Retinal injury mouse model and pathophysiological assessment of the effect of arsenical vesicants

Abstract

The eye is ten times more vulnerable to chemical warfare agents than other organs. Consistently, exposure to vesicant arsenical lewisite (LEW) manifests significant corneal damage leading to chronic inflammation, corneal opacity, vascularization, and edema, culminating in corneal cell death. However, despite the progress has made in the research field investigating arsenical-induced pathogenesis of the anterior chamber of the eye, the retinal damage resulted from exposure to arsenicals has not been identified yet. Therefore, we investigated the effects of direct ocular exposure (DOE) to LEW and phenylarsine oxide (PAO) on the retina. DOE to arsenicals was conducted using the vapor cap method at the MRIGlobal facility or an eye patch soaked in solutions with different PAO concentrations at UAB. Animals were assessed at 1, 3, 14, and 28 days postexposure. Results of the study demonstrated that both arsenicals cause severe retinal damage, activating proinflammatory programs and launching apoptotic cell death. Moreover, the DOE to PAO resulted in diminishing ERG amplitudes in a dose-dependent manner, indicating severe retinal damage. The current study established a prototype mouse model of arsenical-induced ocular damage that can be widely used to identify the key cellular signaling pathways involved in retinal damage pathobiology and to validate medical countermeasures against the progression of ocular damage.

1. Introduction

Arsenicals are highly toxic arsenic-containing inorganic and organic derivatives that cause both acute and chronic tissue damage. Together with mustard agents and phosgene oxime, they constitute a group of chemicals called vesicants. These are included in the priority list of threatening chemical war agents (CWAs) compiled by the US Agency for Toxic Substances and Disease Registry (Tam et al., 2020). Lewisite (LEW) is one of the most powerful arsenical-based CWAs known to readily penetrate clothing and personal protective equipment, making exposed human populations particularly vulnerable. Several countries, including Germany, Italy, the United States, Russia, and Japan, retain significant stockpiles of the arsenical LEW, raising serious concerns over potential public health risks (Tam et al., 2020; Watson and Griffin, 1992). The 2003 accident in Qiqihar in northeast China, which exposed residents to LEW, highlights the severity, spectrum, and late appearance of multiple clinical manifestations of exposure. These manifestations include corneal damage and, most importantly, constricted vision (Isono et al., 2018). Among the few studies examining LEW exposure, those that describe LEW-induced rabbit corneal damage are of particular importance (Goswami et al., 2016; Tewari-Singh et al., 2016, 2017).

Indeed, the eye is ten times more susceptible to CWA exposure and vesicants than other organs (Amini et al., 2020; Ghazanfari et al., 2009). The vesicants LEW, sulfur mustard (SM), and nitrogen mustard (NM) easily penetrate ocular tissue, causing surviving individuals to suffer from delayed effects for decades (Sezigen and Kenar, 2020). Photophobia and inflammation occur at the acute stage, while symptoms including chronic inflammation, corneal opacity, corneal vascularization, and blindness manifest later (Sezigen and Kenar, 2020). While such vesicants can affect the eye systemically, locally, or both, it remains unclear whether direct ocular contact with arsenicals affects the retinal tissue overall. Thus, the retrospective study with survivors of an SM terroristic attack conducted in 2017 reported significant reduction of retinal function in 40 severely intoxicated Iranian veterans (Shoeibi et al., 2017). The max A- and B-wave of the scotopic and photopic ERGs as well as the implicit time of responses were significantly diminished in these individuals overall, supporting earlier research detected central retinal vein occlusion, macular edema, diminished vision acuity, and the overexpression of VEGF-A in the tears of survivors of the SM attack during the Iran–Iraq war (Shoeibi et al., 2017). Altogether, this line of research indicates that the retina could be affected later in exposed individuals, and the damage could be overlooked upon first symptomatic screening; moreover, the signal could be transmitted from the cornea to the retina. Nevertheless, despite the awareness of ocular damage caused by arsenicals, there are no current studies addressing potential retinal problems, as well as the vision of individuals exposed to LEW. Therefore, there is a need to conduct research examining the impact of LEW on the posterior ocular segment.

In contrast to mustard agents, the toxicology of LEW arsenical has been poorly studied in general. Moreover, this blistering agent shares a common downstream molecular pathway, including oxidative stress, inflammatory responses, and VEGF with SM and NM (Goswami et al., 2019; Kadar et al., 2014; Tewari-Singh et al., 2017). On the other hand, the vesicants SM and NM manifest similar structural and functional properties, acting as strong alkylating agents. Notably, neither of these vesicants has previously been carefully tested to explore retinal tissue pathology and the associated molecular mechanisms with the exception of findings reported by Mohan’s group at ARVO meeting describing gliosis in the mice exposed to NM (Umejiego et al., 2022). Such investigations are particularly important because they fill a scientific gap that exists between ocular toxicology and translational pharmacology for the treatment of CWA-related ocular injures.

Arsenic and its derivatives exert toxicity by inhibiting 200 different enzymes involved in multiple cellular pathways (Muzaffar et al., 2022; Shen et al., 2013). Identifying common biological pathways is thus critical in determining the biomarkers of arsenical-related tissue damage. Due to the high toxicity of LEW arsenicals, the opportunity to explore the retinal damage and molecular mechanism of action is rare. The lack of appropriate animal models also causes difficulty in defining the molecular pathogenesis of arsenical vesicants. Previously, we found that the relatively less toxic arsenical phenylarsine oxide (PAO) has a biochemical reaction profile similar to that of LEW and other arsenicals; therefore, PAO can serve as a LEW surrogate to perform safer wet-lab toxicological studies (Srivastava et al., 2016, 2018). Similar to LEW, PAO interacts with thiol groups in cells and inhibits protein-tyrosine phosphatase, resulting in altered signaling pathways (Huang et al., 2017; Shen et al., 2013). For example, our studies on the topical application of PAO on mouse skin demonstrated the involvement of the UPR and inflammatory response in the pathogenesis of skin lesions and acute kidney injury, similar to those in studies of LEW exposure (Srivastava et al., 2016, 2020).

These studies fueled our interest in modeling direct ocular exposure (DOE) in individuals affected by vesicants such as LEW and to investigate whether vesicants induce damage not only to the cornea but also to the retinal tissue. Thus, for the first time, we have demonstrated that DOE to arsenicals triggers retinal damage in exposed mice, and accidental or a battlefield exposure to vesicants damages the posterior chamber of the eye, causing severe retinal injury.

2. Materials and methods

2.1. Cell culture and treatment with PAO

A primary human retinal endothelial cell (HREC) line was purchased from Cell systems (ACBRI 181, Kirkland, WA, USA). Flasks were precoated 30 min before splitting with endothelial attachment factor solution (Cell application, 123–500, San Diego, CA, USA), and cells were grown in basal medium (Lonza, CC-3156, Morristown, NJ, USA) supplemented with growth factors (Lonza, CC-3162 Morristown, NJ, USA). Primary corneal epithelial cells (HCEC) were purchased from ATCC (PCS-700-010, Manassas, VA, USA), along with growth medium (PCS-700-030) supplemented using a growth kit (PCS-700-040). Human corneal keratocytes (HCK) were ordered from Cell Application (632-05a, San Diego, CA, USA) with corneal keratocyte growth medium (6111K-500). Cells were used under 4 to 8 passages and treated when they reached 70–80% confluence. PAO (Sigma-Aldrich, P3075, St.Luis, MO) at a concentration of 200 nM was used for the treatment of cells at different time points (6 h, 12 h, 24 h). DMSO served as a vehicle control. To make a 1 M stock, PAO was dissolved in DMSO under a chemical hood under the conditions required by the Occupational Health Safety Department of the University of Alabama at Birmingham (Srivastava et al., 2016).

2.2. Quantitative real-time PCR analysis (qRT-PCR)

RNA from cells and retinas at different time points was isolated using TRIzol reagent (15596018, Invitrogen, Grand Island, NY, USA). For LEW and PAO exposed tissues we combined 4 individual corneas from two animals, and 2 individual retinas from one animal to obtain a high yield of total RNA. Total RNA (1 μg) was then reversely transcribed into cDNA using Superscript IV Master Mix with ezDNase (11766050, Thermo Fisher Scientific, Waltham, MA, USA). A qRT-PCR was performed using the Quant Studio 3 (Applied Biosystems, Foster City, CA) system with TaqMan primers (Thermo Fisher Scientific, Waltham, MA, USA) (listed in Table A1). The following cycling protocol was used: 95 °C for 20 s, followed by 40 cycles of amplification at 95 °C for 15 s and 60 °C for 60 s. GAPDH was chosen as the housekeeping gene. The ΔΔCt method was applied to calculate the relative expression levels of IL-1B, IL-6, COX2, IFN, and TNFalpha. Each sample was run in two technical repeats.

2.3. Animal studies

In the present study, we investigated the systemic effect of cutaneous PAO exposure on the cornea and the retina using hairless Ptch1^+/−/SKH-1 12-week-old mice. The mice received 150 μg PAO applied over dorsal skin for 6 h. Control group received 100% ethanol (04-355-451, Fisher scientific, Waltham, MA, USA). The animals were euthanized in 72 h. This procedure was described previously (Li et al., 2016). However, DOE was performed using the C57Bl/6J wild type (WT) mice. C57BL/6 mice were purchased from Jackson Laboratory (000664, Bar Harbor, ME, USA). Equal number of both sexes (males and females) were used in this study. All procedures were confirmed by the Institutional Animal Care and Use Committee of the University of Alabama at Birmingham (IACUC UAB) with the IACUC-22105-2-000516358 protocol. The animals were maintained with a standard light and dark cycle (12/12 h) with ambient room temperature and light illumination. All animals were housed in plastic cages with safe environmental pads and standard beddings; water and chow were provided ad libitum. During experimental procedures, all animals were housed in special pre-bedded disposable cages from Innovative (MSX2 set, San Diego, CA, USA) in a quarantine room. Animals were euthanized using CO₂ gas and cervical dislocation without perfusion.

2.4. LEW exposure

The ocular exposure system for LEW was designed using a small vapor cap at MRI Global (Kansas city, MO) facility. Two-month-old C57BL6 males and females were exposed to LEW for the right eye at 2 min with an estimated total dose of 7720 mg* min/m³. Left eyes stayed unexposed. All exposures were conducted within the chemical agent hood line, in compliance with MRIGlobal standard operating procedures. Standard safety procedures and Institutional Animal Care and Use Committee protocols were strictly followed. Animals were sacrificed 24 h post-exposure.

2.5. PAO application to the eye

PAO was dissolved in 100% ethanol by warming at 37 °C for 5–10 min with all required precautions. A half-facepiece respirator equipped with a 3 M 6003 organic vapor/acid gas cartridge (17-986-9B, 2097, Fisher scientific, Waltham, MA, USA) was strictly used during preparation and exposure as personal protective equipment. The animals were anesthetized with a ketamine (74 mg/kg) and xylazine (14 mg/kg) cocktail injected intraperitoneally. To eliminate potential pain, the mice were injected with buprenorphine hydrochloride (NDC12496-0757-5, North Chesterfield, VA, USA) (0.03 mg/kg) approximately 30 min before exposure by subcutaneous (SQ) exposure.

The DOE was performed as follows. Whatman paper eye patches (2 × 2 mm) were applied on the ocular surface of sedated mice for 3 min. The eye patches were soaked in 5 μl solutions containing PAO at different doses (50, 25, and 10 μg/eye). Control eyes received the vehicle only (ethanol, 5 μl). The dose of 50 μg/eye was chosen based on the fact that a mouse ocular surface is about 3 times smaller than the area of topically applied PAO. After PAO exposure, the animals were kept under a continuously operated chemical and biological hood for 3 h for intensive observation and transferred to a quarantine room for daily care. The mice were then euthanized after 24 h and 3 days for retinal RNA isolation and protein isolation, respectively. For corneal and retinal RNA analyses, tissues from four and individual mice, respectively, were combined to generate a biological replicate. For ERG and histological analyses, the animals were analyzed and sacrificed at two and four weeks. The experiments were repeated at least three times.

2.6. Electroretinography (ERG)

Animals were dark adapted overnight. They were then anesthetized using a ketamine/xylazine cocktail, with phenylephrine 2.5% (42702-102-15, Paragon BioTeck, Inc.) drops used topically in the eye to induce mydriasis. Upon completion of dilation, the eyes were coated with Gonak 2.5% (NDC 17478-064-12, AKORN, Lake Forest, IL, USA). The ERG recordings were performed simultaneously on both eyes. A contact electrode was placed on the cornea; reference and ground electrodes were placed SQ on the head and tail areas, respectively. Between animals, the electrodes were sterilized with 70% ethanol. The ERG recording was performed using the UTAS BigShot LKC instrument (Gaithersburg, MD, USA). After the ERG measurements, the animals were euthanized, and the eyes were enucleated for histological examination. The mice were exposed to five flashes of 2.5 cd*s/m², 7.91 cd*s/ m², 25 cd*s/m², and 79.1 cd*s/m² light intensities at intervals of 45 s. The LKC EM software was used to analyze the a- and b-waveforms.

2.7. Draize irritation test

The ocular toxicity test was conducted as described previously (Wilhelmus, 2001). The irritation assessment of the control and PAO-treated eyes was performed at 1, 2, 3, 6 14, and 28 days. Positive responses from the corneal tissues were scored based on the scale from 1 to 4 including A) Opacity (details of iris clearly visible-1; easily discernible translucent areas-2;, size of pupil barely discernible-3; iris invisible-4); B) Involved area of cornea (one quarter of whole area-1; less than half – 2; more than half-3; whole area-4); C) Conjunctiva (redness of palperal conjunctiva: 1–3; chemosis: 1–3; discharge: 1–3). All criteria were scored to generate a maximum total score and a mean. We plotted the mean of the total score in GraphPad to build a curve.

2.8. Histology

Hematoxylin and eosin (HE) staining was performed by fixing the enucleated eyeballs in 4% paraformaldehyde (PFA) as described above (26754-1A, 26762–01, Electron Microscopy Sciences, Hatfield, PA, USA) (Starr and Gorbatyuk, 2019). The cryopreserved eyes were sectioned to obtain 12-μm corneal and retinal sections using a Leica CM1510S cryostat (Leica, Buffalo Grove, IL, USA). The peripheral and central corneal thicknesses were measured in PAO treated eyes.

2.9. TUNEL assay

Terminal deoxynucleotidyl transferase dUTP nick end labeling assay (TUNEL) (C10617, Thermo Fisher Scientific, Waltham, MA, USA) was performed according to the manufacturer’s instructions. The retinal sections were analyzed under 20X magnification using a 200-μm step distance between points starting at the optic nerve head. DAPI (0100–20, Southern biotech, Birmingham, AL, USA) containing a mounting medium was used to cover the slides. TUNEL-positive nuclei were manually counted by a masked researcher. TUNEL-positive retinal ganglion, photoreceptor (outer nuclear layer, ONL) cells and cells of the inner nuclear layer (INL) were detected in six visual fields (200 × 200 μm) of the superior and the inferior retina.

2.10. Statistical analysis

All statistical data were calculated using GraphPad Prism 9 software. Two groups were compared with the Student t-test while multiple groups were analyzed using one-way ANOVA Tukey’s test with multiple comparisons. All data are presented as mean ± standard error of mean (SEM). P values < 0.05 are considered significant.

3. Results

We conducted a pilot study in which we tested whether DOE to LEW could be harmful to retinal tissue. To mimic the exposure of individuals to LEW, we adopted a vapor eye cap method previously validated in rabbits (Tewari-Singh et al., 2016). Given that the mouse eye is significantly smaller that the rabbit eye, we exposed mice to LEW for 2 min to obtain a total concentration of 7, 720 mg* min/m³ LEW placing a vapor cap on the right eye and examined experimental and control eyes 24 h later. We evaluated retinal damage by histological and qRT-PCR analyses.

3.1. DOE to LEW results in severe retinal damages

The analysis of the retinal tissues of the mice 24 h after LEW exposure demonstrated that the expression of cytokines Il-6 and Cox2 was increased over 700- and 5-fold, respectively (Fig. 1A). These dramatic increases were accompanied by a rise in TUNEL-positive cells in the exposed retinas—the numbers of the dying retinal ganglion cells (RGC), cells located in the inner nuclear layer (INL), and the photoreceptor cells (outer nuclear layer, ONL) were dramatically elevated compared to the untreated control retinas.

Fig. 1.

Direct ocular exposure to LEW results in an increase of cytokine expression in corneal and retinal mouse tissues 24 h post-exposure. A: Il-6, Ifnγ, Cox2 and Il 1β cytokines expression measured as mRNA fold change was significantly upregulated in the retina, (n = 3–4, *p < 0.05, **p < 0.009). B: The increase in cytokine expression was accompanied by a widespread distribution of TUNEL-positive staining. Average of 3 superior and 3 inferior areas of each control and experimental eyes were used to calculate numbers of TUNEL-positive cells (in green). Nuclei were stained with DAPI. The blue color was substituted by red for better visualization. The RGC, INL cells, and photoreceptor cells (ONL) of the exposed retinas manifested DNA damage, while no staining was observed in the control eyes (n = 3, ****p < 0.0001). Scale bar is 50 μm.

The retinal injury resulted from DOE to LEW in experimental animals has not been previously investigated. It is also unknown whether systemic exposures to arsenicals trigger the ocular pathogenesis similar to topically applied PAO arsenical in a murine model leading to significant acute kidney injury (Srivastava et al., 2020). Therefore, we became interested in whether cutaneous exposure of hairless Ptch1^+/−/SKH-1 mice to PAO arsenical initiates ocular tissue damage. The Ptch1^+/−/SKH-1 mice were treated with 150 μg PAO for 6 h. Both corneal and retinal tissues of mice with cutaneous PAO exposure were collected. Interestingly, both ocular tissues showed altered cytokine gene expression. Decline and incline in Il-6 and Cox2 expression, respectively, in both types of ocular tissue (*p < 0.05, Fig. A1) were observed in the exposed animals. Therefore, these data imply that similar to PAO cutaneous exposure, either systemic or local LEW exposure could be harmful to both corneal and retinal tissues and serves as a rationale for further investigation of the phenotypic, histological, and molecular consequences of arsenical-induced eye injury.

3.2. Treatment of human corneal and retinal endothelial cells with PAO arsenical results in upregulation of cytokine gene expression and activation of unfolded protein response

This study was conducted using primary corneal epithelial (HCEC) and keratocyte (HCK) cells, as well as human retinal endothelial cells (HREC) (Fig. 2A). The cells were seeded to reach 80% confluency. We then conducted the treatment of cells with 200 nm of PAO. The effect of the treatment was tested at 6, 12, and 24 h by qRT-PCR analysis of proinflammatory cytokine expression.

Fig. 2.

Treatment of corneal and retinal endothelial cells results in the upregulation of cytokine expression and UPR activation. A: Treatment of primary human corneal (HCK and HCEC) and retinal (HREC) cells with 200 nM PAO results in the upregulation of cytokine expression (n = 3–6) measured as mRNA fold change. B and C: Simultaneous with the increase in cytokine gene expression, the activation of the PERK UPR signaling pathway was detected 6 h after the treatment. B: Increases in GADD34, and ATF4 were detected in HCK cell line (n = 3). Immunoblot images are shown on the right. C: Increases in GADD34, ATF4, and p-eIF2α were registered for the treated HREC cell line (n = 4). Immunoblot images are shown on the right. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001.

The experimental data showed that after the first 6 h, all cell types manifested an increase in Cox2 and Ifna2 cytokine expression in response to the treatment. However, the magnitude of their responses was different. For example, while HREC demonstrated a six-fold increase in Ifna expression compared to the untreated control (p < 0.05), HCEC manifested more than a 25-fold increase, and HCK manifested up to a 340-fold change in this cytokine (p < 0.05, p < 0.0001, and p < 0.0001, respectively). In addition, HCEC manifested an increase in the Il1-b cytokine at this time point (p < 0.05). At 12 h post-treatment, HCEC manifested insignificant decline in cytokine expression compared to its own control. However, HCK and HREC showed a continuous increase in Ifna2 expression (for both p < 0.0001). In addition, HREC responded to treatment by the continuous increase in Cox2 cytokine expression (p < 0.0001). At 24 h post-treatment, the HCK still manifested an increase in Ifna2 expression (p < 0.0001), while HCEC and HREC demonstrated no difference in cytokine gene expression compared to controls. Altogether, this data indicates that inflammatory signaling is activated in all types of primary cells in response to treatment with PAO. However, HCK was the cell type with the largest and most sustained upregulation of Ifna2 expression within this experimental setting.

Our previous study identified UPR as a molecular signal contributing to PAO-induced skin cell pathogenesis. Moreover, our published work has indicated that cytokine expression is triggered by the UPR markers ATF4 and CHOP, as well as the PERK arm mediators (Rana et al., 2014). Based on the increase in cytokine expression in PAO-treated cells, we analyzed UPR activation (Fig. 2B). The study results demonstrated that ATF4 and GADD34 mediators of UPR were already upregulated by 6 h post-treatment (p < 0.0001 for HCK and p < 0.001 for HREC) and continued upregulation up to 24 h post-treatment. Overall, these findings indicate that UPR activation in treated cells is concomitant with the upregulation of cytokine gene expression. In this experiment, we learned that as PAO evolves, significant cytokine increases and UPR activation in corneal and retinal endothelial cells occurs. Therefore, we then initiated an in-vivo study to investigate whether DOE to PAO arsenical in mice is harmful to ocular tissues.

3.3. Direct ocular PAO exposure upregulates cytokine gene expression and activates UPR in ocular tissues

To model the DOE of individuals to arsenicals, we adopted an eyepatch method. The corneal surface was in direct contact with an eye patch for 3 min. The eye patch was soaked either in the vehicle (ethanol) or in PAO. Different doses of PAO (50, 25, and 10 μg per eye) were used in the study to verify whether the ocular effects from the exposure were dose-dependent. The dose of 50 μg/eye was chosen based on the fact that the mouse ocular surface is about 3 times smaller than the area of topically applied PAO. The results of the DOE of mice to 50 μg PAO are shown in Fig. 3. The data demonstrated that treatment with PAO caused significant elevation of Il-1β, Il-6, and Cox2 gene expression in both corneal and retinal tissues 24 h post-treatment (Fig. 3A). Interestingly, the retinal tissue of the exposed mice manifested Il-1β elevation (about 2500-fold) to a larger degree compared to corneal tissues. Previously, we have shown that IL-β expression is governed by the UPR mediator ATF4 (Rana et al., 2014). The significance of these findings was emphasized in the current study by the observation of the dramatic upregulation of UPR PERK mediators, ATF4 and GADD34 in the retinas of PAO-treated eyes (Fig. 3B and C). Collectively, these results suggest the “transmission” of pro-death molecular stimuli from the anterior to posterior chambers of the eye.

Fig. 3.

Direct ocular exposure of mice to 50 μg PAO arsenical causes severe corneal and retinal injury. A: Both corneal and retinal tissues of the exposed mice manifested cytokine overexpression 24 h later measured as mRNA fold change (n = 3–4). B: The cytokine overexpression was concomitant with the activation of PERK UPR mediators, ATF4, and GADD34 (n = 4). C: Images of western blots treated with antibodies against ATF4, GADD34, and actin. D: The corneal tissue of mice exposed to PAO manifested significant increases in the thickness 72 h later (n = 3). Images of the H&E-stained cornea are shown at the bottom. *p < 0.05, **p < 0.01, ***p < 0.001, and *****p < 0.0001. Scale bar is 50 μm.

To determine whether PAO affects corneal tissues similarly to LEW (Tewari-Singh et al., 2016), we then analyzed corneal thickness. The results suggested that similar to rabbit’s eyes treated with LEW arsenical (Tewari-Singh et al., 2017), the PAO-treated mice demonstrated a significant—over three-fold—increase in corneal thickness at three days post-treatment (p < 0.01, Fig. 3D).

3.4. Direct ocular exposure to arsenicals results in loss of retinal function and integrity, culminating in apoptotic cell death

Exposure to 50 μg PAO resulted in the dramatic upregulation of cytokine expression and UPR activation, as well as rapid apoptotic cell death suggesting severe damage to retinal tissues. Given that survivors of CWA exposure during the Iran–Iraq War developed reductions in ERG responses and vision acuity 40 years later (Amini et al., 2020;Ghazanfari et al., 2009), we became interested in PAO dose titration to model progressive retinal function changes. Thus, we then tested DOE to PAO at a dose of 25 μg for 3 min in mice (Fig. B1). We dark-adapted two-week post-treated animals and registered the a- and b-waves of the scotopic ERG amplitudes of treated and untreated eyes. We learned that PAO-treated eyes in mice manifested a blindness. Thus, the data demonstrated that the ERG traces of PAO-treated eyes were already flat by two weeks. Therefore, because of our original intention to generate a mouse model providing a window of opportunity to test countermeasures against vesicant arsenicals, we then decided to validate the effects of lower PAO dose (10 μg) applied on the corneal surface of mice. The results of the study with mice treated with 10 μg PAO are shown in Fig. 4.

Fig. 4.

Direct ocular exposure of mice to 10 μg PAO arsenical causes severe corneal and retinal injury. A: Both corneal and retinal tissues of the exposed mice demonstrated cytokine overexpression 24 h later measured as mRNA fold change (n = 3–6). B and C: Treatment with PAO results in the loss of retinal function. The ERG traces of treated and control eyes at two weeks post-treatment are shown (B, n = 12). The a- and b-wave amplitudes were significantly diminished in PAO-treated eyes during the four weeks after exposure (C n = 6). D: TUNEL-positive cells in the retina of PAO-treated mice are shown in green (left). The RGC, cells of the INL, and random photoreceptor cells (white arrows) demonstrated TUNEL-positive staining. The average of six superior and inferior fields of each control and experimental eyes was used to calculate the number of TUNEL-positive cells (in green) at two weeks post-treatment (n = 3). Nuclei were stained with DAPI. The blue color was substituted by red for better visualization. E: The scoring of the PAO-induced corneal injury using the Draize test is shown. Major changes in the cornea and conjunctiva occurred during the first two weeks after exposure. Then, the degree of the damage remained the same. F: Thus, the corneal thickness increased more than five-fold in the PAO-treated eyes at two week after exposure (n = 5). *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. Scale bar is 50 μm

Similar to the mice with DOE to 50 μg PAO, treatment with 10 μg PAO resulted in the dramatic upregulation of cytokine gene expression Cox2, Il-1β, Il-6, and Tnfα cytokines in both ocular tissues in 24 h (Fig. 4A). However, compared to the 50-μg PAO treatment, the magnitude of overexpression in the retinal tissues of animals treated with 10 μg PAO differed. For example, Il-1β expression increased about 150-fold compared to the control eyes (p < 0.05, Fig. 4A), while 50 μg caused about a 2500-fold increase. Another example is Il-6 expression, which was elevated about 450-fold in the retina treated with 10 μg PAO; however, in the eyes treated with 50 μg PAO, the Il-6 expression was lower an approximately 120-fold increase. The corneal tissues, the first line of response to PAO treatment, demonstrated marked upregulation of both cytokines, which was overall compatible with the results obtained in mice treated with 50 μg PAO. The Il-6 mRNA was an exclusion; over 600-fold increase in the Il-6 cytokine expression was observed in corneal tissue treated with 10 μg vs 1200-fold incline detected in 50 μg treated cornea. We then became puzzled whether the treatment with 10 μg PAO results in a slower retinal degeneration. Not only did the retinas manifest a severe inflammatory response in the eyes treated with PAO, but the decline of the a- and b-wave scotopic ERG amplitudes was also detected in the eyes treated with 10 μg PAO at two and four weeks posttreatment. More than a three-fold decrease in both a- and b-wave amplitudes compared to the control eyes treated with the vehicle was registered in these mice (p < 0.001 and p < 0.05 for a-wave at two and four weeks, respectively, and p < 0.01 and p < 0.05 for b-wave at two and four weeks, respectively). The consistency of the decreases in the ERG recordings of the PAO-treated eyes indicates that the mice experienced a loss of retinal function in their right eyes (Fig. 4B and C).

This fact became obvious after the analysis of retinal cell death. The retinal function loss of the PAO- treated eyes agreed with the number of TUNEL-positive cells detected in the vision field (200 × 200 μm) of the retinal section at two weeks post-treatment. Markable cell death was observed in the RGC and INL layers (Fig. 4D). Randomly, TUNEL-positive cells were found in the ONL (white arrows). Compared to treatment with 10 μg PAO, DOE to 50 μg PAO manifested the widespread distribution of TUNEL-positive cells, including photoreceptors (Fig. 1). This data suggests that the damage to retinal cells from vesicant exposure occurs in a dose-dependent manner. Together with the ERG results, this data indicated the PAO-induced retinal degeneration associated with the deterioration of retinal function and cell loss.

We then assessed the corneal and conjunctival tissues using the Draize test and measured the central corneal thickness four weeks (28 days) after treatment. The total score comprised of the criteria such as corneal opacity and area involved in the damage and conjunctival assessment (redness) as described (Wilhelmus, 2001). Results of the eye irritancy test are presented as a mean of the total score, as shown in Fig. 4E. The evaluation of the corneal damage resulting from PAO exposure indicated that the severity of ocular lesions was strongest up to 14 days after initial exposure. Then, the mean of the total score dropped and was sustained from 14 to 28 days, indicating that major corneal damage occurs during the first two weeks post-exposure. In agreement with this data, we found that the corneal thickness of PAO-exposed eyes was markedly greater compared to the control eyes (Fig. 4F). Both the Draize test and the measurement of corneal thickness showed the severity of corneal damage in response to PAO exposure.

Overall, these experiments demonstrated that similar to exposure to LEW, DOE to PAO damages both corneal and retinal tissues, leading to apoptotic cell death.

4. Discussion

The eye is particularly vulnerable to exposure to blistering vesicants. Both systemic and direct ocular contact with CWAs can be harmful to the eye and the circulating and penetrating through ocular surface agents reach the retinal tissue. While our published study demonstrated that topical skin exposure of mice to arsenicals triggers systemic inflammatory responses and the activation of UPR in the skin, kidneys, and lungs (Li et al., 2016; Manzoor et al., 2020; Srivastava et al., 2020, 2021), no study has previously reported the hazardous effects of such exposure to these agents on the retina. Thus, for the first time, we have reported that both systemic and direct ocular contact with arsenicals could activate molecular signaling in retinal tissue and thus magnify the effects of such exposure on the eye. Similar to the kidneys and lungs, systemic PAO circulation from cutaneous airway exposure alters cytokine expression in both ocular tissues emphasizing the indirect vesicant effects on the eye. These findings are of great importance because the ocular damage could be overlooked upon first symptomatic screening. The results of this experiment also imply that the retinal injury in the experimental mice and accidently exposed individuals is secondary to the primary corneal damage or systemic injury; even though personal protective equipment, goggles, or face shields are properly used at the place of exposure, blistering agents, easily penetrating a cloth, may circulate and be delivered to the retina. This is particularly true if time permits the absorption or inhalation of vesicants. In order to assess the vesicant-induced ocular phenotype resulting from local exposure, we validated a mouse model of arsenical-induced retinal injury.

The previously conducted research on arsenical-exposed rabbits did not emphasize the vesicant-induced retinal damage although the vision acuity decline and the diminished ERG amplitudes have been reported in exposed individuals (Shoeibi et al., 2017). Moreover, the lack of a proper animal model allowing for the study of ocular pathogenesis of arsenical exposure may significantly delay the invention of medical countermeasures for the survivals. Therefore, the current research is intended to close a gap in the evaluation of the aftermath CWA damage, overall modeling the consequences of direct ocular exposure to arsenicals in affected individuals.

The most striking discovery of the current research is that the DOE to arsenicals in mice causes severe retinal damage in addition to corneal injury. Thus, the ocular phenotype and progressive retinal dystrophy associated with loss of retinal function and cell death following the DOE to the vesicant are the key findings of the current research. Both the LEW and PAO exposure showed severe retinal damage by activating proinflammatory cytokine cascade and launching the apoptotic cell death program. Progressive retinal functional loss was observed in the mice with DOE to PAO. It is clear that the cornea absorbs arsenicals that activate a signal traveling to the back of the eye and the retinal dystrophy is a secondary injury to the coneal tissue damage. Therefore, the retina become vulnerable to the vesicant exposure and the severity of the damage is dose and time dependent.

The subsequent description of retinal injury can be classified as mild, moderate and severe based on the time and dose of exposure. For example, the treatment of mice with 25 μg PAO results in the flat ERG amplitudes, while exposure to 10 μg leads to significantly reduced ERG recordings persisting over 4 weeks. These data imply that higher arsenical dose induce severe retinal damage associated with blindness, while lower doses trigger progressive retinal function and cells loss. Most likely that a dose of 25 μg/eye triggers decline of ERG amplitudes earlier resulting in blindness in mice by 2 week post treatment. Another example is the inflammatory response that is a molecular phenotype of the LEW and PAO-induced ocular tissue damage. While corneal tissues respond to the treatment with 50 μg PAO by dramatic elevation of IL-6, cornea of the mice exposed to 10 μg manifest lover expression of this cytokine. In support, the Il-1β expression was dramatically higher in the retina of mice treated with 50 μg PAO vs ones exposed to 10 μg. In contrast, the Il-6 expression in mice was significantly elevated at lower dose as compared one found with higher dose. The explanation for this phenomenon could be a different role (neuroprotective vs cytotoxicity) assigned for IL-6 cytokine in neuronal and corneal (epithelial, keratocytes, and endothelial) cells. Finally, mice with DOE to 50 μg PAO manifest markedly higher number of apoptotic RGC, photoreceptors, and the cells of the INL layer as compared to lower dose. Therefore, we propose a dose of 10 μg PAO for DOE in mice to model retinal damage following the exposure to arsenicals.

We previously showed that retinal cell death can occur through sustained UPR activation leading to apoptotic cell death (Bhootada et al., 2016). The involvement of the activated UPR in the pathobiology of mice with cutaneous exposure and treated keratinocytes was demonstrated by means of the application of a chemical chaperone, 4-phenylbutyruc acid (4-PBA) (Srivastava et al., 2016). Thus, we demonstrated that UPR proteins (GRP78, p-PERK, ATF4, and CHOP) were controlled by the application of 4-PBA and that CHOP plays a key role in promoting UPR-mediated caspase-3 activation. The limitation of the current study highlighting no mechanistic role of the UPR in retinal pathobiology should be overcome and addressed in future experiments. The current mouse model of DOE to arsenicals provides an excellent opportunity not only to investigate the molecular driver of retinal pathobiology but also to test new medical countermeasures against retinal dystrophy. Our findings of activated UPR in the retina of exposed eyes support this hypothesis and indicates the necessity to identify a molecular signal transmitted from the cornea to the retina leading to UPR activation. The question of whether the activated UPR is a molecular driver of arsenical-induced retinal pathobiology will be addressed in the future using the proposed model of direct ocular PAO exposure.

Altogether, our findings indicate that a mechanistic study should be conducted to examine the toxic manifestations of arsenicals on the retina and the involved molecular mechanism, although activation of proinflammatory cascade and UPR signaling are most likely associated with the mechanism of retinal pathogenesis. Therefore, the proposed surrogate mouse model of the arsenical-induced retinal injury arises the limitation in the field of current ocular toxicology for a study of the molecular mechanism of retinal pathogenesis and the design of new medical countermeasures to treat retinal injury. Our data also suggest that direct ocular contact with arsenicals either in a battlefield or at the site of accidental exposure may result in severe retinal damage manifesting as vision loss, retinal cell death accompanied by activated inflammatory response and UPR. The retinal damage is both progressive and dose dependent. The establishment of a surrogate mouse model depicting PAO-induced ocular damage permits further investigation the molecular pathological drivers of ocular toxicity.

Data availability

Data will be made available on request.