Matrix Metalloproteinase-9 and Topical Cyclosporine Are Associated with Conjunctival Microbiota Culture Positivity in Korean Patients with Stevens-Johnson Syndrome

Stevens-Johnson syndrome (SJS) is an abnormal immune-response causing extensive exfoliation of the mucocutaneous tissue including conjunctiva. While several factors are associated with the alteration of conjunctival microbiota, the conjunctiva of SJS patients are found to harbor a different microbiota compared to healthy subjects. We investigated the conjunctival microbiota of Korean SJS patients, and identied factors associated with the conjunctival microbiota and its positive culture.

[8, 9,13,14] Higher positive culture rate was observed in elder subjects or with diabetes. [15] Subjects with dry eye or blepharitis were reported to exhibit different conjunctival microbiota compared to healthy subjects. [14][15][16][17] In particular, the conjunctiva of SJS patients were found to harbor a signi cantly different microbiome and have higher culture-positive rate compared to healthy subjects. [8,11,14,18] While the microbiome of several areas of the body have been studied to affect human diseases, the physiological role of ocular surface microbiome is yet unknown. Still, evidence show that an appropriate balance between the ocular surface microbiome and its mucosal immunity helps maintain the homeostasis of commensal bacteria and prevent opportunistic infections. [19] Likewise, a weakened and damaged ocular surface of SJS patients may be more prone to harbor pathobionts and allow opportunistic infections, which can also be aggravated by surgical interventions or medical therapy such as topical immunosuppressant or antibiotics. [19][20][21][22][23] Herein, we investigated the conjunctival microbiota of Korean SJS patients using conventional swab cultures, and identi ed factors associated with the conjunctival culture results.

Subjects and Study Design
This study was approved by the Institutional Review Board of Seoul National University Hospital (IRB No. 2102-014-1193, Seoul, Republic of Korea) and was conducted with adherence to Declaration of Helsinki. The informed consent from patients was waived by the IRB because the study was based on the retrospective review of old charts. This is a retrospective case-series study of SJS patients who had undergone conjunctival swab culture sampling between January 1st, 2019 and December 31st, 2020 at Seoul National University Hospital (Seoul, Republic of Korea). From medical chart review, the following data were collected: 1) general medical history and demographic information, 2) clinical characteristics from ocular examinations including COCS (range, 0 -15), TBUT and tear MMP9 elevation, and 3) conjunctival swab culture results.
The eye with the highest COCS was chosen for analysis and if the scores were the same in both eyes the right eye was included. Excluded from analysis were patients under 18 years of age, with active infectious keratitis and with insu cient clinical data, such as conjunctival swab culture results and COCS.
TBUT was evaluated under slit lamp biomicroscopy with cobalt blue lter after application of uorescein strip. TBUT was measured three consecutive times with a stop watch after each blink. The average of the three measurements was used for analysis.
MMP9 elevation to ≥ 40 ng/ml was tested using In ammaDry test (RPS Diagnostics; Sarasota, FL, USA) according to the manufacturer's instruction. [27,28] The In ammaDry device was gently dabbed at multiple locations of the inferior tarsal conjunctiva with releasing the lid after every 2-3 dabs and allowing the patient to blink. After obtaining su cient tear sample, the device was immediately loaded onto the test cassette and placed directly into the manufacturer's provided buffer solution. After 10 minutes, positivity for MMP9 ≥ 40 ng/ml was indicative when 1 blue line and 1 red line was present in the device's test result window. [27,28] Conjunctival Swab Culture And Drug Sensitivity Test Conjunctival swab culture was performed initially before applying any eyedrops, including topical anesthesia or uorescein strips. Conjunctival swab sampling from each eye from deep portions of the medial and lateral lower conjunctival fornix was carried out using a sterilized cotton tip for each site. [11] Careful caution was taken to avoid the sterilized cotton tip from being in possibly contaminated by factors, such as the eyelid skin or eyelash. After obtaining each swab sample, it was immediately inoculated directly onto either blood agar plate or Sabouraud's agar plate. When possible, a repetition of conjunctival swab culture was performed in the same manner at an interval of at least 3 months since last test.
Incubation for the growth of bacteria and fungus, and drug susceptibility tests were performed at the department of laboratory medicine at Seoul National University Hospital (Seoul, Republic of Korea). Blood agar plate was used to culture a wide range of bacteria, including fastidious microbes and those that are di cult to grow such as Streptococcus and Staphylococcus, and to differentiate hemolytic bacteria. Sabouraud's agar plate was used in cultivating fungus, such as yeasts and molds, and lamentous bacteria. All collected specimens were incubated at 37 ˚C (for bacteria) or 30˚C (for fungi) and examined daily for microorganism growth for 1 week and weekly for up to 1 month. Laboratory analyses consisted of culture, microorganism identi cation, and drug sensitivity tests. Drug sensitivity tests were performed by agar diffusion method using the subsequent antibiotics according to CLSI guidelines: ampicillin, oxacillin, penicillin G, amoxicillin/clavulanic acid, imipenem, gentamicin, rifampicin, cipro oxacin, levo oxacin, moxi oxacin, trimethoprim/sulfamethoxazole, teicoplanin, vancomycin, clindamycin, erythromycin, nitrofurantoin, linezolid, quinupristin/dalfopristin, tetracycline, cefoxitin, ceftriaxone, chloramphenicol, piperacillin/tazobactam, cefotaxime, ceftazidime, cefepime, ertapenem, meropenem, amphotericin, uconazole, voriconazole, ucytosine. [29] Statistical Analysis Statistical analysis was performed using SPSS software for Windows version 22.0 (SPSS, Inc., Chicago, IL). The 2-tailed unpaired t-test and Pearson Chi-square test performed at the 95% con dence interval were used to assess differences in continuous and categorical variables. A probability value of < 0.05 was considered statistically signi cant. The results are presented as the means ± standard deviations (SDs) unless otherwise indicated.

Demographic and ocular characteristics in SJS patients
General demographics and clinical ocular characteristics of the enrolled patients are shown in Table 1. A total of 30 eyes from 30 patients with SJS were assessed. The average age was 47.8 ± 16.5 (18 -71) years and the disease onset age was 34.6 ± 17.8 (7 -66) years. The disease duration was 13.1 ± 12.1 (0 -52) years. 9 (30.0%) and 21 (70.0%) patients were male and female, respectively. Cold medications were the most common cause for SJS, followed by antibiotics and other medications, such as antiepileptic drugs. Fourteen (46.7%) patients had past experience of infectious keratitis. Nearly 70% of all patients were using topical medications, such as corticosteroids, cyclosporine or antibiotics, at the time when conjunctival swab culture sampling was performed. Schirmer test revealed an average of 6.3 ± 6.3 mm while the tear break-up time (TBUT) was 3.4 ± 1.1 seconds. Matrix metalloproteinase 9 (MMP9) was positive in 24 (80.0%) eyes. The average chronic ocular surface complications score (COCS) was 8.4 ± 3.3. Low COCS was considered when the score ranged between 0 and 7, while a score of 8 or higher was de ned as high COCS. [24,25] 10 (33.3%) and 20 (66.7%) eyes were low and high COCS, respectively. The representative photos of low (0 -7) and high (≥ 8) COCS are shown in Figure 1. Figures 1A and 1B represent low COCS patients and are photos of a female subject's left eye with a COCS score of 3 who had been diagnosed with sulfasalazine-related SJS at the age of 49. Her left eye was observed to have 12 clock hours of corneal neovascularization and nasal corneal opacity in absence of abnormal eyelids, chronic conjunctival hyperemia and symblepharon (Fig. 1A). Under cobalt blue lter examination after uorescein application, super cial punctate epithelial erosions were observed in the inferior 2/3 of the cornea (Fig. 1B). Figures 1C and 1D represent high COCS and are photos of a female subject's left eye with a COCS score of 10 who had an onset of SJS at age 44 after taking cold medications. Her left eye showed conjunctivalization at the upper 2/3 of the cornea due to partial limbal stem cell de ciency with diffuse corneal haze, chronic conjunctival hyperemia, severe eyelid meibomian gland dysfunctions with trichiasis and lateral symblepharon at both upper and lower portions (Fig. 1C). Under cobalt blue lter examination after uorescein application, diffuse super cial punctate epithelial erosions were observed ( Fig. 1D).

Isolation And Drug Resistance Results Of Conjunctival Swab Cultures
The initial microbial isolation results of conjunctival swab culture samplings are shown in Table 2. 13 (41.9%) patients had no growth of any microorganisms, while 17 (58.1%) patients were observed to have positive culture results. A total of 27 types of different microorganisms were isolated. Among them, 81.5% and 11.1% were gram positive and negative bacteria, respectively, while 7.4% was fungus origin. Among gram positive bacteria, Coagulase-negative Staphylococci (45.5%) and Corynebacterium species (40.9%) were predominantly observed. Among Coagulase-negative Staphylococci, Staphylococcus epidermidis (60%) was most commonly isolated followed by Staphylococcus hominis (30%). Among gram negative bacteria, Klebsiella pneumonia (33.3%), Stenotrophomonas maltophilia (33.3%) and Escherichia coli (33.3%) were observed. All the isolated fungi were Candida. The number of patients according to isolated number of microorganisms are shown in Table 3. Positive culture of a single type of microorganism was seen in 12 (70.6%) of the 17 patients with positive culture results, in which Corynebacterium species was most commonly isolated followed by Staphylococcus epidermidis. Two and three types of microorganism isolations were observed in 2 (11.8%) and 3 (17.6%) patients, respectively. All fungi were isolated in mixture with other bacteria.  Fig. 2A). Also, no signi cant difference in Schirmer test (negative group 5.4 ± 4.0 mm vs. positive group 6.9 ± 7.7 mm) and TBUT (negative group 3.8 ± 1.0 sec vs. positive group 3.1 ± 1.1 sec) were observed between groups (Unpaired t-test, P = 0.511 and 0.261, respectively, Fig. 2B and 2C). Positive for tear MMP9 was observed signi cantly more in the positive group compared to the negative group (100.0% vs 75.0%, respectively, Pearson Chi-square test, P = 0.040, Fig. 2D). There was no difference in the use of topical medications, such as corticosteroids, cyclosporine or antibiotics, and in the history of prior infectious keratitis (Pearson Chi-square test, P > 0.05, Fig. 2E and 2F).

Discussion
This study presented (1)  microbiome and ocular surface immunity interact is yet unknown, but studies have seen the importance of their balance and, therefore, given that SJS patients have generally abnormal ocular surfaces, they are prone to the imbalance that subsequently can induce commensal bacteria to become pathobionts and lead to possible opportunistic infections. [10,19,20,22,23] Previous studies reported that the conjunctival swab culture positivity of SJS patients was observed in 59 -95%. to most antibacterial agents, they identi ed all isolated bacteria from SJS conjunctiva to be sensitive to gati oxacin and moxi oxacin, but resistant against cipro oxacin in Coagulase-negative Staphylococci, cipro oxacin, tetracycline, cefazolin, and moxi oxacin in Escherichia coli, and tobramycin and gentamicin in Streptococcus pneumonia. [11] Also, they reported that Streptococcus viridans had the highest percentage of drug resistance. [11] Frizon et al. identi ed low sensitivity to neomycin and penicillin in the Staphylococcus group, only 33% sensitivity to chloramphenicol in gram negative bacilli and observed the highest antimicrobial resistance in the Streptococcus group.
[18] However, this study found the highest percentage of drug resistance in Coagulase-negative Staphylococci, where more than half were resistant to penicillin, oxacillin, ampicillin, levo oxacin, moxi oxacin or cipro oxacin, and did not observe any drug resistance in Streptococcus viridans. Also, more than half of the isolated Corynebacterium species exhibited resistance to clindamycin and erythromycin. Aside from Escherichia coli that was resistant to trimethoprim/sulfamethoxazole and cipro oxacin, gram negative bacteria from this study did not reveal any drug resistance. The difference in the drug susceptibility test results regarding the Streptococcus group compared to previous studies may be because of the low culture rate of Streptococcus in this study. Also, the difference in drug resistance of the same kind of microorganism among studies may be attributed to the type of topical antibiotics previously or currently used, in which subjects from Frizon et al.
[18] mainly used chloramphenicol eye drops but the patients from this study mainly applied topical levo oxacin or moxi oxacin. Although the drug susceptibility results may differ according to studies, it is still consistently found that the isolated conjunctival microorganisms from SJS patients have relatively higher percentage of drug resistance compared to healthy subjects. This high percentage of drug resistance in SJS patients may be attributed to the chronic use of topical antibiotics and, moreover, the abnormal ocular surface environment leading to the abundance of more virulent pathogens.
To the best of our knowledge, this study is the rst to identify the association between tear MMP9 positivity and topical cyclosporine use with conjunctival swab culture positivity and its persistency in positive culture in SJS patients, respectively. Despite recent studies concerning conjunctival microbiota, the role of ocular surface microbiota remains unclear. Moreover, most studies have focused on investigating the compositional or alpha/beta-diversity difference of conjunctival microbiota in subjects with various ocular surface diseases compared to healthy subjects rather than possible relative factors. Though a previous study observed higher culture positivity rate in SJS patients with more severe ocular surface scores,[8] this study did not nd any signi cance between COCS and conjunctival swab culture results. In this study, tear MMP9 positivity was more signi cant in patients with culture positive results compared to those with culture negative results. Also, patients with culture positive persistence had higher tear MMP9 positivity rate, though it was not signi cant, and consisted more patients using topical cyclosporine than those who experienced a transition to negative culture results. Despite the chronic use of topical corticosteroid or cyclosporine, the in ammatory nature of SJS may not be fully controlled and may instead promote an environment favoring an abnormal growth of conjunctival microorganisms. positivity may re ect not only the presence of ocular surface in ammation but also the possibility of a more aggressive or persisting abnormal conjunctival microbiota.
There are some limitations to this study. First, it is limited by the relatively small study size. However, given that SJS is a rare disease, a recruitment of 30 patients is not a small number and is similar to previous studies regarding SJS conjunctival microbiota. Also, this study has value in that this is the rst to investigate the conjunctival microbiota in Korean SJS patients. Another limitation is that this study observed the conjunctival microbiota using the conventional swab culture method. Though genetic sequencing methods provide more precise and broader information, conventional swab culture still has the advantageous of simplicity and practicality. Lastly, this study is limited in that there is no control group. However, previous studies have already observed that there is distinct difference in conjunctival microbiota between SJS and healthy subjects. Still, future studies comparing SJS and healthy subjects' conjunctival microbiota and further exploring their associative factors will be bene cial in elucidating the role of ocular surface microbiota.

Conclusion
In conclusion, this study demonstrated the high culture positivity rate in SJS patients' conjunctiva. Coagulase-negative Staphylococci and Corynebacterium species were the most commonly isolated microorganisms. Tear MMP9 positivity was associated with positive culture whereas use of topical cyclosporine was related to positive culture-persistence. These ndings suggest that tear MMP9 positivity in SJS may indicate the presence of an abnormal ocular surface microbiota. Further investigations regarding the associative factors of conjunctival microbiota in SJS patients are necessary in understanding the relation between conjunctival microbiota and ocular surface immunity.  culture results, respectively (unpaired t-test, P = 0.644) (A). No signi cant difference was observed in tear secretion by Schirmer test between groups (negative group 5.4 ± 4.0 mm vs. positive group 6.9 ± 7.7 mm, unpaired t-test, P = 0.511) (B). Also, TBUT was not signi cantly different and were 3.8 ± 1.0 and 3.1 ± 1.1 seconds in the negative and positive groups, respectively (unpaired t-test, P = 0.261) (C). Tear MMP9 positivity was higher in patients with positive culture (100%) compare to those with negative culture (75%) results (Pearson Chi-square test, P = 0.040) (D). There was no difference in the use of topical medications, such as corticosteroids, cyclosporine or antibiotics (E), and in the history of prior infectious keratitis (F) (Pearson Chi-square test, P > 0.05). COCS: Chronic Ocular Surface Complications Score, TBUT: Tear break up time, MMP9: Matrix metalloproteinase 9, vs: Versus * P < 0.05, Pearson Chi-square test Figure 3 Clinical characteristic differences between patients with repetitive cultures resulting in negative transition and positive persistence. Among the 13 patients who had repetitive conjunctival swab cultures, 3 (23.1%) patients experienced a transition from positive to negative culture results (negative-transition group), while 10 (76.9%) patients had persistent positive culture results (positive-persistence group). There was no difference regarding COCS between groups (negative-transition group 9.0 ± 4.0 vs. positive-persistence group 10.2 ± 2.3, unpaired t-test, P = 0.512) (A). No signi cant difference in Schirmer test was observed between groups (negative-transition group 3.0 ± 2.6 mm vs. positive-persistence group 3.8 ± 3.4 mm, unpaired t-test, P = 0.719) (B). Also, TBUT did not differ between groups (negative-transition group 4.0 ± 1.1 sec vs. positive-persistence group 2.8 ± 1.2 sec, unpaired t-test, P = 0.138) (C). Positivity for tear MMP9 was not different between negative-transition group (66.7%) and positive-persistence group (90.0%) (Pearson Chi-square test, P = 0.070) (D). Patients using topical cyclosporine was observed