Safety Evaluation of Far-UV-C Irradiation to Epithelial Basal Cells in the Corneal Limbus

Basal cells in the corneal limbus play an important role in the turnover cycle because they are the source of all cells that constitute the corneal epithelium. We examined the penetration depth of ultraviolet (UV) light in the corneal limbus and assessed the safety of Far-UV-C on stem cells in the basal area of the corneal limbus. Rats were irradiated with UV at peaks of 207, 222, 235, 254 and 311 nm while under anesthesia. The UV penetration depth in the rat corneal limbal epithelium was wavelength dependent: 311 nm UV-B and 254 nm UV-C reached the basal cells of the epithelium, and 235 nm radiation reached the middle area; however, 207 and 222 nm UV-C reached only the super ﬁ cial layer of the epithelium. Porcine cornea, which is similar to the human eye in size and structure, were irradiated with 222 and 254 nm UV-C. As in rats, 222 nm UV-C reached only the super ﬁ cial layer of the porcine corneal limbal epithelium. These results indicate that Far-UV-C, such as radiation of wavelengths of 207 and 222 nm, could not reach corneal epithelial stem cells, i.e. the cells remained intact. It is unlikely that the turnover of the corneal epithelium is obstructed or disrupted by exposure to Far-UV-C.


INTRODUCTION
Ultraviolet (UV) radiation is divided into UV-A (315-400 nm), UV-B (280-315 nm) and UV-C (100-280 nm) in descending order of wavelength. UV radiation, regardless of the wavelength, has been considered harmful to the human body. Chronic UV exposure not only causes premature skin aging, known as photoaging, but also causes skin cancers, such as cutaneous squamous cell carcinoma (cSCC), basal cell carcinoma (BCC) and cutaneous melanoma (1). In the eye, acute UV exposure causes acute keratitis (2), while chronic UV exposure causes pterygium (3), conjunctival tumors (4) and cataracts (5). However, it has recently become clear that short-wavelength UV, called Far-UV-C, has germicidal effects on viruses and bacteria (6), but is much less hazardous to higher organisms, including humans, than previously thought (7)(8)(9)(10).
In the previous studies, we have shown that the UV penetration depth to the rat corneal epithelium is wavelength dependent and Far-UV-C, such as radiation of wavelengths 207 and 222 nm, penetrate only the superficial layers of the corneal epithelium, which is sloughed off within approximately 24 h by the physiological turnover cycle (8). This extremely low penetration into the cornea (8,11) and the rapid turnover cycle of the corneal epithelium may be the primary reasons for the substantially less hazardous Far-UV-C properties. Therefore, to prevent corneal epithelial damage caused by Far-UV-C irradiation, it is necessary for the surface layer cells to slough off naturally, which requires normal turnover.
Corneal epithelial turnover is conceptualized by the XYZ theory proposed by Thoft (12). For corneal epithelial homeostasis to be maintained, an equilibrium must be established between the proliferation and differentiation of basal cells (X), the supply of new basal cells from stem cells in the corneal limbus to the central area (Y) and the shedding of surface cells (Z), expressed as "X + Y = Z" (4,11). For normal turnover to proceed, it is particularly important that the corneal epithelial stem cells, which are located in the area between the cornea and conjunctiva, known as the limbus (4,13), are not injured because all the cells in the corneal epithelium differentiate from basal cells. In this study, we examined the UV penetration depth in the corneal limbus and examined the safety of Far-UV-C on stem cells in the basal area of the corneal limbus. the tissue. The porcine eyes were obtained from slaughterhouses on the day of experimentation and transported to the laboratory on ice to prevent tissue degeneration. The arrived eyeballs were rinsed several times with saline, and excess tissue such as eyelids, fat and muscle were removed with scissors. Since the corneal limbus is located at the transition from the cornea to the conjunctiva, a small cut was made with a scalpel in the conjunctiva on the optic nerve side 1 cm from the corneal edge, and then the cornea was cut with scissors to include the same width of conjunctiva all the way around. Dissected porcine corneas containing the corneal limbus were immersed in cold saline until irradiation. Since the eyes were obtained after enucleation, the strain, sex, age, location (left or right eye) and vertical eye orientation were unknown.
Estimation of light exposure and UV exposure duration. Estimation of light exposure and UV radiation exposure was performed as previously reported (7,8). The irradiation of 207, 222 and 235 nm UV-C were performed using a light source combined with a bandpass filter, which eliminated UV light with irrelevant wavelengths. Information on irradiation equipment and exposure conditions is summarized in Table 1.
Rats. Anesthesia was induced by intramuscular injection of a ketamine (120 mg kg BW À1 ) and xylazine (6 mg kg BW À1 ) mixture, and the pupils were dilated with 0.5% tropicamide and 0.5% phenylephrine hydrochloride eye drops (Santen Pharmaceuticals Co., Ltd., Osaka, Japan). The rats were then exposed to UV radiation with peaks at 207, 222, 235, 254 and 311 nm. Radiant exposure for evaluating the invasion depth to the corneal limbus was 600 mJ cm À2 and eyes were enucleated immediately after irradiation (n = 4 in each group). To evaluate the residual DNA damage after 24 h, rats were irradiated with each wavelength at three different doses (20-10 000 mJ cm À2 , n = 6 in each group). To evaluate the corneal structure and residual DNA damage 7 days after corneal epithelial renewal by turnover, rats were irradiated with 222 nm UV-C or UV with main peak at 254 nm emitted by a mercury lamp at 600 mJ cm À2 or 150 mJ cm À2 of energy, respectively (n = 6 in each group).
Porcine. The dissected porcine cornea (n = 3) stored in cold saline was placed on a culture dish and irradiated with 222 nm UVC or UV with main peak at 254 nm emitted by a mercury lamp with an energy dose of 600 mJ cm À2 . Irradiation durations were 120 s for 222 nm UVC and 546 s for UV with main peak at 254 nm emitted by a mercury lamp.
Rats. Immediately after irradiation, the rats were euthanized by anesthetic overdose followed by cervical dislocation, and both eyes were enucleated for evaluating the invasion depth of the corneal limbus. For evaluating residual corneal limbus DNA damage, eyes were enucleated 24 h after irradiation. For evaluating the corneal structure and residual DNA damage after corneal epithelial renewal by turnover, eyes were enucleated 7 days after irradiation. Enucleated eyes were fixed in 4% paraformaldehyde containing 20% isopropanol, 2% trichloroacetic acid and 2% zinc chloride for 24 h at room temperature. After alcohol dehydration, tissues were embedded in paraffin, and 4 lm thick sections were prepared. The rat eyes were cut vertically to include the optic disc.
Porcine. Immediately after irradiation, porcine corneas were fixed and dehydrated as in rats. The porcine corneas embedded in paraffin were cut into 4 lm-thick slices to include the central portion.
DNA damage and penetration depth of UV to the corneal limbus. We used cyclobutane pyrimidine dimer (CPD) localization in the corneal limbus for the DNA damage index and UV radiation penetration depth, as previously reported (8), because CPDs are DNA mutations resulting from UV radiation absorption. Endogenous peroxidase activity in deparaffinized sections was inactivated with 3% hydrogen peroxide for 10 min. A VECTASTAINâ ABC mouse IgG kit (Vector Laboratories Inc., Burlingame, CA) was used for immunostaining. After blocking with normal horse serum for 30 min at room temperature, the sections were incubated with an anti-CPD (Cosmo Bio Co., Ltd., Tokyo, Japan) (1:400) antibody diluted with antibody diluent (Dako North America, Inc., Carpinteria, CA) for 1 h at 37°C, and then with the anti-mouse IgG biotinylated antibody for 30 min at 37°C. To control for nonspecific antibody-independent signals, the sections were incubated with antibody diluent without a primary antibody for 1 h at 37°C. After washing, the sections were incubated with ABC complex for 30 min at 37°C. Signals were developed with 3 0 ,3 0 -diaminobenzidine (ImmPACT TM DAB Peroxidase Substrate Kit; Vector Laboratories Inc.) in chromogen solution. Table 1.

UV penetration depth in the rat central cornea and limbus
For rat eyes, the localization of CPDs in the central cornea, superior and inferior limbus was observed (Fig. 1a). The thickness of the limbal epithelium was almost the same as that of the corneal epithelium (Fig. 1b,c). UV penetration depth in the corneal limbus and central cornea was observed immediately after irradiation. NoCPDs were observed in unexposed corneas (Fig. 2a), but CPD-positive cells were observed in the most superficial layers of the epithelium in corneas exposed to 207 and 222 nm UV-C (Fig. 2b,c). In central corneas exposed to 235 nm UV-C, CPD-positive cells were observed up to the middle layer of the corneal epithelium (Fig. 2d). In corneas exposed to UV with main peak at 254 nm emitted by a mercury lamp and to 311 nm UV-B, positive cells were observed throughout the epithelium (Fig. 2e,f).
In the limbus, the UV penetration depth at different wavelengths was typically similar to that in the central cornea. In the unexposed eye, a weak reaction was observed throughout the superior limbus, but no positive reaction was observed in the inferior limbus (Fig. 2g,m). In the eyes exposed to 207 nm UV-C, not only was a weak reaction observed throughout the superior limbus but also positive cells were observed in the superficial layers (Fig. 2h). However, no positive reactions were observed in the inferior limbus (Fig. 2n), as in the unexposed eyes. In eyes exposed to 222 nm UV-C, CPD-positive cells were observed in the superficial layer in both the superior and inferior limbus (Fig. 2i,o). In eyes exposed to 235 nm UV-C, strongly positive cells were observed up to the middle layer of the superior limbus, but the staining intensity of the basal cells was comparable to that of the unexposed eyes (Fig. 2j). The entire inferior limbus was weakly stained, but strongly positive cells were observed only in the superficial region (Fig. 2p). In eyes exposed to the 254 nm UV-C and 311-nm UV-B, the entire superior and inferior limbus as well as the cornea were strongly stained, and the basal cells were also strongly positive (Fig. 2k,l, q,r). Regardless of the wavelength, the same or higher amounts of CPDs were observed in the superior limbus as in the inferior limbus ( Fig. 2g-r).

Evaluation of remaining DNA damage in the superior limbus of rats
The localization of CPDs in the upper corneal limbus was observed 24 h after irradiation, and the irradiation energy at which UV damage could not be removed or repaired in 24 h was evaluated. In the superior limbus irradiated with 207 nm UV-C, no CPD-positive reactions were observed in basal cells after UV irradiation with energy doses of 1500 mJ cm À2 (Fig. 3b). Weak CPD-positive cells were observed in the limbus irradiated with a dose of 2500 mJ cm À2 but were comparable to those found in the unirradiated limbus (Fig. 3a,h). At an irradiation dose of 10 000 mJ cm À2 , strong CPD-positive cells were observed in the entire limbus, including the basal cells (Fig. 3m). In the corneal limbus irradiated with 222 nm UV-C, slightly stronger CPD-positive cells were observed near the superficial layer at an irradiation dose of 1500 mJ cm À2 , and moderately positive CPD cells were observed in cells throughout the entire limbus at an irradiation dose of 2500 mJ cm À2 (Fig. 3c,i). At an irradiation dose of 5000 mJ cm À2 , strong CPD-positive reactions were observed throughout the limbus, including the basal cells (Fig. 3n), as well as in the limbus irradiated with 207 nm UV-C at 10 000 mJ cm À2 . In the corneal limbus irradiated with 235 nm UV-C, cells showing weak CPD-positive reactions were observed up to the middle layer even when irradiated with only 30 mJ cm À2 of radiant energy (Fig. 3d), and cells in the superficial layer became even more strongly stained when exposed to 300 mJ cm À2 of radiant energy (Fig. 3j). At a radiant energy dose of 600 mJ cm À2 , even the cells in the middle layer were a strongly CPD-positive, and cells in the basal layer were also weakly stained (Fig. 3o). The staining pattern of the corneal limbus irradiated with 20 mJ cm À2 of UV-C emitted by a mercury lamp with main peak emission at 254 nm was similar to that of the corneal limbus irradiated with 600 mJ cm À2 of 235 nm UV-C, with intermediate layer of cells showing a strong CPDpositive reaction and basal cells also weakly stained (Fig. 3e,o). In the corneal limbus irradiated with 100 mJ cm À2 of 254 nm UV-C, all cells, including the basal area, were CPD-positive (Fig. 3k), and stronger CPD-positive responses were observed when irradiated with 300 mJ cm À2 (Fig. 3p). No CPD-positive cells were observed in the corneal limbus irradiated with 30 mJ cm À2 of 311 nm UV-B (Fig. 3f). After irradiation with an energy dose of 150 mJ cm À2 (Fig. 3l), CPD-positive cells were observed from the middle layer to the basal area, and at 600 mJ cm À2 , CPD-positive cells were observed in the entire limbus (Fig. 3g).

Effects of 222 and 254 nm UV-C irradiation on the corneal epithelial turnover cycle
Seven days after irradiation with 222 nm UV-C at an energy dose of 600 mJ cm À2 , the central corneal and corneal limbus epithelium showed no tissue abnormality (Fig. 4a,e,i,m) and no CPD-positive cells were observed (Fig. 4b,f,j,n) in eyes (n = 6). The central corneal epithelium and corneal limbus epithelium showed no tissue abnormalities 7 days after irradiation with UV with main peak at 254 nm emitted by a mercury lamp with an energy dose of 150 mJ cm À2 (Fig. 5c,g,k,o; n = 6). However, in 3 out of the 6 eyes, the basal cells of the central corneal epithelium were CPD-positive (Fig. 4d,h). No CPD-positive cells were observed in the corneal limbus (Fig. 4l,p; n = 6).
Tissue structure of the porcine corneal limbus and the penetration depth of 222 and 254 nm UV-C In this study we used porcine corneas that had been removed and stored at 4°C until irradiation, and no tissue or cellular degeneration was observed in the HE-stained images (Fig. 5a-c). The porcine corneal limbus was thicker than the central corneal epithelium (Fig. 5a-c), and the basal portion, which is the deepest part, had an orderly arrangement of basal cells (Fig. 5g,h).
No CPD-positive cells were observed in unirradiated corneal limbus (Fig. 5d,j). CPD-positive cells were observed in the most superficial layer of the corneal limbus irradiated with 222 nm UV-C, but not in other areas (Fig. 5e,h). In the corneal limbus irradiated with UV radiant energy emitted by a mercury lamp with main peak emission wavelength of 254 nm, CPD-positive cells were observed in the superficial layers and 50-100 lm deep, but not in the basal areas (Fig. 5f,i).

DISCUSSION
The present study showed that the UV penetration depth in the rat corneal limbal epithelium was wavelength dependent, as was the central corneal epithelium: 311 nm UV-B and presumably UV radiant energy emitted by a mercury lamp with the main peak emission wavelength of 254 nm reached the basal cells of the epithelium, and 235 nm radiation reached the middle area; however, 207 and 222 nm UV-C reached only the superficial layer of the epithelium. Similarly, 222 nm UV-C only reached the superficial layer of the pig corneal limbal epithelium. These results indicate that Far-UV-C, such as wavelengths of 207 and 222 nm, could not reach the corneal epithelial stem cells residing in the basal part of the corneal limbus.
The corneal epithelium is comprised of five to seven cell layers consisting of squamous cells in the outermost part, wing cells in the middle part and basal cells adjacent to the corneal stroma (14). When the superficial epithelial cells are detached (Z), the basal cells of the cornea that exist deep within the epithelium divide and become wing cells, which differentiate into squamous cells (X), and new basal cells are then supplied by the stem   cells in the corneal limbus (Y) due to the limited number of basal cell divisions. A balanced relationship is maintained between the migration and differentiation of these cells, expressed as "X + Y = Z", and the epithelial cells are replaced in 5-7 days (4,12). Therefore, cells in the superficial layer of the epithelium are expected to detach in roughly 24 h. In previous studies, we have shown that Far-UV-C is substantially less hazardous to the cornea, because it only reaches the superficial layers of the corneal epithelium, which is detached within 24 h by the normal physiological turnover cycle (8). Specifically, the less hazardous properties of Far-UV-C are based on maintaining the X + Y = Z balance, as described above, in corneal epithelial turnover and superficial cell layer detachment in 24 h. The most important cells in the normal turnover of the corneal epithelium are the corneal epithelial stem cells, which are the source of all the cells that constitute the corneal epithelium. Since corneal epithelial stem cells are located in the basal area of the corneal limbus between the cornea and conjunctiva (4,13), it is important to evaluate whether UV radiant energy reaches the basal area of the corneal limbus. In this study, we confirmed that Far-UV-C reaches only the superficial layer of the corneal limbal epithelium. In addition, CPD, a marker of UV penetration depth, was observed in the basal area of the rat corneal limbus in eyes irradiated with energies as high as 10 000 mJ cm À2 at 207 nm and 5000 mJ cm À2 at 222 nm. These results suggest that it is extremely unlikely that stem cells in the basal area of the corneal limbus are damaged by Far-UV-C.
Corneal stromal-derived mesenchymal stem cells (CS-MSCs) have self-renewal and multidirectional differentiation potential (15). In vitro studies have shown that CS-MSCs can not only differentiate into keratinocytes (16) that produce types I, V and VI collagen and chondroitin sulfate, which are components of the corneal stroma, but also promote cloning of limbal stem cells through close physiological interactions (17). CS-MSCs mainly present in the anterior of the corneal stroma near the limbal stem cells (15) and may play an important role in maintaining limbal stem cell proliferation. Since CS-MSCs not only contribute to the transparency and stability of the corneal stroma but are also thought to play a supporting role in normal corneal epithelial turnover, injury to CS-MSCs by UV radiant energy could cause significant damage to the cornea. Considering the tissue penetration depth of Far-UV-C, which reaches the superficial layers of the corneal epithelium, the potential for Far-UV-C to injure CS-MSCs would also be extremely low. In other words, Far-UV-C is safer not only for the corneal epithelium but also for the corneal stroma.
The long-term effects of Far-UV-C on the cornea should also be considered because pterygium and conjunctival tumors are corneal diseases caused by long-term UV irradiation injury to the corneal limbal stem cells (2)(3)(4). In the present study, the effects of a single exposure to UV radiant energy on the corneal limbus were investigated, but the long-term safety of Far-UV-C was not examined. However, no injury was observed in the rat cornea and limbus at 1 week post-exposure to 222 nm UV-C, making it unlikely that prolonged Far-UV-C irradiation causes pterygium or conjunctival tumors. Furthermore, it is unlikely that repeated Far-UV-C irradiation causes pterygium or conjunctival tumors because 207 and 222 nm UV-C did not reach the corneal limbal epithelial stem cells, and no damage was observed in the corneas of animals in experiments in which Yamano et al. (9) repeatedly exposed mice to 222 nm UV-C over a long period.
In the present study, we irradiated the corneas of rats and porcine with UV radiant energy and used CPD as a marker to observe the penetration depth into the tissue. The thickness of the rat corneal limbal epithelium was similar to that of the central corneal epithelium, whereas the porcine corneal limbal epithelium was twice as thick as the central corneal epithelium. An energy dose of 600 mJ cm À2 emitted by a mercury lamp with main peak wavelength at 254 nm caused CPD development in the basal cells of the rat corneal limbal epithelium, whereas in the porcine corneal limbus, CPDs were observed in cells in the middle area, but not in the basal area. This suggested that at an energy dose of 600 mJ cm À2 , even UV radiant energy emitted by a mercury lamp with main peak wavelength at 254 nm does not reach the stem cells in the basal area of the limbus. Since the porcine corneal limbus is thicker than that in rats, stem cells in the basal area of the limbus are less susceptible to damage. When porcine and human corneas are compared, the tissue structures are very similar, although the porcine corneal epithelium is slightly thicker (18). While the current study was conducted in animals, Far-UV-C does not reach the basal area of the corneal limbus in human corneas; therefore, stem cells in the basal area of the limbus are presumably not damaged.
A limitation of this study is that the rats irradiated with UV radiant energy were under anesthesia. Although anesthesia has made it possible to irradiate UV radiant energy at a set energy level with some degree of accuracy, the effect of blinking during wakefulness is unknown, and the dynamics of the tear fluid layer on the corneal surface may have been altered by the diluted saline solution applied to prevent drying. Sterilization with 222 nm UV-C irradiators is already in practical use, but the devices are programed for longer periods of time with weaker radiant exposure doses based on working hours. Contrarily, in this study, irradiation was performed for a duration as short as possible to reduce the stress on the animals caused by the anesthesia. The effects of anesthesia on stem cells and corneal epithelial turnover requires further investigation using irradiation on awake animals.
In this study, Far-UV-C, such as radiation at wavelengths of 207 and 222 nm, did not reach the stem cells of the corneal epithelium. This suggests that the turnover cycle of the corneal epithelium is not affected by UV-C, and that pterygium and conjunctival tumors, which are caused by UV exposure, are unlikely to develop. Infection by Coronavirus Disease 2019 (COVID-19) is expected to continue, and the use of Far-UV-C will become increasingly widespread. Although the results of this study indicate that Far-UV-C is safe for the cornea, experimentation was only performed on animals under anesthesia, and the determination of the safety of Far-UV-C for the human cornea will need further investigation.

CONCLUSION
Far-UV-C, such as radiation at wavelengths of 207 and 222 nm, did not reach the stem cells in the basal area of the corneal limbus, i.e. the stem cells were intact. Therefore, the turnover of the corneal epithelium is obstructed or disrupted.