A smaller olfactory bulb in a mouse model of Down syndrome

1 Department of Computer Science, University of Verona, Verona, Italy, 2 Department of Neurosciences Biomedicine and Movement Sciences, University of Verona, Verona, Italy, 3 Centro Piattaforme Tecnologiche dell’Università di Verona,Verona, Italy, 4 Department of Molecular Medicine, University of Padova, Padova, Italy, # These two authors contributed equally to this work, * Email: carla.mucignat@unipd.it


INTRODUCTION
Down syndrome (DS), which affects approximately 1 out of 700 live births (Parker et al., 2010) is characterized by trisomy of chromosome 21. DS is characterized by several conditions including intellectual disability, craniofacial alterations, congenital heart disease, early onset Alzheimer's disease, gastrointestinal disorders, low muscle strength (Korenberg et al., 1994). Olfactory dysfunction has been consistently found in DS (Murphy et al., 1996;Cecchini et al., 2016) together with age-associated development of Alzheimer's disease neuropathology in the olfactory system, starting from the entorhinal and trans-entorhinal areas (Price et al., 1991), as well as neurodegenerative pathology (Mann et al., 1986). The olfactory bulb (OB) is the first central relay station of the peripheral olfactory neurons playing a key role in olfactory processing (Mori et al., 1999). In the OB, multiple types of neurons network to process information and transmit it to the olfactory cortex. Despite several techniques have been used to investigate the DS brain in vivo (review in Neale et al., 2017), information on OB status in living persons with DS is quite limited.
The Ts65Dn mouse is one of the most extensively studied animal models of DS. It presents segmental trisomy for chromosome 16 and involves roughly half of human chromosome 21 orthologs (Davisson et al., 1993). In the Ts65Dn mouse, smell impairments appear in adulthood (Bianchi et al., 2014) since reduced neuro-genesis takes place: this has been suggested to underlie hypocellularity of the OB, ultimately leading to deficit in olfactory learning through a reduction in the number of new granule neurons migrating to the OB from the subventricular zone. This has been challenged by others (López-Hidalgo et al., 2016). Similar alterations have also been observed in persons with DS. Moreover, Ts65Dn mice present impaired refinement of synaptic connections by olfactory neurons in the OB (William et al., 2017). Accordingly, investigating the OB in the Ts65Dn mouse is relevant to clarify the mechanisms of olfactory impairment in DS. Magnetic resonance imaging (MRI) is a recognized, powerful and non-invasive tool to investigate the in vivo morphology and function of organs in both humans and experimental animals. Using the same technology, MRI findings in experimental animals can be easily translated to the patient. In this work we examined trisomic and euploid Ts65Dn mice by means of MRI to highlight possible DS-related changes in the OB. In particular, we focused on an MRI protocol that could provide structural and anatomical insight in the DS-related changes of the OB, while being compatible with in vivo examination in a fragile animal model.

METHODS
Ts65Dn (strain: B6EiC3Sn.BLia-Ts(17<16>)65Dn/DnJ) breeder mice were obtained from the Jackson Laboratory, ME, USA. The colony was maintained by breeding trisomic female mice to euploid B6EiC3Sn.BLiAF1/J males. Pups were weaned at 21 days of age. Tissue for genotyping was obtained from tail clips in p11 mice. Genotyping was accomplished by Mmu17 translocation breakpoint separated PCR. Adult (2-8 months of age) male Ts65Dn mice (8 trisomic and 8 euploid) were used in this work. Male mice were chosen for this study because of lack of variability associated with the estrous cycle. Mice were housed in groups of 3-4 by genotype and maintained under standard conditions (24±1°C temperature, 60±15% relative humidity, and 12 h light/ dark cycle) and fed ad libitum with standard commercial chow.
The experimental protocol was reviewed and authorized according to EU Directive 2010/63/EU by the Italian Ministry of Health (ref.: 538/2015-PR).
For MRI, mice were anesthetized by inhalation of a mixture of air and O 2 containing 0.5-1% isoflurane and placed prone with their head in stereotactic position.
Images were acquired using a Biospec Tomograph (Bruker, Karlsruhe, Germany) equipped with a 4.7 T, 33 cm bore horizontal magnet (Oxford Ltd, Oxford, UK). A double coil configuration was used: the ex-citation r.f. pulses were applied trough a 7.2 cm bird cage volume coil, while the signal was received through a 2-channel surface coil optimized for the mouse brain (Bruker). After a sagittal scout image, 6-7 contiguous 0.5 mm-thick slices were acquired through the whole OB using a RARE T 2 -weighed sequence with TR (repetition time) = 5000 ms, TE (echo time) = 72 ms, FOV (field of view) = 2 × 2 cm 2 , NEX (number of average) = 10, matrix size = 256 × 192, zero-filled at 256 × 256, corresponding to an in plane resolution of 78 × 78 μm 2 . These images were used for calculation of OB volume. For morphometry, the free ImageJ software (National Institutes of Health, USA) was used. The OB area was measured in each 0.5-mm thick axial section containing the OB. Areas were expressed in mm 2 and volumes consequently calculated according to Cavalieri's principle (Roberts et al., 1993). In order to ensure consistence, one operator blind to mouse genetics performed all morphometric measurements. Further on, 1 mm-thick slices of the whole OB were acquired using a multiecho spin-echo sequence, with 10 echoes, TR of 2600 ms, and TE ranging from 20 to 200 ms, for generating quantitative T 2 mapping (Bontempi et al., 2018). T 2 maps were generated with Paravision 5.1 (Bruker, Karlsruhe, Germany).
Diffusion tensor imaging (DTI) is based on the observation of water molecule diffusivity in the brain; DTI is especially employed to characterize the orientation and integrity of the white matter (Basser et al., 2002;Le Bihan et al., 2001). DTI images were acquired with an Echo Planar Imaging (EPI) sequence optimized for the mouse olfactory bulb. Imaging parameters were set as follows: TR=3000 ms, TE=36 ms, FOV=2 × 2 cm 2 , matrix size=128 × 128, 12 transversal 0.5 mm thick slices and 8 EPI segments. Diffusion images were acquired in 12 non-collinear directions with a b-value of 750 s/mm 2 (3 b0 images were acquired). DTI derived maps, Fractional Anisotropy (FA), Apparent Diffusion Coefficient (ADC), Axial Diffusivity (AD) and Radial Diffusivity (RD), were calculated with Paravison 5.1 (Bruker, Karlsruhe, Germany). Images showed no obvious distortion in the area of interest; hence a distortion correction procedure was not applied.
Comparison between the two groups of mice was carried out with the Mann-Whitney test. Results are presented as box-and-whiskers plots. Effect size (ES) was rated according to Cohen (1988) as follows: 0.2 small, 0.5 medium, 0.8 large. Correlation analysis was carried out with Spearman rho.

RESULTS
One euploid mouse died during MRI and was discarded. Age was not significantly different in the two groups of mice (P=0.204), with euploid showing higher body mass (P=0.011; ES=1.7) as previously found (Fructuoso et al., 2017;Giacomini et al., 2018). Fig. 1 shows representative MRI images of euploid and trisomic Ts65Dn mouse head ( Fig. 1A and 1E, respectively) showing the area used for OB morphometry (Fig. 1B and 1F) and T 2 maps (Fig. 1C, 1G) as well as FA (Fig. 1D, 1H). Results of MRI analysis of the olfactory bulb in trisomic and euploid Ts65Dn mice are presented in Fig. 2.
OB volume was reduced in trisomic vs. euploid mice, the difference being statistically significant, P=0.011, ES=1.7 ( Fig. 2A). Mean absolute T 2 value appeared higher in trisomic mice as well, the difference being not significant (P=0.203 with a medium ES=0.7; Fig. 2B). All DTI variables did not differ in the two groups of mice (P>0.05, Fig. 2C-F).
Correlation analysis in the whole sample of Ts65Dn mice (n=15) showed statistically significant association between age and body mass (rho=0.74, P=0.01), while age was not correlated with OB volume (rho=0.09, P=0.742). T 2 significantly correlated in a negative way with both age (rho=-0.866, P<0.001) and body mass (rho=-0.635, P=0.011). After adjusting for body mass, the negative association between age and T 2 maps was maintained (rho=-0.862, P<0.001). No significant correlation was found between DTI variables and age or body mass.

DISCUSSION
In this work, we explored by means of MRI some structural and functional characteristics of the OB in trisomic Ts65Dn mice and compared them with euploid individuals of the same strain.
To the best of our knowledge, OB volume in DS has not been investigated in vivo. In this work, we found that OB volume is significantly lower in trisomic mice with a large ES. This is of special interest insofar over-all brain size/volume is similar in trisomic and euploid Ts65Dn mice (Davisson et al., 1993;Galdzicki et al., 2001;Aldridge et al., 2007;Roubertoux et al., 2017). Intriguingly, the trisomic OB was about 87% the euploid, a reduction superimposable to that found in the cerebellum (about 88%; Baxter et al., 2000). This suggests that trisomy especially affects the OB, making this brain structure of particular interest for investigating the neurofunctional correlates of DS. In humans, OB volume is associated with olfactory function and decreases with age (Buschhutter et al., 2008). Noticeably, DS is characterized by accelerated aging and the same has been found in the Ts65Dn mouse (Vacano et al., 2012). Given the olfactory deficit in DS, it could be hypothesized that reduced OB volume in the Ts65Dn mouse is related to both accelerated aging and reduced olfactory function. This hypothesis is supported by data showing olfactory impairment in the adult Ts65Dn mouse (Guidi et al., 2017). Interestingly, reduced OB volume is also found in neurological diseases like migraine (Aktürk et al., 2019), depression (Rottstädt et al., 2018) and Parkinson's disease (Li et al., 2016), as well as in non-neurological conditions like rheumatoid arthritis (Sayilir et al., 2019) and fibromyalgia (Sayilir et al., 2017), suggesting that OB volume may be related to a variety of pathological stimuli. Further, cell cycle alteration and decreased cell proliferation has been found in brain regions different from the OB in the Ts65Dn mouse (Contestabile et al., 2007), which could contribute to reduction in OB volume. This prompts for future in vivo and ex vivo investigations linking OB volume and structural changes therein.
T 2 maps are able to provide information on the status of water in the brain (Eis et al, 1995). Longer T 2 maps values have been associated with edema (Ellingson et al., 2015), microstructural brain damage (Weissenborn et al., 2013), and myelination disorders (Lee et al., 1998;Ding et al., 2008). We herein found that mean Acta Neurobiol Exp 2020, 80: 375-380 Fig. 1. Representative axial MRI pictures of Ts65Dn mouse head. A-D: euploid mouse. E-H: trisomic mouse. In panels A and E, the olfactory bulb and some other regional structures are indicated. The framed area in panels B and F was used for olfactory bulb morphometry. Panels C and G show T2 maps. Panels D and H are fractional anisotropy images. absolute T 2 maps value is about 8% higher in the OB of trisomic Ts65Dn mice. The difference was not statistically significant, probably due to the limited number of mice in analysis; it was, however, associated with a medium ES (0.7). ES is considered a finding of sub-stantive significance (Cohen, 1988) and is independent of sample size. As a suggestion for future studies, we indicate that changes in the brain parenchyma may be to some extent present therein, albeit the difference in T 2 maps values of trisomic vs. euploid mice was not statistically significant. The finding of similar DTI variables values in euploid and trisomic mice suggests that edema, instead of changes in tissue organization and/or myelin ensheathing of axons had taken place in OB. It would be expected that, in the presence of edema, consensual changes in T 2 maps value and ADC take place since ADC (also called mean diffusivity) is direction-independent, reflecting the total amount of water diffusivity in a voxel. Freedom of water diffusion and lower tissue density contribute to higher ADC (Assaf et al., 2008). Nevertheless, ADC values were similar in the two groups of mice. However, it should be underlined that T 2 maps have a better signal-to-noise ratio than DTI measurements and accordingly show increased sensitivity (Petiet et al., 2016). This may explain such a discrepancy. Further work in a larger number of mice will better characterize DTI pattern in the OB of the Ts65Dn mouse, with the aim at identifying more subtle changes therein. In fact, DTI is considered a microstructural probe operating on the micron scale (Basser et al., 2002), able to evaluate the structural integrity of nervous tissue, showing high concordance with the pathways defined in histological tracer-injection studies (Dyrby et al., 2011). In particular, fractional anisotropy (FA) is a measure of the relative diffusion along versus across fiber tracts, 0 value meaning completely isotropic diffusion and values close to 1 representing diffusion restricted to one direction (Westlin et al., 2002). FA decrease is typically dependent on the number and density of axons and is associated with neurodegenerative changes. AD is also related to axonal damage and fragmentation (Beaulieu et al., 1994). Accordingly, FA and AD are considered as markers of axonal damage. RD is related to the radial diffusivity perpendicular to white matter fascicles, hence changes in RD values are representative of axonal density and diameter, and myelination; RD is considered a marker of myelin damage.
In the whole sample of Ts65Dn mice (n=15), the finding of a negative correlation between age and T 2 of the olfactory bulb is supported by similar findings in the gray matter of normal aging brain in both humans and other species (Dhenain et al., 1997;Falangola et al., 2007;Siemonsen et al., 2008;Kumar et al., 2011;Callaghan et al., 2014). This is possibly explained by hemosiderin deposition in gray matter during the aging process.

CONCLUSIONS
Overall, the current findings show that nervous tissue alterations are detectable in vivo in the OB of the Ts65Dm mouse, an animal model of DS, which are possibly related to the olfactory deficit previously found in these mice. While results need to be confirmed in a larger number of mice and corroborated with ex-vivo morphological techniques, it is suggested that quantitative MRI is a valuable tool to identify trisomy-associated characteristics in mouse model of DS. Being non-invasive, MRI can be used for repeated measurement in the same individual, thereby allowing for longitudinal studies of the OB in DS mice.