Timed pregnant Sprague Dawley (SD) rats were purchased from Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China). They were delivered to the animal facility at West China Hospital (Chengdu, China) at around 17 days of gestation to allow for acclimatization before parturition. Access to water and standard rat chow was given ad libitum, and animals were provided with a 12-h dark, 12-h light cycle. Rats were treated according to institutional and national guidelines. All animal procedures were reviewed and approved by the Ethical Review Committee of Animal Research of West China Hospital, Sichuan University, Chengdu, Sichuan Province, China (AUP# 2018008A), and performed in compliance with the ARVO statement for the use of animals in ophthalmic and visual research.

STZ (Sigma-Aldrich Chemical Co., Saint-Quentin Fallavier, France) was freshly prepared in citrate buffer (pH 4.5) (Solarbio, China) to a final concentration of 60 mg/mL. For intravitreal (IVIT) injection, freshly made STZ solution was diluted to 1 μg/μL, 5 μg/μL, and 10 μg/μL, respectively. One microliter (1 μL) STZ or citrate buffer was injected into the vitreous. Briefly, neonates were anesthetized on ice for 30–60 s, the eyelids were prized apart gently, and forceps were applied. A topical anesthetic (0.5% proparacaine hydrochloride) was applied dropwise to the eye before injection. Eyes were injected under a dissection microscope using a 30-gauge needle attached to a Hamilton syringe. A drop of ophthalmic antibacterial agent (Ofloxacin ophthalmic solution USP, 0.3%) was applied to the injected eyeball. For systemic delivery, 30 or 60 mg/kg body weight (0.5 or 1 μL/g body weight) was administered via subcutaneous (SC) injections in the abdomen. Control animals received an equal volume of citrate buffer (Buffer group). The pups were returned to their mothers after the injection. The animals remained with their mothers until 21 days of age. Animals were weighed, and tail blood was collected for glucose measurement using a standard glucometer (Contour™ Plus, Ascensia).

For hematoxylin and eosin (H&E) staining, eyeballs were fixed for more than 24 h in 10% neutral buffered formalin, dehydrated subsequently in ethanol, embedded with paraffin, and cut into 4 μm sections. The sections were rehydrated in ethanol, stained with hematoxylin and eosin, and finally mounted.

For immunofluorescence, eyeballs were fixed for 45 min at 4 °C in 4% paraformaldehyde, embedded in OCT (TissueTek 4583), frozen at −80°C and cut into 12 μm sections on Superfrost slides. Antigen retrieval was performed as described in our previous report (Chen et al., 2007). Slides were incubated with blocking solution (1% donkey serum and 0.1% Triton X-100 in PBS) for 1 h, and then with primary antibodies including Ap2a (Santa Cruz, SC-8975); Chx10 (Bremner lab, University of Toronto); Cleaved caspase-3 (Cell Signaling 9661); Cone arrestin or ARR3 (Millipore, AB15282); γ-H2ax (Millipore, 05–636); Onecut2 or OC2 (R&D system AF6294); Ki67 (BD science Pharmingen 550609); PH3 (Upstate 05–598); Rhodopsin (Santa Cruz SC-57433) and Sox9 (Millipore AB5535), overnight at 4°C. Vascular endothelial cells were labeled by FITC-IB4 (Sigma, L2895). Primary antibodies or labeled cells were visualized using donkey anti-mouse, donkey anti-rabbit, and donkey anti-goat antibodies conjugated with Alexa-488, Alexa-568, or Alexa-647 (1:1000; Molecular Probes). Nuclei were counter-stained with DAPI (Sigma, D9542) and mounted with Mowiol medium.

Eyeballs were enucleated and incubated for 45 min for whole-mount staining with 4% paraformaldehyde in PBS. A circumferential incision was made around the limbus to harvest the retina. The retinas were incubated at 4°C with primary antibody Brn3 (Santa Cruz, SC-6026) and FITC-IB4 for 2 days, then with secondary antibody (donkey anti-goat Alexa Fluor 488) for a day at 4°C. After briefly washing with PBS, radial cuts were made to divide the retina into four quadrants to flatten the retina, and flat retinas were mounted with Mowiol.

Stained sections and slides were analyzed using a Zeiss Axio Imager Z2 fluorescence microscope and a Nikon C1si confocal microscope. The positive cells of cleaved caspase-3, Ki67, PH3, EdU, γ-H2ax, and cell-type markers (including Brn3, Rho, ARR3, Chx10, Ap2a, OC2, and Sox9) were counted manually or automatically with ImageJ 1.50b with a cell counter plugin (https://imagej.nih.gov/ij/). The minimal number of images for each group was 12 (two images per section, two sections per retina, and three retinas from three animals). Images of peripheral retinas (which have ciliary body and iris in the same image) and central retinas (which have optic nerve in the same image) were separately counted. Most images for cell counting were captured under a fluorescence microscope using a ×20objective lens. For Ki67 and Rho counting, 60 × objective lenses were also used. The quantitative variables of this study were presented as mean ± standard deviation (Mean ± SD) or standard error (SE). The thickness of ONL (outer nuclear layer) was measured using a microscope program. Representative images were analyzed using the AngioTool software (National Cancer Institute, United States) to quantify the average vessel length and branching points for vascular blood vessel analysis.

To label the S phase of retinal progenitor cells, animals were injected subcutaneously once with EdU (Sigma, 900584, 30 μg/g of body weight) to label all cells in the S phase. The retinas were harvested 0.5 h after EdU injection, fixed in 4% paraformaldehyde for 1 h, and dehydrated in 30% sucrose for 24 h. For retinal sections, EdU was detected with the Click-iT^® EdU Alexa Fluor^® 647 Imaging Kit (Thermo Fisher, C10340) according to the manufacturer’s protocol.

Total RNAs were extracted from dissected rat retinas using TRIzol (Thermo Fisher) and treated with RNAse-free DNAse I (New England Biolabs) to remove genomic DNA. The yield of total RNA was assessed using NanoDrop Microvolume Spectrophotometers (NanoDrop Technologies). The cDNA libraries were prepared using an Illumina TruSeq RNA sample preparation kit, and the quality was assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies). For sequencing, the cDNA libraries were loaded on an Illumina HiSeq 2500 at Biomaker (Beijing, China). The raw sequence reads in FASTQ format were processed and analyzed as previously reported (Liu et al., 2014; Gregorieff et al., 2015). Briefly, the sequencing quality was first assessed using FastQC, and poor-quality 5′end reads were trimmed using a Perl script and then mapped onto the rat genome using TopHat2, allowing for up to 2 mismatches as default settings. Reads mapped onto multiple genomic locations were discarded, and a custom R script was used to calculate the fragments per kilobase million (FPKM) of each gene and obtain the expression profile of each sample (Supplementary Table S1). The expression folds change of each coding gene of the retinas of the STZ group compared with the retinas of the buffer group was calculated using the following formula: Fold change = (FPKM ^STZ +1)/(FPKM ^buffer +1). The genes for expression fold changes greater than 1.54 or less than 0.65 were selected as deregulated genes (DEGs). The heatmap was generated using Heatmapper (Babicki et al., 2016). The function enrichment of DEGs was performed using Enrichr (Chen et al., 2013; Kuleshov et al., 2016); the pathways with a p-value less than 0.05 were chosen to report.

Total RNA was isolated from rat retinas using TRIzol reagent (Thermo Fisher). After quantification by NanoDrop, first-strand cDNA was synthesized from 1 µg of total RNA using the RT reagent Kit with gDNA wiper (Vazyme, China). PCR primers are listed in Supplementary Table S2. Real-time quantitative PCR was performed using the qTOWER 2.2 PCR machine (Analytik Jena, Germany). Tests were run in duplicate on three separate biological samples with HiScript II Q RT SuperMix (Vazyme, China). PCR consisted of 40 cycles of denaturation at 95°C for 15 s, annealing, and extension at 55°C for 30 s. An additional cycle (95°C, 15 s) generated a dissociation curve to confirm a single product. Values obtained for test RNAs were normalized to β-actin mRNA levels.

According to the manufacturer’s instructions, single-cell suspension was prepared from the dissected rat retina with a dissociation kit (Papain Dissociation System Kit, Worthington). Then, the retinal cells were fixed in 70% cold ethanol and incubated with propidium iodide solution (Sungene Biotech, China) before running them on the flow cytometer (FACSAria™ III Cell Sorter, BD). Data were analyzed by ModFit LT 5.0 (Verity Software House, Topsham, ME, United States).

Dark-adapted ERG (RetMINER™, IRC, Chongqing, China) were recorded at P21, P28, P42, and P180. The procedures were performed as previously described (Zheng et al., 2018). In brief, rats were dark-adapted overnight and anesthetized by IP injection of ketamine (100 mg/kg) and xylazine (15 mg/kg). Pupils were dilated with 0.5% tropicamide and 0.5% phenylephrine hydrochloride. A full-field ERG was recorded after subcutaneously inserting a ground electrode near the tail and a reference electrode on the head. A golden-ring electrode was gently positioned on the cornea. All procedures were performed under dim red light. We used GraphPad to illustrate and analyze the data recorded by the ERG device. Because there was no light stimulation in the first 25 s, we used the baseline correction function in GraphPad to normalize the baseline to around 0 μV. Before we measured the amplitude of a-wave and b-wave, we used the bandpass filter function in Clampfit 11.1 software (Molecular Devices) to filter clutter and only used waves with a frequency between 0.3 Hz and 65 Hz in the final version of the dark-adapted ERG. The amplitude of the a-wave was measured from the baseline to the peak of the a-wave, and the b-wave was measured from the trough of the a-wave to the apex of the b-wave. Oscillatory potentials (OPs) were studied using Clampfit 11.1 software (Molecular Devices). The extracted OPs were analyzed by Fast Fourier Transform (FFT) and converted to a single-sided smoothed frequency power spectrum with functions built in the Clampfit. To evaluate the OPs magnitude, areas enclosed by the frequency spectrum curve were calculated to assess the total power of OPs using GraphPad.

All data were presented as mean ± SD or SE. Statistical analysis was undertaken using the GraphPad Prism software (GraphPad Prism Software, Inc., San Diego, CA, United States). The results were analyzed by one-way analysis of variance (ANOVA) followed by Bonferroni correction for multiple comparisons. The threshold for significance was set at p < 0.05.

Intravitreal injection (IVIT) of STZ 1 μg, 5 μg, and 10 μg at postnatal day 1 (P1) did not affect glucose and weight (Figures 1A, B). To test if STZ has direct cytotoxicity to neonatal rat retina, we IVIT 1 μg, 5 μg, and 10 μg STZ at P1 to the right eyes. At the same time, an equal amount of citrate buffer was injected into the left eye of the same animal. We then examined the function and morphology of the retina.

ERG measured at P21 showed that intravitreal STZ (10 μg) reduced the amplitudes of dark-adapted a-waves and b-waves (Figures 1C, D), and the total OPs energy at the light stimulation intensity of 3.0 cd*s/m² (Figures 1E,F). H&E staining of P14 retinal sections identified ONL rosettes in the intravitreal STZ (5–10 μg) injected group (Figure 1G), but 1 μg intravitreal STZ has no effects on retinal morphology (Figure 1G). Cleaved caspase 3 (CC3) staining found that IVIT 5–10 μg STZ induced widespread retinal apoptosis at P3, but 1 μg STZ had only induced a few apoptotic retinal cells (Figures 1H, I). IB4 staining at P8 showed that 5–10 μg intravitreal STZ treatments delayed the retinal angiogenesis, as retinal SVP had not reached the peripheral in both groups, but 1 μg STZ had no measurable effects on retinal angiogenesis (Figures 1J, K). Intravitreal injection of 5–10 μg STZ at P8 did not affect retinal function (Figures 2A–D) and morphology (Figure 2E), survival (Figures 2F, G), and angiogenesis (Figures 2H–J). These results suggested that STZ can directly induce neonatal retinal abnormalities independent of hyperglycemia, but only when STZ was injected at an early stage of retinal development (P1).

STZ only affects neonatal rat retinal cells in the early stage (P1); locally administered STZ after P8 had no effects (Figure 2). Thus, it is very likely that STZ mainly affects retinal progenitor cells (RPCs). Indeed, IVIT 5–10 μg STZ at P1 significantly reduced the number of ki67+ cells at P4 in the central and peripheral retina (Figures 3A, B). However, at P8, while most RPCs exit the cell cycle, IVIT 5–10 μg STZ at P1 significantly increased the number of ki67+ cells (RPCs) in the central and peripheral retina (Figures 3B, C). Flow cytometry analysis of P4 retinal cells showed that STZ reduced the G0/G1 phase but increased the G2/M phase while having no effects on the S phase (Figure 3D). PH3 staining indicated that STZ significantly reduced the M phase (PH3+) cells (Supplementary Figure S1). Thus, STZ blocked the cell cycle of RPCs at the G2 phase. This result suggested that STZ might increase the cell cycle length of RPCs and delay their cell cycle exit and differentiation.

Consistent with this notion, IVIT STZ (10 µg) at P1 had no effects on the numbers of early-born retinal cells, including ganglion cells (Brn3+), horizontal cells (OC2+), amacrine cells (AP2a+) and cones (ARR3+) (Figures 3E, F), but significantly reduced late-born retinal cells, including rods (Rho+), bipolar cells (Chx10+) and Müller glia (Sox9+) (Figures 3E, F). Thus, we concluded that STZ mainly affects the survival and proliferation of RPCs in neonatal rats.

Based on the above results, locally administered STZ at P1 could induce retinal cell death, suppress RPCs proliferation, and delay retinal differentiation and angiogenesis. To understand the underlying mechanism, we analyzed the transcriptome of the P4 retinas of rats receiving IVIT STZ (10 g) by RNA sequencing.

We had two control groups: citrate buffer and saline. We found no enrichment of deregulated genes (DEGs) between citrate buffer and saline. Thus, we focused on the differences between the STZ and citrate buffer groups. We identified 608 STZ-related DEGs, including 416 upregulated DEGs and 192 downregulated DEGs (Supplementary Table S1). Gene list enrichment analysis by Enrichr indicated that the most downregulated STZ-related DEGs were in the cancer pathway and PI3K-Akt pathway (Figure 4A), corresponding to RPCs cell cycle arrest and delayed retinal angiogenesis. The most enriched upregulated pathways were related to cell death and inflammation (Figures 4B, C). The p53 pathway, apoptosis pathway, phagosome, and lysosome pathways may be linked to STZ-induced RPCs death (Figures 4B, C).

It is well known that STZ can induce DNA damage via DNA alkylation. DNA damage can activate the p53 pathways, thus activating many genes whose products trigger cell-cycle arrest, apoptosis, or DNA repair. Indeed, many γ-H2ax + cells were in the STZ-treated retinas at P3 and P4 (Figures 4D, E). The expression of Cdkn1a increased four times in the STZ-treated retina (Figure 4C), which may be a major mechanism of cell cycle arrest as Cdkn1a is a well-known cell cycle inhibitor.

Many inflammation pathways were activated, including the complement, chemokine pathway, NF-kB pathway, and phagocytosis pathway (Figure 4C). Thus, likely STZ-induced RPCs death initiated a profound inflammation reaction in the neonatal rat retinas. Indeed, the IBA1+ microglia increased from P4-P21 in STZ-treated retinas (Figures 4F, G). Consistent with the RNA sequencing result, RT-PCR confirmed that p53 pathway genes (Bbc3, Ripk3, Fas, and cdkn1a) were significantly upregulated. In contrast, cell cycle genes (E2f1, E2f2, Pcna, Mcm3, Mycl, and Mycn) were significantly downregulated in STZ-treated retinas (Figure 4H).

In addition, the Notch pathway (for instance, genes of Hes1, Hes5, Hey1, Dll1, Dll3, and Dll4) were also significantly downregulated, explaining the delayed retinal differentiation (Figure 4H). STZ also induced retinal oxidative stress, as the heme oxygenase-1 (Hmox-1) gene was hugely induced (Figure 4H). STZ had no effects on Hif-1 pathway and adiponectin pathway genes (Figure 4H).

Even though locally injected STZ can induce NSIR, we do not know if systemically delivered STZ can induce similar retinal lesions, as BRB (blood-retinal barrier) may block transferring STZ from circulation into the retina. STZ (30 mg/kg of body weight) was subcutaneous (SC) injected at P1 and P3, and rats were analyzed at P4. The weight (Figure 5A) and blood glucose (Figure 5B) were the same between the control and STZ groups. However, STZ suppressed the proliferation of RPCs (Figures 5C, D) and induced retinal apoptosis (Figures 5C, E). Thus, systemically delivered STZ (low dosage) could pass the BRB and induce NSIR without causing hyperglycemia. This suggests that retinal progenitors may be more sensitive to STZ than neonatal pancreatic β-cells.

While locally or systemically injected STZ can induce NSIR without causing hyperglycemia, we do not know if STZ-induced hyperglycemia has a role in the pathogenesis of NSIR. To directly assess the impact of hyperglycemia on the phenotypes of NSIR, we increased the dosage of STZ from 30 mg/kg to 60 mg/kg for SC injection and injected STZ at different time points. For the SC STZ 60 mg/kg body weight cohort, we did not observe any rat pup death in the P1/P3/P5 and P2/P4/P6 injection group, but in the P3/P5/P7, P4/P6/P8 and P8/P10 injection groups, we found that about half of pups receiving this dosage of STZ died in 1–2 days after the second injection at P6-P12, likely due to hyperglycemia.

Although NSIR is not directly related to hyperglycemia, we do not know whether long-term hyperglycemia will aggravate these pathological changes. Thus, we examined retinal morphology and function in adult rats receiving 60 mg/kg STZ SC injection at P1/P3/P5 or IVIT (10 μg) at P1. H&E staining of 4-week, 6-week, and 6-month-old rats showed similar retinal dysplasia with P14 retinas (Figures 1G, 7A, 9A), indicating little progressive retinal degeneration process. The ONL thickness was equal between these three-time points and between systemic STZ and local STZ administration, suggesting there were no significant effects of hyperglycemia on NSIR at these time points (Figure 9B).

The apoptotic peak was around P3 and P4 when STZ was delivered systemically at P1/P3 (Figure 7C) or locally at P1 (Figure 1I). CC3 staining had not identified any apoptotic events in STZ-treated retinas of P28, P42 and P180 (Figure 9C). ERG assay also showed a similar reduction of the amplitude of a-wave (Figure 9D) and b-wave (Figure 9E), and total energy of OPs (Figure 9F) between P28, P42 and P180. Thus, systemically or locally delivered STZ at P1-3 can damage the neonatal rat retina independently of hyperglycemia without progressive secondary retinal degeneration.

Many chemicals, including oxidizing, methylating, and alkylating agents, as well as neurotransmitters and antibiotics, can induce retinal degeneration. Still, each has different chemical properties and special retinal target cells. For instance, NaIO3 is an oxidizing compound, specifically toxic to the RPE cells (Enzmann et al., 2006; Wang et al., 2014). MNU is a methylated nitrosourea compound with alkylating, carcinogenic, and cytotoxic properties. It interacts with DNA to yield various reaction products such as 7-methyldeoxy-guanosine adduct. Systemic administration of MNU induces retinal photoreceptor death in many animal species in adults (Tsubura et al., 2010), but also induces retinal rosettes formation and retinal dysplasia in neonatal rodents (mice and rats), suggesting toxic to retinal progenitors (Nambu et al., 1998a; Nambu et al., 1998b). NMDA is a specific agonist of the NMDA receptor mimicking the action of glutamate, the major excitatory neurotransmitter in the nervous system (Pal, 2021). Intravitreal injection of NMDA induces cell death of retinal ganglion cells in mice (Huang et al., 2014; Niwa et al., 2016).

Like MNU, STZ is an alkylating agent with carcinogenic and cytotoxic properties. Structurally, it is a moiety of MNU linked to the C-2 position of α-D-glucose (Lenzen, 2008; Zhu, 2022). Because of this glucose moiety, STZ has a high affinity for the glucose transporter Glut2, which is highly expressed in pancreatic β-cells. As such, STZ and MNU have similar chemical properties; STZ has much higher toxicity to pancreatic β-cells than MNU and is widely used to model diabetes (Lenzen, 2008; Zhu, 2022).

While MNU mainly targets photoreceptors and retinal progenitors (Tsubura et al., 2010), we found that STZ only targets retinal progenitor cells. IVIT STZ at P1 induces DNA damage, cell death, and cell cycle arrest of RPCs. Many RPCs were blocked at the G2 phase, causing reduced M-phase cells and delayed cell cycle exit, resulting in reduced late-born retinal cells. IVIT and SC injection STZ after P8, when most RPCs exit the cell cycle (Alexiades and Cepko, 1996), do not induce any measurable retinal phenotypes, supporting the notion that STZ mainly targets RPCs. Indeed, numerous studies utilized STZ in adult rats to induce diabetic retinopathy. Still, these studies seldom identified any retinal phenotypes before 2–4 weeks after STZ injection, indicating STZ has no direct toxic effects on adult rat retinal cells or post-mitotic cells (Barber et al., 1998; Asnaghi et al., 2003; Park et al., 2003; Olivares et al., 2017).

The structure and toxic action of MNU and STZ are very similar (Lenzen, 2008; Zhu, 2022), so why do MNU target photoreceptors and RPCs, but STZ only target RPCs in the retina? MNU is lipophilic, and tissue uptake through the plasma membrane is rapid; however, because of the α-D-glucose substitution, STZ is less lipophilic and must be up-taken through Glut2 (Lenzen, 2008). Thus, the possible reason is that the expression levels of Glut2 are higher in RPCs than in post-mitotic retinal cells. Therefore, RPCs can uptake STZ, but post-mitotic retinal cells cannot. It is less likely due to the blood-retinal barrier (BRB), as IVIT injection of STZ after P8 also cannot induce NSIR, which already bypasses the BRB.

Another interesting finding is that NSIR is not related to hyperglycemia. As STZ can damage both pancreatic β-cells and RPCs, it seems hard to exclude the effects of hyperglycemia on NSIR. We overcame this issue in two ways; first, we injected STZ (5–10 μg in total) directly into the eye by IVIT injection, thus avoiding damaging the pancreatic β-cells and hyperglycemia. The second strategy is using a low dosage of STZ (30 mg/kg) for SC injection, which induced NSIR before the appearance of hyperglycemia. Both strategies prove that STZ can induce NSIR without hyperglycemia. This is like systemic-administered MNU-induced neonatal retinopathy in mice and rats, which does not cause the death of pancreatic β-cells and hyperglycemia (Nambu et al., 1998a; Nambu et al., 1998b).

The STZ neonatal rodent models are common animal models of diabetes (Portha et al., 1989; Baig and Panchal, 2019) and are widely used to study diabetic cataract (Patil et al., 2014), diabetic nephropathy (Shinde and Goyal, 2003; Gandhi et al., 2013), diabetic neuropathy (Ashokkumar et al., 2006; Barragán-Iglesias et al., 2018), diabetes-related retinopathy of prematurity (ROP) (Kermorvant-Duchemin et al., 2013; Fu et al., 2018) and diabetic retinopathy (Mancini et al., 2013). While these studies suggested some hyperglycemia-related changes in different organs, the contributions of the direct cytotoxicity of STZ on cells in the lens, kidney, brain, and retina are generally not considered.

For instance, intraperitoneal (IP) injection of STZ on P1 rats induced transient hyperglycemia (P2-P6) but persistent NSIR features. These features were named neonatal hyperglycemia-induced retinopathy (NHIR) (Kermorvant-Duchemin et al., 2013). Similarly, daily IP injection of STZ from P1-P9 to C57BL/6 J mouse pups also induced hyperglycemia (P8-P10) and NSIR, which was named hyperglycemia-related retinopathy (HAR) (Fu et al., 2018). Both NHIR and HAR were considered novel rodent ROP models related to hyperglycemia, but not the direct cytotoxicity of STZ on retinal cells, as IVIT STZ (1 μg) in P3 rat (Kermorvant-Duchemin et al., 2013) or P1 mouse (Fu et al., 2018) did not affect weight nor blood glucose level and did not modify retinal vasculature. However, these studies only used one dose of STZ (a total of 1 μg) for IVIT. The results of IVIT STZ (1 μg) in these two studies were similar to our observation of the same STZ dosage. Still, when we increased STZ to 5–10 μg for IVIT, STZ induced direct retinal damage without hyperglycemia. Interpreting these results from STZ-induced neonatal models requires special cautions to exclude the direct cytotoxicity effects of STZ.

It is well-known that, similar to MNU, STZ can induce DNA alkylation and methylation, resulting in DNA fragmentation, which overstimulates PARP-1, diminishes the cellular NAD+ and ATP pool, and induces cell death (Yamamoto et al., 1981). The observed effects of STZ on neonatal rat retina are typical reactions of retinal progenitors to STZ-induced DNA damages, which activate the p53 pathway, thus inducing cell death and cell cycle arrest by up-regulating Fas and Cdkn1a (p21cip1).

Retinal apoptosis can further cause retinal dysplasia and delay retinal angiogenesis, as it is well known that retinal degeneration or retinal apoptosis can delay the normal development of retinal angiogenesis, as the oxygen demand in the retina is reduced (Lahdenranta et al., 2001; Pennesi et al., 2008; Zhou et al., 2018). Cell death and cell cycle arrest of RPCs reduce the pool of RPCs and delay the cell cycle exit of RPCs, resulting in fewer late-born retinal cell types. The retinal inflammation is likely a response to widespread STZ-induced retinal apoptosis (Massengill et al., 2018).

A previous study suggested that hyperglycemia-induced adiponectin (APN) deficiency, resulting in decreased photoreceptor mitochondrial metabolism and reduced Pdgfb expression, is the underlying mechanism of hyperglycemia-related retinopathy (HAR) (Fu et al., 2018). Our RNA sequencing did not identify DEGs in the APN and Pdgfb pathways. Still, the oxidative stress gene Hmox-1 was hugely induced. It is well known that STZ can cause mitochondrial respiratory defects and oxidative stress (Raza et al., 2011; Genrikhs et al., 2017). Thus, we suggest that STZ cytotoxicity may also contribute to the previously reported HAR phenotype in neonatal mice.

Similar hyperglycemia-independent effects of STZ have been reported in skeletal muscle myoblasts (Johnston et al., 2007). In vitro exposures to STZ significantly impaired the proliferative capacity of myoblasts, increased reactive oxygen species (ROS), and induced G2/M phase cell-cycle arrest (Johnston et al., 2007).

As NSIR is a typical reaction of retinal progenitors to DNA damage, it may be a valuable model to study human diseases with features of RPC DNA damages, such as replication stress response (RSR)-related syndromes, including Seckel syndrome (Matos-Rodrigues et al., 2020; Matos-Rodrigues et al., 2020). NSIR is very similar to the retinal phenotypes of alpha-Cre-mediated ATRIP gene knockout mice (Matos-Rodrigues et al., 2020).