Rose Bengal-Assisted Photodynamic Antimicrobial Therapy for Keratitis

Introduction

Keratitis is an inflammation of the cornea due to microbial infections or noninfectious causes such as peripheral ulcerative keratitis, corneal neuropathy, trauma, or contact lens use.¹ The incidence of microbial keratitis is perceived to be much higher than noninfectious keratitis, although comprehensive global data is not available. Indeed, microbial keratitis is a leading cause of visual morbidity during the second to sixth decades of life,² contributing to 1.5–2 million cases of unilateral blindness worldwide.³ A population-based 1990 report on the incidence of microbial keratitis indicated 27,000 new cases annually in the USA and 840,000 new cases in India. Fungal keratitis constitutes nearly 40%–50% of microbial keratitis globally, and it is a public health concern.^4,5 Incidence of fungal keratitis is 23.6 cases per 100,000 population, with the following incidence across the world in the same units: South Asia 73, China (south, east, central) 15.2, Oceania 14.5, Africa 13.5, North America 6.8, Latin America 5, China (north, west, Hong Kong) 1.3, West Asia 0.12, and Europe 0.02.⁶ The estimated economic burden of keratitis due to all causes is $175 million per year in the USA.⁷ Infectious keratitis may lead to corneal opacity or perforation, requiring surgical management. Surgical treatments for keratitis include debridement of corneal infiltrates,² amniotic membrane placement,^8,9 application of tissue adhesives like cyanoacrylate,^10,11 intrastromal injections of antibiotics,¹² and therapeutic keratoplasty (TPK) or therapeutic corneal transplant of a full-thickness donor cornea.¹³ Corneal transplant may result in corneal scarring or prove nonfunctional, with a loss of best corrected visual acuity. Refractory fungal keratitis may be defined as nonhealing ulcers, if there are no signs of resolution after 2 weeks of maximal therapy.^14,15 Supplementary treatments such as photodynamic therapy, intrastromal injection of antifungals, and amniotic membrane placement may be explored in eyes with refractory fungal keratitis prior to considering TPK. Based on emerging clinical and preclinical evidence, this article discusses rose bengal photodynamic antimicrobial therapy (RB-PDAT) as a potential adjunctive therapy for refractory fungal keratitis.

Photosensitizer Adjunctive Therapy for Refractory Keratitis

There is growing interest in the use of photosensitizers such as riboflavin, rose bengal (RB), methylene blue, hematoporphyrin, chlorin e6, and toluidine blue,¹⁶ which release reactive oxygen species, to kill pathogens responsible for microbial or fungal keratitis. For instance, exposure to riboflavin A followed by ultraviolet A (UVA) light showed promise in treating bacterial keratitis.^17,18 This treatment induces collagen crosslinking (CXL), whose contribution to antimicrobial activity has yet to be fully elucidated. A randomized clinical trial showed that riboflavin with UVA exposure was not effective in treating acanthamoeba or fungal keratitis.¹⁸ In this study, patients with fungal keratitis received natamycin or amphotericin B, with or without riboflavin-UVA. Patients treated with riboflavin-UVA, irrespective of the antifungal received, had 1.32 increased odds of culture positivity at 24 h, when compared to those treated with antifungal alone; however, the result was not statistically significant. Thus, there is a need to explore alternative strategies for the treatment of recalcitrant fungal keratitis.

RB Is a Promising Photosensitizer for Antimicrobial Activity

RB-PDAT, wherein green light is exposed to RB-soaked tissue, is promising as an adjunctive treatment for fungal keratitis.^14,19–22 RB is a tetrachloro and tetraiodo derivative of fluorescein with a molecular weight of 1017.6 g/mol (Fig. 1),²³ which generates reactive singlet oxygen species (ROS) when activated by green light (518–550 nm) in the energy range of 5.25–150 J/cm².^14,20,24 RB stains damaged cells, enabling visualization of ocular surface defects.²⁵ The putative mechanisms of broad-spectrum antimicrobial activity of RB mediated by ROS are illustrated in Figure 2. RB-PDAT kills fungi or inhibits the growth of fungi, including Aspergillus, Fusarium, and Candida, in addition to bacteria such as methicillin-resistant Staphylococcus aureus (MRSA) and Pseudomonas spp., and parasites like Acanthamoeba.^33–35 Thus, RB-PDAT has the potential for broader application in treating microbial keratitis. In addition to the generation of ROS, RB-PDAT may enhance corneal CXL and/or corneal rigidity, which might inhibit microbial growth. In deepithelialized rabbit corneas, Cherfan et al.³⁶ and Kochevar et al.³⁷ reported that RB-PDAT increases corneal stiffness by 3.8- and 2.7-fold, respectively, without causing keratocyte toxicity. On the contrary, another rabbit study reported no significant increase in corneal stiffness with RB-PDAT, unlike UVA-CXL.³⁸ In additional rabbit studies, RB-PDAT using 0.1% RB and green light at 5.4–150 J/cm² over 15 min^20,24,39 showed apoptosis in the corneal stromal keratocytes for up to a depth of 125 µm of the stroma.²⁴ These studies evaluated the effects of RB in rabbit tissues from 1 to 36 days.^20,24 Treatment protocols and methods of measurement may account for these differences between studies.

Role of RB-PDAT as an Adjunctive Treatment for Fungal Keratitis

Interest in RB-PDAT as an adjunctive treatment for fungal keratitis is based on the observations in a patient suffering from drug-resistant Fusarium keratitis (Table 1).¹⁹ This report suggested in vitro and in vivo efficacy, with a complete in vitro zone of inhibition to RB-PDAT. Lawn cultures of Fusarium keratoplasticum isolated from the patient showed inhibition upon exposure to 0.1% RB and 518 nm green light at 3 energy settings: 10.8 J/cm² over 30 min, 5.4 J/cm² over 15 min, or 3.6 J/cm² over 10 min. The fungus was resistant to natamycin, amphotericin B, and voriconazole. The patient had fungal graft infiltrates, adequate contact lens hygiene, and antifungals for 7 weeks. Two doses of RB-PDAT were administered, for a total energy of 2.7 J/cm² at an interval of 2 weeks. Keratitis was resolved with reduced scarring and neovascularization at about 4 months after the first dose of RB-PDAT, with a stable ocular surface and endothelial function and no limbal stem cell deficiency or further side effects.¹⁹

Clinical pilot trials with adjunctive RB-PDAT using 0.1% RB and 525 nm, 6 mW/cm², green LED light that delivers 5.4 J/cm² over 15 min have shown a favorable resolution of fungal keratitis.²¹ Complete clinical resolution was observed in 3 out of 3 patients with Fusarium keratitis at 39.3 ± 5.13 days after RB-PDAT.²¹ Patients with Aspergillus and Acremonium keratitis (n = 4) worsened and underwent TPK.²¹ This study had an in vitro arm, which indicated inhibition of Fusarium and Acremonium. However, Aspergillus was not affected by RB-PDAT. Fusarium isolates from patients of keratitis were resistant to antifungals but susceptible to RB-PDAT.²¹ The lack of response of Aspergillus to RB-PDAT may be due to a variation in responsiveness of this organism or deep keratitis with a more posterior stromal involvement.

Outcomes of adjunctive 0.1% or 0.2% RB and exposure to greenlight at 5.2 J/cm² over 15 min have been studied in a series of 19 eyes of 18 patients.¹⁴ The study showed that Acanthamoeba spp. was the most common (n = 10 patients), followed by Fusarium (n = 4), Pseudomonas (n = 2), Curvularia (n = 1), and culture-negative (n = 1; presumed to be Acanthamoeba). The duration of symptoms was 2–64 weeks prior to the first RB-PDAT. The procedure was performed once in 9 patients, 2 times in 8 patients, and 3 times in 1 patient, as decided by the treating clinician. Resolution without the need for TPK occurred in 13 patients at 46.9 ± 26.4 days. Corneal perforation occurred in 5 patients at 15.8 ± 8 days of the first RB-PDAT. Improvement in best corrected visual acuity post RB-PDAT was not significant. No differences were noted in patients treated with 0.1% or 0.2% RB.

The experiences from the studies described above suggest a favorable response of microbial keratitis and fungal keratitis to RB-PDAT.^19,21,39 Failure of treatment is reported in 27%¹⁴ to 57%²¹ of eyes and could be due to low depth of penetration of RB. Chen et al.⁴⁰ measured the depth of penetration of RB in the human cornea using confocal microscopy. A depth of 100-micron penetration was estimated at the end of 30 min, following the application of 0.1% RB. This depth of penetration increased to 177 microns with the application of iontophoresis at 2.5 mA over 15 min. Since the penetration depth is 100 microns, whereas infiltrates in eyes with keratitis might be deeper at 230 to 280 microns or beyond, the low depth of penetration of RB may explain the lack of efficacy of RB-PDAT in deep keratitis.^14,40 Crosslinking of corneal collagen induced by RB-PDAT increases the strength of collagen, decreases the lytic action of collagenases, and arrests melting.^36,38,41,42

In another retrospective study,²² of the 31 patients who had RB-PDAT at any point of time during an episode of microbial keratitis, 77.4% of the patients achieved resolution at 2.72  $\pm$ 1.85 months after the first RB-PDAT. Therapeutic penetrating keratoplasty was required in 22.5% of the patients. About 35.4% of the patients underwent optical penetrating keratoplasty (PKP) at 9.29 ± 4.78 months, and 19.3% underwent deep anterior lamellar keratoplasty at 9.71 ± 1.92 months after the first RB-PDAT. Details of clinical reports supporting RB-PDAT are summarized in Table 1.

Facilitation of RB Penetration into Stroma by Iontophoresis and Ultrasound

The depth of RB penetration was ∼16% in rabbit cornea compared to ∼40% with riboflavin, which is 2.5 times lower, when estimated using a phase-stabilized swept-source optical coherence elastography system.³⁸ This limitation may be affecting the ability of RB to reach the site of action. Strategies employed include the use of iontophoresis to increase the permeability of RB into the corneal stroma. A study of iontophoresis-induced RB (iRB) penetration into human corneas noted greater depth compared to RB alone (177 ± 9.5 μm vs. 100 ± 5.7 μm, p < 0.001).⁴³ There are 2 advantages of iRB: an increase in tissue penetration and greater retention of laser light in the cornea, with less transmittance to deeper layers. A study on iRB using 0.1% RB exposed to green light, 250 mW/cm², delivering energy of 100 J/cm² over 7 min in rabbit eyes showed a diffusion depth of 42.22% ± 4.77% compared to 26.63% ± 3.84% without iontophoresis. Mean elastic modulus and corneal stiffness in the corneas undergoing iRB were 4.7 times higher compared to control corneas where no RB was administered and 2.2 times higher than the corneas exposed to plain RB with green light.⁴⁴ Drawbacks of iRB still include limited penetration of the drug into the corneal stroma, with possible residual infection posttreatment.

In tissues that are poorly penetrated by light such as dermis, a 28 KHz ultrasound source has been reported to inactivate gram-positive and gram-negative bacteria.⁴⁵ It led to 3–4 log₍₁₀₎ reduction in the number of microorganisms after exposure. A sonic and light-assisted photoinduction of RB has shown a reduction in dose and toxicity of RB and has been deployed for treating skin cancers.⁴⁶ These approaches in ocular tissue may facilitate the development of suitable photosensitizer compositions.

Potassium Iodide as a Potentiating Agent for RB

RB dye can be an effective photosensitizer against gram-positive bacteria.⁴⁷ Killing effect of RB for gram-negative organisms and yeasts can be potentiated by adding potassium iodide (KI) in the following manner: 25 mM of KI, 100 nM RB attained 7 logs of killing P. aeruginosa, and 100 mM of KI, 10 µM RB provided 6 logs of killing C. albicans after exposure to 10 J/cm² of 540 nm green light.⁴⁷ Singlet oxygen created by light excitation of RB could oxidize the iodide anions to peroxy iodide, which breaks into free iodine and H₂O₂, which is an antimicrobial agent. In addition, oxidized iodine also produces I^2.− and HOO^. that are highly reactive, short-lived species that might add to the antimicrobial activity.^48,49

RB diacetate (RBDA), a lipophilic derivative of RB, has been studied for antimicrobial activity.⁵⁰ Due to its lipophilic nature, it is well internalized into bacterial cells and host cells.⁵⁰ This study reported potentiated killing of E. coli and Candida with potassium iodide supplemented with RBDA after 2 h of incubation and exposure to 10 J/cm² of 540 nm green light (100 mM KI and 14 µM RBDA resulted in over 6 logs of killing of E. coli; 100 mM KI and 15 µM RBDA resulted in over 4 logs of killing C. albicans). In addition to killing the microorganism, healing of an infected skin wound was noted as suggestive of reduced phototoxicity despite the high lipophilicity of the RBDA molecule.⁵⁰

Photodynamic therapy with RB nanoparticles exhibited lower dark cytotoxicity, while phototoxicity remained unaffected in human liver carcinoma (HepG2) cells.⁵¹ Further, in a Wistar albino rat study, RB-loaded nanoparticles improved cure rates of fungal keratitis and reduced stromal penetration of fungi at 3 weeks.⁵² In this study, RB (500 µM) was loaded in polypyrrole gold-coated nanoparticles and excited with 530 nm green light for 5 min, followed by 840 nm of infrared light for 5 min.

Future Directions for Human Trials with RB-PDAT

Patient data discussed in this article are summarized in Table 1, suggesting the potential adjunctive role of RB-PDAT in patients with microbial keratitis and fungal keratitis, reducing the need for TPK. Fungal or acanthamoeba keratitis has a poorer clinical course. The availability of an adjunctive treatment of RB-PDAT may benefit the outcomes and reduce the need for TPK. An international, NIH-funded, randomized control two-arm trial of patients with smear-positive acanthamoeba and fungus and smear-negative ulcers, assigned to either topical antimicrobial plus sham or RB-PDAT, which is underway, is likely to yield definitive results for the role of RB-PDAT in microbial keratitis.⁵³ The results of this study are anticipated to have favorable outcomes in microbial keratitis and improve the best spectacle-corrected visual acuity, decrease the chance of perforation and the need for TPK. Similarly, innovative approaches for enhancing the delivery of RB may improve the depth of action of RB-PDAT in fungal keratitis.⁵²

Conclusions

RB-PDAT is showing promise as an adjunctive treatment in cases of refractory fungal keratitis. While the role of RB-PDAT’s collagen crosslinking in antimicrobial activities needs to be confirmed, its Type II reactive oxygen species after a photoexcitation reaction have potent antimicrobial effects. It must be understood that RB-PDAT may not be universally effective and limited by corneal penetration. Based on evidence and treatment protocols to date, genera such as Aspergillus and Acremonium are unaffected by RB-PDAT. Alternative strategies such as iontophoresis and sonic activation hold promise by increasing the depth of penetration of RB to deep corneal layers. However, these strategies remain in the realm of research and have limited clinical applications. The ultimate test is randomized controlled trials in humans. Thus, the upcoming trial of RB-PDAT trial⁵³ may yield answers regarding the utility of this approach. RB-PDAT has the potential to become an adjunctive therapy for keratitis induced by fungi, bacteria, and parasites.

Author Disclosure Statement

None of the authors have disclosures and conflicts of interest to declare.

Funding Information

This work was supported in part by the NIH grant R21EY035764.