Browse the corpus

Effects of Repeated Early-life Anesthesia Exposure on Visual System Development in Mice. BACKGROUND: General anesthesia is administered to millions of children annually, yet its long-term effects on neurodevelopment remain a concern. It was previously reported that even a single early exposure to general anesthesia for minor surgery impairs visual attention in children. This study investigates the effects of early repeated general anesthesia exposure on visual system maturation in mice and explores the role of tissue-type plasminogen activator in mediating these effects during development. METHODS: Male SWISS and C57BL/6J mice (wild-type or deficient for tissue-type plasminogen activator) were exposed to general anesthesia with 1.3% isoflurane in 50% oxygen for 90 min per day from postnatal days 4 to 10. Control animals received 50% oxygen alone. Visual system integrity and inflammation were assessed at postnatal day 15 and at 6 weeks using behavioral tests, high-resolution imaging, and immunohistochemistry. In SWISS mice, circulating tissue-type plasminogen activator levels were measured using a biochemical approach, and neurovascular coupling was evaluated by functional ultrasound imaging. RESULTS: Early repeated general anesthesia exposure delayed eyelid opening (median postnatal day 13 [95% CI, 0.52 to 1.45] vs. postnatal day 15 [95% CI, 0.62 to 1.92]; P < 0.0001), caused lasting visual function deficits (depth perception and oculomotor reflex), and reduced retinal (0.2627 ± 0.04 mm vs. 0.1667 ± 0.03 mm; P < 0.0001) and primary visual cortex thickness (0.8000 ± 0.08 mm vs. 0.7282 ± 0.05 mm; P = 0.0235). Notably, lower circulating tissue-type plasminogen activator levels were observed in general anesthesia-exposed SWISS mice (11.580 ± 2.19 ng/ml vs. 7.654 ± 1.31 ng/ml; P = 0.0082). Tissue-type plasminogen activator-deficient mice exhibited attenuated or absent general anesthesia-induced visual alterations. CONCLUSIONS: These findings indicate that early repeated exposure to general anesthesia disrupts visual system maturation in mice and suggest that altered tissue-type plasminogen activator pathways may contribute to these effects, identifying tissue-type plasminogen activator as a potential marker of anesthesia-related neurodevelopmental vulnerability. Additional experimental work will be required to further support this association and clarify its underlying mechanisms.

In preclinical experimental studies, early-life exposure to general anesthetics have been shown to have a variety of both morphologic and functional effects on central nervous system development, but the translational relevance of these observations remains unclear Recent retrospective human observations suggest an association between early-life anesthesia exposure and visual attention bias in a small number of children The question of whether early-life anesthesia exposure alters visual system development in preclinical experimental models remains incompletely explored In newborn mice, repeated exposure during 7 consecutive days to 1.5 h of isoflurane anesthesia induced lasting visual function deficits that were associated with reduced retinal and visual cortical thickness, as well as with lower plasma levels of tissue plasminogen activator Genetically modified mice lacking tissue plasminogen activator displayed a phenotype in the absence of anesthesia comparable to that of wild-type mice exposed to repeated anesthesia early in life In contrast to wild-type counterparts, neonatal genetically modified mice lacking tissue plasminogen activator did not experience further impairment of the visual system after repeated isoflurane exposure

Genetically modified mice lacking tissue plasminogen activator displayed a phenotype in the absence of anesthesia comparable to that of wild-type mice exposed to repeated anesthesia early in life In contrast to wild-type counterparts, neonatal genetically modified mice lacking tissue plasminogen activator did not experience further impairment of the visual system after repeated isoflurane exposure While early general anesthesia is considered clinically safe in the short term, with a low incidence of perioperative complications,1 growing evidence suggests potential long-term effects on central nervous system (CNS) development, particularly within the visual system. As numerous visual functions mature within the first year of life, this system may be vulnerable to early external interventions, including general anesthesia.2 Preclinical and clinical studies have reported persistent brain structure and behavioral alterations after early general anesthesia exposure,3,4 and some human studies suggest adverse neurodevelopmental outcomes.5–10 Large-scale trials such as General Anesthesia Spinal (GAS), Pediatric Anesthesia Neurodevelopment Assessment (PANDA), and Mayo Anesthesia Safety in Kids (MASK), as well as two recent randomized studies, found no major long-term impairments in tests assessing global intelligence after single or repeated exposures,11–13 yet subtle or domain-specific alterations cannot be excluded. Our previous retrospective analysis revealed modified visual attention in children exposed to general anesthesia early in life, marked by a stronger visual attention bias toward global processing strategies, which may affect learning capacity.14 Experimental work in rodents and primates further supports similar general anesthesia–induced neurodevelopmental effects, showing increased apoptosis in the retina and visual cortex.15–18 Collectively, these findings indicate that early general anesthesia may affect structural and functional maturation within the developing visual system.

rodents and primates further supports similar general anesthesia–induced neurodevelopmental effects, showing increased apoptosis in the retina and visual cortex.15–18 Collectively, these findings indicate that early general anesthesia may affect structural and functional maturation within the developing visual system. When looking for potential mechanistic biomarkers of general anesthesia–induced neurodevelopmental alterations,10 tissue-type plasminogen activator (tPA) emerges as a compelling candidate, as this serine protease exerts pleiotropic functions in the CNS. By interacting with N-methyl-d-aspartate receptors and regulating extracellular matrix remodeling, tPA modulates key neurodevelopmental processes including neuronal migration, synaptogenesis, and cortical lamination.19–21 In the developing eye, tPA is expressed in the lens and retina, where it governs extracellular matrix dynamics, cell proliferation, and excitotoxic responses.22–24 Its regulation by the early growth response protein 1 (Egr1) and by epidermal growth factor receptor signaling, both of which are involved in eyelid and corneal development,25–28 suggests that tPA could influence developmental events coinciding with eyelid opening and visual system maturation, potentially susceptible to disruption by general anesthesia. Moreover, its implication in excitotoxic retinal injury and retinal ganglion cell survival underscores its relevance to neuronal vulnerability and general anesthesia–induced neurotoxicity.23,24 Together, these observations led us to identify tPA as a candidate mediator of early general anesthesia–related neurodevelopmental effects in the visual system.

otoxic retinal injury and retinal ganglion cell survival underscores its relevance to neuronal vulnerability and general anesthesia–induced neurotoxicity.23,24 Together, these observations led us to identify tPA as a candidate mediator of early general anesthesia–related neurodevelopmental effects in the visual system. Although early general anesthesia exposure is known to affect neurodevelopment, the underlying cellular and molecular mechanisms within the visual system remain poorly understood. In particular, the potential contribution of tPA—whose activity-dependent regulation is essential for synaptic remodeling and maturation—to general anesthesia–induced alterations in developmental timing, structural organization, and functional integrity has never been explored. We hypothesized a cross-talk between tPA and general anesthesia–induced neurodevelopmental alterations along the visual pathway that we investigated from the retina to cortical visual areas. To test this hypothesis, we employed a murine model of repeated early-life general anesthesia exposure in both wild-type (tPAWT) and tPA-deficient (tPAnull) mice. The primary outcome was defined as assessing the effects of early general anesthesia on visual system development across macrodevelopmental, anatomic, functional, and molecular levels. Secondary outcome aimed to determine whether tPA contributes to these general anesthesia–induced alterations, through parallel assessments.

In preclinical experimental studies, early-life exposure to general anesthetics have been shown to have a variety of both morphologic and functional effects on central nervous system development, but the translational relevance of these observations remains unclear Recent retrospective human observations suggest an association between early-life anesthesia exposure and visual attention bias in a small number of children The question of whether early-life anesthesia exposure alters visual system development in preclinical experimental models remains incompletely explored

In newborn mice, repeated exposure during 7 consecutive days to 1.5 h of isoflurane anesthesia induced lasting visual function deficits that were associated with reduced retinal and visual cortical thickness, as well as with lower plasma levels of tissue plasminogen activator Genetically modified mice lacking tissue plasminogen activator displayed a phenotype in the absence of anesthesia comparable to that of wild-type mice exposed to repeated anesthesia early in life In contrast to wild-type counterparts, neonatal genetically modified mice lacking tissue plasminogen activator did not experience further impairment of the visual system after repeated isoflurane exposure

Detailed methods are provided in the Supplemental Digital Content (https://links.lww.com/ALN/E324). Experiments were conducted on postnatal day 15 (P15) and 6- to 8-week-old male SWISS mice (Janvier Labs, France) and 6- to 8-week-old male tPAnull and tPAWT mice (C57BL/6J; Normandy University, Caen, France). All protocols were approved by CENOMEXA (agreement No. D14118001) in compliance with European Union directive 2013/63/UE and French regulations (French Ministry of Research, license Nos. 8972 and 46951). Only healthy pups showing normal development before anesthesia exposure were included in the study. Clinical signs were monitored twice daily by animal facility staff and/or experimenters. Animals were excluded if they exhibited visible signs of pain or prolonged prostration, a high grimace scale score (score higher than 7 for more than 48 h), absence of response to external stimuli (e.g., paw pinch), cessation of food or water intake, or body weight loss exceeding 15% of initial weight. The animals were randomized at birth, with control and general anesthesia–exposed pups mixed within each litter. After weaning (postnatal day 21), the animals were separated by condition to avoid long-term confusion. Behavioral experiments and analyses were performed blind to condition whenever possible. Blinding was not feasible for magnetic resonance imaging (MRI) and functional ultrasound (fUS) experiments due to practical constraints requiring animal identification after imaging. Two experimental designs were used. For the SWISS mice, the animals were assigned to either a control group or a group exposed to early general anesthesia. In C57BL/6J mice, four groups were included (tPAWT control, tPAWT general anesthesia, tPAnull control, and tPAnull general anesthesia). Each individual animal was considered as an independent experimental unit (see Supplemental Digital Content for details, https://links.lww.com/ALN/E324).

d to early general anesthesia. In C57BL/6J mice, four groups were included (tPAWT control, tPAWT general anesthesia, tPAnull control, and tPAnull general anesthesia). Each individual animal was considered as an independent experimental unit (see Supplemental Digital Content for details, https://links.lww.com/ALN/E324). To model repeated early general anesthesia exposure, SWISS, tPAnull, and tPAWT mice were randomly assigned to once daily exposure from P4 to P10, during 7 consecutive days, replicating the paradigm described in a recently published study.3 The general anesthesia group received 1.3% isoflurane (1 minimum alveolar concentration) in 50% O2 for 90 min; controls received 50% O2 alone: control pups were separated from the dams for the same duration as anesthetized pups and exposed to 50% O2 under identical environmental conditions, without anesthesia. Body temperature was monitored throughout general anesthesia exposure. Anesthetic depth was assessed by the loss of righting reflex and paw withdrawal reflex, as well as by monitoring heart and respiratory rates using a dedicated small animal monitoring system (SA Instruments Inc., USA; see Supplemental Digital Content for details, https://links.lww.com/ALN/E324). To assess repeated early developmental outcomes after general anesthesia, body weight was monitored daily from P4 to P10 (preexposure) and then weekly. Eyelid opening was scored daily from P12 to P18, with ocular development tracked until behavioral and imaging experiments (see Supplemental Digital Content for details, https://links.lww.com/ALN/E324).

ental outcomes after general anesthesia, body weight was monitored daily from P4 to P10 (preexposure) and then weekly. Eyelid opening was scored daily from P12 to P18, with ocular development tracked until behavioral and imaging experiments (see Supplemental Digital Content for details, https://links.lww.com/ALN/E324). At P15 and 6 weeks, mice underwent four noninvasive behavioral tests to evaluate their visual abilities, including oculomotor reflex, depth perception, and visual acuity (see Supplemental Digital Content for more details, https://links.lww.com/ALN/E324). Oculomotor reflexes were assessed by the eye blinking reflex (EBR) test. A fine tip (0.7 mm) was brought close to the center of the cornea while the mouse head movements were limited by light hand restraint. Precautions were taken to avoid contact with the whiskers to prevent sensory stimulation. A score ranging from 0 to 9 for each eye was assigned. This scoring system was adapted from a previously published scale.29 Depth perception was assessed using the vertical placement test (VPT; scored 0 to 2) and the visual cliff test (VCT), where mice explored a two-zone arena (“safe” vs. “deep”; 0 vs. 30 cm floor drop). Anxiety-like behavior was assessed using the VCT open field and was recorded for 5 min and analyzed with Noldus EthoVision (distance, velocity, and zone preference; supplemental data 1A, https://links.lww.com/ALN/E316).

cliff test (VCT), where mice explored a two-zone arena (“safe” vs. “deep”; 0 vs. 30 cm floor drop). Anxiety-like behavior was assessed using the VCT open field and was recorded for 5 min and analyzed with Noldus EthoVision (distance, velocity, and zone preference; supplemental data 1A, https://links.lww.com/ALN/E316). Visual acuity was assessed via the optokinetic reflex test, using head-tracking responses to rotate sine wave gratings (12°/s) in a virtual cylinder. Spatial frequency thresholds were determined by increasing grating frequency until tracking stopped. The test was conducted separately for each eye by reversing the cylinder’s rotation direction. This test was only performed on tPAnull and tPAWT mice because of their healthy visual acuity, unlike SWISS mice, which are known to have an albino phenotype.

determined by increasing grating frequency until tracking stopped. The test was conducted separately for each eye by reversing the cylinder’s rotation direction. This test was only performed on tPAnull and tPAWT mice because of their healthy visual acuity, unlike SWISS mice, which are known to have an albino phenotype. At 6 weeks, the mice were subjected to various MRI sequences under general anesthesia (isoflurane 2% and 30%/70% O2/N2O), including injection of contrast agents for molecular MRI, delivered through a caudal catheter. To do so, a catheter was placed in the tail vein under anesthesia (5% isoflurane induction, maintained at 1 to 2% isoflurane with 30% O2/70% N2O during acquisitions). The mice were thermoregulated at 38.0°C with a self-regulated heating mat, and ophthalmic gel was applied before catheter insertion. Imaging exams were performed with a 7T MRI (Bruker, Germany), and the data were acquired with Paravision 6.0.1 software (Bruker). The sequences provided anatomic, functional, and molecular data. MRI imaging was performed as a terminal procedure, as it required additional anesthesia and the administration of contrast agents (see Supplemental Digital Content for more details, https://links.lww.com/ALN/E324). For the anatomic assessment of the brain and eyes, T2-weighted images were acquired using a multislice multiecho sequence. Areas and thickness of multiple brain and eye parts were manually segmented on T2 acquisitions using ImageJ software (National Institutes of Health, USA).

At 6 weeks, the mice were subjected to various MRI sequences under general anesthesia (isoflurane 2% and 30%/70% O2/N2O), including injection of contrast agents for molecular MRI, delivered through a caudal catheter. To do so, a catheter was placed in the tail vein under anesthesia (5% isoflurane induction, maintained at 1 to 2% isoflurane with 30% O2/70% N2O during acquisitions). The mice were thermoregulated at 38.0°C with a self-regulated heating mat, and ophthalmic gel was applied before catheter insertion. Imaging exams were performed with a 7T MRI (Bruker, Germany), and the data were acquired with Paravision 6.0.1 software (Bruker). The sequences provided anatomic, functional, and molecular data. MRI imaging was performed as a terminal procedure, as it required additional anesthesia and the administration of contrast agents (see Supplemental Digital Content for more details, https://links.lww.com/ALN/E324). For the anatomic assessment of the brain and eyes, T2-weighted images were acquired using a multislice multiecho sequence. Areas and thickness of multiple brain and eye parts were manually segmented on T2 acquisitions using ImageJ software (National Institutes of Health, USA). For the functional assessment of specific brain areas, arterial spin labeling sequences were acquired, translating tissue perfusion through the measurement of cerebral blood volume in the somatosensory cortex, the hippocampus, the amygdala, the visual primary cortex, and the periaqueductal gray matter.

For the anatomic assessment of the brain and eyes, T2-weighted images were acquired using a multislice multiecho sequence. Areas and thickness of multiple brain and eye parts were manually segmented on T2 acquisitions using ImageJ software (National Institutes of Health, USA). For the functional assessment of specific brain areas, arterial spin labeling sequences were acquired, translating tissue perfusion through the measurement of cerebral blood volume in the somatosensory cortex, the hippocampus, the amygdala, the visual primary cortex, and the periaqueductal gray matter. Molecular MRI assessed cerebral and retinal vascular inflammation through intravenous injection of iron oxide microparticles (MPIOs) linked to antibodies against endothelial adhesion molecules (P-selectin and vascular cell adhesion molecule-1 [VCAM-1]). The animals were first injected with 100 μl MPIOs–P-selectin (1 mg/kg iron) and second with 100 μl MPIOs–VCAM-1 (1 mg/kg iron) through the catheter placed in the caudal vein. The injection of MPIOs coupled to IgG served as a negative control by revealing the nonspecific fixation of the particles to the endothelium (Supplemental Digital Content, https://links.lww.com/ALN/E324; table 1). Labeled MPIOs were used as a contrast agent, were obtained in the laboratory, and have been previously described.30,31 Antibodies Used in the Project Biotechne (USA); BD Biosciences (USA); Abcam (USA); Jackson Immunoresearch (USA); R&D Systems (USA): BD Pharmingen (USA); Miltenyi Biotec (Germany); BioLegend (USA); Invitrogen (USA).

Molecular MRI assessed cerebral and retinal vascular inflammation through intravenous injection of iron oxide microparticles (MPIOs) linked to antibodies against endothelial adhesion molecules (P-selectin and vascular cell adhesion molecule-1 [VCAM-1]). The animals were first injected with 100 μl MPIOs–P-selectin (1 mg/kg iron) and second with 100 μl MPIOs–VCAM-1 (1 mg/kg iron) through the catheter placed in the caudal vein. The injection of MPIOs coupled to IgG served as a negative control by revealing the nonspecific fixation of the particles to the endothelium (Supplemental Digital Content, https://links.lww.com/ALN/E324; table 1). Labeled MPIOs were used as a contrast agent, were obtained in the laboratory, and have been previously described.30,31 Antibodies Used in the Project Biotechne (USA); BD Biosciences (USA); Abcam (USA); Jackson Immunoresearch (USA); R&D Systems (USA): BD Pharmingen (USA); Miltenyi Biotec (Germany); BioLegend (USA); Invitrogen (USA). Anesthetized mice (isoflurane 1 to 2% and 30% O2/70% N2O) were placed in a stereotaxic frame and thermoregulated at 38.0°C, and an ophthalmic gel was applied to the eyes. Analgesia was achieved by subcutaneous injection of buprenorphin (0.1 mg/kg). The skull was exposed for acquisition. Whiskers were cut to a length of 1 cm. Medetomidine (Domitor, Vetoquinol, France), 0.08 mg/kg induction flowed by 0.16 mg · kg−1 · h−1 subcutaneous infusion) was used for maintenance.

pplied to the eyes. Analgesia was achieved by subcutaneous injection of buprenorphin (0.1 mg/kg). The skull was exposed for acquisition. Whiskers were cut to a length of 1 cm. Medetomidine (Domitor, Vetoquinol, France), 0.08 mg/kg induction flowed by 0.16 mg · kg−1 · h−1 subcutaneous infusion) was used for maintenance. Cerebral blood volume and cerebral blood flow (CBF) were measured by transcranial Power Doppler ultrasound using an ultrafast scanner (Iconeus System, France). Neurovascular coupling (NVC) was assessed via CBF changes in the primary visual cortex (PVC), lateral geniculate nucleus (LGN), and superior colliculus (SC) after visual stimulation and in the ventral posterior–medial nucleus (VPM), posterior complex (PC) of the thalamus, and barrel cortex after whisker stimulation (see Supplemental Digital Content for details, https://links.lww.com/ALN/E324). Visual stimulation was applied using an LED light positioned 20 cm from the mouse eye, with a 60-s baseline followed by four 30-s stimuli (200-ms pulses at 4 Hz, repeated every 90 s) for a total of 420 s. Cerebrovascular hemodynamic responses were assessed by averaging cerebral blood volume variations across four trials within an imaging session. Both eyes were stimulated. Sensory responses in the barrel cortex were assessed through whisker stimulations. The stimulation pattern and the cerebrovascular hemodynamic responses are the same as previously described under “Visual Stimulation” (with mechanical whisker stimulation, 4 Hz).

Visual stimulation was applied using an LED light positioned 20 cm from the mouse eye, with a 60-s baseline followed by four 30-s stimuli (200-ms pulses at 4 Hz, repeated every 90 s) for a total of 420 s. Cerebrovascular hemodynamic responses were assessed by averaging cerebral blood volume variations across four trials within an imaging session. Both eyes were stimulated. Sensory responses in the barrel cortex were assessed through whisker stimulations. The stimulation pattern and the cerebrovascular hemodynamic responses are the same as previously described under “Visual Stimulation” (with mechanical whisker stimulation, 4 Hz). Blood and tissue samples were collected at P15 and 6 to 8 weeks. After induction (isoflurane 5%, 30% O2/70% N2O), anesthesia was reinforced with local anesthesia (lidocaine 5%). Analgesia was achieved by subcutaneous injection of buprenorphin (0.1 mg/kg). After deep anesthesia, ventricular blood was drawn, plasma was isolated, and samples were stored at −80°C. After perfusion, brains, eyes, and spleens were harvested (see Supplemental Digital Content for more details, https://links.lww.com/ALN/E324). For detection of total tPA in plasma, tPA levels were measured in P15 and 6- to 8-week-old mice using a total tPA enzyme-linked immunosorbent assay kit according to the manufacturer’s instructions (MI-MTPAKT-TOT; Gentaur, Inc., Belgium; see Supplemental Digital Content for more details, https://links.lww.com/ALN/E324).

For detection of total tPA in plasma, tPA levels were measured in P15 and 6- to 8-week-old mice using a total tPA enzyme-linked immunosorbent assay kit according to the manufacturer’s instructions (MI-MTPAKT-TOT; Gentaur, Inc., Belgium; see Supplemental Digital Content for more details, https://links.lww.com/ALN/E324). The retinas were flattened and cut into quarters. Brain sections (10 μm) were mounted on polylysine-coated slides and stored at −80°C before processing. Sections were incubated overnight at room temperature with primary antibodies diluted in a blocking solution (1× phosphate-buffered saline + 0.025% Triton X-100). Antibodies for immunohistochemistry are detailed in the Supplemental Digital Content (https://links.lww.com/ALN/E324) and table 1. Sagittal 10-μm cryosections were hematoxylin and eosin–stained (Biognost kit) and mounted on polylysine-coated slides. Retinal thickness was measured in central, proximal, and distal regions across all layers: ganglion cell layer, inner plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer, outer nuclear layer (ONL), photoreceptors, retinal pigmented epithelium, and choroid; using ImageJ (National Institutes of Health; see Supplemental Digital Content for details, https://links.lww.com/ALN/E324).

ll layers: ganglion cell layer, inner plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer, outer nuclear layer (ONL), photoreceptors, retinal pigmented epithelium, and choroid; using ImageJ (National Institutes of Health; see Supplemental Digital Content for details, https://links.lww.com/ALN/E324). To assess peripheral inflammation, the spleens were collected for flow cytometry. After Fc blocking (anti-CD16/CD32), the cells were stained with fluorochrome-conjugated antibodies and 7-aminoactinomycin D, including intracellular markers (kit from Miltenyi Biotec, Germany), and analyzed on a FACSVerse (BD Biosciences, USA) using FlowJo 7.6.5. Antibody details are in the Supplemental Digital Content (https://links.lww.com/ALN/E324) and table 1.

were stained with fluorochrome-conjugated antibodies and 7-aminoactinomycin D, including intracellular markers (kit from Miltenyi Biotec, Germany), and analyzed on a FACSVerse (BD Biosciences, USA) using FlowJo 7.6.5. Antibody details are in the Supplemental Digital Content (https://links.lww.com/ALN/E324) and table 1. Statistical analyses were performed using Prism 10.1.0 software (GraphPad, USA). Data normality was assessed using the Shapiro–Wilk test. Depending on data distribution, appropriate parametric or nonparametric tests were applied. For comparisons between two nonparametric groups (e.g., SWISS general anesthesia–exposed vs. control animals), the P values were calculated using the Mann–Whitney test. Aggregated data sets were analyzed using two-way analysis of variance, with the experimental protocol as the between-subject factor and a repeated measure (e.g., postnatal day for weight curves, or retinal layer identity for thickness measurements) as the within-subject factor. For comparisons involving four groups (tPAWT control, tPAWT general anesthesia, tPAnull control, and tPAnull general anesthesia), analyses were performed using a two-way analysis of variance with “experimental protocol” and “genotype” as independent variables. When appropriate, Tukey’s or Sidak’s post hoc test were applied. Outliers were identified using the ROUT method (Q = 1%). The results are presented as mean ± SD or median [interquartile range, 25th to 75th percentile], as specified in the figure legends. The differences were considered statistically significant for P < 0.05 (two-sided).

Tukey’s or Sidak’s post hoc test were applied. Outliers were identified using the ROUT method (Q = 1%). The results are presented as mean ± SD or median [interquartile range, 25th to 75th percentile], as specified in the figure legends. The differences were considered statistically significant for P < 0.05 (two-sided). The protocols were approved by the institutional ethics committee CENOMEXA (Comité Normand d’Éthique En Matière d’Expérimentation Animale, Normandy University) following European directive No. 2013/63/UE (agreement No. D14118001), with the French and regional guidelines for animal experimentation and the use of genetically modified organisms (French Ministry of Research, project licenses Nos. 8972 and 46951). All data supporting the findings of this study are available from the corresponding author upon reasonable request.

Experiments were conducted on postnatal day 15 (P15) and 6- to 8-week-old male SWISS mice (Janvier Labs, France) and 6- to 8-week-old male tPAnull and tPAWT mice (C57BL/6J; Normandy University, Caen, France). All protocols were approved by CENOMEXA (agreement No. D14118001) in compliance with European Union directive 2013/63/UE and French regulations (French Ministry of Research, license Nos. 8972 and 46951). Only healthy pups showing normal development before anesthesia exposure were included in the study. Clinical signs were monitored twice daily by animal facility staff and/or experimenters. Animals were excluded if they exhibited visible signs of pain or prolonged prostration, a high grimace scale score (score higher than 7 for more than 48 h), absence of response to external stimuli (e.g., paw pinch), cessation of food or water intake, or body weight loss exceeding 15% of initial weight. The animals were randomized at birth, with control and general anesthesia–exposed pups mixed within each litter. After weaning (postnatal day 21), the animals were separated by condition to avoid long-term confusion. Behavioral experiments and analyses were performed blind to condition whenever possible. Blinding was not feasible for magnetic resonance imaging (MRI) and functional ultrasound (fUS) experiments due to practical constraints requiring animal identification after imaging. Two experimental designs were used. For the SWISS mice, the animals were assigned to either a control group or a group exposed to early general anesthesia. In C57BL/6J mice, four groups were included (tPAWT control, tPAWT general anesthesia, tPAnull control, and tPAnull general anesthesia). Each individual animal was considered as an independent experimental unit (see Supplemental Digital Content for details, https://links.lww.com/ALN/E324).

To model repeated early general anesthesia exposure, SWISS, tPAnull, and tPAWT mice were randomly assigned to once daily exposure from P4 to P10, during 7 consecutive days, replicating the paradigm described in a recently published study.3 The general anesthesia group received 1.3% isoflurane (1 minimum alveolar concentration) in 50% O2 for 90 min; controls received 50% O2 alone: control pups were separated from the dams for the same duration as anesthetized pups and exposed to 50% O2 under identical environmental conditions, without anesthesia. Body temperature was monitored throughout general anesthesia exposure. Anesthetic depth was assessed by the loss of righting reflex and paw withdrawal reflex, as well as by monitoring heart and respiratory rates using a dedicated small animal monitoring system (SA Instruments Inc., USA; see Supplemental Digital Content for details, https://links.lww.com/ALN/E324).

To assess repeated early developmental outcomes after general anesthesia, body weight was monitored daily from P4 to P10 (preexposure) and then weekly. Eyelid opening was scored daily from P12 to P18, with ocular development tracked until behavioral and imaging experiments (see Supplemental Digital Content for details, https://links.lww.com/ALN/E324).

At P15 and 6 weeks, mice underwent four noninvasive behavioral tests to evaluate their visual abilities, including oculomotor reflex, depth perception, and visual acuity (see Supplemental Digital Content for more details, https://links.lww.com/ALN/E324). Oculomotor reflexes were assessed by the eye blinking reflex (EBR) test. A fine tip (0.7 mm) was brought close to the center of the cornea while the mouse head movements were limited by light hand restraint. Precautions were taken to avoid contact with the whiskers to prevent sensory stimulation. A score ranging from 0 to 9 for each eye was assigned. This scoring system was adapted from a previously published scale.29 Depth perception was assessed using the vertical placement test (VPT; scored 0 to 2) and the visual cliff test (VCT), where mice explored a two-zone arena (“safe” vs. “deep”; 0 vs. 30 cm floor drop). Anxiety-like behavior was assessed using the VCT open field and was recorded for 5 min and analyzed with Noldus EthoVision (distance, velocity, and zone preference; supplemental data 1A, https://links.lww.com/ALN/E316). Visual acuity was assessed via the optokinetic reflex test, using head-tracking responses to rotate sine wave gratings (12°/s) in a virtual cylinder. Spatial frequency thresholds were determined by increasing grating frequency until tracking stopped. The test was conducted separately for each eye by reversing the cylinder’s rotation direction. This test was only performed on tPAnull and tPAWT mice because of their healthy visual acuity, unlike SWISS mice, which are known to have an albino phenotype.

Oculomotor reflexes were assessed by the eye blinking reflex (EBR) test. A fine tip (0.7 mm) was brought close to the center of the cornea while the mouse head movements were limited by light hand restraint. Precautions were taken to avoid contact with the whiskers to prevent sensory stimulation. A score ranging from 0 to 9 for each eye was assigned. This scoring system was adapted from a previously published scale.29

Depth perception was assessed using the vertical placement test (VPT; scored 0 to 2) and the visual cliff test (VCT), where mice explored a two-zone arena (“safe” vs. “deep”; 0 vs. 30 cm floor drop). Anxiety-like behavior was assessed using the VCT open field and was recorded for 5 min and analyzed with Noldus EthoVision (distance, velocity, and zone preference; supplemental data 1A, https://links.lww.com/ALN/E316).

Visual acuity was assessed via the optokinetic reflex test, using head-tracking responses to rotate sine wave gratings (12°/s) in a virtual cylinder. Spatial frequency thresholds were determined by increasing grating frequency until tracking stopped. The test was conducted separately for each eye by reversing the cylinder’s rotation direction. This test was only performed on tPAnull and tPAWT mice because of their healthy visual acuity, unlike SWISS mice, which are known to have an albino phenotype.

For the anatomic assessment of the brain and eyes, T2-weighted images were acquired using a multislice multiecho sequence. Areas and thickness of multiple brain and eye parts were manually segmented on T2 acquisitions using ImageJ software (National Institutes of Health, USA).

For the functional assessment of specific brain areas, arterial spin labeling sequences were acquired, translating tissue perfusion through the measurement of cerebral blood volume in the somatosensory cortex, the hippocampus, the amygdala, the visual primary cortex, and the periaqueductal gray matter.

Molecular MRI assessed cerebral and retinal vascular inflammation through intravenous injection of iron oxide microparticles (MPIOs) linked to antibodies against endothelial adhesion molecules (P-selectin and vascular cell adhesion molecule-1 [VCAM-1]). The animals were first injected with 100 μl MPIOs–P-selectin (1 mg/kg iron) and second with 100 μl MPIOs–VCAM-1 (1 mg/kg iron) through the catheter placed in the caudal vein. The injection of MPIOs coupled to IgG served as a negative control by revealing the nonspecific fixation of the particles to the endothelium (Supplemental Digital Content, https://links.lww.com/ALN/E324; table 1). Labeled MPIOs were used as a contrast agent, were obtained in the laboratory, and have been previously described.30,31 Antibodies Used in the Project Biotechne (USA); BD Biosciences (USA); Abcam (USA); Jackson Immunoresearch (USA); R&D Systems (USA): BD Pharmingen (USA); Miltenyi Biotec (Germany); BioLegend (USA); Invitrogen (USA).

Anesthetized mice (isoflurane 1 to 2% and 30% O2/70% N2O) were placed in a stereotaxic frame and thermoregulated at 38.0°C, and an ophthalmic gel was applied to the eyes. Analgesia was achieved by subcutaneous injection of buprenorphin (0.1 mg/kg). The skull was exposed for acquisition. Whiskers were cut to a length of 1 cm. Medetomidine (Domitor, Vetoquinol, France), 0.08 mg/kg induction flowed by 0.16 mg · kg−1 · h−1 subcutaneous infusion) was used for maintenance. Cerebral blood volume and cerebral blood flow (CBF) were measured by transcranial Power Doppler ultrasound using an ultrafast scanner (Iconeus System, France). Neurovascular coupling (NVC) was assessed via CBF changes in the primary visual cortex (PVC), lateral geniculate nucleus (LGN), and superior colliculus (SC) after visual stimulation and in the ventral posterior–medial nucleus (VPM), posterior complex (PC) of the thalamus, and barrel cortex after whisker stimulation (see Supplemental Digital Content for details, https://links.lww.com/ALN/E324). Visual stimulation was applied using an LED light positioned 20 cm from the mouse eye, with a 60-s baseline followed by four 30-s stimuli (200-ms pulses at 4 Hz, repeated every 90 s) for a total of 420 s. Cerebrovascular hemodynamic responses were assessed by averaging cerebral blood volume variations across four trials within an imaging session. Both eyes were stimulated.

ioned 20 cm from the mouse eye, with a 60-s baseline followed by four 30-s stimuli (200-ms pulses at 4 Hz, repeated every 90 s) for a total of 420 s. Cerebrovascular hemodynamic responses were assessed by averaging cerebral blood volume variations across four trials within an imaging session. Both eyes were stimulated. Sensory responses in the barrel cortex were assessed through whisker stimulations. The stimulation pattern and the cerebrovascular hemodynamic responses are the same as previously described under “Visual Stimulation” (with mechanical whisker stimulation, 4 Hz).

Visual stimulation was applied using an LED light positioned 20 cm from the mouse eye, with a 60-s baseline followed by four 30-s stimuli (200-ms pulses at 4 Hz, repeated every 90 s) for a total of 420 s. Cerebrovascular hemodynamic responses were assessed by averaging cerebral blood volume variations across four trials within an imaging session. Both eyes were stimulated.

Sensory responses in the barrel cortex were assessed through whisker stimulations. The stimulation pattern and the cerebrovascular hemodynamic responses are the same as previously described under “Visual Stimulation” (with mechanical whisker stimulation, 4 Hz).

Blood and tissue samples were collected at P15 and 6 to 8 weeks. After induction (isoflurane 5%, 30% O2/70% N2O), anesthesia was reinforced with local anesthesia (lidocaine 5%). Analgesia was achieved by subcutaneous injection of buprenorphin (0.1 mg/kg). After deep anesthesia, ventricular blood was drawn, plasma was isolated, and samples were stored at −80°C. After perfusion, brains, eyes, and spleens were harvested (see Supplemental Digital Content for more details, https://links.lww.com/ALN/E324). For detection of total tPA in plasma, tPA levels were measured in P15 and 6- to 8-week-old mice using a total tPA enzyme-linked immunosorbent assay kit according to the manufacturer’s instructions (MI-MTPAKT-TOT; Gentaur, Inc., Belgium; see Supplemental Digital Content for more details, https://links.lww.com/ALN/E324). The retinas were flattened and cut into quarters. Brain sections (10 μm) were mounted on polylysine-coated slides and stored at −80°C before processing. Sections were incubated overnight at room temperature with primary antibodies diluted in a blocking solution (1× phosphate-buffered saline + 0.025% Triton X-100). Antibodies for immunohistochemistry are detailed in the Supplemental Digital Content (https://links.lww.com/ALN/E324) and table 1.

t −80°C before processing. Sections were incubated overnight at room temperature with primary antibodies diluted in a blocking solution (1× phosphate-buffered saline + 0.025% Triton X-100). Antibodies for immunohistochemistry are detailed in the Supplemental Digital Content (https://links.lww.com/ALN/E324) and table 1. Sagittal 10-μm cryosections were hematoxylin and eosin–stained (Biognost kit) and mounted on polylysine-coated slides. Retinal thickness was measured in central, proximal, and distal regions across all layers: ganglion cell layer, inner plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer, outer nuclear layer (ONL), photoreceptors, retinal pigmented epithelium, and choroid; using ImageJ (National Institutes of Health; see Supplemental Digital Content for details, https://links.lww.com/ALN/E324). To assess peripheral inflammation, the spleens were collected for flow cytometry. After Fc blocking (anti-CD16/CD32), the cells were stained with fluorochrome-conjugated antibodies and 7-aminoactinomycin D, including intracellular markers (kit from Miltenyi Biotec, Germany), and analyzed on a FACSVerse (BD Biosciences, USA) using FlowJo 7.6.5. Antibody details are in the Supplemental Digital Content (https://links.lww.com/ALN/E324) and table 1.

For detection of total tPA in plasma, tPA levels were measured in P15 and 6- to 8-week-old mice using a total tPA enzyme-linked immunosorbent assay kit according to the manufacturer’s instructions (MI-MTPAKT-TOT; Gentaur, Inc., Belgium; see Supplemental Digital Content for more details, https://links.lww.com/ALN/E324).

The retinas were flattened and cut into quarters. Brain sections (10 μm) were mounted on polylysine-coated slides and stored at −80°C before processing. Sections were incubated overnight at room temperature with primary antibodies diluted in a blocking solution (1× phosphate-buffered saline + 0.025% Triton X-100). Antibodies for immunohistochemistry are detailed in the Supplemental Digital Content (https://links.lww.com/ALN/E324) and table 1.

Sagittal 10-μm cryosections were hematoxylin and eosin–stained (Biognost kit) and mounted on polylysine-coated slides. Retinal thickness was measured in central, proximal, and distal regions across all layers: ganglion cell layer, inner plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer, outer nuclear layer (ONL), photoreceptors, retinal pigmented epithelium, and choroid; using ImageJ (National Institutes of Health; see Supplemental Digital Content for details, https://links.lww.com/ALN/E324).

To assess peripheral inflammation, the spleens were collected for flow cytometry. After Fc blocking (anti-CD16/CD32), the cells were stained with fluorochrome-conjugated antibodies and 7-aminoactinomycin D, including intracellular markers (kit from Miltenyi Biotec, Germany), and analyzed on a FACSVerse (BD Biosciences, USA) using FlowJo 7.6.5. Antibody details are in the Supplemental Digital Content (https://links.lww.com/ALN/E324) and table 1.

Statistical analyses were performed using Prism 10.1.0 software (GraphPad, USA). Data normality was assessed using the Shapiro–Wilk test. Depending on data distribution, appropriate parametric or nonparametric tests were applied. For comparisons between two nonparametric groups (e.g., SWISS general anesthesia–exposed vs. control animals), the P values were calculated using the Mann–Whitney test. Aggregated data sets were analyzed using two-way analysis of variance, with the experimental protocol as the between-subject factor and a repeated measure (e.g., postnatal day for weight curves, or retinal layer identity for thickness measurements) as the within-subject factor. For comparisons involving four groups (tPAWT control, tPAWT general anesthesia, tPAnull control, and tPAnull general anesthesia), analyses were performed using a two-way analysis of variance with “experimental protocol” and “genotype” as independent variables. When appropriate, Tukey’s or Sidak’s post hoc test were applied. Outliers were identified using the ROUT method (Q = 1%). The results are presented as mean ± SD or median [interquartile range, 25th to 75th percentile], as specified in the figure legends. The differences were considered statistically significant for P < 0.05 (two-sided).

The protocols were approved by the institutional ethics committee CENOMEXA (Comité Normand d’Éthique En Matière d’Expérimentation Animale, Normandy University) following European directive No. 2013/63/UE (agreement No. D14118001), with the French and regional guidelines for animal experimentation and the use of genetically modified organisms (French Ministry of Research, project licenses Nos. 8972 and 46951).

We first assessed the impact of early repeated general anesthesia exposure on SWISS mice weight gain and eyelid opening as landmarks of overall growth (fig. 1A). Early repeated general anesthesia exposure in SWISS mice did not affect weight gain (fig. 1B) but significantly delayed eyelid opening (median day: control P13 [95% CI, 0.52 to 1.45], general anesthesia P15 [95% CI, 0.62 to 1.92]; P < 0.0001; fig. 1C). At P15, general anesthesia–exposed mice showed impaired oculomotor reflexes and depth perception, scoring lower in EBR (control: 10 ± 3.35, general anesthesia: 3.36 ± 3.00; P < 0.0001) and VPT (control: 1.46 ± 0.52, general anesthesia: 0.75 ± 0.62; P = 0.0119; fig. 1D). At adulthood, deficits persisted in EBR (control: 14.11 ± 1.75, general anesthesia: 4.45 ± 4.23; P < 0.0001) and VPT (control: 1.71 ± 0.47, general anesthesia: 0.73 ± 0.70; P = 0.0006), while VCT performance and anxiety-like behaviors were unaffected (fig. 1E; supplemental data 1, A to E, https://links.lww.com/ALN/E316). From P15 to adulthood, EBR scores increased by +83% in general anesthesia mice versus +40% in controls; VPT scores remained stable in controls and improved slightly in general anesthesia mice (+25%), indicating modest sensorimotor recovery. These results highlight delayed eyelid opening and persistent visual processing deficits after early repeated general anesthesia.

% in general anesthesia mice versus +40% in controls; VPT scores remained stable in controls and improved slightly in general anesthesia mice (+25%), indicating modest sensorimotor recovery. These results highlight delayed eyelid opening and persistent visual processing deficits after early repeated general anesthesia. General anesthesia delays eyelid opening and alters visual-related behavior. (A) Schematic representation of the experimental timeline for exposure protocol to anesthetic agent, developmental monitoring and assessment of visual abilities during development. Mice were exposed from postnatal day (P)4 to P10 to 90 min of isoflurane anesthesia (1.3% + 50% O2) or to 50% O2 alone (controls). Eyelid opening was monitored daily from P12 to P18. Behavioral testing included P15 and adult assessments using the eye blinking reflex (EBR), vertical placement test (VPT), and visual cliff test (VCT). (B) Trajectory of body weight from baseline to P35 was comparable between groups (control, n = 38; general anesthesia, n = 37; two-way ANOVA, Tukey’s post hoc test, mean ± SD). (C) General anesthesia–exposed mice showed a delayed eyelid opening compared to controls (control, n = 31; general anesthesia, n = 28; Gehan–Breslow–Wilcoxon test, 95% CIs). (D) At P15, general anesthesia–exposed mice showed lower EBR and VPT scores (EBR: control, n = 24; general anesthesia, n = 22; VPT: control, n = 14; general anesthesia, n = 15; Mann–Whitney test, mean ± SD). (E) At 6 weeks, general anesthesia–exposed mice still showed lower EBR and VPT scores. No differences are found during the VCT (EBR: control, n = 18; general anesthesia, n = 20; Mann–Whitney test, mean ± SD. VPT: control, n = 14; general anesthesia, n = 15; Mann–Whitney test, mean ± SD. VCT: control, n = 14; general anesthesia, n = 14; Wilcoxon test, mean ± SD). Scoring details for EBR and VPT are provided in the Supplemental Digital Content (https://links.lww.com/ALN/E324).

nesthesia, n = 20; Mann–Whitney test, mean ± SD. VPT: control, n = 14; general anesthesia, n = 15; Mann–Whitney test, mean ± SD. VCT: control, n = 14; general anesthesia, n = 14; Wilcoxon test, mean ± SD). Scoring details for EBR and VPT are provided in the Supplemental Digital Content (https://links.lww.com/ALN/E324). To evaluate long-term structural effects of repeated early general anesthesia, eye and brain analyses were performed 5 weeks after exposure (fig. 2A). MRI revealed significant retinal thinning and reduced area in general anesthesia–exposed mice (thickness: control, 0.26 ± 0.04 mm; general anesthesia, 0.17 ± 0.03 mm; P < 0.0001; and area: control, 0.84 ± 0.14 mm²; general anesthesia, 0.60 ± 0.12 mm²; P < 0.0001; fig. 2B). At P15, no differences in retinal layer thickness were found (fig. 2, C and D). In adulthood, structural alterations appeared: central IPL was thinner and ONL was thicker in general anesthesia mice (IPL: control, 77.89 ± 2.31 µm; general anesthesia, 46.77 ± 13.11 µm; and ONL: control, 24.33 ± 2.04 µm; general anesthesia, 55.09 ± 3.68 µm; P < 0.0001). In the proximal retina, INL and total thickness were reduced (INL: control, 50.93 ± 0.76 µm; general anesthesia, 31.89 ± 5.47 µm; P = 0.0330; and total: control, 254.80 ± 17.92 µm; general anesthesia, 221.80 ± 14.26 µm; P < 0.0001; fig. 2D). No changes were found in anterior chamber, vitreous, or lens (supplemental data 2, A and B, https://links.lww.com/ALN/E317). To determine whether nuclear layer thinning was linked to vascular defects, retinal vasculature was analyzed at 6 weeks (supplemental data 3, A and D, https://links.lww.com/ALN/E318). Immunohistochemistry revealed no significant differences between general anesthesia and control groups.

https://links.lww.com/ALN/E317). To determine whether nuclear layer thinning was linked to vascular defects, retinal vasculature was analyzed at 6 weeks (supplemental data 3, A and D, https://links.lww.com/ALN/E318). Immunohistochemistry revealed no significant differences between general anesthesia and control groups. Early general anesthesia exposure alters retinal structure. (A) Schematic representation of the experimental timeline for exposure protocol to anesthetic agent, assessment of anatomic eye development. Retinal structure was analyzed at postnatal day (P) 15 using histologic analysis (hematoxylin and eosin [H&E] staining) and at 6 weeks using T2-weighted magnetic resonance imaging (MRI) and histologic analysis. (B, left) Representative T2-weighted MRI images of the eye acquired at 6 weeks. (Right) General anesthesia–exposed mice showed reduced retinal thickness and retinal area (control, n = 15; general anesthesia, n = 14; Mann–Whitney test, mean ± SD). (C) Representative micrographs of H&E-stained retinal sections from a control animal (top) and a general anesthesia animal (bottom) at 6 weeks. (D) Retinal layer thickness was quantified at P15 and 6 weeks. No major differences were observed at P15. At 6 weeks, general anesthesia–exposed mice showed reduced thickness of the inner plexiform layer (IPL) and increased thickness of the outer nuclear layer (ONL) in the central retina. In the proximal retina, the inner nuclear layer (INL) appeared thinner, and the total retinal thickness was reduced in general anesthesia–exposed mice (P15 and 6 weeks: control, n = 3; general anesthesia, n = 4; two-way analysis of variance; Sidak’s post hoc test, mean ± SD). C, choroid; GCL, ganglion cell layer; INL, inner nuclear layer; MRI, magnetic resonance imaging; OPL, outer plexiform layer; P, photoreceptors; PE, retinal pigmented epithelium.

sia–exposed mice (P15 and 6 weeks: control, n = 3; general anesthesia, n = 4; two-way analysis of variance; Sidak’s post hoc test, mean ± SD). C, choroid; GCL, ganglion cell layer; INL, inner nuclear layer; MRI, magnetic resonance imaging; OPL, outer plexiform layer; P, photoreceptors; PE, retinal pigmented epithelium. To assess general anesthesia effects on visual brain regions, PVC and LGN structures were measured (fig. 3A). General anesthesia–exposed mice showed reduced PVC thickness and area (control: 0.80 ± 0.08 mm/0.67 ± 0.05 mm2; general anesthesia: 0.73 ± 0.05 mm/0.63 ± 0.05 mm2; P = 0.0235/0.0487; fig. 3B), while LGN was unaffected, indicating selective cortical vulnerability.

regions, PVC and LGN structures were measured (fig. 3A). General anesthesia–exposed mice showed reduced PVC thickness and area (control: 0.80 ± 0.08 mm/0.67 ± 0.05 mm2; general anesthesia: 0.73 ± 0.05 mm/0.63 ± 0.05 mm2; P = 0.0235/0.0487; fig. 3B), while LGN was unaffected, indicating selective cortical vulnerability. Early general anesthesia exposure alters primary visual cortex (PVC) structure and increases vascular area without immune changes. (A) Schematic representation of the experimental timeline for exposure protocol to anesthetic agent, assessment of anatomic visual brain areas, and vascular and inflammatory changes. At 6 weeks, anatomic and molecular imaging (magnetic resonance imaging [MRI]) accompanied with histologic analysis were performed (immunohistochemistry [IHC]). (B) Manual delineation of visual brain regions on T2-weighted anatomic MRI revealed reduced thickness and surface area of the PVC in general anesthesia–exposed mice, while the lateral geniculate nucleus (LGN) remained unaffected (PVC thickness: control, n = 14; general anesthesia, n = 11; PVC area: control, n = 14; general anesthesia, n = 11; LGN area: control, n = 14; general anesthesia, n = 11; Mann–Whitney test, mean ± SD). (C) Representative images of brain sections immunohistochemistry labeled with vascular cell adhesion molecule 1 (VCAM-1), podocalyxin (PDLX), Iba1, and 4′,6′-diamino-2-phenylindole (DAPI) in the dentate gyrus, cortex, PVC, and LGN. (D to G) Quantification of vascular (PDLX) and inflammatory marker (VCAM-1) coverage (%) showed a significant increase in vascular area (PDLX labeling) within the PVC after early general anesthesia exposure (F, P = 0.0480), while no significant differences were observed for other brain regions or for VCAM-1 coverage (control, n = 3; general anesthesia, n = 4; Mann–Whitney test, mean ± SD). (H) Representative images of high-resolution T2*-weighted molecular MRI performed under baseline (before injection) and after injection of different contrast-agents (iron oxide microparticles [MPIOs] coupled to immunoglobulin G [IgG], P-selectin, or VCAM-1).

3; general anesthesia, n = 4; Mann–Whitney test, mean ± SD). (H) Representative images of high-resolution T2*-weighted molecular MRI performed under baseline (before injection) and after injection of different contrast-agents (iron oxide microparticles [MPIOs] coupled to immunoglobulin G [IgG], P-selectin, or VCAM-1). (I) Quantification of induced hyposignal (signal void) revealed higher VCAM-1-associated signal compared to IgG and P-selectin conditions in mice exposed to general anesthesia, but no significant differences were found between control and general anesthesia–exposed groups (IgG: control, n = 3; general anesthesia, n = 4; P-selectin: control, n = 10; general anesthesia, n = 10; VCAM-1: control, n = 11; general anesthesia, n = 10; two-way ANOVA; Tukey’s post hoc test, mean ± SD).

ral anesthesia, but no significant differences were found between control and general anesthesia–exposed groups (IgG: control, n = 3; general anesthesia, n = 4; P-selectin: control, n = 10; general anesthesia, n = 10; VCAM-1: control, n = 11; general anesthesia, n = 10; two-way ANOVA; Tukey’s post hoc test, mean ± SD). Podocalyxin and VCAM-1 staining showed preserved vascular and inflammatory profiles in LGN, cortex, and dentate gyrus (fig. 3, C to G). In the PVC, general anesthesia–exposed mice showed a slight but significant increase in podocalyxin-postive vessel area (control, 2.68 ± 0.65%; general anesthesia, 3.49 ± 0.39%; P = 0.0480; fig. 3F), coinciding with reduced cortical thickness (fig. 3B). This local vascular change was not seen in other regions and was not linked to increased VCAM-1 expression. Retinal vasculature remained unaffected (supplemental data 3, A to D, https://links.lww.com/ALN/E318). Microglial and nuclear morphology (area, density, and perimeter/diameter) were unchanged (supplemental data 4, A to D, https://links.lww.com/ALN/E319), as were in vivo vascular inflammation markers (IgG, P-selectin, and VCAM-1; fig. 3, H and I) and systemic immune cell profiles by flow cytometry (supplemental data 5, A to N, https://links.lww.com/ALN/E320).

, density, and perimeter/diameter) were unchanged (supplemental data 4, A to D, https://links.lww.com/ALN/E319), as were in vivo vascular inflammation markers (IgG, P-selectin, and VCAM-1; fig. 3, H and I) and systemic immune cell profiles by flow cytometry (supplemental data 5, A to N, https://links.lww.com/ALN/E320). Since general anesthesia exposure altered vascular area, we next assessed CBF. At rest, no basal CBF changes were detected by arterial spin labeling MRI or fUS in visual or nonvisual regions (supplemental data 6, A to C, https://links.lww.com/ALN/E321). NVC, measured by fUS during visual and somatosensory stimulation (fig. 4A), was preserved in PVC, SC, and LGN (fig. 4, B and C). In contrast, CBF responses were reduced in the thalamic PC (control, 20.33% [95% CI, 17.92 to 26.24%]; general anesthesia, 14.75% [95% CI, 9.20 to 18.17%]; P = 0.0303) and barrel cortex (control, 24.44% [95% CI, 20.11 to 27.20%]; general anesthesia, 19.18% [95% CI, 9.80 to 19.92%]; P = 0.0101) but not in the VPM (fig. 4, D and E).

CBF responses were reduced in the thalamic PC (control, 20.33% [95% CI, 17.92 to 26.24%]; general anesthesia, 14.75% [95% CI, 9.20 to 18.17%]; P = 0.0303) and barrel cortex (control, 24.44% [95% CI, 20.11 to 27.20%]; general anesthesia, 19.18% [95% CI, 9.80 to 19.92%]; P = 0.0101) but not in the VPM (fig. 4, D and E). Early general anesthesia exposure alters neurovascular coupling (NVC) in the thalamus but not along the visual pathway. (A) Schematic representation of the experimental timeline for exposure protocol to anesthetic agent and analysis of brain cerebral blood flow (CBF) in response to stimuli. At 6 weeks, functional ultrasound (fUS) imaging was performed during visual and whisker stimulations to assess CBF changes. Left-sided visual stimuli activated the right visual cortex due to optic chiasm crossing, while left whisker stimulation elicited ipsilateral (left hemisphere) responses; measurements were accordingly performed in the corresponding cortical regions. (B) Representative CBF activation shows responses in the primary visual cortex (PVC), superior colliculus (SC), and lateral geniculate nucleus (LGN) after visual stimulation in control and general anesthesia–exposed mice. The color scale ranges from cool to warm colors, indicating increasing levels of CBF changes during visual stimulation, with warm colors reflecting greater CBF elevations. (C) Time course and box plots show the variation in the percentage of CBF increase from baseline during visual stimulations in brain areas involved in control and general anesthesia–exposed groups. CBF responses in the PVC, SC, and LGN reveal no significant differences between groups (control, n = 5; general anesthesia, n = 4; Mann–Whitney test, 95% CIs). (D) Representative CBF maps after whisker stimulation in the thalamus and barrel cortex. The color scale ranges from cool to warm colors, indicating increasing levels of CBF changes during visual stimulation, with warm colors reflecting greater CBF elevations. (E) Time course and box plots show the variation in the percentage of CBF increase from baseline during whisker stimulations in brain areas involved in control and general anesthesia–exposed groups.

ndicating increasing levels of CBF changes during visual stimulation, with warm colors reflecting greater CBF elevations. (E) Time course and box plots show the variation in the percentage of CBF increase from baseline during whisker stimulations in brain areas involved in control and general anesthesia–exposed groups. CBF responses in the thalamic ventral posterior–medial nucleus (VPM) remain unaffected in general anesthesia–exposed mice, while responses in the thalamic posterior complex (PC) and in the barrel cortex are reduced in general anesthesia–exposed mice (control, n = 5; general anesthesia, n = 7 or n = 6 for VPM area); Mann–Whitney test, 95% CIs). Early repeated general anesthesia exposure alters structural and vascular development in the PVC without triggering immune activation. While basal CBF remains unchanged, NVC is impaired in somatosensory but not visual regions, suggesting pathway-specific vulnerability. In the PVC, reduced cortical thickness and increased vascularization may indicate disorganization, yet preserved NVC suggests a compensatory mechanism supporting functional integrity. In contrast, impaired NVC in somatosensory areas may limit experience-dependent plasticity critical for sensorimotor learning.32,33

ity. In the PVC, reduced cortical thickness and increased vascularization may indicate disorganization, yet preserved NVC suggests a compensatory mechanism supporting functional integrity. In contrast, impaired NVC in somatosensory areas may limit experience-dependent plasticity critical for sensorimotor learning.32,33 Given tPA’s role in neural and vascular development, we quantified plasma tPA in SWISS mice after general anesthesia (fig. 5A). At P15, levels were unchanged, but at 6 weeks, general anesthesia–exposed mice showed a significant reduction (control, 11.58 ± 2.19 ng/ml; general anesthesia, 7.65 ± 1.31 ng/ml; P = 0.0082). Controls showed a developmental increase in tPA from P15 to adulthood (supplemental data 7A, https://links.lww.com/ALN/E322), absent in general anesthesia mice, suggesting disrupted tPA regulation after early repeated exposure.

reduction (control, 11.58 ± 2.19 ng/ml; general anesthesia, 7.65 ± 1.31 ng/ml; P = 0.0082). Controls showed a developmental increase in tPA from P15 to adulthood (supplemental data 7A, https://links.lww.com/ALN/E322), absent in general anesthesia mice, suggesting disrupted tPA regulation after early repeated exposure. Early general anesthesia exposure and tissue-type plasminogen activator (tPA) deficiency similarly impair growth, visual responses, and cortical structure. (A) Plasma tPA concentration measured by enzyme-linked immunosorbent assay at postnatal day (P) 15 and in adult SWISS mice shows reduced circulating tPA in general anesthesia–exposed animals (P15: control and general anesthesia, n = 6; adult: control, n = 7; general anesthesia, n = 6; Mann–Whitney test, mean ± SD). (B) Schematic representation of the experimental timeline for exposure protocol to anesthetic agent, developmental monitoring, assessment of visual abilities, and anatomic analysis of brain visual areas. tPAWT and tPAnull mice were exposed to anesthesia (1.3% isoflurane + 50% O2) or 50% O2 (controls) from P4 to P10. The mice were monitored for weight, eyelid opening, and behavioral performance at P15 and adulthood, followed by anatomic brain magnetic resonance imaging (MRI) at 6 weeks. (C) Trajectory of body weight from P4 to P35. General anesthesia–exposed tPAWT mice showed significantly reduced body weight compared to tPAWT controls between P10 and P21. tPAnull control mice also displayed lower body weight compared to their tPAWT counterparts at multiple time points between P10 and P35. No significant differences in body weight were observed between tPAnull general anesthesia and tPAWT general anesthesia mice nor between tPAnull general anesthesia and tPAnull controls (tPAWT: control, n = 15; general anesthesia, n = 13; tPAnull: control, n = 15; general anesthesia, n = 18; two-way ANOVA, Tukey’s post hoc test, mean ± SD). (D) Eyelid opening incidence showed a trend toward delay in general anesthesia–exposed tPAWT mice, whereas a significant delay was observed in control tPAnull mice compared to their respective counterparts (tPAWT: control, n = 22; general anesthesia, n = 16; tPAnull: control, n = 11; general anesthesia, n = 12; Gehan–Breslow–Wilcoxon test, 95% CIs). (E) Optokinetic response, eye-blinking reflex (EBR), and vertical placement test (VPT) were assessed at P15 and 6 weeks in tPAWT and tPAnull mice.

eir respective counterparts (tPAWT: control, n = 22; general anesthesia, n = 16; tPAnull: control, n = 11; general anesthesia, n = 12; Gehan–Breslow–Wilcoxon test, 95% CIs). (E) Optokinetic response, eye-blinking reflex (EBR), and vertical placement test (VPT) were assessed at P15 and 6 weeks in tPAWT and tPAnull mice. Reduced visual responses were observed in general anesthesia–exposed tPAWT mice at both P15 and 6 weeks, while control tPAnull mice showed comparable reductions relative to tPAWT controls, particularly at 6 weeks (optokinetic: P15 tPAWT: control, n = 11; general anesthesia, n = 11; tPAnull: control, n = 8; general anesthesia, n = 12; and adult tPAWT: control, n = 14; general anesthesia, n = 13; tPAnull control, n = 15; general anesthesia, n = 18. EBR: P15 tPAWT: control, n = 12; general anesthesia, n = 12; tPAnull control, n = 10; general anesthesia, n = 15; EBR: adult tPAWT: control, n = 13; general anesthesia, n = 11; tPAnull: control, n = 14; general anesthesia, n = 16. VPT: P15 tPAWT: control, n = 9; general anesthesia, n = 12; tPAnull: control, n = 15; general anesthesia, n = 17. Adult tPAWT: control, n = 9; general anesthesia, n = 13; tPAnull: control, n = 13; general anesthesia, n = 20; two-way ANOVA; Tukey’s post hoc test, mean ± SD). Scoring details for EBR and VPT are provided in the Supplemental Digital Content (https://links.lww.com/ALN/E324). (F) Retinal thickness and area were measured by T2-weighted anatomic MRI in tPAnull versus tPAWT mice, general anesthesia–exposed or controls, at 6 weeks. Both measurements were reduced in general anesthesia–exposed tPAWT compared to tPAWT control and in tPAnull control compared to tPAWT control (tPAWT: control, n = 5; general anesthesia, n = 7; tPAnull: control, n = 6; general anesthesia, n = 6; two-way ANOVA; Tukey’s post hoc test, mean ± SD). MRI acquisition and analysis followed the same procedures as for the SWISS cohort, illustrated in figures 2B and 3B. (G) T2-weighted anatomic MRI revealed reduced thickness and area of the primary visual cortex (PVC), as well as reduced lateral geniculate nucleus (LGN) thickness in tPAWT general anesthesia–exposed and tPAnull mice compared to their counterparts (PVC measurements: tPAWT: control, n = 5; general anesthesia, n = 5; tPAnull: control, n = 6; general anesthesia, n = 5. LGN measurements: tPAWT control, n = 5; general anesthesia, n = 5; tPAnull control, n = 5; general anesthesia, n = 5; two-way ANOVA, Tukey’s post hoc test, mean ± SD).

compared to their counterparts (PVC measurements: tPAWT: control, n = 5; general anesthesia, n = 5; tPAnull: control, n = 6; general anesthesia, n = 5. LGN measurements: tPAWT control, n = 5; general anesthesia, n = 5; tPAnull control, n = 5; general anesthesia, n = 5; two-way ANOVA, Tukey’s post hoc test, mean ± SD). MRI acquisition and analysis followed the same procedures as for the SWISS cohort, illustrated in figures 2B and 3B. To assess general anesthesia effects in the absence of tPA, tPAnull and tPAWT mice were exposed to general anesthesia or control conditions (fig. 5B). Control tPAnull mice had reduced weight versus tPAWT at all time points (P10: 5.67 ± 0.65 g vs. 4.58 ± 0.78 g, P = 0.0331; P15: 7.17 ± 1.13 g vs. 5.93 ± 0.25 g, P = 0.0168; P21: 10.01 ± 1.31 g vs. 8.12 ± 1.30 g, P = 0.0002; P35: 19.02 ± 2.46 g vs. 17.25 ± 2.31 g, P < 0.0001; fig. 5C), suggesting a role for tPA in growth. Surprisingly, general anesthesia–exposed tPAnull mice showed improved weight versus nonexposed tPAnull at later stages (P21 to P35: all P < 0.02), indicating that general anesthesia may partially compensate for tPA deficiency. No differences were seen between general anesthesia and control tPAWT mice or between general anesthesia–exposed tPAWT and tPAnull, suggesting that general anesthesia alters growth only in the absence of functional tPA.

P35: all P < 0.02), indicating that general anesthesia may partially compensate for tPA deficiency. No differences were seen between general anesthesia and control tPAWT mice or between general anesthesia–exposed tPAWT and tPAnull, suggesting that general anesthesia alters growth only in the absence of functional tPA. Eyelid opening was delayed in both general anesthesia–exposed tPAWT and control tPAnull mice versus control tPAWT (median P16 to P17; P = 0.0616 and P = 0.0516; fig. 5D). No additional delay was seen in general anesthesia–exposed tPAnull mice, suggesting that general anesthesia and tPA deficiency converge via a shared, nonadditive mechanism. In tPAWT mice, general anesthesia exposure impaired visual acuity, oculomotor responses, and depth perception at P15 and adulthood (fig. 5E). Visual acuity dropped at P15 (0.11 ± 0.04 vs. 0.05 ± 0.03 cycles/degree; P = 0.0003) and remained lower in adulthood (0.41 ± 0.09 vs. 0.35 ± 0.07; P = 0.0501). EBR scores declined (P15: 9.58 ± 3.50 to 5.17 ± 4.04; P = 0.0414; adult: 11.38 ± 5.82 to 5.00 ± 4.00; P = 0.0008), as did VPT scores (P15: 1.44 ± 0.73 to 0.58 ± 0.56; P = 0.0072; adult: 1.44 ± 0.60 to 0.62 ± 0.58; P = 0.0052). tPAnull mice showed similar impairments under control conditions, and general anesthesia exposure did not further exacerbate these deficits, suggesting nonadditive effects of tPA deficiency and early repeated general anesthesia on visual function.

0.58 ± 0.56; P = 0.0072; adult: 1.44 ± 0.60 to 0.62 ± 0.58; P = 0.0052). tPAnull mice showed similar impairments under control conditions, and general anesthesia exposure did not further exacerbate these deficits, suggesting nonadditive effects of tPA deficiency and early repeated general anesthesia on visual function. MRI revealed retinal thinning and area reduction in general anesthesia–exposed tPAWT mice (thickness: 0.27 ± 0.03 vs. 0.20 ± 0.05 mm; P = 0.0274; area: 0.89 ± 0.20 vs. 0.65 ± 0.25 mm2; P = 0.0492; fig. 5F). Similar defects were observed in control tPAnull mice (thickness: P = 0.0235; area: P = 0.0031), with no additive effect of general anesthesia, suggesting that tPA deficiency accounts for the retinal phenotype. Other ocular compartments were unaffected (supplemental data 7B, https://links.lww.com/ALN/E322).

; fig. 5F). Similar defects were observed in control tPAnull mice (thickness: P = 0.0235; area: P = 0.0031), with no additive effect of general anesthesia, suggesting that tPA deficiency accounts for the retinal phenotype. Other ocular compartments were unaffected (supplemental data 7B, https://links.lww.com/ALN/E322). In tPAWT mice, repeated early general anesthesia reduced PVC thickness (0.71 ± 0.02 vs. 0.67 ± 0.03 mm; P = 0.0117) and showed a trend toward reduced area (P = 0.0897; fig. 5G). Similar reductions were found in control tPAnull mice (thickness: 0.65 ± 0.02 mm; P = 0.0003; area: 0.55 ± 0.04 mm2; P = 0.0068), with no additive general anesthesia effect. In the LGN, both general anesthesia in tPAWT and tPA deficiency under control reduced area (tPAWT control, 0.32 ± 0.02 mm2 vs. general anesthesia, 0.30 ± 0.02 mm2; P = 0.0440; or tPAnull control, 0.29 ± 0.01 mm2; P = 0.0003), again with no further reduction in general anesthesia–exposed tPAnull mice. These results support a shared, saturable mechanism by which general anesthesia and tPA loss affect visual brain structures. Molecular MRI showed no P-selectin or VCAM-1 upregulation in tPAWT or tPAnull mice, under general anesthesia or control conditions, indicating an absence of CNS vascular inflammation (supplemental data 8, A to C, https://links.lww.com/ALN/E323).

To evaluate long-term structural effects of repeated early general anesthesia, eye and brain analyses were performed 5 weeks after exposure (fig. 2A). MRI revealed significant retinal thinning and reduced area in general anesthesia–exposed mice (thickness: control, 0.26 ± 0.04 mm; general anesthesia, 0.17 ± 0.03 mm; P < 0.0001; and area: control, 0.84 ± 0.14 mm²; general anesthesia, 0.60 ± 0.12 mm²; P < 0.0001; fig. 2B). At P15, no differences in retinal layer thickness were found (fig. 2, C and D). In adulthood, structural alterations appeared: central IPL was thinner and ONL was thicker in general anesthesia mice (IPL: control, 77.89 ± 2.31 µm; general anesthesia, 46.77 ± 13.11 µm; and ONL: control, 24.33 ± 2.04 µm; general anesthesia, 55.09 ± 3.68 µm; P < 0.0001). In the proximal retina, INL and total thickness were reduced (INL: control, 50.93 ± 0.76 µm; general anesthesia, 31.89 ± 5.47 µm; P = 0.0330; and total: control, 254.80 ± 17.92 µm; general anesthesia, 221.80 ± 14.26 µm; P < 0.0001; fig. 2D). No changes were found in anterior chamber, vitreous, or lens (supplemental data 2, A and B, https://links.lww.com/ALN/E317). To determine whether nuclear layer thinning was linked to vascular defects, retinal vasculature was analyzed at 6 weeks (supplemental data 3, A and D, https://links.lww.com/ALN/E318). Immunohistochemistry revealed no significant differences between general anesthesia and control groups.

To assess general anesthesia effects on visual brain regions, PVC and LGN structures were measured (fig. 3A). General anesthesia–exposed mice showed reduced PVC thickness and area (control: 0.80 ± 0.08 mm/0.67 ± 0.05 mm2; general anesthesia: 0.73 ± 0.05 mm/0.63 ± 0.05 mm2; P = 0.0235/0.0487; fig. 3B), while LGN was unaffected, indicating selective cortical vulnerability.

Given tPA’s role in neural and vascular development, we quantified plasma tPA in SWISS mice after general anesthesia (fig. 5A). At P15, levels were unchanged, but at 6 weeks, general anesthesia–exposed mice showed a significant reduction (control, 11.58 ± 2.19 ng/ml; general anesthesia, 7.65 ± 1.31 ng/ml; P = 0.0082). Controls showed a developmental increase in tPA from P15 to adulthood (supplemental data 7A, https://links.lww.com/ALN/E322), absent in general anesthesia mice, suggesting disrupted tPA regulation after early repeated exposure.

compared to their counterparts (PVC measurements: tPAWT: control, n = 5; general anesthesia, n = 5; tPAnull: control, n = 6; general anesthesia, n = 5. LGN measurements: tPAWT control, n = 5; general anesthesia, n = 5; tPAnull control, n = 5; general anesthesia, n = 5; two-way ANOVA, Tukey’s post hoc test, mean ± SD). MRI acquisition and analysis followed the same procedures as for the SWISS cohort, illustrated in figures 2B and 3B.

To assess general anesthesia effects in the absence of tPA, tPAnull and tPAWT mice were exposed to general anesthesia or control conditions (fig. 5B). Control tPAnull mice had reduced weight versus tPAWT at all time points (P10: 5.67 ± 0.65 g vs. 4.58 ± 0.78 g, P = 0.0331; P15: 7.17 ± 1.13 g vs. 5.93 ± 0.25 g, P = 0.0168; P21: 10.01 ± 1.31 g vs. 8.12 ± 1.30 g, P = 0.0002; P35: 19.02 ± 2.46 g vs. 17.25 ± 2.31 g, P < 0.0001; fig. 5C), suggesting a role for tPA in growth. Surprisingly, general anesthesia–exposed tPAnull mice showed improved weight versus nonexposed tPAnull at later stages (P21 to P35: all P < 0.02), indicating that general anesthesia may partially compensate for tPA deficiency. No differences were seen between general anesthesia and control tPAWT mice or between general anesthesia–exposed tPAWT and tPAnull, suggesting that general anesthesia alters growth only in the absence of functional tPA. Eyelid opening was delayed in both general anesthesia–exposed tPAWT and control tPAnull mice versus control tPAWT (median P16 to P17; P = 0.0616 and P = 0.0516; fig. 5D). No additional delay was seen in general anesthesia–exposed tPAnull mice, suggesting that general anesthesia and tPA deficiency converge via a shared, nonadditive mechanism.

In tPAWT mice, general anesthesia exposure impaired visual acuity, oculomotor responses, and depth perception at P15 and adulthood (fig. 5E). Visual acuity dropped at P15 (0.11 ± 0.04 vs. 0.05 ± 0.03 cycles/degree; P = 0.0003) and remained lower in adulthood (0.41 ± 0.09 vs. 0.35 ± 0.07; P = 0.0501). EBR scores declined (P15: 9.58 ± 3.50 to 5.17 ± 4.04; P = 0.0414; adult: 11.38 ± 5.82 to 5.00 ± 4.00; P = 0.0008), as did VPT scores (P15: 1.44 ± 0.73 to 0.58 ± 0.56; P = 0.0072; adult: 1.44 ± 0.60 to 0.62 ± 0.58; P = 0.0052). tPAnull mice showed similar impairments under control conditions, and general anesthesia exposure did not further exacerbate these deficits, suggesting nonadditive effects of tPA deficiency and early repeated general anesthesia on visual function. MRI revealed retinal thinning and area reduction in general anesthesia–exposed tPAWT mice (thickness: 0.27 ± 0.03 vs. 0.20 ± 0.05 mm; P = 0.0274; area: 0.89 ± 0.20 vs. 0.65 ± 0.25 mm2; P = 0.0492; fig. 5F). Similar defects were observed in control tPAnull mice (thickness: P = 0.0235; area: P = 0.0031), with no additive effect of general anesthesia, suggesting that tPA deficiency accounts for the retinal phenotype. Other ocular compartments were unaffected (supplemental data 7B, https://links.lww.com/ALN/E322).

The current study provides insights into how early repeated exposure to general anesthesia affects the visual system, revealing key findings implicating disrupted regulation of tPA. General anesthesia delays developmental milestones, impairs visual abilities like oculomotor reflexes and depth perception, and reduces retinal and PVC thickness, disrupting visual processing areas. Exposure to isoflurane-based general anesthesia alters LGN structure in tPAWT but not in SWISS mice, while both PVC and LGN are affected in tPAnull mice, indicating overlap between tPA deficiency and general anesthesia exposure. These results identify tPA as a key modulator of visual system development and vulnerability to general anesthesia–induced alterations and a potential biomarker in future studies.

SWISS mice, while both PVC and LGN are affected in tPAnull mice, indicating overlap between tPA deficiency and general anesthesia exposure. These results identify tPA as a key modulator of visual system development and vulnerability to general anesthesia–induced alterations and a potential biomarker in future studies. Our results show that early repeated general anesthesia exposure alters structural and functional maturation of the visual system. Delayed eyelid opening in general anesthesia–exposed male mice suggests impaired maturation, potentially contributing to later retinal and cortical alterations. Although delayed eyelid opening after early repeated general anesthesia exposure may transiently affect visuomotor development, the persistence of functional and structural deficits into adulthood, along with neurovascular alterations in nonvisual cortical regions, suggests that general anesthesia induces broader and lasting neurodevelopmental disruptions beyond transient sensory deprivation. These changes may be driven by neurotoxic mechanisms including apoptosis or oxidative stress, as previously reported in early general anesthesia models.17,34–43 In the central retina, early repeated general anesthesia exposure led to IPL thinning and ONL thickening, indicating disrupted synaptic organization, compensatory remodeling, or impaired cell differentiation. INL thinning, which contains neuronal (amacrine, bipolar, and horizontal) and glial (Müller) cell bodies, likely reflects reduced cell numbers or altered morphology.44 As the IPL is primarily formed by synapses from INL and ganglion cell neurons, INL thinning may secondarily contribute to IPL loss, potentially through amacrine cell degeneration and reduced synaptic connectivity. This vulnerability may be exacerbated by insufficient trophic support from local vascular deficits, given the INL’s position between intermediate and deep vascular plexuses,45 and by apoptosis or delayed maturation during sensitive developmental periods.46–48 Early repeated general anesthesia also induced cortical alterations, including reduced PVC thickness and area, likely reflecting neuronal loss or disrupted migration and lamination.

ion between intermediate and deep vascular plexuses,45 and by apoptosis or delayed maturation during sensitive developmental periods.46–48 Early repeated general anesthesia also induced cortical alterations, including reduced PVC thickness and area, likely reflecting neuronal loss or disrupted migration and lamination. These findings align with previous reports of neocortical degeneration after prolonged general anesthesia exposure in young mice49 and with a meta-analysis of greater than 440 preclinical studies linking early anesthesia to neurotoxic and behavioral effects.50 Such structural changes likely underlie the observed functional deficits, including impaired depth perception linked to PVC alterations and oculomotor reflex deficits possibly involving brainstem pathways.51,52 Although the SC appeared unaffected, contributions from other subcortical structures cannot be excluded. The translational relevance of these findings is supported by human studies showing that global visual processing deficits correlate with reduced gray matter volume in visuospatial cortical areas.53 Interestingly, cerebral perfusion in visual regions remained stable 5 weeks after exposure, suggesting structural rather than hemodynamic alterations. In contrast, neurovascular responses were reduced in somatosensory regions, including the barrel cortex and posterior thalamus, indicating selective circuit vulnerability possibly related to developmental timing, metabolic demand, or tPA-mediated mechanisms. Overall, early repeated general anesthesia exposure induces persistent structural alterations within the visual system and region-specific neurovascular dysfunctions, which may disrupt sensory-motor and cognitive networks. Consistently, early general anesthesia has been associated with object recognition deficits in primates and altered visuospatial attention in children,14,54 suggesting long-term consequences for adaptive behavior.

em and region-specific neurovascular dysfunctions, which may disrupt sensory-motor and cognitive networks. Consistently, early general anesthesia has been associated with object recognition deficits in primates and altered visuospatial attention in children,14,54 suggesting long-term consequences for adaptive behavior. Although structural and vascular changes were limited to the PVC, immune profiles appeared largely unaffected. The local increase in vascular area may reflect compensatory remodeling due to reduced cortical thickness, as seen after injury.55 No other vascular or inflammatory changes were detected, suggesting that these alterations may occur independently of global immune activation. However, since analyses were performed at a late time point, transient periexposure inflammation, such as microglial activation or cytokine release, cannot be excluded.47 Early disruptions like delayed retinal vascularization or apoptosis may also contribute, warranting future studies at earlier stages to clarify timing and nature of these events.

performed at a late time point, transient periexposure inflammation, such as microglial activation or cytokine release, cannot be excluded.47 Early disruptions like delayed retinal vascularization or apoptosis may also contribute, warranting future studies at earlier stages to clarify timing and nature of these events. To explore systemic effects of early repeated anesthesia, we measured circulating levels of tPA in SWISS mice. A significant reduction in plasma tPA 6 weeks after exposure suggests a long-lasting disruption, contrasting with previous reports of transient increases shortly after anesthesia.56 The delayed reduction in circulating tPA levels observed in adult rodents may reflect progressive alterations in neuronal and endothelial regulation of tPA expression. Given that tPA pathways are sensitive to Egr1 and epidermal growth factor receptor signaling during eyelid and visual development25–28 and that early anesthesia affects neuronal activity and oxidative balance,57,58 these mechanisms may underlie its long-term downregulation and related neurodevelopmental effects. Since endothelial cells are the main source of circulating tPA, its sustained reduction may reflect subtle, widespread endothelial dysfunction. Consistently, impaired NVC in the somatosensory cortex indicates a dissociation between neuronal activity and vascular response. These alterations may stem from disrupted long-term neurovascular remodeling involving endothelial signaling, such as N-methyl-Dd-aspartate receptor–mediated tPA regulation recently described in neurovascular control.59 Interestingly, a slight vascular surface increase was observed in the PVC of anesthetized mice, possibly reflecting an adaptive response to restore tPA production. This supports the existence of an endothelial feedback loop promoting local angiogenesis to rebalance systemic tPA availability. Altogether, these findings identify circulating tPA as a mechanistic link between early repeated general anesthesia exposure and persistent vascular and neuronal alterations.

tPA production. This supports the existence of an endothelial feedback loop promoting local angiogenesis to rebalance systemic tPA availability. Altogether, these findings identify circulating tPA as a mechanistic link between early repeated general anesthesia exposure and persistent vascular and neuronal alterations. This study reveals a phenotypic convergence between anesthesia-exposed and tPA-deficient mice, suggesting that early repeated general anesthesia disrupts tPA regulation and interferes with shared neurodevelopmental pathways. In this context, tPA modulation by early repeated anesthesia may represent a key driver of neurodevelopmental disruption. tPA is essential for CNS maturation, regulating proliferation, extracellular matrix remodeling, neuronal migration, and synaptogenesis.19–21,60 It is also expressed in developing ocular structures, including the retina,22 supporting its involvement in visual system plasticity and susceptibility to stressors such as general anesthesia. Our results show that general anesthesia alters PVC structure (reduced thickness and area) and vasculature (increased area) in SWISS mice without affecting the LGN, indicating selective vulnerability within the visual pathway. In contrast, anesthesia reduced LGN area in tPAWT mice on a C57BL/6J background, suggesting strain-dependent sensitivity. Even in the absence of anesthesia, tPAnull mice displayed smaller LGN areas than tPAWT controls, indicating that both tPA deficiency and general anesthesia independently affect LGN development. Their combination did not produce additive effects, reinforcing the view that tPA acts as a susceptibility factor rather than a direct mediator of injury. Although general anesthesia has been reported to suppress tPA and promote apoptosis, other studies suggest potential protective or anti-inflammatory effects depending on context.61–65 In the retina, tPA has been linked to excitotoxic injury, yet reduced tPA activity may limit retinal ganglion cell death. Collectively, these observations support a modulatory role for tPA in anesthesia-induced neurotoxicity, particularly in visual structures.24 By influencing excitatory/inhibitory balance and developmental plasticity, tPA emerges as a central regulator of visual system vulnerability during critical windows. Altogether, our findings position tPA as a key integrator of disrupted neurodevelopmental signaling in sensitive tissues such as the retina and cortex.

24 By influencing excitatory/inhibitory balance and developmental plasticity, tPA emerges as a central regulator of visual system vulnerability during critical windows. Altogether, our findings position tPA as a key integrator of disrupted neurodevelopmental signaling in sensitive tissues such as the retina and cortex. This work shows that repeated early exposure to isoflurane perturbs endogenous tPA pathways, leading to lasting impairments in visual system maturation. These insights identify the retina as a sensitive model for studying anesthesia-induced neurodevelopmental vulnerability, with implications for pediatric anesthesia safety.

work shows that repeated early exposure to isoflurane perturbs endogenous tPA pathways, leading to lasting impairments in visual system maturation. These insights identify the retina as a sensitive model for studying anesthesia-induced neurodevelopmental vulnerability, with implications for pediatric anesthesia safety. Our study highlights the long-term neurodevelopmental effects of early repeated general anesthesia exposure while acknowledging key limitations and translational perspectives. The design followed the protocol of Salaün et al.3 to ensure reproducibility and direct comparison with previous findings. Therefore, we used a repeated general anesthesia exposure paradigm, which should be considered when interpreting translational relevance, as preclinical and clinical protocols are not perfectly aligned in duration, frequency, or developmental timing. To maintain methodologic consistency and minimize variability related to sex-dependent developmental trajectories, only male rodents were included, in accordance with ethical guidelines. This choice is supported by previous evidence showing marked sex-dependent differences in anesthetic susceptibility and outcomes.66–68 Together, these studies highlight the complexity of sex-related effects and justify controlling for sex as a biologic variable when establishing mechanistic models. Nonetheless, future work will be required to determine whether the mechanisms identified here similarly apply to females. Additional limitations include the use of albino SWISS mice and the inherent variability of behavioral testing,69 as strain-specific differences (e.g., SWISS vs. C57BL/6J) may influence anesthesia sensitivity and tPA signaling. Furthermore, the absence of a gain-of-function or rescue approach to determine whether restoring tPA activity could mitigate general anesthesia–induced effects represents an important limitation. Although technically challenging in early postnatal models, such experiments would be instrumental in establishing a causal link between tPA signaling and general anesthesia–induced visual alterations. Moreover, further studies are needed to explore broader time windows, other molecular pathways (e.g., apoptosis, oxidative stress) and to validate findings in human or nonhuman primate models. These clarifications will enhance our understanding of general anesthesia’s developmental impact and inform safer pediatric anesthesia protocols.

studies are needed to explore broader time windows, other molecular pathways (e.g., apoptosis, oxidative stress) and to validate findings in human or nonhuman primate models. These clarifications will enhance our understanding of general anesthesia’s developmental impact and inform safer pediatric anesthesia protocols. In conclusion, our study demonstrates that, in a male mice model, early repeated general anesthesia exposure induces structural and functional changes in the developing visual system, affecting retinal and cortical regions without vascular inflammation. These include delayed eyelid opening, impaired visual abilities, and reduced retinal and PVC thickness, indicating lasting disruptions in the visual pathway. Similar effects in tPA-deficient mice, along with reduced circulating tPA after general anesthesia, suggest a shared mechanism by which general anesthesia may impair visual maturation through tPA signaling. This study offers new insights into anesthesia-related neurodevelopmental alterations. A key question emerges: could visual screening offer a window into broader neurodevelopmental effects, by predicting or reflecting long-term anesthesia-related brain changes? The authors thank the Comete Lab (UMR1075 UNICAEN/INSERM) for granting us access to the optokinetic reflex test and Prof. Francis Veyckemans, M.D., Ph.D. (Université de Louvain, Louvain-la-Neuve, Belgium), for insightful feedback, careful editing, and valuable guidance throughout this work.

In conclusion, our study demonstrates that, in a male mice model, early repeated general anesthesia exposure induces structural and functional changes in the developing visual system, affecting retinal and cortical regions without vascular inflammation. These include delayed eyelid opening, impaired visual abilities, and reduced retinal and PVC thickness, indicating lasting disruptions in the visual pathway. Similar effects in tPA-deficient mice, along with reduced circulating tPA after general anesthesia, suggest a shared mechanism by which general anesthesia may impair visual maturation through tPA signaling. This study offers new insights into anesthesia-related neurodevelopmental alterations. A key question emerges: could visual screening offer a window into broader neurodevelopmental effects, by predicting or reflecting long-term anesthesia-related brain changes? The authors thank the Comete Lab (UMR1075 UNICAEN/INSERM) for granting us access to the optokinetic reflex test and Prof. Francis Veyckemans, M.D., Ph.D. (Université de Louvain, Louvain-la-Neuve, Belgium), for insightful feedback, careful editing, and valuable guidance throughout this work. Supported by Normandy University (Caen, France; UMR1237), the Institut National de la Santé et de la Recherche Médicale (INSERM, UMR-S U1237), the Fondation de France (grant No. 00110944, young researcher grants in ophthalmology and neuro-ophthalmology, 2020), and the Norman Research Council (SeeBrain project, 2023). Dr. Bonnet was supported by Caen–Normandie University (Caen, France) for teaching duties during the project course (24 h/yr).

ERM, UMR-S U1237), the Fondation de France (grant No. 00110944, young researcher grants in ophthalmology and neuro-ophthalmology, 2020), and the Norman Research Council (SeeBrain project, 2023). Dr. Bonnet was supported by Caen–Normandie University (Caen, France) for teaching duties during the project course (24 h/yr). The authors declare no competing interests.

The authors thank the Comete Lab (UMR1075 UNICAEN/INSERM) for granting us access to the optokinetic reflex test and Prof. Francis Veyckemans, M.D., Ph.D. (Université de Louvain, Louvain-la-Neuve, Belgium), for insightful feedback, careful editing, and valuable guidance throughout this work.

Supported by Normandy University (Caen, France; UMR1237), the Institut National de la Santé et de la Recherche Médicale (INSERM, UMR-S U1237), the Fondation de France (grant No. 00110944, young researcher grants in ophthalmology and neuro-ophthalmology, 2020), and the Norman Research Council (SeeBrain project, 2023). Dr. Bonnet was supported by Caen–Normandie University (Caen, France) for teaching duties during the project course (24 h/yr).

Supplemental Data 1. General anesthesia has no effect on locomotion/anxiety, https://links.lww.com/ALN/E316 Supplemental Data 2. General anesthesia does not alter ocular morphology, https://links.lww.com/ALN/E317 Supplemental Data 3. General anesthesia does not alter retinal vasculature, https://links.lww.com/ALN/E318 Supplemental Data 4. General anesthesia does not change brain cell density, https://links.lww.com/ALN/E319 Supplemental Data 5. General anesthesia does not alter peripheral immune cell profile, https://links.lww.com/ALN/E320 Supplemental Data 6. General anesthesia does not affect resting CBF, https://links.lww.com/ALN/E321 Supplemental Data 7. General anesthesia alters plasma tPA, not eye anatomy, https://links.lww.com/ALN/E322 Supplemental Data 8. General anesthesia and tPA loss do not affect immuno-MRI signal, https://links.lww.com/ALN/E323 Supplemental materials, https://links.lww.com/ALN/E324