVA compared to IGCs: p = 0.518 main effect of cell type, p = 0.0267 interaction between cell type and stimulus intensity). I, Input-output curves for the ratio of peak voltage-gated Na⁺ to peak voltage-gated K⁺ currents (N = 29, two-way ANOVA compared to IGCs: p = 0.0048 main effect of cell type, p = 0.0916 interaction between cell type and stimulus intensity). J, Input-output curve for normalized voltage-gated K⁺ current inactivation (N = 29, two-way ANOVA compared to IGCs: p = 0.0381 main effect of cell type, p = 0.1823 interaction between cell type and stimulus intensity). K, Representative recordings of spontaneous synaptic currents (command potential –70 mV) before and during application of gabazine, AP5, NBQX, and MCPG (N = 5). L, Blockade of net spontaneous synaptic currents during pharmacological blockade (holding potential –70 mV). Asterisks indicate p < 0.05 by multiple comparisons of mean ranks, one-way ANOVA. M, sEPSC frequency (N = 29, p = 2.84 × 10⁻⁶ Wilcoxon rank-sum test). Purple bars refer to EGCs, green bars to IGCs. N, sEPSC amplitude (N = 29, p = 9.53 × 10⁻¹³ Wilcoxon rank-sum test).

Figure 5.

Multidimensional analysis of IGC and EGC physiologic properties. A, Cluster analysis of IGCs and EGCs with mitral cells included for comparison. Normalized values were calculated as per Figure 3. 29/35 recorded EGCs had information for all 26 parameters. EGC data were combined with the population of neurons in Figure 3 and re-clustered. Row labels refer to intrinsic properties listed in Table 1. Each column represents a single cell, and column labels are color coded based on genetically defined type and morphologically defined type. Clusters are separated by solid vertical lines. B, C, Nonclassical multidimensional scaling of the 83 cells shown in A, colorized by genetically defined type (B) and morphologically defined type (C). Dashed outlines indicate approximate cluster boundaries. C1–C8 refer to the cluster definitions in A.

Figure 6.

Electrophysiological properties of AOB JGCs. A, JGC morphologic reconstructions. B1–D1, Responses of three representative JGCs (top) to current clamp challenges (bottom). Scale bars: 10 mV, 500 ms. B2–D2, Responses of the same three JGCs (top) to a series of command potential steps from –70 mV in voltage clamp (bottom). Hyperpolarizing responses were also recorded, but are not shown here. Scale bars: 500 pA, 100 ms. E, Spike rate input-output curves for JGCs. Dashed red line indicates the mean ± SE for JGCs, dashed gray for IGCs, dashed black from EGCs (N = 27; two-way ANOVA compared to IGCs: p = 0.414 main effect of cell type, p = 0.850 interaction between cell type and stimulus intensity; two-way ANOVA compared to EGCs: p = 0.032 main effect of cell type, p = 0.0254 interaction between cell type and stimulus intensity). F, Spike accommodation index for JGCs (N = 27, two-way ANOVA compared to IGCs: p = 0.013 main effect of cell type. p = 0.078 interaction between cell type and stimulus intensity; two-way ANOVA compared to EGCS: p = 0.041 main effect of cell type, p = 0.007 interaction between cell type and stimulus intensity). G, Maximal I_H sag potential (N = 30, compared to IGCs p = 1.07 × 10⁻¹⁰, to EGCs 0.07 by Wilcoxon rank-sum test). H, Input-output curves for peak inward voltage-gated Na⁺ currents (N = 27, two-way ANOVA compared to IGCs: p = 0.065 main effect of cell type, p = 0.323 interaction between cell type and stimulus intensity; two-way ANOVA compared to EGCs: p = 0.346 main effect of cell type, 0.456 interaction between cell type and stimulus intensity). I, Input-output curves for the ratio of peak voltage-gated Na⁺ to peak voltage-gated K⁺ currents (N = 28, two-way ANOVA compared to IGCs: p = 0.0001 main effect of cell type, p = 0.0059 interaction between cell type and stimulus intensity; two-way ANOVA compared to EGCs: p = 0.052 main effect of cell type, p = 0.148 interaction between cell type and stimulus intensity). J, Input-output curve for normalized voltage-gated K⁺ current inactivation (N = 28, two-way ANOVA compared to IGCs: p = 0.0001 main effect of cell type, p = 0.001 interaction between cell type and stimulus intensity; two-way ANOVA compared to EGCs: p = 0.054 main effect of cell type, p = 0.0752 interaction between cell type and stimulus intensity). K, sEPSC frequency (N = 27, p = 0.21 to IGCs p = 4.77 × 10⁻⁶ to EGCs by Wilcoxon rank-sum test). L, sEPSC amplitude (N = 27, p = 2.61 × 10⁻⁸ to IGCs, p = 6.10 × 10⁻⁴ to EGCs by Wilcoxon rank-sum test).

Figure 7.

Multidimensional analysis across AOB neuronal types. A, Cluster analysis of the 118 interneurons and mitral cells for which all 26 physiologic properties were obtained. Normalized values were calculated as per Figure 3. Row labels refer to intrinsic properties listed in Table 1. Each column represents a single cell, and column labels are color coded based on genetically defined type and morphologically defined type. Clusters are separated by solid vertical lines. B, C, Nonclassical multidimensional scaling of the cell properties shown in A, colorized by genetically defined type (B) and morphologically defined type (C). Dashed outlines indicate approximate cluster boundaries. C1–C11 refer to the cluster definitions in A. D–M, Current clamp responses to the same current injection series shown in Figures 2, 4, 6 from selected cells as noted below the columns in A. Scale bars: 10 mV, 500 ms.