Real-Time Two-Photon Imaging of Brain Endothelial NAD+ Metabolism in Mice

Summary

This protocol enables real-time monitoring of brain endothelial nicotinamide adenine dinucleotide metabolism in living mice using endothelial-targeted biosensor expression, cranial window preparation, and two-photon imaging.

Abstract

We present a reproducible workflow for real-time visualization of nicotinamide adenine dinucleotide (NAD⁺) biosensor-channel responses in brain microvascular endothelial cells in living mice using cranial-window two-photon microscopy and an endothelial-targeted fluorescent NAD⁺ sensor. The protocol includes (1) adeno-associated virus (AAV)-mediated expression of the NAD⁺ sensor selectively in cerebrovascular endothelium using AAV-X1.1 under the Cdh5 promoter (vascular endothelial cadherin), (2) surgical preparation of a stable 3 mm × 3 mm cortical cranial window, and (3) dual-wavelength two-photon imaging to simultaneously capture the 920 nm-excited green sensor-channel signal and the 1040 nm-excited intravascular tetramethylrhodamine (TMR)–dextran reference channel. The TMR–dextran channel provides a vascular lumen reference and supports vessel selection, motion assessment, and vascular-integrity evaluation. As an application example, we describe nicotinamide mononucleotide–associated changes in endothelial sensor-channel fluorescence following drinking-water, oral-gavage, or intravenous delivery. This protocol emphasizes critical steps, recommended viral-dosing and titer considerations, troubleshooting, and quantitative-analysis strategies, enabling laboratories to implement in vivo monitoring of cerebrovascular NAD⁺ biosensor-channel dynamics for studies of neurometabolism in health and disease.

Introduction

Nicotinamide adenine dinucleotide (NAD⁺) is a central cofactor for redox balance, mitochondrial activity, DNA repair, and stress-responsive signaling. In the cerebral vasculature, endothelial NAD⁺ homeostasis is increasingly recognized as an active regulator of barrier integrity, vascular tone, and neurovascular coupling rather than a passive metabolic readout¹. Disruption of endothelial NAD⁺ metabolism has been linked to aging-associated blood–brain barrier (BBB) leakage and impaired neurovascular homeostasis, highlighting the need for methods that can resolve endothelial metabolic dynamics directly in living brain tissue².

Most currently used approaches to quantify NAD⁺ rely on endpoint biochemical measurements, fixed-tissue staining, or ex vivo sensor readouts. Although informative, these approaches do not preserve intact microvascular architecture, cannot readily distinguish rapid within-animal changes over time, and provide limited cell-type specificity when applied to heterogeneous brain tissue. By contrast, chronic cranial-window two-photon imaging enables repeated visualization of cortical microvessels in living mice with high spatial and temporal resolution, making it well suited for monitoring metabolically responsive endothelial signals under defined physiological or pharmacologic interventions^3,4,5.

The overall goal of this protocol is to provide an end-to-end workflow for real-time monitoring of brain endothelial NAD⁺ dynamics in vivo. The method combines systemic delivery of an endothelial-targeted, enhanced green fluorescent protein (eGFP)-based NAD⁺ sensor packaged in adeno-associated virus (AAV)-X1.1 and driven by the Cdh5 promoter, a stable 3 mm × 3 mm cranial window for optical access^6,7,8, and dual-wavelength two-photon imaging with an intravascular tetramethylrhodamine (TMR)–dextran reference channel¹. The sensor mechanism has been clarified as follows: the Cambronne et al. LigA-cpVenus cytoplasmic NAD⁺ biosensor is an inverse-response cpVenus-based sensor. NAD⁺ binding reduces cpVenus/eGFP-like fluorescence in the standard 488 nm excitation readout, whereas a 405 nm excitation channel serves as a reference for ratiometric normalization. Therefore, raw green fluorescence should not be interpreted as a direct positive measure of NAD⁺ concentration. In this two-photon implementation, 920 nm excitation is used to acquire the green sensor-channel signal, whereas 1040 nm excitation is used for the TMR–dextran vascular reference channel; thus, the present 920/1040 nm acquisition is not equivalent to a calibrated 488/405 nm ratiometric measurement. The Cambronne et al. biosensor was selected because it has been extensively validated for monitoring intracellular NAD⁺ dynamics and is compatible with the current dual-wavelength imaging configuration. Although newer indicators such as FiNad provide alternative approaches for monitoring NAD⁺ dynamics, their implementation would require modification of the red fluorescence channel currently used for TMR–dextran vascular-reference labeling. The Cambronne sensor reports relative changes in NAD⁺ availability within a finite dynamic range and has been validated in cellular and subcellular applications; however, signal interpretation may be influenced by expression level, pH, photobleaching, and optical-path variability, and absolute NAD⁺ quantification requires calibration or orthogonal biochemical validation. AAV-X1.1 was selected because endothelial-tropic AAV variants have been reported to transduce the central nervous system vasculature efficiently^6,7. In our preliminary comparison of several AAV serotypes, including AAV9, AAV-BR1, AAV-ENT, and AAV-X1.1, using CAG- or Cdh5-driven enhanced green fluorescent protein expression as a reporter, AAV-X1.1 produced the most favorable vascular-associated expression pattern among the tested serotypes. The Cdh5 promoter was subsequently used to further restrict transgene expression to vascular endothelial cells, consistent with previous reports demonstrating endothelial-selective transgene expression in vivo⁶.

Compared with ex vivo NAD⁺ assays, this approach preserves the intact vascular network, enables longitudinal measurements in the same animal, and permits simultaneous assessment of endothelial metabolic signals and a vascular lumen reference. The workflow is particularly useful for studies of aging, inflammatory stress, BBB dysfunction, and rapid responses to metabolic interventions such as nicotinamide mononucleotide (NMN) administration^{9,10,11,12,13}. This method is most appropriate when the experimental question requires repeated visualization of cortical microvessels in living mice and relative changes in an endothelial sensor-channel signal. It is less suitable for absolute NAD⁺ quantification without additional ratiometric calibration or biochemical validation. Practical limitations include cortical optical accessibility, imaging-depth constraints, dependence on successful endothelial AAV transduction, potential cranial-window-associated inflammation, and the requirement for a dual-wavelength two-photon microscope. Chronic cranial-window implantation can induce local inflammation, vascular reactivity, gliosis, opacity, or tissue remodeling, all of which may influence endothelial physiology and sensor-channel measurements^1,3,4. To minimize these confounders, the protocol requires aseptic surgery, intermittent drilling with frequent cooling, careful hemostasis, avoidance of adhesive contact with the dura, postoperative analgesia, daily monitoring, and exclusion of animals with persistent opacity, inflammatory debris, infection, unstable vascular morphology, or poor baseline signal stability.

Protocol

All procedures were approved by the Animal Ethics Committee of the University of Health and Rehabilitation Sciences under the project entitled “Mechanistic Study of Metabolic Probe-Based Observation of Blood–Brain Barrier Integrity” (approval number KFDX: NO. 2023-1023). Perform all surgeries under aseptic conditions and provide perioperative analgesia and supportive care according to the approved animal protocol. In the representative workflow, administer buprenorphine hydrochloride (0.05–0.1 mg/kg, subcutaneously) before or immediately after surgery and then every 8–12 h for 24–48 h as needed. Recover mice in a warmed cage and monitor them until normal posture, spontaneous movement, and stable respiration are restored.

1. Experimental overview

1. Use adult C57BL/6J mice aged 10–12 weeks and weighing 23–25 g for the representative experiments.
2. Maintain mice under specific-pathogen-free conditions with a 12 h light/dark cycle, an ambient temperature of 22°C–24°C, a relative humidity of 40%–60%, and ad libitum access to food and water.
3. Acclimatize animals for at least 7 days before procedures.
4. Use either sex as appropriate for the experimental design. Report sex, animal numbers, and group allocation for each experiment. The representative imaging dataset shown here used N = 3 mice.
5. Administer systemic AAV first and allow a 3–4 week expression period before imaging.
6. Perform cranial-window surgery 3–5 days before imaging.
7. Anesthetize mice on the imaging day, inject TMR–dextran intravenously, and perform imaging after adequate tracer circulation.
8. Administer NMN according to the selected intervention schedule. Provide drinking-water supplementation for 5 days before imaging, perform oral gavage at the predefined interval before imaging, or administer intravenous NMN during the imaging session after baseline acquisition.
   NOTE: Pause the protocol after AAV administration during the 3–4 week expression period and maintain animals under routine health monitoring and standard housing conditions. Pause the protocol after cranial-window surgery during the 3–5 day recovery period. Before proceeding to imaging, confirm stable body weight, normal grooming and activity, absence of infection or excessive inflammation, and a clear cranial window.

2. Cranial window surgery for longitudinal cortical imaging

1. Pre-surgical preparation
   1. Sterilize all surgical instruments and establish a sterile field before anesthesia induction.
   2. Set the homeothermic heating system to maintain body temperature at 36.5°C–37.5°C, preferably using a rectal-probe feedback controller.
   3. Prepare respiratory-monitoring equipment and confirm stable operation before surgery.
   4. Prepare sterile saline, absorbent swabs, hemostatic materials, a sterile No. 1.5 glass coverslip, a metal headplate, and dental cement before starting the procedure.
   5. Use a sterile No. 1.5 glass coverslip sized to completely cover the 3 mm × 3 mm cranial window and sit flush on the skull rim.
   6. Prepare a high-speed microdrill fitted with a 0.5–1.0 mm burr.
   7. Operate the drill at approximately one-half of the maximum speed and confirm that sterile saline irrigation can be applied immediately during skull thinning.
   8. Apply room-temperature sterile saline with a cotton swab or dropper during drilling to cool the skull and remove bone dust.
2. Anesthesia, scalp preparation, and skull exposure
   1. Anesthetize mice with freshly prepared Avertin working solution (1.25%, 12.5 mg/mL) by intraperitoneal injection at 200–250 mg/kg.
      1. Prepare the Avertin working solution from 2,2,2-tribromoethanol and 2-methyl-2-butanol according to institutional standard operating procedures, protect the solution from light, and discard preparations showing precipitation or discoloration.
   2. Confirm a stable surgical plane by loss of pedal-withdrawal reflex and stable respiration.
   3. Apply ophthalmic ointment to both eyes immediately after anesthesia induction.
   4. If inhalation anesthesia is used, induce anesthesia with 3%–4% isoflurane.
   5. Maintain anesthesia with 1%–2% isoflurane in oxygen at 0.5–1.0 L/min and adjust according to respiration and reflexes.
      CAUTION: Handle anesthetic agents according to institutional safety procedures and use appropriate scavenging systems when administering inhalation anesthesia.
   6. Fix the mouse in a stereotaxic frame or custom head holder that provides stable access to the dorsal skull.
   7. Maintain body temperature at 36.5°C–37.5°C using a feedback-controlled heating pad and rectal probe when available.
   8. Shave the scalp and remove loose hair from the surgical field.
   9. Disinfect the scalp using alternating povidone-iodine or iodophor solution and 70% ethanol for three cycles.
   10. Avoid allowing disinfectant to enter the eyes or cranial opening.
   11. Make a midline scalp incision and gently retract the skin.
   12. Expose the skull sufficiently to allow both cranial-window creation and headplate fixation.
   13. Remove the periosteum and residual connective tissue using fine forceps or a scalpel blade.
   14. Dry the skull gently with sterile cotton swabs for approximately 10–20 s until the surface appears matte rather than wet.
   15. Avoid excessive dehydration of exposed tissue.
3. Marking a 3 mm × 3 mm craniotomy
   1. Identify bregma and lambda under the surgical microscope.
   2. Use these anatomical landmarks to orient the midline and target cortical region.
   3. Mark a 3 mm × 3 mm square over one hemisphere of the dorsal somatosensory cortex.
   4. Position the anterior edge approximately 0.5 mm posterior to bregma and the medial edge approximately 1.0 mm lateral to the midline.
   5. Avoid major skull sutures and large surface vessels whenever possible.
4. Drilling and bone-flap removal
   1. Thin the skull along the marked perimeter using intermittent drill contact rather than continuous pressure.
   2. Apply sterile saline repeatedly during drilling to prevent thermal injury.
   3. Deliver approximately 20–50 µL sterile saline per application or enough to keep the skull surface moist and cool.
   4. Repeat irrigation every 10–20 s or whenever bone dust accumulates or heat buildup is suspected.
   5. Continue thinning until the bone becomes uniformly translucent and slightly flexible.
   6. Confirm adequate thinning by observing visible dural vessels, preserved dural pulsation, and absence of cortical compression or thermal discoloration.
   7. Stop drilling immediately if excessive heat buildup, bleeding, or reduced dural pulsation occurs.
   8. Irrigate the area and reassess before continuing.
   9. Lift the bone flap gently from one corner using fine forceps or a needle tip.
   10. Re-thin the perimeter if resistance remains.
   11. Preserve the dura intact during bone-flap removal.
   12. Rinse away bone dust using sterile saline.
   13. Control minor bleeding using gentle saline irrigation and absorbent swabs.
   14. Consider bleeding acceptable if it resolves within 1–2 min after irrigation and absorbent-swab contact.
   15. Exclude the animal or terminate the procedure if bleeding persists for more than approximately 5 min.
   16. Exclude the animal or terminate the procedure if the dura is torn, cortical compression occurs, or blood obscures the imaging area.
       NOTE: The procedure may be paused temporarily after complete hemostasis and before coverslip placement. Maintain the exposed dura in a moist condition using sterile saline and resume only after confirming the absence of active bleeding.
5. Window sealing and headplate fixation
   1. Confirm that the exposed dura is clean, moist, and free of active bleeding.
   2. Place a sterile coverslip over the craniotomy.
   3. Position the coverslip on the skull rim surrounding the craniotomy without compressing the dura.
   4. Verify correct positioning by confirming a flush coverslip surface, absence of trapped air bubbles, preserved dural pulsation, and a clear optical path.
   5. Prepare dental cement or Super-Bond C&B-type adhesive according to the manufacturer’s instructions.
      1. For the representative Super-Bond C&B preparation, mix 1 scoop of powder with 1 drop of catalyst and 4 drops of monomer in a pre-cooled ceramic dish maintained at approximately −20°C immediately before application.
   6. Apply adhesive only around the outer skull edge surrounding the coverslip.
   7. Avoid adhesive contact with the dura and avoid capillary wick-back beneath the glass.
   8. Allow the adhesive to harden for approximately 5–10 min at room temperature.
      CAUTION: Dental cement and adhesive components may contain volatile or irritant chemicals. Handle them according to institutional safety procedures and use appropriate ventilation.
   9. Position a custom titanium headplate around the cranial window.
   10. Use a headplate containing an approximately 10 mm central opening compatible with the imaging head holder.
   11. Center the cranial window within the headplate opening.
   12. Apply dental cement only to skull-contacting regions outside the optical window.
   13. Allow the cement to harden completely for approximately 5–10 min before moving the animal.
6. Post-operative recovery and window quality control
   1. Administer postoperative analgesia according to the approved animal protocol.
   2. In the representative workflow, administer buprenorphine hydrochloride at 0.05–0.1 mg/kg subcutaneously every 8–12 h for 24–48 h.
   3. Recover mice in a warmed cage.
   4. Monitor respiration, posture, movement, grooming, and body weight daily.
   5. Accept cranial windows that exhibit a transparent glass surface, visible cortical vasculature, intact dura, minimal residual blood, no adhesive beneath the coverslip, no infection, and stable headplate fixation.
   6. Exclude animals with marked opacity, persistent bleeding, inflammatory debris, infection, window displacement, severe weight loss (>15%), or poor grooming and activity.
   7. Proceed to imaging after 3–5 days of recovery.
   8. Confirm normal activity, normal grooming behavior, stable body weight, absence of pain or infection, and a transparent cranial window with clearly visible vessels before imaging.
      NOTE: The protocol may be paused during the postoperative recovery period. Maintain mice under standard housing conditions, monitor body weight and window clarity daily, and resume imaging only after all recovery criteria have been met.

3. AAV-mediated endothelial NAD⁺ sensor expression

1. Viral construct and preparation
   1. Use an eGFP/cpVenus-based genetically encoded NAD⁺ biosensor driven by the endothelial Cdh5 promoter.
   2. Use the Cambronne et al. LigA-cpVenus cytoplasmic NAD⁺ biosensor sequence (for example, Addgene plasmid #186787 or an equivalent sequence-verified insert).
   3. Clone the Cdh5 promoter upstream of the NAD⁺ biosensor coding sequence using the CV232 vector backbone.
   4. Package the expression cassette into the AAV-X1.1 capsid as a customized viral preparation.
      1. Record the viral-production provider, purification method, viral-genome quantification method, production lot number, and certificate of analysis for each AAV preparation.
   5. Record the plasmid map, sequence-verification results, viral-production information, purification information, viral-genome titer, and certificate of analysis for each viral preparation.
      1. Record the construct map, sequence-verification report, viral-production batch number, quantitative polymerase chain reaction (qPCR) titer report, purification documentation, and certificate of analysis for each viral preparation used in the study.
   6. Refer to the Table of Materials for all construct details.
   7. Use high-titer AAV preparations with lot-specific qPCR titers typically greater than 2 × 10¹³ vg/mL.
   8. Store single-use aliquots at −80°C.
   9. Avoid repeated freeze–thaw cycles.
   10. Record the production lot, qPCR titer, storage duration, and acceptable lot-to-lot variation before use.
2. Systemic AAV administration
   1. Thaw a viral aliquot on ice.
   2. Dilute the virus in sterile phosphate-buffered saline (PBS) to a final injection volume of 100–200 µL per mouse.
   3. Maintain a final dose of 2 × 10¹² vg per mouse across animals.
      1. Calculate the injection volume from the lot-specific viral-genome titer. For the representative preparation with a qPCR titer of 1.39 × 10¹³ vg/mL, inject approximately 144 µL to deliver a dose of 2 × 10¹² vg per mouse.
   4. Deliver the virus by retro-orbital venous sinus injection using a sterile 29 G insulin syringe.
   5. Inject the viral suspension slowly over approximately 10–20 s.
   6. Confirm successful delivery by verifying the absence of leakage, excessive bleeding, and ocular injury.
   7. Validate and report any alternative administration route separately.
   8. Observe the mouse until normal posture and spontaneous movement are restored.
   9. Confirm stable respiration and absence of bleeding or distress.
   10. Return the animal to its home cage only after full recovery.
   11. Allow 21–28 days for sensor expression before imaging.
   12. Maintain a consistent expression interval within each comparison set.
   13. Validate expression intervals outside the 21–28 day range separately.
       NOTE: The protocol may be paused during the AAV-expression period. Maintain routine husbandry and health monitoring and proceed to cranial-window surgery only after the planned expression interval has elapsed and the animal remains healthy.
3.  Validation of endothelial restriction
   1. Screen animals before formal data collection.
   2. Accept animals when the green sensor-channel signal is detectable and follows the vascular network.
   3. Confirm that the signal is enriched along vessel walls adjacent to the TMR–dextran–positive lumen.
   4. Confirm the absence of dominant diffuse parenchymal fluorescence.
   5. Apply identical acceptance criteria across all experimental groups.
   6. Perfuse mice with PBS followed by 4% paraformaldehyde when post hoc validation is required.
   7. Collect the brain and prepare 30–40 µm coronal or sagittal sections.
   8. Stain sections with endothelial markers such as tomato lectin, CD31, or vascular endothelial cadherin (VE-cadherin).
      1. Record the antibody source, catalog number, lot number, host species, dilution, incubation conditions, and detection method for each validation experiment.
   9. Assess nicotinamide phosphoribosyltransferase (NAMPT) and poly(ADP-ribose) polymerase 1 (PARP1) expression when required.
   10. Use primary-antibody dilutions optimized for the tissue preparation.
   11. Acquire fluorescence or confocal images using identical acquisition settings across groups.
   12. Quantify colocalization as the fraction of sensor-positive signal overlapping with or immediately adjacent to vascular-marker-positive regions.
   13. Exclude animals when most detectable signal is located outside vascular structures.
   14. Exclude animals when vessel-wall enrichment is absent.
   15. Exclude animals when off-target parenchymal fluorescence prevents reliable region-of-interest selection.
   16. Exclude animals when sensor-channel fluorescence is saturated or unstable during baseline acquisition.

4. Two-photon imaging and quantitative analysis

1. Imaging-day preparation and intravascular reference labeling
   1. Anesthetize mice with freshly prepared Avertin working solution (1.25%, 12.5 mg/mL) by intraperitoneal injection at 200–250 mg/kg, or use the anesthetic regimen approved by the institutional animal protocol.
   2. Confirm a stable surgical plane by loss of pedal withdrawal and stable respiration.
   3. Apply ophthalmic ointment immediately after induction.
   4. If inhalation anesthesia is used, induce anesthesia with 3%–4% isoflurane.
   5. Maintain anesthesia with 1%–2% isoflurane in oxygen at 0.5–1.0 L/min and adjust according to respiration and reflexes.
   6. Prepare 70 kDa TMR–dextran in sterile PBS at 10–25 mg/mL.
   7. Protect the TMR–dextran solution from light and filter the solution if required by institutional practice.
   8. Inject TMR–dextran intravenously at 20 mg/kg.
   9. Adjust the injection volume according to the stock concentration. For a 25 g mouse, administer approximately 0.5 mg total dye.
   10. Prepare the tracer fresh or store aliquots protected from light according to the manufacturer's recommendations.
   11. Allow TMR–dextran to circulate for 5–10 min before image acquisition.
2. Microscope configuration
   1. Use a dual-wavelength two-photon microscope configured for 920 nm excitation of the eGFP-based sensor channel.
   2. Use a 1040 nm excitation path for the TMR–dextran vascular reference channel.
   3. Representative imaging was performed on a Nikon A1 MP/A1R MP+ multiphoton microscope controlled by NIS-Elements AR/AR Analysis software (version 5.42.06).
   4. Record the laser sources, pulse parameters, beam-combination configuration, detector type, and software version for each imaging system.
   5. Configure detection channels to separate the sensor-channel signal from the TMR–dextran signal.
   6. Use non-descanned photomultiplier tube (PMT) or gallium arsenide phosphide (GaAsP) detectors.
   7. Collect green sensor-channel emission between approximately 500 and 550 nm.
   8. Collect TMR–dextran emission between approximately 570 and 650 nm.
      1. In the representative dataset, collect green sensor-channel emission using a 525/50 nm emission filter and collect TMR–dextran emission using a 575/25 nm or 610/75 nm emission filter.
   9. Minimize spectral bleed-through between channels.
   10. Maintain detector gain and offset settings constant within each comparison set.
   11. Verify channel separation using single-channel controls when available.
   12. Use a Nikon CFI75 Apochromat 25XC W 1300 water-immersion objective or equivalent (25×, numerical aperture 1.10, working distance 2.0 mm).
   13. Confirm channel co-registration using fixed vascular landmarks such as bifurcations and lumen edges.
   14. Repeat co-registration verification after any optical-path adjustment.
   15. Measure laser power at the sample plane before imaging.
   16. Maintain 920 nm laser power constant within each comparison set.
   17. Maintain 1040 nm laser power constant within each comparison set.
   18. Use approximately 10–20 mW at 920 nm and 8–18 mW at 1040 nm as representative working ranges.
   19. Adjust laser power according to imaging depth and cranial-window quality.
   20. Exclude fields showing greater than 20% signal loss during the 120 s baseline period.
   21. Exclude fields showing visible phototoxic changes.
3. Acquisition settings
   1. Select fields of view with a clear optical window, sharply focused microvessels, and minimal motion.
   2. Accept windows with a transparent coverslip, visible cortical vasculature, intact dura, minimal residual blood, no adhesive beneath the coverslip, no infection, and stable headplate fixation.
   3. Exclude animals with marked opacity, persistent bleeding, inflammatory debris, infection, window displacement, severe weight loss (>15%), or poor grooming/activity.
   4. Acquire time-lapse images at 512 × 512 pixels.
   5. Acquire representative still images at 1024 × 1024 pixels when higher spatial detail is required.
   6. Maintain pixel size, zoom, scan speed, and field of view constant within each comparison set.
   7. Acquire z-stacks using a step size of 1–3 µm.
   8. Image cortical microvessels at depths of approximately 200–250 µm below the pial surface.
   9. Extend imaging depth to approximately 400 µm when window quality permits.
   10. Maintain imaging depth constant for paired comparisons.
   11. Acquire a 120 s baseline at 1 s intervals before acute interventions.
   12. Continue acquisition for the predefined post-intervention period.
   13. Record intravenous NMN responses for approximately 10–30 min after injection.
   14. Acquire drinking-water and oral-gavage datasets at predefined experimental endpoints.
   15. Maintain identical acquisition settings across all comparison groups.
4. NMN intervention (application example)
   1. Dissolve NMN in sterile drinking water.
   2. Adjust the NMN concentration to deliver approximately 100 mg/kg/day.
   3. Protect the drinking bottle from light.
   4. Replace NMN-containing drinking water every 24–48 h.
   5. Monitor water consumption using bottle-weight measurements.
   6. Calculate estimated NMN intake using water consumption and body weight.
   7. Continue supplementation for 5 days before imaging.
   8. Prepare NMN at 1 mg/mL for oral gavage.
   9. Administer 200 µL per mouse using a 20G, 38 mm rounded-tip gavage needle.
   10. Perform imaging 30–60 min after gavage.
   11. Prepare NMN at 1 mg/mL for intravenous delivery.
   12. Inject 100 µL per mouse through the retro-orbital venous sinus or tail vein.
   13. Deliver the injection slowly over 10–20 s.
   14. Begin imaging immediately or within 5 min after injection.
   15. Acquire drinking-water datasets after 5 days of supplementation.
   16. Acquire oral-gavage datasets at a fixed post-gavage time point.
   17. Acquire a baseline immediately before intravenous injection.
   18. Record acute responses for approximately 10–30 min after intravenous injection.
5. Image processing and signal extraction
   1. Perform motion correction when motion artifacts are present.
   2. Use Fiji/ImageJ (version 1.8.0 or later) with StackReg, TurboReg, or an equivalent rigid-body registration algorithm.
   3. Use vascular landmarks in the TMR–dextran channel to guide registration.
   4. Accept datasets only when residual displacement is less than approximately 1–2 pixels after correction.
   5. Confirm that vascular region of interest (ROI) boundaries remain stable throughout the acquisition.
   6. Use the TMR–dextran channel to segment vessel lumens.
   7. Define endothelial ROIs immediately adjacent to the lumen.
   8. Select microvessels with stable lumen labeling, minimal motion, and visible sensor-channel signal.
   9. Analyze vessel diameters of approximately 5–80 µm.
   10. Define endothelial ROIs as vessel-wall-associated regions approximately 2–5 µm wide.
   11. Analyze 3–5 vessels per field of view when available.
   12. Analyze 1–3 fields of view per mouse when available.
   13. Define background ROIs in tissue regions lacking tracer and sensor signal.
   14. Place background ROIs at least 20 µm away from analyzed vessels and image artifacts.
   15. Subtract mean background intensity from each channel before normalization.
   16. Apply the same background-subtraction strategy to all samples.
   17. Extract time-series mean intensity values for F_sensor(t).
   18. Extract time-series mean intensity values for F_ref(t).
   19. Perform measurements using NIS-Elements AR/AR Analysis or Fiji/ImageJ.
   20. Export raw image files as ND2 files.
   21. Export analysis copies as 16-bit TIFF stacks when possible.
   22. Export ROI time-series data as CSV files.
   23. Calculate background-subtracted fluorescence values before normalization.
   24. Calculate ΔF/F₀ using ΔF/F₀ = (F(t) − F₀)/F₀.
   25. Define F₀ as the mean signal during the 120 s baseline period.
   26. Calculate vascular-reference normalization as F_sensor(t)/F_ref(t).
   27. Report peak change, time-to-peak, area under the curve, and recovery kinetics when acquisition duration permits.
   28. Do not interpret the 920 nm sensor-channel signal as a calibrated NAD⁺ concentration.
   29. Perform additional ratiometric calibration or biochemical validation when quantitative NAD⁺ measurements are required.
6. Statistics and reporting
   1. Define the mouse as the biological experimental unit.
   2. Average vessel-level measurements within each mouse before group-level analysis or apply a nested statistical model.
   3. Use paired two-tailed t-tests for normally distributed within-mouse comparisons.
   4. Use Wilcoxon matched-pairs signed-rank tests when normality assumptions are not met.
   5. Use repeated-measures ANOVA or mixed-effects models for analyses involving more than two related conditions.
   6. Apply appropriate multiple-comparison corrections.
   7. Perform statistical analyses using GraphPad Prism or equivalent software.
   8. Define statistical significance as p < 0.05.
   9. Report the exact statistical test in each figure legend.
   10. Code image files before ROI selection and quantification whenever feasible.
   11. Maintain analyst blinding until ROI measurements and quality-control decisions are complete.
   12. Use N = 3 mice as the biological experimental unit in the representative dataset.
       NOTE: Do not pause image acquisition during active time-lapse imaging. Image processing and quantitative analysis may be paused after raw data export. Store ND2 files, TIFF analysis files, ROI files, metadata, and CSV measurement tables in a backed-up archive. Record laser power, detector settings, imaging depth, animal ID, treatment assignment, and exclusion decisions for each imaging session.

Results

A successful preparation yields a transparent and mechanically stable cranial window suitable for longitudinal imaging. Figure 1A summarizes the procedural workflow from anesthesia and skull exposure through craniotomy, coverslip placement, headplate fixation, and post-operative recovery. Figure 1B presents representative serial photographs of the major surgical steps. Windows suitable for imaging exhibit a clear optical path, visible cortical vasculature, an in...

Discussion

This protocol provides an end-to-end workflow to monitor brain endothelial NAD⁺ metabolism in vivo. Key determinants of success include: (i) window clarity and stability, including minimization of thermal damage and control of bleeding during drilling; (ii) endothelial specificity and adequate sensor expression achieved through appropriate selection of the AAV capsid, promoter, dose, and expression interval; and (iii) consistent imaging and analysis settings, particularly when performing longitudinal ...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
2,2,2-Tribromoethanol (Avertin)	Sigma-Aldrich	T48402	CAUTION: Toxic/harmful chemical. Used to prepare Avertin anesthesia solution for cranial-window surgery. Typical working solution: 1.25% (12.5 mg/mL), administered intraperitoneally at 200–250 mg/kg.
2-Methyl-2-butanol	Sigma-Aldrich	152463	CAUTION: Toxic/harmful chemical. Solvent used for preparation of concentrated Avertin stock solution. Protect from light and discard preparations showing precipitation or discoloration.
Absorbent swabs	Winner Medical	Sterile medical cotton swabs	Used for hemostasis, skull drying, irrigation, and surgical-field maintenance.
AAV-X1.1-Cdh5-cytoplasmic NAD+ biosensor-SV40 pA	Shanghai GeneChem Co., Ltd.; Addgene sources for plasmids	GeneChem construct GCPV5052075; Addgene #186787; Addgene #196836	Endothelial-targeted NAD+ biosensor packaged in AAV-X1.1. Expression cassette: Cdh5 promoter–cytoplasmic NAD+ biosensor–SV40 poly(A). Representative injection dose: 2 × 10¹² vg/mouse. Store aliquots at −80 °C and avoid repeated freeze–thaw cycles. Record lot-specific qPCR titer, batch number, plasmid map, sequence verification, and certificate of analysis.
Analgesic agent (buprenorphine hydrochloride)	MCE	HY-B0071	Perioperative and postoperative analgesia. Representative dose: 0.05–0.1 mg/kg subcutaneously every 8–12 h for 24–48 h.
Anesthesia delivery system	RWD Life Science	R500 with R580 vaporizer	Used for isoflurane delivery during imaging. Typical oxygen flow: 0.5–1.0 L/min.
Anesthetic agent (isoflurane)	RWD Life Science	R510-22 series	Used for imaging anesthesia. Typical settings: 3–4% induction and 1–2% maintenance.
Coverslip	Custom-made	No catalog number	Sterile No. 1.5 glass coverslip sized to completely cover a 3 × 3 mm cranial window and sit flush on the skull rim.
Dental cement/adhesive	Sun Medical	Super-Bond C&B; VZB/JAP8147	Used for cranial-window sealing and titanium headplate fixation. Representative preparation: 1 scoop powder, 1 drop catalyst, and 4 drops monomer. Harden approximately 5–10 min at room temperature.
Dental drill	RWD Life Science	78001 microdrill	High-speed drill used for skull thinning and craniotomy. Use approximately half of maximum speed with frequent saline cooling.
Disinfectant solution	Winner Medical	Povidone-iodine/iodophor and 70% ethanol	Used for scalp preparation before surgery. Apply alternating disinfectants for three cycles.
Fine forceps	JZ Brand	JD1050/JD1070	Used for periosteum removal, tissue manipulation, and bone-flap lifting.
Gavage needle	RWD Life Science	20G, 38 mm, ball tip	Used for oral NMN administration.
Headplate	Custom-made	No catalog number	Titanium headplate with approximately 10 mm central opening compatible with the imaging holder.
Heating pad/homeothermic controller	RWD Life Science	69020 with 69023 pad	Used to maintain body temperature at approximately 36.5–37.5 °C during surgery and imaging.
Image acquisition software	Nikon	NIS-Elements AR/AR Analysis v5.42.06	Used for microscope control and image acquisition. Representative settings include 512 × 512 or 1024 × 1024 pixels, 1–3 µm z-step, and 1 s time-lapse interval.
Image-analysis software	NIH	Fiji/ImageJ v1.8.0 or later	Used for motion correction, ROI selection, background subtraction, intensity extraction, and CSV export.
Intravenous injection supplies	BD	329461	Sterile 1 mL insulin syringe with 29G needle for AAV, TMR–dextran, and NMN injections.
NMN (nicotinamide mononucleotide)	Selleck Chemicals	S5259	NAD+ precursor used for drinking-water, oral-gavage, and intravenous administration experiments.
Ophthalmic ointment	Local certified pharmacy	Not applicable	Used to prevent corneal drying during anesthesia.
Phosphate-buffered saline (PBS)	Servicebio	G4202-100ML or G4202-500ML	Sterile 1× PBS used for viral dilution, tracer preparation, tissue rinsing, and reagent preparation.
Respiratory-monitoring system	RWD Life Science	Integrated monitoring module	Used to monitor respiration and anesthetic depth during surgery and imaging.
Saline (0.9% NaCl, sterile)	Servicebio	G4702-500ML	Used for skull cooling, irrigation, tissue moistening, and solution preparation.
Stereotaxic frame/head holder	Custom-made	No catalog number	Used for stable positioning during cranial-window surgery and imaging.
Surgical microscope	Olympus	SZ61	Used during cranial-window surgery for skull thinning, craniotomy, and coverslip placement.
TMR–dextran (70 kDa)	Thermo Fisher Scientific / Invitrogen	D1818	Intravascular fluorescent tracer used for vascular reference labeling. Typical concentration: 10–25 mg/mL. Inject at 20 mg/kg and allow 5–10 min circulation before imaging.
Tomato lectin (DyLight 488)	Vector Laboratories	DL-1174	Used for post hoc vascular/endothelial validation.
Two-photon microscope	Nikon	A1R MP+ multiphoton microscope	Dual-wavelength imaging system using 920 nm excitation for the sensor channel and 1040 nm excitation for the TMR–dextran reference channel. Record laser power, detector settings, filters, and acquisition metadata for each experiment.
Water-immersion objective	Nikon	CFI75 Apochromat 25XC W 1300	25× objective, NA 1.10, working distance 2.0 mm, used for in vivo two-photon imaging.