Neurofibrillary tangle-bearing neurons are functionally integrated in cortical circuits in vivo

Alzheimer''s disease (AD) is pathologically characterized by the deposition of extracellular amyloid-β plaques and intracellular aggregation of tau protein in neurofibrillary tangles (NFTs) (1, 2). Progression of NFT pathology is closely correlated with both increased neurodegeneration and cognitive decline in AD (3) and other tauopathies, such as frontotemporal dementia (4, 5). The assumption that mislocalization of tau into the somatodendritic compartment (6) and accumulation of fibrillar aggregates in NFTs mediates neurodegeneration underlies most current therapeutic strategies aimed at preventing NFT formation or disrupting existing NFTs (7, 8). Although several disease-associated mutations cause both aggregation of tau and neurodegeneration, whether NFTs per se contribute to neuronal and network dysfunction in vivo is unknown (9). Here we used awake in vivo two-photon calcium imaging to monitor neuronal function in adult rTg4510 mice that overexpress a human mutant form of tau (P301L) and develop cortical NFTs by the age of 7–8 mo (10). Unexpectedly, NFT-bearing neurons in the visual cortex appeared to be completely functionally intact, to be capable of integrating dendritic inputs and effectively encoding orientation and direction selectivity, and to have a stable baseline resting calcium level. These results suggest a reevaluation of the common assumption that insoluble tau aggregates are sufficient to disrupt neuronal function.Neurofibrillary tangles (NFTs) containing aggregated tau protein (1) have long been considered key players in the progressive neural dysfunction and neurodegeneration observed in Alzheimer’s disease (AD) (2, 3) and other tauopathies (4, 5). It is commonly assumed that NFT-bearing neurons exhibit deficits in synaptic integration and eventually lead to neurodegeneration (11, 12). However, the actual functional properties of NFT-bearing neurons in intact neural circuits have not been explored previously (13). We addressed this question directly using awake in vivo two-photon calcium imaging in a mouse model of NFT formation (rTg4510) by applying recently developed imaging approaches allowing for single-neuron–level and population-level assessment of neural activity in awake mice (14). Because two-photon calcium imaging allows for measurement of response properties in many neurons simultaneously, we were able to directly isolate the impact of NFT deposition in a neuronal microcircuit by evaluating population-level network dynamics and, more specifically, by differentiating the function of individual NFT-bearing and neighboring non–NFT-bearing neurons.To assess the functional properties of neurons in the visual cortex, we used a genetically encoded ratiometric calcium indicator, yellow cameleon 3.6 (YC3.6), packaged in an adeno-associated viral vector (15, 16). To assess functional responses, we exploited the well-characterized functional architecture of visual cortex whereby neurons in mouse visual cortex modulate their activity during presentation of drifting gratings moving at specific orientations and directions (orientation and direction selectivity) (17, 18). We further used YC3.6 as a FRET-based ratiometric indicator to make quantitative measurements of resting calcium (15). Resting calcium is tightly regulated in the brain, and slight deviations can trigger chronic and severe degenerative pathways (15). Thus, measurement of resting calcium is an important and complementary functional assay for evaluating neuronal health. Importantly, performing experiments in awake, head-fixed animals eliminates the impact of anesthesia on response properties, resting calcium, and tau aggregation (Fig. 1 A and B and Movie S1) (19).Open in a separate windowFig. 1.YC3.6 calcium imaging in awake cortex of control mice reveals robust visual response tuning. (A) Awake and head-fixed mice expressing AAV-YC3.6 in visual cortex were presented with drifting gratings to record visually evoked calcium responses through a chronic cranial window (Movie S1). (B) In vivo two-photon image of YC3.6-expressing neurons in layer 2/3 (∼180 µm below the brain surface) with four example neurons (white circles). (Scale bar: 50 μm.) (C) Drifting gratings (eight stimulus types, 45° apart) were presented for 7–10 s, followed by 7–10 s without stimulation (gray screen). This sequence was repeated 10 times. For cell 1 (orientation- and direction-selective), three representative single-trial time courses (YFP:CFP ratio) and the average trace across 10 trials reveal its predominant response to a grating direction of 225°. SEM is shown in gray; arrowhead indicates preferred direction. (D) Polar plot of cell 1 demonstrating selectivity of this neuron for the 225° stimuli. (E) Average traces of three other neurons (white circles in A) illustrate the diversity in orientation and direction tuning. Cell 2 is orientation-selective but not direction-selective, cell 3 responds to all directions (broadly tuned), and cell 4 does not respond to any direction (not responding). (F) OSI and DSI for cells 1–4.