Trigeminal stereotactic electrolysis induces dry eye in mice

Abstract

orneal nerves are known to con-tribute significantly to ocularsurface integrity. A normal cornealinnervation is required for a normaltear secretion and flow. In fact, dryeye disease is defined as a disorder ofthe lacrimal functional unit, an inte-grated system comprising the sensoryand motor nerves (DEWS 2007). Cor-neal nerve depletion reduces tearsecretion by decreasing reflex-inducedlacrimal secretion and by reducing theblink rate. Moreover, corneal nervessecrete up to 17 different mediatorssuch as nerve growth factor and oth-ers, which have profound effects oncorneal surface homoeostasis (Mu¨lleret al. 2003).Damage to the sensory nerves in theocular surface, particularly the cornea,is common in a number of conditions,including refractive surgery, keratopl-asty and even normal ageing (He et al.2010). Corneal nerve damage oftenoccurs in concomitance with inflam-mation elicited by surgery or drugs;hence, it is difficult to precisely dissectthe role of corneal nerves in dry eye.We recently developed a murinemodel of corneal denervation (TSE,Trigeminal Stereotactic Electrolysis)where we selectively ablate the oph-thalmic branch of the trigeminal nerve(Ferrari et al. 2011). We used thismodel to test whether and to whichextent selective sensory nerve ablationcould induce dry eye in terms of (i)tear secretion and (ii) corneal fluores-cein staining.Eight eyes of eight C57BL⁄6 micewere used. Animals were examinedbefore and 7 days after TSE. Theeffectiveness of the procedure wasconfirmed by the absence of blinkreflex (Ferrari et al. 2011). At the endof the procedure, a complete tars-orrhaphy was performed to minimizethe effect of absent blink reflex. Theanimals were examined using a biomi-croscope, and tear production wasquantified with phenol red thread test.Finally, the corneas were stained withfluorescein and scored according tothe National Eye Institute GradingSystem.Before TSE was performed, biomi-croscopy revealed a normal cornea.Corneal fluorescein staining was nega-tive before TSE (score, 0) and phenoltest was 2.59 ± 0.44 mm (mean ±SD). Corneal reflex was tested with acotton thread and found to be presentin all the eyes. Following TSE, corneafluorescein staining score was10.39 ± 2.20 (mean ± SD), phenoltest decreased to 0.52 ± 0.36 mm(mean ± SD) and corneal reflex wasabsent in all the eyes (Fig. 1). AfterTSE, we found significant differencesin both the fluorescein score (p <0.0001) and the phenol test(p < 0.0001) compared to the tests atbaseline before TSE (paired t-test).In conclusion, sensory cornealdenervation induces a form of dry eyein the TSE murine model. Thisoccurred despite complete tarsorrha-phy, suggesting that reduction in theblink reflex was not responsible forthis finding. Functional inhibition ofcorneal nerves with topical anaesthesiais able to reduce tear flow up to 75%(Jordan & Baum 1980). In this model,tear production was reduced byapproximately five times.Our findings may be due to thereduced sensory drive from the ocularsurface, which, together with areduced trophic influence of the nerveson the ocular surface epithelium,could result in a dry eye phenotype.We suggest that the TSE modelmay be useful in the study of dryeye disease, in particular as amodel for the dry eye patients whoalso demonstrate significant ‘neuro-trophic’ disease, such as viral kerati-tis, diabetes, or after refractivesurgery.