3PEF at 1320-nm excitation wavelength enabled greater imaging depths than 2PEF at 800-nm excitation wavelength in the mouse spinal cord in vivo.

(A) 2PEF and (B) 3PEF image stacks of fluorescently labeled blood vessels in the spinal cord of a live mouse. (C) 2PEF (left column), 3PEF (middle column), and THG images (right column) at selected depths into the mouse spinal cord. (D) Line profiles across selected capillaries, comparing 3PEF (magenta) and 2PEF (blue) image contrast at different depths. Lines in (C) indicate location of lineouts. (E) SBR for 2PEF and 3PEF imaging of the spinal cord vasculature. In the center region of the frame, the SBR was calculated as the average pixel intensity of the upper 1% to the lower 5%. The line show fits to exponential decays, with the characteristic attenuation length (CAL) of the fit indicated.

Representative images of fluorescein-labeled vasculature stack under 2PEF at 800-nm versus 3PEF at 1320-nm excitation source.

(A) 2PEF (left column), 3PEF (middle column), and THG images (right column) at selected depths into the mouse spinal cord. Scale bar: 100 μm. (B) Line profiles across selected capillaries, comparing 3PEF (magenta) and 2PEF (blue) image contrast at different depths. Lines in (A) indicate location of lineouts. These data include the three depths shown in Figs. 1C and D and add three additional depths. Scale bar: 100 μm.

Chronic 3PEF measurement of spinal blood flow from surface venules to deep arterioles in vivo.

(A) 4x magnification image taken from the dorsal aspect of the mouse spinal cord using 2PEF at 800 nm; spinal cord vasculature was labeled with i.v. injection of FITC-dextran dye (dSV: dorsal spinal vein; dAV: dorsal ascending venule). (B) Volumetric reconstruction of a ~500-μm deep stack of spinal vascular architecture using 3PEF at 1320-nm. (C) Representative images of vascular features at different depths into the spinal cord. (D) Schematic representation defining vascular branch order. (E) Selected vessel segments and (F) corresponding linescan profile from vessels that were one, four, and seven branches upstream from the dAV. (G) RBC speed in arterioles, capillaries (vessel diameter < 10 μm) and venules, expressed as a function of vessel diameter with diameter decreasing in arterioles then increasing in venules from left to right on the graph (n = 166 vessels from 17 mice). Capillaries were assigned to the arteriole or venule side based on their closest topological proximity to an arteriole or venule. (H) RBC speed of arterioles as a function of vascular branch order starting from dLA and going downstream. (I) RBC speed of venules as a function of vascular branch order starting from dAV and going upstream (n = 6 mice).

Representative images demonstrating spinal cord vasculature, including dAV at the surface and dLA at depth.

Selected images from a continuous vasculature stack indicate the dAV at the top 50 μm followed by a dense capillary network spanning across the spinal cord tissue column, which is fed by dLA that runs along the rostral-caudal axis of the mouse spinal cord.

3PEF at 1320-nm enabled multicolor imaging of diverse cellular events in response to a local injury in vivo.

(A) Experiment timeline and schematics for 3PEF imaging of neural dieback and inflammatory response following a photothrombotic venule occlusion. (B) 3D reconstruction of a z-stack taken from a Thy1-YFP x Cx3Cr1-GFP double transgenic mouse spinal cord; vessels were labeled with i.v. injection of QtrackerTM 655 red-emitting quantum dots. Arrow points to the occluded vessel segment. (C) A cessation of blood flow after a green laser-induced photothrombotic occlusion was followed by rapid axon dieback (yellow, seen via 3PEF) and myelin degeneration (grey, seen via THG) nearby the targeted vessel segment. (D) Time-series of representative images across different laminar depths showing progressive neural dieback, inflammatory response, and capillary disruption after a surface venule occlusion.

Rapid neurite dieback after a surface venule occlusion.

(A) Time-series of representative Thy1-YFP labeled neural structure dynamics from intact, swollen, broken to disappeared across different depths after a photothrombotic stroke. Stacked bar graphs showing the proportion of neural structures that were intact, swollen, broken, or disappeared at different time points from depths of (B) 0-100 μm, (C) 100-200 μm, (D) 200-300 μm, and (E) >300 μm.

Perivascular Cx3Cr1+ cells marched towards vessel lumens, invaded and disrupted small vessels under ischemic conditions.

(A) Perivascular Cx3Cr1+ cells marched towards vessel walls and extended processes, infiltrating into the vessel lumen (white arrows) and disrupting blood flow after a photothrombotic stroke in an upstream dAV. (B) Summary chart of the behavior of 43 perivascular Cx3Cr1+ cells. Perivascular microglia behaviors at depths of (C) 0-150 μm, (D) 150-200 μm, (E) 200-250 μm, (F) 250-300 μm before, 30 min, 60 min, and 120 min after a surface venule occlusion.