Pauci- and Multibacillary Leprosy: Two Distinct, Genetically Neglected Diseases

Abstract

After sustained exposure to Mycobacterium leprae, only a subset of exposed individuals develops clinical leprosy. Moreover, leprosy patients show a wide spectrum of clinical manifestations that extend from the paucibacillary (PB) to the multibacillary (MB) form of the disease. This "polarization" of leprosy has long been a major focus of investigation for immunologists because of the different immune response in these two forms. But while leprosy per se has been shown to be under tight human genetic control, few epidemiological or genetic studies have focused on leprosy subtypes. Using PubMed, we collected available data in English on the epidemiology of leprosy polarization and the possible role of human genetics in its pathophysiology until September 2015. At the genetic level, we assembled a list of 28 genes from the literature that are associated with leprosy subtypes or implicated in the polarization process. Our bibliographical search revealed that improved study designs are needed to identify genes associated with leprosy polarization. Future investigations should not be restricted to a subanalysis of leprosy per se studies but should instead contrast MB to PB individuals. We show the latter approach to be the most powerful design for the identification of genetic polarization determinants. Finally, we bring to light the important resource represented by the nine-banded armadillo model, a unique animal model for leprosy.

Fig 1. Concordance of leprosy polarization phenotypes according to WHO-82, WHO-96, or physician definitions in a Vietnamese sample.

Venn diagram for the number of patients with PB leprosy (left) and MB leprosy (right) according to three definitions: WHO-82, WHO-96, and physician in a Vietnamese sample. The surfaces of overlap are approximately proportional to the number of individuals identically classified by one, two, or three definitions. There was no individual classified PB by a physician and MB by WHO-82 + WHO-96 as well as no individual classified MB by a physician and PB by WHO-82 + WHO-96.

Fig 2. Proportion of multibacillary patients according to age and sex.

Proportion of multibacillary patients according to age and sex are given for two distinct geographic areas: (1) in Brazil, 40,544 PB and 29,764 MB cases (Madrid classification) were declared over the 2006–2010 period (national data) [28]; (2) in Tamil Nadu (India), during the 1962–1970 period, 3,963 PB and 1,258 MB cases were collected out of a total of 276,568 persons screened for leprosy (Madrid classification) [31].

Fig 3. Distribution of 91 published associations between a genetic variant and leprosy per se, any leprosy subtype (i.e., MB or PB), or polarization.

The + and–symbols refer to the presence or the absence of association of a given gene with one of the leprosy subtypes in the study. The red quadrant refers to studies for which the primary outcome was the association with leprosy polarization or the PB subtypes; the blue quadrant refers to studies that could not conclude an association with leprosy per se but did conclude an association with one of its subtypes; the green quadrant refers to studies that identified an association for both leprosy per se and one of its subtypes; and the grey quadrant refers to studies for which an association with only leprosy per se was reported.

Fig 4. Statistical impact of the hypothesis used to test for an association between a genetic variant and leprosy polarization.

In the common case-control study design, a standard test of association between a genetic variant and the phenotype under study is to compare allelic frequencies (f) between the group of cases and the group of controls. Here, we consider a sample including an even number of controls and leprosy cases, equally distributed between PB and MB. In addition, we fixed the frequency of the allele of interest to 0.5 among the controls (f_C = 0.5). In each panel, f_PB and f_MB (x and y axis) are the allele frequencies among PB and MB individuals, respectively. The first strategy directly compares MB to PB cases (Panel A). The second strategy first tests PB versus controls (Panel B) and MB versus controls (Panel C) before performing a heterogeneity test MB versus PB, possibly after testing leprosy per se versus controls (panel D). The red SNP is a variant for which f_MB ≠ f_PB and for which the association is significant with the first strategy but not with the second (the red SNP is in the gray area where the null hypothesis H₀ cannot be rejected). This second strategy is indeed hampered by correcting for multiple testing (testing MB versus PB after testing PB versus controls and MB versus controls) as well as the usual low power of heterogeneity tests.