Rotavirus is a major cause of infectious diarrhea in infants and young children and is one of the main reasons for their hospitalization (1,2). The disease is of major concern in childcare centers (3–5) and pediatric hospitals (6,7), where rotavirus cases may occur sporadically as outbreaks of diarrhea. Several studies have shown that, when given passively, hyperimmune bovine colostrum (HBC) derived from cows immunized with rotavirus can prevent or treat rotavirus diarrhea in humans (8–11). This indicates that orally administered antibodies must be able to survive the hostile environment of the upper gastrointestinal tract, where protein degradation can occur as a result of acid hydrolysis and enzyme digestion (12). The extent of such degradation is thought to be influenced by factors such as the buffering capacity of the product formulation, product resistance, time of ingestion in relation to food intake, and pH in various sections of the gastrointestinal tract. Petschow et al. (12) reported that the biological activity of bovine colostrum antibodies was reduced by exposure to acid and trypsin. In his study, limited proteolysis of bovine colostral immunoglobulin G by trypsin was found to cause loss of specific antibody activity with relatively little fragmentation into smaller molecular weight components.
Little work has been done to determine the survival of antibodies after their passage through the gastrointestinal tract. Hilpert et al. (13) administered concentrated antirotavirus hyperimmune bovine immunoglobulin G to 164 children up to 2 years of age who were hospitalized with acute gastroenteritis. The results suggested that ∼10% of the orally administered antiviral activity was detectable in diarrheal stools and that it took approximately 36 hours for the concentrate to pass through an the digestive tract of an infant. Blum et al. (14) reported that between 4–12% of an oral dose of immunoglobulin G prepared from human serum and administered to 6 immature infants was recovered undigested or partially digested in stools in all but the control infants.
To utilize the full potential of hyperimmune bovine colostrum containing rotavirus antibodies (HBC-R), it was important to determine whether the specific antibody activity survives transit through the gastrointestinal tract. Therefore, this study was conducted to measure the level of rotavirus antibody activity in the feces of subjects receiving different doses of HBC-R, using a novel enzyme-linked immunosorbent assay (ELISA)-based assay.
MATERIALS AND METHODS
Subjects and Study Design
Written informed consent was obtained from parents or guardians of all study participants. The study protocol was approved by the Women's and Children's Hospital Research and Ethics Committee, Adelaide, South Australia, Australia. This study was conducted in accordance with Good Clinical Research Practice (GCRP) in Australia (1991), the Declaration of Helsinki (1989) and ICH Harmonized Tripartite Guidelines for Good Clinical Practice (1996). Subjects participating in the study were aged up to but not including 4 years of age. The mean age of subjects across all study groups was ∼25 months, (range, 8–48 months). The subjects were recruited from nine childcare centers in Adelaide, South Australia. Subjects were eligible for recruitment if they: 1) could tolerate whole cow's milk, 2) routinely opened their bowels once a day, 3) were not breast fed, 4) did not experience diarrhea within the previous 72 hours, and 5) had no detectible levels of either rotavirus antigen or antibody in their stools. A total of 105 subjects met the eligibility criteria. Over half of the subjects involved in the study were male, 60 (57%) of 105. There were five experimental groups in the study (Table 1). Three groups received specific amounts of HBC-R in liquid form, one group received a reconstituted powder form of HBC-R, and the control group received whole cow's milk. Each center had its own control group, and data from all the control groups were combined for analysis. Subjects were asked to drink 100 ml of whole milk supplemented with HBC-R 3 times a day throughout the day, for a period of six days. Stool samples were collected from the subjects before, during, and after consumption of the study product to assess the concentration of rotavirus antibody activity as well as rotavirus infection. Samples taken up to 48 hours before the study began were used as negative controls for each subject. No product related serious adverse events, adverse reactions, or adverse events were reported during the study period.
Preparation of the Study Product
Pregnant cows were vaccinated with an approved inactivated rotavirus vaccine. The liquid form of HBC-R was prepared by pooling colostrum to obtain a minimum ELISA titer of 1:52,000, then subjected to low heat treatment, skimmed, and centrifuged to reduce bioburden. The powder form of HBC-R was prepared by pooling colostrum to obtain a minimum ELISA titer of 1:32,000. This was followed by downstream ultrafiltration and spray drying of the liquid colostrum. The liquid or powder colostral products were mixed with the whole milk to achieve the appropriate rotavirus antibody titer (Table 1). The milk was sampled and tested to ensure microbial quality met food standards code for fresh milk and for rotavirus antibody titer.
Determination of Rotavirus Antibody Activity
The concentrations of rotavirus-specific immunoglobulin antibody in colostrum were routinely measured using an antibody capture ELISA described by Duermeyer et al. (15), with minor modification. Briefly, immuno plates (96 well, MaxiSorpC, NUNC, Roskilde, Denmark) were coated with goat antibovine immunoglobulin G (H+L, Pierce, Carpinteria, California, U.S.A.), to bind bovine immunoglobulin G in the colostral samples. Rotavirus and horseradish peroxidase-labeled rabbit antirotavirus (human) conjugate (P 0219 DAKO, Denmark) were each added sequentially, with washing between each reagent. The substrate was TMB, (3,3´5,5´-tetramethyl-benzidine dihydrochloride, Sigma). Absorbency was read at 450 nm (reference 610 nm) using the automated Dynatech MR5000 spectrophotometer. Antibody titers were calculated from standard curves generated using colostrum samples that contained high concentrations of antibody specific for rotavirus.
Fecal Extracts Preparation
A 40% suspension of each fecal sample was prepared in phosphate-buffered saline (PBS), vortexed, and centrifuged for 30 min at 5,480 g at 4°C. The supernatant was collected and stored at −20°C.
Human rotavirus culture (Wa strain serotype G1, National Institute of Health, Bethesda, MD, U.S.A.) was produced using the method previously described by Kok et al. (16). The virus was harvested and virus titration was performed to assess viral titer. The viral concentration was also determined by modification of the IDEIA™ rotavirus kit (DAKO, K6020), as described below. The virus was concentrated using the Amicon concentrating system to obtain a final concentration of 4.2 × 108 rotavirus particles (rv) ml. The rotavirus stock was stored at −70°C.
Virus Reduction (VR) ELISA
The rotavirus antibody activity in feces was detected by the novel VR ELISA that was set up for the purpose of the study. The assay measures the capacity of fecal material to block the detection of rotavirus in an antigen capture ELISA. Briefly, the rotavirus antiserum was produced using previously published procedures (16). MaxiSorp plates were coated with rabbit rotavirus antiserum in a coating buffer (0.05M carbonate buffer, pH 9.6). Plates were sealed and stored overnight at 4 °C. Fecal extracts were thawed and mixed (1:1) with PBS + 0.2% Tween 20 (TW), sonicated at low frequency ∼ 15 kHz and stored at 4 °C overnight. The standard curve covering the range from 105–l08 rv/ml, was prepared in PBS + 0.05% TW containing 0.25% bovine serum albumin (BSA, CSL, Melbourne, Australia) and added in duplicate onto the Microtest plate (SARSTEDT, Adelaide, Australia). Three dilutions of the rotavirus were prepared in diluent 0.5% BSA. Each virus dilution was added (60 μl/well) into three columns on the plate. The diluent was added to the last column (60 μl/well) to detect a naturally occurring rotavirus infection in each fecal extract. Sonicated fecal extracts were then added at 60 μl/well to obtain one fecal extract per row. The resulting rotavirus concentrations after the addition of fecal samples were: 5 × 106, 1 × 107, and 5 × 107 rv/ml. The first fecal extract on each plate was the negative control followed by the test fecal extracts. The plates were shaken and the virus-fecal extract mixture incubated at 37°C for 1 hour. Following the incubation, the virus-fecal extract was mixed using a multichannel pipette before being transferred (100 μl/well) to precoated washed MaxiSorp plates.
The plates were incubated at 37°C for 1 hour and then washed three times with the diluent to remove unreacted sample material. Rabbit antirotavirus peroxidase-conjugated (P0219, DAKO) was added to each well (100 μl/well). Plates were incubated at 37°C for 1 hour and washed twice with diluent and the substrate buffer (0.2M sodium acetate, pH 6.0–6.1). Chromogen TMB was added to each well at 100μl/well. The plates were incubated for 5 min at 25 °C, and the reaction stopped by addition at 50 μl/well of 2M sulphuric acid. Absorbency was read at 450 nm (reference, 610 nm). The viral concentration was calculated from the absorbance readings of the viral suspensions and the standard curve, using Biolinx software (Dynatech Laboratories, Inc., Chantilly, VA, USA).
Calculations for VR ELISA
Readings from the standard curve represented the amount of rotavirus left after the reaction had taken place between the antigen delivered to the assay and the antibody activity present in the fecal extracts. Percent virus reduction (% VR) was calculated using the formula: % VR = 100 − [(virus titer for test sample/virus titer for negative control) × 100]. The negative control rather than 100% rotavirus was used in the above calculations to take into account the possible “blocking effect” of the fecal extracts on the virus used in the assay.
Virus Neutralization (VN) Assay
The VN assay was adapted from the method previously described by Ward et al. (17), with minor modifications. Four twofold dilutions (40% →5%) of fecal extract, rotavirus (serotype G1 Wa ) with 50 μg trypsin (1:250 GIBCO), and Ma104 cells were used in the VN assay. After overnight incubation, the rotavirus antigen titer was determined as described previously in this paper, as part of VR ELISA method.
Calculations for VN
Readings from the standard curve represented the yield of rotavirus detected in each well after inoculation with virus-fecal extract mixtures. Percent virus neutralization (% VN) was calculated using the following formula: % VN = 100 − [(virus titer for test sample/virus titer for negative control) × 100]. The negative control rather than 100% rotavirus was used in the previous calculations to take into account the possible “blocking effect” of the fecal extracts on the yield of rotavirus used in the assay.
Analysis was performed using Systat for Windows (version 7.0, Hearne Scientific Software, Melbourne, Australia). An average of 20 subjects per group was sufficient to give a statistically significant result for power of test = 0.8 and P ≤ 0.05. Before the main statistical analysis, the results were subjected to stem-and-leaf plots designed to assess the distributional shape of the data and to identify outliers. For regression and correlation analyses, stem-and-leaf plots also were performed on the residuals to identify further outliers in the dependent variable. A total of 18 of 105 subjects identified as outliers, and they were removed from further analysis. All tests of significance were two-tailed at P ≤ 0.05. In the final analysis, the following variables were used: 1) the dose of HBC-R ingested, i.e. rotavirus antibody titer administered to each child and averaged over a period of 5 days (days 2–6), and 2) the rotavirus antibody activity from each fecal sample expressed as the percent reduction of 5 × 106 rv/ml and averaged over a period of 5 days (day 2–6). Two experiments were conducted to determine the relationship between the percent reduction and percent neutralization of the rotavirus by fecal extracts. Data for each experiment were analyzed by two-way analysis of variance to ascertain whether the results of the two experiments could be pooled, and then further analysis was performed on the combined data.
A summary of the demographics of the subjects, i.e., the number of subjects per group and mean product intake is shown in Table 1. Across all 5 of the study groups, the mean milk consumption was 277 ml (± 7) and the mean number of bowel actions recorded per day was 1.8 (± 0.16), which is similar to that reported by Tham et al. (18). During the study ∼1,500 fecal samples were tested for the presence of rotavirus antigen or antibodies. Rotavirus antibody activity was detected in 25% of all fecal extracts tested as early as 8 hours after the first ingestion of HBC-R. The percent of fecal extracts with detectible antibody activity increased to 86% during the next 5 days of the study while HBC-R was administered. Rotavirus antibody activity could still be detected in 63% of fecal extracts up to 72 hours after the final ingestion of the HBC-R product.
Fecal extracts from the five study groups had detectable levels of rotavirus antibody activity when tested against the 3 concentrations of rotavirus (Table 1). The very low concentrations of antibody activity detected in the control group were caused apparently by the presence of antibody or other naturally occurring proteins, which caused a “blocking effect”. The HL and P groups had relatively high rotavirus antibody activity detected in the feces even at the highest rotavirus concentration tested, i.e. 5 × 107 rv/ml. Dose equivalence for fecal rotavirus antibody activity was statistically demonstrated between the study groups HL and P (P = 0.44, n = 7 and 11, respectively).
There was a strong and positive relationship (r = 0.81) between the percent reduction of rotavirus by fecal extracts containing specific antibody activities and the titer of rotavirus antibody administered to subjects (Fig. 1). This relationship was one of functional dependence as indicated by the calculated regression line, which suggested that 65% of variation in the percent reduction of rotavirus could be explained by the variation in the administered dose of the rotavirus antibody. The relatively lower value of the coefficient of determination (R2) in comparison to that of the correlation coefficient is thought to be caused by natural variation in responses to the study product, which, in turn, affected the linearity of the data and, thus, its predictive strength.
It also was apparent from the results in Figure 1 that even within the same group the level of rotavirus antibody activity in fecal samples varied markedly from one individual to another. The pattern of fecal frequency and the level of rotavirus antibody activity in stools for 4 representative subjects receiving powder are shown in Table 2. For the majority of individuals, the levels of rotavirus antibody activity in feces varied from one stool collection to the next. Despite the variation in virus reduction activity (27–96%) observed in Table 2, 521 (86%) of 602 fecal samples collected from subjects consuming milk product supplemented with HBC-R retained sufficient activity to neutralize up to 81% of 5 × 106 rv/ml.
The relationship between the percent reduction of rotavirus by VR ELISA and the percent of rotavirus neutralization by VN, using fecal extracts of subjects taking HBC-R, was also investigated (Figure 2). The results of the two assays correlated strongly (r = 0.96) and were of high functional dependence (R2 = 0.92), providing evidence for the rotavirus neutralizing capacity of the fecal extracts under the test conditions.
The results of this study clearly show that rotavirus antibody activity survived passage through the gastrointestinal tract and can be detected in the feces of the subjects who ingested HBC-R supplemented milk. We also have shown that the rotavirus antibody activity detected in feces is directly proportional to the titer of rotavirus antibody administered to subjects. Furthermore, we have provided evidence for the rotavirus neutralizing capacity of the fecal extracts of subjects consuming HBC-R.
During the trial, the level of rotavirus antibody activity in feces varied considerably between subjects, and from one fecal sample to the next. Such variation is thought to reflect differences in dietary intake and gut motility as well as the variable effects of gastric acid and proteolytic enzymes (12). Considering this variation, however, 74 (88%) of 84 subjects who consumed milk supplemented with HBC-R retained sufficient activity to neutralize up to 81% of 5 × 106 rv/ml. This result was considered to be extremely encouraging because the subjects were not supervised closely during this community-based trial, and occasional lapses with product intake were not uncommon. Even more encouraging was the fact that rotavirus antibody activity could be detected as early as 8 hours after ingestion of HBC-R (in contrast to 36 hours as reported by Hilpert et al (13)) and up to 72 hours after consumption had ceased. That alone is an added benefit of HBC-R because it may provide not only an early but also an extended protective effect after its consumption.
Because of difficulties associated with complete recovery of fecal specimens in this community-based trial, it was not possible to fully determine the proportion of rotavirus antibody activity surviving transit through the gut. However, from the recovered fecal specimens, we calculated that ∼5% (mean activity, four groups, 602 samples) of relative rotavirus antibody activity had survived transit through the gut. This result is in agreement with the findings of Hilpert et al. (13) and Blum et al. (14), who reported similar survival of antibody activity (4–12%).
We have also determined that the form of colostrum (fresh liquid vs. spray-dried powder and reconstituted) seemed to have little effect on survival of antibody activity in vivo. The level of rotavirus antibody activity detected in fecal samples from subjects who consumed reconstituted powder colostrum (P group) was found to be not significantly different from that of subjects who consumed liquid colostrum (HL group). Therefore, assuming that the subjects from P and HL groups consumed the same volume of study product with the equivalent antibody titer, the expected rotavirus antibody activity in feces should be similar for the two groups.
Previous attempts to detect immunoglobulin and rotavirus antibody in feces using conventional immunoassays, such as ELISA and immunoblots, were unsuccessful. This was taken as an indication that the active moiety may have been an antibody fragment Fab and F(ab`)2 rather than intact immunoglobulin (13). The importance of functional assays in demonstrating antiviral activity in feces cannot be overstated. Our novel virus reduction ELISA, with its high sensitivity, specificity, accuracy, and precision, allows detection of rotavirus antibody activity in feces. The assay is cost and time effective, and allows a large number of samples to be tested at a given time. The assay is also highly comparable with the standard virus neutralization assay.
The findings presented in this paper suggest that passive immunotherapy (pathogen specific) may be used to prevent or treat infectious gut diseases along the entire length of the gastrointestinal tract. Furthermore, it is likely that rotavirus antibody activity in the small intestine (where most viral and bacterial infections have their major impact) will be considerably higher than the activity in the feces, providing even greater protection or treatment against gastrointestinal diseases. Passive protection is at present the only effective means of preventing rotavirus infection in children, especially in hospitals, childcare centers, and in the community.
The authors thank the study nurses, childcare center staff, and parents for their participation in the study, and Professor D. Rowley and Dr. P. A. Marshall, NorthField Laboratories Pty Ltd, for their advice and assistance in reviewing the manuscript. Associate Professor Geoffrey Davidson was the principal investigator for the study and also is a member of the NorthField Laboratories Pty Ltd Scientific Advisory Committee.
1. Glass RI, Kilgore PE, Holman RC, et al. The epidemiology of rotavirus diarrhea in the United States: surveillance and estimates of disease burden. J Infect Dis 1996; 174 (Suppl 1):S5–11.
2. Parashar UD, Holman RC, Clarke MJ, et al. Hospitalizations associated with rotavirus diarrhea in the United States, 1993 through 1995: surveillance based on the new ICD-9-CM rotavirus-specific diagnostic code. J Infect Dis 1998; 177:13–17.
3. Ferson MJ, Stringfellow S, McPhie K, et al. Longitudinal study of rotavirus infection in child-care centres. J Paediatr Child Health 1997; 33:157–60.
4. Bartlett AV, Moore M, Gary GW, et al. Diarrheal illness among infants & toddlers in day care centers. I. Epidemiology and pathogens. J Pediatr 1985; 107:495–502.
5. Hjelt K, Paerregaard A, Nielsen OH, et al. Acute gastroenteritis in children attending day-care centers with special reference to rotavirus infections. I. Aetiology and epidemiological aspects. Acta Paediatr Scand 1987; 76:754–62.
6. Ringenbergs ML, Davidson GP, Spence J, et al. Prospective study of nosocomial rotavirus infection in a paediatric hospital. Aust Paediatr J 1989; 25:156–60.
7. Carlin JB, Chondros P, Masendycz P, et al. Rotavirus infection and rates of hospitalisation for acute gastroenteritis in young children in Australia, 1993–1996. Med J Aust 1998; 169:252–6.
8. Davidson GP, Whyte PBD, Daniels E, et al. Passive immunisation of children with bovine colostrum containing antibodies to human rotavirus. Lancet 1989; 2:709–12.
9. Turner RB, Kelsey DK. Passive immunization for prevention of rotavirus illness in healthy infants. Pediatr Infect Dis J 1993; 12,718–22.
10. Davidson GP. Passive protection against diarrheal disease. J Pediatr Gastroenterol Nutr 1996; 23:207–12.
11. Ylitalo S, Uhari M, Rasi S, et al. Rotaviral antibodies in the treatment of acute rotaviral gastroenteritis. Acta Paediatr 1998; 87:264–7.
12. Petschow BW, Talbott RD. Reduction in virus-neutralizing activity of a bovine colostrum immunoglobulin concentrate by gastric acid and digestive enzymes. J Pediatr Gastroenterol Nutr 1994; 19:228–35.
13. Hilpert H, Brussow H, Mietens C, et al. Use of bovine milk concentrate containing antibody to rotavirus to treat rotavirus gastroenteritis in infants. J Infect Dis 1987; 156:158–66.
14. Blum PM, Phelps DL, Ank BJ, et al. Survival of oral human immune serum globulin in the gastrointestinal tract of low birth weight infants. Pediatr Res 1981; 15:1256–60.
15. Duermeyer W, Van der Veen J. Specific detection of IgM-antibodies by ELISA, applied in Hepatitis-A. Lancet 1978; 2:684–5.
16. Kok TW, Sharley A, Payne L, et al. A composite, in-house enzyme immunoassay for detection of rotavirus and adenovirus in fecal samples. Aust J Med Sci 1998; 19:56–9.
17. Ward RL, Kapikian AZ, Goldberg KM, et al. Serum rotavirus neutralizing-antibody titer compared by plaque reduction and enzyme-linked immunosorbent assay-based neutralization assays. J Clin Microbiol 1996; 34:983–5.
18. Tham EBA, Nathan R, Davidson GP, et al. Bowel habits of healthy Australian children aged 0–2 years. J Paediatr Child Health 1996; 32:504–7.