Skip Navigation

A Novel Single-Nucleotide Polymorphism in the Lactoferrin Gene Is Associated with Susceptibility to Diarrhea in North American Travelers to Mexico

  1. Jamal A. Mohamed1,
  2. Herbert L. DuPont1,3,4,
  3. Zhi Dong Jiang3,
  4. Jaime Belkind-Gerson5,
  5. Jose Flores Figueroa1,
  6. Lisa Y. Armitige1,
  7. An Tsai1,
  8. Parvathy Nair1,
  9. Francisco J. Martinez-Sandoval6,
  10. Dong-chuan Guo2,
  11. Patricia Hayes1, and
  12. Pablo C. Okhuysen1,3
  1. 1Division of Infectious Diseases, Houston, Texas
  2. 2Division of Medical Genetics, The University of Texas Medical School, Houston, Texas
  3. 3Center for Infectious Diseases, School of Public Health, Houston, Texas
  4. 4St. Luke's Episcopal Hospital, Houston, Texas
  5. 5Instituto Nacional de Salud Publica, Cuernavaca, Morelos
  6. 6Universidad Autonoma de Guadalajara, Guadalajara, Mexico
  1. Reprints or correspondence: Dr. Pablo C. Okhuysen, Div. of Infectious Diseases, The University of Texas Medical School, 6431 Fannin St., MSB 2.112, Houston, TX 77030 (Pablo.c.okhuysen{at}uth.tmc.edu).
  1. Presented in part: 44th Annual Meeting of Infectious Diseases Society of America, Toronto, Canada, 12–15 October 2006 (abstract 63).

Abstract

Background. Diarrhea affects 40%–60% of travelers from industrialized nations who visit developing countries and is due to bacterial, viral, and parasitic agents. Lactoferrin is bactericidal to enteric pathogens, modulates the intestinal immune response, and is excreted in stool in response to infection with intestinal organisms. We investigated the impact that selected single-nucleotide polymorphisms (SNPs) in the human lactoferrin gene have on susceptibility to traveler's diarrhea.

Methods. Adults who had recently arrived in Mexico were studied prospectively for the occurrence and causal agent(s) of traveler's diarrhea, and genotyping was performed for 9 distinct lactoferrin SNPs.

Results. Of the 9 SNPs studied, only 1 SNP (located in exon 15) was associated with traveler's diarrhea (P = .004). When compared with healthy travelers, and after adjustment for known risk factors for traveler's diarrhea (such as age and duration and season of travel), subjects with the T/T genotype in amino acid position 632 were more likely to develop traveler's diarrhea (67% vs. 33%; relative risk [RR], 1.4; 95% CI, 1.2–1.7; P < .001), to have diarrhea with a pathogen identified (RR, 1.3; 95% CI, 1.1–1.6; P = .03), and to have a marker of intestinal inflammation in stool specimens (blood, mucus, or white blood cells; 52% vs. 38%; P = .036). The association was also significant when norovirus was not identified in stool samples (RR, 1.34; 95% CI, 1.06–1.34; P = .01).

Conclusions. The T/T genotype in position codon 632 of the lactoferrin gene is associated with susceptibility to diarrhea in North Americans traveling to Mexico.

It is estimated that 40%–60% of US travelers to Mexico develop diarrheal illness during short periods of travel [1]. Bacterial pathogens are the causal agents in up to 85% of cases. Enterotoxigenic Escherichia coli, which produces heat-stable and/or heat-labile enterotoxins, and enteroaggregative E. coli account for the majority of the cases, whereas Shigella, Campylobacter, and Salmonella species and parasites are identified less frequently. Recently, norovirus has been identified in ∼15% of cases [2]. Lactoferrin is a glycoprotein secreted in human milk and mucosal secretions and in the secondary granules of neutrophils. When found in stool specimens obtained from travelers with diarrhea, lactoferrin is a marker of infection with agents that induce an inflammatory response. Fecal lactoferrin has been shown to be elevated in patients with Clostridium difficile–associated colitis [3] and shigellosis, in 26% of patients with enterotoxigenic E. coli infection, and in 60% of patients with enteroaggregative E. coli infection [46], as well as in patients with diarrhea due to enteropathogenic E. coli or to Campylobacter and Cryptosporidium species [7]. Levels are absent or low in patients with diarrhea due to norovirus [8].

In addition to iron binding and its role in iron homeostasis [9], lactoferrin has antibiotic [10], anti-inflammatory [11], and immunoregulatory properties [12, 13]. Lactoferrin has been shown to prevent endotoxic shock associated with parenteral E. coli infection in piglets [14], to interfere with adhesion of enteroaggregative E. coli to epithelial cell lines [15], and to inactivate the type III secretory system of enteropathogenic E. coli [16]. Lactoferrin also protect rabbits from the inflammatory effects of Shigella species [17] and can also block the binding of LPS to CD14 [16, 18], decreasing the inflammatory response.

Traveler's diarrhea is acquired through the ingestion of contaminated food and water. Risk factors include the region of the world visited [1], the age of the traveler, the site of food preparation [19, 20], excessive alcohol consumption [21], and the ingestion of risky food items. It has long been suspected that host genetics determine susceptibility to illness. Our research group previously demonstrated that presence of a single-nucleotide polymorphism (SNP) in the -251 position of the IL-8 gene promoter is associated with diarrhea due to enteroaggregative E. coli [22], and others have found associations with SNPs in the FUT2 gene and susceptibility to norovirus [23]. In this study, we sought to determine whether SNPs in the lactoferrin gene were associated with an increased risk of traveler's diarrhea among North American visitors to Mexico.

Subjects and Methods

Human subjects. Adult North American students aged ⩾18 years or adolescents aged >16 years with parental consent who attended language programs in Guadalajara and Cuernavaca, Mexico, during 2002–2005 participated in the study. Students were eligible if they were otherwise in good health and were not eligible if they had preexisting known irritable bowel syndrome or lactose intolerance or were taking medications to prevent traveler's diarrhea. Students were enrolled within 1 week after their arrival in Mexico and were required to have a duration of stay of ⩾5 days. Subjects were observed prospectively for the occurrence, duration, and etiology of diarrhea while in Mexico. This study was approved by the University of Texas Committee for the Protection of Human Subjects (Houston).

Definitions. Acute diarrhea was defined as the passage of ⩾3 unformed stools within a 24-h period plus ⩾1 symptom of enteric infection (abdominal pain, cramping, excessive gas, nausea, vomiting, fever, urgency, bloody or mucous stool, or tenesmus).

Genotyping by PCR and pyrosequencing. DNA was purified from blood specimens (Gentra Systems), and SNPs in the lactoferrin gene were determined by DNA pyrosequencing. PCR and pyrosequencing primers were designed using Primer3 program (available at http://fokker.wi.mit.edu/primer3) and SNP Primer Design software (Biotage AB and Biosystems), respectively. Primers, sequences, and locations of the SNPs are listed in table 1. PCR was performed using HotStar Taq DNA polymerase (Qiagen). SNP detection was performed by pyrosequencing (PSQ 96; Biotage).

Figure 1

Human lactoferrin concentrations in fecal samples obtained from individuals with SNP, LTFEx15 codon 632 (T/C), homozygous for T/T (n = 40), heterozygous for T/C (n = 25), and homozygous for C/C (n = 7), as measured by ELISA. Values are means ± SD. Human lactoferrin standards ranging from 6.25 to 100 µg/mL were used to generate a standard curve. The lactoferrin concentration in a test sample was determined by plotting absorbance values versus standard lactoferrin concentrations. NS, not significant

Table 1

Primer sequences for PCR and pyrosequencing and single-nucleotide polymorphisms (SNPs) of the human lactoferrin gene.

Microbiologic testing and fecal lactoferrin levels. Stool samples were obtained from subjects with diarrhea and were transported directly to the laboratory or in Cary-Blair transport media. Stool samples were studied for enteric pathogens as previously described [24]. Five E. coli colonies were studied for the presence of heat-stable and heat-labile toxins by use of oligonucleotide probes [25]. Enteroaggregative E. coli was identified by HEp-2 adherence assay [26]. Noroviruses were identified by RT-PCR [27]. Fecal lactoferrin content was measured by ELISA (IBD-SCAN; Techlab). Stool samples were evaluated for the presence of mucus, fecal leukocytes, and occult blood by conventional methods [6].

Statistical analysis. Univariate and multivariate analyses were performed using SPSS software, version 14 (SPSS). The associations between clinical, microbiological, and genotypic data were determined by χ2 analysis, and linear and nonlinear outcomes were compared using SPSS software, version 14. For multiple comparisons, Bonferroni correction was applied, and data sets related to genotypes and alleles were considered independently.

Results

Seven hundred sixty-two students were enrolled during the 4-year study period and had ⩾1 genotype in the lactoferrin gene determined (table 1). Nine SNPs were studied. Only 1 SNP (LTFEx15), which is in exon 15, was found to be associated with diarrhea (table 2). This SNP (rs 7645243), which codes for a transition of T to C, causes a corresponding synonymous amino acid change at codon 632 from TTG (leucine) to CTG (leucine). The genetic variation at position LTFEx15 codon 632 (T/C) showed a T allele frequency of 79% and a C allele frequency of 21%. The allele frequencies for the T and C alleles were 67.2% and 32.8%, respectively, for healthy control subjects and 74.9% and 25.1%, respectively, for persons with diarrhea (P = .0016). The relative genotype frequencies were found to be in Hardy-Weinberg equilibrium. Genetic and clinical information were available for 718 of the students whose samples were genotyped at codon 632. As shown in table 3, female students accounted for 68% of the participants. Six hundred fifty-six students (91%) were white, 34 (5%) were black, 19 (3%) were Asian, and 9 (1%) were of other origins. Among the participants, 106 (15%) were Hispanic. The mean age of the participants was 30.7 years, and the mean length of stay in Mexico was 25.9 days.

Table 2

Human lactoferrin single-nucleotide polymorphisms (SNPs) studied and their association with diarrhea in North American travelers to Mexico.

Table 3

Univariate logistic regression analysis of risk factors for the acquisition of diarrhea in North American travelers to Mexico in subjects whose samples were genotyped for LTFEx15 codon 632 (T/C) in the lactoferrin gene.

In our study, 50% of the students developed diarrhea. We examined the potential for confounding variables to contribute to the effect of the LTFEx15 SNP on diarrhea (table 3). In univariate analysis, male subjects were 12% more likely to report traveler's diarrhea in their diaries than were female subjects. Younger age, longer duration of stay, and summertime travel (51% for subjects who traveled during June–August, compared with 25% for those who traveled during nonsummer months; relative risk [RR], 1.52; 95% CI, 1.37–2.04; P = .04) had significant influences on rates of diarrhea. A multivariate analysis demonstrated that male sex (RR, 1.48; 95% CI, 1.07–2.05; P = .02), age <25 years (RR, 1.74; 95% CI, 1.25–2.43; P = .001), season of travel of summer (RR, 4.04; 95% CI, 1.25–13.03; P = .02), and presence of the LTFEx15 SNP T/T genotype (RR, 1.9; 95% CI, 1.138–3.505; P = .016) were independently associated with diarrhea.

To eliminate other potential confounders, we examined the relative distribution of the different alleles for this SNP among the various racial groups. Substantial differences in the distributions of the genotypes were noted. Among white subjects, 7.8% belonged to the C/C genotype group, 39.5% belonged to the C/T genotype group, and 52.7% belonged to the T/T genotype group. In contrast, 17.6% of black subjects were in the C/C genotype group, 55.9% were in the C/T genotype group, and 26.9 were in the T/T genotype group (P = .005 for white vs. black race, and P < .001 for white race vs. all other racial groups). Despite differences in the genotype distributions in the racial composition of our study, subjects with the T/T genotype demonstrated an increased susceptibility to diarrhea as a group, regardless of race (RR, 1.28; 95% CI, 1.10–1.48; P = .001). We then conducted an analysis in a subgroup of 397 white travelers who were at highest risk for diarrhea (i.e., those who had traveled for >2 weeks during the summer season and who were <50 years of age) (table 4). In this subgroup, 228 subjects (57%) experienced traveler's diarrhea, 158 (69%) of whom provided a stool for microbiologic examination (table 4). At least 1 pathogen was identified in 103 subjects (65%). Bacterial cases consisted mostly of enterotoxigenic E. coli (48%), with heat-labile, heat-stable and heat-labile/heat-stable strains reported equally (18%, 17%, and 11%, respectively). Enteroaggregative E. coli were found alone or with other agents in 29% of subjects. Norovirus was identified in 13 (33%) of 39 samples, and Cryptosporidium species were found in 16 (15%) of the 103 subjects.

Table 4

Clinical outcomes in a subgroup of travelers at risk for diarrhea, according to LTFEx15 codon 632 genotype.

When the effect of the LTFEx15 T/T genotype was compared with the T/C plus C/C genotypes in the subgroup of 397 subjects, the associations between T/T genotype and the study outcomes were also significant (table 5). Subjects homozygous for the T/T genotype were more likely to experience diarrhea than were subjects who were not of the T/T genotype (67 vs. 47%; P < .001; corrected P = .001) and to have diarrhea with an agent identified in their stool specimens (30 vs. 21%; P < .001; corrected P = .003). Of interest, subjects with stool samples that tested negative for norovirus were also more likely to have the T/T genotype (8 vs. 2%; P < .001; corrected P = .01). Because diarrhea was more common among male subjects, we sought to determine the effect of the subjects' sex. Despite an increased frequency of diarrhea among male subjects, the associations of the LTFEx15 codon 632 (T/C) in the exon 15 T/T genotype were significant only in female participants (63% vs. 42%; P < .001; corrected P = .01).

Table 5

Clinical outcomes among white participants in the study and results of microbiological evaluations of stool specimens submitted for analysis, according to the LTFEx15 codon 632 T/T versus non-T/T genotype.

We observed a relationship between the LTFEx15 T/T genotype and fecal inflammatory markers. Nineteen (65%) of 29 subjects with diarrhea and fecal leukocytes in their stool specimen had the T/T genotype, compared with 64 (38%) of 169 healthy control subjects (P = .005; corrected P = .08). Similarly, 23 (64%) of 36 subjects with mucus in their stool samples had the T/T genotype (P = .004 [corrected P = .07], when compared with healthy control subjects). No differences were noted in the proportions of subjects with the T/T genotype among individuals with fecal occult blood present (6 [67%] of 9; P = .08; corrected P = not significant), compared with healthy control subjects. In aggregate, an overrepresentation of the T/T genotype was seen in subjects with diarrhea and any of the 3 markers of inflammation (48 [52%] of 92 vs. 64 [38%] of 169 subjects; P = .002; corrected P = .036).

Mean fecal lactoferrin levels (±SD) were significantly higher in subjects with pathogen-positive diarrhea than in those without a pathogen identified (43.2 ± 47 vs. 83.5 ± 91.8 µg/mL; P = .02, by analysis of variance). Fecal lactoferrin levels in the stool specimens obtained from patients in the C/C group ranged from 0.34 to 49 µg/mL (mean ± SD, 20 ± 16 µg/mL), and for patients in the C/T group, the mean concentration (±SD) was 93.0 ± 88.22 µg/mL (range, 0–277 µg/mL). Among subjects in the T/T group, the mean concentration (±SD) was 65.57 ± 19.19 µg/mL (range, 0–276 µg/mL). Although a trend toward higher fecal lactoferrin levels was noted for the T/C group, compared with the C/C group (P = .029), there were no statistically significant differences between the C/C and T/T groups or between the T/C and T/T groups (figure 1).

Discussion

Many of the ∼1.42 million SNPs [28] in the human genome have important implications in human disease, and the impact of SNP intestinal infections are now becoming evident. We choose to study SNPs in the lactoferrin gene, because this is an important element of the intestinal innate immune system, and SNPs in various regions of this gene are associated with cancer [29], periodontitis [30], and corneal amyloidosis [31].

In this study, North American visitors to Mexico with traveler's diarrhea—specifically, those in whom diarrhea was due to an enteric pathogen or who had markers of intestinal inflammation—were more likely to carry the LTFEx15 632 T/T genotype. The association of this SNP with diarrhea remained significant after adjusting for confounding variables, such as age, race, and duration of travel, in a multivariate analysis and after using a conservative approach by correcting for multiple comparisons (i.e., the Bonferroni correction).

The rates of diarrhea were higher among male than female students. It has been hypothesized that male subjects are at a higher risk of acquiring traveler's diarrhea because of consumption of larger amounts of food (and, therefore, a larger inoculum of enteric pathogens) or because of other risk factors, such as alcohol consumption [21], which may lower the threshold for the consumption of risk food items. Because of this, we studied the effect of the SNP with the subjects' sex and the association with diarrhea. Unexpectedly, an association between the T/T genotype and diarrhea was found to be significant only in female students. When examined for known confounders, such as length of stay, age, and duration of travel, female and male subjects shared the same risks and were as likely to have an enteric pathogen identified in their stool specimens. The molecular basis for this sex-specific association with the T/T genotype is unknown but may relate to the fact that lactoferrin expression is regulated differently in females than males, owning to the well-characterized estrogen-responsive elements in the DNA sequence [32], which could be relevant for female-specific roles of lactoferrin, such as during lactation or reproduction.

The association of the LTFEx15 T/T genotype and diarrhea may reflect the fact that production of lactoferrin forms part of a stereotypical response to enteric pathogens. Given our sample size, we were unable to establish associations with specific pathogens, except for subjects who had negative results of RT-PCR assays for norovirus, consistent with the lack of lactoferrin response to viral pathogens [33].

Although fecal lactoferrin levels were higher in samples obtained from subjects with pathogens present, levels correlated poorly with the SNP genotype. A trend toward a lower level of lactoferrin was found for the C/C genotype than for the T/C, but not when compared with the T/T genotype. Of note, the location of the LTFEx15 codon 632 (T/C) SNP is not in the region that codes for the bactericidal activity of lactoferrin or its subcomponent lactoferricidin. The lack of an association with quantitative differences in lactoferrin production could be associated with a dilution effect or degradation during processing. It is also possible that a functional genomic polymorphism that is in linkage disequilibrium with this SNP alters lactoferrin level or function, because the LTF Ex15 SNP is located in a 21-kb linkage disequilibrium block.

The LTFEx15 632 SNP codes for a leucine to leucine substitution. A number of studies have identified synonymous SNPs of human relevance [3436]. Such SNPs can alter biological function at the mRNA or protein level. This may occur by creating mRNA errors, altering its stability and expression [37, 38] or affecting the splicing of mRNA. Furthermore, at the protein level, there may be differences in translational efficiency associated codon use bias [39, 40]. Uncertainty remains on the exact mechanisms underlying the function of many synonymous substitutions, including the one reported herein, and additional studies are warranted.

Our estimate may be an underrepresentation of this effect, because subjects who may have considered themselves at risk for traveler's diarrhea from past experience or family history may have elected to take prophylactic measurers and not to take part in our study. This study has other limitations. First, not all travelers with diarrhea provided a stool specimen. We also found a high frequency of coinfection with bacterial, viral, and parasitic agents. These confounders may limit the power to correlate the relationship of the different genotypes with individual agents of diarrhea. Second, we did not include a control group that could be monitored for asymptomatic colonization to further confirm the association of the LTFEx15 632 genotypes with enteric infection. Third, the exposure to agents of diarrhea may have not been uniform, because travelers demonstrate variable degrees of adherence to the avoidance of risk food items. The microbiologic assessment of all food items consumed by travelers would have been impractical and cost-prohibitive.

The contribution that this SNP has on the risk of diarrhea in travelers was modest. However, this may be an underestimation of the true risk, because most of the cases of traveler's diarrhea in this study were due to noninvasive organisms, and this SNP may be of higher relevance in other regions of the world where invasive organisms are more common. Nevertheless, it is likely that, in addition to unidentified environmental and immune factors, other genes contribute to the risk of acquiring traveler's diarrhea. A larger study investigating other SNPs in genes that mediate susceptibility, injury, and healing in traveler's diarrhea are underway by our group.

Acknowledgments

We are indebted to Dorothy Ruelas, Judy Guillen, Margaret DuPont, Lily Carlin, Jackie Vaca, Stephanie A. Lee, and the administration and staff of Universidad Internacional in Cuernavaca, Morelos, Mexico, for their assistance with this project.

Financial support. National Institutes of Health (ULIRR024148; R01 AI54948-01 and M01 RR002558 to the General Clinical Research Center of the University of Texas Medical School at Houston, and DK56338, which funds the Texas Gulf Coast Digestive Diseases Center).

Potential conflicts of interest. All authors: no conflicts.

  • Received November 17, 2006.
  • Accepted December 23, 2006.

References

| Table of Contents