Reply to Zhang et al.: The differential role of LRRK2 variants in nested leprosy phenotypes

We thank Dr. Zhang et al. (1) for their thought-provoking comments and addition of exciting results to our recent study that identifies an overlap in the genetic control of type-1 reaction (T1R) and Parkinson’s disease (PD) (2). To put the comments by Zhang et al. into context, the purpose and design of our study is to analyze the contribution of rare variants to T1R by contrasting T1R-affected with T1R-free leprosy patients. Our study is not designed to test for common variants or the impact of genetic variants on leprosy per se susceptibility. Rather, our study deals with the role of specific genetic variants in the development of T1R among leprosy patients, and we never state or imply that the PRKN and LRRK2 genes have an exclusive impact on T1R but not leprosy. Indeed, we have published extensively on the role of common PRKN variants in leprosy susceptibility (3, 4)

As for PRKN, the two markers listed by Zhang et al. (1) do not show evidence for significant association with T1R in our study (2). We find that the V380L marker, which is not significant on its own, contributes to the rare variant analysis but not to an extent that impacts the overall conclusion about the contribution of rare PRKN variants to T1R risk. Given the nonsignificant effect of V380L, we do not attempt to do any functional validation of this particular variant.

Stimulated by the data of Zhang et al. (1) for LRRK2, we compared the R1628P distribution between T1R-affected and T1R-free leprosy cases of our sample against 472 ethnically matched healthy controls. For T1R-free leprosy patients, we observed a trend for a risk effect of R1628P (odd ratios [OR] = 1.64, P = 0.07) of similar effect size to the one reported by Zhang et al. when using ethnically matched controls (OR = 1.6; table 1 in ref. 1). Conversely, for T1R-affected cases, a protective effect was detected (OR = 0.44, P = 0.04). Hence, R1628P is not a leprosy per se (i.e., leprosy irrespective of the clinical manifestation) variant, but is highly dependent on clinical specifics of leprosy. Since the incidence of T1R in leprosy can be as low as 8 to 9% depending on patient characteristics, it is possible that the sample of Zhang et al. comprises a small number of T1R cases, which would result in the observed impact of the R1628P polymorphism on leprosy risk (5, 6). The opposite effect for R1628P between T1R-affected and T1R-free patients stresses the importance of assessing subgroup heterogeneity and the increased power to detect associations with secondary phenotypes when comparing extremes, as we discussed previously (7–9).

We conclude, from our results, that LRRK2 is an inhibitor of host inflammation (2). This conclusion is supported by the interesting observation, also quoted by Zhang et al. (1), that Mycobacterium tuberculosis (Mtb) induced neuroinflammation in Lrrk2 KO mice (10). Similarly, a Lrrk2 knockout had been shown to enhance the proinflammatory cytokine response to Mtb (11). Given that LRRK2 R1628P is a gain of function mutation, collectively, the results of multiple studies support a role of LRRK2 in the down-regulation of host inflammatory responses.