Patterned optogenetic modulation of neurovascular and metabolic signals

Abstract

The hemodynamic and metabolic response of the cortex depends spatially and temporally on the activity of multiple cell types. Optogenetics enables specific cell types to be modulated with high temporal precision and is therefore an emerging method for studying neurovascular and neurometabolic coupling. Going beyond temporal investigations, we developed a microprojection system to apply spatial photostimulus patterns in vivo. We monitored vascular and metabolic fluorescence signals after photostimulation in Thy1-channelrhodopsin-2 mice. Cerebral arteries increased in diameter rapidly after photostimulation, while nearby veins showed a slower smaller response. The amplitude of the arterial response was depended on the area of cortex stimulated. The fluorescence signal emitted at 450/100 nm and excited with ultraviolet is indicative of reduced nicotinamide adenine dinucleotide, an endogenous fluorescent enzyme involved in glycolysis and the citric acid cycle. This fluorescence signal decreased quickly and transiently after optogenetic stimulation, suggesting that glucose metabolism is tightly locked to optogenetic stimulation. To verify optogenetic stimulation of the cortex, we used a transparent substrate microelectrode array to map cortical potentials resulting from optogenetic stimulation. Spatial optogenetic stimulation is a new tool for studying neurovascular and neurometabolic coupling.

Figure 1

Schematic of the microprojection and fluorescence imaging system. A high-powered laser is reflected off a digital micromirror device (DMD) and combined into an epifluorescence light path. This optical system is automated with a data acquisition (DAQ) card, which also receives real-time information from the electrophysiology system's digital signal processor.

Figure 2

Arterial vasodilation depends on area of optogenetic stimulation. Large (860 × 860 μm), medium (570 × 570 μm), and small (320 × 320 μm) photostimulus squares were applied (A to C, overlaid in blue). Each of these photostimulus patterns were applied as a pulse train with the irradiance held constant (20 ms pulsewidth at 25 Hz for 1 second, 4.5 mW/mm2, and 445 nm). The average prestimulus image is shown in gray with poststimulus percentage of change shown in red (A to C). The large (A) and medium (B) photostimuli caused the middle cerebral artery branches to dilate, while the veins showed little response. (D) Based on anatomic reference, branches of the middle cerebral artery, which originate lateral and anterior were identified as a1 and a2, while medially originating veins were labeled v1 and v2. (E) Time-series diameter quantification of the middle cerebral artery (MCA) branches (a1 and a2) showed a 30% increase after the large area photostimulus, while the medium area photostimulus caused a 15% diameter increase, and the small area photostimulus caused little to no vasodilation. In comparison to the MCA, the venous response (v1 and v2) was much smaller. Seven trials of each stimulus area were conducted per animal. (F) The study was replicated in four ChR2-positive mice and four wild-type (WT) mice, and the median percentage change of the middle branch of the MCA was significantly larger in ChR2 mice than in WT mice at the α=0.05 confidence level (one-tailed signed rank test).

Figure 3

Projection avoiding major blood vessels comparison. We projected either a solid square (A) or pattern with the blood vessels omitted (B). The other photostimulation parameters (20 ms pulsewidth at 25 Hz for 1 second, 4.5 mW/mm2, and 445 nm) were held constant. The vessels were measure at the sites labeled in (C). The hemodynamic response (D) was similar for the solid and blood vessel avoiding stimulus patterns. This experiment helps test for the effect of directly photostimulating the blood vessels and their associated perivascular nerves.

Figure 4

Transient metabolic changes after channelrhodopsin-2 (ChR2) stimulation. Exciting at 365 nm and imaging at 450/100 nm, we recorded this fluorescent signal before and after ChR2 stimulation (5 ms pulsewidth at 66 Hz for 1 second, 445 nm, and 4.5 mW/mm²) of different cortical areas. The large area stimulus (D) caused a correspondingly large decrease (F) in the fluorescence intensity of the region of interest (ROI) displayed in (E). The smaller area photostimulus (A) caused a smaller NADH (nicotinamide adenine dinucleotide) response (C) in the region of interest (ROI) displayed in (B). Ten trials were conducted at each location per animal. The study was replicated in three ChR2 mice and three wild-type (WT) mice with the peak fluorescence change for the large photostimulus plotted in (G). A P value of 0.05 was found for a one-tailed signed rank test.

Figure 5

Optogenetically evoked potentials colocalized with the photostimulus location. We implanted microfabricated electrocorticography electrode arrays under a cranial window in Thy1-ChR2/H134R mice (A and D). Photostimulation of small areas (A and D) caused optogenetically evoked potentials (B and E, respectively) that were largest at sites nearest the stimulus. The stimulus was easily localized with cubic spline interpolation (C and F) of the peak potentials. Nonfunctional electrode sites (impedance >5 M Ω at 1 kHz) were omitted from the analysis.