NEGR1 Modulates Mouse Affective Discrimination by Regulating Adult Olfactory Neurogenesis

Abstract

Background: Affective recognition and sensory processing are impaired in people with autism. However, no mouse model of autism comanifesting these symptoms is available, thereby limiting the exploration of the relationship between affective recognition and sensory processing in autism and the molecular mechanisms involved.

Methods: With Negr1 ^-/- mice, we conducted the affective state discrimination test and an odor habituation/dishabituation test. Data were analyzed using the k-means clustering method. We also employed a whole-cell patch clamp and bromodeoxyuridine incorporation assay to investigate underlying mechanisms.

Results: When encountering mice exposed to restraint stress or chronic pain, wild-type mice discriminated between them by either approaching the stressed mouse or avoiding the painful mouse, whereas Negr1 ^-/- mice showed unbiased social interactions with them. Next, we demonstrated that both wild-type and Negr1 ^-/- mice used their olfaction for social interaction in the experimental context, but Negr1 ^-/- mice showed aberrant olfactory habituation and dishabituation against social odors. In electrophysiological studies, inhibitory inputs to the mitral cells in the olfactory bulb were increased in Negr1 ^-/- mice compared with wild-type mice, and subsequently their excitability was decreased. As a potential underlying mechanism, we found that adult neurogenesis in the subventricular zone was diminished in Negr1 ^-/- mice, which resulted in decreased integration of newly generated inhibitory neurons in the olfactory bulb.

Conclusions: NEGR1 contributes to mouse affective recognition, possibly by regulating olfactory neurogenesis and subsequent olfactory sensory processing. We propose a novel neurobiological mechanism of autism-related behaviors based on disrupted adult olfactory neurogenesis.

Keywords: Adult neurogenesis; Autism olfaction; Autism sensory processing; Olfactory bulb; Subventricular zone neurogenesis.

Figure 1

Negr1^−/− mice demonstrate impaired affective discrimination for stress and pain. (A, E) Schematic designs of emotional discrimination test for stressful/neutral (WT, n = 16; Negr1^−/−, n = 15) and painful/neutral (WT, n = 17; Negr1^−/−, n = 17) demonstrators. (B, F) Left: time ratio between the right and left social areas (1-sample t test, μ = 0.5, stress/neutral: WT, p = .390; Negr1^−/−, p = .812; painful/neutral: WT, p = .882; Negr1^−/−, p = .839). Right: total distance traveled in the test chamber (stress/neutral, Student’s t test, ∗p <.05; painful/neutral, Wilcoxon signed-rank test, p = .357). (C) Left: the 6 minutes of the test with stress/neutral demonstrators is divided into 3 sessions (each of 2 minutes), the first 2 minutes (left; 2-way mixed ANOVA, interaction F_1,29 = 4.559, ∗p < .05; Bonferroni post hoc test, ∗p < .05, ∗∗p < .01, ∗∗∗p < .001), the second 2 minutes (middle; 2-way mixed ANOVA, interaction F_1,29 = 2.423, p = .130; Bonferroni post hoc test, ∗∗p < .01), and the last 2 minutes (right; 2-way mixed ANOVA, interaction F_1,29 = 0.003, p = .960). Right, the total 6 minutes (2-way mixed ANOVA, interaction F_1,29 = 2.768, ∗p < .05; Bonferroni post hoc test, ∗p < .05, ∗∗∗p < .001). The within-group differences are indicated by being color matched. (G) Left: the 6 minutes of test with painful/neutral demonstrators is divided into 3 sessions (each of 2 minutes), the first 2 minutes (left; 2-way mixed ANOVA, interaction F_1,32 = 4.032, p = .053; Bonferroni post hoc test, ∗p < .05), the second 2 minutes (middle; two-way mixed ANOVA, interaction F_1,32 = 2.321, p = .137; Bonferroni post hoc test, ∗p < .05), and the last 2 minutes (right; two-way mixed ANOVA, interaction F_1,32 = 0.190, p = .666). Right: the total of 6 minutes (2-way mixed ANOVA, interaction F_1,32 = 4.560, ∗p < .05; Bonferroni post hoc test, ∗∗p < .01). (D, H) The moving traces during the early phase (0–2 minutes) of the tests with both stress/neutral (D) and painful/neutral (H). ANOVA, analysis of variance; NS, nonsignificant; SNT, sciatic nerve transection; WT, wild-type.

Figure 2

Olfactory processing for social odors is dysregulated in Negr1^−/− mice. (A) Schematic design of the sensory-controlled social test (WT, n = 8; Negr1^−/−, n = 7). (B) Sniffing time during the sensory-controlled social test (each 2 minutes; upper: WT, bottom: Negr1^−/−). Left: the first 2 minutes (2-way RM ANOVA; WT, interaction F_2,14 = 27.6, p < .0001; Negr1^−/−, interaction F_1.07,6.42 = 16.2, p = .006; Bonferroni post hoc test, ∗p < .05, ∗∗p < .01, ∗∗∗p < .001). Middle: the second 2 minutes (2-way RM ANOVA; WT, interaction F_1.03,7.19 = 3.70, p = .094; Negr1^−/−, interaction F_1.04,6.42 = 2.61, p = .156). Right: the last 2 minutes (2-way RM ANOVA; WT, interaction F_1.03,7.18 = 2.18, p = .183; Negr1^−/−, interaction F_1.09,6.52 = 1.88, p = .218). (C) Sniffing time for the olfaction-only condition for the comparison between WT and Negr1^−/−. Top: 0 to 2 minutes (2-way mixed ANOVA, interaction F_1,13 = 0.315, p = .584; Bonferroni post hoc test, ∗∗p < .01, ∗∗∗p < .001). Bottom: total 6 minutes (2-way mixed ANOVA, interaction F_1,13 = 0.381, p = .548; Bonferroni post hoc test, ∗∗p < .01, ∗∗∗p < .001). (D) The schematic design of the experimental schedule (WT, n = 7; Negr1^−/−, n = 6). (E) Raw sniffing time during the odor habituation-dishabituation test (2-way mixed ANOVA, interaction F_8,88 = 5.318, p < .0001, Bonferroni post hoc test, ∗p < .05, ∗∗∗p < .001). (F) Left: olfactory habituation index for social odors (2-way mixed ANOVA, interaction F_2,22 = 8.603, p = .008; Bonferroni post hoc test, ∗∗∗p < .001). Right: olfactory dishabituation index for social odors (Welch’s t test, ∗∗p = .004). (G, H) The ADT in the early stage and total duration. Left: sniffing ratio for the stress demonstrator (Student’s t test, early, p = .092; total, ∗p = .034). Right: scatter plot with olfactory dishabituation index on the x-axis and sniffing ratio on the y-axis (Pearson correlation, early vs. sniff, r = 0.526, p = .065; total vs. sniff, r = 0.512, p = .037). (I) The input behavioral measures used for k-means clustering. (J) The scatter plot representing clusters on the x-y plane of the reduced dimensionality by the principal component analysis (Dim1 = 54.3%, Dim2 = 22.3%). Enlarged plots indicate the computed means of each group. ADT, affective state discrimination test; Dim, dimension; NS, nonsignificant; RM ANOVA, repeated-measures analysis of variance; WT, wild-type.

Figure 3

Negr1^−/− mice fail to associate a novel odor with a reward. (A) The schematic design of the odor discrimination learning test (WT, n = 8; Negr1^−/−, n = 8). (B) Digging time for the area under which odors are buried. Top: the first-day test for odor discrimination (2-way mixed ANOVA, interaction F_1,14 = 5.740, p = .031; Bonferroni post hoc test, ∗∗p < .01). Middle: the second-day test for reward-odor association (2-way mixed ANOVA, interaction F_1,14 = 7.309, p = .017; Bonferroni post hoc test, ∗p < .05). Bottom, the last-day test with the neutral odor (2-way mixed ANOVA, interaction F_1,14 = 0.043, p = .839). The within-group differences are indicated by being color matched. ANOVA, analysis of variance; NS, nonsignificant; WT, wild-type.

Figure 4

Inhibitory inputs are increased in the mitral cells of Negr1^−/− mice with a decrease in excitability. (A) The schematic design of the whole-cell recording experiment. (B) The amplitude and frequency of sEPSCs (24 cells from 4 WT mice and 20 cells from 3 Negr1^−/− mice; Student’s t test, p = .508 for amplitude and p = .647 for frequency). (C) The amplitude and frequency of sIPSCs (22 cells from 4 WT mice and 20 cells from 3 Negr1^−/− mice; Student’s t test, p = .452 for amplitude and p = .00088 for frequency). (D) Representative voltage traces and quantified graphs of excitability (22 cells from 4 WT mice and 20 cells from 3 Negr1^−/− mice; 2-way mixed analysis of variance, F_9,288 = 2.526, p = .008, Bonferroni post hoc test, ∗p < .05). EPL, external plexiform layer; GCL, granule cell layer; GL, glomerular layer; IEI, interevent interval; MCL, mitral cell layer; NS, nonsignificant; sEPSC, spontaneous excitatory postsynaptic current; sIPSC, spontaneous inhibitory postsynaptic current; WT, wild-type.

Figure 5

The number of neural progenitor cells in the SVZ and newly generated olfactory neurons in the GL are reduced in Negr1^−/− mice. (A) Along the rostral migratory pathway, anterior SVZ, posterior SVZ, and OB were chosen (WT, n = 8; Negr1^−/−, n = 7). Dashed lines on the brain sagittal image indicate collected sections, and dashed line squares show regions of interest. (B) Immunostaining showing Ki67 (orange) in the posterior SVZ (top; scale bar = 500 μm, 100 μm), anterior SVZ (middle; scale bar = 500 μm, 100 μm), and OB (bottom; scale bar = 200 μm, 50 μm) and the number of Ki67⁺ cells in the posterior SVZ (top; Student’s t test, ∗∗p = .005), anterior SVZ (middle; Student’s t test, p = .227) and OB (bottom; Student’s t test, p = .915). (C) The schematic design of the BrdU incorporation assay. (D) The whole image of the OB delineating GCL and GL. Scale bar = 500 μm. (E) Representative immunohistochemistry images showing BrdU (red), DAPI (blue), and GAD67 (orange). Scale bars = 50 μm. (F, G) The number of GAD67⁺ and/or BrdU⁺ cells in the GL (F) and GCL (G). In panel (F) (GL; WT, n = 5; Negr1^−/−, n = 5), GAD67 and BrdU double-positive cells (left; Student’s t test, ∗p = .421), GAD67-negative but BrdU-positive cells (middle; Student’s t test, p = .963), and total number of BrdU-positive cells (right; Student’s t test, ∗p = .048). In panel (G) (GCL; WT, n = 5; Negr1^−/−, n = 5), GAD67 and BrdU double-positive cells (left; Student’s t test, p = .222), GAD67-negative but BrdU-positive cells (middle; Student’s t test, p = .243), and total number of BrdU-positive cells (right; Student’s t test, p = .605). AP, anteroposterior; BrdU, bromodeoxyuridine; GCL, granule cell layer; GL, glomerular layer; OB, olfactory bulb; SCZ, subventricular zone; WT, wild-type.