Schistosoma-associated Salmonella resist antibiotics via specific fimbrial attachments to the flatworm

Background Schistosomes are parasitic helminths that infect humans through dermo-invasion while in contaminated water. Salmonella are also a common water-borne human pathogen that infects the gastrointestinal tract via the oral route. Both pathogens eventually enter the systemic circulation as part of their respective disease processes. Concurrent Schistosoma-Salmonella infections are common and are complicated by the bacteria adhering to adult schistosomes present in the mesenteric vasculature. This interaction provides a refuge in which the bacterium can putatively evade antibiotic therapy and anthelmintic monotherapy can lead to a massive release of occult Salmonella. Results Using a novel antibiotic protection assay, our results reveal that Schistosoma-associated Salmonella are refractory to eight different antibiotics commonly used to treat salmonellosis. The efficacy of these antibiotics was decreased by a factor of 4 to 16 due to this association. Salmonella binding to schistosomes occurs via a specific fimbrial protein (FimH) present on the surface on the bacterium. This same fimbrial protein confers the ability of Salmonella to bind to mammalian cells. Conclusions Salmonella can evade certain antibiotics by binding to Schistosoma. As a result, effective bactericidal concentrations of antibiotics are unfortunately above the achievable therapeutic levels of the drugs in co-infected individuals. Salmonella-Schistosoma binding is analogous to the adherence of Salmonella to cells lining the mammalian intestine. Perturbing this binding is the key to eliminating Salmonella that complicate schistosomiasis.


Background
Schistosomes are parasitic helminths that infect humans, with life cycles involving snails as intermediate hosts.
Schistosomiasis occurs in 74 developing tropical and subtropical countries in which over 200 million people are infected. Of those, 120 million patients show symptoms with 20 million severely infected. There are 14,000 deaths per year due to schistosomiasis [1].
Salmonella spp. is a common water-and food-borne cause of gastrointestinal and systemic diseases worldwide. Approximately 2 million individuals die each year from diarrheal disease and Salmonella is a leading cause of this malady. In the U.S. alone, Salmonella causes about 1.4 million infections per year [2].
Concurrent Schistosoma-Salmonella infections occur when enteroinvasive Salmonella enter the systemic circulation and attach to the tegument of adult Schistosoma [3,4] present in the mesenteric vasculature. This interaction apparently provides a refuge in which the bacterium can evade systemic antibiotic therapy. For example, chloramphenicol-sensitive S. typhi were demonstrated to be refractory to chloramphenicol treatment in co-infections [5]. Furthermore, therapy with the anthelmintic praziquantel can lead to a massive release of schistosome-associated Salmonella causing peracute septicemia if the appropriate antibiotic is not co-administered to co-infected children [6,7]. Finally, the use of ineffective antibiotics contributes to antibiotic resistance development and the phenomenon of bacterial persistence.
To assess the nature and extent of this phenomenon, we developed an in vitro antibiotic protection assay in which anti-Salmonella antibiotic efficacy is evaluated using various Salmonella incubated with adult schistosomes and a variety of antibiotics. These studies identify Salmonella factors that facilitate attachment to Schistosoma while also cataloguing the antibiotics that are ineffective against Salmonella adhering to schistosomes.

Results
Assessment of qualitative amoxicillin resistance in Salmonella adhering to schistosomes To determine if Salmonella can evade the effects of antibiotics through adherence to schistosomes, we developed a novel antibiotic protection assay in which antibiotic-sensitive Salmonella typhimurium (strains listed in Table 1) were incubated with a therapeuticallyrelevant concentration of amoxicillin (32 μg/ml) known to be 100% lethal to these strains in the absence of schistosomes [2]. Salmonella typhimurium strains included SL1344 (invasive), BJ68 (non-invasive), and EE419 (hyperinvasive). As part of the control conditions, bacterial killing was determined prior to co-incubation with schistosomes and following adherence to schistosomes. Additional controls included assessing amoxicillin-mediated killing of Salmonella adhering to Girardia tigrina (free-living flatworms) and to HEp-2 mammalian tissue culture cells. A laboratory strain of E. coli (Top10) was also used as a control.
As shown in Figure 1, bacterial survival was significantly higher (p <0.01) during adherence to schistosomes when compared to the other conditions for Salmonella. Specifically, whereas 32 μg/ml amoxicillin kills essentially 100% of the Salmonella strains incubated alone, it kills less than 12% of the same strains incubated with adult schistosomes. This schistosome protective effect was not dependent upon the ability of Salmonella to invade the eukaryotic cells since the non-invasive strain was protected from amoxicillin in a manner similar to that observed for invasive and hyperinvasive strains. The protective effect could not be replicated by co-incubation with the free-living flatworm G. tigrina or with HEp-2 mammalian tissue culture cells. Further, S. mansoni afforded no protection to E. coli.
Quantitative evaluation of amoxicillin resistance in Salmonella adhering to schistosomes To assess the concentration-dependent nature of schistosome-mediated Salmonella resistance to antibiotic killing, we used the antibiotic protection assay to obtain amoxicillin kill curves for Salmonella adhering to schistosomes. Salmonella and Schistosoma were co-incubated in the presence of various concentrations of amoxicillin. Control conditions include determining bacterial killing in the presence of HEp-2 mammalian tissue culture cells.
For most concentrations of amoxicillin, bacterial survival was significantly higher (p <0.01) during adherence to schistosomes when compared to HEp-2 cells adherence ( Figure 2). The apparent minimum inhibitory concentration (MIC), i.e. the lowest concentrations that kills or inhibits the growth of 100% of Salmonella in an in vitro inoculum containing 10 6 bacteria/ml, was 16 μg/ml in the presence of HEp-2 cells but 16 times higher for Salmonella adhering to Schistosoma.

Evaluation of Schistosoma-mediated protection of Salmonella from other antibiotics
To determine if Salmonella can evade the effects of antibiotics other than amoxicillin, studies presented in Figure 1 were repeated using the invasive strain of Salmonella (SL1344) and clinically-relevant concentrations of seven other antibiotics. Other antibiotics included cefepime, cefpodoxime, chloramphenicol, ciprofloxacin, Invasive; adherent to mammalian cells [12] Salmonella typhimurium strain BJ68 Non-invasive; adherent to mammalian cells [13] Salmonella typhimurium strain EE419 Hyperinvasive; adherent to mammalian cells [14] Salmonella paratyphi Adapted to humans; adherent to mammalian cells [9] Salmonella pullorum/gallinarum Adherent to avian cells [8] Salmonella pullorum/pFimH Engineered to be adherent to mammalian cells via heterologous expression of FimH from SL1344

This study
Salmonella typhimurium/dFimH Engineered to be non-adherent to mammalian cells via transposon-mediated deletion of fimH This study

Salmonella javiana
Adherent to mammalian and reptilian cells [15] Shigella flexneri Human pathogen related to Salmonella [19] E. coli Top10 Non-invasive and non-adherent E. coli (ETEC, EPEC, and EHEC) Invasive and adherent streptomycin, sulfadimethoxine, or tetracycline. Cefepime (4 th generation cephalosporin), cefpodoxime (3 rd generation cephalosporin), and ciprofloxacin (fluoroquinolone) represent the antibiotics currently recommended for use against systemic salmonellosis [2]. Chloramphenicol was chosen because of documented treatment failures in Salmonella-Schistosoma co-infections despite chloramphenicol sensitivity exhibited by the Salmonella isolate [5]. Streptomycin, sulfadimethoxine, and tetracycline are other antibiotics that have been previously used to treat systemic salmonellosis [2]. As shown in Figure 3, bacterial survival was significantly higher (p <0.01) during adherence to schistosomes when compared to the other conditions for all antibiotics. Table 2 illustrates the quantitative nature of the resistance where adherence to Schistosoma significantly increased the apparent MIC values beyond the clinical breakpoint, i.e. MIC value that is ascribed as the threshold for sensitivity/resistance, for each antibiotic. Again, this protection was not provided by co-incubation with either G. tigrina or mammalian cells, where in every instance killing was not significantly lower than 100%. Also, this protection was not afforded to E. coli incubated with schistosomes.

Assessment of FimH as a determinant of Salmonella adherence to Schistosoma
Previous studies indicated that Salmonella fimbrial proteins participate in the binding to Schistosoma [3,4]. Fimbrial proteins, specifically FimH, confer selective binding properties to mammalian versus avian hosts. The T78⇒I mutation in FimH confers avian cell binding to Salmonella Figure 1 Amoxicillin resistance in Salmonella adhering to schistosomes (black boxes). Approximately 10 4 CFUs of antibiotic-sensitive Salmonella typhimurium were incubated with a single adult Schistosoma for two hours followed by the addition of the "breakpoint" concentration [16] of amoxicillin (32 μg/ml). Following washing of the antibiotic, Salmonella were liberated from eukaryotes by trypsinization and bacteria were then recovered and enumerated for determination of percent bacterial killing. Salmonella typhimurium strains include SL1344 (invasive), BJ68 (non-invasive), and EE419 (hyperinvasive). Control conditions include determining bacterial killing prior to co-incubation with schistosomes (open boxes) and following adherence to schistosomes (vertical hatching). Additional controls included the assessment of antibiotic-mediated killing of Salmonella adhering to Girardia free-living flatworms (horizontal hatching) and HEp-2 mammalian tissue culture cells (gray boxes). A laboratory strain of E. coli (Top10) was also used as a control. Data presented are the mean ± SEM for three independent experiments each performed in triplicate. *, bacterial survival was significantly higher (p <0.01) during adherence to schistosomes when compared to the other conditions for Salmonella.
pullorum and S. gallinarum [8] whereas the wild-type FimH confers mammalian cell binding to many Salmonella including S. typhimurium and the human-adapted S. paratyphi. Therefore, this mutation may play a role in Salmonella-Schistosoma adherence. In the present study, schistosome-mediated protection from antibiotics was evaluated in Salmonella expressing mammalian cell-binding FimH (threonine at amino acid 78) or avian cell-binding FimH (isoleucine at amino acid 78).
As shown in Figure 4, protection from antibiotics was only afforded to Salmonella expressing mammalian cellbinding FimH. Deletion of FimH abrogated the schistosome-mediated protection for Salmonella typhimurium. Further, Salmonella pullorum, which was not protected by co-incubation with schistosomes, became protected with the heterologous expression of the mammalian cell-binding FimH. Table 3 summarizes the relationship between expression of mammalian cell binding FimH and the antibiotic protection afforded by schistosomes.

Discussion
A previous study documents antibacterial treatment failures in Schistosoma-Salmonella co-infections despite in vitro sensitivity to the drug in the absence of schistosomes [5]. Other studies revealed that Salmonella can adhere to the tegument of Schistosoma and that this adherence is dependent upon fimbriae present on the surface of Salmonella [3,4]. The study herein describes experiments addressing the hypothesis that Schistosomaassociated Salmonella can evade a number of antibiotics by binding to adult schistosomes using fimbrial proteins implicated in mammalian cell adherence.
Antibiotic protection studies revealed that eight different anti-Salmonella antibiotics were incapable of effectively killing or inhibiting the bacterium when adhered to Schistosoma, despite efficacy of these drugs against Salmonella that were not associated with schistosomes. This effect was not related to Salmonella invading the tegumental cells of schistosomes since a non-invasive isostrain was recalcitrant to the antibiotics and a hyperinvasive isostrain did not exhibit any significant elevation of antibiotic resistance. This phenomenon was not observed for Salmonella binding to other eukaryotic cells such as HEp-2 mammalian epithelial cells or tegumental cells on the surface of the free-living flatworm G. tigrina. Bacterial growth was minimal and equivalent * g * Figure 2 Kill-curve obtained for Salmonella incubated with Schistosoma or HEp-2 cells in the presence of various concentrations of amoxicillin. Approximately 10 4 CFUs of antibiotic-sensitive Salmonella typhimurium were incubated with a single adult Schistosoma or HEp-2 cells for two hours followed by the addition of various concentrations of amoxicillin. Data presented are the percentages of Salmonella killed by amoxicillin during attachment to schistosomes (black circles) or HEp-2 cells (red boxes). The native breakpoint for amoxicillin (32 μg/ml [17]) is indicated by the vertical dashed green line. Data presented are the mean ± SEM for three independent experiments each performed in triplicate. *p <0.01 versus HEp-2 controls for 0.5-128 μg/ml of amoxicillin. Similar results were observed for other antibiotics (summarized in Table 2). Schistosoma-mediated protection of Salmonella exposed to amoxicillin, cefepime, cefpodoxime, chloramphenicol, ciprofloxacin, streptomycin, sulfadimethoxine, or tetracycline. Approximately 10 4 CFUs of antibiotic-sensitive Salmonella typhimurium were incubated with a single adult Schistosoma for two hours followed by the addition of the "breakpoint" concentration [17] of an antibiotic. Data presented are the percentages of Salmonella killed or inhibited by antibiotics prior to incubation with schistosomes (open boxes) or in the presence of Schistosoma (black boxes). Data presented are the mean ± SEM for three independent experiments each performed in triplicate. Bacterial survival was significantly higher (p <0.01) during adherence to schistosomes when compared to the schistosome-free condition for a given antibiotic. No significant differences were noted for survival of bacteria prior to schistosome attachment, following attachment to schistosomes, in the presence of HEp-2 cells, or in the presence of Girardia (data not shown). Italicized number indicate resistance to the respective antibiotic. Adherence to HEp-2 cells or Girardia had no effect on the apparent MIC values for Salmonella (not shown). Adherence to schistosomes had no effect on the apparent MIC values for pathogenic and avirulent E. coli (not shown).
for Salmonella adhering to schistosomes, HEp-2 cells, or when incubated with only RPMI media (data not shown). The binding of Salmonella to Schistosoma is dependent upon a specific fimbrial protein expressed by Salmonella that attach to mammalian cells. Various Salmonella serotypes have specific host tropisms and fimbrial proteins dictate this host range phenomenon [8]. Salmonella pullorum and Salmonella gallinarum are avian-adapted Salmonella and schistosomes protected Figure 4 Amoxicillin killing of various Salmonella strains adhering to schistosomes. Salmonella and Schistosoma were co-incubated, exposed to 32 μg/ml of amoxicillin, recovered and enumerated as described above. Data presented are the percentages of Salmonella killed or inhibited by antibiotics prior to incubation with schistosomes (open boxes) or in the presence of Schistosoma (black boxes). Bacterial survival was significantly higher (p <0.01) during adherence to schistosomes for Salmonella expressing the FimH that confers mammalian cell adherence. No significant differences were noted for survival of bacteria prior to schistosome attachment, following attachment to schistosomes, in the presence of HEp-2 cells, or in the presence of Girardia (data not shown). neither of these serotypes whereas broad host range serotypes, like S. typhimurium and S. javiana, were protected from the antibiotics. Mammalian-adapted Salmonella paratyphi, which is adapted to humans [9], is refractory to the antibiotics upon association with Schistosoma. FimH is a fimbrial protein that endows selective eukaryotic cell adherence [8] and our studies revealed that schistosomal adherence is absent for Salmonella that express the avian cell-specific version of this fimbria. It would therefore appear that Schistosoma bear a cell surface epitope that is analogous to the mammalian cell-docking site for FimH.
Since the antibiotic resistance is transient and only occurs during adherence, the mechanism for the resistance is likely a physical barrier that may include a biofilm. The glycocalyx of Schistosoma is thick, highly immunogenic, and fucose-rich [10] suggesting that antibiotics may poorly penetrate this milieu. Given the chemical diversity of the antibiotics used in this study, it is unlikely that the schistosome glycocalyx is capable of chemically altering and inactivating the drugs. Bacterial quiescence is another possibility but Salmonella depend on folate auto-deprivation and the temporal nature of this phenomenon [11] is not likely to occur during the protection assay. The same can be said for the activation of other metabolic changes in Salmonella physiology, i. e., the two hour incubation period is likely insufficient for inducing resistance. Further studies will address the physical basis for the phenomenon described herein.

Conclusions
Schistosoma-associated Salmonella can evade certain antibiotics during adherence to the flatworm. As a result, traditional anti-Salmonella drugs are not useful in coinfected individuals. Bacteria-flatworm binding mimics the adherence of Salmonella to cells lining the mammalian intestine. Perturbing this binding is the key to eliminating Salmonella that complicate schistosomiasis.

Summary of microbes used in this study
Bacterial strains are summarized in Table 1 with Salmonella typhimurium strain SL1344 [12] serving as the model invasive strain. Non-invasive isostrain BJ68 [13] and hyperinvasive strain EE419 [14] are both derivatives of SL1344. Other Salmonella include human-adapted S. paratyphi [9] and the broad host-range S. javiana that infects mammals and reptiles [15]. Avian-adapted S. pullorum was also used in these studies. Bacteria were stored in cryopreservation tubes containing 50% glycerol:50% culture medium at -80°C and grown in LB broth (Sigma) without antibiotics.
To assess the role of specific fimbriae in the adherence of Salmonella to schistosomes, the fimH gene was PCR-amplified from SL1344 and cloned into the prokaryotic expression vector pCR2.1 (Invitrogen). This plasmid, designated as pFimH, was transformed into Salmonella pullorum and the transformant is designated as Salmonella pullorum/pFimH.
Additionally, S. typhimurium was engineered to express the S. pullorum-specific FimH bearing the T78⇒I mutation that abrogates adherence to mammalian cells. First, the TnZeo [16] transposon was inserted into fimH of SL1344 thus preventing expression of native fimH. Since this deletion will have polar effects on the polycistronic fim transcript, the fim operon (fimABCDHFZYW) was PCR-amplified from S. pullorum, cloned into the pCRXL prokaryotic expression vector, and transformed into SL1344 bearing the TnZeo insertion into fimH. This strain is designated as Salmonella typhimurium/dFimH.
Non-Salmonella bacteria included pathogenic Shigella, pathogenic E. coli, and a laboratory strain of E. coli. These bacteria are related to Salmonella but they express fimbriae that are divergent from those found in Salmonella.

Isolation of parasites
Adult male Schistosoma mansoni worms were recovered 45-60 days post-infection from portal and mesenteric veins of Swiss Webster female mice provided by the Biomedical Research Institute, Rockville MD, USA. All animal procedures were conducted in accordance with Iowa State University's approved animal care protocol #07-I-025-A/H.
Following washing of the antibiotic after 2 hrs, trypsin (0.05%) was then added in order to cleave the Salmonella fimbriae that mediate the attachment to Schistosoma [3,18]. Media was then removed and plated on Salmonella-selective agar that was incubated at 37°C overnight. Salmonella colonies were then enumerated for determination of percent of bacteria recovered, i.e. an indirect measurement of bacterial killing. Percent bacterial killing equals 100 ×(10 4 minus the number of CFUs recovered)/10 4 .
Control conditions included determining bacterial killing prior to co-incubation with schistosomes, during adherence to schistosomes, and following adherence to schistosomes. The latter situation refers to worms that were removed from the schistosomes and then examined for antibiotic susceptibilities, a measurement of the duration (or lack thereof) of the resistance afforded by schistosomes. Additional controls included the assessment of antibiotic-mediated killing of Salmonella adhering to Girardia free-living flatworms and HEp-2 mammalian tissue culture cells.
Concentration-response analysis of antibiotic resistance exhibited by Schistosoma-associated Salmonella typhimurium SL1344 Various concentrations of antibiotics (0-1,024 μg/ml) were used in the antibiotic protection assay described above. Percent bacterial killing was determined for SL1344 adhering to schistosomes, Girardia, or HEp-2 cells. As a control, bacterial killing was assessed in eukaryotic cell-free assays and in the presence of HEp-2 cells. Apparent antibiotic minimum inhibitory concentrations (MICs) were ascribed to the lowest concentration of antibiotic that inhibited or killed 100% of the bacterial inocula.

Statistics
Statistical analysis was performed using ANOVA with Scheffe's F test for multiple comparisons. Comparisons were made between strains and between incubation conditions.