Striking Differences in the Neuroanatomical Phenotype of the Neuroligin3 R451C Knock-in and the Neurexin1α Knock-out

Background: Neurexin and neuroligins are synaptic cell adhesion genes that have been independently associated with autism (Jamain et al., 2003; Kim et al., 2008).  Neurexins are found on the pre-synaptic side and they bind to neuroligin on the post synaptic side.  It is thought that alterations in either neuroligins or neurexins could alter the excitatory/inhibitory balance and possibly cause the autistic symptoms (Etherton et al., 2009, 2011).

Objectives:   To compare and contrast the volumetric differences of the NL3 R451C knock-in (NL3 KI) mouse with the neurexin1α (NRXN1α) knock-out.

Methods: In total, 48 fixed mouse brains were examined.  Sixteen of which were NL3 KI, 8 wild-type (WT, B6/129F2) and 8 NL3 KI. The other 32 were NRXN1α, 10 WT (B6/SV129), 13 NRXN1α (+/-), and 9 NRXN1α (-/-). The NL3 KI mice were 15 weeks and the NRXN1α mice were 11-13 weeks old.

MRI Acquisition – A multi-channel 7.0 Tesla MRI scanner was used to acquire images of the brain. To acquire images for an anatomical analysis, a T2-weighted, 3-D fast spin-echo sequence was used. This sequence yielded an image with 56 μm isotropic voxels (3D pixel) ~12 h (Lerch et al. 2011).

Data Analysis – To visualize and compare any differences the images from each group are registered together. The goal of the registration is to model how the deformation fields relate to genotype (Lerch et al., 2008). Volume differences are then calculated either in individual voxels or for 62 different regions in both groups (Dorr et al. 2008). Multiple comparisons were controlled for using the False Discovery Rate (FDR) (Genovese et al., 2002).

Results:   Fourteen significant regional differences were found in both the NL3 KI and the NRXN1α (-/-), but the differences were in opposing directions, with the regions in the NL3 KI smaller, and in the NRXN1α larger.  Several large white matter structures were affected indicative of structural connectivity differences between groups. One of the largest differences in the NL3 KI is the 12% decrease in the hippocampus, and this is seen clearly in Figure 1.  Overall the hippocampus was not affected in the NRXN1α model, although it was larger (+3%, FDR of 16%) and there were a few voxel-wise increases found (Figure 1E). In addition to the neuroanatomical differences in the hippocampus, the electrophysiology is quite different (Etherton et al., 2009, 2011), with the NL3 KI showing an enhancement of excitatory synaptic transmission, and the NRXN1α showing a reduction in spontaneous excitatory transmission.  Differences are also found behaviourally between the two models, with NL3 KI mice showing social deficits that are not seen in the NRXN1α.

Conclusions:   One might hypothesize due to their close connection that the neuroanatomical phenotype of the NL3 KI and NRXN1α model may be similar, but they are quite different. It is clear that multiple factors can lead to an autistic phenotype and therefore care must be taken in assessing genetic manipulations involving similar genes or genetic pathways.