MAJOR ARTICLE

Fernando P. Lázaro,1,a Renata I. Werneck,1,a Ciane C. O. Mackert,1 Aurélie Cobat,3,4 Flávia C. Prevedello,1
Raphaela P. Pimentel,1 Geraldo M. M. Macedo,2 Marco A. M. Eleutério,1 Guilherme Vilar,1 Laurent Abel,3,4,5
Marı́lia B. Xavier,2 Alexandre Alcaı̈s,3,4,5,b and Marcelo T. Mira1,b
1

Advanced Molecular Investigation Core, Graduate Program in Health Sciences, Pontifical Catholic University of Paraná, Curitiba, Paraná,
and 2Tropical Medicine Core, Federal University of Pará, Belém, Brazil; 3Laboratoire de Génétique Humaine des Maladies Infectieuses,
Institut National de la Santé et de la Recherche Médicale U550, and 4Université Paris René Descartes, Faculté Médecine Necker,
Paris, France; and 5Laboratory of Human Genetics of Infectious Diseases, Rockefeller Branch, Rockefeller University, New York, New York

Background. Leprosy is a chronic infectious disease that affects 250,000 new individuals/year worldwide.
Genetic analysis has been successfully applied to the identification of host genetic factors affecting susceptibility
to leprosy; however, a consensus regarding its mode of inheritance is yet to be achieved.
Methods. We conducted a complex segregation analysis (CSA) on leprosy using data from the Prata Colony,
an isolated, highly endemic former leprosy community located at the outskirts of the Brazilian Amazon. The
colony offers large multiplex, multigenerational pedigrees composed mainly by descendents of a small number of
original leprosy-affected families. Our enrollment strategy was complete ascertainment leading to the inclusion of
the whole colony (2005 individuals, 225 of whom were affected) distributed in 112 pedigrees. CSA was performed
using REGRESS software.
Results. CSA identified a best-fit codominant model, with a major gene accounting for the entire familial
effect observed. The frequency of predisposing allele was estimated at 0.22. Penetrance for homozygous individuals
for the predisposing allele 130 years old ranged from 56% to 85%, depending on sex.
Conclusions. A strong major gene effect in the isolated, hyperendemic Prata Colony indicates enrichment of
genetic risk factors, suggesting a population particularly suitable for leprosy gene identification studies.
Leprosy is a chronic infectious disease caused by Mycobacterium leprae that affects 250,000 new individuals
worldwide every year, with the majority of cases concentrated in India and Brazil [1]. On exposure, most
individuals develop efficient immunity against M. leprae
with no signs of clinical disease. However, in a small
proportion of exposed individuals, leprosy manifests in
a spectrum of clinical forms, ranging from the localized,
tuberculoid to the systemic, lepromatous disease [2],

Received 21 August 2009; accepted 3 December 2009; electronically published
13 April 2010.
a
F.P.L. and R.I.W. contributed equally to this work.
b
A.A. and M.T.M. share senior authorship.
Reprints or correspondence: Dr Marcelo T. Mira, Advanced Molecular
Investigation Core, Pontifical Catholic University of Paraná, Imaculada Conceição
1155, Prado Velho, Curitiba, Paraná, CEP 80215–901, Brazil (m.mira@pucpr.br).
The Journal of Infectious Diseases 2010; 201(10):1598–1605
  2010 by the Infectious Diseases Society of America. All rights reserved.
0022-1899/2010/20110-0021$15.00
DOI: 10.1086/652007

1598 • JID 2010:201 (15 May) • Lázaro et al

associated with a Th1 or Th2 type of immune response
presented against the pathogen, respectively [3]. Despite
a global leprosy elimination effort coordinated by the
World Health Organization since 1991, the disease persists in 118 countries, and Brazil, Nepal, and TimorLeste have still not achieved the elimination goal of a
prevalence rate !1 case per 10,000 persons [1].
The clinical outcome of infection is the result of interaction between variables related to the host, the path-

Potential conflicts of interest: none reported.
Financial support: UNICEF/UNDP/World Bank/WHO Special Program for Research
and Training in Tropical Diseases; Brazilian Conselho Nacional de Desenvolvimento
Cientı́fico e Tecnológico (CNPq) (projects 410470/2006-6 [MCT/CNPq/MS-SCTIE-DECIT
25/2006, Estudo de Doenças Negligenciadas] and 401057/2005-4 [MCT/CNPq/MSSCTIE-DECIT 35/2005, Estudo da Hansenı́ase]); Order of Malta Grants for Leprosy
Research—2008; INSERM-CNPq joint program (project 490844/2008-1 [ASCIN/CNPq
61/2008, Convênios Bilaterais Europa]); Rockefeller University Center for Clinical and
Translational Science (grant 5UL1RR024143-03 to Laboratory of Human Genetics of
Infectious Disease) and Rockefeller University (Laboratory of Human Genetics of
Infectious Disease); Agence Nationale de la Recherche, Ministère Français de
l’Éducation Nationale de la Recherche et de la Technologie (A.A., L.A.); CAPES/PROSUP
(scholarships to F.P.L. and R.I.W.); and CAPES/PDEE (scholarship to R.I.W.).

Downloaded from https://academic.oup.com/jid/article-abstract/201/10/1598/993909 by guest on 11 January 2020

A Major Gene Controls Leprosy Susceptibility
in a Hyperendemic Isolated Population
from North of Brazil

 mogeneously exposed to the disease. In addition, enrollment
strategies are always prone to ascertainment bias that is completely controlled only if the entire target population is included. For example, complete ascertainment was applied in 2
CSA studies performed for susceptibility to human T lymphotropic virus type 1 and human herpes virus type 8 in an isolated
hyperendemic population from French Guiana, with notable
results [33, 34]. In leprosy, the CSA conducted in 1988 in the
Desirade Island involved 1600 individuals recruited from an
area with disease frequency of ∼30 cases per 1000 persons, the
highest at the time [25]. To date, the Desirade study is the
closest to a theoretical ideal design for a CSA on leprosy, in
which the entire population from an isolated hyperendemic
area would be recruited.
Here we present the results of a CSA performed in a unique
collection of leprosy multiplex, multigenerational families corresponding to the complete population of the Colony of Santo
Antônio do Prata (the Prata Colony), located in the Amazonic
state of Pará, in the north of Brazil. The Prata is a former
leprosy colony created in the early 1920s to isolate leprosyaffected individuals. Isolation was compulsory until 1962; however, the population of the colony remains isolated, probably
owing to the strong stigma still associated with the disease.
Preliminary assessment indicated the highest disease prevalence
ever reported worldwide, homogenous environmental and socioeconomic variables, and a predominance of a mixed ethnic
group.
MATERIALS AND METHODS
Population recruitment. The entire population of the Prata
Colony was contacted by the research team over a period from
April 2006 to December 2007. The investigation was approved
by the Research Ethics Committee of the Pontifical Catholic
University of Paraná, the Brazilian National Board for Ethics
in Research, and the Ethics Research Committee of the World
Health Organization.
To assure that all households were contacted, visits were
planned according to a previously existing subdivision of the
colony into 6 sectors. All households in 1 sector were visited
before the entire research team moved on to the next sector.
The procedure was repeated until all sectors were systematically
included. All adult individuals were independently interviewed
by trained personnel. Information regarding individuals !18
years old was provided by the parents.
Epidemiological data collection and phenotype definition.
Demographic data (sex, age, ethnicity, and location of household), and affection status was collected for all individuals, who
were classified as affected or not by leprosy per se. Ethnicity
was defined by trained personnel based on the following phenotypic characteristics: skin color, hair type, and conformation
of the nose and lips [35]. Age and disease status were self-

Genetics of Leprosy in an Area of Hyperendemicity • JID 2010:201 (15 May) • 1599

Downloaded from https://academic.oup.com/jid/article-abstract/201/10/1598/993909 by guest on 11 January 2020

ogen, and the environment [4]. In this complex scenario, an
important role for host genetic risk factors controlling susceptibility to disease has become increasingly evident. For example,
the following have been implicated with leprosy phenotypes:
variants of HLA class I and II; HLA-linked genes MICA and
MICB [5, 6], TNFA [7, 8], and KIR [9], among others; and
non-HLA genes IL-10 [10], VDR [11], and SLC11A1 (formerly
NRAMP1) [12, 13]. In addition, 3 model-free genome-wide
linkage studies involving different leprosy phenotypes have
been performed, localizing loci harboring susceptibility genes
on chromosomal regions 10p13 [14], 17q11–q21 [15], 6q25–
q26, and 6p21 [16]. However, new leprosy candidate genes have
emerged only from chromosomes 6q25–q27 and 6p21: highdensity association mapping demonstrated susceptibility variants for leprosy per se (ie, the disease regardless its clinical
form) located on the shared regulatory region of the PARK2
and PACRG genes on chromosome 6q25–q27 [17, 18] and a
functional regulatory site of the LTA gene located on 6p21 [19].
In addition to molecular, DNA-based studies, classic observational genetic epidemiology tools, such as complex segregation analysis (CSA), have been used to advance our understanding of the genetic basis of complex diseases. Several CSA
studies have been performed in leprosy-affected population
samples of different ethnic background using a variety of analytical methods, with the objective to identify the best-fit
model of inheritance for leprosy phenotypes. Although a few
studies could not distinguish between environmental and genetic effects or suggested a predominant environmental effect
controlling susceptibility to disease [20, 21], more indicate the
existence of a recessive major gene (MG) controlling susceptibility to lepromatous disease [22], nonlepromatous disease
[23–25], and leprosy per se [25, 26].
Of note, the term “major gene” means that its effect is important enough to be distinguished from other genes effects
but does not assume it is the only gene involved. In contrast,
2 studies concluded in favor of the existence of a dominant or
codominant MG related to susceptibility to leprosy per se in
families from Thailand [27] and Vietnam [28]. More recently,
Shaw et al [7] described a best-fit model for susceptibility to
leprosy per se that included 2 loci, one a recessive MG and the
other a recessive modifier. The model was then used in parametric linkage followed by association analysis, using the same
family sample that confirmed the role of HLA class II and III
alleles, located on chromosome 6p21, in leprosy control. Such
a strategy (ie, CSA followed by model-based linkage analysis)
was also successful in identifying human genetic factors controlling other infectious diseases, as in the seminal works performed in schistosomiasis [29–31] and tuberculosis [32], highlighting the power of this approach.
Importantly, none of the other leprosy studies have been
performed using an isolated, hyperendemic population ho-

 Table 1. Segregation Analysis of Leprosy Per Se in the Prata Population
gPO
Model

Qa

Unaffected

aAA

aAa

aaa

(0)

⫺3.67

[aAA]

[aAA]

(0)

a. PO

(0)

⫺3.78

[aAA]

[aAA]

⫺0.44

b. SS

(0)

⫺3.85

[aAA]

[aAA]

(0)

c. PO + SS

(0)

⫺3.90

[aAA]

[aAA]

⫺0.23

⫺7.62

⫺5.38

0.29

⫺0.23

I. Sporadic

b sex

b
log
age

tAAa

tAaa

taaa

⫺2lnL
+C

gSS
Affected
(0)

Unaffected

Affected

(0)

(0)

⫺0.52

2.27

…

…

…

55

(0)

(0)

42

II. FD

III. MG and FD (PO + SS)

0.21

⫺0.51

2.33

…

…

…

⫺0.06

0.61

⫺0.51

2.24

…

…

…

46

0.56

⫺0.08

0.49

⫺0.48

2.31

…

…

…

33

0.28

⫺0.06

0.06

⫺0.66

4.50

0.62
(0)

(0)

(0.5)

(1)

7

a. Codominant

0.22

⫺7.98

⫺5.94

⫺0.68

(0)

(0)

(0)

(0)

⫺0.75

4.26

(0)

(0.5)

(1)

5

b. Recessive

0.27

⫺6.49

[⫺6.49]

⫺0.97

(0)

(0)

(0)

(0)

⫺0.72

3.82

(0)

(0.5)

(1)

14

V. Absence of transmission

0.06

⫺6.76

⫺6.75

⫺0.87

(0)

(0)

(0)

(0)

⫺0.69

4.32

0.68

[tAAA]

[tAAA]

27

VI. General transmission

0.15

⫺7.59

⫺5.60

⫺1.05

(0)

(0)

(0)

(0)

⫺0.64

4.13

0.00

0.31

1.00

0

NOTE. C p ⫺1015, corresponding to twice the logarithm of the likelihood (2lnL) of the best-fitting model (model VI); FD, familial dependency; MG, major
gene; PO, parent-offspring; Q frequency of leprosy predisposing allele a; SS, sibling-sibling; a, baseline risk of being affected on a logit scale corresponding to
3 genotypes: AA (aAA), Aa (aAa), and aa (aaa); b, covariable regression coefficients; gPO and gSS, regression coefficients associated with familial dependencies
PO and SS, respectively; tAAa, tAaa, and taaa, probabilities of transmitting a for individuals AA, Aa, and aa, respectively. Terms in brackets represent parameters
fixed to the same value as the preceding estimated parameter; terms in parentheses, fixed parameters for hypothesis; ellipses, irrelevant parameters in the
model.

reported at the interview. For self-reported affected individuals,
both affected status and age were confirmed on cross-checking
using 3 independent sources available at the local health care
center: the patient’s medical record, a copy of the compulsory
notification form, and a registry book used for treatment follow-up.
Pedigree reconstruction. To allow for pedigree reconstruction, parental information was also obtained for all individuals.
Pedigree drawing was performed using Cranefoot software (version 3.2.2) [36]. Information regarding members of informative pedigrees who were not living in the colony (moved or deceased) and therefore not personally interviewed was used on
confirmation of epidemiological and clinical status, as described
above. If disease status could not be checked, these individuals
were coded as “unknown.” One very large pedigree was decomposed, due to computational limitations, according to systematic criteria that reduced the loss of information [37], before
being included in the CSA. If necessary to link nuclear pedigrees, individuals were duplicated or created (“dummies”), a
procedure that did not affect the outcome of either the epidemiological description or the CSA.
Statistical methods. The phenotype of interest for the CSA
was a binary trait—that is, affected or not affected by leprosy.
Age was coded as the natural logarithm of age in years, which
was the best-fitting age function based on Akaike’s information
criterion [38]. Ethnicity was coded as a categorical variable with
3 classes: white (reference class), black, and mixed race. The
analyses were performed using logistic regression analysis as
implemented in the LOGISTIC procedure of SAS software, version 9.1 (SAS Institute).
Segregation analysis was done by using the regressive logistic
1600 • JID 2010:201 (15 May) • Lázaro et al

model [39], which specified a regression relationship between
the probability that a person would be affected and a set of
explanatory variables, including MG, phenotype of preceding
relatives, and other covariates. The regressive model allowed
analysis of large families as a whole, simultaneous estimation
of genetic and risk factor effects, and consideration of different
patterns of familial correlations for leprosy status. Sporadic
transmission (Table 1, model I) includes only the nongenetic
covariates with a significant effect on disease susceptibility (for
example, the question whether the phenotype of an individual is influenced by those of other family members is not
addressed).
In addition to the significant covariates, familial correlations
(Table 1, model II) are included in the model, using the class
D pattern of familial dependencies (FD) [39]. Four types of
phenotypic FD were considered: father-mother (FM), fatheroffspring (FO), mother-offspring (MO), and sibling-sibling (SS),
with corresponding regression coefficients denoted as G FM,
G FO, G MO, and G SS, respectively. To account for the phenotype
“unaffected,” each G  parameter is a vector of 2 coefficients [39]:
G FM p gFMunaf and gFMaff, G FO p gFOunaf and gFOaff,
G MO p gMOunaf and gMOaff, and G SS p gSSunaf and gSSaff.
gXunaf and gXaff can be interpreted as the impact on the risk
that an individual will develop leprosy conditional on the observation that he or she has a relative of type X affected and
unaffected, respectively. The higher the parameter, the higher the
increase in the risk of developing leprosy and symmetrically.
The presence of significant FDs can be due to any source of
unmeasured shared environmental factor (eg, higher exposure
to the microbe due to the geographic location of the family).
In particular, this should be suspected in the presence of sig-

Downloaded from https://academic.oup.com/jid/article-abstract/201/10/1598/993909 by guest on 11 January 2020

IV. MG

 nificant spouse-spouse dependencies. To rule out this possibility, an MG effect is included in the model (Table 1, model
III). The estimated parameters of the MG are Q (the frequency
of allele a predisposing someone to be affected by leprosy) and
aAA, aAa, and aaa (the 3 baseline risks of being affected on the
logit scale for the 3 genotypes AA, Aa, and aa, respectively).
Several studies [40–42] have shown that the identification
of an MG effect (ie, a mix of 3 distributions, 1 per genotype,
explains the data better than a single distribution) was not
specific enough to demonstrate the existence of an MG. They
have suggested that 2 additional tests were needed. These 2 tests
rely on the parent-offspring (PO) transmission pattern of the
major effect. The PO transmission is parameterized in terms
of the 3 classic transmission probabilities, as defined by Elston
and Stewart [43]: tAAa, tAaa, and taaa, which denote the
probabilities of transmitting a for individuals AA, Aa, and aa,
respectively [43]. Mendelian transmission is obtained by setting
taaa p 1, tAaa p 0.5, and tAAa p 0; in this case the major
effect is actually an MG.
Two additional models including a major effect were considered: (1) an “absence of transmission” model (model IV) in
which 3 types of individuals (aa, Aa, and AA) are specified but
in which absence of PO transmission is obtained by setting
taaa p tAaa p tAAa; and (2) a more general transmission
model (model V) in which the 3 t values are estimated. Segregation of an MG can be inferred if we fail to reject mendelian

transmission of the major effect in comparison with the general
transmission model and we reject the nontransmission hypothesis, also in comparison with the general transmission
model (this latter test rules out the possibility that the failure
to reject mendelian transmission of the major effect was due
to lack of power). Parameter estimation and hypothesis testing
were performed using classic likelihood strategies. Because of
the study design (ie, exhaustive collection of the whole population), there was no need for ascertainment correction. CSA
was performed as implemented in REGRESS software, version
4.8 [44], which incorporates the regressive approach into the
LINKAGE package [45].
RESULTS
Population description. Of the 2007 individuals invited, only
2 declined to participate; therefore, a total of 2005 individuals
(1012 male and 993 female) signed an informed consent and
were enrolled in the study. Leprosy prevalence was 12.82% (257
confirmed cases among 2005 individuals), distributed equally
throughout the colony (Figure 1). For the CSA, after removal
of single subjects and uninformative families, a total of 1867
individuals (225 affected and 1642 unaffected) were included,
distributed in 112 pedigrees. Therefore, the observed prevalence
in the CSA families was 12.05%, very close to the global prevalence. All clinical forms of disease were represented, with 104

Genetics of Leprosy in an Area of Hyperendemicity • JID 2010:201 (15 May) • 1601

Downloaded from https://academic.oup.com/jid/article-abstract/201/10/1598/993909 by guest on 11 January 2020

Figure 1. Distribution of households, all individuals, and leprosy-affected individuals across the 6 sectors of the colony. Values inside the cells
correspond to the exact percentages in each sector of the colony. An apparent increase in leprosy cases in sector 3 is due to the location in this
sector of the permanent shelter for elderly patients with leprosy who have no relatives living in the colony.

 DISCUSSION
Even though several initiatives have successfully identified genes
associated with host susceptibility to leprosy phenotypes [5,
1602 • JID 2010:201 (15 May) • Lázaro et al

Table 2. Factors Influencing the Onset of Leprosy in Univariate
and Multivariate Logistic Regression Analysis

Covariate

P
Sample
No. (%) of
Univariate Multivariate
size,
a
affected subjects analysis
analysis
no.
!10⫺4

Age, years
0–20

1019

21–40
41–60

520
233

160
Sex

Male
Female
Ethnicity
White
Mixed race
Black

!10⫺4

28 (2.7)
54 (10.38)
86 (36.9)

95

57 (60)

933
934

129 (13.8)
96 (10.2)

187
1451
229

20 (10.6)
161 (11)
44 (19.2)

.01

.002

.002

.365

.08b
.01c

.57
.67

a

Samples represent only informative individuals included in the complex
segregation analysis.
b
Comparison between mixed-race and white subjects.
c
Comparison between black and white subjects.

46], basic answers regarding the genetic model involved are still
not precisely known, as reflected by discordant results of several
CSA studies [7, 20–28]. These discrepancies may be due to the
tremendously difficult task of accounting for all environmental,
socioeconomic and cultural variables involved, as well as limitations regarding population enrollment. For example, the
large number of individuals required for a powerful CSA often
imply population sample heterogeneity and ascertainment bias,
unless the population is exposed to very high disease risk and/
or is entirely included in the analysis.
The objective of this study was to conduct a CSA using data
obtained from the entire population of the Colony of Santo
Antônio do Prata, a former leprosy colony with unique characteristics, including very high disease frequency with equal
distribution across the community (indicating homogenous exposure), a high degree of isolation, and homogenous socioeconomic, environmental, and ethnic backgrounds. We hypothesize
that, because the colony was founded almost exclusively by leprosy-affected individuals, genetic risk factors are enriched within
the community, composed today of a large number of extended,
multiplex, multigenerational families. In this context, the Prata
CSA detected the existence of an MG controlling susceptibility
to leprosy per se, inherited after a codominant model with the
frequency of the predisposing allele a estimated at 0.22.
Our results are not far from a recessive model, as observed
in a previous CSA using a population from the Desirade Islands
[25]—small discrepancies may be due to ascertainment correction procedures adopted in the Desirade and not necessary
in the Prata study—and such recessive-like models may be
specific to isolated population that descended from inhabitants

Downloaded from https://academic.oup.com/jid/article-abstract/201/10/1598/993909 by guest on 11 January 2020

(40.5%) of the cases being lepromatous, 53 (20.6%) tuberculoid, 53 (20.6%) borderline, and 47 (18.3%) of indeterminate
clinical form. The mean and median ages at diagnosis were
27.2 and 25 years, respectively.
Analysis of covariates. In the univariate analysis, age had
a strong effect on disease status (Table 2). The distribution of
disease according to age class shows an increase in the proportion of leprosy-affected individuals in each age class as age
increases up to 160 years old. Males were affected more often
than females (13.8% vs 10.2%, respectively). The distribution
of leprosy across ethnicity shows a higher proportion of affected among black subjects (19.2%) than among those who were
white (10.6%) or mixed race (11%) (P p .002). Multivariate
logistic regression analysis confirmed the strong effect of age
(P ! 10⫺4) and sex (P p .002); however, the ethnicity effect became not significant, because it was due to an age-confounding effect.
CSA results. Results of the CSA are shown in Table 1. Because we never observed significant FM correlation, these results were not included in the table. Moreover, in the different
analyses we never observed a significant difference between FO
and MO dependency; therefore, only a global PO dependency
(model IIa) (gFOunaf/aff p gMOunaf/aff p gPOunaf/aff) was considered.
There was evidence of a strong FD, because the sporadic
model without FD was rejected against the model that included
PO plus SS correlation (model I vs IIc, x2(2df) p 22; P p
10⫺5). When the model with SS correlation was compared with
the model with PO plus SS correlation, the first model was
rejected (model IIb vs IIc, x2(2df) p 13; P p .0015). Of note,
the absence of FM correlation is a good indicator that the
significant SS and PO correlations are not caused by an unmeasured environmental factor shared by the family members.
The inclusion of a codominant major effect to PO plus SS FDs
resulted in a highly significantly better fit (model IIc vs III,
x2(3df) p 26; P p 10⫺5). Interestingly, removal of residual PO
plus SS dependency did not significantly affect the fitness
(model III vs IVa, x2(4df) p 2; P p .73). The codominant
model was significantly better than a recessive model (IVb vs
IVa, x2(1df) p 9; P p .0027). Finally, the transmission of the
codominant major effect was compatible with the mendelian
hypothesis (IVa vs VI; x2(3df) p 5; P p .17), and the hypothesis
of no transmission was rejected (V vs VI, x2(2df) p 27; P p
10⫺6); In summary, the CSA favored a codominant MG influencing leprosy per se. Under this codominant MG model (aAA
( aAa ( aaa), the frequency of the predisposing allele a was
estimated at 0.22. The same analysis was performed using the
mixed population only, with similar results (data not shown).

 dominant gene with a very strong effect and several codominant
genes with milder effects that play additively on the risk. For
example, an exciting possibility is that the MG effect observed
is largely due to the PARK2/PACRG [17, 18] and LTA [19]
effects, described elsewhere in different populations. Alternatively—and not less exciting—new major susceptibility genes
are yet to be described. For further investigation, molecular,
DNA-based studies are necessary.
It is widely accepted that genetic control of the pathogenesis
of leprosy follows a 2-step model, in which different genes affect
susceptibility to leprosy per se and clinical manifestation of disease [49]. Our study focused only on the phenotype for leprosy
per se, because of 2 limitations: (1) the distribution of clinical
forms of disease must be interpreted as an approximate estimate, given the impossibility of confirming this very complex
phenotype retrospectively, based exclusively on heterogeneous
clinical information recorded by different physicians over a
time period that spans almost 100 years; and (2) the smaller
number of cases in each clinical class significantly decreases the
power of the analysis. Therefore, the MG effect detected here
is likely to modify the onset of leprosy independent of its clinical form [5].
The results of the Prata CSA support the hypothesis of an
enrichment of genetic risk factors in this particular population,
making it ideal for future gene mapping analysis. Parameters
of the genetic model generated by the CSA may be used for

Figure 2. Penetrance (ie, probability of being leprosy affected) according to age, genotype, and sex, as predicted in model IVa (Table 2); a is the
leprosy-predisposing allele. Exact penetrance values at age 30 years are as follows: females, 0.56 for aa and 0.005 for Aa; males, 0.85 for aa and
0.024 for Aa. Penetrance for homozygous AA individuals at age 30 years was 0.003 for males and 0.0006 for females (not shown).
Genetics of Leprosy in an Area of Hyperendemicity • JID 2010:201 (15 May) • 1603

Downloaded from https://academic.oup.com/jid/article-abstract/201/10/1598/993909 by guest on 11 January 2020

isolated because they had leprosy. Under this codominant, recessive-like model, ∼5% of the population is aa homozygous
and therefore highly predisposed to leprosy. As shown in Figure
2, in addition to the strong effect of the underlying genotype,
disease penetrance was also influenced by sex and age; as an
example, at age 30 years, penetrance ranged from 0.85 for aa
men to 0.0006 for AA women.
A previous, classic study of a population from Malawi demonstrated an increased risk for leprosy among dwelling contacts
of a leprosy case patient. Disease risk decreased for household
contacts and was the lowest for no contacts, leading the authors
to conclude that there was an environmental or behavioral
cause for familial aggregation [47]. In the Prata colony, the MG
effect was detected in the presence of FD; however, when FD
was removed from the model, the MG effect became better
noticed, indicating that the observed familial aggregation was
entirely explained by genetics, as first suggested by the absence
of FM correlation. Interestingly, the median age at diagnosis
of leprosy in the Prata (25 years) is lower than the Brazilian
national median of 39 years [48]. Although early diagnosis of
cases in a hyperendemic population under constant monitoring
is somewhat expected, this observation is compatible with recent findings suggesting that early leprosy cases are more likely
to have a genetic basis [19]. Unfortunately, more precise inference about the nature of the MG effect observed is limited
by the inability of CSA to distinguish between 1 single, co-

 12.
13.

14.

15.

16.

17.
18.
19.

Acknowledgments
We thank all affected individuals and their families for agreeing to participate in the study. We also thank the local team from the Federal University of Pará, as well as the leaders and health workers of the Prata Colony
for their help in guiding our team through the community, socially and
geographically. A special acknowledgment is addressed to Dr Erwin Schurr
for his critical reading and valuable suggestions during the development
of the study and preparation of the manuscript.

20.
21.
22.
23.
24.

References
1. World Health Organization. Global leprosy situation. Wkly Epidemiol
Rec 2008; 83:293–300.
2. Ridley DS, Jopling WH. Classification of leprosy according to immunity: a five-group system. Int J Lepr Other Mycobact Dis 1966; 34:
255–273.
3. Scollard DM, Adams LB, Gillis TP, Krahenbuhl JL, Truman RW, Williams DL. The continuing challenges of leprosy. Clin Microbiol Rev
2006; 19:338–381.
4. Casanova JL, Abel L. The human model: a genetic dissection of immunity to infection in natural conditions. Nat Rev Immunol 2004; 4:
55–66.
5. Mira MT. Genetic host resistance and susceptibility to leprosy. Microbes
Infect 2006; 8:1124–1131.
6. Tosh K, Ravikumar M, Bell JT, Meisner S, Hill AV, Pitchappan R.
Variation in MICA and MICB genes and enhanced susceptibility to
paucibacillary leprosy in South India. Hum Mol Genet 2006; 15:2880–
2887.
7. Shaw MA, Donaldson IJ, Collins A, et al. Association and linkage of
leprosy phenotypes with HLA class II and tumour necrosis factor genes.
Genes Immun 2001; 2:196–204.
8. Santos AR, Suffys PN, Vanderborght PR, et al. Role of tumor necrosis
factor-alpha and interleukin-10 promoter gene polymorphisms in leprosy. J Infect Dis 2002; 186:1687–1691.
9. Franceschi DS, Mazini PS, Rudnick CC, et al. Association between
killer-cell immunoglobulin-like receptor genotypes and leprosy in Brazil. Tissue Antigens 2008; 72:478–482.
10. Pereira AC, Brito-de-Souza VN, Cardoso CC, et al. Genetic, epidemiological and biological analysis of interleukin-10 promoter singlenucleotide polymorphisms suggests a definitive role for ⫺819C/T in
leprosy susceptibility. Genes Immun 2009; 10:174–180.
11. Fitness J, Floyd S, Warndorff DK, et al. Large-scale candidate gene

1604 • JID 2010:201 (15 May) • Lázaro et al

25.

26.

27.

28.
29.

30.

31.

32.

33.

34.

study of leprosy susceptibility in the Karonga district of northern Malawi. Am J Trop Med Hyg 2004; 71:330–340.
Abel L, Sánchez FO, Oberti J, et al. Susceptibility to leprosy is linked
to the human NRAMP1 gene. J Infect Dis 1998; 177:133–145.
Alcaı̈s A, Sanchez FO, Thuc NV, et al. Granulomatous reaction to intradermal injection of lepromin (Mitsuda reaction) is linked to the human NRAMP1 gene in Vietnamese leprosy sibships. J Infect Dis 2000;
181:302–308.
Siddiqui MR, Meisner S, Tosh K, et al. A major susceptibility locus for
leprosy in India maps to chromosome 10p13. Nat Genet 2001; 27:439–
441.
Miller EN, Jamieson SE, Joberty C, et al. Genome-wide scans for leprosy
and tuberculosis susceptibility genes in Brazilians. Genes Immun 2004;
5:63–67.
Mira MT, Alcaı̈s A, Van Thuc N, et al. Chromosome 6q25 is linked to
susceptibility to leprosy in a Vietnamese population. Nat Genet 2003;
33:412–415.
Mira MT, Alcaı̈s A, Nguyen VT, et al. Susceptibility to leprosy is associated with PARK2 and PACRG. Nature 2004; 427:636–640.
Alcaı̈s A, Mira M, Casanova JL, Schurr E, Abel L. Genetic dissection
of immunity in leprosy. Curr Opin Immunol 2005; 17:44–48.
Alcaı̈s A, Alter A, Antoni G, et al. Stepwise replication identifies a lowproducing lymphotoxin-alpha allele as a major risk factor for earlyonset leprosy. Nat Genet 2007; 39:517–522.
Shields ED, Russell DA, Pericak-Vance MA. Genetic epidemiology of
the susceptibility to leprosy. J Clin Invest 1987; 79:1139–1143.
Serjeantson S, Wilson SR, Keats BJ. The genetics of leprosy. Ann Hum
Biol 1979; 6:375–393.
Smith DG. The genetic hypothesis for susceptibility to lepromatous
leprosy. Hum Genet 1979; 50:163–177.
Haile RW, Iselius L, Fine PE, Morton NE. Segregation and linkage
analysis of 72 leprosy pedigrees. Hum Hered 1985; 35:43–52.
Demenais F, Goulet V, Feingold N, et al. Genetic study of leprosy in
a Caribbean island: Desirade. In: Skamene E, ed. Genetic control of
host resistance to infection and malignancy. New York: Alan R. Liss,
1985:319–324.
Abel L, Demenais F. Detection of MGs for susceptibility to leprosy and
its subtypes in a Caribbean island: Desirade island. Am J Hum Genet
1988; 42:256–266.
Feitosa MF, Borecki I, Krieger H, Beiguelman B, Rao DC. The genetic
epidemiology of leprosy in a Brazilian population. Am J Hum Genet
1995; 56:1179–1185.
Wagener DK, Schauf V, Nelson KE, Scollard D, Brown A, Smith T.
Segregation analysis of leprosy in families of northern Thailand. Genet
Epidemiol 1988; 5:95–105.
Abel L, Vu DL, Oberti J, et al. Complex segregation analysis of leprosy
in southern Vietnam. Genet Epidemiol 1995; 12:63–82.
Abel L, Demenais F, Prata A, Souza AE, Dessein A. Evidence for the
segregation of a major gene in human susceptibility/resistance to infection by Schistosoma mansoni. Am J Hum Genet 1991; 48:959–970.
Marquet S, Abel L, Hillaire D, et al. Genetic localization of a locus
controlling the intensity of infection by Schistosoma mansoni on chromosome 5q31-q33. Nat Genet 1996; 14:181–184.
Dessein AJ, Hillaire D, Elwali NE, et al. Severe hepatic fibrosis in
Schistosoma mansoni infection is controlled by a major locus that is
closely linked to the interferon-gamma receptor gene. Am J Hum Genet
1999; 65:709–721.
Greenwood CM, Fujiwara TM, Boothroyd LJ, et al. Linkage of tuberculosis to chromosome 2q35 loci, including NRAMP1, in a large aboriginal Canadian family. Am J Hum Genet 2000; 67:405–416.
Plancoulaine S, Gessain A, Joubert M, et al. Detection of a major gene
predisposing to human T lymphotropic virus type I infection in children among an endemic population of African origin. J Infect Dis
2000; 182:405–412.
Plancoulaine S, Gessain A, van Beveren M, Tortevoye P, Abel L. Evidence for a recessive major gene predisposing to human herpesvirus

Downloaded from https://academic.oup.com/jid/article-abstract/201/10/1598/993909 by guest on 11 January 2020

model-based linkage analysis followed by high-density association mapping using the already-mapped Prata families. Specifying the nature of the genetic component controlling susceptibility to leprosy may have a profound impact on the development of new strategies for prevention, diagnosis, and treatment of the disease. For example, individuals genetically at risk
and living under high exposure would be candidates for chemoprophylaxis, a highly debated issue today in Brazil and other
countries of endemicity [50]. Genetic susceptibility could also
help in the understanding of issues such as disease relapse and
drug resistance, the mechanisms of which are still widely unknown in leprosy. Finally, a deeper insight into the molecular
basis of leprosy pathogenesis would certainly improve the understanding not only of this particular disease but of infection
in general.

 35.

36.

37.

38.

40.

41.

42.

43.
44.

45.
46.

47.

48.

49.

50.

formulations of the logistic regressive models with the mixed model in
segregation analysis of discrete traits. Genet Epidemiol 1992; 9:419–435.
Elston RC, Stewart J. A general model for the genetic analysis of pedigree data. Hum Hered 1971; 21:523–542.
Demenais F, Lathrop GM. REGRESS: a computer program including
the regressive approach into the LINKAGE programs. Genet Epidemiol
1994; 11:291.
Lathrop GM, Lalouel JM, Julier C, Ott J. Strategies for multilocus linkage
analysis in humans. Proc Natl Acad Sci USA 1984; 81:3443–3446.
Alter A, Alcaı̈s A, Abel L, Schurr E. Leprosy as a genetic model for susceptibility to common infectious diseases. Hum Genet 2008; 123:227–
235.
Fine PE, Sterne JA, Pönnighaus JM, et al. Household and dwelling contact as risk factors for leprosy in northern Malawi. Am J Epidemiol 1997;
146:91–102.
Ministério da Saúde. Hansenı́ase. http://portal.saude.gov.br/portal/
saude/Gestor/visualizar_texto.cfm?idtxtp27445. Accessed 21 August
2009.
de Vries RR, Fat RF, Nijenhuis LE, van Rood JJ. HLA-linked genetic
control of host response to Mycobacterium leprae. Lancet 1976; 2:1328–
1330.
Goulart IM, Goulart LR. Leprosy: diagnostic and control challenges
for a worldwide disease. Arch Dermatol Res 2008; 300:269–290.

Genetics of Leprosy in an Area of Hyperendemicity • JID 2010:201 (15 May) • 1605

Downloaded from https://academic.oup.com/jid/article-abstract/201/10/1598/993909 by guest on 11 January 2020

39.

8 (HHV-8) infection in a population in which HHV-8 is endemic. J
Infect Dis 2003; 187:1944–1950.
Krieger H, Morton NE, Mi MP, Azevêdo E, Freire-Maia A, Yasuda N.
Racial admixture in north-eastern Brazil. Ann Hum Genet 1965; 29:
113–125.
Mäkinen VP, Parkkonen M, Wessman M, Groop PH, Kanninen T, Kaski
K. High-throughput pedigree drawing. Eur J Hum Genet 2005; 13:987–
989.
Liu F, Kirichenko A, Axenovich TI, van Duijn CM, Aulchenko YS. An
approach for cutting large and complex pedigrees for linkage analysis.
Eur J Hum Genet 2008; 16:854–860.
Akaike H. A new look at statistical model identification. IEEE Trans
Autom Control 1974; 19:716–723.
Bonney GE. Regressive logistic models for familial disease and other
binary traits. Biometrics 1986; 42:611–625.
Demenais F, Lathrop M, Lalouel JM. Robustness and power of the
unified model in the analysis of quantitative measurements. Am J Hum
Genet 1986; 38:228–234.
Demenais F, Abel L. Robustness of the unified model to shared environmental effects in the analysis of dichotomous traits. Genet Epidemiol 1989; 6:229–234.
Demenais FM, Laing AE, Bonney GE. Numerical comparisons of two