Highlights
•
The mammillothalamic tract (MTT) supports neural activity in an extended memory system.•
Optogenetic activation of neurons in the anterior thalamus acutely improves memory after MTT lesions.•
Rescued memory associates with system-wide neuronal activation and enhanced EEG.•
Anterior thalamus actively sustains memory and is a feasible therapeutic target.
Abstract
A hippocampal-diencephalic-cortical network supports memory function. The anterior thalamic nuclei (ATN) form a key anatomical hub within this system. Consistent with this, injury to the mammillary body-ATN axis is associated with examples of clinical amnesia. However, there is only limited and indirect support that the output of ATN neurons actively enhances memory. Here, in rats, we first showed that mammillothalamic tract (MTT) lesions caused a persistent impairment in spatial working memory. MTT lesions also reduced rhythmic electrical activity across the memory system. Next, we introduced 8.5 Hz optogenetic theta-burst stimulation of the ATN glutamatergic neurons. The exogenously-triggered, regular pattern of stimulation produced an acute and substantial improvement of spatial working memory in rats with MTT lesions and enhanced rhythmic electrical activity. Neither behaviour nor rhythmic activity was affected by endogenous stimulation derived from the dorsal hippocampus. Analysis of immediate early gene activity, after the rats foraged for food in an open field, showed that exogenously-triggered ATN stimulation also increased Zif268 expression across memory-related structures. These findings provide clear evidence that increased ATN neuronal activity supports memory. They suggest that ATN-focused gene therapy may be feasible to counter clinical amnesia associated with dysfunction in the mammillary body-ATN axis.
1. Introduction
Anatomical, clinical, and experimental lesion evidence suggests that the anterior thalamic nuclei (ATN) form a critical subcortical hub within a hippocampal-diencephalic-cortical network that supports memory function (Harding et al., 2000; Aggleton, 2014; Aggleton et al., 2016; Bubb et al., 2017; Ferguson et al., 2019; Nelson et al., 2020; Nelson, 2021). Participation in a broad memory system is supported by many reports that ATN injury is also associated with metabolic and microstructural dysfunction in cortical and hippocampal structures (Reed et al., 2003; Caulo et al., 2005; Aggleton, 2008; Aggleton and Nelson, 2015; Dalrymple-Alford et al., 2015; Perry et al., 2018). Together, this evidence corroborates the idea that the ATN are responsible for more than the simple transfer of information to other key structures in the memory network (Wolff and Vann, 2019).
The negative impact of ATN lesions does not directly answer the question of whether ATN neurons actively facilitate memory, but there is some support for this claim from human studies. Functional magnetic resonance imaging data suggest that both ATN and mediodorsal thalamus activity is associated with memory in healthy adults (Pergola et al., 2013; Zotev et al., 2018). Memory performance in recovered Wernicke patients correlates with the degree of functional connectivity between the mammillary bodies (MB) and the ATN (Kim et al., 2009). This recovery might reflect information conveyed by the mammillothalamic tract (MTT), which constitutes axons from the MB that project to the ATN. More direct evidence comes from intracranial electrophysiological recordings made in epilepsy patients, which suggests that the ATN integrates information from diverse cortical sources during encoding to guide successful recall and that 50 Hz stimulation of the ATN improves working memory (Liu et al., 2021; Sweeney-Reed et al., 2021). In rats, however, there is mixed evidence for the impact of electrical stimulation of the ATN on memory. Hamani and colleagues reported that high current electrical stimulation at 130 Hz focused on the ATN impaired fear conditioning and nonmatching-to-sample in intact rats when stimulated during training; low current stimulation had no effect (Hamani et al., 2010). Low current stimulation improved delayed (4 min) visual nonmatching-to-sample performance in corticosterone-treated, but not saline-treated, intact rats when memory was tested at remote but not recent time points; this improvement was linked with increased hippocampal neurogenesis (Hamani et al., 2011).
Optogenetics allows for the selective stimulation of specific neuronal populations, whereas electrical stimulation activates all surrounding cells and fibers of passage. If ATN projections to memory structures actively support memory, stimulating these neurons optogenetically should impact the other structures within the distributed memory system. Moreover, we reasoned that selective stimulation could improve impaired memory caused by brain injury such as MTT lesions. MTT injury is the most consistent predictor of diencephalic amnesia due to thalamic infarcts (Carlesimo et al., 2011). In support of this clinical evidence, deficits in spatial working memory and temporal (i.e. recency) memory are found after experimental MTT lesions in rats (Aggleton, 2014; Nelson and Vann, 2017; Perry et al., 2018; Dillingham et al., 2015, 2021). The longevity of the deficit after MTT lesions in rats is uncertain, so we used a demanding 12-arm radial arm maze (RAM) and first confirmed that these lesions produce a long-lasting spatial working memory impairment. The majority of ATN efferents derive from glutamatergic neurons (Zakowski, 2017), so we next used a viral vector to transduce these specific neurons and render them responsive to light stimulation. Medial MB neurons, which project to the AV and the anteromedial subregion (AM) of the ATN, modulate their firing rate at theta frequency (Nelson et al., 2018). Similarly, there are theta-modulated cells in the AV and AM, which fire rhythmically with hippocampal theta rhythm (Jankowski et al., 2013; Zakowski et al., 2017). There is also evidence that the integrity of the MB-ATN axis influences electrophysiology across the extended hippocampal memory network (Dillingham et al., 2019, 2021). During spatial working memory testing, therefore, we used optogenetic theta-burst stimulation (TBS) in the dorsolateral region of the ATN, focused primarily on the AV, and we recorded electrophysiology to determine the theta-range power spectrum density (PSD) within, and coherence across, the ATN-hippocampal-prefrontal cortex axis (ATN-HPC-PFC axis). We compared the effects of optogenetic TBS of the ATN with both no stimulation and control stimulation by an ineffective light wavelength, while rats were tested in the RAM.
The most salient evidence that the MB-ATN axis impacts a network of memory-related brain regions is that MTT and ATN lesions reduce the expression of immediate early gene (IEG) activity in these distal brain structures (Aggleton, 2008; Dillingham et al., 2015; Loukavenko et al., 2016; Perry et al., 2018). We know that projections from the ATN to cortical regions and the hippocampal subiculum are almost exclusively ipsilateral in the rat (Mathiasen et al., 2017). So, we measured the IEG response in distal brain structures when rats received optogenetic stimulation of the ATN using effective (i.e. blue-light) TBS in one hemisphere and ineffective (i.e. orange-light) TBS simultaneously in the contralateral ATN. To avoid confounds related to the recruitment of these areas due to memory demands, we applied this stimulation while rats foraged for food in an open field, conducted 90 min before sacrifice, and removal of the brain to analyse Zif268 expression. Zif268 was selected as this marker is associated with spatial memory formation and long-term plasticity (Jones et al., 2001; Penke et al., 2014; Farina and Commins, 2016; Gallo et al., 2018) and has successfully revealed reduced IEG activity in the extended memory system following both ATN and MTT lesions (Dumont et al., 2012; Frizzati et al., 2016; Perry et al., 2018). Regions of interest for IEG analysis included key structures within the broader memory system that receive direct innervation from the ATN (Aggleton and Brown, 1999; Bubb et al., 2017; Nelson, 2021).
2. Methods
2.1. Animals
The experiment used 27 male Piebald Virol Glaxo cArc hooded rats, bred in the Animal Facility at the University of Canterbury. Rats weighed between 290 and 320 g and were aged 8–9 months at the time of surgery (Fig. 1). Twenty-three rats were randomly allocated to opsin-Sham or opsin-MTT lesion groups and 3 rats were assigned to a non-opsin MTT group. Rats were housed in mixed-condition groups of three or four rats per standard opaque plastic cage (50 cm length, 30 cm wide, and 23 cm high) in a vivarium. The vivarium lights were off between 8 a.m. and 8 p.m., when behavioural testing was conducted; observation of the rats confirmed that they remained relatively inactive during the lights-on period. Following surgery, rats were housed individually for seven to ten days. Food was available ad libitum just before surgery and during recovery. For behavioural testing, rats received restricted food access to maintain 85% of their free-feed body weight. Water was always available. All procedures were approved by the University of Canterbury Animal Ethics Committee, New Zealand (2016/24R) and comply with the ARRIVE guidelines.
