Water Buffalo were purchased from Huilong county, Hunan province, from an area with no history of schistosomiasis transmission, and transported to the field site. Age, sex and weight of the different water buffaloes used in this study are summarized in the Supplementary Tables 1–4. Upon arrival at the field site, the buffalo were quarantined for 4 weeks, confirmed schistosome-free by the miracidial hatching test and ELISA, and treated with levamisole to eliminate other gastrointestinal geohelminths, as described (16). Water buffalo were then tagged with an identification number on the collar, the right ear, and into one of the horns, then divided randomly into different trial groups such that buffalo in each trial cohort had similar numbers of male and female buffalo and were of similar age and weight (Supplementary Table 1). The study site selected for the vaccine trials was Huaqiao village, Linxiang city, Hunan province.

All DNA vaccine constructs were constructed as described in Da'dara et al. (16). GLP standard plasmids (SjC23-Hsp70 and SjTPI-Hsp70) including control plasmid pVAX1, and pUMVC3-hIL-12 plasmid DNA encoding human IL-12 (pIL-12) were manufactured by Aldevron (Fargo, ND). Recombinant SjTPI (rSjTPI) and SjC23 (rSjC23) were purified using similar protocols as previously described (16). Briefly, for SjC23, the large extracellular hydrophilic domain (amino acids 108–183) of SjC23 (SjC23-loop) was expressed in Escherichia coli using the prokaryotic expression plasmid pET32b (Novagen, Madison, WI, USA). For SjCTPI, the pTrcHisB plasmid (Invitrogen) was used to produce full-length recombinant SjCTPI. Recombinant SjC23 and SjCTPI proteins were purified using His-Trap HP columns (GE Healthcare, Piscataway, NJ, USA). Purified recombinant proteins were analyzed by SDS-PAGE and Western blotting.

Initially, we compared immunogenicity and efficacy of in-lab produced SjC23 plasmid DNA vaccines to commercially produced, GLP grade, SjC23 plasmid DNA vaccines. The trial was a randomized double-blinded study, comprised of 45 buffalo divided randomly into three groups. Plasmid DNAs were coded by an independent laboratory at the Harvard School of Public Health (HSPH), then shipped to China. Animals were immunized intramuscularly with 1 ml of 300 μg of schistosome-DNA or control pVAX vaccine plus 300 μg of IL-12 plasmid DNA for priming (Figure 1). One month after the last immunization, water buffalo were challenged with 1,000 freshly shed cercariae kindly provided by Jiangsu Institute of Parasitic Diseases, China. Eight weeks post-challenge, water buffalo were euthanized and adult worm burdens determined by perfusion. Additionally, when possible, identical portions of liver lobes were obtained to determine liver egg burdens, and fresh fecal samples were obtained to determine fecal egg counts and miracidial hatching rates as previously described (16).

The second goal of this study was to determine if extending the time between prime and boost would improve DNA vaccine efficacy. Animals were immunized intramuscularly with schistosome-DNA or control DNA vaccines as previously described (Figure 2A). We compared 1- and 3-month intervals between prime and boost. One month after the last immunization, water buffalo were challenged with 800 cercariae freshly shed from native/local, infected snails.

We then examined the importance of the IL-12 plasmid DNA for induction of optimal levels of efficacy. Buffalo were immunized as previously described with schistosome-DNA or control DNA vaccines with or without the pUMVC3-hIL-12 plasmid DNA (Figure 3). The pUMVC3-hIL12 was administered at the prime only. One month after the last immunization, water buffalo were challenged with 800 cercariae.

Finally, we evaluated efficacy of a heterologous plasmid DNA prime with a recombinant protein boost. In this trial, we employed combination SjTPI/SjC23 vaccines. Buffalo were primed as described with 300 μg of both the SjC23 and SjTPI DNA vaccines or control DNA vaccine. Booster vaccination was with plasmid DNAs or with 100 μg each, recombinant SjC23 and SjTPI proteins (Figure 4A). One control group received control pVAX DNA as prime and were then boosted with recombinant antigens. One month after the last immunization, water buffalo were challenged with 1,000 freshly shed cercariae.

Two months post-challenge, water buffalo were weighed then euthanized. Adult worms were obtained from the portal vein by perfusion of the descending thoracic aorta with physiologic saline. Worms trapped in intestinal tissue or caught up in tissues near where the portal vein was cut, were collected by hand and counted. All adult worms recovered from each water buffalo were counted and recorded as total worm burdens.

Two samples from the left lobe and one sample from the right lobe were taken from the livers of each water buffalo and weighed. Each piece was placed in 20 ml of 4% (w/v) KOH for 2 days at 37°C. The suspension was then agitated to re-suspend the mixture, 1 ml of the liver suspension was collected, centrifuged at 3,000 rpm for 1 min, the pellet re-suspended in 200 μl of PBS and the total eggs present were counted under a microscope. Using the average of three separate samples, the number of eggs per gram of liver tissue were calculated for each animal.

Three fecal samples were collected from the rectum of each water buffalo 2 days and 1 day prior to perfusion and again on the day of perfusion. For each fecal sample, the number of eggs per gram (EPG) of stool was determined by microscopy and the number of hatched miracidia per gram (MPG) of stool was determined as previously described (20).

Serum samples from four different buffalo from each group were randomly selected then analyzed by Western blot and ELISA to determine the antigen-specific antibody levels in sera from vaccinated buffalo as described (16).

GraphPad Prism version 5 was used for statistical analysis. One-way ANOVA with Tukey's post-test or two-way ANOVA and Bonferroni's post-test were used to identify the significance between various groups. Statistical significance was accepted at p < 0.05.

In order to meet potential larger vaccine doses and reduce costs, we contracted with Aldevron, a leading company in the production of plasmid DNA and recombinant proteins, to produce each DNA vaccine and the control plasmid DNA under GLP conditions. To determine if GLP grade production would influence vaccine efficacy, we compared the Aldevron GLP grade vaccines to our in-house DNA vaccines using the regimen outlined in Figure 1A. Vaccination of water buffalo with the Sj23-Hsp70 construct produced at the Harvard School of Public Health or from Aldevron resulted in nearly identical levels of efficacy, 23.5 and 25.7% respectively, with no significant differences of mean adult worm burdens (p = 0.816) between these two vaccine groups as determined by One-ANOVA with Turkey's post-test (Figure 1B). Interestingly, we also observed lower reductions in miracidial hatching rates of 19 and 10% respectively (Figure 1C). Table 1 summarizes this data.

Previous studies have reported that extending the interval between priming and boosting can influence vaccine efficacy (21–24). To test this, we increased the time between prime and boost from 1 to 3 months. Using the SjC23 DNA/DNA prime boost-boost with 1 or 3-month intervals, we found that efficacy, as determined by worm burden, liver eggs/gram and fecal eggs/gram, was higher in animals where boosts were given at 1-month intervals (Table 2 and Figure 2). We observed significant reductions in adult worm burdens of 37.2 and 41.2% and reductions in eggs/gram liver of 30.4 and 37.5% for SjC23 DNA vaccinated groups at extended (3 month) or regular (1 month) intervals, respectively (Figures 2B,C). The reduction in eggs per gram stools of 31.5 and 45.8% for extended and regular month interval groups (Figure 2D). However, there was no statistically significant difference between the extended and the regular immunization regimen in all measured parameters.

IL-12 is an important cytokine that has the potential to activate natural killer cells and promote cytolytic T cell proliferation and has been used as an adjuvant to promote Th1-type and cellular immune responses in various DNA vaccine trials (25–29). We previously reported that administration of IL-12 with the boost enhanced host immunity and better protective efficacy in water buffalo (16, 19). However, we did not know if our DNA vaccines would retain similar protection without IL-12. The vaccination trial in water buffalo performed here showed that addition of Plasmid DNA encoding IL-12 enhanced all aspects of DNA vaccine efficacy (Table 3). Adult worm burdens were reduced by 46.6% in buffalo co-administered plasmid DNA IL-12 compared to 30.2% in buffalo vaccinated with SjC23 plasmid DNA and no IL-12 (Figure 3B). Similarly, the number of eggs/gram liver tissue was reduced 45% in buffalo co-administered plasmid DNA IL-12 compared to 31% in buffalo vaccinated with SjC23 plasmid DNA and no IL-12 (Figure 3C). A similar influence of co-administering plasmid DNA encoding IL-12 was seen in eggs/gram stool (Figure 3D). Overall, efficacy data from the DNA prime-boost experiments evaluating interval between prime and boost or value of the IL-12 plasmid were similar to what we previously reported for our plasmid DNA vaccines (16) and show that co-administration of the IL-12 encoding plasmid enhances DNA vaccine efficacy, yet not statistically significant from the no IL-12 plasmid.

Boosting with recombinant proteins following a DNA prime immunization elicits a broad humoral immune response and enhances the cell-mediated immune responses in mice and non-human primates (30–34). Therefore, here we moved from a three-dose plasmid DNA vaccine regimen to a two-dose plasmid DNA prime and recombinant protein boost regimen with a 3-month interval between prime and boost (Figure 4A). Groups of water buffalo were primed with pVAX control plasmid DNA, SjC23-Hsp70 plasmid DNA, or with a combination of SjC23-Hsp70 and SjTPI-Hsp70 plasmid DNAs. Three months post-prime, groups of water buffalo were boosted with recSjC23 or a combination of recSjC23 and recSjTPI (Figure 4). The results of these trials demonstrated that buffalo only receiving recSjC23 following the control pVAX prime had adult worm burdens reduced by 32.2% compared to pVAX primed and boosted buffalo (Table 4; Figure 4B). Similarly, eggs/gram liver were reduced by 31.0% in buffalo that only received the SjC23 boost following control pVAX prime (Figure 4C). Vaccine efficacy significantly increased in the two-dose regimen when buffalo were primed with SjC23-Hsp70 DNA then boosted with recSjC23. Here, adult worm burdens were reduced by 50.0% compared to controls and eggs/gram liver were reduced 47.0% compared to control vaccinated buffalo. However, we observed the highest levels of efficacy in this two-dose regimen trial when buffalo were primed with the combination of SjC23-Hsp70 and SjTPI-Hsp70 plasmid DNAs followed by boost with a combination of recSjC23 and recSjTPI. This combination yielded the highest level of vaccine efficacy that we have observed to date for these candidate vaccines in water buffalo. Adult worm burdens were reduced by 55.1% and eggs/gram liver by 53.2% compared to levels in control vaccinated buffalo (Figures 4B,C). Using DNA prime/protein boost, the reduction in worm burdens (Figure 4B), eggs in liver tissues (Figure 4C) and hatched miracidia in stool (Figure 4D) were dramatic compared to that in the control vaccinated group (p < 0.0001).

We next evaluated sera from control and immunized animals for anti-SjC23 or SjTPI specific antibodies using Western blots and ELISA (Figure 5). We randomly picked four or more individual serum samples from each group to detect antigen-specific antibodies using Western blot. Sera from animals primed with SjC23-Hsp70 alone in in combination with SjTPI-Hsp70 then boosted with the correlate recombinant antigen elicited specific antibody responses to the respective antigens (Figure 5A). In contrast, sera from animals primed with the control pVAX plasmid and boosted with either pVAX or recombinant proteins did not react with rSjC23 (Figure 5A left panel, lanes 1-8) or rSjTPI (Figure 5A right panel, lanes 1-8). Individual sera collected at different time points from animals were used in ELISA. ELISA results showed that antibody levels are significantly increased after boosting with recombinant proteins in group 3 and 4 (Figure 5B).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.