1. Generation of human wild-type and P56S VAPB transgenic mice We have obtained three independent lines of WT (A6, B3, and B7) and P56S (C8, C11, and D3) VAPB transgenic (Tg) mice, respectively. More than a 20-fold over-expression of VAPB mRNA was observed in the brain of line B3 of WT (25.6 1.9 fold) and line D3 of P56S VAPB Tg mice (23.7 0.7 fold) as compared to littermate non-transgenic (nTg) controls. However, the steady level of P56S VAPB protein (7-fold vs. endogenous) was significantly less compared to WT VAPB protein (20-fold vs. endogenous) in whole brain lysate, indicating that the P56S mutation might impair stability and solubility of VAPB protein in vivo. We later refer to these two highest VAPB expression lines as WT and P56S VAPB Tg mice in the downstream behavioral and pathological studies. To examine the cellular and subcellular distribution of VAPB, we performed immunohistochemistry with a VAPB antibody, and found that endogenous VAPB protein is highly expressed by spinal motor neurons (Fig. 1D) and is also abundant in corticospinal motor neurons. Over-expression of WT VAPB led to a diffuse cytosolic distribution pattern of VAPB protein in neurons of WT VAPB Tg mice. By contrast, widespread large punctate staining of VAPB was observed in neurons of P56S VAPB Tg mice. This apparent abnormal clustering of P56S VAPB protein may potentially affect the function and survival of neurons. 2. P56S VAPB Tg mice developed abnormal motor behavioral phenotypes A cohort of 39 male mice (15 for nTg, 12 for WT, and 12 for P56S VAPB) were closely monitored for motor and other behavioral phenotypes. Both P56S and WT VAPB Tg mice were viable, developed normally, and lived a normal life span. However, the gain of body weight of P56S VAPB mice was significantly less compared to littermate nTg and WT VAPB mice after 15 months of age. P56S VAPB mice unexpectedly started to exhibit significant hyperactivity in both horizontal and vertical movements in the Open-field test at 12 months of age. In addition, Rotarod test revealed a significant impairment of motor coordination/balance in P56S VAPB mice beginning at 12 months of age, too. Since gait abnormalities are an early sign of motor dysfunction in other mouse models of motor neuron degeneration, we quantified the gait parameters of P56S VAPB mice using the Treadscan Gait Analysis system. Interestingly, both the stride length and stride time of P56S VAPB mice were significantly shorter compared to littermate nTg controls and WT Tg mice. The shorter stride of P56S VAPB Tg mice was first detected at two months of age and was persistent without significant deterioration through the rest of their lives. A similar shortening of stride length was also observed in line C11 of P56S VAPB Tg mice; which however, displayed normal locomotor activities. By contrast, no significant difference in the stride length was found between WT VPAB and control nTg mice. The early onset and non-progressive alteration of gait parameters in P56S VAPB mice indicates P56S VAPB may affect the development of neural circuitry regulating the gait properties of mice. 3. P56S VAPB Tg mice developed progressive degeneration of corticospinal motor neurons Given that the P56S mutation in VAPB causes motor neuron degeneration, it seems counterintuitive to discover that P56S VAPB Tg mice developed a progressive hyperactivity. However, a previous report shows that Fez-like (Fezl) knockout mice, which lack the corticospinal motor neurons, also exhibit significant hyperactivity. This phenomenon could be attributed to different wiring or properties of corticospinal motor neurons in human and mouse motor control systems. To explore the pathological basis of hyperactivity in aged P56S VAPB Tg mice, we examined the viability of corticospinal motor neurons in the cerebral cortex of P56S VAPB Tg mice. The corticospinal motor neurons can be easily identified within layerV of the cerebral cortex by immunostaining with an antibody against transcription factor COUP TF1-interacting protein 2 (CTIP2). We found a significant reduction of CTIP2-positive corticospinal motor neurons in the cortex of 18-month old P56S VAPB Tg mice as compared to age-matched WT and control nTg mice (Fig. 3). In addition to the loss of corticospinal motor neurons, a significant increase of GFAP-positive reactive astrocytes was also found mainly in the cerebral cortex of 18-month old P56S mice. Noticeably, no significant degeneration of corticospinal motor neurons or reactive gliosis was detected in the cerebral cortex of C11 line P56S VAPB Tg mice, which showed lower level expression of mutant VAPB in the cortex and only exhibited gait abnormalities. Together, we document a rather selective loss of corticospinal motor neurons in aged P56S VAPB Tg mice, which may contribute to the progressive hyperactivity developed in these mutant animals. 4. No significant degeneration of spinal motor neurons in aged P56S VAPB Tg mice To investigate whether the P56S mutation in VAPB leads to spinal motor neuron degeneration, we counted motor neurons in the lumbar spinal cord of 18-month old P56S and WT VAPB Tg mice as well as nTg littermate controls. The number of spinal motor neurons per section in P56S VAPB Tg mice (11.65 +/- 0.62) was comparable with those in WT VAPB (12.15 0.57, p=0.8) and control nTg mice (13.01+/- 0.77, p=0.6). We also examined the occurrence of reactive gliosis in the spinal cord of 18 month-old P56S VAPB Tg mice. However, we did not observe any significant increase of reactive gliosis in the spinal cord of P56S VAPB Tg mice. 5. The P56S mutation induced a translocation of VAPB protein to the postsynaptic site of C-boutons in spinal motor neurons In contrast to our observations in corticospinal motor neurons, no apparent co-staining of calnexin in VAPB-positive large punctuate structures was observed in the spinal motor neurons of P56S VAPB mice. Instead, we found a small fraction of VAPB-immunoreactivity was juxtaposed with the synaptophysin (SYN) staining in the spinal motor neuron of P56S VAPB Tg mice when we co-stained VAPB with SYN, a marker for presynaptic terminals. We further confirmed the postsynaptic location of the VAPB protein in the spinal motor neuron of P56S VAPB mice by immuno-EM with a VAPB antibody. As controls, no postsynaptic location of VAPB was found in the spinal motor neuron of WT VAPB Tg and control nTg mice. The mutant VAPB-associated synapses were large and restricted to soma and proximal dendrites of spinal motor neurons, which belongs to a special class of synapses, the C-bouton of spinal motor neurons. C-boutons receive cholinergic innervation from a group of cholinergic interneurons near the central canal of spinal cord. The type 2 muscarinic (M2) receptors evenly distributed along the plasma membrane of large spinal motor neurons mediate the postsynaptic response of C-boutons and modulate the excitability of spinal motor neurons. Accordingly, we found the immunoreactivity of VAPB was juxtaposed with choline acetyltransferase (CHAT) staining in the spinal motor neuron of P56S VAPB mice. Furthermore, VAPB protein showed specific co-localization with M2 receptors at the cell surface of spinal motor neurons, indicating a potential gain of function of P56S VAPB in affecting the normal structure and function of C-boutons in spinal motor neurons.
