In Vivo Three-Dimensional Two-Photon Microscopy to Study Conducted Vascular Responses by Local ATP Ejection Using a Glass Micro-Pipette

Podsumowanie

We present an optimized local ejection procedure using a glass micro-pipette and a fast two-photon hyperstack imaging method, which allows precise measurement of capillary diameter changes and investigation of its regulation in three dimensions.

Streszczenie

Maintenance of normal brain function requires a sufficient and efficient supply of oxygen and nutrition by a complex network of vessels. However, the regulation of cerebral blood flow (CBF) is incompletely understood, especially at the capillary level. Two-photon microscopy is a powerful tool widely used to study CBF and its regulation. Currently, this field is limited by the lack of in vivo two-photon microscopy studies examining (1) CBF responses in three-dimensions, (2) conducted vascular responses, and (3) localized interventions within the vascular network. Here, we describe a 3D in vivo method using two-photon microscopy to study conducted vascular responses elicited by local ejection of ATP with a glass micro-pipette. Our method uses fast and repetitive hyperstack two-photon imaging providing precise diameter measurements by maximal intensity projection of the obtained images. Furthermore, we show that this method can also be used to study 3D astrocytic calcium responses. We also discuss the advantages and limitations of glass micro-pipette insertion and two-photon hyperstack imaging.

Wprowadzenie

The brain has a high energy consumption rate. About 20% of the oxygen and 25% of the glucose consumed by the human body are dedicated to brain function, while the brain only occupies 2% of the total body mass. Maintenance of normal brain function requires a sufficient and efficient supply of oxygen and nutrition by blood flow in a complex network of vessels. Local brain activity and cerebral blood flow (CBF) are robustly coupled, depending on the functional properties of neurons, astrocytes, pericytes, smooth muscle cells (SMCs) and endothelial cells (ECs)¹. Recently, the first few orders of capillaries branching from penetrating arterioles have emerged as a 'hotspot'², showing active regulation of capillary blood flow. A slow conducted vascular response (CVR) was discovered at this 'hotspot' in mouse somatosensory cortex during both whisker stimulation and local ejection (puffing) of ATP with a glass micro-pipette³.

Although in vivo imaging by two-photon laser scanning fluorescent microscopy has been widely used for studying neurovascular responses in cerebral cortex, most of the studies measured blood vessel diameters and investigated their regulation in a two-dimensional (2D) x-y plane. The challenges are: Firstly, cerebral blood vessels and their embracing astrocytes, pericytes and SMCs construct branches in three dimensions (3D). It is therefore crucial to study their interactions in 3D. Secondly, even a small amount of drift in focus will affect the precise measurement of both vessel diameters and cellular fluorescent signals. Finally, CVRs are fast and far-reaching in three dimensions. 3D volume scanning is optimal for detecting CVRs and unveiling their mechanisms. We implemented a piezo motor objective in a two-photon microscope to study mouse somatosensory cortex in vivo, allowing precise diameter measurements by maximal intensity projections of the obtained images.

Glass micro-pipettes have frequently been used for in vivo brain studies, e.g., to bulk-load organic dyes⁴, record EEGs⁵ and for patch clamping⁶. Nonetheless, limitations remain. Commonly, the tip of the glass micro-pipette is imprecisely placed, or the micro-pipette is not used for local interventions. Here, we have optimized the procedure of micro-pipette insertion and local ejection.

Furthermore, the combination of 3D two-photon microscopy and genetically-encoded fluorescent indicators offers an unprecedented opportunity to investigate neurovascular coupling in a 3D scope. In this study, we took advantage of this and injected viral vectors carrying astrocyte specific genetically-encoded calcium indicators into the mouse somatosensory cortex. Astrocytes as well as vessel diameters were imaged simultaneously by combining different fluorescent markers.

Overall, we present an optimized method of local ejection (puffing) by glass micro-pipette and fast two-photon hyperstack imaging, which allows precise measurement of capillary diameter changes. In addition, our method provides a novel tool to simultaneously study 3D profiles of Ca²⁺ responses in astrocytes and vascular diameter responses.

Protokół

All procedures involving animals were approved by the Danish National Ethics Committee according to the guidelines set forth in the European Council’s Convention for the Protection of Vertebrate Animals Used for Experimental and Other Scientific Purposes and were in compliance with the ARRIVE guidelines. This is a terminal procedure with the mice being euthanized prior to anesthetic recovery.

1. Pre-surgical preparation

1. Clean the surgical table and all the surrounding area with 70% ethanol. Thoroughly clean and dry the surgical tools before the surgery.
2. Anesthetize an NG2DsRed mouse (Tg(Cspg4-DsRed.T1)1Akik/J; both genders; 4-8 months old) by a peritoneal injection of ketamine + xylazine (60 mg/kg + 10 mg/kg) dissolved in sterilized water, pH 7.4. Administer supplemental doses of ketamine (30 mg/kg) every 25 min until completion of the surgical procedure. Check tail and toe pinch reflexes regularly to verify the depth of anesthesia.
3. Inject saline subcutaneously at four different locations on the back of the mouse for a total volume of 1 mL to prevent dehydration during the experiment. To protect the eyes from drying out, apply eye lubricant. Maintain body temperature at 37 °C using a rectal temperature probe and heating blanket.

2. Surgical procedure

1. Tracheotomy
   1. Place the mouse on its back. Inject 0.04 mL of 0.5% lidocaine subcutaneously under the planned incision and wait for 2 min. Make a 10 mm long incision above the chest bone (manubrium). Separate the submandibular glands and sternothyroid muscles, and then expose the trachea.
   2. Make tracheostomy between two tracheal rings with scissors. Insert a small metal tube with a length of 16.2 mm and a diameter of 1 mm into the trachea. Connect the tube to a mechanical ventilator and capnograph to monitor the end-expiratory CO₂ (Figure 1A).
2. Catheter insertion
   1. Inject 0.5% lidocaine (0.04 mL) subcutaneously under the planned incision and wait for 2 min.
   2. Using fine tip forceps gently separate the artery from the vein. Stop the blood flow upstream with a microvascular clamp. Constrict the blood flow downstream with a 10-0 nylon suture ligating both blood vessels.
   3. Use micro-dissecting scissors to make a small incision in both the artery and the vein. Insert plastic catheters into both blood vessels. Secure the catheters by tightening the sutures (8-0 or 10-0 nylon sutures dependent on the vessel size) without compromising catheter lumens. Remove the microvascular clamp and close the skin.
   4. Monitor blood pressure via the arterial catheter. Measure levels of blood gases in arterial blood samples (50 μL) before and after each experiment (normal values: pO₂, 90-120 mmHg; pCO₂, 35-40 mmHg; pH, 7.25-7.45) using a blood gas analyzer. Use the vein catheter for infusion of fluorescein and anesthesia.
3. Craniotomy
   1. Shave the fur off the head of the mouse between the ears. Inject 0.04 mL of 0.5% lidocaine subcutaneously under the planned incision and wait for 2 min.
   2. Wipe the skull using a 10% iron-chloride solution to remove the periosteum. Rinse the skull thoroughly with saline. Glue a metal head bar to the skull with cyanoacrylate glue and activator (Figure 1A).
      NOTE: The metal head bar is 75 mm long and 13.5 mm wide, with a round hole of 5 mm diameter in the center.
   3. Drill a 4-mm-diameter craniotomy, centered at 0.5 mm behind and 3 mm to the right of the bregma over the right sensory barrel cortex. Remove the dura with a fine-tip vessel dilator.
   4. Cover the exposed cortex with 0.75% agarose gel, dissolved in artificial cerebrospinal fluid (aCSF; pH = 7.4) and cooled to 35 °C. Cover 80-90% of the craniotomy with a glass coverslip at an angle of 10-15° to the metal head bar (Figure 1B) that permits insertion and placement of the glass micro-pipette.
   5. To minimize brain pulsation, glue the two corners of the glass coverslip onto the metal head bar with cyanoacrylate glue and activator. Apply the activator carefully to prevent getting onto the exposed brain. Rinse the glued glass coverslip thoroughly with saline afterwards.
4. Perform insertion of the whisker pad stimulation probe. Insert a set of custom-made bipolar electrodes (8 mm length and 0.25 mm thickness) percutaneously into the left side of face to stimulate the ramus infraorbitalis of the trigeminal nerve contralateral to the craniotomy. Position the cathode close to the hiatus infraorbitalis, and insert the anode into the masticatory muscles⁷. Perform the stimulation with an electrical stimulator at an intensity of 1.5 mA for 1 ms in trains of 20 s at 2 Hz.
5. Upon completion of surgery, administer a bolus of 0.05 mL fluorescein isothiocyanate-dextran (FITC-dextran, 4% w/v, molecular weight [MW] 50,000) into the femoral vein to label the blood plasma. Discontinue ketamine and switch the anesthesia to continuous intravenous infusion of α-chloralose (17% w/v, final concentration) mixed with FITC-dextran (2% w/v, final concentration) at 0.02 mL/10 g/h. Wait for about half an hour for stabilization of the new anesthetic state.
   NOTE: Both FITC-dextran and α-chloralose are dissolved in 0.9% saline.

3. First two-photon imaging session

1. Transfer the mouse to the stage of a commercial two-photon microscope (Table of Materials). Under both red fluorescent protein (RFP, Figure 1C) and green fluorescent protein (GFP, Figure 1C) modes of epi-fluorescent illumination, take a picture of the exposed cortex with a 5x objective. Use the RFP picture as a 'map' for insertion of the glass micro-pipette in the next session.
   NOTE: The GFP 'map' is used to distinguish veins/venules from arteries/arterioles.
2. Perform two-photon imaging using the two-photon microscope and a 25x 1.0 numerical aperture (NA) water-immersion objective with piezo motor. Set the excitation wavelength to 900 nm. Search the cortex and follow each penetrating arteriole and find its horizontal branches (1^st order capillaries).
3. Image penetrating arterioles and their 1^st order capillaries at rest and during whisker pad stimulation. See the detailed imaging parameters in section 5.
4. Mark locations of 1^st order capillaries with >5% vasodilation during whisker pad stimulation on the RFP 'map' (Figure 1C). Locations with <5% vasodilation are considered as being outside of the whisker barrel cortex region.
   NOTE: If using wild-type mice instead of NG2DsRed mice in the experiment, the GFP image is used as ‘map’ instead.

4. Insertion of the glass micro-pipette

1. Prepare glass micro-pipettes for puffing using a pipette puller (Table of Materials). The glass micro-pipettes have a resistance of 3-3.5 MΩ and a taper length of 4.5-5 mm. Load a pipette with a mixture of 1 mM ATP and 10 µM Alexa 594 in order to visualize the pipette tip under the epi-fluorescent lamp and two-photon microscope.
2. Mount the glass micro-pipette onto a patch clamp holder and connect it to an air pump. Set the pump holding pressure to 0 psi.
3. Using red epi-fluorescent illumination, focus on the pipette tip with the 5x objective. Place the glass micro-pipette into the aCSF above the agarose. Adjust the holding pressure of the air pump to 0.2 psi. A small red cloud can be observed ejecting out of pipette tip in the aCSF. This is to prevent clogging during pipette insertion.
4. Choose one of the marked locations in the RFP 'map' as destination. Move the pipette tip to the horizontal plane of the cover glass edge with 30 µm distance, roughly pointing towards the destination in the x-y plane.
5. Carefully lower the pipette tip in z-axis under the cover glass slip. Then start advancing the glass pipette towards the destination until ~500 µm above the brain surface. Switch focus frequently between the cover glass edge, the pipette tip and the brain surface, making sure there is enough free space for moving the pipette.
6. Switch the objective to 25x. Re-center the pipette tip and advance the pipette tip further towards the destination on the RFP 'map'. Keep the pipette tip in the agarose layer.
7. When the pipette tip is ~100 µm above the brain surface, switch the imaging mode to two-photon microscopy. Focus on a target location at which to puff. Write down its x, y coordinates (x₀, y₀) and its depth below the surface (z₀). Calculate the coordinates of the insertion point on the brain surface (x_i, y_i) as follows:
   x_i = x₀ + z₀ / tan θ
   y_i = y₀
   where θ is the angle between the pipette holder and the horizontal plane.
8. Position the pipette tip to the coordinates (x_i, y_i) on the brain surface. Gently and slowly insert the pipette until it reaches (x₀, y₀, z₀). If a vessel is in the way of an insertion path, withdraw the pipette and re-calculate other insertion coordinates and path; or turn the stage plate slightly (as indicated in Figure 1A).
9. Set the pump holding pressure at 0 psi and ejection pressure at 10-15 psi. Puff ATP for 200-400 ms at the target location during two-photon imaging. Adjust the pressure and duration of puffing for each pipette and over time as explained in the discussion section.

5. Hyperstack two-photon imaging

1. Perform experiments using the two-photon microscope and a 25 x 1.0 NA water-immersion objective with piezo motor. Set the excitation wavelength to 900 nm. Filter the emitted light to collect red light from DsRed (pericytes)/tetramethylrhodamine isothiocyanate-dextran (TRITC-dextran) staining blood plasma and green light from FITC-dextran (blood plasma)/GCaMP6 (GFP, astrocytic calcium).
2. Produce a high-resolution z-stack at the location of interest containing a penetrating arteriole, a 1^st order capillary, a 2^nd order capillary and possibly neighboring fluorescent cells (e.g., pericytes, astrocytes). Determine the x*y*z volume of hyperstack imaging by including the 3D structure of the blood vessels as much as possible. The total height of each stack may vary from 30-50 μm, while the x-y plane roughly covers an area of 60 µm x 40 µm.
3. Set the image stack in the imaging software to be comprised of 8-10 planes with an inter-plane distance of 4-5 µm. The piezo-motor objective stops at each level on the z-axis to acquire the image of each plane. The sampling rate is 1 s per stack and pixel resolution in the x-y plane is 0.2-0.3 µm (Figure 2A). One recording normally includes 10 pre-puff image stacks and 150 post-puff image stacks.

6. Data processing

1. Process data using custom-made analytical software. Flatten each image stack into one image by maximal intensity projection (Figure 2B). Draw a rectangular region of interest (ROI) with a width of 3 µm perpendicularly across the vessel of interest (Figure 2C).
2. To minimize interference from shadows of red blood cells and from minor vibrations of the cortex, average the rectangular ROI by projecting it into one line, which then represents the average profile of the vessel segment. Do this for each maximal intensity image.
3. Plot the profile lines as a 2D image with the x-axis representing maximal intensity images in chronological order (Figure 2D, upper panel). Use an active contour algorithm (Chan-Vese segmentation) to find the edges of the vessel^8,9 (indicated by red curves; Figure 2D, upper panel).
4. Calculate the time course of the diameter changes as the difference between the edges of the blood vessel (vertical distance between the upper and lower red curves in Figure 2D, lower panel). Normalize diameter changes to pre-puffing diameter baseline and plot curves of diameter change over time. Do this for each ROI (Figure 2E).
5. The vessel response amplitude is defined as the highest/lowest peak amplitude after puffing. The response latency is defined as the latency of half-max amplitude (Figure 2G). Manually trace the skeletonized vascular structure by placing nodes along the vessels (Figure 2F) and measuring geographical distance between each ROI and the penetrating arteriole. Calculate CVR speed by dividing the distances by latency differences.

7. Viral vector injection

1. In order to simultaneously study astrocytic calcium responses, inject a volume of 0.6 µL viral vector (pZac2.1 gfaABC1D-lck-GCaMP6f, Addgene) into the somatosensory cortex of NG2DsRed mice, following the standard stereotactic viral vector injection procedure¹⁰.
2. After three weeks, astrocytes in mouse somatosensory cortex express the genetically encoded calcium indicator, GCaMP6f. Follow the aforementioned surgical procedure and two-photon imaging for mice. To distinguish between astrocytes and vessel lumina, TRITC-dextran instead of FITC-dextran is used to stain vessel lumina red.

Wyniki

Once the surgery was complete, mice were transported to two-photon microscope (Figure 1A). A glass micro-pipette containing 1 mM ATP was inserted in proximity of the destination blood vessel at the target location (Figure 1B).

We performed hyperstack imaging while giving a puff of 1 mM ATP (Figure 2A, Supplementary Video 1). Each image st...

Dyskusje

One challenge for vascular studies is the precise measurement of vessel diameters. The method we describe here used a motorized piezo objective to make fast and repetitive hyperstack imaging by two-photon microscopy. Firstly, this method allows repeated examinations of the penetrating arteriole, 1^st order and 2^nd order capillaries without loss of focus and led to the discovery of slowly conducted vascular responses in capillaries in vivo. Secondly, in combination with a viral vector injection techni...

Materiały

Name	Company	Catalog Number	Comments
Agarose	Sigma–Aldrich	A6138	Apply upon exposed cortex for protection
Alexa 594	Life Technologies	A-10438	Stain puffing compound to red fluorescent color
ATP	Sigma-Aldrich	A9187	Vasodilator and vasoconstrictor, puffing compound
Cyanoacrylate glue and activator	Loctite	Adhesives and SF7452	Glue for metal piece and coverglass
Eye lubricant	Neutral Ophtha, Ophtha A/S, Denmark		Keep the mouse eyes moisterized
FITC-dextran	Sigma-Aldrich	FD500S	Blood serum dye, green fluorescent color
NG2DsRed mice	Jackson Laboratory	8241	These transgenic mice express an red fluorescent protein variant (DsRed) under the control of the mouse NG2 (Cspg4) promoter
pZac2.1 gfaABC1D-lck-GCaMP6f	Addgene	52924-AAV5	Astrocyte specific viral vectors carrying genetically encoded calcium indicators
TRITC-dextran	Sigma-Aldrich	52194	Blood serum dye, red fluorescent color
List of Equipments
Air pump	WPI	PV830	Give air pressure to pipette puffing
Blood gas analyzer	Radiometer	ABL 700	Measure levels of blood gases
Blood pressure monitor	World Precision Instruments	BP-1	Monitor aterial blood pressure
Body temperature controller	CWE	Model TC-1000	Keep the mouse body temperature in physiological range
Capnograph	Harvard Apparatus	Type 340	Monitor the end-expiratory CO2 from the mouse
Electrical stimulator	A.M.P.I.	ISO-flex	Apply whisker pad stimulation
Mechanical ventilator	Harvard Apparatus	D-79232	Mechanically ventilate the mouse in physiological range
Micropipette puller	Sutter Instrument	P-97
Two-photon microscope	Femtonics Ltd	Femto3D-RC
List of Surgical Instruments
Anatomical tweezer	Lawton	09-0007
Angled and balanced tweezer	S&T AG	00595 FRAS-18 RM-8
Iris scissor	Lawton	05-1450
Micro vascular clamp	S&T AG	462
Mouse vascular catheters	Verutech	100828