Hypoxic–ischemic injury causes functional and structural neurovascular degeneration in the juvenile mouse retina

Ischemic stroke is a major cause of long-term disabilities, including vision loss. Neuronal and blood vessel maturation can affect the susceptibility of and outcome after ischemic stroke. Although we recently reported that exposure of neonatal mice to hypoxia–ischemia (HI) severely compromises the integrity of the retinal neurovasculature, it is not known whether juvenile mice are similarly impacted. Here we examined the effect of HI injury in juvenile mice on retinal structure and function, in particular the susceptibility of retinal neurons and blood vessels to HI damage. Our studies demonstrated that the retina suffered from functional and structural injuries, including reduced b-wave, thinning of the inner retinal layers, macroglial remodeling, and deterioration of the vasculature. The degeneration of the retinal vasculature associated with HI resulted in a significant decrease in the numbers of pericytes and endothelial cells as well as an increase in capillary loss. Taken together, these findings suggest a need for juveniles suffering from ischemic stroke to be monitored for changes in retinal functional and structural integrity. Thus, there is an emergent need for developing therapeutic approaches to prevent and reverse retinal neurovascular dysfunction with exposure to ischemic stroke.

HI induction in juvenile mice. Hypoxia and ischemia were induced in postnatal day 30 (P30) C57BL/6J mice. P30 is considered a juvenile stage of mouse development, during which the vascular and the nervous systems are reaching maturity. The animals were anesthetized with isoflurane (Butler Schein Animal Health Supply, Reno, NV) (5% for induction, 2-3% for maintenance) in 30% oxygen mixed with nitrous oxide. The body temperature of the mice was maintained at 36 °C using a heated surgical table. Under a surgical microscope, a midline skin incision was made in the ventral neck, and the trachea was visualized through the muscle overlying it. The left common carotid artery was freed from the left common jugular vein and vagus nerve by blunt dissection, electrically cauterized and cut. In addition to the brain, the hypoxic-ischemic insults generated by this procedure apply to the eye as well; the ophthalmic artery is a branch from the common carotid artery. The incision was injected with 0.5% bupivacaine and closed with a single 6-0 silk suture. Animals were returned to their cages and monitored continuously for a 2 h recovery period. To induce unilateral ischemic injury, animals were then placed after 2 h of the left common artery occlusion in a hypoxia chamber (BioSpherix Ltd, Redfield, NY) equilibrated with 10% O 2 and 90% N 2 at 36 °C for 50 min. This is a well-characterized model of neonatal HI and results in reproducible brain injury ipsilateral to the electrocauterized and transected left common carotid artery. We have successfully used this model to study the effect of HI on the neurovascular retina in mouse neonates 54 . Full-field electroretinography. Twenty-one P30 animals from three cohorts (N = 6, 7, and 8 animals) were subjected to electroretinography (ERG) recording 7 days following the HI procedure induction (P30D7) that was repeated on D14, D21, D45, and D65. The first time point (D7) was chosen because (1) the animals had to be transferred to a different animal facility for ERG recording after the induction of the HI procedure, (2) animals were allowed to recover from the HI procedure for transfer before ERG recording. Since the first ERG recoding was done at D7 (1 week after HI induction), we then decided to perform additional ERG recoding weekly, twice (D14 and D21), followed by approximately triweekly, twice (D45 and D65). Collectively, these consecutive five-time points were chosen to help determine whether the visual functional integrity recovers or deteriorates with time after HI exposure. Mice were housed in a room with controlled temperature, humidity, and light-dark cycle and were dark-adapted overnight before ERG recording. Under dim-red illumination, animals were anesthetized using an intraperitoneal injection of ketamine (80 mg/kg) and xylazine (16 mg/kg). For local anesthesia, a drop of 0.5% proparacaine hydrochloride was topically applied, and the pupil was dilated with an application of a drop of 1% tropicamide. While under anesthesia, animals were kept on a heating pad (37 °C) to prevent hypothermia. Corneal full-field flash ERG was recorded from mouse eyes using Espion system (Diagnosys LLC, MA) in accordance with the standards of International Society for Clinical Electrophysiology of Vision (ISCEV) (Doc Ophthalmol (2015) 130:1-12) adapted for mice. A drop of sterile 2.5% hypromellose ophthalmic solution (Goniovisc, HUB pharmaceuticals LLC, CA) was applied to the cornea of the dilated eyes to prevent desiccation and dehydration of the eye, and for electrical contact with the recording electrode. The reference needle electrode was inserted through the cheek, and the ground electrode was subcutaneously inserted near the base of the tail. ERG was recorded using Espion system colordome Ganzfeld for uniform illumination. Full-field ERG recording was achieved by exposing both eyes (left eye HI injured; right eye, control) simultaneously to increasing flash intensities (0.03-30 cd s/m 2 ) for 400 ms. An interval 60 s was maintained between two different flash intensities. At lower flash intensities (0.03, 0.1, 0.3, 1 and 3 cd s/m 2 ) 10 flashes were presented and averaged, with an interval of 4 s between the flashes and at higher flash intensities (10 and 30 cd m 2 ), 4 flashes were presented and averaged, with an interval of 10 s between flashes.
Analyses of the data were carried out using Espion software (Diagnosys LLC, MA) and analyzed using Orig-in2018b (OriginLab Corp., MA). For quantification, the a-wave amplitude was measured from baseline to the trough of the negative deflection of the response; the b-wave amplitude was measured from the negative trough to the maximum positive peak of the response. For the acquisition of c-wave, the eyes were stimulated with light flashes of 25 cd s/m 2 intensity for 4 s. The amplitude of the c-wave was measured from the trough of the b-wave to the next positive peak. The stimulus-response exponential fit of the dark-adapted b-wave amplitude was derived using standard Naka-Rushton function: www.nature.com/scientificreports/ where R is the response amplitude at stimulus intensity (I), R max is the maximum response amplitude, K is the stimulus intensity (I) that produces a response amplitude that is half of R max , and n is proportional to the slope of the curve at the point where the contrast is taken to be K 55 . The R max , K, and n parameters were determined using commercial software OriginPro 2020 (OriginLab Corp., Northampton, MA).
The grading of retinal damage. Exposure to HI conditions can result in mild, moderate, or severe injuries in human, rat, and mouse brain [56][57][58] and mouse and rat retina 54,59 . While the inner retina (responsible for b-wave) is especially susceptible to various kinds of ischemic conditions, the outer retina (responsible for a-wave) is not as susceptible 54,59 . The ratio values of the b-wave/a-wave amplitudes (b/a ratio) from injured eyes were used to categorize retinal damage into mild, moderate, or severe groups. The b/a ratio above 2 is typically considered normal, which is the case with control eyes 60 . To categorize the HI injured eyes in this study into three groups, the following cutoff values of b/a wave ratio were used. HI exposed eyes with ratios above 1.75 were considered mildly injured, between 1.35-1.75 were considered moderately injured, and eyes with ratios below 1.35 were considered severely injured.
Optical coherence tomography (OCT). Acquisition. OCT images of the retinal microarchitecture were obtained using the Spectralis HRA + OCT system (Heidelberg Engineering Inc., Heidelberg, Germany). Animals were anesthetized with xylazine (10 mg/kg, Akorn) and ketamine (100 mg/kg, Akorn, Lake Forest, IL). Pupils were dilated using 0.5% tropicamide (Akorn), and contact lens was used to prevent corneal dehydration. While anesthetized, animals were kept on a heating pad (37 °C) to prevent hypothermia. OCT volume scans (30° × 25° degrees with 61 individual b-scans, 120 µm distance between B-scans) were obtained, centered on the optic nerve head of each eye (left eye, HI injured; right eye, control) using the instrument's automatic real-time tracking mode (ART) averaging ten frames per b-scan. Immunofluorescence staining of retinal wholemounts. Immunofluorescence staining of retinal wholemounts was performed as we previously described with little modification 54,61 . Eyes were enucleated immediately post-mortem following euthanasia and fixed in 4% PFA for 10 min at room temperature, washed three times in PBS, and then transferred to methanol and kept at − 20 °C until stained. On the day of staining, eyes were rehydrated in PBS for 1 h on a rocker at room temperature. Retinas were dissected in PBS and then washed in PBS three times, 10 min each, and incubated in blocking solution (3% protease-free bovine serum albumin (BSA), and 0.3% Triton X-100 in PBS) for 1 h. Retinas were then incubated with rabbit anti-collagen IV (Millipore, AB756P) (1/500) and Guinea pig anti-GFAP (Synaptic Systems, 173004) (1/500) in the blocking solution at 4 °C overnight. Biotinylated Griffonia Simplicifolia Lectin I (GSL I) isolectin B4 (Vector Laboratories, B-1205) (1/50) was also used to stain the vasculature in the blocking solution at 4 °C overnight. Following incubation, retinas were washed three times with PBS, 10 min each, incubated with fluorescently conjugated secondary antibody, diluted (1/500), for 5 h at room temperature (RT), washed four times with PBS, 30 min each, and mounted with inner retina uppermost on a slide with DAPI Fluoromount-G (Southern Biotech). Both secondary antibodies were obtained from Jackson ImmunoResearch Laboratories, including donkey anti-rabbit-Cy3 (cat. No. 711-165-152), donkey anti-guinea pig-Alexa Fluor 488 (cat. No. 706-545-148). Streptavidin-Alexa Fluor 647 was obtained from ThermoFisher (ThermoFisher; cat. No. 016-600-084). Retinas were viewed by fluorescence microscopy, and images were captured in digital format using Nikon confocal microscope system A1+. Captured images were analyzed using NIS elements viewer (Nikon). Images from the right eye served as control for the left eye (HI injured eye).

Trypsin-digested retinal vessel preparations.
Eyes enucleated immediately post-euthanasia with carbon dioxide were fixed in 4% PFA in 0.1 M PBS for a minimum of 1 week. The cornea and lens were removed, and the whole retina was obtained under the dissecting microscope and rinsed three times in PBS, twice in distilled water, and then washed in fresh distilled water overnight at room temperature. The next day, distilled water was aspirated, and the retinas incubated in 3% trypsin (Trypsin 1:250, Difco) prepared in pH 7.8 containing 0.2 M NaF, 0.1 M maleic acid, 0.1 M Tris, at 37 °C overnight. The next morning, the whole retina was retrieved and was beaten by a thick hairbrush to loosen and separate away the nonvascular cells from the vasculature. Clean retinal blood vessels were then radially cut proximal to the optic nerve four times and then flat-mounted on glass slides for hematoxylin and periodic acid-Schiff (PAS) staining. Morphology of the nucleus was used to distinguish endothelial cells from pericytes. The endothelial cell nucleus is elongated or oval and positioned within the vessel wall along the axis of the capillary, while the pericyte nucleus is spherical, small but stain densely, and generally have a bulgy position on the capillary wall. The intact retinal wholemounts with highquality staining were coded and subsequently used for counting. One sampling image (0.273 mm 2 ) was captured from the periphery of each of the four quadrants of the retina. Thus, the total area of 0.273 × 4 = 1.09 mm 2 per retina was used for counting. Locations where the vasculature folded on itself on the slide were not used for counting. Only retinal capillaries, but not large blood vessels, were included in the cell counts. www.nature.com/scientificreports/ acellular capillaries, trypsin-digested retinal vessels were prepared as described above. Acellular capillaries were counted in the same field areas used to count the endothelial cells and the pericytes. Acellular capillaries were defined as collapsed capillary-sized tubes but without any nuclei along their length. After the cells and acellular capillaries were quantified, all four images were averaged to determine the mean number of cells and vascular density in the retina. Data from the right eye served as control to the HI injured left eye.

Statistical analysis.
Each experimental group (mild, moderate, and severe HI group) was compared to the control group using the Student's unpaired t test (two-tailed). A p-value < 0.05 was considered significant, with *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. For a single data point, mean ± standard deviation is reported for each condition. Statistics were calculated using GraphPad Prism version 8 (GraphPad Software, La Jolla, CA).

Results
Retinal function is impaired at variable levels in juvenile mice subjected to hypoxiaischemia. Juvenile C57BL/6J mice, postnatal day 30 (P30), were exposed to hypoxia-ischemia conditions according to the Rice-Vannucci model 57,62 . The left, but not the right, common carotid artery was occluded and after 2 h the animals were exposed to hypoxic conditions (10% oxygen) for 50 min. The resulting injury was limited to the left eye. The right eye from these animals was used as control. Electroretinography (ERG) was used to assess the functional integrity of the retina in HI exposed juvenile mice. After HI exposure at P30, 21 mice from 3 cohorts were assessed by ERG on post-HI day 7 (D7) (P30D7), and at timed intervals thereafter, designated P30D14, P30D21, P30D45 and P30D65. ERG was performed on dark-adapted mice and eyes were subjected to light flashes of sequentially increasing intensities. Figure 1A shows representative ERG a-and b-waves generated using 10 cd s/m 2 flashes, and the c-wave generated using 25 cd s/ 25 cd s/m 2 intensities (c-waves, right panel) representing a control (right) eye (black) and HI exposed (left) eyes with different waveform profiles. At 10 cd s/m 2 ; some eyes were almost completely normal (green), some other eyes showed a slight reduction in b-wave amplitude (blue) and a third group of eyes showed a substantial reduction in both a-and b-wave amplitudes (red). The c-waves of the same representative HI exposed eyes as in left panel were unaffected. (B) Shows the average plot of b/a ratio of control eye group and the three groups (mild, moderate, or severe groups) of HI injured eyes. (C) Shows amplitude (µV) of a-waves and b-waves in response to a series of intensities of single light flashes in control eyes, and in eyes with mild, moderate, and severe P30D21 HI injuries. Solid line is the Naka Rushton exponential fit of the data as an intensity-response function. (D) Shows summary results for amplitudes of a-wave, b-wave and c-wave obtained from control eyes, and eyes with mild, moderate, and severe HI at P30D7, P30D14, P30D21, P30D45 and P30D65. Peak amplitudes were collected at flash intensity 10 cd s/m 2 for a-wave and b-wave and 25 cd s/m 2 for c-wave. (E) shows amplitude responses of the four major oscillatory potential (OP) peaks of control eyes, and from eyes with mild, moderate, and severe HI exposure. Data from 21 animals from three cohorts (N = 6, 7, 8 animals) were analyzed and presented as mean ± SD. www.nature.com/scientificreports/ m 2 flash intensities and recorded over 4 s. Compared with the control eye response (black traces), either there was no reduction in a-or b-wave forms (green traces); moderate reduction in the b-wave only (blue traces), or substantial functional impairment with mild reduction in a-wave and substantial reduction in b-wave (red traces) observed in the HI exposed eyes. The c-wave appeared normal for all HI exposed eyes ( Fig. 1A; right panel). The inner retina, responsible for the ERG b-wave, is especially susceptible to various kinds of ischemic conditions, while the outer retina, responsible for the ERG a-wave, is less susceptible to HI 54,59 . Thus, the ratio of b-wave/a-wave amplitudes (b/a ratio) from P30D21 injured eyes were used to categorize retinal damage into mild, moderate, or severe groups 60 . The b/a ratio above 2 is typically considered normal, which was the case with control eyes (Fig. 1B; black). To categorize the HI injured eyes in this study into three groups, the following cutoff values of b/a wave ratio were used. HI exposed eyes with ratios above 1.75 were considered mildly injured ( Fig. 1B; green), between 1.35-1.75 were considered moderately injured ( Fig. 1B; blue), and eyes with ratios below 1.35 were considered severely injured ( Fig. 1B; red). These differences in severity of injury are well documented responses of the central nervous system, including the retina and brain, in different individuals after HI exposure in both humans and preclinical Rice-Vannucci rodent model 58,59,63,64 . The four main components of an ERG response from each eye were further analyzed. The peak amplitudes of the a-wave derived from photoreceptors, the b-wave derived from the inner nuclear layer; mainly bipolar cells, the oscillatory potentials, OP; potentially derived from the feedback pathways among ganglion cells, amacrine cells and bipolar cells, and the c-wave derived from retinal pigment epithelium (RPE). Figure 1C shows a-wave and b-wave amplitudes (µV) as a function of flash intensities in mild, moderate, and severe P30D21 HI injury. As compared to the control eyes, the a-wave amplitude ( Fig. 1C; upper panels) remained unperturbed in eyes with mild (green curve) and moderate (blue curve) injury, while eyes with severe injury (red curve) showed reduced a-wave amplitudes only at higher flash intensities between 1 to 10 cd s/m 2 . For b-wave comparisons ( Fig. 1C; lower panels), mildly injured eyes showed slightly reduced amplitudes (green curve). In contrast, eyes with moderate (blue curve) and severe (red curve) injuries showed significantly lower b-wave amplitudes.
To more comprehensively examine retinal function, we used Naka-Rushton equation fitting 55 . Utilizing different parameters, the Naka-Rushton equation quantifies selective changes in b-wave responses to a range of light intensities. Values of different Naka-Rushton equation parameters, R max, K , and n, are shown in Table 1. R max reflects both the cell number and the increment (μV/quanta) related to each b-wave generating cell. The HI injured eyes showed reduced R max values that correspond with the severity of the injury. K represent the sensitivity responses of the retina to flash stimulus; an increase in K value means a stronger flash stimulus is needed to generate b-wave of similar amplitude. K values of the mildly injured group exhibited slight enhancement in sensitivity response as compared to their control eyes. In contrast, the moderately injured eyes showed a moderate reduction in sensitivity, and the severely injured eyes showed a dramatic reduction in light sensitivity as compared to their corresponding control eyes (Table 1). Together, results of the b-wave amplitude fitting to Naka-Rushton equation further confirmed our categorization, whereby the three HI injury groups showed distinct values for the different parameters of the Naka-Rushton equation.
Amplitudes of a-wave and b-wave collected at additional time points (P30D7, P30D14, P30D45, and P30D65) were similar to those obtained for P30D21 (Fig. 1D). C-wave amplitudes appeared to be less affected over time by HI conditions (Fig. 1D). Amplitude values for the four OPs were calculated and summarized in Fig. 1E. Both control eyes, and mild HI injury eyes had similar OP amplitudes ( Fig. 1E; black vs. green plots). In contrast, the eyes with moderate ( Fig. 1E; blue) and severe ( Fig. 1E; red) injuries showed significantly reduced OP1, OP2, and OP3 amplitudes compared with control eyes. OP4 amplitude was not affected by HI injury relative to the control eyes, and all eyes with varying degrees of HI injury had similar OP4 amplitudes. Taken together, our ERG data demonstrated that exposure of the juvenile mice to HI conditions irreversibly compromised the functional integrity of retinal neurons, especially in the inner retina.

Retinal layers show thinning in juvenile mice subjected to hypoxia-ischemia.
To examine the effect of HI on the morphology of the retinal layers, we used optical coherence tomography (OCT) to image the eyes in vivo. OCT is widely used to diagnose many eye diseases, including various ischemic eye conditions, including diabetic retinopathy, glaucoma, and macular degeneration. Because ERG studies showed that the HI induced damage is irreversible, we reasoned that assessing the integrity of the retinal structure using OCT at a single time point is sufficient. Eleven additional P30 mice were subjected to HI conditions, and OCT was performed 90 days afterward (P30D90). The OCT images allowed for segmentation of eleven major layers of Table 1. Naka-Rushton parameters (means ± SD) for dark-adapted b-wave amplitudes of P30D21 control and HI injured eyes. R max is the maximum response amplitude to a range of flash intensities (I)s given in cd s/m 2 . K is the stimulus intensity (I) that produces a response amplitude that is half of R max . n is proportional to the slope of the curve at the point where the contrast is taken to be K. www.nature.com/scientificreports/ the retina identifiable in hematoxylin and eosin (H&E) stained adult mouse retina ( Fig. 2A). For the purposes of this study, full retinal thickness along with the thickness of ganglion cell layer (GCL), inner plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer (OPL), outer nuclear layer (ONL), and retinal pigment epithelium (RPE) layer were measured in scans acquired from retina superior to the optic nerve head of the right (control) and the left (injured) HI eyes of each animal. Animals with mild, moderate, and severe injuries were identified here based on whole retinal thickness. The mild HI group showed similar thickness to control eyes. In contrast, both the moderate and the severe HI groups demonstrated significantly different retinal thick- www.nature.com/scientificreports/ nesses from control eyes. The severe HI group displayed significant thickness differences in GCL, IPL, and INL. The OPL was significantly thinner in both moderate and severe HI groups, but not in the mild HI group. No significant differences between eyes and groups were observed in either the ONL or RPE thickness for the three HI groups (Fig. 2B).

Development of astrogliosis in the retinas of juvenile mice subjected to hypoxiaischemia.
Astrocytes reside in the retinal GCL and nerve fiber layer. They interact with the retinal ganglion cells and blood vessels in the layer 65 . Müller cells span the retina from inner to outer limiting membranes with cell bodies that reside in the inner nuclear layer of the retina. Both astrocytes and Müller cells are referred to as macroglia or glial cells of the retina. Glial cells play crucial roles in homeostasis and function of neurons and blood vessels 66 . Under ischemic conditions in the central nervous system, glial cells become reactive. Cardinal feature of glial reactivity are the upregulation of glial fibrillary acidic protein (GFAP) and changes in the morphology of glial somas and processes 67,68 . To study the reactivity of the glial cells in the retina, GFAP expression was examined in immunolabeled wholemount retinas from control (right eye) and HI injured (left eyes), P30D65 animals. Representative images from the superior periphery of a control retina, and retinas with mild, moderate, and severe HI injury are shown in Fig. 3. Astrocytes in the control retina displayed a stellate shape with distinct soma and thin cytoplasmic processes, with a morphology typical of normal retinal astrocytes. The astrocytes in the control retina tended to touch each other but without any bundling or forming scars. Furthermore, no detectable GFAP expression was observed in Müller cells. With mild HI, retinal astrocyte morphology grossly appeared similar to their control counterparts. However, Müller cells at the retinal periphery edge showed reactivity, as GFAP expression was detectable in their trunks and end-feet; the trunk normally extends throughout the inner retina until the end-feet reach the ganglion cell/nerve fiber layer and inner limiting membrane (Fig. 3). In the moderately and severely affected retinas, astrocytes showed features typical of reactivity. They lost their stellate shape; cytoplasmic processes became thicker and tended to bridge together with processes from other cells and formed scars at the retinal periphery (Fig. 3). Furthermore, Müller cells appeared reactive with enhanced GFAP expression.

HI exposure results in retinal vascular degeneration in juvenile mice.
To investigate the effect of HI on the integrity of blood vessels in the retina of juvenile mice, wholemount retinas from control (right eye), www.nature.com/scientificreports/ mild, moderate, and severe HI (left eyes) of P30D65 mice were immunolabelled with anti-collagen IV antibody and stained for isolectin B4. Collagen IV is expressed in basement membrane of blood vessels regardless of whether the vessel is perfused (live vascular cells) or not (loss of vascular cells). Isolectin B4 labels perfused blood vessels with viable vascular cells. Representative images of blood vessels in the superior retinas of control and HI injured eyes are shown in Fig. 4. Retinas from control eyes displayed normal vascular structure. The blood vessels in mildly HI injured retinas had a similar vascular structure to those in the control eyes. In contrast, retinas of both moderately and severely HI injured eyes showed a significant vascular degeneration, especially at the periphery. This damage resulted in fewer blood vessels with viable vascular cells (co-labeled for collagen IV and isolectin B4) and an abundance of acellular capillaries (staining only with anti-collagen IV) (Fig. 4).
To quantify the effects of HI on retinal vasculature integrity and endothelial cell and pericyte numbers, trypsin digests were prepared from retinas of control eyes and eyes with mild, moderate, and severe P30D9 HI injury (Fig. 5). Grossly, unlike retinas from animals with mild HI injury, retinas with moderate and severe HI injury in P30D9 animals had lower vascular density as compared with retinas from the control eyes. The numbers of pericytes, endothelial cells, and acellular capillaries and the pericyte/endothelial (PC/EC) ratios were quantified (summarized in Fig. 5). The number of pericytes was significantly lower in animals with severe HI. The number of endothelial cells was lower in both the moderately and severely HI injured retinas. Pericyte/ endothelial cell (PC/EC) ratios were significantly higher in mild and moderate injury HI groups because EC numbers were reduced more than the PC numbers in these two groups when compared with the control group. PC/EC ratio was not different in the severe HI group because both the EC and PC equally decreased in this group when compared with the control group. The number of degenerated capillaries was significantly higher in all three HI groups as compared to the control group. Thus, HI caused retinal vascular degeneration in juvenile mice, which varied in magnitude and was consistent with the severity of the retinal injury as determined by the reduced retinal function.

Discussion
Here we determined that HI in juvenile mice disrupts retinal neurovascular integrity and function. Our data demonstrated interindividual variability in retinal injury after exposure to HI. This is in line with previous reports that demonstrated exposure to HI conditions resulted in various levels of injuries (mild, moderate and severe) in different individuals including human, rat and mouse neonatal retinas 54,59,63,69,70 and brains [56][57][58]64 . The exact reasons for these interindividual variabilities in HI-induced injuries are not completely understood. Age may play a role as the magnitude of severe and moderate forms of the retinal vascular injury after exposure to HI are more in neonatal mice 54 as compared to juvenile mice (this report). Hemodynamic variations may affect the www.nature.com/scientificreports/ severity of HI-induced injury 71 as different retinas have unique vascular structures. Sex of the animal may also contribute to the severity of injury of the retina as in the brain after exposure to HI conditions 72,73 . Although this was not directly addressed here, the impact of sex on the severity of retinal injury deserves a more careful evaluation and is a subject of future investigation.  www.nature.com/scientificreports/ We noted retinal macroglia reactivity and retinal capillary degeneration in moderate and severe HI groups, especially in the peripheral retina. Functional and structural studies using ERG and OCT analysis in vivo demonstrated preservation of the outer retinal structure and function including the a-wave, which reflects the activity of photoreceptors 74 . The c-wave, which derives from RPE cells 75 and depends on photoreceptors integrity was normal in all animals 76 . Together, these results suggest that HI does not affect photoreceptor and RPE physiology, or their interactions, in juvenile mice. The OCT results corroborated these findings, as both the OPL and RPE layers were preserved in all examined animals. These studies will further benefit by carefully evaluating the expression and localization of synaptic markers like PSD95, Gephyrin, Ctbp2, mGluR6 and VGlut1 in future studies.
Both ERG and OCT studies supported inner retinal damage in animals subjected to HI. Both moderate and severe HI groups displayed a reduced ERG b-wave amplitude and a shift in light sensitivity. The ERG b-wave response mainly reflects the electrical activity of bipolar cells 77 , an activity that initiates at its dendrites in the OPL. In addition, pharmacological and genetic mouse studies showed that horizontal cells 78 and the third-order neurons, amacrine cells, and ganglion cells [79][80][81] contribute to the b-wave response. Therefore, a loss in amplitude and reduction of light sensitivity indicates degeneration of inner retina cells. Three of the four ERG oscillatory potentials were also attenuated in moderate and severe HI groups. Oscillatory potentials reflect the inhibitory feedback activity among amacrine cells, bipolar cells, and ganglion cells 82 . The intermediate OP responses which we found to be severely reduced following HI insult in mice, are generated by action-potential-independent interactions between third-order neurons in the ON pathway of the rabbit retina 83 . Thus, exposure to HI leads to the impaired neuronal activity of all neurons in the inner retina. This is supported by the OCT data, which showed a significant reduction in the overall thickness of the whole retina, and specifically the OPL, in moderate and severe HI groups. The OPL layer harbors synapses between photoreceptors and the bipolar cell second-order retinal neurons.
The amplitude of b-wave can be significantly reduced under severe or even subtle ischemic conditions in humans and mice as a result of blocking the central retinal artery [84][85][86] . The amplitudes of OPs are responsive to even subtle retinal ischemia when other ERG components remain unchanged, as recognized in diabetic retinopathy [87][88][89] . Furthermore, several mouse knockouts for genes critical for normal vascular development and homeostasis in the retina (e.g., frizzled-4, and Lrp5) result in compromised b-wave and OP amplitudes 90,91 . Our findings are consistent with these reports that further underscore the sensitivity of retinal neurons, specifically those in the inner retina to ischemia. Neuronal degeneration following HI is postulated to result from glutamate excitotoxicity, free oxygen radical accumulation, inflammation, and disruption of the blood-retinal barrier 51 .
Here we observed chronic macroglia activation in the retinas of mice with moderate and severe HI injury. Astrogliosis is noted under ischemic conditions in the central nervous system, including the retina. Prolonged astrogliosis can compromise blood-retinal barrier integrity, which in turn contributes to neurodegeneration 92,93 . Thus, the reactivity of both astrocytes and Müller cells may contribute to the neurovascular degeneration in retinas with HI injury.
In conclusion, this report demonstrates that both neurons and blood vessels in the inner retina of juvenile mice are susceptible to damage by HI. As in the case of ischemic stroke in the pediatric population, our studies showed that the severity of the injury resulting from HI varied among individual mice. Our data suggest that ischemic stroke in pediatric patients (children from 29 days to 18 years old) is very likely to cause retinal damage in addition to brain damage, and therefore warrants clinical follow up to identify and manage potentially debilitating ocular pathology. www.nature.com/scientificreports/