Gene-specific sex effects on eosinophil infiltration in leishmaniasis

Background Sex influences susceptibility to many infectious diseases, including some manifestations of leishmaniasis. The disease is caused by parasites that enter to the skin and can spread to the lymph nodes, spleen, liver, bone marrow, and sometimes lungs. Parasites induce host defenses including cell infiltration, leading to protective or ineffective inflammation. These responses are often influenced by host genotype and sex. We analyzed the role of sex in the impact of specific gene loci on eosinophil infiltration and its functional relevance. Methods We studied the genetic control of infiltration of eosinophils into the inguinal lymph nodes after 8 weeks of Leishmania major infection using mouse strains BALB/c, STS, and recombinant congenic strains CcS-1,-3,-4,-5,-7,-9,-11,-12,-15,-16,-18, and -20, each of which contains a different random set of 12.5% genes from the parental “donor” strain STS and 87.5% genes from the “background” strain BALB/c. Numbers of eosinophils were counted in hematoxylin-eosin-stained sections of the inguinal lymph nodes under a light microscope. Parasite load was determined using PCR-ELISA. Results The lymph nodes of resistant STS and susceptible BALB/c mice contained very low and intermediate numbers of eosinophils, respectively. Unexpectedly, eosinophil infiltration in strain CcS-9 exceeded that in BALB/c and STS and was higher in males than in females. We searched for genes controlling high eosinophil infiltration in CcS-9 mice by linkage analysis in F2 hybrids between BALB/c and CcS-9 and detected four loci controlling eosinophil numbers. Lmr14 (chromosome 2) and Lmr25 (chromosome 5) operate independently from other genes (main effects). Lmr14 functions only in males, the effect of Lmr25 is sex independent. Lmr15 (chromosome 11) and Lmr26 (chromosome 9) operate in cooperation (non-additive interaction) with each other. This interaction was significant in males only, but sex-marker interaction was not significant. Eosinophil infiltration was positively correlated with parasite load in lymph nodes of F2 hybrids in males, but not in females. Conclusions We demonstrated a strong influence of sex on numbers of eosinophils in the lymph nodes after L. major infection and present the first identification of sex-dependent autosomal loci controlling eosinophilic infiltration. The positive correlation between eosinophil infiltration and parasite load in males suggests that this sex-dependent eosinophilic infiltration reflects ineffective inflammation.


Background
Sex influences susceptibility to many infectious diseases [1], including some manifestations of leishmaniasis [2], a disease that threatens several hundred million people in 98 countries [3]. Disability-adjusted life years (DALYs) due to leishmaniasis are globally increasing [4]. The disease is caused by intracellular protozoan parasites of the genus Leishmania and is transmitted to the vertebrates by the bite of female phlebotomine sand flies.
Leishmania parasites infect the so-called professional phagocytes (neutrophils, monocytes, and macrophages), as well as dendritic cells and fibroblasts. The major host cell is the macrophage, where parasites multiply, eventually rupturing the cell and spread to the uninfected cells (reviewed in [5]). Infected monocytes and macrophages circulating in the peripheral blood are believed to be carriers of the parasite to distal sites [6]. In the dermis, parasites cause the cutaneous form of the disease (which can be localized or diffuse), whereas infection of the mucosa gives rise to mucocutaneous leishmaniasis. The metastatic spread of the infection to the spleen and liver results in visceral leishmaniasis. Although these are the major sites of visceral disease, parasites can also enter other organs, such as the bone marrow, lymph nodes, and lungs (reviewed in [5]). Presence of parasites in organs usually induces inflammation through cascade of signals that leads to recruitment of inflammatory cells, such as neutrophils, macrophages, eosinophils, and dendritic cells. These innate immune cells might phagocytose parasites and/or produce cytokines and chemokines that activate both innate and adaptive immune responses. Resulting responses can be protective and eliminate parasites, or ineffective and lead to chronic inflammation [7].
In general, sex bias is observed after infection with Leishmania parasites, and men are more frequently infected than women ( [13][14][15]; reviewed in [11,12]), although in certain areas no sex bias in prevalence of disease was observed [16]. The higher susceptibility of males also applies to hamster [17] and mouse [18,19]; reviewed in [12] models of leishmaniasis. The effect of male orchidectomy and female testosterone replacement studies suggests that the hormone testosterone can modulate systemic L. major infection in BALB/cAnPt, DBA/2N, DBA/2J, and F 1 hybrids (BALB/cAnPt x DBA/2N) mouse strains [18].
Importantly, the host genes, including those regulated differently in males and females, play a significant role in determining susceptibility and organ tropism for infectious diseases. Experimental data have shown different sex influence on susceptibility to relatively closely related pathogen species [20,21], different sex biases in susceptibility to the same Leishmania species in different host genotypes [21,22], and different sex and genetic influence on organ-specific pathology [21,23,24]. For example, high resistance to skin lesions induced by L. mexicana was observed in females but not in males of DBA/2 mice, but the sex effect was opposite in L. major infection [20].
Genotype influence on sex differences was defined in the studies of L. major infection [22,24]. Giannini [22] found no sex effect on L. major-induced skin pathology and mortality in BALB/cJ mice, but a higher susceptibility of B10.129(10M)ScSn females than males. The comparison of L. major susceptibility in two strains, BALB/cHeA and CcS-11 [24], has shown that there is no significant sex influence on skin lesion development, splenomegaly, and hepatomegaly in these strains. However, parasite numbers in lymph nodes are higher in both BALB/c and CcS-11 males; moreover, CcS-11 males have higher parasite load in spleens, showing an organ-specific, sex-, and genotype-dependent pathology [24].
In the present study, we address influence of genotype and sex on infiltration of eosinophil leukocytes into the inguinal lymph nodes of L. major-infected mice. Eosinophils are granulocytes that develop in the bone marrow from pluripotent progenitors. They are released into the peripheral blood in phenotypically mature state and can be activated and recruited into tissues in response to appropriate stimuli, most notable IL-5, and the eotaxin chemokines [25].
Eosinophils contribute to the initiation of inflammatory and adaptive responses due to their bidirectional interactions with dendritic cells and T cells, as well as their large spectrum of secreted cytokines and soluble mediators. They have key immunoregulatory roles as professional antigen-presenting cells and modulators of functions of CD4 + T cells, dendritic cells, B cells, mast cells, neutrophils, and basophils [26].
Eosinophil-associated disorders can affect practically all tissues and organs in the body, either individually or in combination. They are involved in inflammatory conditions affecting the skin, cardiovascular, nervous and renal system, gastrointestinal tract, and upper and lower airways [27,28], are key effector cells in eosinophilic asthma [29], and their interaction with peripheral nerves has impact on pathology of many diseases. In addition, they are also involved in regulatory mechanisms modulating local and systemic immune responses and remodeling and repair mechanisms [30].
Eosinophils may have an important role in maintaining host survival in life-threatening viral infections [31]. They combat worms such as Angiostrongylus cantonensis [32], Nippostrongylus brasiliensis [33], Litomosoides sigmodontis [34], and Brugia pahangi [35]; but their role in response to other nematoda is more complex. Eosinophils have no role in protection against Schistosoma mansoni [36]. They even promote larval growth in primary infection with Trichinella spiralis [37], but they mediate protective immunity against secondary infection with this nematode [38].
Activated eosinophils can kill [39] or support killing of L. major parasites [40]; however, in chronic disease, eosinophil infiltration might be a consequence of an ineffective elimination of these parasites and/or an excessive inflammatory response to the present pathogens [41].
Here, we analyzed genetic influence on eosinophil infiltration after L. major infection into the lymph nodes of strains BALB/cHeA (BALB/c), STS/A (STS), and selected 12 (out of 20) RC strains of CcS/Dem series [42]. Each of the 20 RC CcS/Dem strains contains a different unique set of approximately 12.5% genes of the donor strain STS on the genetic background of BALB/c. We found surprisingly high numbers of eosinophils in the inguinal lymph nodes of the strain CcS-9, males containing higher numbers of eosinophils than females. We analyzed genetics of this infiltration using microsatellite DNA markers and mapped four loci that control eosinophil numbers after L. major infection, one of them being strongly influenced by sex. We also found that the numbers of eosinophils in the lymph nodes correlate positively with the parasite load and that this correlation is partly genetically controlled and is higher in males than in females.
A linkage study of eosinophil infiltration: F 2 hybrids between CcS-9 and BALB/c (age 11 to 21 weeks at the time of infection, mean and median age 14.8 and 15 weeks, respectively) were produced at the Institute of Molecular Genetics. When used for these experiments, the CcS-9 was in the 40th generation of inbreeding and therefore highly homozygous. Two hundred fifty-four F 2 hybrids between BALB/c and CcS-9 comprised 139 females and 115 males. Mice of the background parental strains BALB/c (18 females, 17 males) and STS (8 females, 6 males) and the RC strain CcS-9 (16 females, 14 males), 7 to 20 weeks old at the time of infection (mean 13 weeks, median 13 weeks), were used as controls. During the experiment, male and female mice were placed into separate rooms and males were caged individually. F 2 mice were tested in three independent experimental groups.

Ethical statement
All experimental procedures in this study comply with the Czech Government  Amastigotes were transformed to promastigotes using SNB-9 [43]. 10 7 promastigotes from the passage, two cultivated for 6 days were inoculated in 50 μl sterile saline s.c. into mouse rump [44].

Disease phenotype
The size of the primary skin lesion was measured weekly using a Vernier caliper gauge. The mice were killed 8 weeks after infection and inguinal lymph nodes draining the site of infection were collected for further analysis.

Histological analysis
Inguinal lymph nodes of female and male mice were fixed in 10% neutral buffered formalin (NBF; approximately 4% formaldehyde) and embedded in paraffin using automatic tissue processor. Tissue sections (5-7 μm) were stained with hematoxylin, differentiated into 1% acid alcohol, stained with 1% alcoholic eosin, dehydrated, assembled with permanent mounting medium, and analyzed under a light microscope (Olympus BX51; Olympus Optical Co. (EUROPA) GMBH., Hamburg, Germany).
In F 2 mice, as well as the parental strains BALB, STS, and CcS-9, eosinophil numbers were determined quantitatively. The total number of eosinophils was counted in the node section and each lymph node was assessed in four independent sections. The mean value of these four counts was used to calculate the role of genetic factors in control of eosinophil infiltration. Sixty slides from 15 mice were blindly recounted by an independent investigator with concordant results (R = 0.913, P value = 5.66 × 10 −29 ).
Genotyping of F 2 mice by PCR DNA was isolated from tails using a standard proteinase procedure. The strain CcS-9 differs from BALB/c at STS-derived segments on eight chromosomes ( [45] and unpublished results). These differential segments were typed in the F 2 hybrid mice between CcS-9 and BALB/c using 18 microsatellite markers (Research Genetics, Huntsville, FL, USA): D2Mit283, D2Mit148, D4Mit172, D4Mit23, D4Mit53, D4Mit17, D5Mit24, D5Mit143, D6Mit122, D6Mit274, D9Mit15, D11Mit141, D11Mit242, D11Nds18, D11Nds10, D16Mit19, D17Mit120, and D17Mit122. The markers were selected because their genomic location makes them suitable to detect linkage. The maximum distance between any two markers in the chromosomal segments derived from the strain STS or from the nearest BALB/c derived markers was 12.46 cM, and mean distance was 4.67 cM. The PCR genotyping for markers with fragment length difference more than 8 bp was performed using unlabeled primers as in [46,47]. The PCR genotyping for markers with fragment length difference less than 8 bp was performed using [γ-32 P]ATP end-labeled primers as described elsewhere [48].

Measurement of parasite load in lymph nodes
Total DNA was isolated from the frozen lymph nodes, and parasite load was measured using PCR-ELISA according to the previously published protocol [49]. Briefly, for detection of Leishmania parasite DNA, in total DNA, PCR was performed using two primers (digoxigenin-labeled F 5′-ATT TTA CAC CAA CCC CCA GTT-3′ and biotin-labeled R 5′-GTG GGG GAG GGG CGT TCT-3′ (VBC Genomics Biosciences Research, Austria). The 120-bp fragment within the conserved region of the kinetoplast minicircle of Leishmania parasite was amplified. In each PCR reaction, 50 ng of extracted total DNA was used. As a positive control, 20 ng of L. major DNA per reaction was amplified as a highest concentration of the standard. A 26-cycle PCR reaction was used for quantification of parasites. Parasite load was determined by measurement of the PCR product with the modified ELISA protocol (Pharmingen, San Diego, USA). The concentration of Leishmania DNA was measured at the ELISA Reader Tecan with the curve fitter program KIM-E (Schoeller Pharma, Prague, Czech Republic) using least squares-based linear regression analysis [24,49].

Statistical analysis
The differences among BALB/c, STS, and CcS/Dem strains in eosinophil numbers in lymph nodes were evaluated by the analysis of variance (ANOVA) and Newman-Keuls multiple comparison test at 95% significance using the program Statistica for Windows 12.0 (StatSoft, Inc., Tulsa, OK, USA).
The role of genetic factors in control of eosinophil infiltration in F 2 hybrids was examined by ANOVA (Statistica for Windows 12.0; StatSoft, Inc., Tulsa, OK, USA). In order to obtain normal distribution of the analyzed parameter required for ANOVA, the obtained values were transformed as shown in the legends of tables. Markers and interactions with P < 0.05 were combined in a single comparison. In all ANOVA analyses strain or genotype, sex, and age were fixed factors, and the experiment was considered a random parameter.
These studies showed mild and no infiltration into the lymph nodes of parental strains BALB/c (Fig. 1a, b) and STS (Fig. 1c, d), respectively. Strains CcS-9 (P = 0.00020) (Fig. 1e, f ) and CcS-12 (P = 0.0024) exhibit significantly higher eosinophil infiltration in their lymph nodes than the background parental strain BALB/c. BALB/c and CcS-9 males presented higher eosinophil infiltration than females of these strains P = 0.0089 and P = 0.016, respectively. 80% of examined CcS-9 males in comparison with 26.67% of CcS-9 females contained infiltrating eosinophils, 50% of males having 7 and more eosinophils in their lymph nodes (Table 1). Sex difference in strains CcS-7,-11, and -18 was not significant. Strain CcS-9 with the highest eosinophil infiltration (Table 1) was selected for further genetic studies.

Four novel loci control eosinophil infiltration in leishmaniasis
We examined eosinophil numbers in lymph nodes in 254 F 2 hybrids between the strains BALB/c and CcS-9. The strain CcS-9 differs from BALB/c at STS-derived genetic regions located at eight chromosomes ( [45], Šíma unpublished data). These differential STS-derived segments were genotyped in the F 2 hybrid mice using 18 microsatellite markers. A statistical analysis of linkage revealed four genetic loci that influence eosinophil infiltration into the inguinal lymph nodes after L. major infection.  Eosinophil numbers in lymph nodes depending on genotype and sex. Eosinophil infiltration was evaluated as described in the "Methods" section. Numbers higher than 75% are shown in italics The effects of Lmr14 (L. major response 14) linked to D2Mit283 (corrected P value = 0.0081) and Lmr25 linked to D5Mit143 (corrected P value = 0.044) were detectable and significantly independent of each other or other genes (the main effects) ( Table 2). Lmr14 operated only in males (corr. P value of marker and sex interaction =0.0085)( Table 2, Fig. 2), higher numbers of eosinophils were associated with presence of BALB/c (C) allele (Fig. 2c). The P value for Lmr14 was significant only in cross (CcS-9 × BALB/c)F 2 (where the mother of the F 1 hybrids was CcS-9 and the father was BALB/c) (Fig. 2c), but not in cross (BALB/c × CcS-9)F 2 (where the mother was BALB/c and the father was CcS-9) (Fig. 2d). However, interaction between the cross and marker D2Mit283 was not significant (corr. P = 0.6). The effect of Lmr25 was not influenced by sex, and higher numbers of eosinophils were observed in heterozygotes ( Table 2).
In contrast to the main effects of Lmr14 and Lmr25, Lmr15 (linked to D11Nds10) and Lmr26 (linked to D9Mit15) operated in cooperation with each other (nonadditive, epistatic, interaction) (corrected P = 0.010). F 2 male mice of the cross (BALB/c × CcS-9)F 2 with homozygous BALB/c (CC) alleles at both Lmr26 and Lmr15 had nearly nine times higher numbers of eosinophils in the lymph nodes than mice with homozygous STS (SS) alleles at both these loci, and nearly 90 times higher than mice with homozygous CC alleles at Lmr26 and CS alleles at Lmr15 (Table 3). The linkage was detected only in males, but the interaction between sex and marker was not significant (corr. P = 0.19).

Positive correlation between parasite numbers and eosinophils in the inguinal lymph nodes
We have determined parasite load in the lymph nodes of the F 2 hybrids between BALB/c and CcS-9 and analyzed the relationship between parasite numbers in lymph nodes and eosinophil infiltration to this organ. In both sexes pooled, there was a positive correlation between parasite numbers and eosinophil infiltration R = 0.39, P = 1.3 × 10 −10 , and the correlation was significant in males R = 0.29, P = 0.0017, but not in females R = 0.14, P = 0.10. This correlation is at least partly controlled by Lmr loci, because in F 2 hybrid mice, this correlation was positive in male homozygous for the Lmr14 (D2Mit283) BALB/c allele (CC) (R = 0.51, P = 0.016) and STS allele (SS) (R = 0.50, P = 0.00088), but no correlation was observed in heterozygotes (R = −0.013, P = 0.92).

Eosinophil infiltration in strain CcS-9 exceeds that of both parents
Strain CcS-9 that contains a set of approximately 12.5% genes of the donor strain STS and 87.5% genes of the background strain BALB/c exhibited numbers of infiltrating eosinophils (Fig. 1, Table 1) exceeding those in both parental strains BALB/c and STS. The observations of progeny having a phenotype, which is beyond the range of the phenotype of its parents, are not rare in traits controlled by multiple genes. It was detected in different tests of immune responses of RC strains in vitro [51][52][53][54][55][56] and in vivo [21,[57][58][59][60], and in analysis of  Mean and SE values were obtained by analysis of variance. In order to obtain normal distribution required for analysis of variance, the value of eosinophil numbers in the inguinal lymph nodes was transformed by using the 0.1th power of natural logarithm of the (observed value ×1000). The numbers in bold give the average non-transformed values. C and S indicate the presence of BALB/c and STS allele, respectively n number of mice  Mean and SE values were obtained by analysis of variance. In order to obtain normal distribution required for analysis of variance value of eosinophil numbers in serum inguinal lymph nodes was transformed by using the 0.1th power of natural logarithm of the (observed value ×1000). The numbers in bold give the average non-transformed values. C and S indicate the presence of BALB/c and STS allele, respectively n number of mice some of them and eosinophil low infiltration alleles at others, and some progeny may receive predominantly eosinophil high infiltration alleles from both parents.

Sex influence on eosinophil infiltration
Our data show a sex influence on eosinophil numbers in the inguinal lymph nodes. Differences between the immune system of females and males have been well documented [62][63][64] and could result in differences in susceptibility to diseases with immune component. Immune responses including those involving eosinophils might be modulated by steroid hormones [65,66]. Moreover, some of the differences between females and males might be due to sexspecific genetic architecture, characterized by profound gene-sex interactions [67,68]. This would mean that some genes controlling response to L. major might operate differently in the two sexes. Indeed, locus Lmr14 controls eosinophil infiltration only in males. Genes controlling infections that appear to be sex dependent have been observed also with other infectious agents such as viruses [69][70][71], bacteria [72], parasites [58], and fungi [73] and helminths [74]. Some of sex-dependent QTLs exhibit a higher or exclusive influence on susceptibility in females [58,69,[71][72][73] or males [69,[71][72][73][74], phenotypic effect of other genes is present in both sexes, but with opposite direction of effect [69,70]. All these reported loci are situated on autosomal chromosomes. In contrast to the sex chromosomes, the autosomal genome is shared by both sexes. However, although the DNA sequence, gene structure, and frequency of polymorphism on the autosomes do not differ between males and females, the regulatory genome is sexually dimorphic [68].
Future genetic and functional studies will help to establish the mechanistic basis of the observed gene-sex interactions.

Loci controlling eosinophil infiltration and other immune traits
The Lmr loci influencing eosinophil infiltration may be related to QTLs that determine certain immunologically relevant traits, because they co-map with other immunological functional polymorphisms.
Interestingly, two of the eosinophil controlling loci, Lmr15 and Lmr26 co-localize with loci that determine hemopoietic cell cycling measured by cobblestone areaforming cell (CAFC) assay using cells from the bone marrow [75]. Lmr15 encompasses the mouse ortholog of human gene IL5, whose polymorphism was found to be associated with eosinophil counts in the blood [76], and Lmr26 co-localizes also with locus Tria5 that modifies in vitro proliferation of mouse splenocytes stimulated by soluble anti-CD3 [77].
The four described loci comprise several genes (Fig. 3), whose biological function is compatible with the effects on eosinophil infiltration [78][79][80][81][82][83][84][85][86][87][88][89][90] and their potential role can now be investigated. However, the effects of these Lmr loci might be also caused by genes that are at the present not considered as candidates. The issue of identity of eosinophil controlling genes and their possible relationship to other immune traits will be resolved by a recombinational analysis.
We have found positive correlation between eosinophil infiltration and parasite numbers in Lmr14 in homozygous (CC or SS), but not in heterozygous (CS) F 2 hybrid males. The lack of positive correlation between eosinophil infiltration and parasite load in Lmr14 heterozygotes (CS) may reflect a more effective inflammation process, perhaps facilitated by other phenotypic effects of Lmr14 that include circulating levels of IFNγ, TNF, IgE, and IL-12 [91] and possibly other as yet undetected regulatory effects. This possibility has to be tested in future experiments.

Conclusions
This is the first demonstration of genetic loci and sex influence controlling infiltration of eosinophils into the lymph nodes and its relationship with parasite load. Some of these loci comprise genes with broader biological and immunological effects, so they might be relevant also in control of other diseases and symptoms mediated by eosinophils.
Our data also suggest that ignoring sex in gene mapping might prevent detection of sex-dependent QTLs.