Patient-specific molecular signature may be predictive of CXL outcomes

September 26, 2018

A study of patients undergoing corneal crosslinking (CXL) for keratoconus suggests gene expression and tear inflammatory profile may be used as biomarkers to predict CXL outcomes.

Reviewed by Pallak Kusumgar, MS

Findings of a study associating the preoperative gene expression of certain molecular factors in corneal epithelial tissue and tears of keratoconus (KCN) patients with outcomes of corneal crosslinking (CXL) point to the possibility of developing an assay for predicting CXL response and complications.

“Our study suggests that the outcome of CXL may be determined by the levels of basal inflammation, collagen, and cellular components,” said Pallak Kusumgar, MS, Cornea, Cataract, and Refractive surgery fellow, Narayana Nethralaya Eye Hospital, Bangalore, India.

“Now our aim is to apply the results from this study and ongoing research to develop a rapid point-of-care kit for detecting predictive biomarkers in preoperative tears,” Kusumgar said. “Such an assay could provide evidence-based prognostication and customized treatment selection for improved patient outcomes.”

To investigate associations between molecular markers and CXL outcomes, Kusumgar and colleagues undertook a prospective study enrolling patients with progressive keratoconus. Eligible eyes had grade 1 or 2 keratoconus with >1 D increase in keratometry in the previous 6 months, corneal thickness ≥400 μm at the thinnest location, and contact lens intolerance, Kusumgar noted.

Patients with active allergic eye disease or ocular inflammation, central or paracentral scarring, or other ocular comorbidities were excluded.

Samples from the corneal epithelium over the ectatic cone area and the corneal periphery were obtained pre-operatively from KC patients when performing accelerated CXL. The tissue samples were analyzed for total mRNA levels of lysyl oxidase (LOX), matrix metalloproteinase 9 (MMP 9), bone morphogenic protein 7 (BMP7), tissue inhibitor of metalloproteinase 1 (TIMP1), and two collagens [collagen type I, alpha 1 (COL 1A1 and collagen type IV, alpha 1 (COL IVA1)].

Patients were followed for at least 6 months and categorized as achieving an optimal or suboptimal response defined by whether or not they had a ≥0.5 D decrease from the preCXL maximum keratometry (Kmax).

Of 37 evaluable patients, 26 (70%) had an optimal outcome; they had a mean Kmax change from baseline of 1.4 D.

The mRNA analyses showed a statistically significant difference in expression levels between groups for LOX expression, with the ratio of cone: periphery expression being significantly higher in the optimal versus suboptimal group (0.66 versus 0.25; p = 0.001).

Compared with the suboptimal group, there were trends for the optimal response group to have a higher cone:periphery ratio of BMP7, COL IVA1, and TIMP1 and a lower cone:periphery ratio of MMP9.

Two patients who underwent bilateral CXL developed a unilateral sterile infiltrate. Analyses of a number of inflammatory cytokines in their preoperative tear samples showed that in both patients, the levels of these cytokines were >1.5-fold higher in the eye with the sterile infiltrate compared with the fellow eye that did not develop a complication.

“This finding suggests that detecting and managing increased inflammation of the ocular surface prior to CXL might help to mitigate the risk of developing a sterile infiltrate,” Kusumgar said.

Rationale for methodology

The molecular factors investigated as potential biomarkers for predicting outcomes after CXL were chosen based on previous research showing that their level of expression in the corneal epithelium or tears correlated with clinical severity of keratoconus, said Kusumgar.

Knowing that KCN is a localized disease and that the molecular signature is different in ectatic and non-ectatic areas of the cornea, the investigators hypothesized that CXL outcomes may be associated with the expression levels of the molecular markers within the corneal cone versus in the periphery

Therefore, they analyzed the gene expression changes at the cone normalized to the periphery.
“We believed this methodology might also control for expression characteristics that were specific to individuals and provide better insight into tissue parameters that were predictive of CXL outcomes,” Kusumgar said.

Disclosures:

Pallak Kusumgar, MSE: MSpallak3004@gmail.com
This article was adapted from a presentation at the 2018 meeting of the American Society of Cataract and Refractive Surgery. A paper reporting on data from 35 eyes in the series has been published [Shetty R, et al. Cornea. 2018;37:369-374]. Kusumgar has no relevant fi nancial interests to disclose.