The exogenous crosslinking of tissular collagens accomplished in vivo by sensitizer-assisted exposure to selected regions of electromagnetic radiation has found a major clinical application in ophthalmic medicine. Keratoconus, an ectatic corneal disorder leading to severe visual complications, is currently treated non-surgically by crosslinking the corneal stromal collagen in vivo upon irradiation with ultraviolet A (UVA) radiation (315 to 400 nm wavelengths) in the presence of a water-soluble derivative of riboflavin (RF) as a photosensitizer. The procedure, known by the acronym CXL, has become the routine non-surgical treatment for keratoconus [1]-[6]. As a result of subsequent investigations, the UVA/RF crosslinking system has also been proposed for preventing progressive myopia [7]-[9] and for treating eyelid laxity conditions [10]-[20].

The UVA/RF protocols appear biologically safe in CXL applications, which would be expected considering that UVA is non-ionizing radiation and the photosensitizer RF is a biocompatible chemical compound (vitamin B2), and their potentially harmful side effects have been duly investigated. Although the standard CXL procedure involves a radiant exposure (fluence) of 5.4 J/cm², which is higher than the limit of 1 J/cm² as advised by existing guidelines [21], it was shown [22] [23] that by absorbing radiation, RF actually provides a ‘shield’ that reduces the total exposure to levels below those recommended. Furthermore, it was found that only when UVA and RF act together can they damage in vitro the corneal limbal epithelial cells [24] or harm in vivo (mouse) the endothelial cells [25], suggesting that riboflavin alone does not affect the corneal cells. That was contradicted by an in-vivo study (rabbits) showing that by preventing RF solution from reaching the corneal limbus during CXL, limbal epithelial stem cells were preserved, but they were damaged in the corneas where RF solution was freely infused [26]. Likely because of such equivocal findings, alternative crosslinking systems have been suggested for CXL. Thus, the pharmacologic alternative involves using various chemical compounds as crosslinking agents, with no need for radiation [27] [28]. However, many of these agents may display cytotoxicity at certain doses, and there is little evidence so far to support their potential clinical performance.

In order to circumvent the UVA/RF protocol, UVA radiation was replaced with visible light, which allowed both lower photon energies and the utilization of photosensitizers different from RF. For instance, rose bengal dye was employed as a photosensitizer during the irradiation of animal corneas with green light radiation (532 nm). Experiments performed ex vivo (rabbit, lamb, pig) [29]-[32] or in vivo(rabbit) [33] have shown that mechanical reinforcement was equal to, or even better than, that obtained with the UVA/RF procedure. More recently reported, and of relevance to our investigation, was the use of a ruthenium compound, Ru(II)(bpy)₃Cl₂, as a photosensitizer activated by blue light radiation for crosslinking corneal collagen in animal models. The method led to results superior to the UVA/RF procedure regarding multiple aspects (biomechanical, biochemical, optical) in both ex-vivo (bovine) [34] and in-vivo (rat) corneas [35], all at relatively low radiant exposures.

We report here the evaluation of a ruthenium-based photosensitizer associated with blue light (470 nm) as it was applied for crosslinking the tarsal collagen in order to enhance the mechanical characteristics (strength, stiffness) of the upper eyelid in ex-vivo animal (goat) specimens.

As seen in Table 1, at radiant exposures of 0.9, 1.8, and 3.6 J/cm², respectively, there were statistically significant increases in both M_Y (stiffness) and strength. However, at the highest exposure tested (6 J/cm²), involving an irradiance of 20 mW/cm², the mechanical enhancement was very small and not statistically significant.

Table 1. Effects of ruthenium-assisted irradiation with blue light on mechanical properties of the eyelid in an ex-vivo goat model.

^ARegulated name for “fluence”. ^Bn = 8. ^CStress recorded at a strain of 10%.

For a more suggestive evaluation, a comparison between the percentage increases against native untreated samples of M_Y and strength following the radiation-induced crosslinking is illustrated as bar graphs in Figure 3. Although the increases as such are not statistically significant between them, this representation confirms the enhancing effect of crosslinking and its marked decline at the highest radiant exposure applied.

A method to reduce the pathological laxity of eyelids has been developed in our laboratories [10]-[17], and its conceptual validity and potential efficacy have been confirmed by other investigators [18]-[20]. The method is based on the use of a UVA/RF crosslinking protocol. Our goal in the current study was to evaluate a ruthenium compound as a photosensitizer during exposure of the palpebral tissue to blue light radiation (470 nm) to achieve the crosslinking of the eyelid’s tarsal collagen.

Photochemistry of the complexes with bipyridine ligands of ruthenium (and other metals) has experienced spectacular progress over the past decades, as described in several seminal reviews [36]-[39]. The prototype coordination divalent cation, [Ru(bpy)₃]²⁺, displays a distinctive combination of chemical and photochemical properties that play important roles in photochemistry, photophysics, and related fields, and generated research works reported in about 9000 publications between 1985 and 2020 [39]. It is no surprise that this cation has been deservedly qualified as “one of the most influential molecules for the development of chemistry in the last five decades” [39], and effusive reviews have been recently dedicated to it [40]-[42].

Drawing upon previous research on the oxidative crosslinking of proteins [43]-[47] and the need to find rapid and facile methods for crosslinking proteins as a tool to investigate protein-protein interactions in multiprotein complexes, the first use of Ru(bpy)₃Cl₂ as a photosensitizer in such processes was reported in 1999 [48]. A number of selected proteins were crosslinked upon irradiation with light generated by a xenon lamp over very short exposure durations. According to the mechanism proposed [48]-[50], and generally accepted, the photochemical reactions also require the “sacrificial” presence of the persulfate anion ( S 2 O 8 2 − ) as an electron acceptor that contributes to the formation of the [Ru(bpy)₃]³⁺ trication by the transfer of an electron from the excited dication, alongside a sulfate anion and a sulfate anion radical. It was found [49] that in the absence of a persulfate anion, the yields of crosslinked reactions were very low. The [Ru(bpy)₃]³⁺ trication is a strong oxidant able to subtract electrons from amino acids such as collagen’s tyrosine residues, thus generating tyrosine radicals that produce dityrosine crosslinks, while the [Ru(bpy)₃]²⁺ dication is recovered into its ground state. The rapid crosslinking of proteins assisted by ruthenium coordination compounds in association with electron acceptor molecules (e.g., persulfate anions) is included in the category of “photo-induced crosslinking of unmodified proteins” (PICUP). In the field of analytical crosslinking for encoding information related to the supramolecular structure of noncovalent multiprotein complexes in living tissues, PICUP constitutes one of the main approaches [51] [52]. In a distinct approach, aiming at creating novel biomaterials such as tissue adhesives or sealants, or templates for tissue engineering applications, the Ru(bpy)₃Cl₂/persulfate/blue light protocol was employed for crosslinking proteins containing tyrosine residues, such as resilin, fibrinogen, or gelatin [53]-[58], where the purpose was different but the Ru complex played the same role of a photosensitizer.

As mentioned, Gulzar et al. [34] [35] were the first to apply, both ex vivo and in vivo, the Ru(bpy)₃Cl₂/persulfate/blue light system to an ocular tissular component (the deepithelialized cornea) in animal models. The aim was to find a safer alternative for implementing CXL that would be associated with fewer complications and less clinical failure [34]. Radiant exposures of 7.2 J/cm² (irradiance 3 mW/cm² for a 40-min exposure time) for ex-vivo experiments, and 0.9 or 1.8 J/cm² (3 mW/cm² for 5 or 10 min) for in-vivo animal (rat) trials were used by investigators [34] [35]. They determined in cell cultures of three different cell lines that a concentration of 1 mM (~0.08% wt) of ruthenium compound was biologically safe. These investigators’ conclusive finding was that the new protocol generated results superior to those produced by the UVA/RF protocol regarding safety, biocompatibility, and functionality [35].

In our present study, the Ru(bpy)₃Cl₂/persulfate/blue light protocol was applied to achieve the crosslinking of collagen(s) in the tarsal plate of an ex-vivo animal eyelid, upon activation with blue light (wavelength 470 nm). By using radiant exposures (fluences) between 0.9 and 3.6 J/cm², generated through irradiances of 3 or 6 mW/cm² for exposure durations of 5 or 10 min, the expected mechanical reinforcement occurred with statistical significance (Table 1). In the context of our previous results on the sheep upper eyelids using the UVA/RF protocol (at 7.2 or 9 J/cm²) [10] [11], both final stiffness and strength are numerically higher in the current study, and they were recorded at lower radiant exposure. Importantly, based on the results recorded at 6 J/cm², our findings confirm that the mechanical enhancement declines with increasing radiant exposures, as previously reported in sheep eyelids [10] [11].

Based on rather scant reporting, there may be a difference between the outcomes of crosslinking the corneal stromal collagen and crosslinking the tarsal collagen. Relatively high radiant exposures have been used on the human cornea with the UVA/RF protocol (12 J/cm²) [59] and on the animal cornea with the Ru(bpy)₃Cl₂/persulfate/blue light protocol (7.2 J/cm²) [34], with acceptable biomechanical enhancement. However, we have recorded a decline in this trend with both the UVA/RF protocol at ~9 J/cm² in sheep eyelids [11] and at ~6 J/cm² in goat eyelids with the Ru(bpy)₃Cl₂/persulfate/blue light protocol as shown in the present study. Hypothetically, this can be a result of the unique composition and properties of the tarsal plate (known also as the palpebral tarsus), which is composed of networks of several collagen types that surround the meibomian glands. In a detailed summary [13] of the current knowledge on this eyelid component, we have suggested that its formal inclusion within the fibrocartilage category might be wrong, as it does not contain collagen type II (typical of cartilages [60]), but contains collagen type VI, which is found in reduced amounts in other cartilages where is mainly located within the pericelluar matrix [60]-[63]. It is reasonable to assume that the tarsal plate is also substantially different from the corneal stroma.

Based on the findings presented here, the Ru(bpy)₃Cl₂/persulfate/blue light protocol is able to induce the crosslinking of the eyelid’s tarsal collagen at relatively low radiant exposure levels and short exposure durations, resulting in the mechanical reinforcement of the eyelid. Accordingly, this procedure can be applied as a potential non-surgical treatment for eyelid laxity conditions, also offering some advantages over UVA/RF, our previously employed protocol.

None applicable. Neither experiments with live animals nor trials on human subjects were involved in this study.

T.V.C. initiated the project, suggested the methodology, conducted the literature search, and wrote the first draft. S.S. designed, coordinated, and supervised the experimental protocols. M.M. carried out the experiments. All authors were involved in the interpretation of results. S.S. and M.M. contributed to and reviewed the final draft, and approved its submission. The authors assume full responsibility for the veracity of the results reported in this article.

All relevant data supporting this study are included in this published article. Additional information is available from the corresponding author upon reasonable request.