Microfluidic Chip for Axonal Injury Models Construction and Enabling Multi-Omics Analysis

Summary

This protocol describes a microfluidic system modeling neuronal metabolic dynamics post-axonal injury, enabling imaging, multi-omics analysis, and mechanistic studies of intrinsic metabolic remodeling.

Abstract

Neurons polarize to form dendrites and axons, enabling intercellular communication. Axonal injury disrupts these connections and transmits damage signals to the soma, often leading to neuronal degeneration. Thus, maintaining axonal homeostasis is essential for promoting local axon regeneration and protecting against neurodegeneration. This process relies on cellular metabolism to supply energy and biosynthetic precursors and is sustained by mechanisms that regulate metabolic balance and eliminate by-products. However, neuronal metabolism is compartmentalized between the soma and axon and is further influenced in vivo by the surrounding microenvironment, such as astrocyte-derived metabolic activity (e.g., the astrocyte-neuron lactate shuttle). These factors complicate the investigation of neurons' intrinsic metabolic mechanisms. To address these challenges, here we developed a microfluidic platform for culturing primary cortical neurons in vitro that preserves key metabolic characteristics observed in vivo, including physiological glycolytic flux and mitochondrial respiration. This system provides a simplified model for investigating intrinsic metabolic remodeling in neurons after axonal injury. Conventional microfluidic chips support in vitro axonal injury models and are compatible with live-cell imaging, immunofluorescence staining, and hypoxia treatment. To accommodate large-scale transcriptomic and metabolomic analyses involving millions of cells, we further designed and fabricated high-throughput microfluidic chips with optimized operational protocols. The device features alternately arranged soma and axon chambers connected by microchannels, and axonal injury is induced by vacuum aspiration of fluid from the axon compartment. This platform enables rapid assessment of metabolite and enzyme dynamics, improving the accuracy and reproducibility of multi-omics investigations.

Introduction

Neurons establish two distinct types of protrusions with specialized functions, axons and dendrites, via polarization during development^1,2. Nerve regeneration refers to the process of axonal regrowth after neural injury³, with a research focus on central nervous system (CNS) repair. Unlike the peripheral nervous system, the CNS has limited regenerative capacity^4,5,6, often resulting in persistent functional impairments or permanent disabilities in patients with spinal cord injuries or traumatic brain injuries^7,8. Thus, developing strategies to enhance CNS repair and elucidating the mechanisms underlying axonal regeneration are crucial.

Microfluidic devices, functioning as integrated, cost-effective, and high-throughput in vitro culture systems, exhibited remarkable advantages in cell and neuron research in recent years^9,10. Compared with traditional methods, microfluidic technology enables precise regulation of fluid shear stress, concentration gradients, and spatial structures, thereby closely simulating authentic physiological and pathological microenvironments to promote cell growth, migration, differentiation, and interactions^11,12,13,14. Microfluidics-based in vitro axonal injury models enable compartmentalized culture of neuronal cell bodies and processes through spatial polarization, providing an indispensable tool for exploring the intrinsic regenerative capacity of neurons and their metabolic adaptations following injury^15,16,17,18. It provides a critical window into how neurons respond to injury, regulate metabolic processes¹⁹, and promote axonal regeneration²⁰. However, existing microfluidic platforms for axonal injury and metabolic research still face challenges, including low neuronal cell yields, reduced accuracy in metabolic measurements, and a lack of standardized operational procedures.

In response to these limitations, this study developed a novel large-scale microfluidic platform that enables standardized, high-throughput induction of axonal injury and provides sufficient cell material for compartmentalized multi-omics analyses. This integrated design directly confronts the critical unmet need in neural injury research. Specifically, the microfluidic chip we designed features an alternating arrangement of somal and axonal chambers connected by microgrooves, with axonal transection achieved through controlled vacuum aspiration. This approach is ideal for studies that require reproducible, high-throughput axonal injury models and precise metabolic profiling, such as investigating the metabolic mechanisms underlying axonal regeneration. Compared to traditional methods, this technique offers higher throughput, greater metabolic precision, and the capability to establish standardized injury models. For example, previous studies have demonstrated the importance of mitochondrial transport in axonal regeneration^20,21,22,23, whereas our platform further reveals the role of glucose metabolism in axonal regeneration following injury. This protocol establishes a technical system for large-scale axotomy and metabolic analysis that is both visually verifiable and operationally standardized.

Protocol

All experiments were conducted in accordance with the guidelines of the Ethics Committee of Beihang University and were approved by the committee (approval code: BM20210060). Figure 1 outlines the experimental timeline.

1. Design of the conventional and large-scale microfluidic device

1. Design the microfluidic device using AutoCAD to create two-dimensional (2D) schematics and three-dimensional (3D) structural models.
2. Use these designs to fabricate a master mold on a silicon wafer with SU-8 negative photolithography.
3. Replicate the master mold by casting polydimethylsiloxane (PDMS), resulting in a high-precision, functional microfluidic chip.
4. Design each microchannel to be 5 µm high, allowing only axons to pass while blocking somas.
5. Make the microchannels 10 µm wide to increase axonal capacity.
6. Design the chambers to be 120 µm in height and 220 µm in width, providing enough space for neuronal soma.
7. Use this microfluidic platform to simulate in vivo axonal injury by controlling fluid flow and neuron placement.
8. Prepare metabolomic and transcriptomic samples using the device for high-precision cell manipulation, and enable high-throughput processing to improve sample accuracy and biological validity.

2. Fabrication of conventional and large-scale microfluidic devices

1. Weigh the PDMS base and curing agent at a 10:1 (w/w) ratio and place them into a centrifuge tube.
2. Place the centrifuge tube containing the PDMS mixture into a centrifugal stirrer mixer. Centrifuge at 2000 × g for 4 min to mix, then centrifuge again at 2000 × g for 4 min to degas.
3. Pour 13-15 g of the degassed PDMS into the microfluidic mold, ensuring the bottom of the mold is completely covered.
4. Place the mold in a vacuum desiccator. Use a vacuum pump to evacuate air for 5-10 min to thoroughly remove bubbles from the PDMS.
5. Use a rubber bulb to gently tap the surface of the PDMS to break any remaining surface bubbles.
6. Place the mold in a convection oven and bake at 80 °C for 100 min to cure the PDMS.
7. Once the PDMS is fully cured, gently slide the tip of a scalpel under the edge of the PDMS to carefully lift and separate it from the mold.
8. Use a biopsy punch with a diameter of 2.0-2.5 mm to create holes in the PDMS for culture medium infusion and cell loading.
9. Remove any surface impurities from the PDMS using adhesive tape, then place it in a clean glass dish and wrap it with aluminum foil for storage.
10. (Optional) Treat the PDMS surface with oxygen plasma (30 W, 20 s) and then press it against another surface for irreversible bonding.
11. Prior to the experiment, autoclave the microfluidic device at 121 °C and 101 kPa for 5 min to ensure sterility.

3. Preparation of cortical neurons

1. Anesthetize 3 postnatal Sprague-Dawley (SD) rats (within 12 h after birth) by placing them on ice for 5 min, then rapidly decapitate.
2. Dissect the motor area of the cerebral cortex and place the tissue in pre-cooled dissection buffer (e.g., 2 mL of HBSS without calcium and magnesium).
   NOTE: Perform all tissue dissection procedures on ice using pre-chilled solutions.
3. Mince the tissue into approximately 1 mm³ pieces using sterile scissors and transfer to a 15 mL centrifuge tube.
4. Let stand for 5 min and then carefully aspirate the supernatant to remove residual blood and buffer.
5. Add pre-warmed papain solution (1 mg/mL) in 5 mL of HBSS with calcium and magnesium that has been heated to 37 °C.
6. Incubate in a 37 °C, 5% CO₂ incubator for 40 min, gently inverting the tube 3 times every 10 min.
7. After digestion, gently triturate the suspension 100 times using a 200 µL pipette.
   NOTE: Avoid introducing air bubbles to prevent cell stress.
8. Centrifuge at 300 × g for 3 min and discard the supernatant.
9. Add a termination solution containing 1 mg/mL protease inhibitor (e.g., 600 µL of papain dissociation system kit inhibitor vial in 5 mL of HBSS with calcium and magnesium) to resuspend the cells.
10. Filter the cell suspension through a 40 µm cell strainer and collect the filtrate in a new centrifuge tube.
11. Centrifuge the filtrate at 200 × g for 10 min and discard the supernatant.
12. Resuspend the pellet in 500 µL of complete neuronal culture medium containing 2% B27 and 1% GlutaMAX.
13. Count the cells (approximately 3 million/rat) and adjust the density to 1 × 10⁶ cells/mL for subsequent use.

4. Coating

NOTE: Treat the coverslips or culture dishes with poly-D-lysine (PDL) solution to improve neuron adhesion and create an optimal environment for neuronal growth.

1. Coating of coverslips
   1. Place a coverslip intended for conventional microfluidic devices (Φ25 mm, thickness 0.17 mm) in a 35 mm culture dish, and add 2 mL of 0.1 mg/mL PDL solution.
   2. Incubate at 37 °C for 6 h or overnight.
   3. After incubation, carefully retrieve the PDL solution for reuse or proper disposal.
   4. Add 2 mL of sterile ultrapure water to the dish and rotate it 50 times clockwise, followed by 50 times counterclockwise (for multiple dishes, rotate them simultaneously in a larger dish).
   5. Aspirate the water using a vacuum pump connected to a sterile pipette tip, collecting waste in a liquid waste container.
   6. Repeat the washing procedure (steps 4.1.4-4.1.5) three times in total.
   7. Allow the coverslips to air-dry naturally in a biosafety cabinet before use.
2. Culture dish coating
   1. Add an appropriate amount of 0.1 mg/mL PDL solution to the culture dish or well plate (e.g., add 2 mL to a 35 mm dish for conventional microfluidics, and 5 mL to a 10 cm dish for large-scale microfluidics).
      NOTE: Ensure the biosafety cabinet is in proper operation with sterile airflow to prevent contamination.
   2. Ensure the solution completely covers the bottom, and incubate it at 37 °C for 6 h or overnight.
   3. After recovering the PDL solution, wash the culture dish three times with an appropriate amount of sterile ultrapure water as described in step 4.1.4.
   4. Remove any residual ultrapure water and air-dry the culture dish in a laminar flow hood for later use.

5. Preparation of neuronal medium

1. Prepare the complete medium for routine neuron culture.
   1. Remove the plunger from a disposable 50 mL syringe. Place the syringe with the rubber head facing upward on a clean bench within a biosafety cabinet.
   2. Add 500 µL of penicillin streptomycin (PS), 1 mL of B27, 125 µL of GlutaMAX, and 500 µL of fetal bovine serum (FBS) into the syringe in sequence. Change the pipette tip for each component to avoid cross-contamination.
      NOTE: Add each component in the specified order to maintain consistency.
   3. Take a new sterile 50 mL centrifuge tube. Label it as "Neurobasal-A (NBA) + 2% B27 + 0.25% GlutaMAX + 1% FBS + 1% PS".
      NOTE: Use NBA if not embryonic neurons; use neurobasal (NB) for embryonic neurons.
   4. Attach a 0.22 µm syringe filter to the end of the syringe.
      NOTE: Ensure the filter is tightly connected to the syringe to prevent leakage.
   5. Use NBA to fill the syringe with liquid up to 50 mL, including the volume of additives. Reinstall the plunger.
   6. Slowly push the plunger at approximately 2 mL/s to filter the liquid into the labeled centrifuge tube.
   7. Tighten the cap of the centrifuge tube. Invert the tube 3-5 times to mix the medium.
2. Complete medium containing [U-¹³C₆]-glucose (for metabolic flux analysis): For metabolic flux analysis of glucose metabolism in damaged neurons, use sugar-free NBA to prepare a complete medium containing [U-¹³C₆]-glucose.
   1. Use glucose-free NBA as the base solution and add [U-¹³C₆]-glucose to a final concentration of 25 mM before filtration (if [U-¹³C₆]-glucose is solid, calculate and weigh it in advance). Follow the same sequence for adding other components, as well as the filtration and mixing steps, as described in section 5.1.

6. Cell counting

1. Wipe the hemocytometer thoroughly with 75% ethanol. Ensure all surfaces are completely dry before proceeding.
2. Add 90 µL of sterile phosphate buffered saline (PBS) buffer to a 1.5 mL sterile centrifuge tube.
3. Gently resuspend the collected neuronal cell suspension by pipetting up and down 3-5 times to ensure thorough mixing.
4. Using a 10 µL pipette, transfer 10 µL of the mixed cell suspension into the centrifuge tube containing 90 µL of PBS (1:10 dilution).
5. Use a 200 µL pipette to gently mix the solution 10 times, ensuring no bubbles form.
6. Replace the pipette tip with a new 10 µL tip. Draw 10 µL of the mixture and add it to the sample chamber of the hemocytometer.
7. Place the hemocytometer on the microscope stage and focus using the 20x objective. Locate the central 4 × 4 grid and count cells in all 16 squares following standard hemocytometer rules.
8. Record total cell counts from all 16 squares. Calculate cell concentration using the formula: (Total count × 10) × 10⁴ cells/mL.
   NOTE: Always work in a biosafety cabinet with proper personal protective equipment.

7. In vitro culture of cortical neurons in conventional or large-scale microfluidic devices

1. Protocol A (Conventional microfluidic devices).
   1. Verify that the glass surface in the 35 mm culture dish is completely dry.
   2. Using sterile fine-tip forceps, place the autoclaved microfluidic chip at the center of the glass surface with microchannels facing downward.
   3. Gently press the chip onto the glass surface using a 200 µL pipette tip to ensure complete adhesion.
   4. Mark the left chamber with a dot (·) using a permanent marker to designate the soma compartment.
   5. Aspirate 10 µL of neuronal suspension (10 million cells/mL) using a 10 µL pipette.
   6. Slowly dispense the suspension into the loading port above the marked soma compartment.
   7. Confirm proper fluid flow into the lower reservoir of the left chamber.
   8. Transfer the culture dish to a humidified incubator (37 °C, 5% CO₂) for 20 min.
   9. Remove the culture dish from the incubator.
   10. Using a 20 × phase-contrast microscope, confirm media perfusion from the soma compartment to the axonal compartment through the microchannels.
   11. Add 150 µL of neuronal basal medium to the upper port of the axonal compartment (right chamber), allowing gravity-driven flow to the lower reservoir.
   12. Similarly, add 15 µL of complete neuronal medium to the upper port of the soma compartment (left chamber).
       NOTE: At this stage, neurons should adhere visibly, with most cell debris flushed into the lower well. Do not replace the medium during culture; use a single addition of complete neuronal medium as described in step 5.1.3.
   13. Pour 20 mL of ultrapure water into a 150 mm culture dish to create a humidity chamber. Place the 35 mm culture dish inside the large dish.
   14. Transfer the entire culture system to the incubator and maintain for the required duration (e.g., 7 days).
       NOTE: Perform all cell culture procedures under sterile conditions in a biosafety cabinet. Wear appropriate personal protective equipment (lab coat and gloves). Decontaminate work surfaces before and after procedures.
2. Protocol B (Large-scale microfluidic chips).
   NOTE: The standard protocol of conventional specifications was followed with two modifications.
   1. Use a 10 cm Petri dish for device placement.
   2. Choose either full-channel or interval-channel seeding based on experimental needs, as each method requires different axonal injury protocols.
   3. After seeding is completed, add 5 mL of complete neuron culture medium to the 10 cm Petri dish to maintain neuron growth.
   4. Add 20 mL of ultrapure water to a 150 mm large dish to create a humidity chamber. Then, place the 10 cm dish inside the large dish (Figure 3F).
   5. Same as step 7.1.14.

8. Establishing an in vitro axonal injury model in conventional microfluidic devices

1. Mark the left side of the microfluidic device as the somal compartment (as indicated by the "·" mark), and the right side as the axonal terminal compartment.
2. Seed cortical neurons into the somal compartment.
3. Culture the neurons in vitro until day 7 (DIV7), allowing axons to extend through the microchannels to the axonal terminal compartment.
4. At DIV7, perform axonal injury procedures in the axonal terminal compartment to establish an in vitro axonal injury model.
5. Take an appropriate amount of NBA medium and transfer it to a new 15 mL centrifuge tube for later use.
6. Culture cortical neurons in the microfluidic device for the specified period. Then transfer the device to a clean bench.
7. Use lint-free paper to dry the bottom of the 35 mm culture dish. Use a 200 µL pipette to aspirate the medium from the two right-side holes of the microfluidic device.
8. Connect the vacuum pump tubing to a sterile, filter-free 200 µL pipette tip, turn on the vacuum pump (suction rate 60 L/min), and aim the tip at the connection point between the lower hole of the axon terminal chamber and the chamber to aspirate the old medium (at this point, axons break due to negative pressure).
9. Replace the 200 µL pipette tip and draw 150 µL of NBA, slowly adding it through the lower hole of the right chamber.
   NOTE: If the liquid flow is too slow and bubbles form, quickly expel the liquid from the tip to rapidly fill the chamber.
10. Alternate using the two right-side holes for liquid addition: after each NBA addition, immediately use the vacuum pump to aspirate liquid from the other hole to sever axons. Repeat this operation 4 times, and after the last aspiration, replace with fresh complete neuronal medium.
11. Aspirate the old medium from the cell body side hole and add fresh complete neuronal medium containing drugs (if needed).
12. Place the microfluidic device along with the culture dish in a 150 mm culture dish and continue culturing for the specified time (e.g., 24 h).
    NOTE: The axon damage operation requires NBA medium, as neuronal complete medium will produce bubbles affecting the severing effect.

9. Large-scale microfluidic device axonal injury

1. Method 1: Manual axonal injury.
   1. Seed neurons into all chambers of the microfluidic device.
   2. Incubate the device at 37 °C in a humidified incubator with 5% CO₂ for three days, then remove the entire microfluidic device.
      NOTE: To avoid mechanical stress or tearing of axons grown into the microchannels, remove the microfluidic device at DIV3 instead of DIV7. This is particularly important if the experimental design includes an undamaged control group.
   3. At DIV7, remove the culture dish containing the microfluidic device from the incubator and place it on a sterile stage.
   4. Prepare a microscope with 20x magnification and sterilize the work area with 75% ethanol.
   5. Hold a 10 µL sterile filtered pipette tip, and identify axon bundles in the original microchannels under microscope guidance.
   6. Perform longitudinal scratches along each axon bundle using the pipette tip.
   7. Observe the axon breakage under the microscope in real-time after scratching.
      NOTE: When handling live neurons, wear nitrile gloves and work in a UV-sterilized laboratory environment
2. Method 2: Vacuum-assisted axonal injury.
   1. Seed cortical neurons alternately in the large-format microfluidic device.
   2. Turn on the vacuum pump and set it at 60 L/min.
   3. Connect the end of the vacuum pump tube to the axon chamber port using a sterile pipette tip.
   4. Complete axon severance by applying vacuum pressure through conventional microfluidic protocols.
   5. (Optional) To specifically collect injured axons, at a defined time point after axotomy, rapidly aspirate the contents of the cell body compartment using a vacuum pump (with a suction rate of 60 L/min) to remove the cell bodies, thereby leaving only the axons within the microchannels.

10. Hypoxic treatment of conventional or large-scale microfluidic devices

1. Seed primary neurons in the microfluidic device and culture until DIV7 under normoxic conditions (37 °C, 5% CO₂).
2. Place the microfluidic device in a hypoxia incubator chamber and supply it with a gas mixture consisting of 1% O₂, 5% CO₂, and balanced N₂.
3. Use a rotameter to control the gas flow rate and set it to 25 L/min.
4. Flush the chamber for 5 min to displace air.
5. While clamping the inlet of the hypoxic chamber, close the outlet pressure-reducing valve.
6. Clamp the outlet tubing of the hypoxic chamber, and finally, close the gas cylinder valve tightly.
7. Incubate the microfluidic device in the hypoxic chamber at 3 °C for 1 h.

11. Sample preparation of neurons cultured in vitro using large-scale microfluidic devices for omics analysis

1. Sample for transcriptomics.
   1. Transfer the microfluidic device with the damaged axons to a laminar flow hood 6 h after axonal injury is complete.
   2. After discarding the old medium, wash the neuronal cells 2-3 times with sterile PBS.
   3. Use a standard RNA extraction kit to isolate and purify RNA according to the manufacturer's instructions.
   4. Perform RNA sequencing. Conduct follow-up bioinformatics analysis on the acquired sequencing data.
2. Sample for metabolic flux analysis.
   NOTE: To monitor changes in glucose metabolic flux in neurons after injury using a large-scale microfluidic device, we employed [U-¹³C₆]-glucose tracing metabolic flux analysis²⁴. The protocol for preparing samples for LC-MS analysis is as follows:
   1. Use P0 SD rat pups for the experiments. Dissect cortical neurons and culture them in vitro either short-term (DIV5-DIV7) or long-term (DIV15-DIV30).
   2. For both young (DIV5-DIV7) and mature (DIV15-DIV30) neurons, establish injury groups and uninjured control groups.
   3. Replace the medium with one containing 25 mM [U-¹³C₆]-glucose (Neurobasal-A without glucose, 1% FBS, 2% B27, 25 mM [U-¹³C₆]-glucose) and incubate for 4 h at 37 °C.
      NOTE: To avoid neuronal stress and apoptosis caused by osmotic or pH changes during medium exchange, prepare additional neurons cultured in medium containing 25 mM [U-¹³C₆]-glucose at the time of seeding, and use this medium for the exchange.
   4. After gently rinsing twice with 5 mL of pre-warmed HBSS (containing calcium and magnesium), flash-freeze the samples in liquid nitrogen and store at -80 °C.
      NOTE: Wear nitrile gloves and a lab coat when handling liquid nitrogen.

12. Neuronal immunofluorescence staining based on conventional and large-scale microfluidic devices

NOTE: When culturing neurons in vitro using microfluidic devices, one must consider the potential damage to neuronal axons within microchannels that may occur during the disassembly of these devices for live cell fixation.

1. Use a micro waste vacuum pump to remove the culture medium from the four wells of the microfluidic device.
   ​NOTE: Keep the liquid in two chambers throughout the immunofluorescence staining process. The suction parameters should not be too high (2.5 L/min is optimal) to avoid damaging or aspirating neuronal cell bodies and axons.
2. Add 100 µL of 1x PBS to each of the two upper wells of the microfluidic device (PBS will naturally flow to the lower wells).
3. Replace the filterless tip at the end of the vacuum pump, aspirate PBS from the lower wells, and repeat the process three times. On the final iteration, aspirate all PBS from the four wells.
4. Add 150 µL of 4% Paraformaldehyde (PFA) solution to each of the two upper wells, and fix at room temperature for 10-15 min.
   NOTE: Avoid device disassembly during live cell fixation to maintain neuronal structural integrity. This approach helps distinguish the locations of neuronal somas and axons more clearly.
5. Wash three times with 1x PBS, then incubate neurons with blocking buffer containing 5% goat serum, 0.3 M glycine, 2% bovine serum albumin (BSA), PBS, and 0.1% (or 0.5%) Triton X-100 for 1 h.
6. Dilute the primary antibody (e.g., βIII-tubulin, mouse, 1:2000) in the blocking buffer, add 100 µL to the upper left well, and 50 µL to the upper right well, and incubate with neurons overnight at 4 °C.
7. After aspirating the primary antibody, wash three times with 1x PBS as described in steps 12.2-12.3. Then add the secondary antibody prepared with blocking buffer into the upper wells of the microfluidic device as described in step 12.6, and incubate in the dark for 30 min (cover with aluminum foil if necessary).
8. After washing three times with 1x PBS, rinse once with ultrapure water. Carefully use curved tweezers to grasp and remove the device from the upper right of the microfluidic setup.
9. Add 1 mL of PBS to the culture dish to separate the large coverslip from the dish.
10. Use a pipette to draw 15 µL of mounting medium and drop it onto the coverslip at the designated location for the coverslips.
11. Gently lower the cell side onto the mounting medium, place the coverslip, and remove excess liquid.
12. Allow the mounting medium to dry completely at room temperature.
13. Once the mounting medium has dried at room temperature, use a spinning disk confocal microscope to perform imaging.
14. Select a 20x air objective, set the exposure time to 100-200 ms per frame, and set the laser power to 5-10% of the maximum with 488 nm excitation for Green Fluorescent Protein (GFP) and/or 561 nm for Red Fluorescent Protein (RFP).

13. Live cell imaging in conventional microfluidic devices

1. Culture cortical neurons in the microfluidic device until DIV7.
2. Prepare live-cell dye (e.g., a probe for mitochondrial membrane potential) according to the manufacturer's instructions.
3. Dilute dye in pre-warmed (37 °C) neuronal maintenance medium.
4. Load dye solution into the somatic compartment of the microfluidic device.
5. Incubate for 30 min in a 37 °C, 5% CO₂ incubator.
6. Gently wash the compartments 3 times with 3 mL of pre-warmed (37 °C) HBSS containing calcium and magnesium.
7. Add phenol red-free live cell imaging solution to the microfluidic chip.
8. (Optional) Include GlutaMAX (0.25%) and B27 (2%) for prolonged imaging.
9. Place the microfluidic chip on the microscope stage with the somatic chamber on the left and the axonal terminal chamber on the right.
10. Add a drop of immersion oil to the objective lens, ensuring the oil makes contact with the coverslip beneath the chip.
11. Use a 40x objective with 150 ms exposure and 10-15% maximum laser power for imaging.
    NOTE: Adjust these parameters based on dye photostability.

14. Data analysis

1. Use ImageJ software to analyze microscopy images. Assess RNA quality with a bioanalyzer. Perform and analyze RNA-seq using an appopriate platform.
2. For metabolomic studies, extract raw mass spectrometry data using the MRMPROBS program (version 2.60), and collect information on identified metabolites, including isotopic details and peak areas.
3. Use GraphPad Prism (version 8.4.3) for all statistical analyses. Present data as mean ± SEM. Use unpaired two-tailed t-tests or Mann-Whitney tests for group comparisons, and apply two-way ANOVA for comparisons among multiple time points within groups (Significance: ns, not significant; *p < 0.05; **p < 0.01; ***p < 0.001).

Results

To establish a comprehensive research platform for studying the metabolic regulation mechanisms in cortical neurons, we first cultured cortical neurons in a conventional microfluidic device, precisely controlling axon growth along microchannels (Figure 2A). With an appropriate design, the microchannels were constructed with a low height (5 µm) and a larger width (10 µm), allowing each channel to host more axons without allowing cell bodies to enter. Immunofluorescence staining for βIII-tubulin confirmed axonal growth through the microchannels into the axonal terminal chamber (Figure 2B). To further determine whether axonal injury affects neuronal survival, we quantitatively analyzed neuronal density before and after axonal injury. The results showed that there was no significant difference in neuronal density following axonal injury, indicating that axonal injury did not lead to notable neuronal death (Figure 2C). To meet the sample size requirements for multi-omics analysis (typically around 3 × 10⁶ cells per sample), we developed a large-scale microfluidic device with an expandable layout and three-dimensional structure, ensuring accurate experimental data (Figure 3A-C). Both the unpunched and perforated versions of the PDMS replica molds adhere well to PC boards or 10 cm culture dishes (Figure 3D,E).

To verify axonal injury in the large-scale microfluidic device, two injury methods were used. For conventional axotomy (Figure 4A,B), axons were injured using a vacuum pump. For scratch injury (Figure 4C,D), the microfluidic mold was removed at DIV3, and axons were manually scratched with a pipette tip at DIV7. βIII-tubulin immunofluorescence staining was used to visualize axonal morphology before and after injury. Prior to injury, intact axons were observed in the axonal terminal chamber (Figure 4A,C). Immediately after injury, severed axonal stumps with disrupted morphology appeared (Figure 4B,D), confirming successful axonal injury and enabling subsequent analysis of regeneration. Microscopic observations before and after manual scratch assays showed that the axons of uninjured neurons remained intact, while axon regeneration to varying extents was observed at 6 h and 24 h after injury (Figure 5). Due to the small scale of conventional culture specifications (approximately 1 × 10⁵ per chip), protein concentration is difficult to detect; whereas in large-scale microfluidic neuron cultures (approximately 3 million per chip), the protein concentration is higher, making protein analysis more feasible and accurate. Figure 6 shows that the concentrations of neuronal proteins in the control group, 6 h and 24 h after axonal injury are 1420.4230, 1748.9397, and 1823.7007 µg/mL, respectively.

To elucidate the molecular mechanisms of axonal injury responses, we used a large microfluidic device for neuron culture with 3 million cells per chip. After 5 days (DIV5) of culture, axonal injury was induced, and cells were lysed to extract total RNA 6 h post-injury. Upon testing, the total RNA quantities in the control and experimental groups were 5.28 µg and 5.27 µg, respectively, meeting the quality requirements for RNA-sequencing (RNA-seq) (Figure 7A). Subsequently, RNA-seq analysis identified 595 injury-specific genes and 609 uninjured neuron-specific genes (Figure 7B). KEGG analysis further indicated significant upregulation of the TCA cycle, mitophagy, reactive oxygen species, oxidative phosphorylation, and cholesterol metabolism post-axonal injury (Figure 7C). These results suggest that young neurons actively regulate energy metabolism and the intracellular environment to cope with injury-induced stress. Further metabolic flux analysis showed that after [U-¹³C₆]glucose labeling, the levels of TCA cycle intermediates citrate and fumarate significantly increased, consistent with the activation trend of the TCA cycle pathway in the transcriptome, supporting the hypothesis of neuronal energy metabolism reprogramming following axonal injury (Figure 8).

In summary, these results demonstrate the utility of our microfluidic device in researching cortical neuron culture, injury, and regeneration, providing a reproducible platform for multi-omics analysis.

Figure 1: Workflow diagram of the experimental procedure. First, fabricate microfluidic devices using PDMS. Next, culture primary neurons in vitro and seed them into the microfluidic devices. Place the devices in a CO₂ incubator for further culturing. After the desired number of days, induce axonal injury using a vacuum pump or apply hypoxic treatment in a hypoxic chamber. Imaging or sample preparation is performed at designated time points following these treatments. Please click here to view a larger version of this figure.

Figure 2: Culturing cortical neurons using conventional microfluidic devices. (A) Standard microfluidic device mounted in a 35 mm dish during neuronal culture. Scale bar: 1 mm. (B) Immunofluorescence staining of cortical neurons cultured in a conventional-sized microfluidic device for 7 days. The axons are marked in green using βIII-tubulin (green). Scale bar: 50 µm. (C) Representative images (left) and corresponding quantitative analysis (right) of somatic side cortical neurons (DIV7) in microfluidic devices are shown for both uninjured and 24 h post-axonal injury conditions. The data present the number of surviving neurons per unit area, normalized by the control group. Scale bars: 20 µm. DAPI (blue), MAP2 (green). Data were presented as mean ± SEM. N = 3; n_(-)Ax = 30, n_(+)Ax = 33. p-values were calculated using a Two-tailed unpaired Student's t-test. Significance: ns, not significant. Please click here to view a larger version of this figure.

Figure 3: Design drawings of large-scale microfluidic devices. (A) The AutoCAD layout of the ultra-large-scale microfluidic chip, with dimensions in millimeters. (B) Schematic diagram of the three-dimensional overall structure of the ultra-large-scale microfluidic chip. (C) Top view of the ultra-large-scale microfluidic chip, with an enlarged view showing the microchannels. (D) Image of a large-scale microfluidic chip device, molded and cured from a PDMS replica of a silicon wafer master, connected to a PC board (without holes). Scale bar: 5 mm. (E) Image of the large-scale microfluidic chip device (with cell and axon chambers, with holes) connected to a 10 cm dish. Scale bar: 10 mm. (F) Schematic of the humidity chamber. 20 mL of ultrapure water are added to a 150-mm dish to create a humidified environment. A 10-cm cell culture dish containing the PDMS device and neuronal cultures is then placed inside the larger dish. Please click here to view a larger version of this figure.

Figure 4: Immunofluorescence images of cortical neurons (DIV7) cultured in large-scale microfluidic devices. (A) Neurons in the ultra-large-scale microfluidic chip labeled with βIII-tubulin (green) at DIV7, with an enlarged view showing axons growing in the microchannels and axon terminal chambers. Scale bars: 2000 µm (left), 500 µm (right). (B) Neurons in the ultra-large-scale microfluidic chip underwent axotomy at DIV7, and were labeled with βIII-tubulin (green). Enlarged view shows axonal stumps post-axotomy in the microchannels, suctioned by vacuum. Scale bars: 2000 µm (left), 100 µm (right). (C) Immunofluorescence images of neurons seeded in complete microchannels within large-scale microfluidic devices (Left: panorama, scale bar: 2000 µm; Right: enlarged view, scale bar: 100 µm). (D) Immunofluorescence images following manual scratch injury under a microscope (Left: injury panorama, scale bar: 2000 µm; Right: enlarged view of injury area, scale bar: 100 µm). Please click here to view a larger version of this figure.

Figure 5: Microscopic images of cortical neurons (DIV15) cultured in large-scale microfluidic devices. (A) Uninjured group; (B) 6 h post-axonal injury; (C) 24 h post-axonal injury. Scale bar: 100 µm. Please click here to view a larger version of this figure.

Figure 6: Protein concentration in injured neurons (DIV21) within large-scale microfluidic devices. (A) Standard curve for determining protein concentration using the Bicinchoninic Acid (BCA) method. A standard curve was established using 0-2000 µg/mL BSA standards (562 nm detection). The linear regression equation is: y = 0.0004637x + 0.002350 (R² = 0.9916). (B) Quantitative analysis of total neuronal protein yield collected from the large-scale microfluidic device in the control group and at 6 and 24 h after axonal injury. Data were presented as mean ± SEM. N = 3. Please click here to view a larger version of this figure.

Figure 7: RNA-seq analysis of cortical neurons (DIV5) after axonal injury in large-scale microfluidic devices. (A) Total RNA extracted 6 h after axonal injury in cortical neurons cultured within large-scale microfluidic devices. Data were presented as mean ± SEM. N = 3. (B) The Venn diagram of gene expression levels before and 6 h after axonal injury shows that 595 genes are specific to post-injury, while 609 genes are specific to the uninjured group. (C) KEGG pathway enrichment analysis on the upregulated genes in the axonal injury group (DIV5_Ax). Please click here to view a larger version of this figure.

Figure 8: Metabolic flux analysis of injured neurons cultured in large-scale microfluidics traced with [U-¹³C₆]-glucose. Cortical neurons were cultured in a large-scale microfluidic device, with both uninjured and injured groups (20 h post-injury) being continuously cultured for 4 h in medium containing [U-¹³C₆]-glucose, followed by liquid chromatography-mass spectrometry (LC-MS) detection of isotopic distribution of citrate (A) and succinate (B) in neurons (DIV5). All data were presented as mean ± SEM. N = 6, n_[(-)Ax] = 7, n_[(+)Ax24h] = 6; p-values were calculated using two-way ANOVA. Significance: ***p < 0.001. Please click here to view a larger version of this figure.

Discussion

The establishment of advanced in vitro models for studying axonal injury and regeneration is essential for uncovering the molecular and metabolic mechanisms involved in neuronal repair^16,19. In this study, we present a standardized workflow that integrates microfluidic devices, large-scale neuronal cultures, and precise axonal injury protocols, enabling robust multi-omics analysis of injured cortical neurons.

The success of this approach relies heavily on several critical procedural steps. In microfluidic chip preparation, precise control of the PDMS curing time (100-120 min at 80 °C) and sterilization duration (5-10 min) is crucial to prevent chip over-hardening and subsequent issues with slippage or leakage²⁵. In addition, accurate parameter settings for the vacuum pump during axotomy are essential: excessive suction can damage neuronal somas, while insufficient suction may result in incomplete axon severance. Notably, both the conventional and large-scale devices were manufactured using the same protocol. The main differences are in the design of the molds and the punching positions and diameters after curing. For the conventional microfluidic device, the diameter of the holes is 5 mm, while for the large-scale microfluidic device, the hole diameter is 3 mm. Another crucial operational detail is ensuring the right channel is free of any complete medium components (such as B27 or serum) before axotomy, as these substances easily generate bubbles during suction, interfering with axon severance.

We developed two methods for inducing axonal injury in neurons cultured within large-scale microfluidic devices. One method involves manual scratching under a microscope²⁶. Using this approach, the device must be removed 48 h to 72 h after seeding to avoid neurites damaging that have not yet extended into the channels. This step enables more accurate analysis of metabolic changes, including the effects of residual axons on metabolism. However, manual operation may result in unintended injury to neuronal cell bodies. In the future, this could potentially be addressed by employing computer-guided axon injury using laser ablation or controlled mechanical force. The second method utilizes the vacuum pump²⁰, which reduces human error and increases reproducibility through a standardized process. Importantly, this axonal injury method of a large-scale microfluidic device allows for the collection of substantially greater quantities of axonal material, thereby greatly facilitating omics analyses that demand large amounts of axons. Nevertheless, a major limitation is that the severed axons cannot be effectively retained in the terminal chamber. Operational failures are mainly related to bubble formation within the channels, which must be resolved by removing these bubbles with a high-suction vacuum pump before adding the culture medium.

Our protocol still has limitations. Although manual injury methods are highly suitable for high-throughput data acquisition on entire neurons, including damaged axons and dendrites, this technique, however, demands exceptional precision, and the entire injury process must be completed within 10 min to avoid unfavorable cell stress.

Compared with traditional axotomy models, the standardized workflow described here dramatically improves reproducibility, scalability, and compatibility with multi-omics platforms, ensuring sufficient detection of low abundance RNA, protein, and metabolites. This technical platform also provides a promising foundation for high-throughput screening of neuroregenerative compounds, systematic exploration of axonal metabolic regulation, and dissecting the local translation mechanisms after injury^27,28. By further enhancing the accessibility and compatibility of the equipment, this technology is expected to gain wider application in basic neuroscience research and therapeutic development.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Neurobasal-A Medium	Gibco	10888022	Culture medium
Alexa Fluor 488 goat anti-mouse	Invitrogen	A-21042	Immunofluorescence
Analytical balance	Shanghai Shunyu Hengping Scientific Instrument Co., Ltd.	FA1004	Precise weighing of reagents/samples
B27 (50x)	Gibco	17504044	Culture medium
BCA protein assay	Boster	AR1189	Protein concentrations
Bioanalyzer	Agilent	https://www.agilent.com/en/product/automated-electrophoresis/bioanalyzer-systems/bioanalyzer-instrument
Bovine serum albumin	Genview	FA016	Immunofluorescence
Centrifugal stirrer mixer	THINK	ARE-310	Mix PDMS and degasification
CLARIOstarPlus-ACU analyzer	BMG LABTECH	CLARIOstar Plus	protein concentrations
CO2 incubator	Thermo Fisher Scientific (Asheville) LLC	3111	Cell culture incubation at 37 °C, 5% CO?
Corrosion-resistant Diaphragm Vacuum Pump	Shanghai Lichen Bangxi Instrument Technology Co., Ltd.	LC-85DLC	Aspirate waste liquid and provide suction for axotomy
D-GLUCOSE (U-13C6, 99%)	Cambridge Isotope Laboratories	CLM-1396	Metabolic flux
DNase I	Roche	11284932001	Neuronal culture
Dr.TOM2 platform	BGI Genomics		Perform and analyze RNA-seq
Electric hot-air drying oven	BIOM Huahong Bonn / Huaying Bosi	DHG-9070A	Curing PDMS
Electronic Balance	Shanghai Huachao Industrial Co., Ltd.	HC313	Routine weighing of materials
Fluoromount Aqueous Mounting Medium	Sigma-Aldrich	F4680	Immunofluorescence
GlutaMAX	Gibco	35050061	Culture medium
Glycine	Solarbio	G8200	Immunofluorescence
HBSS	Gibco	14025092	Neuronal culture
HBSS, no calcium, no magnesium	Gibco	14175095	Neuronal culture
High-pressure steam sterilizer	STIK Instrument (Shanghai) Co., Ltd.	HMJ-54A	Sterilization of lab equipment and media
Hypoxic chamber	STEMCELL Technologies	27310	Hypoxic treatment
ImageJ	NIH		Used to analyze microscopy images
Leica Microscope	Leica Microsystems CMS GmbH	DMIL LED Fluo	Live and fluorescence observation
Live cell imaging solution	Invitrogen	A14291DJ	Live cell imaging
Medical low-temperature storage box (-80 °C)	Qingdao Haier Biomedical Co., Ltd.	DW-86L388J	Long-term sample storage at -80 °C
Mini desktop vacuum pump	Haimen Qilin-Bell Instrument Manufacturing Co., Ltd.	GL-805	Small-scale vacuum applications
MitoTracker Orange CMTMRos	Invitrogen	M7510	Mitochondrial membrane potential staining
MRMPROBS program		Version 2.60	To extract raw mass spectrometry data
Neurobasal Medium	Gibco	21103049	Culture medium
Neurobasal-A without glucose	Gibco	A2477501	Culture medium
Papain	Worthington	LS003127	Neuronal culture
Paraformaldehyde	Sigma-Aldrich	P6148	Cell fixation
PDMS	Dow Corning	DC184	Device fabrication
PDS Kit Inhibitor Vial	Worthington	LK003182	Neuronal culture
Penicillin Streptomycin	Gibco	15140122	Culture medium
Phosphate Buffered Saline	Biosharp	BL302A	Dilute the substances and clean the cell containers
Plasma cleaner	Harrick Plasma	ODC-32G-2	Activation
Poly-D-lysine	Sigma-Aldrich	P6407	Neuronal culture
Prism	GraphPad	Version 8.4.3	Statistical analyses
Refrigerated centrifuge	Eppendorf AG (Germany)	5430R	Sample separation at controlled temperatures
RNArep Pure Cell/Bacterial Total RNA Extraction Kit	Tiangen	dp430	Transcriptomics
Rotameter	Dwyer	RMA-23-SSV	Control the low oxygen flow
Stereo microscope	Olympus Corporation	SZ2-ILST	Sample observation and dissection
Stuart Orbital Shaker	Kylin-Bell	TS-200	Mixing cell cultures/solutions at adjustable speeds
Triton X-100	Sigma-Aldrich	X100	Immunofluorescence
ULVAC Rotary Vane Vacuum Pump	ULVAC (Ningbo) Co., Ltd.	GLD-N202	Creating vacuum for filtration/drying
βIII-tubulin	Promega	G7121	Immunofluorescence