Thy1-GCaMP6 transgenic mice for neuronal population imaging in vivo

Abstract

Genetically-encoded calcium indicators (GECIs) facilitate imaging activity of genetically defined neuronal populations in vivo. The high intracellular GECI concentrations required for in vivo imaging are usually achieved by viral gene transfer using adeno-associated viruses. Transgenic expression of GECIs promises important advantages, including homogeneous, repeatable, and stable expression without the need for invasive virus injections. Here we present the generation and characterization of transgenic mice expressing the GECIs GCaMP6s or GCaMP6f under the Thy1 promoter. We quantified GCaMP6 expression across brain regions and neurons and compared to other transgenic mice and AAV-mediated expression. We tested three mouse lines for imaging in the visual cortex in vivo and compared their performance to mice injected with AAV expressing GCaMP6. Furthermore, we show that GCaMP6 Thy1 transgenic mice are useful for long-term, high-sensitivity imaging in behaving mice.

Figure 1. Thy1 transgenic mice expressing GCaMP6s or GCaMP6f.

a. Schematic of the transgene cassettes used to generate GP4.x (top) and GP5.x (bottom) lines. WPRE = Woodchuck hepatitis virus post-transcriptional regulatory element, pA = poly-adenylation tail. b. Wide-field images of coronal sections showing GCaMP fluorescence in various transgenic lines. c. Representative confocal images (tiled and stitched to show larger field of view) from the somatosensory cortex of the same lines as in b. All images show GCaMP6 fluorescence.

Figure 2. Stable GCaMP expression in GP mice over months.

a. Confocal microscope images of fixed coronal sections from the motor cortex of GP5.17 mice at different ages (F1- first generation, F2- second generation). b. Somatic GCaMP6f brightness for all neurons inside the white rectangles in a (51–65 neurons from each animal). For each mouse, the box indicates the 25^th to 75^th percentile distribution, red line indicates the median, and whisker length is 150% of the 25^th to 75^th percentile distance, or until it touches the last sample position. Outliers are marked in red crosses. c. In vivo two-photon microscopy images of GP4.3 mice taken at different days after cranial window implantation show similar expression pattern without filled nuclei. Arrowheads point to three individual cells in both images.

Figure 3. Quantification of GCaMP expression.

a. Demonstration of the analysis method used for calculating single-neuron brightness distribution across brain regions. Confocal microscopy images of fixed brain slices were used for segmentation into cell bodies (red rings, nuclei were excluded from somata). Somatic brightness was calculated by averaging all pixels in each segmented cell. b–d. Neuronal (somatic) GCaMP6 brightness of labeled neurons in various transgenic and AAV infected mice. NIH line 10 is a Thy1-transgenic mouse line expressing GCaMP3 . Ai38 is a Cre-dependent reporter mouse expressing GCaMP3 in the ROSA26 locus , here injected with synapsin1-Cre AAV. Each box indicates the 25^th to 75^th percentile distribution with different colors for each brain region, red line indicates the median, and whisker length is 150% of the 25^th to 75^th percentile distance, or until it touches the last sample position. Outliers are marked in red crosses. b, L2/3 pyramidal cells (86–289 cells per line; median, 181). c, L5 pyramidal cells (45–230 cells per line; median, 148). d, Hippocampal pyramidal cells (25–342 cells per line; median, 113). e. Confocal image of GP5.17 fixed tissue (green) counterstained with NeuN (red). f–h. Fraction of neurons that are GCaMP6-positive, estimated by counterstaining with NeuN, corresponding to b–d, respectively (54–186 cells per line; median, 102).

Figure 4. Functional imaging in the visual cortex (V1) of transgenic and AAV infected mice.

a. Schematic of the experimental setup. b. Responses of three GP5.17 example cells to eight oriented grating stimuli. c. Responses of three GP4.12 example cells. d. Mean ΔF/F₀ responses to the preferred stimulus for all cells with peak ΔF/F₀>1. Cells were aligned according to their response maximum to one of four time points (1, 2, 3, or 4 s), and each stimulus lasts 4 s (average of 91 cells for GP5.17, 124 cells for GP4.3, 362 cells for GP4.12, 83 cells for AAV-6f, and 224 cells for AAV-6s). e. Fourier transform of the response to the preferred stimulus (median across cells). The 1 Hz peak corresponds to the frequency of the drifting grating. f. Half-decay time (mean±s.d.) after the last response peak during stimulus presentation (n = 52, 75, 446, 136, and 235 cells for GP5.17, GP4.3, GP4.12, AAV-6f, and AAV-6s respectively). g. Distribution of ΔF/F₀ responses to the preferred stimulus. h. Fraction of statistically significant responsive cells (mean±s.d., n = 731,1130, 1325, 871, and 672 cells, for GP5.17, GP4.3, GP4.12, AAV-6f, and AAV-6s respectively).

Figure 5. ALM functional imaging using GP4.3 during behavior.

a. Schematic of the object localization behavior (see ref. for details). b. An image of GCaMP6s labeled neurons in a GP4.3 mouse, 160 µm below the pia. Blue and red colors indicate cells that responded during lick-right and lick-left trials, respectively. Cue = auditory “go” signal. c. Fluorescent responses of three example cells indicated in b. d. Top, single-trial responses of the same neurons in b sorted according to trial type (blue: lick-right, red: lick-left). Bottom, trial-averaged response. e. The responses of cell 3 measured over 4 behavioral sessions spanning more than one month. f. Peak normalized fluorescent transients of GCaMP expressing neurons in GP5.17, GP4.3 and AAV-GCaMP6s injected mice (0.2±0.1 s, 0.65±0.35 s, and 1.8±1.1s mean±s.d; n = 207, 771, and 369 cells for GP5.17, GP4.3, and AAV-6s respectively).