Hospital in Central Peru

Transcription

1 JCM Accepts, published online ahead of print on August 00 J. Clin. Microbiol. doi:./jcm Copyright 00, American Society for Microbiology and/or the Listed Authors/Institutions. All Rights Reserved. 1 Molecular Epidemiology of Brucella Genotypes in Patients at a Major Hospital in Central Peru Karsten Nöckler, 1 Ryan Maves, David Cepeda, Angelika Draeger, 1 Anne Mayer- Scholl, 1 Jesus Chacaltana, María Castañeda, Benjamin Espinosa, Rosa Castillo, Eric Hall, Sascha Al Dahouk, Robert H. Gilman, Franco Cabeza, and Henk L. Smits Federal Institute for Risk Assessment, Berlin, Germany 1 ; United States Naval Medical Research Center Detachment, Lima, Peru ; Hospital Nacional Daniel Alcides Carrión, Callao, Peru ; RWTH Aachen University, Aachen, Germany ; Johns Hopkins University, Baltimore, MD, USA ; Asociación Benéfica Proyectos en Informática, Salud, Medicina y Agricultura (A.B. PRISMA), Lima, Peru ; Royal Tropical Institute / Koninklijk Instituut voor de Tropen (KIT), Amsterdam, The Netherlands * Running title: Brucella MLVA genotypes from Peru *Corresponding author. Mailing address: KIT Biomedical Research, Royal Tropical Institute / Koninklijk Instituut voor de Tropen (KIT), Meibergdreef, 0 AZ Amsterdam, The Netherlands. Tel: 1 (0)0 0. Fax: 1 (0) h.smits@kit.nl. Web: 1 1

2 Multiple Locus Variable Number Repeats Analysis of 0 human B. melitensis isolates from a large urban area in central Peru revealed variations at five (Bruce0, 0,, 1 and ) out of sixteen loci investigated, of which one (Bruce) is also used for species identification. Ten genotypes were identified, separated by the number of Bruce repeats into two groups that may have distinct phenotypic characteristics. Whereas genotypes with five or six Bruce repeats were cultured mainly from adult patients, genotypes with three Bruce repeats were isolated from children and young adolescents as well as from adults. In addition, the isolates with three Bruce repeats were obtained more often from patients with splenomegaly (P=0.0) or hepatomegaly (P=0.00). An annual variation in the diversity of genotypes was observed, possibly reflecting changes in sources of fresh dairy products, supply routes to city shops and markets, and the movement of infected dairy goat herds. Brucellosis is endemic in Peru and continues to be transmitted to humans. The number of reported cases of human brucellosis has decreased in recent years, presumably due to ongoing vaccination campaigns of goats and the promotion of the pasteurization of dairy products (, ). Between and % of the human cases are reported from Callao, the harbour city of the capital Lima (1). Human brucellosis is a non-specific but serious febrile illness without distinctive signs or characteristic symptoms that may affect multiple organ systems and cause various debilitating complications (, 0). Treatment

3 consists of a combination of antibiotics (), most commonly doxycycline in combination with an aminoglycoside, rifampicin, or both. Brucellosis is caused by infection with bacteria of the genus Brucella. Several species are pathogenic to man, most notably Brucella melitensis, B. abortus and B. suis (1). In Peru, human brucellosis is almost exclusively caused by infection with B. melitensis, a species that principally infects goats and sheep. Based on their agglutinating properties with specific antisera, B. melitensis may be differentiated into three biovars, biotypes 1, and, of which biotype 1 is known to be present in Peru (1,, ). Recently, a highly discriminatory method for the genotyping of Brucella known as Multiple Locus Variable Number Repeats Analysis (MLVA) has become available (). This method makes use of various loci on the Brucella genome composed of repeats of short nucleotide sequences. These tandem repeat units tend to occur in varying numbers, and various alleles can be observed in different species and isolates. The recently published MLVA- assay developed for genotyping of Brucella makes use of eight minisatellite loci for species identification supplemented with a selection of eight more polymorphic microsatellite loci for further characterization and differentiation of isolates (, ). Whereas the MLVA- assay may be used for biovar classification in B. abortus and B. suis, no correlation between biovar and genotype has been observed for B. melitensis (, ). MLVA- typing of animal and human Brucella isolates has revealed that clusters of individual genotypes within a species may show a distinct geographic distribution. For instance, human isolates of B. melitensis from Europe and North Africa can be divided according to their geographic origin into a West and an East Mediterranean cluster.

4 Within the West Mediterranean cluster (which includes isolates from France, Switzerland, Tunisia and Algeria), a clearly separate cluster originating from Italy can be identified (). Genotypes are relatively stable, and isolates with identical MLVA patterns have been obtained from the same geographic area over a time-span of almost three decades (). A considerable number of distinct B. melitensis genotypes have already been identified (). MLVA typing additionally has some practical clinical applications, such as tracing sources of infections and discriminating relapse from re-infection (,, ). Recently, we showed that B. melitensis biotype 1 from Peru is clearly distinct from the East and West Mediterranean groups, forming a Latin American cluster that is most closely related to two previously genotyped isolates from Mexico (). Here, we investigate the genotypes of a series of consecutive human Brucella isolates cultured from patients hospitalized at a major hospital of Callao. MATERIALS AND METHODS Ethical considerations. Ethical approval was obtained from the ethical committees of the United States Naval Medical Research Center and from the Hospital Nacional Daniel Alcides Carrión. Written informed consent was obtained from all patients, their parents, or guardians. Patients and clinical samples. All patients hospitalized between November 00 and April 00 at the Hospital Nacional Daniel Alcides Carrión in Callao with a Brucella positive blood culture were entered in the study. All patients were Rose Bengal positive

5 with reactive serum agglutination tests (median titer, 1:00; range, 1:0 to 1:1,00). During this.-year period, 1 isolates were cultured from patients. Two patients experienced second disease episodes, and two isolates each were obtained from these patients. Patients were classified according to the duration of illness in acute (< months of illness), subacute (-1 months of illness) and chronic (>1 months of illness). Genotyping. Genotyping was performed using MLVA- panel 1 for species identification (Bruce0, 0,, 1,,, and ) and MLVA- panels A (Bruce1, 1, 1) and B (Bruce0, 0, 0, and 0) for further subspecies differentiation (). Primer pairs and PCR conditions were as described by Le Flèche and coworkers (). PCR products were separated by electrophoresis on 1. to % agarose gel. For each run, DNA from the B. melitensis M reference strain was included. In anticipation of the expected tandem repeat unit length, a 0-bp or -bp molecular marker ladder (hyper ladder no. IV and V, Bioline, Berlin, Germany) was used. Ethidium bromide stained gels were visualized by UV light and graphically evaluated (GeneGenius from Syngene, Cambridge, UK). Genotyping was performed for all but one isolate. The MLVA- patterns from the Peruvian isolates were compared with the public database Brucella00 ( in order to identify the most closely related isolates. The cluster analysis was performed by unweighted pair group method with an arithmetic mean (UPGMA) algorithm, and a rooted tree was generated. In this study, we define a genotype as an isolate with a distinct MLVA- pattern. The distance between two genotypes is defined as the minimum number of changes in the number of repeats of any loci that converts one genotype to the other.

6 Statistical analysis. Chi-square analysis was used to correlate patient characteristics with genotype RESULTS Patient characteristics, disease presentation and risk-factors. The mean age of the patients was years (range 1 to ), and the male (n=) to female (n=) ratio was 0. (Table 1). Male patients (mean age,. years; range, 1-) were slightly older than female patients (mean age,. years; range, -1). Most (.%) patients presented with acute disease. All patients presented with one or more significant signs and/or symptoms, of which fever, malaise, sweats, headache, hyporexia, shaking chills, and arthralgia were each observed in over 0% of the patients. None of the patients presented with focal disease or complications requiring intervention other than antibiotic treatment. Most (.%) patients came from Callao, with other patients coming from different districts of greater Lima. Almost all (.1%) patients reported the consumption of unpasteurized dairy products. As expected for patients from an urban area, contact with livestock was rare. Conserved and polymorphic tandem repeat loci. A total of 0 isolates could be typed by MLVA- and patterns obtained for all 0 isolates were consistent with B. melitensis. The results revealed that the panel 1 locus Bruce, the panel A locus Bruce1, and the panel B loci Bruce0 and Bruce0 were polymorphic (Table a). Isolates with three, five and six repeats were detected for panel 1 locus Bruce, isolates

7 with seven and eight repeats for panel A locus Bruce1, isolates with five and six repeats for panel B locus Bruce0, and isolates with four, five, six and seven repeats for panel B locus Bruce0. All isolates were monomorphic for the panel 1 loci Bruce0 ( repeats), 0 ( repeats), ( repeats), 1 (1 repeats), ( repeats) ( repeats), and ( repeats), as well as for the panel loci Bruce0 ( repeats), ( repeats), 1 (1 repeats), 1 ( repeats), and 0 ( repeats). In total, three unique and seven genotypes each comprising two to 0 isolates could be distinguished. Four genotypes together constituted 0% of the isolates: genotype G1 with 1 isolates, G with 0 isolates, and genotypes G and G with six isolates each. Genotypes G1, G, and G were identical to three out of fifteen previously described human B. melitensis genotypes isolated in 000 and 001 from Peruvians also residing in Callao and Lima (). All other genotypes found in this study differed from the fifteen previously characterized Peruvian B. melitensis genotypes, which were polymorphic for Bruce0, 0, and (Table b). The MLVA- results are illustrated in figure 1, showing the length of the PCR products for the four polymorphic loci Bruce0, 0, 1 and for each of the ten genotypes of Brucella isolates obtained during the period Phylogenetic relationship of Peruvian B. melitensis genotypes. Of all published non-peruvian isolates characterized by MLVA- genotyping, the Brucella isolate key bru1 showed the closest relationship with a maximum distance of to with the ten Peruvian genotypes isolated in the period (). According to database information, bru1 is a B. melitensis biovar 1 isolate from Mexico. Furthermore, genotype G closely matched with the Brucella isolate key bru1 (distance of ), a B.

8 melitensis biovar 1 isolate from Spain. The relationship of these ten Peruvian genotypes with bru1 is depicted in the dendrogram presented in Figure A. Bruce variation and the relationship with patient characteristics and clinical presentation. According to the dendrogram, the ten genotypes separate into two branches with branch 1 including genotypes G1 (No. of type strain, 1), G (No. ), G (No. ), G (No. ), G (No. 0), and G (No. ), and branch including genotypes G (No. ), G (No. ), G (No. 0), and G (No. ). Isolates with a genotype grouped in branch were isolated throughout the study period, whereas isolates comprising branch 1 were isolated mainly in the period from 00 to 00 (Table ). No differences were observed between the spectrum of genotypes isolated from males and females and the ratio of branch 1 and branch genotypes was similar for the two sexes (Table ). Consistent with the larger number of patients from Callao, the diversity of genotypes derived from patients from this city was higher than from those from Lima (Table ). Interestingly, a clear difference was observed between the spectrum of genotypes isolated from children and young adolescence and those isolated from adults (Table ). Very few branch 1 genotypes were obtained for children and adolescence compared with adults. Twenty-six branch 1 and branch genotypes were isolated from adult patients (age 0 years) compared with four branch 1 and branch genotypes isolated from children and adolescent. The difference was significant (Relative Risk,.; % Confidence Interval, 1.-.: P=0.00). Branch 1 genotypes differ from branch genotypes in the number of Bruce repeats. Branch 1 genotypes are characterized by the presence of five or six Bruce repeats, whereas branch genotypes have three Bruce repeats (Table A). The pattern

9 of panel 1 loci repeats for the genotype with five Bruce repeats (e.g., Bruce0, Bruce0, Bruce, 1 Bruce1, Bruce, Bruce, Bruce and Bruce repeats) was previously determined for a B. melitensis biovar 1 (BCNN 1b; bru) isolate cultured from a patient from Argentina (). B. melitensis genotypes with three or six Bruce repeats and with a pattern of panel 1 loci alleles similar to the Peruvian isolates have not been described for isolates from other countries. We did not observe a relationship with the consumption of specific dairy products and genotype (data not shown). Also, we did not find a relation between genotype and specific symptoms and signs except that a branch genotype was isolated from.% of the patients presenting with splenomegaly, compared with the isolation of a branch genotype from.0% of the patients with no evidence of splenomegaly (P=0.0). Additionally, a branch genotype was isolated from.0% of the patients with hepatomegaly, compared with.% of the patients without evidence of hepatomegaly (P=0.00). The response to treatment was not recorded in the patient database, and hence we could not determine a relationship between genotype and recovery after treatment. Relevance of genotyping for patient management and source identification. The genotype of isolates obtained from one patient during the subacute phase and after a relapse were identical (genotype G). In another case, two genotype G isolates were cultured from a single patient s blood during the acute and subacute phase of a single illness episode. In a third case, genotype G was cultured from the blood of two patients living at the same address and hospitalized at the same day. These two patients recalled that they had consumed fresh cheese and eaten papa a la Huancaína, a popular local dish

10 prepared with potatoes and fresh cheese. No evidence for a common source of infection was obtained for any of the other patients. Cluster analysis for Peruvian B. melitensis genotypes. A dendrogram constructed for all twenty-two B. melitensis genotypes currently indentified in Peru is presented in Figure B. This dendrogram shows a more complex structure which is related to the relatively high degree of variation in the MLVA- panel B loci Bruce0, 0 and observed in these genotypes. DISCUSSION Evolution and origin of B. melitensis in Peru. Consistent with our previous molecular studies describing the genotypes of human Brucella isolates from Peru isolated in 000 and 001 (), all 0 isolates obtained between 00 and 00 were identified by MLVA- typing as B. melitensis and showed a closest homology with a B. melitensis biovar 1 isolate (bru1) from Mexico. Essentially all previously described Brucella isolates from Peru were determined to be B. melitensis biovar 1 (1,, ). Classical biovar typing however was not performed for all current isolates, and MLVA- typing does not allow biovar classification (, ). The homology of the Peruvian isolates with the B. melitensis biovar 1 isolate from Mexico further supports our notion that the majority of the B. melitensis strains circulating in Peru form a Latin American cluster distinct from the West and East Mediterranean clusters (). The similarity of two Peruvian isolates with an isolate originating from Spain may indicate that either the

11 source of infection of the latter case came from Peru or that Brucella has recently been re-imported into Peru. This could have occurred with the import of replacement animals for herds sacrificed in the course of the brucellosis control programme. If so, this illustrates a risk of global trade and transboundary transport of livestock. Replacement animals imported into Peru include Anglo-Nubians, a breed developed in the United Kingdom, Alpine and Saanen breeds of Swiss origin and imported from Chile, and Murcina and Granadina breeds from Spain (, Dr. Vargas personal communication). Also, semen from the Murcina and Granadina breeds has been imported during the last decade (Dr. Arroyo, personal communication). Historically, goats and other livestock animals were introduced into Latin America as early as with the second journey of Columbus to the Antilles (). These goats were of Spanish origin and have developed over time into Creoles goats kept by most farmers. Genetic analysis has indicated the phylogenetic relationship of Creoles goats from Peru with certain goat breeds from Spain and with Creoles goats from other Latin American countries (). Brucella may have evolved with these Creoles goats, thereby separating from the present day West and East Mediterranean clusters of B. melitensis genotypes and forming a distinct cluster of Latin American B. melitensis genotypes (). However, brucellosis is common in camels and may already have been present in indigenous Peruvian camelids (including the llama, vicuña, and alpaca) at the time of the Spanish conquest (1, ). Marine mammals may be infected with B. pinnipedialis, but this species is clearly distinct from B. melitensis (). Marine mammals are abundantly present in the pacific ocean along the Peruvian coast and infection of Peruvians with B. pinnipedialis has been reported, but transmission of this pathogen to humans is very rare (1, ). Most Brucella patients in Peru belong to

12 the lower socio-economic classes, and infection during international travel or through the consumption of imported food is highly unlikely. Using the same typing method as we have used, Kattar and coworkers identified a series of unique B. melitensis isolates in patients from Lebanon that were closely related to isolates from neighbouring countries (). The genotypic variation of these Lebanese isolates was limited with most of the variation occurring in panel B loci, and it was postulated that these isolates had evolved from indigenous strains with a common ancestor. The genotypic variation in the Peruvian isolates is similarly restricted, with variation occurring in one panel 1 locus (Bruce), one panel A locus (Bruce1) and three panel B loci (Bruce0, 0 and ) indicating that these Latin American isolates may have evolved from a small number of ancestors. Genotypic variation revealed by MLVA- typing was also observed for Rev-1 isolates, providing an example of the evolution of Brucella (1). Currently, a total of twenty-two genotypes have been identified among human Peruvian B. melitensis isolates. Most likely changes in the repeat number of in particular the more heterogenous panel B loci may have occurred at different occasions. Thus, identical or similar panel B loci patterns may have developed independently in strains with dissimilar panel 1 and or panel A loci patterns. Therefore, the genetic relationship between the twenty-two genotypes as shown in the dendrograms does not reveal their phylogenetic relation. In our hypothesis the different genotypes have developed by microevolution from a small number of ancestors characterized by differences in the panel 1 Bruce locus. Further, it is possible that the genotype (G) similar to the isolate from Spain has evolved separately in Peru. 1

13 Genotypic variation of B. melitensis in Peru and source of transmission. Brucella strains showing a specific MLVA pattern may circulate in a geographic area and may be transmitted to the human population over many decades (). Our results show that the spectrum of strains isolated from patients from central Peru changes over time. Of the fifteen B. melitensis genotypes isolated in 000 and 001, only three were identified among the ten genotypes isolated between 00 and 00 (). Also, the percentage of isolates with five or six Bruce repeats (branch 1) and with three Bruce repeats (branch ) has varied considerably over the years. The majority (.%) of the isolates obtained in 000 and 001 had five Bruce repeats and just.% had three Bruce repeats, but almost all (.%) isolates obtained in 00 and 00 had three Bruce repeats and.1% had five Bruce repeats. This trend continued through the period 00 to 00, by which time.% of the isolates had five or six Bruce repeats, and.% had three Bruce repeats. Although the isolates presented here and in our previous study () were obtained from blood cultures from patients initially hospitalized at three different hospitals, these three hospitals serve patients from the same districts of Callao and Lima, with most patients coming from specific districts within Callao. Most likely, this variation over time in the spectrum of genotypes is the result of movements of migratory herds and perhaps changes in sources of dairy products. The major route of infection with Brucella in cities in Peru is through the ingestion of contaminated dairy products, the origin of which products is unclear. Contaminated dairy products may be obtained from remote farms, from migratory goat herds in the Andean highlands that have not yet been reached by the vaccination programs, or from farms with goat herds that have been 1

14 partially vaccinated but still include infected animals. Thus, herds from different areas may be infected with different genotypes. Shops may also obtain their products from multiple vendors and areas depending on available supplies. There is no evidence to suggest that shops and markets in Callao obtain their dairy products from clearly distinct areas than those in Lima; Lima and Callao form one large metropolis. The migration of herds makes the epidemiology and control of brucellosis more difficult. Identification of the supply routes may help to trace the sources of infection, and these sources then may be confirmed by genotyping of isolates obtained from milk samples or other animal materials. Changes in the supply and marketing of dairy products originating from goat herds infected with different Brucella genotypes may cause a change in the predominant genotype isolated from patients. Alternatively, an outbreak of brucellosis in certain herds may have caused a predominance of human infections with a certain genotype during a particular period. Gender and risk of infection. More than half of the Brucella patients in Peru are females. We earlier speculated that Peruvian women who do most of the grocery shopping may have an increased risk of infection due to sampling different products while shopping (). Women may also become exposed to the pathogen while preparing meals. This is typical for urban exposure. Studies in rural, pastoral communities in the Balkans (, ) and rural Spain () show a much higher rate of disease, or at least seroprevalence, in men in herding communities. Is Bruce a marker for pathogenic variation of B. melitensis? One of the major challenges in brucellosis research is to delineate differences in pathogenicity and host-susceptibility. A growing number of studies have shown a relation between host-

15 susceptibility and gene polymorphism of various human genes such as the transforming growth factor beta1 gene, major histocompatibility complex class I genes, and certain interleukin genes (,, ). Here, we provide evidence that susceptibility to different B. melitensis biovar 1 genotypes is age-related and that specific Brucella genotypes are more frequently isolated from Brucella patients presenting with specific clinical manifestations such as splenomegaly and hepatomegaly. The genotypes G1, G, G, G, G and G comprising branch 1 of the dendrogram drawn for the Peruvian isolates were isolated almost exclusively from adult patients, and a relatively high proportion of the patients infected with these six genotypes presented with splenomegaly and or hepatomegaly. Studies in mice may help to elucidate difference in organ tropism and pathological effects (0). The ten genotypes comprising branch 1 and genotypes showed variation at four (Bruce0, 0, 1 and ) of the tandem repeat loci investigated of which variation in one (Bruce) locus separated branch 1 from branch. Bruce is used for species identification in combination with other MLVA- panel 1 loci and is considered relatively invariable. The number of Bruce alleles for a panel of 1 reference strains including B. melitensis biotypes 1 to, B. abortus biotypes 1 to and, B. suis biotypes 1 to, B. ovis, B. canis, B. neotemae, B. pinnipedialis and B. ceti was five with either one, two, three, four or six repeats (). Three Bruce repeats were found for B. pinnipedialis, B. ceti and B. suis biotypes and, and six Bruce repeats were found for B. suis biotype. Branch 1 of the Peruvian isolates is characterized by the presence of five or six Bruce repeats and branch isolates by three Bruce repeats. A B. melitensis isolate with three Bruce repeats was isolated from (.%) of the 0

16 children and adolescent patients and from (.%) of the 0 adult patients. From the other patients an isolate with five or six Bruce repeats was obtained. In our previous study, all patients were 1 years or older (mean, 0 years of age) and 0 of the isolates had five Bruce repeats (). The other four isolates had three Bruce repeats. These results may suggest that the number of Bruce repeats could mark a phenotypical characteristic that relates to age-dependent host susceptibility. Differences in the mode of exposure need to be considered. Occupational exposure is rare in patients residing in Lima or Callao, and food borne transmission is the most common mechanism for both adults and children. This seems to speak against the route of infection as being the principle discriminating factor. The possibility that adult patients had acquired brucellosis from a common point source (for instance, from meal served during a wedding ceremony) is considered highly unlikely. In Peru, brucellosis is a reportable disease, and each case requires detailed epidemiological investigations. We have not found any such demographic relationship between patients, excepting only two patients living at the same address. Furthermore, most patients in this study came from Callao, a large urban area with almost 00,000 inhabitants or from Lima, a city with almost. million inhabitants. More extensive analysis of the genomes of branch 1 and genotypes may help to identify the bacterial factors that determine differences in host susceptibility. Several bacterial factors are involved in the virulence and pathogenicity of Brucella, including the lipopolysaccharide, the type IV secretion system, and the BvrR/BvrS two component signaling system (, ). However, genetic differences that result in a modification of pathogenicity have not yet been identified. In B. melitensis, Bruce is located within the region of the trna modification GTPase gene, which stretches from nucleotide

17 to nucleotide of chromosome I of B. melitensis strain M (). The outer membrane proteins thought to be involved in virulence are encoded on this chromosome (1), are amongst others involved in cell adhesion and cell entry. It is possible that branch 1 and isolates might differ in a mutation in one of the outer membrane proteins, causing a critical change in a functional activity. Most, but not all, of the other virulenceassociated genes are also located on chromosome I; the genes encoding the type IV secretion system are located on chromosome II (1). As noted, the clinical signs and symptoms of brucellosis include such non-specific features as fever, malaise, headache and sweats (). A B. melitensis genotype belonging to branch was isolated from most of the brucellosis patients presenting with hepatomegaly and or splenomegaly. As branch isolates were obtained from.% of all Brucella patients, further studies will be needed to demonstrate a correlation of disease severity or of specific disease presentations with the genotype of the infecting strain. (It should also be noted that hepatosplenomegaly was assessed clinically, and radiographic confirmation with sonography or computed tomography was not obtained.) With regard to data analysis we consider it unlikely that an age related difference in the diagnosis could have led to a selection bias as the percentage of adults and children in whom splenomegaly or hepatomegaly was observed was similar. Nevertheless, authors are aware that there is intrinsic bias in the study due to the nature of the data at hand, which was not obtained through a structured or planned survey, and of the consequences which could result from this for the interpretation of the findings. Therefore, confirmation of our findings should come from future investigations into the genotypic diversity of human B. melitensis isolates between patient groups. Because Brucella is a fastidious 1

18 organism that is more easily cultured from the blood of acute or subacute patients than from patients with more persistent disease (1) this led to the inclusion of patients with mostly early stage disease. However, this unlikely will have resulted in the selection of a specific group of genotypes unless one assumes that a separate group of genotypes cause predominantly persistent disease and that some genotypes are more easily cultured than others. Implications of genotyping for patient management. Recurrence of brucellosis is a major problem in Peru and worldwide, and may be due to either relapse or reinfection (). In relapsing patients, the same genotype can be expected to be isolated during both the first episode and relapse (). Conversely, the isolation of a different genotype from a patient with a repeat episode of brucellosis would be consistent with reinfection. The limited variation among genotypic characteristics in a given patient population, however, makes it more difficult to discriminate between relapse and reinfection in a patient experiencing two or more disease episodes. Further studies that investigate the inclusion of other more highly variable tandem repeat loci or the use of other potentially more discriminatory genotyping methods such as single nucleotide polymorphism analysis could potentially help to discriminate in such cases between these two options (,, 0). By using MLVA genotyping for the same panel of loci, we previously demonstrated probable re-infection in three out of three brucellosis patients from Peru with recurrent disease (). The diversity of Brucella genotypes isolated from that particular group of patients hospitalized in 000 and 001 was larger than among the isolates obtained from the patients hospitalized in the period Finally, the isolation of the less-frequently observed B. melitensis genotype G from two patients 1

19 living at the same address and hospitalized on the same day supports a common source of infection in these two patients DISCLAIMER The views expressed in this article are those of the authors and do not necessarily reflect the official policy or position of the Department of the Navy, Department of Defense, or the U.S. government. Drs. E. Hall, B. Espinosa, and R. Maves are U.S. military service members working at the U.S. Naval Medical Research Center Detachment in Lima, Peru. This work was prepared as part of their official duties. Title 1 U.S.C. provides that 'Copyright protection under this title is not available for any work of the United States Government'. Title 1 U.S.C. 1 defines a U.S. Government work as a work prepared by a military service member or employee of the U.S. Government as part of that person's official duties. REFERENCES 1. Abbas, B., and H. Agab. 00. A review of camel brucellosis. Prev. Vet. Med. :-.. Al Dahouk, S., P.L. Flèche, K. Nöckler, I. Jacques, M. Grayon, H.C. Scholz, H. Tomaso, G. Vergnaud, and H. Neubauer. 00. Evaluation of Brucella MLVA typing for human brucellosis. J Microbiol Methods. :1-.. Al Dahouk, S., K. Nöckler, A. Hensel, H. Tomaso, H.C. Scholz, R.M. Hagen, and H. Neubauer. 00. Human brucellosis in a nonendemic country: a report from Germany, 00 and 00. Eur. J. Clin. Microbiol. Infect. Dis. :0-. 1

20 Avdikou I., V. Maipa, and Y. Alamanos. 00. Epidemiology of human brucellosis in a defined area of Northwestern Greece. Epidemiol. Infect. 1:0-.. Azor, P.J., M. Valera, J. Sarria, J.P. Avilez, J. Nahed, M. Delgado, and J.M. Castel. 00. Estimación de las Relaciones Genéticas Entre Razas Caprinas Españolas y Criollas Utilizando Microsatélites. ITEA :-.. Bosilkovski M., L. Krteva, M. Dimzova, and I. Kondova. 00. Brucellosis in 1 patients from the Balkan Peninsula: exposure-related differences in clinical manifestations, laboratory test results, and therapy outcome. Int. J. Infect. Dis. :-.. Bravo, M.J., J.D. Colmenero, J. Martín, A. Alonso, and A. Caballero. 00. Polymorphism of the transmembrane region of the MICA gene and human brucellosis. Tissue Antigens. :-0.. Bravo, M.J., J.D. Colmenero, M.I. Queipo-Ortuño, A. Alonso, and A. Caballero. 00. TGF-beta1 and IL- gene polymorphism in Spanish brucellosis patients. Cytokine :1-1.. Bricker, B.J., D.R. Ewalt, and S.M. Halling. 00. Brucella 'HOOF-Prints': strain typing by multi-locus analysis of variable number tandem repeats (VNTRs). BMC Microbiol. :.. Budak, F., G. Göral, Y. Heper, E. Yilmaz, F. Aymak, B. Baştürk, O. Töre, B. Ener, and H.B. Oral. 00. IL- and IL- gene polymorphisms as potential host susceptibility factors in Brucellosis. Cytokine. :-. 0

21 Castel, J.M., J.A. Sarria, M.J. Alcalde, O. Arroyo, Y. Mena, D. Bedotti, and D.Y. Fernandez-Cabanas. Estudio de algunos aspectos productivos y reproductivos en una explotación caprina experimental situada en el valle del río chillón (Lima-Perú) Corbel, J.M. 11. Identification of dye-sensitive strains of Brucella melitensis. J. Clin. Microbiol. :-. 1. DelVecchio, V.G., V. Kapatral, R.J. Redkar, G. Patra, C. Mujer, T. Los, N. Ivanova, I. Anderson, A. Bhattacharyya, A. Lykidis, G. Reznik, L. Jablonski, N. Larsen, M. D'Souza, A. Bernal, M. Mazur, E. Goltsman, E. Selkov, P.H. Elzer, S. Hagius, D. O'Callaghan, J.J. Letesson, R. Haselkorn, N. Kyrpides, and R. Overbeek. 00. The genome sequence of the facultative intracellular pathogen Brucella melitensis. Proc. Natl. Acad. Sci. U S A. :-.. Escalante, J.A., and J.R. Held.. Brucellosis in Peru. J. Am.Vet. Med. Assoc. :-.. Espinosa B.J., J. Chacaltana, M. Mulder, M.P. Franco MP, D.L. Blazes, R.H. Gilman, H.L. Smits HL, and E.R. Hall. 00. Comparison of culture techniques at different stages of brucellosis. Am. J. Trop. Med. Hyg. 0:-.. Foster, J.T., R.T. Okinaka, R. Svensson, K. Shaw, B.K. De, R.A. Robison, W.S. Probert, L.J. Kenefic, W.D. Brown, and P. Keim. 00. Real-time PCR assays of single-nucleotide polymorphisms defining the major Brucella clades. J. Clin. Microbiol. : Franco, M.P., M. Mulder, R.H. Gilman, and H.L. Smits. 00. Human brucellosis. Lancet Infect. Dis. :-. 1

22 Godfroid, J., A. Cloeckaert, J.P. Liautard, S. Kohler, D. Fretin, K. Walravens, B. Garin-Bastuji, and J.J. Letesson. 00. From the discovery of the Malta fever's agent to the discovery of a marine mammal reservoir, brucellosis has continuously been a re-emerging zoonosis. Vet. Res. : García-Yoldi, D., P. Le Fleche, M.J. De Miguel, P.M. Muñoz, J.M. Blasco, Z. Cvetnic, C.M. Marín, G. Vergnaud, and I. López-Goñi. 00. Comparison of multiple-locus variable-number tandem-repeat analysis with other PCR-based methods for typing Brucella suis isolates. J.Clin. Microbiol. : Gopaul, K.K., M.S. Koylass, C.J. Smith, and A.M. Whatmore. 00. Rapid identification of Brucella isolates to the species level by real time PCR based single nucleotide polymorphism (SNP) analysis. BMC Microbiol. :. 1. Guerra, H. 00. The brucellae and their success as pathogens. Crit. Rev. Microbiol. :-1.. Guillen, A., A.M. Navarro, M. Acosta, and M. Arrelucé. 1. Brucellosis epidemiology. In: Annals of the National Seminary of Zoonosis and Diseases of Nourishing Transmission. Lima: Ministry of Health... Kattar, M.M., R.F. Jaafar, G.F. Araj, P. Le Flèche, G.M. Matar, R. Abi Rached, S. Khalife, and G. Vergnaud. 00. Evaluation of a multilocus variable-number tandem-repeat analysis scheme for typing human Brucella isolates in a region of brucellosis endemicity. J. Clin. Microbiol. :-0.. Lapaque, N., I. Moriyon, E. Moreno, and J.P. Gorvel. 00. Brucella lipopolysaccharide acts as a virulence factor. Curr. Opin. Microbiol. :0-.

23 Le Flèche, P., I. Jacques, M. Grayon, S. Al Dahouk, P. Bouchon, F. Denoeud, K. Nöckler, H. Neubauer, L.A. Guilloteau, G. Vergnaud. 00. Evaluation and selection of tandem repeat loci for a Brucella MLVA typing assay. BMC Microbiol. :.. Lucero, N.E., S.M. Ayala, G.I. Escobar, and N.R. Jacob. 00. Brucella isolated in humans and animals in Latin America from to 00. Epidemiol. Infect. :-0.. Maas, K.S.J.S.M., M. Méndez, M. Zavaleta, J. Manrique, M.P. Franco, M. Mulder, N. Bonifacio, M.L. Castañeda, J. Chacaltana, E. Yagui, R.H. Gilman, A. Guillen, D.L.Blazes, B. Espinosa, E. Hall, T.H. Abdoel, and H.L. Smits. 00. Evaluation of brucellosis by PCR and persistence after treatment in patients returning to the hospital for follow-up. Am. J. Trop. Med. Hyg. :- 0.. Marianelli, C., C. Graziani, C. Santangelo, M.T. Xibilia, A. Imbriani, R. Amato, D. Neri, M. Cuccia, S. Rinnone, V. Di Marco, and F. Ciuchini. 00. Molecular epidemiological and antibiotic susceptibility characterization of Brucella isolates from humans in Sicily, Italy. J. Clin. Microbiol. :-.. Mendoza-Núñez, M., M. Mulder, M.P. Franco, K.S. Maas, M.L. Castañeda, N. Bonifacio, J. Chacaltana, E. Yagui, R.H. Gilman, B. Espinosa, D. Blazes, E. Hall, T.H. Abdoel, and H.L. Smits. 00. Brucellosis in household members of Brucella patients residing in a large urban setting in Peru. Am. J. Trop. Med. Hyg. :-.

24 Mense, M.G., L.L Van De Verg, A.K. Bhattacharjee, J.L. Garrett, J.A. Hart, L.E. Lindler, T.L. Hadfield TL, and D.L. Hoover Bacteriologic and histologic features in mice after intranasal inoculation of Brucella melitensis. Am. J. Vet. Res. : Moreno, E., and I. Moriyon. 00. Brucella melitensis: a nasty bug with hidden credentials for virulence. Proc. Natl. Acad. Sci. USA. :1-.. Moreno, E., A. Cloeckaert, and I. Moriyón. 00. Brucella evolution and taxonomy. Vet. Microbiol. 0:0-.. Ministry of Health, Peru. 00. Brucellosis in Callao. Wkly. Epidemiol. Bull. 1:1-.. Ministry of Health, Peru. 00. Brucellosis cases registered in outpatient clinics in Peru General Office of Statistics and Informatics, 00.. Rodero, A., J.V. Delgado, and E. Rodero. 1. Primitive andalusian livestock and their implications in the discovery of America. Arch. Zootec 1:-00.. Serra Alvarez J., and P. Godoy García Incidence, etiology and epidemiology of brucellosis in a rural area of the province of Lleida. Rev. Esp. Salud Publica. :-.. Smits, H.L., B. Espinosa, R. Castillo, E. Hall, A. Guillen, M. Zevaleta, R.H. Gilman, P. Melendez, C. Guerra, A. Draeger, A. Broglia, and K. Nöckler. 00. Brucella MLVA genotyping of human Brucella isolates from Peru. Trans. Roy. Soc. Trop. Med. Hyg. :-0.. Sohn, A.H., W.S. Probert, C.A. Glaser, N. Gupta, A.W. Bollen, J.D. Wong, E.M. Grace, and W.C. McDonald. 00. Human neurobrucellosis with

25 intracerebral granuloma caused by a marine mammal Brucella spp. Emerg. Infect. Dis. :-.. Tibary, A., C. Fite, A. Anouassi, and A. Sghiri. 00. Infectious causes of reproductive loss in camelids. Theriogenology. :-. 0. Young, E.J.. Brucellosis: current epidemiology, diagnosis, and management. Curr. Clin. Top. Infect. Dis. :-1. Legend to the figures Figure 1. MLVA patterns of B. melitensis genotypes from Peru PCR bands for the four polymorphic MLVA- loci are shown for the ten genotypes (G1-G) identified among 1 consecutive human isolates obtained in the period C, Bruce0; E, Bruce0; I, Bruce1; M, Bruce. Lanes to the side of each of the groups of four lanes labelled C, E, I and M show the 0bp ladder marker. Figure. Dendrogram of Peruvian B. melitensis genotypes showing their relationship with strain Bru1 from Mexico. MLVA- assay results for the B. melitensis isolates from Peru were compared with Bru1 (queried strain), a human B. melitensis biovar 1 isolate from Mexico, by the use of UPGMA algorithm to construct a dendrogram. A. Genotypes obtained in the period A dendrogram was constructed for the ten genotypes (G1-G) obtained during this period with numbers referring to

26 1 1 the isolate number of the type strain. The ten genotypes divide into two branches named B1 (closed circle) and B (open circle) with B1 encompassing G1 (type strain 1), G (strain ), G (strain ), G (strain ), G (strain 0) and G (strain ) and B encompassing G (strain ), G (strain ), G (strain 0) and G (strain ). Genotype (type strain 0) in turn closely matches with Bru1, a B. melitensis biovar 1 isolate from Spain. Branch 1 genotypes characterized by five or six Bruce repeats and Branch genotypes characterized by three Bruce repeats. B. Genotypes obtained in the periods and A dendrogram was constructed for all twenty-two Peruvian genotypes currently characterized by MLVA- genotyping. Type strains 1,,,, 0,,,,, and 0 were detected in the period and type strains 1,,,,,, -, 1-, and were detected in the period Genotypes with five or six Bruce repeats are enclosed by closed circles and genotypes with three Bruce repeats are enclosed by open circles. Type strain 0 closely matches with Bru1, a B. melitensis biovar 1 isolate from Spain.

27 TABLE 1. Clinico-epidemiological characteristics of Brucella patients diagnosed in the period Characteristic of patients No. (%) Sex Male Female Course of disease a Acute Subacute Persistent Relapse Disease symptom or sign Fever Malaise Myalgias Lumbalgias Sweats Headache Hyporexia Shaking chills Arthralgia (.) (0.) (.) 1 (.) (.) (.) (.) (.1) (.) (.) (1.) (.0) (.) (.) (.)

28 Abdominal pain (.) Diarrhoea (1.1) Hepatomegaly 0 (.) Splenomegaly (1.1) Residence Callao Lima Unknown Consumption of unpasteurized dairy produce Milk Cheese Papa a la Huancaína Ice-cream 0 (.) (1.) (.) (.) (.) (.) (1.) Downloaded from 1 Yogurt Other Any fresh dairy product Contact with livestock or animal products Goats Animal organs 1 (1.) (.) (.1) 1 (1.1) (.) on October, 01 by guest a Including second isolates of two patients with Brucella isolated during two disease episodes

29 TABLE. Polymorphic MLVA- assay loci of B. melitensis isolates from Peru and comparison of genotypes identified in the period with those obtained in 000 and 001 A. Isolates obtained in Number of repeats at each of the following loci Bruce panel and locus 1 A B No. of type strain 1 Number of isolates Genotype a Branch 1 1 G1 B1 0 G B G B G B1 0 G B G B 1 G B1

30 G B1 0 1 G B1 1 G B1 B. Isolates obtained in Number of repeats at each of the following loci Bruce panel and locus 1 A B No. of type strain b (No. used for construction of the dendrogram of figure Number of isolates b Genotype b No. of identical type strain isolated in the period (Genotype, Branch) 0

31 B) 1 (1) G1 No. 1 (G1, B1) () G None () G No. (G, B1) () G No. (G, B1) 1 () G None () G None 1 () 1 X None () 1 X None () 1 X None 1 () 1 X None 1 (1) 1 X None () 1 X None () 1 X None () 1 X None 1

32 0 () 1 X None a This study. b As coded for genotypes isolated in the period by Smits et al, 00 (), with X presenting unique genotype isolated in the period not specified by code in that study.

33 TABLE. Variation in spectrum of human B. melitensis genotypes identified in subsequent years. Number of isolates in each year Genotype and branch G1 (B1) 1 G (B1) 1 1 G (B1) 1 G (B1) 1 G (B1) 1 G (B1) 1 G (B) 1 G (B) 1 G (B) G (B) 1 1 B1 1 1 B 1

34 TABLE. Distribution of the different B. melitensis genotypes among the two sexes Number of isolates in each group Genotype and branch Female Male G1 (B1) 1 G (B1) 1 1 G (B1) 1 G (B1) G (B1) 1 G (B1) 1 G (B) 1 G (B) G (B) G (B) B1 1 1 B

35 TABLE. Residence of patients and B. melitensis genotypes Number of isolates in each city Genotype and branch Callao a Lima G1 (B1) G (B1) G (B1) 1 G (B1) G (B1) 1 G (B1) 1 G (B) G (B) G (B) 1 1 G (B) B1 B

36 a Including Bellavista, La Perla, Ventenilla and Chalaca. Three isolates from patients with unknown residency are not included.

37 TABLE. Age of patient and genotype of the B. melitensis isolate Number of isolates from patients in the following age groups (years) Genotype and branch 1 to 0 1 to 0 1 to 0 1 to 0 >0 G1 (B1) 1 G (B1) 1 1 G (B1) 1 G (B1) 1 1 G (B1) 1 G (B1) 1 G (B) 1 G (B) 1 1 G (B) 1 1 G (B) 1 1 B1 B 1

38

39