Animals

All male C57BL/6 J mice were purchased from GemPharmatech Co., Ltd. PV-IRES-Cre transgenic mice and tdTomato Cre reporter Ai14 mice were provided by Zhou Yifeng’s research group (Hefei, China). PV-IRES-Cre; Ai14 mice were derived from crosses of the PV-IRES-Cre and Ai14 genotypes. Adult male mice (8–13 weeks old) were used in all experiments. For the amblyopia model, mice underwent monocular deprivation beginning at postnatal day 21 (P21) and were subsequently used in experiments after reaching adulthood. All animals were collectively housed in standardized enclosures with a regulated 12-h light/dark cycle. The animals had unrestricted access to both food and water, ensuring their basic needs were consistently met at all times. To minimize potential hormonal interference and maintain experimental control, only adult male mice were used in this study. All animal procedures were approved by the Institutional Animal Care and Use Committees of University of Science and Technology of China (protocol number: USTCACUC22240122017).

Light flickering stimulation protocol

The light flickering stimulation device used in this experiment consisted of a microcontroller and white light-emitting diodes (LEDs). White LEDs were arranged in parallel in a line. The frequencies were set at 20, 40, and 80 Hz (50% duty cycle). We applied visual stimulation to the mice based on a light flickering protocol previously reported [15]. Mice were placed in a dim and quiet experimental room for 30 min of acclimatization, followed by exposure to light flickering in the testing cage for 1 h. During the light flickering stimulation, mice were allowed to move freely within the cage. Three sides of the testing cage were wrapped with black plastic or fabric, and the transparent side faced the LED light source. The cage had no bedding or access to water. After the light flickering, mice were transferred to the housing cage for 30 min of rest. The light flickering exposure was administered daily from 9:00 to 11:00 a.m. The experiment continued for 7 consecutive days.

Eyelid suture

Short-term monocular deprivation involved suturing the contralateral eye of mice for 4 days before in vivo electrophysiological recordings. Mice were kept under anesthesia with 1–3% isoflurane. To establish an amblyopia model, mice underwent long-term deprivation, starting from postnatal day 21 (P21) and continuing until adulthood (> P60). In adulthood, reverse suturing was performed on amblyopic mice, opening the long-term deprived eye while suturing the normal eye that was not deprived. Following the surgery, the mice underwent daily inspections, and chloramphenicol eye drops and erythromycin eye ointment were administered to guard against ocular opacification. Any mice that exhibited premature eyelid opening or cloudiness in their eyes before recording were excluded from subsequent experiments.

Drug administration

Minocycline was diluted in 0.9% saline and administered to mice via intraperitoneal injection, while the control group received injections of saline. Mice were injected with minocycline for 14 consecutive days prior to the start of electrophysiological recordings, at a dose of 50 mg/kg per day.

Viral intracranial injection

In this experiment, adeno-associated virus (AAV) carrying the CMV promoter was used. The RNA sequence for interfering with LCN2 was CCAGTTCACTCTGGGAAAT, and the control group was designed with a non-specific sequence TTCTCCGAACGTGTCACGT. For the experimental group of mice, AAV-U6-shRNA(lcn2)-CMV-EGFP (5.00 × 1012 vg/ml) was injected into the visual cortex, while the control group received an injection of AAV-U6-shRNA (scramble)-CMV-EGFP (3.36 × 1012 vg/ml). Mice received intracortical injections of rAAV-CMV-LCN2-P2A-EGFP-WPRE-Hgh pA (5 × 1012 vg/ml) to upregulate LCN2, while the control group received injections of rAAV-CMV-EGFP-WPRE-hGH pA (5.19 × 1012 vg/ml). AAV viruses were bilaterally delivered into the V1 region, 250 nL for each hemisphere. All viruses were purchased from BrainVTA.

Mice were anesthetized with 2–4% isoflurane, and their heads were secured on a stereotaxic apparatus. The mouse skin was incised, tissues were dissected, and a hole in the skull at the position of the visual cortex was drilled using a skull drill. The specific coordinates were as follows: 1 mm anterior to the Lambda point, 2.75 mm lateral to the midline, and 0.5 mm below the cortex. Using a 1-μL microsampler, the virus was injected at a rate of 20 nL/min to the desired depth, with an injection volume of 250–300 nL. After completing the injection, the microsampler was slowly withdrawn after waiting for 10 min to prevent viral overflow. The incision on the mouse’s scalp was then sutured with a sterilized needle. After surgery, daily observations were conducted to ensure the well-being of the mice. To evaluate the knockdown efficacy, we quantified the expression level of LCN2 protein in the visual cortex, 3 weeks following the administration of LCN2-AAV or Control-AAV.

In vivo electrophysiological recording

Adult mice underwent anesthesia using urethane (2 g/kg, i.p.) and chlorprothixene (5 mg/kg, i.m.). Following anesthesia, the mice were positioned in a stereotaxic apparatus to ensure immobility during the recording. Concurrently, body temperature was constantly monitored and adjusted to maintain at 37 °C using a heating blanket. To facilitate recording, a craniotomy was conducted over the V1b region.

Recording of the electrophysiological properties of V1b region cells was conducted in adult mice without MD. The visual stimulus, generated using MATLAB, was displayed on a monitor positioned 23 cm in front of the eyes of the mouse. A sinusoidal drifting grating was presented across 0° to 330° in 12 steps of 30° intervals (spatial frequency: 0.05 cycles per degree, temporal frequency: 2 Hz) to measure orientation/direction selectivity. The formula for calculating the Orientation Selectivity Index (OSI) was (Rpref − Rorth)/(Rpref + Rorth), where Rpref was the response to the preferred direction, and Rorth represented the average response to two directions orthogonal to the preferred direction. The formula for calculating the Direction Selectivity Index (DSI) was (Rpref − Ropp)/(Rpref + Ropp), where Ropp was the average response to the opposite direction.

To examine the ODI for each cell, the optimal parameter of the visual stimuli such as direction, moving speed, and bar width were firstly assessed within its receptive field, then the responses driven by each eye were recorded under conditions where one eye was alternately occluded, while identical optimal visual stimuli were displayed on the monitor. All recordings underwent spike sorting to ensure that over 90% of spikes originated from the same neuron. Recordings were conducted throughout the entire thickness of the V1 region. Each mouse underwent recording from 18–24 cells across 5 vertical penetrations, ensuring even spacing (with a minimum interval of 200 μm) across the mediolateral extent of V1b to avoid sampling bias. Only cells with receptive fields located within the binocular area (0° to 20°) were included in the analysis. Cells were categorized into OD groups based on seven category schemes (n1-n7) according to the responses to input from both eyes: C/I > 10, 10 > C/I > 3, 3 > C/I > 3/2, 3/2 > C/I > 2/3, 2/3 > C/I > 1/3, 1/3 > C/I > 0.1, and C/I < 0.1. Cells in the n1 group were primarily driven by contralateral input, while those in the n7 group were predominantly influenced by ipsilateral input. The response of each cell was calculated by subtracting its spontaneous activity from its visually driven response. The ocular dominance index (ODI) was then determined using the formula: (I − C)/(I + C), where C and I represented the evoked responses of contralateral and ipsilateral, respectively.

Cortical slice preparation and mIPSC recording

Briefly, mice were anesthetized using isoflurane and perfused transcardially with an ice-cold buffered solution containing the following: 212.7 mM sucrose, 3 mM KCl, 1.25 mM NaH2PO4, 3 mM MgCl2, 1 mM CaCl2, 26 mM NaHCO3, and 10 mM dextrose, bubbled with 95% O2/5% CO2. Using a tissue slicer (Vibratome 3000; Vibratome), coronal slices from V1b (300 μm) were precisely crafted in cold dissection buffer identical to the perfusion solution. After preparation, the slices were promptly transferred to a solution of artificial cerebrospinal fluid (ACSF) maintained at 35 °C and allowed to equilibrate for a period of 30 min prior to initiating the recordings. The composition of ACSF mirrored that of the dissection buffer, with the notable variation: substituting 124 mM NaCl in place of sucrose, while adjusting the MgCl2 and CaCl2 concentrations to 1 and 2 mM, respectively. Throughout the recording process, the temperature was maintained at 28–30 °C. Under infrared differential interference contrast optics, pyramidal neurons situated within layer 2/3 of the V1b region were visually discerned and selectively targeted for the purpose of electrophysiological recording. Miniature inhibitory postsynaptic currents (mIPSCs) were recorded in voltage-clamp mode at a holding potential of − 60 mV using a CsCl-based internal solution (as in eIPSC recordings). To pharmacologically isolate inhibitory synaptic currents and eliminate network activity, 1 μM tetrodotoxin (TTX), 100 μM 2-amino-5-phosphonovaleric acid (APV), and 20 μM 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) were added to the ACSF. All events in a given cell (ranging from 500 to 2000) were analyzed for frequency measurements using Mini Analysis software (Synaptosoft, Decatur, GA), while “bursts” with superimposed events were excluded from amplitude analysis. The decay constant was calculated from the average of the first 100 isolated events and fitted with a single-exponential function. Electrophysiological recordings were captured utilizing an Integrated Patch-Clamp Amplifier (Sutter Instrument, Novato, CA, USA), which was operated by lgor 7 software (WaveMetrics, Portland, OR, USA). The sampling rate for these recordings was 20 kHz, with filtering applied at 5 kHz. Furthermore, Igor 7 software facilitated both the acquisition and subsequent analysis of the recorded data.

RNA-seq

After being anesthetized with isoflurane, the mice underwent intracardial perfusion with ice-cold phosphate-buffered saline (PBS). Subsequently, brain slices were precisely prepared for the purpose of electrophysiological recording, and V1 samples were manually dissected on ice-cold surface. The extraction of total RNA from the tissue was performed using TRIzol® Reagent strictly according the manufacturer’s instructions (Invitrogen). To remove genomic DNA, DNase I (TaKara) was employed.

Subsequently, the quality of RNA was assessed utilizing the 2100 Bioanalyzer (Agilent), while quantification was performed with the ND-2000 (NanoDrop Technologies). High-grade RNA samples were then selected for the construction of sequencing libraries. These RNA-seq transcriptome libraries were prepared according to the protocol outlined in the TruSeqTM RNA Sample preparation Kit provided by Illumina (San Diego, CA), employing 1 μg of total RNA as the starting material. The process of cDNA synthesis, end repair, A-base addition, and ligation of Illumina-indexed adapters performed according to Illumina’s protocol. Following this, the libraries underwent size selection to isolate cDNA target fragments ranging from 200 to 300 bp on a 2% Low Range Ultra Agarose gel. Subsequently, these libraries were PCR-amplified using Phusion DNA polymerase (NEB) for a total of 15 cycles. After quantification using TBS380, the paired-end libraries were subjected to sequencing on the Illumina NovaSeq 6000 platform (150 bp*2, Shanghai BIOZERON Co., Ltd).

The foundation of all downstream analyses rested on high-quality, thoroughly cleaned data. To bolster the utility of these reads, we leveraged hisat2 software (https://ccb.jhu.edu/software/hisat2/index.shtml) to map the clean data against the reference genome sequence. Strictly, only those reads exhibiting a perfect match or a single mismatch were considered for further analysis and annotation, ensuring precision and accuracy in our findings. We used the edgeR package (Empirical analysis of Digital Gene Expression in R) to identify differentially expressed genes in all the transcriptomic data. To control the false positive rate, the Benjamin-Hochberg (FDR) method was applied to test the p-values, resulting in corrected p-values known as FDR, where a smaller FDR indicates greater significance. The DEGs between the two samples were identified based on rigorous criteria. Specifically, a gene was considered differentially expressed if it exhibited a logarithmic fold change greater than 2 and had a false discovery rate (FDR) below 0.05, ensuring that the detected differences were statistically significant and biologically meaningful.

Real-time PCR

Mice were rapidly dissected under deep anesthesia with ether to obtain V1 tissues, which were then placed into grinding tubes containing RNA extraction solution. After thorough grinding until no visible tissue chunks were present, RNA was extracted, washed, and dissolved in RNA dissolution solution. The concentration and purity of the RNA were assessed using Nanodrop 2000. Following the determination of RNA content for each sample, reverse transcription was performed to synthesize cDNA, and the synthesized cDNA was subsequently subjected to PCR amplification. Quantitative analysis of the target gene was conducted using the 2-∆∆Ct method.

Immunohistochemistry

Mice were anesthetized through intraperitoneal administration of a ketamine (0.1 mg/g) and xylazine (0.01 mg/g) mixture. After complete anesthesia, cardiac perfusion was performed. The brain was initially subjected to perfusion with ice-cold PBS to rinse out any residual blood or contaminants. Subsequently, it was immersed in 4% formaldehyde for fixation. Following overnight post-fixation in 4% formaldehyde, the brain was transferred to a solution of 30% sucrose dissolved in 0.01 M PBS for 48 h. Brain tissues were sliced into 40-μm coronal sections using a freezing microtome (Leica). After treatment with 0.5% Triton X-100 and blocking (5% BSA in 0.01 M PBS) for 1 h, free-floating sections were co-incubated overnight at 4 °C with primary antibody (Anti-Iba1 antibody, 1:500, Abcam, ab178846; Anti-GFAP antibody, 1:400, Cell Signaling Technology, 12,389) or biotin-labeled lectin from Wisteria floribunda agglutinin (WFA) (1:500, Sigma-Aldrich, L1516). On the following day, brain slices were washed with PBS and co-incubated at room temperature with the corresponding Alexa Fluor 488 secondary antibody (1:500, ThermoFisher) for 90 min. The sections were then mounted on slides using Antifade Polyvinylpyrrolidone Mounting Medium (Beyotime). Subsequently, images of the slices were captured using a Leica high-resolution confocal microscope with constant parameters. The density of cells surrounded by WFA-positive perineuronal nets was quantified within a 550 × 300 μm2 region covering the entire thickness of V1b (10 fields of view per mouse).

Sholl analysis

Following acquisition of confocal images of microglia in the visual cortex, branch length and branch endpoints were analyzed using the AnalyzeSkeleton (2D/3D) plugin in ImageJ software (Fiji edition, NIH). Complexity analysis of microglia was performed using the Sholl plugin in ImageJ software. For this, individual microglia with relatively intact morphology were selected, binary conversion was applied to convert signals to 1 and background to 0, and then saved. Subsequently, the Sholl analysis plugin was employed to set the initial shell at 6 μm, with subsequent shells incremented by 2 μm each. This allowed for the determination of intersection counts at each defined Sholl radius. The complexity of microglia was determined by analyzing the number of intersections between elongated branches and concentric circles.

Western blot analysis

The primary visual cortex of adult mice was rapidly dissected after deep anesthesia with ether and placed in Eppendorf tubes containing RIPA lysis buffer. Following SDS-PAGE electrophoresis, proteins were transferred onto a polyvinylidene difluoride (PVDF) membrane. The membrane was blocked with 5% non-fat milk (w/v in 20 mM TBST) for 1 h at room temperature, followed by overnight incubation at 4 °C with rabbit anti-LCN2 (1:1000, Affinity, DF6816), rabbit anti-GAD65 (1:2,000, Proteintech, 20,746–1-AP), mouse anti-GAD67 (1:5,000, Millipore, MAB5406), and mouse anti-β-actin (1:10,000, Abcam, ab6276) antibodies. Subsequently, the membrane was exposed to the corresponding HRP-conjugated secondary antibodies (1:5000, Promega) for 1 h at room temperature. Membrane visualization was achieved using Super Signal West Pico chemiluminescent substrate (Pierce). ImageJ was employed to determine the optical density of each band, normalized to β-actin.

Statistical analysis

GraphPad Prism 8.4.2 (GraphPad Software, California, USA) was used to evaluate all of the data. Data are presented as mean ± standard error of the mean (SEM). Statistical significance between two groups was analyzed using unpaired Student’s t test or one-way analysis of variance (ANOVA) with post hoc Tukey test. Differences between data influenced by two factors were analyzed using two-way ANOVA with Bonferroni post hoc test. The cumulative distribution of the groups was compared using the Kolmogorov–Smirnov test (K-S test). Differences were considered significant when the p-value was less than 0.05.