Deletion of Munc18-1 in 5-HT Neurons Results in Rapid Degeneration of the 5-HT System and Early Postnatal Lethality

The serotonin (5-HT) system densely innervates many brain areas and is important for proper brain development. To specifically ablate the 5-HT system we generated mutant mice carrying a floxed Munc18-1 gene and Cre recombinase driven by the 5-HT-specific serotonin reuptake transporter (SERT) promoter. The majority of mutant mice died within a few days after birth. Immunohistochemical analysis of brains of these mice showed that initially 5-HT neurons are formed and the cortex is innervated with 5-HT projections. From embryonic day 16 onwards, however, 5-HT neurons started to degenerate and at postnatal day 2 hardly any 5-HT projections were present in the cortex. The 5-HT system of mice heterozygous for the floxed Munc18-1 allele was indistinguishable from control mice. These data show that deletion of Munc18-1 in 5-HT neurons results in rapid degeneration of the 5-HT system and suggests that the 5-HT system is important for postnatal survival.


Introduction
The 5-HT system consists of clusters of cell bodies in the midbrain raphe nuclei, with the largest clusters in the median raphe nucleus and the dorsal raphe nucleus (DRN). Several brain areas receive dense 5-HT innervation and 5-HT is released both synaptically and as volume transmission [1,2]. Due to this and to the several 5-HT receptor subtypes which are present in the brain, 5-HT has many roles and influences many processes in the brain [3].
Neurogenesis of 5-HT neurons in the mouse brain occurs in the ventral rhombencephalon around embryonic day (E) 10 [4]. One day later, 5-HT neurons begin to synthesize and secrete 5-HT and start growing out axons. Around birth, target areas such as the forebrain and the hippocampus are densely innervated with 5-HT projections. Only after birth, the maturation of the 5-HT network is completed.
Several studies have addressed the role of 5-HT on the development of the 5-HT system and brain development. In a conditional Lmx1b knockout (KO) mouse, almost all 5-HT neurons fail to survive, resulting in a significant decrease in brain tissue 5-HT levels [5]. However, these mice do not show an overt phenotype and survive to adulthood [5]. In contrast, it was shown that maternal 5-HT is required for embryonic development [6]. Furthermore, in tryptophan hydroxylase 2 (Tph2) KO mice 5-HT neurons are completely devoid of 5-HT, but the morphology and neurite distribution of the 5-HT system is not affected and these mice do show only a subtle behavioural phenotype [7,8,9]. Neonatal depletion of 5-HT by the neurotoxin 5,7-Dihydroxytryptamine results in rather subtle changes in behavioural response and brain development [10,11].
In this study we silenced the 5-HT system by conditional deletion of Munc18-1 in 5-HT neurons. Munc18-1 is a presynaptic protein which is essential for vesicle release and neurons that lack Munc18-1 have a complete absence of neurotransmitter secretion [12]. Via interactions with Syntaxin1A and the SNARE complex, Munc18-1 is involved in vesicle docking and fusion [13]. Munc18-1 knockout mice are born paralyzed and die immediately after birth [12]. In these mice, initially synapses are formed and the assembly of the brain is normal. However, in later stages of brain development there is massive neuronal cell death and brain degeneration [12]. Since Munc18-1 knockout mice die immediately after birth, we have generated Munc18-1 lox/lox mutant mice in order to conditionally delete Munc18-1. Crossing these mice with a L7-Cre line, with Cre expressed in Purkinje neurons in the cerebellum, resulted in mice which developed severe ataxia, suggesting a cerebellar phenotype [14].
In SERT-Cre cre/wt Munc18-1 lox/lox mice 5-HT neurons were initially generated and 5-HT projections innervated the midbrain and cortex, later followed by degeneration and loss of 5-HT projections in the cortex. The majority of these mice died within a few days after birth. These data suggest that the 5-HT system contributes importantly to postnatal brain development.

Deletion of Munc18-1 in SERT expressing neurons results in postnatal lethality
To assess the effect of deletion of Munc18-1 in 5-HT neurons, we crossed Munc18-1 lox/lox mice with SERT-Cre mice, which express Cre in SERT expressing neurons. These are the 5-HT neurons in the raphe nuclei, but also some hippocampal neurons and thalamocortical neurons which express SERT transiently during development [15,16]. Crossing SERT-Cre mice with Munc18-1 lox/lox mice results in mice in which Munc18-1 is specifically removed in SERT expressing neurons (Fig. 1A). We crossed SERT-Cre cre/wt Munc18-1 lox/wt mice with SERT-Cre wt/wt Munc18-1 lox/wt mice which should result in 12.5% of offspring which have SERT-Cre cre/wt and Munc18-1 lox/lox genotypes. Genotyping 101 mice three weeks after birth revealed that genotype frequencies were not distributed at Mendelian ratio (p = 0.027). Only three SERT-Cre cre/wt Munc18-1 lox/lox mice were found, whereas based on Mendelian ratio the expected number of SERT-Cre cre/wt Munc18-1 lox/lox mice should be 13, giving a 77% mortality rate in SERT-Cre cre/wt Munc18-1 lox/lox mice (Fig. 1B). These three remaining SERT-Cre cre/wt Munc18-1 lox/lox mice all died in the fourth postnatal week. This shows that Munc18-1 deletion in SERT expressing neurons results in postnatal lethality.

5-HT neurons are initially generated but then quickly degenerate
To assess the effect of Munc18-1 deletion in 5-HT neurons on the development of the 5-HT system, we made coronal sections of paraformaldehyde (PFA) fixed brains at different developmental stadia and performed immunohistochemistry for 5-HT. As developmental time points we chose E16, E18 and postnatal day (P) 2. Expression of SERT, and thus of Cre recombinase, starts at E11 and the earliest recombination observed in the DRN in SERT-Cre mice was at E12.5 [16]. Since it will take some days before all remaining Munc18-1 mRNA and protein is degraded, we chose E16 as the first time point. We first focused on the 5-HT cell bodies in the DRN. Immunohistochemistry for 5-HT on brain slices from E16 SERT-Cre cre/wt Munc18-1 lox/lox and control mice showed that in the DRN 5-HT neurons are present and these showed the characteristic DRN topology ( Fig. 2A,B). The number of 5-HT neurons did not differ from control (control 312617.67, SERT-Cre cre/wt Munc18-1 lox/lox 305625.43, Fig. 2O). However, morphological analysis of the neurons showed that these were already degenerating, as assessed by the round morphology and reduced number of primary neurites (Fig 2G,H). At E18, there was a ,80% decrease in the number of 5-HT neurons in SERT-Cre cre/wt M18 lox/lox mice compared to control (control 358610.82, SERT-Cre cre/wt Munc18-1 lox/lox 5964.84, Fig . At E18 and P2 the remaining 5-HT neurons in SERT-Cre cre/wt Munc18-1 lox/lox brain sections displayed an altered morphology compared to control 5-HT neurons ( Fig. 2I,J,K,L). 5-HT neurons in control brains had a characteristic fusiform or ovoid shape and grew out several primary neurites (Fig. 2M). In contrast, the mutant 5-HT neurons had a rounded morphology and the majority did not contain primary neurites, indicative of degenerating neurons (Fig. 2N). This showed that between E16 and P2 there is a massive degeneration of 5-HT neurons in the DRN.

5-HT projections innervate the DRN and cortex but are degenerated at P2
Next, we focused on the innervation of the midbrain and cortex with 5-HT projections. Midbrain and cortex are the first regions which are densely innervated with 5-HT projections during development. We quantified 5-HT projection density first in DRN midbrain sections at E16, E18 and P2. At E16 the sections from both control and SERT-Cre cre/wt Munc18-1 lox/lox mice contained several 5-HT projections and there was no significant difference in 5-HT innervation density (1.3160.51% and 0.6360.24% respectively, Fig. 3A,B). At E18, in control sections the 5-HT innervation density was slightly increased compared to E16 (Fig. 3C). However, in SERT-Cre cre/wt Munc18-1 lox/lox sections the 5-HT innervation density was decreased compared to control (control 2.3260.86%, SERT-Cre cre/wt Munc18-1 lox/lox 0.3160.27%, Fig. 3D).
Development of the 5-HT system is not affected in SERT-Cre cre/wt Munc18-1 lox/wt mice Finally, we investigated whether deletion of one allele of Munc18-1 in SERT expressing neurons affected the number of 5-HT cell bodies in the DRN, or the 5-HT fiber density in the DRN and cortex. Analysis of midbrain sections from control and SERT-Cre cre/wt Munc18-1 lox/wt (hz) mice showed that 5-HT neurons were distributed in the characteristic DRN topology, and control and hz were indistinguishable (Fig. 4A,B). 5-HT cell bodies in both control and hz had a fusiform or ovoid shape (Fig. 4C,D). Quantification of the number of 5-HT cell bodies in the DRN revealed no difference between control and hz (control 364614.78, hz 361615.01, Fig. 4I). Likewise, the 5-HT projection density in the midbrain and cortex were similar (midbrain control 4.3860.93%, hz 4.6760.55%; cortex control 2.1860.35%, hz 2.8760.99%, Fig. 4E-H,J). Thus, deletion of one allele of Munc18-1 does not affect 5-HT neuronal survival, 5-HT neurite outgrowth or 5-HT projection density.

Discussion
In this study, we showed that SERT-Cre cre/wt Munc18-1 lox/lox mice display a postnatal lethality phenotype, accompanied by a rapid degeneration of the 5-HT system. Inactivation of one allele of Munc18-1 in SERT expressing neurons does not affect 5-HT cell number or 5-HT innervation density.
Crossing Munc18-1 lox/lox mice with SERT-Cre mice results in deletion of Munc18-1 from SERT expressing cells. It was shown previously that crossing SERT-Cre mice with loxSTOPlox-EYFP mice results in .99% of 5-HT neurons in the DRN which are EYFP positively labelled [15]. However, SERT is also transiently expressed in some other brain regions during development, such as the thalamocortical neurons [16,17]. In SERT-Cre ROSA-loxSTOPlox-LacZ double positive mice, strong LacZ expression was not only observed in the DRN, but also in the thalamus, cingulate cortex and in the CA3 region of the hippocampus [15,16]. Therefore, in SERT-Cre Munc18-1 lox/lox mice also in some non-5-HT neurons Munc18-1 will be deleted. Genotyping sacrificed mice at E16 or E18 revealed that there were more SERT-Cre cre/wt Munc18-1 lox/lox mice than expected, showing that these mice do not die during embryonic development. However, genotyping mice three weeks after birth showed that the majority of SERT-Cre cre/wt Munc18-1 lox/lox mice had died, and the remaining SERT-Cre cre/wt Munc18-1 lox/lox mice died within the fourth postnatal week.
Previously we showed that in Munc18-1 KO mice the neuromuscular junction initially develops, but at later stages most of the motor neuronal cell bodies in the spinal cord degenerate [18]. In analogy with this, in the absence of regulated secretion the hypothalamoneurohypophysial system is initially normally formed, but in later stages of development degenerates [19]. This is consistent with our observation that Munc18-1 is dispensable for initial generation and outgrowth of 5-HT neurons, but required for their survival.
Since also in some other neurons Munc18-1 is deleted, it is unclear whether the observed phenotypes can be attributed to the degeneration of the 5-HT system. In Tph2 KO mice, 5-HT neurons are completely devoid of 5-HT, yet there are no alterations in 5-HT neuron morphology and 5-HT neurite distribution [7]. Although these mice can survive into adulthood, approximately 50% does not survive the first four weeks and surviving mice have growth retardation [8]. Additionally, in a conditional Lmx1b KO mouse, 5-HT neurons are initially formed but almost all central 5-HT neurons fail to survive and brain 5-HT levels are reduced to ,10% [5]. Remarkably, these mice display normal locomotor behaviour and show no abnormalities in gross brain morphology. Recently, it was shown that these mice have severe apnea and have a ,20% mortality rate during the neonatal period, although the majority of these mice survive beyond P28 [20,21].
Apparently, even the absence of central 5-HT synthesis or an almost complete removal of 5-HT neurons does not result in a severe (postnatal) lethality phenotype. Thus, although several of these mice display high (postnatal) mortality, the phenotypes observed are not as severe as in the SERT-Cre cre/wt Munc18-1 lox/lox mice.
On the other hand, mice which lack GAP43 display disrupted barrel cortex formation and thalamocortical connections fail to form. These mice display a postnatal lethality phenotype, with 50% of homozygotes which die between P0 and P2, and .95% of homozygotes which die before P21 [22]. Although these mutant mice display a reduced 5-HT innervation in the cortex and hippocampus, they have no significant difference in the number of 5-HT neurons, and other areas such as the piriform cortex and amygdale receive normal 5-HT innervation [23]. Thus, based on these data it is unclear whether the postnatal lethality phenotype observed in this study is attributable to a degeneration of thalamocortical or 5-HT neurons. It would be very interesting to delete the Munc18-1 gene specifically in 5-HT neurons using the Pet-Cre transgenic line, in which expression of Cre is restricted to 5-HT neurons [24]. This approach could be used to investigate whether the phenotype we observed could indeed be attributed to degeneration of the 5-HT system rather than degeneration of thalamocortical neurons.
In conclusion, deletion of both alleles of Munc18-1 in SERT expressing neurons results in a rapid degeneration of the 5-HT system and postnatal lethality.

Laboratory animals
Generation and characterization of SERT-Cre mice has been described previously [15]. Briefly, immediately upstream of the SERT translational start codon a cassette containing Cre and a FRT-flanked neomycin cassette was inserted. To generate conditional Munc18-1 mice [14], floxed Munc18-1 mice were generated by insertion of LoxP sites flanking the second exon using homologous recombination. To obtain SERT-Cre cre/wt Munc18-1 lox/lox mice, SERT-Cre cre/wt Munc18-1 lox/wt mice were crossed with SERT-Cre wt/wt Munc18-1 lox/wt mice. Pregnant females were sacrificed by cervical dislocation. Mouse embryos were obtained by caesarean section of pregnant females from timed matings. To investigate the effect of deletion of Munc18-1 in the 5-HT system on the development of the 5-HT system, SERT-Cre cre/wt Munc18-1 lox/lox mice were used. SERT-Cre wt/wt Munc18-1 lox/lox , SERT-Cre wt/wt Munc18-1 lox/wt , or SERT-Cre cre/wt Munc18-1 wt/wt littermates were not different from SERT-Cre wt/wt Munc18-1 wt/wt mice and were used as controls. SERT-Cre mice were genotyped as has been previously described [15]. For genotyping of Munc18 lox/lox mice the following primers were used: 59-ttggtggtcgaatgggcaggtag-39, 59-cctgtatgggtactgttcgttcactaaaata-39 and 59-ttctgaacttgaggccagtctgagacacag-39. These studies were approved by the institutional ethical committee of the VU University (Protocol FGA 06-11-2). Animals were housed and bred according to institutional and Dutch guidelines.

Immunohistochemistry
For immunohistochemical analysis of 5-HT, embryonic mouse brains were dissected and immediately fixed by immersion in 4% PFA in phosphate buffered saline (PBS, pH 7.4). For analysis of 5-HT neurons and projections in adult brains, mice were anaesthetized and transcardially perfused with 4% PFA in PBS. For post-fixation, brains were incubated in 4% PFA in PBS overnight. For cryoprotection, brains were incubated in increasing concentrations of sucrose. Subsequently, coronal slices of 40 mm were made. For immunohistochemistry, brain slices were first incubated in blocking buffer containing PBS supplemented with 0.5% Triton X-100 and 10% normal goat serum for two hours. Subsequently, slices were incubated overnight with primary antibodies in PBS containing 0.5% Triton X-100 at 4uC under gentle agitation. The next day, the slices were washed three times two hours in PBS and incubated with secondary antibodies in PBS for one hour under gentle agitation. Finally, slices were washed again three times two hours in PBS and mounted in Dabco Mowiol for analysis. All procedures were performed at room temperature unless otherwise stated. Polyclonal anti-5-HT (1:1000) (Immunostar/Diasorin) was used as the primary antibody, and as a secondary antibody GAR-546 (1:1000) was used (Invitrogen). To analyze number of 5-HT cell bodies in DRN sections, at least three sections per experimental group were imaged using a CLSM 510 microscope. Number of 5-HT cell bodies was manually counted in the slices. For analysis of 5-HT fiber density in DRN and cortex Z-stacks of at least three positions per group was made. To analyze the 5-HT fiber density, the Zstacks were projected in a single image and binarized. The 5-HT fiber density was quantified as the area in the image occupied by 5-HT projections compared to the total area.

Statistical analysis
Data were analyzed using SPSS 17.0. For analysis of number of 5-HT cell bodies in DRN and 5-HT projection density in DRN and cortex, data was analyzed using the independent samples ttest. To analyze whether observed genotype frequencies differed significantly from a Mendelian ratio, a one-way chi-square test was performed. Data shown are mean6SEM. Significance level was set at p,0.05.