Spatial memory training induces morphological changes detected by manganese-enhanced MRI in the hippocampal CA3 mossy fiber terminal zone

Highlights

• Translaminar sprouting of hippocampal mossy fiber after learning observed in Wistar but not LH rat.

• Consistent change of CA3a area detected by MEMRI compared to swim control.

• MEMRI enhancement correlated with Timm's staining of mossy fiber.

• We demonstrated the potential of imaging plasticity in neuro-architectures.

Abstract

Hippocampal mossy fibers (MFs) can show plasticity of their axon terminal arbor consequent to learning a spatial memory task. Such plasticity is seen as translaminar sprouting from the stratum lucidum (SL) of CA3 into the stratum pyramidale (SP) and the stratum oriens (SO). However, the functional role of this presynaptic remodeling is still obscure. In vivo imaging that allows longitudinal observation of such remodeling could provide a deeper understanding of this presynaptic growth phenomenon as it occurs over time. Here we used manganese-enhanced magnetic resonance imaging (MEMRI), which shows a high-contrast area that co-localizes with the MFs. This technique was applied in the detection of learning-induced MF plasticity in two strains of rats. Quantitative analysis of a series of sections in the rostral dorsal hippocampus showed increases in the CA3a′ area in MEMRI of trained Wistar rats consistent with the increased SO + SP area seen in the Timm's staining. MF plasticity was not seen in the trained Lister-Hooded rats in either MEMRI or in Timm's staining. This indicates the potential of MEMRI for revealing neuro-architectures and plasticity of the hippocampal MF system in vivo in longitudinal studies.

Introduction

The hippocampus plays a central role in memory formation. It is composed of three types of principal neurons: granule cells (GCs) of the dentate gyrus (DG), CA3 pyramidal cells, and CA1 pyramidal neurons. These cells, together with the interconnected glutamatergic synapses, form the classical trisynaptic hippocampal circuits. The hippocampal mossy fibers (MFs) pathway comprises the second synapse of the circuit. Axons of the hippocampal MFs arise from GCs in the DG and project to neurons in the hilus and CA3 area, providing synaptic inputs to the proximal apical dendrites of CA3 pyramidal cells. In this trisynaptic model of the hippocampus, the MF pathway provides a strong excitatory input to CA3 pyramidal cells (Blackstad and Kjaerheim, 1961, Andersen et al., 1966, Andersen et al., 1971). In rodents, each of the GCs in the DG gives rise to a single MF axon and the main MF axon forms fine collaterals to provide input to the polymorphic neurons of the hilus and pass through the CA3 area in a narrow band which is called the stratum lucidum (SL), corresponding to the proximal (~ 100 μm) apical dendrites of CA3 pyramidal cells (Henze et al., 2000). There are a large number of active zones and associated post-synaptic densities in the MF synaptic complex (Acsady et al., 1998, Chicurel and Harris, 1992). Owing to its unique and basic structure, the hippocampal MF pathway is usually considered as a good model for investigations into the links between presynaptic axonal readjustments and spatial learning.

MF sprouting has been observed following periods of strong activity, such as after epilepsy as induced by kainite injection (Represa and Ben-Ari, 1992a, Wuarin and Dudek, 1996) or after kindling (Represa and Ben-Ari, 1992b, Van-der-Zee et al., 1995). Beyond diseases, it has also been reported that LTP induced by high-frequency stimulation causes MF synaptogenesis (Adams et al., 1997, Escobar et al., 1997). Moreover, there is strong evidence that hippocampal MF synaptogenesis is associated with spatial long-term memory after overtraining in the Morris water maze. Overtraining in a hidden platform task leads to the growth of hippocampal granule cell MF terminal fields from the SL of CA3 into the stratum oriens (SO) and the stratum pyramidale (SP). These terminal fields have a spatial extent of about 100 μm approximately 7 days after the last day of training in the Wistar rats and 1 day after training in the Long Evans rats (Ramirez-Amaya et al., 1999, Ramirez-Amaya et al., 2001, Holahan et al., 2006, Holahan et al., 2007). In these studies, Timm's staining of zinc was used to reveal the highly zinc-concentrated MF buttons in the brain (Frederickson et al., 1983). Therefore, Timm's-stained granule cells correspond to the morphology of MFs. However, the sprouting of MFs is a time dependent process; thus, the in vivo method is more desirable for detecting activity dependent plasticity to understand its functional role in the long-term memory.

Magnetic resonance imaging (MRI) has been widely used to visualize anatomical and functional characteristics of the brain in vivo in both animal and human studies based on intrinsic longitudinal and transverse relaxation times (T₁ and T₂, respectively) of tissues (Budinger et al., 1999). In the hippocampus, although T₂ and diffusion changes can be seen after kainic acid-induced pathology, the T₁ and T₂ of the cellular layers and MFs are normally indistinguishable (Hsu et al., 2007). Therefore, the neuroplastic changes of MFs cannot be differentiated using conventional MRI (Watanabe et al., 2004). Recently, manganese-enhanced MRI (MEMRI) has been increasingly used for visualizing structure, function, and neuronal circuits in the brain at a high spatial resolution (for review, see Silva et al., 2004, Silva and Bock, 2008). Mn²⁺ ion is a Ca²⁺ analog and can enter cells via voltage-gated Ca²⁺ channels and other transport mechanisms for Ca²⁺ (Crossgrove and Zheng, 2004). Since Mn²⁺ is a paramagnetic substance, its presence can be visualized by T₁-weighted MRI. MEMRI has been applied to the imaging of the cytoarchitecture of the brain (Aoki et al., 2004, Watanabe et al., 2004, Silva et al., 2008, Chuang et al., 2010), specific neuronal pathways (Pautler et al., 1998, Leergaard et al., 2003, Chuang and Koretsky, 2006, Tucciarone et al., 2009), and activated neural areas (Lin and Koretsky, 1997, Yu et al., 2005, Chuang et al., 2009). Large-scale plasticity of neural structure has been studied in the vocal control center of the songbird (Van der Linden et al., 2004, Van Meir et al., 2006), in the auditory midbrain (Yu et al., 2007), after unconditioned fear (Chen et al., 2007), and in conditioned taste aversion (Inui-Yamamoto et al., 2010). Since MEMRI shows preferential enhancement of the subregions in the hippocampus (Aoki et al., 2004, Watanabe et al., 2004), whether this particular enhancement can reflect structural and functional changes of the MF plasticity in vivo has been tested in epilepsy models (Nairismägi et al., 2006, Alvestad et al., 2007, Immonen et al., 2008, Kuo et al., 2008). It was shown that the Mn²⁺-enhanced area in the DG and CA3 of rats increased with kainic acid injection and correlated with histologically verified MF sprouting (Nairismägi et al., 2006). In another study, MEMRI signal changes were attributed to increased axonal density of MFs rather than to neurodegeneration, astrogliosis, or microgliosis (Immonen et al, 2008). These studies suggest that MEMRI could potentially be useful for the observation of MF plasticity in vivo. However, compared to the large-scale changes seen in epilepsy, the MF remodeling after maze learning is much smaller and poses a particular challenge for imaging.

In this study, we investigated the use of MEMRI for the in vivo detection of MF remodeling in spatial memory formation after maze learning. After the rats received 5 days of training to locate a hidden platform in the Morris water maze, MnCl₂ was injected intravenously on the 6th day post-training to enhance the neural cytoarchitecture. High-resolution three-dimensional (3D) MRI was acquired on the 7th day to measure the lamellar organization of the MFs. In particular, the differences between the strain dependent remodeling of the MF system was examined using Wistar and Lister-Hooded (LH) rats. The results show an increased MEMRI area in the SO of the Wistar rats but not the LH rats. Consistent with Timm's staining, the learning-induced plasticity was observed in vivo.