Expression and diversity of Echinococcus multilocularis AgB genes in secondarily infected mice: evaluating the influence of T-cell immune selection on antigenic variation

The T-cell-mediated immune response exhibits a crucial function in the control of the intrahepatic proliferation of Echinococcus multilocularis larvae in mice and humans, both being natural intermediate hosts of the parasite. Antigen B ( AgB ), a metabolized Echinococcus spp. lipoprotein, contributes to the modulation of the T-cell immune response, and distinct sites of the corresponding AgB1 , AgB3 and AgB4 genes were shown to be under positive selection pressure. Since several AgB gene variants are present in a single Echinococcus metacestode, we used secondary E. multilocularis infections in BALB/c and in athymic nude mice (devoid of T-cell responses) to analyze the effect of the cellular immune response on the expression and diversity of EmAgB1 – EmAgB4 genes. We demonstrated hereby that EmAgB transcripts were less abundant in nude mice during the early phase of infection (at one month post-infection), and that EmAgB2 is simultaneously down-regulated when compared to the other three genes. A negative relationship exists between the level of transcription and diversity of EmAgB genes. Moreover, no excess of nonsynonymous substitutions was found among the distinct EmAgB alleles from a single host. Together, these results pointed to the effect of purifying selection, which seemed to eliminate the detrimental AgB variants generated during the development of the metacestode within the peritoneal cavity of its intermediate host.


Introduction
Alveolar echinococcosis is a severe human life-threatening zoonosis caused by the infection with the larval stage of the fox tapeworm Echinococcus multilocularis. It is characterized by the growth of a tumor-like larval tissue, the so-called metacestode, which primarily infiltrates the host target organ (liver) and secondarily may form metastases in other organs or sites. Because of the growth behavior and the potential of the E. multilocularis metacestode to form metastases, alveolar echinococcosis is considered the most lethal helminthic infection in humans (Craig et al., 1996). The common intermediate hosts harboring the larval stage of E. multilocularis are rodents, but humans can accidentally become infected as well (Thompson, 1995). Since E. multilocularis uses mice and humans as natural intermediate hosts, this parasite represents a suitable model to study the infective process in echinococcosis at the experimental murine laboratory level.
Host immunity can contribute to the control of the E. multilocularis metacestode growth. Thus, upon experimental secondary infections it became apparent that the metacestode grows faster in athymic nude mice or in severely compromised immunodeficient (SCID) mice than in fully T-cell-competent wild type control animals (Playford et al., 1992;Dai et al., 2004). Furthermore, T-cell immunity significantly contributes to the control of AE in human patients as well, as demonstrated by the rapid, fatal outcome of the infection in an immunodeficient patient co-infected with human immunodeficiency virus (HIV) (Sailer et al., 1997). Upon restoration of CD4-immunocomptence in another AIDS-patient co-infected with E. multilocularis, the course of disease could be positively influenced again (Zingg et al., 2004). Gottstein et al. (1994) demonstrated that mouse strains resistant to E. multilocularis infections show a stronger cellular immune response than susceptible mouse strains. The main role of the cellular immune response in the control of metacestode growth was confirmed later, by experiments showing that mouse strains which do not develop cellular immune response do not control metacestode growth, while mice deficient in humoral immunity control parasite growth up to a certain level (Dai et al., 2004).
Antigen B (AgB) is a lipoprotein secreted during the larval stage of Echinococcus spp. parasites. It was first isolated from E. granulosus hydatid fluid (Oriol et al., 1971) where it appeared as the most abundant protein (Oriol and Oriol, 1975). The native protein of E. granulosus is a 160 kDa polymer composed by 8 kDa monomers (González-Sapienza et al., 1996) encoded by the AgB gene family (Chemale et al., 2001). There are five AgB genes characterized in E. multilocularis, EmAgB1-EmAgB5 (Frosch et al., 1994;Fernández et al., 2002;Mamuti et al., 2006), but only EmAgB1-EmAgB4 show similarity (larger than 90%) with ortologous sequences of E. granulosus. Antigen B sequence variability inside a single metacestode was described for E. granulosus (Arend et al., 2004 and E. multilocularis (Mamuti et al., 2004). We have suggested that this variability could be adaptive, being related to the evasion of immune responses , but this had never been confirmed to date.
Several approaches have been used so far to examine AgB function. In vitro studies demonstrated that AgB is a protease inhibitor with an ability to inhibit the recruitment of neutrophils (Shepherd et al., 1991). Riganò et al. (2001) evaluated the cytokine secretion profile after AgB stimulation in peripheral mononuclear blood cells (PMNC) collected from patients with progressive versus stationary cystic echinococcosis (infection with E. granulosus). Their findings showed that AgB elicits a Th2 non-protective response preferentially in patients with progressive disease, and it suggested that AgB may contribute to the modulation of the host immune response. Chemale et al. (2005) analyzed the AgB lipid-binding properties and demonstrated that recombinant peptides encoded by EgAgB1 and EgAgB2 were able to bind but not to release fatty acids, fitting with a lipid detoxification role for the AgB protein. Haag et al. (1998) verified an excess of non-synonymous over synonymous substitutions for EgAgB1 sequences from different strains. Additional analyses of positive selection on genes encoding AgB identified particular codons under selection pressure in AgB1, AgB3 and AgB4, but not in AgB2 (Haag et al., 2006b). Since the variable sites occur in sequences from distinct strains/species of the genus Echinococcus, which manifestly differ in host specificity, we think that positive selection on AgB (among other factors, of course) is important for the adaptation to a particular host species. Interestingly, the distribution of nucleotide polymorphisms in sequences obtained from a single E. granulosus metacestode also departs from neutrality, but only for EgAgB2, and not for the other EgAgB genes . The adaptive role of nucleotide polymorphisms present inside a single E. multilocularis metacestode is the focus in the present study.

Echinococcus culture and mice infection
Maintenance of E. multilocularis metacestodes (KF5 isolate) was done by serial passages in gerbils (Meriones unguiculatus). Microcysts were prepared according to Hemphill and Gottstein (1995). Fourteen 8-week-old female athymic nude mice and 14 BALB/c mice were infected by intraperitoneal inoculation of approximately 100 freshly prepared microcysts suspended in 100 μl of Hank's Balanced Salt Solution. The 14 infected mice from each strain were separated in two groups (n = 7 each), and sacrificed by CO 2 -euthanasia 1 or two 2 months post-infection (mpi). The intraperitoneal metacestode tissue was immediately recovered upon euthanasia, weighted and submitted to molecular analysis.

RNA extraction and cDNA synthesis
Total parasite RNA was isolated by the single-step guanidinium thiocyanate procedure using TRIzol Reagent (Gibco) according to the manufacturer's instructions. Residual genomic DNA was removed from sample RNA by performing 180 min incubation at 37°C in the presence of RQ1 RNase-free DNase I (Promega). After heat inactivation (10 min at 80°C) of the DNase, RNA preparations were subjected to cDNA synthesis by applying a random oligonucleotide-hexamer primer (Promega) and AMV reverse transcriptase (Promega). Finally, complementary DNA (cDNA) was purified by using a spin column from High Pure PCR product purification kit (Roche Diagnostics) and subsequently eluted with 100 μl of water before use in RT-PCR analyses.

Quantification of AgB expression
Four primer pairs were designed to specifically amplify the four E. multilocularis AgB genes (EmAgB1-EmAgB4). We used a specific forward primer annealing to the 5′ region inside the second exon and a reverse primer annealing to the 3′ flanking region of AgB1 and AgB3 or AgB2 and AgB4 (Table 1).
The amplification was performed on a LightCycler (Roche Diagnostics) using SYBR Green I as a dsDNA-specific fluorescent dye and a continuous monitoring. Amplifications of cDNA molecules derived from either different AgB genes or a "housekeeping" act gene (encoding a β-actin homolog) were carried out in triplicates. A control PCR for each sample included RNA equivalents from an aliquot that had not been treated with reverse transcriptase, to confirm that no residual genomic DNA might have resisted to DNase I digestion. PCR was initiated by a "hot start" of 15 min at 95°C followed by 40 cycles of denaturation (94°C, 15 s), annealing (50°C, 30 s) and extension (72°C, 30 s). Fluorescence was measured at each cycle at 72°C for AgB2, 76°C for AgB1 and AgB3 and at 79°C for AgB4. The quantification of the "housekeeping" act gene mRNA was used in order to compensate for variations in AgB input RNA amounts and efficiencies of reverse transcription. Mean values from triplicate determinations were taken for the calculation of relative AgB-mRNA levels, using the ratio of AgB-mRNA to act-mRNA.

Cloning, SSCP screening and sequencing
The RT-PCR products of AgB1-AgB4 obtained from E. multilocularis of one randomly chosen mouse for each group (nude 1 mpi, nude 2 mpi, BALB/c 1 mpi and BALB/c 2 mpi) were purified from solution using appropriate spin columns (Qiagen). One microliter of each purified RT-PCR product was cloned in a TOPO TA Vector for Sequencing (Invitrogen).
About 33-44 recombinant plasmids of each EmAgB cDNA were screened for sequence polymorphism by single strand conformation polymorphism (SSCP) after PCR amplification. Each recombinant colony was eluted in 50 μl water and heat denatured at 95°C for 5 min. The PCR mix was added to 5 μl of the lysate in a final volume of 25 μl. The PCR was carried out with the same primers and conditions as used for RT-PCR. Subsequently, 3 μl of each PCR product were denatured for 3 min in a buffer containing 95% formamide, 0.0125% xylene cyanole and 0.025% bromophenol blue, and immediately chilled on ice. The electrophoresis was conducted in a GenePhor Electrophoresis Unit (GE Healthcare) using precasted gels (GeneGel SSCP, 12.5%), at 200 V for 2.5 h and SSCP buffer A (GE Healthcare). The electrophoresis temperature was standardized for each EmAgB gene: 12°C for AgB1, AgB2 and AgB3, and 5°C for AgB4. The gels were silver stained, digitalized and analyzed with the ImageMaster package (GE Helthcare). The plasmids containing inserts from different alleles, as identified by the SSCP technique, were sequenced by cycle sequencing using primers M13 F and R (Invitrogen) and BigDye v3.0 (Applied Biosystems). The electrophoresis was performed in a capillary ABI-PRISM 3100 sequencer.

Data analyses
We used a factorial analysis of variance to search for significant differences in expression between months of infection (one or two) and mouse strains (athymic nude or BALB/c) for each gene using the SPSS (version 10.0) package. The AgB-mRNA quantifications relative to act-mRNA for each gene (EmAgB1-EmAgB4) and mouse were log-transformed (natural logarithm) to normalize their distribution. A one-way analysis of variance was performed to evaluate significant expression differences among genes.
Sequence quality was assessed by Phred software at www. unb.br/electropherogram/ and by visual inspection. The sequences were manually aligned and polymorphism analyses were done with DNAsp v 4.1 (Rozas et al., 2003). The allele diversity for each AgB gene in each mouse was estimated by the Past Software (Hammer et al., 2001) using the Shannon Diversity Index (SDI), where the total number of alleles is Fig. 1. Average EmAgB1-EmAgB4 mRNA amounts in parasites collected from BALB/c (black bars) and nude (grey bars) mice at 1 and 2 mpi, quantified in relation to act-mRNA. Asterisks display the statistically significant differences (p b 0.05). The horizontal dashed lines indicate the highest mean, and bars represent the standard deviation.
weighted by the frequency of each allele (Poole, 1974). SDIs for each gene/mouse were compared using a t test, where means and variances were obtained by a randomization process based on 1000 permutations (Hammer et al., 2001). We calculated the nonparametric Spearman correlation between EmAgB1-EmAgB4 expression level and the corresponding SDIs in each mouse using the SPSS (version 10.0) package.
Gene networks were obtained using the TCS software, version 1.21 (Clement et al., 2000). Both Tajima's D Test (Tajima, 1989) and the Fisher's exact test for excess of nonsynonymous over synonymous substitutions (Nei and Kumar, 2000) were carried out with DNAsp v 4.1 (Rozas et al., 2003) using the sequences sampled from each metacestode. An additional test, based on Fitch et al. (1997), was used to calculate the probability of a given set of mutations in a coding sequence. In this binomial probability analysis we determined the probability for a synonymous and non-synonymous mutation set to be observed, considering the number of synonymous and non-synonymous sites present in the sequence. Therefore, in a sequence with m nucleotides in which s is the number of synonymous sites and n (m-s) is the number of non-synonymous sites (Nei and Gojobori, 1986) we considered that s/m was the probability of each mutation to be synonymous and n/m the probability from this mutation to be non-synonymous. Then, we used the binomial distribution properties to estimate the probability of a certain set of mutations for a given sequence.

Parasite mass recovered from experimentally infected mice
To obtain parasites that underwent different immune challenges from the host, 14 T-cell-deficient nude mice and 14 BALB/c controls were intraperitoneally infected. At 1 mpi the mass of parasites recovered from 7 mice of each strain was relatively low and not significantly different for both nude and BALB/c mice (0.57 g ± 0.22 and 0.94 g ± 0.28, respectively). However, when the parasite was recovered 2 mpi, the nude mice contained a significantly larger amount of metacestode mass than the BALB/c (6.21 g ± 2.87 and 1.85 g ± 1.34; p b 0.05). These results confirm that E. multilocularis metacestode growth is suppressed in a host with an intact immune system when compared to a T-cell-deficient situation.

AgB-mRNA quantification
To evaluate the relative expression of E. multilocularis AgB genes and to investigate if the host T-cell immune response has an effect on EmAgB transcription, we quantified the abundance of mRNAs encoding the four known AgB subunits in metacestodes recovered from BALB/c and nude mice at 1 and 2 mpi. First, it became apparent that EmAgB2 was significantly less transcribed than the other three AgB genes (p b 0.01; Fig. 1). Note that the amount of mRNA is expressed in a logarithmic (ln) scale, meaning that the difference from EmAgB2 mRNA quantities to those obtained for the other genes corresponds to about e 2 . Our results also show differences in expression between the two mouse strains, most evidently visible at 1 mpi (Fig. 1). Within 1 mpi, the closely related EmAgB2 and EmAgB4 genes were significantly more expressed (p b 0.05) in the BALB/c host. Although EmAgB1 and EmAgB3 also show more mRNA transcripts during 1 mpi in the BALB/c mice, the difference detected was statistically not significant, probably due to the large variances and small sizes in our samples. At 2 mpi, parasites from both host strains did not show differences in the amount of EmAgB transcripts, except for EmAgB2, which was consistently less transcribed in the nude mice (see Fig. 1).

AgB diversity and host T-cell immune response
The role of the T-cell dependent immune response on AgB gene diversification was evaluated by cloning and sequencing transcripts (cDNA) from the four AgB genes, sampled from parasites of four randomly selected mice from each group. The total number of alleles, and the diversity index for each gene/ mouse, are shown in Table 2. The lowest allele diversity, as estimated by SDI, was obtained for EmAgB4 from the BALB/c mouse at 1 mpi and for EmAgB1 and EmAgB3 from nude mice at 1 and 2 mpi, respectively, whereas the highest diversity values were from EmAgB2, on both months and both mouse strains. However, the differences of SDI between genes were statistically significant (p b 0.01) only when EmAgB2 was compared to EmAgB3 and EmAgB4 in the BALB/c mouse at 1 mpi. There was a negative relationship between AgB transcription levels and transcript diversity, but all Spearman coefficients were non-significant (data not shown). The best relationship between expression and diversity was obtained for parasites at 2 mpi (ρ = − 0.667; p = 0.071; n = 8). Fig. 3. EmAgB1-EmAgB4 allele networks. The larger circles indicate the alleles obtained in our sample and dots represent the putative intermediate states, which were not encountered. Each line represents a single mutational step, as inferred according to a 95% confidence parsimony criterion. The standard sequences are always the most abundant, and variants appear only once, except for two variants of EmAgB2, which appeared twice in the same mouse or in separate mice (indicated with bold circles). Variants are denoted according to the mouse strain (w denotes the BALB/c and n the nude mice) and month of parasite collection (1 or 2 mpi). Numbers after the underline sign indicate the respective clone. GenBank accessions are DQ826464-DQ826476 (EmAgB1), DQ826477-DQ826503 (EmAgB2), DQ831832-DQ831843 (EmAgB3) and DQ831844-DQ831855 (EmAgB4).
The diversity values of EmAgB1-EmAgB3 were higher in the BALB/c mice at both time points (1 and 2 mpi), while in the nude mouse the values were higher for the EmAgB4 only. However, differences between the EmAgB1-EmAgB4 diversities and between parasites sampled from nude and BALB/c mice were all not statistically significant (Table 2). These results do not suggest a clear relationship between EmAgB diversity and host T-cell immune response in the frame of secondary infections in mice. Differences in EmAgB diversity between 1 and 2 mpi for each gene were also not statistically significant.

EmAgB alleles within the host are originated by point mutations
The SSCP technique efficiently detected single nucleotide polymorphisms from the EmAgB amplicons obtained in the present study (Fig. 2). We sequenced ∼10 clones showing the same standard SSCP pattern for each gene, and no nucleotide variation was encountered. Most of the variant SSCP patterns found in the present study, which we call alleles, differed by a point mutation.
As expected, most nucleotide substitutions were transitions (44 out of 50 substitutions). A surprising result was that six out of 26 EmAgB2 and two out of 12 EmAgB3 variants contained indels, which would result in non-functional proteins (data not shown). The parsimony networks generated from our set of sequences (Fig. 3) showed that each EmAgB variant originated independently by a single mutation from the most frequent, standard allele. There were only a few exceptions, four occurring for EmAgB2, one for EmAgB3 and one for EmAgB4. A single EmAgB2 variant was shared by parasites from different mice (clones b2w1_29 and b2n1_1) and another two appeared twice in the same mouse (Fig. 3). All other variant alleles occurred only once and in a single mouse.

Natural selection and murine T-cell immunity
In order to determine the evolutionary significance of EmAgB diversity inside a single E. multilocularis metacestode, we looked for evidences of natural selection on transcript variability. A Tajima's D Test was performed, using the total cluster of EmAgB alleles collected from each mouse. The values were all negative and statistically significant, except for EmAgB1 at 2 mpi of a nude mouse and EmAgB4 1 mpi of a BALB/c mouse (Table 3). Negative and significant Tajima's D values, caused by an excess of rare alleles in the sample, could be explained either by an adaptive sweep, or by a rapid population expansion. Trying to discriminate between these two causes, we carried out additional tests, classifying each mutation from the rare variant alleles as synonymous or nonsynonymous. Both the Fisher exact test and the binomial probability analysis (see Materials and methods section) would provide additional evidences about the occurrence of negative or positive (diversifying) selection, given a significant excess of synonymous or non-synonymous mutations, respectively. However, our results did not indicate departure from neutrality.

Discussion
Considering that, as a basic principle, the host controls Echinococcus metacestode growth predominantly via cellular immune responses, and that parasitic AgB modulates the host's T-cell response, we analyzed secondary E. multilocularis infections of athymic nude and T-cell-competent BALB/c mice, to assess the effect of a host T-cell response on AgB expression and respective diversity. For this, we quantified the transcription of EmAgB1-EmAgB4, hereby reporting that the AgB mutations originated in both mouse strains during parasite development. At variance with the former view that AgB variation would be adaptive even within a single host, we concluded that it is better explained by the combined effect of diversifying selection at the population level and relaxation of purifying selection at the level of individual metacestodes.
Antigen B is among the most abundant and immunogenic antigens of Echinococcus (Oriol et al., 1971;Lightowlers et al., 1989). There is accumulating evidence that AgB modulates the intermediate host immune response abrogating protective immunity and consequently maintaining a continuous status of parasite proliferation (Shepherd et al., 1991;Riganò et al., 2001Riganò et al., , 2002. Parasite antigens which interact directly with the host immune system are usually subject to strong selective pressures and can co-evolve with host immunity (Baum et al., 2003). Primarily the cellular immune responses, but not the humoral responses, are able to contribute to the control of the Echinococcus metacestode growth in mice (Fotiadis et al., 1999;Dai et al., 2004); therefore BALB/c mice are anticipated to challenge the E. multilocularis metacestode cell proliferation more importantly than athymic nude mice.
Intriguingly, the quantity of EmAgB transcripts was higher in the T-cell proficient BALB/c mice during the beginning of the infection. It seemed that during the early period of metacestode proliferation, EmAgB2 and EmAgB4 genes were downregulated in the athymic nude mice, which lacked of T-helper cell interaction. Two alternative explanations are possible for these findings: AgB genes were either 1) directly or indirectly induced by T-cell immune responses in the BALB/c mice; or 2) down-regulated in the immune deficient nude mice. In another gene expression study, using the same infected mice analyzed here, Matsumoto et al. (2006) had found lower levels of 14-3-3 transcripts in the nude mice infected with E. multilocularis at 1 mpi, but not for II/3-10, a protein related to the mammalian ezrins, radixins and moezins, all involved in the cellular architecture. Both AgB and 14-3-3, which are known to be involved in the control of cell proliferation, must have critical roles in parasite infectivity and patogenicity.
The first step to understand the differential control of AgB genes during the E. multilocularis life cycle was made by Mamuti et al. (2006), by demonstrating that EmAgB3 is constitutively expressed, while the other AgB genes are mainly transcribed in protoscoleces. The key factors linking the host immune response and the control of Echinococcus gene expression should be a matter for further study, since they are the basis of a host-parasite cross-talk, which determines the fate of the infection. It has already been shown for schistosome parasites that the host cellular immune responses in the definitive host have major effects on parasite early development (Davies et al., 2001). The lack of a Tcell immune response alters the normal parasite development in the early stages, leading to attenuated parasite forms which would prolong host survival.
A high sequence (allele) diversity of E. granulosus AgB genes inside a single metacestode was previously reported (Arend et al., 2004;Haag et al., 2004) and its origin was supposed to be related to the host-mediated immune selection pressure . Our results demonstrated that E. multilocularis AgB genes can also reach high diversities in a single metacestode. Although we found no significant difference between the SDI of AgB genes from parasites recovered from nude and BALB/c mice, we think that the higher values found for EmAgB1-3 in the BALB/c mice and for EmAgB4 in the nude model could have a biologic impact. However, due to the limited number of infected mice analyzed (one mouse per strain and mpi), we cannot draw definitive conclusions about any direct relationship between sequence diversity and T-cell immune challenge exerted by the host.
We are also aware that it is not possible to ascertain if the observed mutations had originated before the experimental infection (during gerbil passages or microcyst proliferation inside the in vitro culture) or after the infection (during metacestode growth inside host peritoneal cavity). Most alleles arose by a single mutation from the standard sequence, but metacestodes from distinct mice seldom shared a non-standard allele (this happened only once for EmAgB2), which could be an indication that the mutations occurred after the infection. Moreover, since a similar pattern of diversity (excess of rare alleles in a single metacestode) was also described for E. granulosus , where the cyst develops from only one oncosphere, we suggest that the mutations were mostly originated inside the host, although we can not rule out that some mutations were already present in the microcyst population injected into the mice.
Another scenario which would explain both the present findings for E. multilocularis and the previous results obtained for E. granulosus would be an extremely high redundancy of AgB genes inside the genome. In this case, a single cell could bear a number of distinct AgB gene variants expressed at the metacestode stage. However, genomic Southern Blots for E. granulosus AgB genes did not indicate a high redundancy (Chemale et al., 2001;Haag et al., 2004), but we hypothesized that AgB genes can increase in number through gene rearrangements during metacestode growth (Haag et al., 2006a).
The present results failed to confirm our previous hypothesis that the Echinococcus within-metacestode AgB variability would represent some form of antigenic variation mechanism. There is no straightforward proof that AgB variation is directed to T-cell-mediated immune response. Moreover, our analyses showed that the substitution process in EmAgB1-EmAgB4 alleles followed the neutral pattern, since the frequency of nonsynonymous and synonymous mutations was proportional to the number of non-synonymous and synonymous sites. Two questions remain, however.
First, if particular amino acids changes do not seem to be favored by natural selection when in contact with the host immune system, how are mutations maintained? Our results of transcript quantification of EmAgB1-EmAgB4 genes demonstrated that the lowest expressed gene is EmAgB2. In our mutation analysis EmAgB2 was also the most diverse gene, and a large fraction of the variant alleles contained missense frameshift mutations (data not shown). Although not statistically significant, the inverse relationship between the amount of mRNA and genetic diversity suggests that variation in AgB genes within a metacestode is principally maintained through a relaxation of purifying selection.
Second, what is, then, the biological significance of the AgB positively selected sites previously mapped by Haag et al. (2006b)? Although the positively selected sites identified in our previous study did not seem to have a biologic significance in terms of antigenic variation within a single host, it is possible that AgB variation is needed at the population level. Particular AgB alleles might be involved in the adaptation to a particular host species, or even genetic variable individuals within a host species. Indeed, the sequences analyzed in our previous study came from different Echinococcus strains or species, which are specialized to distinct intermediate hosts. Being able to infect a wide range of herbivorous mammals, Echinococcus parasites must be able to face quite distinct immune challenges, and AgB diversity appears to have an important role in this process.
Our expression studies suggest that there are significant differences in the AgB gene transcription levels between BALB/c and nude mice, especially during the early periods of infection. However, we were not able to conclude from the present AgB diversity data, if the antigenic variation is directed or not to T-cell response evasion. Further studies focusing on this particular question should, therefore, rely on larger sample sizes. Furthermore, reconstituting nude mice with purified immune cell populations from normal mice will also help to address this problem in the frame of future experiments. Another aspect which might have influenced our results is that we worked with secondary infections, starting from microcysts showing a rudimental laminated layer, which might already protect the parasite from the direct attack of host immune cells. The biological role of AgB diversity during oncosphere establishment at the very early stage of primary infection, where innate immunity leads into adaptive immunity, remains an open question, especially also in view of the subsequent events in the target organ where differentiation into a mature metacestode occurs.