Oligophrenin-1 regulates number, morphology and synaptic properties of adult-born inhibitory interneurons in the olfactory bulb

Abstract

Among the X-linked genes associated with intellectual disability, Oligophrenin-1 (OPHN1) encodes for a Rho GTPase-activating protein, a key regulator of several developmental processes, such as dendrite and spine formation and synaptic activity. Inhibitory interneurons play a key role in the development and function of neuronal circuits. Whether a mutation of OPHN1 can affect morphology and synaptic properties of inhibitory interneurons remains poorly understood. To address these open questions, we studied in a well-established mouse model of X-linked intellectual disability, i.e. a line of mice carrying a null mutation of OPHN1, the development and function of adult generated inhibitory interneurons in the olfactory bulb. Combining quantitative morphological analysis and electrophysiological recordings we found that the adult generated inhibitory interneurons were dramatically reduced in number and exhibited a higher proportion of filopodia-like spines, with the consequences on their synaptic function, in OPHN1 ko mice. Furthermore, we found that olfactory behaviour was perturbed in OPHN1 ko mice. Chronic treatment with a Rho kinase inhibitor rescued most of the defects of the newly generated neurons. Altogether, our data indicated that OPHN1 plays a key role in regulating the number, morphology and function of adult-born inhibitory interneurons and contributed to identify potential therapeutic targets.

Introduction

Intellectual disability (ID) is a neurodevelopmental disorder characterized by impaired cognitive performance. Among the X-linked genes associated with ID, Oligophrenin-1 (OPHN1) encodes for a Rho GTPase activating protein (Rho GAP) (1), that is expressed in areas with a high degree of plasticity such as the cortex, the hippocampus and the olfactory bulb (2). As a Rho GAP, OPHN1 modulates Rho GTPases, a family of proteins that regulate actin cytoskeleton dynamic, a process essential to orchestrate cell morphology, synaptic connectivity and gene expression (3,4). Inhibitory interneurons play a critical role in circuit formation and function (5,6). The impact of OPHN1 mutation on inhibitory interneurons remains unknown. To address this open question, we investigated, in a well-established mouse model of X-linked ID, OPHN1 ko (i.e. OPHN1^-/y) mice (7), the development of adult generated inhibitory interneurons in the olfactory system, where OPHN1 is highly expressed (2).

The olfactory system (OS) is one of the two unique neurogenic niches of the adult mammalian brain, along with the hippocampus (8–10). The olfactory bulb (OB) receives every day thousands of neuronal precursors that undergo a series of morphological and functional changes to become mature interneurons (i.e. granule cells (GC), and periglomerular cells) and integrate into functional circuits (11). Adult neurogenesis in the OS is therefore a useful system to investigate mechanisms underlying neurodevelopmental disorder, not only because it retains developmental properties but also because it exhibits a unique form of network plasticity.

Here we focused on adult-born GC. Briefly, new cells generated in the subventricular zone (SVZ) migrate to the OB along the rostral migratory stream (RMS). By day 9, most of the newly generated cells have reached the OB and by day 14 they begin to form synapses and integrate with pre-existing circuits (12,13). Between 15 and 50 days, the number of newly generated cells reduced by one half. In the OB, the adult-born cells undergo morphological maturation consisting in the elaboration of dendrites and spines, to become mature interneurons (14).

Using a cell division marker, 5-Bromo-2′-deoxyuridine (BrdU), we found that the number of adult-born interneurons in the OB was dramatically reduced in OPHN1^-/y mice. Employing lentiviral vectors expressing green fluorescent protein (GFP) to label the neuronal precursors in the SVZ, we reported that newly generated GC exhibited a higher proportion and density of filopodia-like spines in OPHN1^-/y mice. Electrophysiological recordings showed that some features of inhibitory inputs, IPSC, on the postsynaptic cells (i.e. the mitral cells (MC)) of GC, were increased in OPHN1^-/y mice. Furthermore, we found that olfactory behaviour was hampered in OPHN1^-/y mice. Most of the abnormalities of the adult-born GC and the functional properties of the OB circuitry were successfully reverted or improved by chronic treatment with the clinically approved Rho kinase (ROCK) inhibitor fasudil. Data obtained in the present work showed that OPHN1 has a key role in regulating morphology and function of GABAergic inhibitory interneurons in the OB, whose alterations could contribute to the pathophysiology of the disease.

Results

Number of adult-born GC is reduced in OPHN1^-/y mice

New cells are constantly generated in the SVZ. To investigate whether mutation in OPHN1 could regulate cell proliferation, we quantified the number of newly generated cells in the SVZ, 24 hours after injection of BrdU, a cell division marker (Fig. 1A and F). We found no difference between OPHN1^-/y mice and controls (control mice n = 3; OPHN1^-/y mice n = 3 P = 0.9).

Figure 1.

Adult-born cells in the subventricular zone and in the olfactory bulb of control and OPHN1^-/y mice. (A) Examples of coronal sections of the subventricular zone (SVZ) labelled with antibodies against BrdU (red) 24 hours after injections, in control (CTRL; left) and OPHN1^-/y (right) mice. Bar = 200 µm. (B-C) Images of coronal sections of the granule cells layer of the olfactory bulb (OB), double labelled with antibodies against BrdU (red) and the neuronal marker NeuN (green) 15 days post injection (15 dpi) in control (B) and OPHN1^-/y (C) mice. Bar = 200 µm. (D-E) Higher magnification of the dashed rectangular area in (B) and (C), respectively. Bar = 50 µm. (F) Number of newborn cells in the SVZ 24 hour after the BrdU injections. Bar = SEM. (G) Number and (H) percentage of double labelled cells (BrdU and NeuN) in the OB 15 dpi. Bar = SEM. (I) Number and (J), percentage of double labelled cells (BrdU and NeuN) in the OB 9 dpi. Bar = SEM. (K-L) Number of double labelled cells (BrdU and NeuN) in the OB 15 and 50 dpi in control (K) and OPHN1^-/y (L) mice. Adult-born GC reduced to ∼ one half at 50dpi, both in controls and OPHN1^-/y mice. (M-N) Number (M) and percentage (N) of double labelled cells (BrdU and NeuN) in the OB 50 dpi. Bar = SEM.

From the SVZ, progenitors migrate along the RMS to reach their final destination, the olfactory bulb (OB). The number of neuronal precursors, i.e. cells double positive for BrdU and the neuronal marker NeuN, was dramatically reduced (about to one third) in the OB of OPHN1^-/y mice respect to controls, 15 day-post injection (dpi) (Fig. 1, B-E and G; control mice n = 3; cell n = 2468 ± 243; OPHN1^-/y mice n = 3, cell n = 741.6 ± 164; P = 0.003**). Cell fate was however not affected, since the proportion of cells double labelled for BrdU and for NeuN (Fig.1H) was similar in OPHN1^-/y mice and controls (BrdU + NeuN/BrdU_tot= 80.2% in OPHN1^-/y mice and 81% in controls).

To assess whether the reduced number of interneurons could be ascribed to the altered integration of the new cells into OB circuitry, we counted the adult generated cells at 9dpi, prior to synaptic contact formation (13,15). We found that also at this time point, the number of adult-born cells in the OB was again significantly reduced and not significantly different from the number of progenitors counted at 15 dpi, in OPHN1^-/y mice respect to controls (Fig. 1I, BrdU +NeuN positive cells 9 dpi, control cell n = 2406.67 ± 171.96, OPHN1^-/y cell n = 734 ± 23.7, P < 0.001***, BrdU +Neu N positive cells 9 dpi versus 15 dpi, control P = 0.84, OPHN1^-/yP = 0.95). Cell fate remained unaffected in OPHN1^-/y mice at 9 dpi, as already observed at 15 dpi (Fig. 1J, BrdU+ NeuN/BrdU tot cells control = 80%; OPHN1^-/y = 80%). To explore the destiny of the missing cells, sagittal sections of the brain were immunolabelled with an antibody against cleaved Caspase 3 (CC3), a marker of apoptosis (Supplementary Material, Fig. S1A–H). The number of adult-born cells in apoptosis (i.e. double positive for CC3 and double cortin, DCX, a marker of neuronal precursors) was significantly higher in the SVZ and along the RMS in the OPHN1^-/y mice than in controls (Supplementary Material, S1I, SVZ control cell n = 14.7 ± 4.33, OPHN1^-/y cell n = 69.3 ± 2.9; P < 0.001***; RMS control cell n = 46.3, OPHN1^-/y cell n = 144.7 ± 12.3 P = 0.002**). To investigate whether cell death hit preferentially a cell type, i.e. neuronal precursors versus non neuronal cells, we estimated the proportion of cells double positive for CC3 and DCX. We found that a higher proportion of neuronal precursors went into apoptosis before reaching the OB in OPHN1^-/y mice than in controls (Supplementary Material, Fig. S1J; CC3+ DCX/CC3+ tot (%), SVZ, controls, 34 ± 1.7; OPHN1^-/y mice = 54 ± 0.6, P < 0.001***; RMS, controls = 49 ± 1.7, OPHN1^-/y mice= 64 ± 1.1, P < 0.003**). Noteworthy we reported a reduction to about 1/3 of the newly generated GC in the OB, 15 dpi and 9 dpi, in OPHN1^-/y mice. Accordingly, we observed an increase of about 2/3 of apoptotic neuronal precursors in OPHN1^-/y mice.

The question then arose on the survival of the adult-born GC in the OB. By counting cells double positive for BrdU and NeuN in the OB 50 dpi, we discovered that adult-born cells were reduced to one half in OPHN1^-/y mice as in controls (Fig. 1K-L, control mice n = 3, cell n = 1206 ± 57.8; OPHN1^-/y mice n = 4, cell n = 286.8 ± 106.8). The total number of newly generated GC at 50dpi was, however, significantly lower in the OB of OPHN1^-/y mice than in controls (P < 0.001***), since only one third of cells reached the OB (Fig. 1M). The proportion of neuronal cells that survived at 50dpi was similar in controls and OPHN1^-/y (Fig. 1N, BrdU + NeuN/BrdUtot %, controls = 90.7, OPHN1^-/y= 84.3).

Altered morphology of adult-born GC in OPHN1^-/y mice

Neuronal precursors were labelled in the SVZ with lentiviral vectors expressing GFP and the morphology of the adult-born GC was analysed 30 dpi in the OB (Fig. 2A–E). At this time point newly generated GC exhibit the morphology of mature interneurons. GC are axonless inhibitory interneurons with an apical dendrite and basal dendrites. The apical dendrite has been subdivided in an unbranched portion (the proximal domain) emerging from the soma, followed by a branched segment (distal dendrites, see scheme in Fig. 2A). The basal and the proximal dendritic domains receive synaptic inputs from axon collaterals of the MC (i.e. OB projecting neurons) and from fibres from the cortex (16,17). The distal dendrites form dendro-dendritic synapses with the lateral dendrites of the MC (18,19). We found that the length of the dendritic tree and the number of branching points were not statistically different in OPHN1^-/y mice and in controls (dendritic length, in basal domain, controls versus OPHN1^-/y mice P = 0.61; in apical domain, P = 0.87; in distal domain P = 0.42. Branching points, in controls versus OPHN1^-/y mice P = 0.43; Fig. 2F–I). However, analysis of dendritic spine morphology revealed a statistically significant increase in filopodia in the basal and, even stronger, in the distal dendritic domains in OPHN1^-/y mice respect to controls. This increase in filopodia resulted in a higher number of total spine density and in a higher proportion of filopodia-like spines respect to mushroom-stubby spines in the distal and in the basal domain in OPHN1^-/y mice than in controls (Supplementary Material, Fig. S2 and Fig. 2J–L distal domain: control spines (%) = 50.5 ± 2.4; filopodia (%) = 49.5 ± 2.5; OPHN1^-/y spines (%) = 36.8 ± 1.8; filopodia (%) = 63.2 ± 1.8; P < 0.001***; basal domain: control spines (%) = 53 ± 3.1; filopodia (%) = 47 ± 3.1; OPHN1^-/y spines (%) = 44 ± 2; filopodia (%) = 56 ± 2; P < 0.001***, mice n = 3 per condition).

Figure 2.

Morphology of adult-born granule cells. (A) Schematic representation of the different dendritic domains of granule cells: basal domain, (= basal dendrites); proximal domain, (=15% of the apical dendrite); distal domain, (= distal dendrites). (B,C) Examples of GFP- labelled newborn cells in control (CTRL, B) and OPHN1^-/y (C) mice, bar 200 = µm. (D-E) Higher magnification of the dashed rectangular area in (B) and (C), respectively. Bar = 10 µm. (F–L) summary of results related to: dendritic length in the basal (F), apical (G), distal (H) dendritic domains; number of branching points (I) and spine morphology presented as ratio between spines (s) and filopodia (f) in the basal (J), proximal (K) distal (L) dendritic domains in control and OPHN1^-/y mice. s = mushroom and stubby spines, f = filopodia and thin spines. Bar = SEM.

To investigate whether filopodia like spines in OPHN1^-/y mice established functional synaptic contacts, we analysed the distribution of glutamatergic inputs onto GC by estimating PSD-95, a scaffold protein of the post synaptic density of glutamatergic synapses (Fig. 3A–D). We found that in the basal and in the distal domains, the density of PSD-95 was higher in filopodia-like spines of OPHN1^-/y mice than in controls (Fig. 3E–G, basal domain, PSD-95, control spines = 0.13 ± 0.013; OPHN1^-/y spines = 0.13 ± 0.025, P = 0.7; control filopodia = 0.068 ± 0.01; OPHN1^-/y filopodia = 0.12 ± 0.02, P = 0.018*; control shaft = 0.17 ± 0.02, OPHN1^-/y shaft = 0.16 ± 0.027, P = 0.7; distal domain, control spines = 0.09 ± 0.006; OPHN1^-/y spines = 0.09 ± 0.007; P = 0.09; control filopodia = 0.09 ± 0.008; OPHN1^-/y filopodia = 0.13 ± 0.009, P = 0.01*; control shaft = 0.24 ± 0.02; OPHN1^-/y shaft = 0.23 ± 0.02, P = 0.6; proximal domain = control spines = 0.11± 0.014; OPHN1^-/y spines = 0.09 ± 0.02, p= 0.48; control filopodia = 0.098 ± 0.013, OPHN1^-/y filopodia = 0.095 ± 0.019 P = 0.9; control shaft =  0.28 ± 0.037, OPHN1^-/y shaft = 0.4 ± 0.07, P = 0.12). In a similar way, filopodia-like -spines of OPHN1^-/y mice expressed a higher density of the vesicular neurotransmitter transporter of GABA, vGAT in the distal domain (i.e. the only dendritic domain of GC that release GABA, Fig. 3H–L, vGAT spines control = 0.115 ± 0.001; spines OPHN1^-/y = 0.10 ± 0.015, P =0.47; shaft control =  0.14 ± 0.015; shaft OPHN1^-/y = 0.13 ± 0.017, P = 0.57; filopodia control = 0.086 ± 0.008; OPHN1^-/y = 0.13 ± 0.02, P = 0.03*). To analyse the density of GABA_A receptors on MC, immunostaining against the α1 subunit of GABA_A receptors was performed. We found no difference between control and OPHN1^-/y mice (GABA_A receptors in controls versus OPHN1^-/y mice P = 0.91; Supplementary Material, Fig. S3).

Figure 3.

Expression of PSD-95 and vGAT in adult generated cells. (A,B) PSD-95 positive clusters (PSD ⁺ Cs) in the dendritic domains of adult-born granule cells in control (CTRL, A) and OPHN1^-/y mice (B). Bar = 20 µm. (C,D) Higher magnification of the boxed areas in A and B, respectively. Arrowheads = example of PSD-95 cluster in granule cells. Bar = 2 µm. (E–G) Summary of results. Bar = SEM. (H,I) vGAT puncta in the distal dendritic domain of adult-born granule cells in control (H) and OPHN1^-/y mice (I). Bar = 20 µm. (J,K) Higher magnification of boxed area in (H) and (I), respectively. Arrowheads = example of vGAT puncta in granule cells. Bar = 2 µm. (L) Summary of results. Bar = SEM. Spines (m + s) = mushroom and stubby spines, filopodia (f + t) = filopodia and thin spines.

Functional consequences of altered newly generated GC in OPHN1^-/y mice

We then analysed the functional consequences of altered morphology of adult-born GC, by electrophysiological recordings in their target neurons, the MC (Fig. 4). Distal dendrites of GC establish dendro-dendritic synapses with lateral dendrites of MC, consisting of an excitatory input from MC to GC adjacent to an inhibitory synapse from GC to MC (18,19). As shown in Fig. 4A–C, MC exhibited a single apical dendrite (proper of mature MC, (20) in control and OPHN1^-/y mice (mice n = 3, per control and OPHN1^-/y mice). Miniature IPSCs (mIPSCs) recorded from MC exhibited larger amplitude and higher frequency in OPHN1^-/y than in control mice (Fig. 4D–F, amplitude in control = 28.8 ± 4.8 pA amplitude in OPHN1^-/y = 53.0 ± 9.6 pA, control cell n = 15, OPHN1^-/y cells n = 14, P = 0.03*; frequency in control = 0.8 ± 0.1 Hz frequency in OPHN1^-/y =1.9 ± 0.5 Hz control cells =15, OPHN1^-/y cells =14 P =0.04*). Also the decay time and the area of mIPSCs were bigger in OPHN1^-/y mice than in controls (Fig. 4F; decay time control = 20.4 ± 2.3 ms; decay time in OPHN1^-/y = 35.0 ± 6.8 ms; control cell n = 15; OPHN1^-/y cell n = 14 P = 0.04*; area in control = 413.5 ± 84.4 pA*ms: area in OPHN1^-/y = 871.5 ± 202.4 pA*ms; control cells n =15; OPHN1^-/y cells n = 14; P = 0.03*). The amplitude and the area of spontaneous IPSCs (sIPSCs), recorded from MC, were larger in OPHN1^-/y than in control mice (Fig. 4G–I; amplitude in control = 44.5 ± 5.9 pA amplitude in OPHN1^-/y = 69.7 ± 8.2 pA, control cell n =10, OPHN1^-/y cells n =17, P = 0.04*; area in control = 438.3 ± 83.8 pA*ms; area in OPHN1^-/y = 738.0 ± 93 pA*ms; control cells n = 10; OPHN1^-/y cells n = 17 P = 0.04 *). However the frequency and the decay time of sIPSC were not significantly different in OPHN1^-/y and control mice (Fig. 4I, frequency control = 2.48 ± 0.49, n = 8; OPHN1^-/y= 3.27 ± 0.42, n = 15; decay time control =17.3 ± 2.4 ms, cell n =11, OPHN1^-/y = 15.6 ± 1.3 ms, cell n = 17).

Figure 4.

Electrophysiological recording from mitral cells. (A) Schematic of the dendro-dentritic synapses between MC and granule cells (GC). (B) Examples of adult mitral cells (MC) labelled with DiI, in control (CTRL, left) and OPHN1^-/y mice (right). Mature MC exhibit a single apical dendrite (arrows) in OPHN1^-/y as in control mice. Bar = 120 μm. (C) Percentage of MC with single (s, white bar) or double (d, black bar) apical dendrite. (D) Representative traces of mIPSCs recordings from MC in control (top) and OPHN1^-/y mice (bottom). (E) Mean of a population of mIPSCs in control and OPHN1^-/y mice. (F) Histograms showing basal properties quantification of mIPSCs recorded from MC in control (white bar) and OPHN1^-/y (black bar) mice. (G) Representative traces of sIPSCs recordings from MC in control (top) and OPHN1^-/y (bottom) mice. (H) Mean of a population of sIPSCs in control and OPHN1^-/y mice. (I) Histograms showing basal properties quantification of sIPSCs recorded from MC in control (white bar) and OPHN1^-/y (black bar) mice. Bar = SEM.

Pharmacological rescue of altered morphological and functional features of new GC in OPHN1^-/y mice

Loss-of-function mutation of OPHN1 (see OPHN1 expression in Supplementary Material, Fig. S5) is thought to lead to overactivation of the ROCK pathway (Supplementary Material, Fig. S5). We reasoned that pharmacological inhibition of ROCK could therefore counteract the loss of function of OPHN1 and rescue the neuronal abnormalities triggered by the mutation. OPHN1^-/y mice were treated, for three weeks, with fasudil, a clinically approved kinase inhibitor, namely an inhibitor of protein kinase A (PKA) and ROCK, that exhibits a significant more potent inhibition on ROCK (21) .We found that chronic treatment with fasudil (see methods for details) successfully rescued the number of newly generated cells 15 dpi in OPHN1^-/y mice (Fig. 5, control mice n = 3, cell n = 2468 ± 243; OPHN1^-/y mice n= 3, cell n = 2345 ± 446 P = 0.9). Fasudil treatment did not affect the number of cells in the OB, 15dpi in control mice (Fig. 5). Accordingly, we found that upon fasudil treatment, the overactivation of ROCK target, MYPT1, was reduced in OPHN1^-/y mice and in controls (see Supplementary Material, Fig S5).

Figure 5.

Treatment with Rho Kinase (ROCK) inhibitor fasudil restored the number of adult-born granule cells in the olfactory bulb. (A,B) Images of coronal sections of the granule cells layer of the olfactory bulb, double labelled with antibodies against BrdU (red) and the neuronal marker NeuN (green) 15 dpi in control (CTRL, A) and OPHN1^-/y+fasudil (B) mice. Bar = 200 µm. (C,D) Number (C) and percentage (D) of adult-born granule cells (BrdU and NeuN positive cells) in the olfactory bulb 15 dpi in controls (white bar), OPHN1^-/y mice (black bar), in controls + fasudil (light grey bar) and in OPHN1^-/y mice + fasudil (dark grey bar). Bar = SEM.

To explore whether treatment with the ROCK inhibitor fasudil was able to revert also the alterations in spine morphology, mice were injected in the SVZ with lentiviral vectors expressing GFP to label the newly generated cells and, at the same time, treated with fasudil for 30 days (see methods for details). By analysing the morphology of the new GC 30 dpi (Fig. 6A–D), we found that the length of dendrites, that was not perturbed by OPHN1 mutation (see above), was not affected by treatment with fasudil in control and OPHN1^-/y mice (Fig. 6E–G).The proportion between filopodia-like spines and mushroom – stubby spines was successfully reverted in OPHN1^-/y treated with fasudil (n = 3 per condition, Fig. 6 H–J, basal domain, control spines (%) = 53 ± 3.1; filopodia (%) = 47 ± 3.1; OPHN1^-/y+fasudil, spines (%) = 52.3 ± 1.7; filopodia (%) = 47.6 ±.1.7; P = 0.07; distal domain, controls spines (%) = 50.5 ± 2.4, filopodia (%) = 49.5 ± 2.5; OPHN1^-/y+fasudil, spines (%) = 50.1 ± 1.78; filopodia (%) = 49.9 ± 1.78, P = 0.95). The number of spines, in particular of mature spine (i.e. mushroom-stubby) was, however, increased both in control and OPHN1^-/y mice treated with fasudil (Supplementary Material, Fig. S4).

Figure 6.

Treatment with Rho Kinase (ROCK) inhibitor fasudil rescued the proportion of spines (s)/filopodia (f) in adult-born granule cells. (A,B) Examples of GFP- labelled adult-born granule cells (GC) in control (CTRL, A) and OPHN1^-/y+fasudil (B) mice. Bar = 200 µm. (C,D) Higher magnification of the dashed rectangular area in A and B respectively. Bar = 10 µm. (E–G) Dendritic length in the basal (E), apical (F) and distal dendritic domain (G) of GC in controls (CTRL, white bar), in OPHN1^-/y mice (black bar), in controls treated with fasudil (light grey bar), in OPHN1^-/y mice treated with fasudil (dark grey bar). (H–J) Proportion of spine-filopodia (%) in the basal (H), proximal (I) and distal (J) dendritic domains in control, OPHN1^-/y and OPHN1^-/y+fasudil mice. S = mushroom and stubby spines, f = filopodia and thin spines. Bar = SEM.

Finally, to investigate whether also the functional properties of neuronal circuits in the OB of OPHN1^-/y mice could be reverted, we choose to record mIPSC, the most affected parameter, in MC in OPHN1^-/y mice treated with fasudil and in control mice. Fasudil was administered to mice for 30 days, the same time period used to ascertain the effect of the inhibitor on the morphology of adult-generated GC (see above). Electrophysiological recordings in MC revealed that the basal properties of mIPSC, i.e. the amplitude, the area and the decay time, were not significantly different in chronically treated OPHN1^-/y mice respect to controls (Fig. 7; amplitude in control = 29.02 ± 3.5 pA; amplitude in OPHN1^-/y treated =  25.6 ± 2.0 pA; control cells n = 23; OPHN1^-/y treated cells n = 11 P = 0.52; area in control = 438.9 ± 64.2 pA*ms; area in OPHN1^-/y treated = 302.4 ± 35.1 pA*ms control cells n = 23; OPHN1^-/y treated cells n = 11 P = 0.18; decay time in control = 21.4 ± 1.9 ms; decay time in OPHN1^-/y treated = 21.2 ± 0.9 ms; control cells n = 23; OPHN1^-/y treated cells n =11 P = 0.93). The frequency of mIPSC remained, however, higher in OPHN1^-/y than in control mice (Fig. 7C, frequency in control = 0.8 ± 0.1 Hz frequency in OPHN1^-/y = 2.5 ± 0.6 Hz control cells =23; OPHN1^-/y cells =11 P = 0.001***).

Figure 7.

Rescue of mIPSC in OPHN1^-/y mice treated with Rho kinase (ROCK) inhibitor fasudil. (A) representative traces of mIPSCs recorded from mitral cells (MCs) in controls (CTRL, top), in OPHN1^-/y mice (middle) and OPHN1^-/y mice + fasudil (bottom). (B) Mean of a population of mIPSCs in control, OPHN1^-/y and OPHN1^-/y+fasudil mice. (C) Histograms showing basal properties quantification of mIPSCs recorded from MCs in control (white bar), OPHN1^-/y (black bar), and OPHN1^-/y+fasudil (grey bar) mice. Bar = SEM.

To explore the functional outcome of the altered number and morphology of adult-born GC, we performed an olfactory behaviour test, namely the habituation-dishabituation test. Briefly, mice were exposed to filter paper with double distilled water, for three consecutive times, for 3 min each, with 1 min inter-trial interval. The fourth time, the mice were exposed to filter paper scented with Eugenol (1 mM) (Fig. 8A). Control mice exhibit a progressive reduction of sniffing time in the first three (habituation) trials. A significant increase in the sniffing time is observed in the fourth trial (dishabituation trial, P = 0.03*, mice n = 11, Fig. 8B). OPHN1^-/y mice did not exhibit a significant increase in sniffing time between the third and the fourth trial, indicating their inability to recognize the new odour (Fig. 8C, mice n = 11, P = 0.6). To assess whether treatment with fasudil could rescue the impaired olfactory behaviour, OPHN1^-/y mice were treated with fasudil for 30 days (mice n = 11), and then tested with the habituation-dishabituation test. We found that OPHN1^-/y mice treated with fasudil spent more time (although not statistically significant, P = 0.24) sniffing the scented paper in the fourth trial than in the third trial, than OPHN1^-/y mice (Fig. 8E). Fasudil treatment did not affect the olfactory behaviour in control mice (Fig. 8D, mice n = 8, P = 0.005**).

Figure 8.

The olfactory habituation-disabituation test. (A) First, filter papers with double distilled water were presented to control and OPHN1^-/y mice three times (habituation trials, 3 min each trial) with 1 min interval between trials. In the fourth dishabituation trial, filter paper scented with Eugenol (1mM) where presented to the mice for 3 min. (B–E) Mean investigation time (s) in the habituation (trials 1–3) and dishabituation (trial 4) in controls (B) and OPHN1^-/y mice (C), controls treated with fasudil (D) and OPHN1^-/y mice treated with Fasudil (E) Bar = SEM.

Discussion

The etiopathogenesis of ID remains poorly understood. Several studies on brain of patients with ID revealed abnormalities in dendrites and dendritic spines (22,23) suggesting that impaired cognitive abilities could be ascribed to abnormal connectivity, neuronal network function and plasticity. Inhibitory interneurons provide a significant contribution in neuronal wiring and network dynamics. The role of OPHN1 in the development and function of inhibitory interneurons remained however unknown.

In this work, we investigated the impact of OPHN1 in the generation and maturation of adult–born inhibitory interneurons in the OS, using OPHN1 ko mice. The OS is one of the few regenerative niches of the adult mammalian brain. Adult neurogenesis is a useful model to analyse neuronal development and function, since it offers a window into a continuous developmental process and it is a unique form of network plasticity (24,25). These features are of relevance studying the pathophysiology of ID, since alterations in neuronal wiring, synaptic activity and network plasticity are critical aspects of the disease and major targets of therapeutic interventions.

Significant reduced number of adult-born interneurons in the olfactory bulb (OB) of OPHN1^-/y mice

Proliferation of new cells in the SVZ was not affected by loss of function mutation of OPHN1.

The number of the adult-born cells resulted however dramatically reduced in the OB at 15 dpi. We envisioned two hypotheses to explain the reduced number of adult-born GC: 1. newly generated GC did not reach the OB; 2. Adult-born GC reached the OB, but failed to integrate into neuronal circuits. At 14 days, GC just started to receive input synaptic contacts on the proximal dendritic domain. Input-output dendro-dendritic synapses on the distal dendritic domain appear several days later and reach their final density by 28–30 days (13). Therefore, it seemed unlikely that at 15 days, when synapses on the proximal domain of GC just started to be established, altered synaptic integration could result in such a striking reduction in cell number. Our hypothesis was confirmed by the finding of a significant lower number of cells in the OB also at 9 dpi, a time point at which GC has not established yet synaptic contacts. The number of GC in the OB was not significantly different at 9 and 15 days, suggesting that a significant portion of the newly generated GC did not reach the OB. By quantitative immunohistological analyses of the expression of the marker of apoptosis CC3, we found that a higher number of neuronal precursors went into apoptosis before reaching the OB, in OPHN1^-/y. Noteworthy, we found that only ∼ 1/3 of newly generated GC reached the OB and that ∼ 2/3 of neuronal precursors went into apoptosis in the SVZ and along the RMS, indicating that the neuronal precursors that did not reach the OB went into apoptosis. The GC that reached the OB, however, followed a similar destiny in OPHN1^-/y and control mice. In both cases, the complement of adult-born GC reduced to one-half between 15 and 50 dpi. The total number of new inhibitory interneurons resulted, however, significantly lower in OPHN1^-/y mice, since fewer cells reached the OB.

Morphological and functional abnormalities of adult-born GC in OPHN1^-/y mice

The geometry of the dendritic tree is critical for neuronal information processing, since this structure integrates synaptic inputs (26,27) that in turn regulate neuronal network connectivity, synaptic activity and plasticity. Spines, elective sites of synaptic contacts, are highly plastic elements that can change in number, size and shape in response to the constantly changing flow of information (28–31). Changes in spine morphology require remodelling of membrane and actin cytoskeleton and are assumed to reflect synaptic activity modifications. These changes are particularly evident during highly plastic period of life, such as during development (32,33,35). The adult-born GC represent an interesting model to analyse changes in neuronal morphology and function, since they maintain a high degree of plasticity throughout life (25,34). Alterations in dendritic tree architecture and spines are a hallmark of ID (22) and are thought to play a critical role in cognitive impairments (35). OPHN1 contributes to regulate the dynamic of actin cytoskeleton, a critical process in dendritic elaboration and spine formation (36,37). Previous works reported that mutation in OPHN1 perturbed spine density and morphology in excitatory neurons in the hippocampus, although different results were observed according to the experimental conditions and the developmental stage of the analysed neurons (38,7). Consistent with the increased in filopodia observed by Purpura (1974) in the brain of children affected by ID, Khelfaoui (2007) observed that OPHN1 mutation induced an increase in filopodia, in primary culture of embryonic excitatory neurons of hippocampus. These results indicated that abnormal spines derived from altered development. The impact of OPHN1mutation on inhibitory interneuron morphology, remained unknown. Here, analysing adult-generated inhibitory interneurons, GC, in the OS, we found that the lack of OPNH1 did not affect the length nor the number of branching points of GC dendrites. New GC exhibited, however, a significant higher proportion and density of filopodia-like spines (as observed by Purpura 1974 and Khelfoui et al. 2007, see above) in the basal and even more strikingly, in the distal dendritic domains (basal, P < 0.001***, distal P < 0.001***) in OPHN1^-/y than in controls. Accordingly, the synaptic markers PSD-95 and vGAT were expressed in a higher number of filopodia-like spines in OPHN1^-/y mice than in controls, suggesting that these filopodia-like spines can establish synaptic contacts. The increase in filopodia-like spines in specific dendritic domains is relevant to the functional consequences on information processing in the OB. Each GC dendritic domain establishes synaptic contacts with specific postsynaptic targets to exert distinct functions in neuronal information processing. Dendrites of the distal domains provide the mechanism for lateral inhibition on MC, that regulates the tuning, the amplitude and the temporal pattern of MC activity (39,40). In the basal domain, GC receive only input synapses from a collateral of MC axons and from centrifugal fibres. The latter provide top-down modulation of sensory information, that reflects the behavioural state of the individual (41,42). Alterations in each dendritic domain are therefore likely to affect the specific aspects of information processing in the OB circuitry.

Recording synaptic activity impinging on MC we found that some features of mIPSC and sIPSC, were increased in OPHN1^-/y mice with respect to controls. These results suggest a compensatory upregulation of inhibitory neurotransmission following the reduced number of new GC. In particular, the increased frequency and amplitude of mIPSC could be ascribed to: an increase in GABA release from GC that could be related to (i) a higher number of sites of release Indeed, we found a significant increase of filopodia-like spines, mostly on the distal dendritic domain, where GABA is released. (ii) A higher amount of release of GABA/per site. We found a higher expression of vGAT on the distal filopodia-like spines. Expression of vGAT is related to the expression of GAD 65/67 and therefore to the availability and release of GABA (43). (iii) The sensitivity or the number of GABA_A receptors on the postsynaptic target, i.e. MC, could be higher and that could explain the larger amplitude. We favour the former hypothesis since the number of GABA_A receptors on MC was similar in OPHN1^-/y and control mice. The fact that the frequency of sIPSC was not increased, suggested that the compensatory upregulation of inhibitory neurotransmission is mostly a subthreshold phenomenon that likely did not achieve a real compensation of the reduced number of adult-born inhibitory interneurons.

Abnormalities of inhibitory interneurons can affect sensory information processing and inhibitory-excitatory balance. These factors are thought to play a critical role in the etiopathogenesis of several neurodevelopmental disorders (44,45). The results here presented seem to corroborate that hypothesis.

Inhibition of ROCK rescued the altered adult-born GC

OPHN1 can interact with RhoA, Rac and Cdc 42, although it seems to act electively on RhoA (7,38). As a consequence of the loss of function of OPHN1, Rho GTPases persist in the active GTP bound state, leading to overactivation of the downstream signalling pathway, such as the Rho kinases (46). Inhibition of ROCK was reported to revert morphological (38) or functional (47) abnormalities triggered by OPHN1 mutation in excitatory neurons, in vitro. Whether ROCK inhibitors could rescue abnormalities related to loss of function mutation of OPHN1 also in inhibitory interneurons, in vivo, in OPHN1^-/y mice, remained unknown. Here, we found that chronic treatment in vivo with the clinically approved ROCK inhibitor fasudil (21) was able to revert most of the morphological and functional alterations in OPHN1^-/y mice. Namely, upon 3 weeks of treatment in vivo with fasudil, the number of adult-born GC in the OB was superimposable in OPHN1^-/y mice and in controls. Furthermore, we found that fasudil treatment reverted the proportion of mushroom-filopodia spines in OPHN1^-/y mice. The number of spines, in particular of mature spines, was however higher both in control and in OPHN1^-/y treated mice. In previous works (48,38) it was reported that treatment with a different inhibitor of ROCK, i.e. 27632, rescued the spine defects triggered by OPHN1 mutation, but did not affect the controls. The different results can likely be ascribed to several and significant differences between our experimental conditions and those of previous studies (48,38). In previous works a different inhibitor of ROCK, i.e. 27632 was applied for 48 hours on organotypic culture of rat hippocampus. In our work we performed a chronic treatment (3–4 weeks) with fasudil, in mice carrying a null mutation for OPHN1. Fasudil was chosen since it can be conveniently administered in drinking water (i.e. it is able to pass the brain barrier) and therefore it is ideal for chronic treatment in vivo in mice (49). Different ROCK inhibitors were shown to exert different effects also on other systems (21). Furthermore, analysis was performed on different type of neurons, i.e. newborn inhibitory GC of the OB, in our case, versus excitatory hippocampal neurons, in previous works.

Electrophysiological recordings targeted to MC revealed that amplitude, decay time and area of mIPSC, were successfully rescued in fasudil treated animals, although the mIPSC frequency persisted to be higher. The latter result is likely related to the high number of spines induced by fasudil treatment (see above).

To explore the behavioural outcome of the altered neuronal circuitry in the OB, we examined an odour-guided behaviour. Specifically, we employed a task that allows for odour investigation, odour learning and memory and odour discrimination to be assessed within a single behavioural test, i.e. habituation dishabituation task (50,51). We found that in OPHN1^-/y mice the olfactory behaviour was significantly perturbed. Fasudil treatment resulted in a consistent improvement in the olfactory task, that was not however totally reverted to control level.

These data suggest that ROCK inhibitors, such fasudil, have the potential to provide benefits to some of the alterations triggered by OPHN1 mutations, although not to all, as observed also in previous work both in vivo and in vitro (5,47). This is conceivable considering the complexity of the disease, the different brain areas affected and the different types of defects observed. A multi therapeutic approach could be therefore required.

In the present work, we unravelled the impact of OPHN1 in regulating morphology and functions of adult generated inhibitory interneurons in the OB and we identified possible therapeutic strategies to revert or improve abnormalities induced by OPHN1 mutation.

Materials and Methods

All animal procedures were performed under protocols approved by the ethical committee of the University of Padua and of the University of Milan.

Mutant line mice

Experiments were performed on the genetically modified line of mice, OPHN1 knock-out mice, that was generously provided by Pierre Billuart, and described in details previously (7). Experiments were performed on male OPHN1 knock-out mice indicated as OPHN1^-/y and control littermates indicated as control mice.

BrdU labelling

To estimate the number of newly generated cells, a DNA synthesis marker, 5-Bromo-2′-deoxyuridine (BrdU) (Sigma-Aldrich, 50 mg/Kg body) was administered intraperitoneally. Four injections, every 2 hours, were performed in OPHN1^-/y mice and wild type C57B/6 mice, postnatal day (P) > 45. To assess the neurogenesis and the survival of newborn GC, animals were sacrificed by Xilor-Zoletil overdose 24 h, 9, 15 and 50 days, respectively, after BrdU injections and then transcardially perfused with 0.9 saline, followed by 4% Paraformahaldehyde in 0.1M phosphate buffer saline (PBS). Brains were then sectioned at vibratome (Leica, VT 1000S, 60 µm thickness section for the olfactory bulb, 45 µm thickness section for the subventricular zone). Sections were pretreated with the DNA denaturating agent HCl (2N, at 37°C). To detect the new born cells 24 h after brdU injection, sections of the subventricular zone were incubated with a rat monoclonal anti BrdU (1:200, Abcam) primary antibodies, that were then visualized using Cy3 conjugate anti-rat IgG (Jackson laboratories). To detect adult-born GC 9, 15 and 50 days after Brdu injections, olfactory bulb sections were incubated with a rat monoclonal anti BrdU (1:200 Abcam) and with a mouse monoclonal anti NeuN, a neuronal marker, (1:200, Millipore) primary antibodies. The bound primary antibodies were visualized using Cy3 conjugate donkey anti-rat IgG and DyLight 488 conjugate donkey anti-mouse IgG, respectively (Jackson laboratories).

Stereotaxic injections

GFP-expressing lentiviruses (gift of Angela Gritti, San Raphael Hospital, Milan) were stereotaxically injected in the subventricular zone in OPHN1 knock-out mice and wild type C57B/6, at postnatal day (P) > 45. Stereotaxic injections were performed using a picopump (WPI), at the following coordinates (relative to bregma): anterior = 1 mm, lateral = 1 mm, depth = 2.7 mm (Petreanu and Alvarez-Buylla, 2002). The mice were killed by Xilor-Zoletil overdose 30 days after lentivirus injection and then transcardially perfused with the 0.9% saline solution, followed by 4% paraformahaldehyde in 0.1M phosphate buffer saline (PBS). Brains were sectioned at vibratome (Leica, VT 1000S; 60 µm thickness section) and mounted (Elvanol).

Image analysis and quantification

To estimate the number of adult-born GC, Brdu and NeuN double-labelled cells were counted in the entire GC layer on every two sections (120 µm) of the olfactory bulb (OB). To determine the total number of newly generated cells, all BrdU immunopositive nuclei were counted in the entire GC layer on every two sections of the OB. The ratio Brdu-NeuN/BrdU tot cells (%) where the Brdu –NeuN = the number of newly generate neurons, BrdU tot = total number of newly generated cells, was used to determine the fate of the newly generated cells. Immunofluorescent sections were analysed using a Leica SP5 confocal microscope with HC PL Fluotar 20X/0.50 NA objective (Leica). To analyse morphology of dendrites and spines of adult-born GFP-labelled GC, 60 µm thick sections of the OB were analysed at the confocal microscope (Leica SP5) with a HC PL Fluotar 20X/0.5 NA and with a HCxPL Apo LambdaBlu 63X/1.40 NA oil immersion objectives (Leica). Quantitative data analysis on BrdU, dendrite and spines was performed by hand using ImageJ (NIH) and Neuronstudio program (52,53). All counts were done blindly.

Drug administration

OPHN1 knock-out mice and controls were treated with the Rho kinase (ROCK) and Protein Kinase A (PKA) inhibitor Fasudil also known as HA1077 (Selleckchem.com)

Fasudil was dissolved in daily drinking water at 0.65 mg ml⁻¹ and given orally i) for 3 weeks, for BrdU experiments, starting fasudil treatment 1 week before BrdU injection; ii) for 4 weeks for morphological analysis of GFP labelled adult-born GC. In this case the treatment began the day of Lentivirus injection in the subventricular zone.

Immunohistochemistry and quantitative analysis

Immunohistochemistry was performed on free floating sections using the following primary antibody: Rabbit anti PSD-95 (1:1000, Invitrogen); Rabbit anti vGAT (1:250, Synaptic System, Gottingen, Germany); Rabbit anti GABA_A a1 subunit (1:100, Alomone labs); Rabbit anti cleaved Caspase-3, CC3 (1:250, Cell Signaling); Goat anti doublecortin, DCX (1:1000, Santa Cruz) at 4°C overnight. The primary antibodies were visualized with Cy3-conjugated donkey anti rabbit (Jackson Laboratories); DyLight 488- conjugated donkey anti goat (Jackson Laboratories) antibodies. Images were acquired at the confocal microscope (Leica SP 5) using a HCxPL Apo LambdaBlu 63X/1.40 NA oil immersion objectives (Leica).

Quantitative analysis of different markers was performed with ImageJ (NIH).

Western blot

Mouse tissues were dissected and solubilized with RIPA lysis buffer supplemented with protease inhibitor cocktail (Sigma-Aldrich) and then cleared at 14,000 g at 4°C for 15 min. Protein concentrations were determined using the BCA protein concentration assay as manufacturer’s instructions (Thermo Scientific) and subjected to Western blotting using standard procedures.

The following antibodies were used: rabbit anti-MYPT1 (1:500, Cell Signaling), rabbit anti-Phospho-MYPT1 (Thr696, 1:500, Cell Signaling). Membranes with GAPDH were probed with anti-GAPDH antibody to assure equal loading. Protein levels were normalized to GAPDH levels.

Quantification of the ratio of phospho-MYPT1 against total MYPT1 density, was calculated using densitometric analysis with ImageJ software.

Electrophysiology

C57Bl/6 OPHN1^+/+and OPHN1^-/y mice were anaesthetized in a chamber saturated with chloroform and then decapitated. The brain was rapidly removed and placed in an ice-cold solution containing 220 mM sucrose, 2 mM KCl, 1.3 mM NaH₂PO₄, 12 mM MgSO₄, 0.2 mM CaCl₂, 10 mM glucose, 2.6 mM NaHCO₃ (pH 7.3, equilibrated with 95% O₂ and 5% CO₂), and 3 mM kynurenic acid. Horizontal olfactory bulb slices (thickness, 250 μm) were prepared with a vibratome VT1000 S (Leica) and then incubated first for 40 min at 37°C and then for 40 min at room temperature in artificial CSF (aCSF), consisting of (in mM) 125 NaCl, 2.5 KCl, 1.25 NaH₂PO₄, 1 mM MgCl₂, 2 mM CaCl₂, 25 mM glucose, and 26 mM NaHCO₃ (pH 7.3, equilibrated with 95% O₂ and 5% CO₂). Slices were transferred to a recording chamber perfused with aCSF at a rate of ∼2 ml/min and at room temperature.

Whole-cell patch-clamp electrophysiological recordings were performed with a Multiclamp 700B amplifier (Axon CNS molecular devices, USA) and using an infrared-differential interference contrast microscope. Patch microelectrodes (borosilicate capillaries with a filament and an outer diameter of 1.5 μm; Sutter Instruments) were prepared with a four-step horizontal puller (Sutter Instruments) and had a resistance of 3–5 MΩ.

Spontaneous and miniature inhibitory post synaptic currents (sIPSCs/mIPSCs) were recorded at a holding potential of −65 mV with an internal solution containing 140 mM CsCl, 2 mM MgCl₂, 1 mM CaCl₂, 10 mM EGTA, 10 mM HEPES-CsOH, 2 mM ATP (disodium salt) (pH 7.3), and 5 mM QX-314 (lidocaine N-ethyl bromide), for the sIPSCs only. Access resistance was between 10 and 20 MΩ if it changed by >20% during the recording, the recording was discarded.

All GABAergic currents were recorded in the presence of kynurenic acid (3 mM) in the ACSF. For mIPSCs, lidocaine (500 μM) was added in the ACSF. Currents through the patch-clamp amplifier were filtered at 2 kHz and digitized at 20 kHz using the Clampex 10.1 software. Analysis was performed offline by the Clampfit 10.1 software.

Olfactory habituation-dishabituation test

Mice were habituated to the cage (35 × 20×14 cm) for 3 min. Subsequently, filter paper (2 × 2 cm) with 20 µl of double distilled water (ddW) was presented to the mice for 3 times, 3 min each time, with 1 min inter-trial interval. On the fourth trial, a filter paper with 20 µl of Eugenol (1mM, Sigma Aldrich) was presented for 3 min. Investigation times during the 3-min test period were measured on the third and fourth trials. The mouse behaviour was recorded with a digital video camera (30 frames per second, 2592 × 1944 pixels) for the analysis. We defined ‘an investigation’ as a nasal contact with the filter paper within a 1mm distance.

Statistical analysis

Data are presented as means ± S.E.M. Statistical comparisons of pooled data were performed by Student's t-test with the use of Prism software (GraphPad, San Diego, CA, USA). A P value of <0.05 was considered statistically significant. * = 0.01< P < 0.05; ** = 0.001 <P < 0.01 ***= P < 0.001.

Supplementary Material

Supplementary Material is available at HMG online.

Conflict of Interest statement. None declared.

Funding

This work was supported by Telethon grant GGP11116 and Cariparo Foundation grant and NANOmax to C.L.