Introduction
Bovine respiratory disease continues to impair animal welfare and cause substantial economic losses owing to morbidity, reduced performance, and death of cattle in both the beef and dairy industries.1 Total prevention of BRD remains an elusive goal. Vaccination is an important adjunct to improvements in management, nutrition, and biosecurity. For some animals, such as young calves with circulating maternal antibodies, vaccine administration by the IN route may result in a superior immune response than vaccine administration by other routes. Indeed, IN administration of MLV vaccines is used in many cattle production systems, and several commercial vaccines have been licensed for IN administration to cattle.2 Administration of MLV vaccines by the IN route induces acquired immune responses including antibody and cell-mediated immunity to specific pathogens,3,4,5,6,7,8,9 and some MLV vaccines labeled for IN administration also invoke innate immune responses characterized by the production of IFNs.3,4
In bovids, there are 3 distinct subtypes of IFNs (type I [9 α and 3 β genes], type II [1 γ gene], and type III [1 λ gene]), which have the capacity to inhibit the replication of many viruses.4,10,11 In some situations, it may be beneficial to stimulate the innate immune system (manifested in part by the production of IFNs) of animals to increase their resistance to viral infections.12 In a study11 in which BHV1-naïve calves were experimentally infected with BHV1, concentrations of IFN-α, IFN-β, and IFN-γ in nasal secretions became increased from baseline concentrations within 2 to 3 days after infection, peaked at approximately 5 days after infection, and returned to baseline concentrations by 7 to 10 days after infection. Similarly, in cattle that receive an IN MLV vaccine with a BHV1 component, measurable IFN responses typically become detectable 2 to 3 days after vaccination, peak approximately 5 days after vaccination, and decline to baseline levels at approximately 10 days after vaccination.3,4 Thus, the kinetics of the IFN response to a primary MLV vaccine against BHV1 is similar to that observed following a virulent BHV1 infection of the respiratory tract.
Currently, there are 2 commercially available, multivalent MLV vaccines that contain a BHV1 component and are approved for IN administration to cattle. One of those vaccines (vaccine A) contains modified-live strains of BHV1 (strain JSAV) and parainfluenza virus type 3. The other vaccine (vaccine B) contains a modified-live strain of BRSV and temperature-sensitive strains of BHV1 (strain RLB 106) and parainfluenza virus type 3. Results of a study9 in which cows received vaccine B at either 14 days prior to their projected calving date or on the day of calving indicate that the vaccine did not induce significant increases in IFN-α, IFN-β, or IFN-γ concentrations in nasal secretions of either group, although the mean concentration of IFN-γ in nasal secretions differed significantly between the 2 treatment groups on the day of vaccination (baseline) and throughout the observation period. In that study,9 the cows vaccinated on the day of calving had a significant increase in anti-BHV1 IgA titers following vaccination, whereas the cows vaccinated 14 days before their projected calving date did not. Those observations suggest that the BHV1 fraction of vaccine B induces an acquired immune response but does not induce a detectable innate immune response in adult cows.
The primary goal of the study reported here was to compare the innate (IFN-α and IFN-γ) and acquired (secretory IgA antibody) immune responses in the upper respiratory tract (ie, nasal secretions) between calves that were IN vaccinated with vaccine B and those that were IN vaccinated with vaccine A. Neonatal calves were used for the study, so the immunogenicity of primary vaccination could be evaluated. A secondary goal of the study was to investigate whether activation of innate immune responses of newborn calves following IN vaccination might alter the host immune response to potential respiratory pathogens within the respiratory tract microbiome. Results of microbiome analyses indicate that many bacterial pathogens of the bovine respiratory tract, such as Mannheimia haemolytica and Pasteurella multocida, colonize the upper respiratory tract of Holstein calves within the first week of life.13 Interferon-γ, also known as immune IFN, has potent immunostimulatory activity and can enhance the IgG response in serum14,15 and IgA response in nasal secretions.15 Therefore, secretory IgA antibody titers against M haemolytica and P multocida were monitored throughout the observation period to investigate whether IN vaccination against viral pathogens might alter the host immune response against bacterial pathogens.
Materials and Methods
Animals
All experimental procedures were reviewed and approved by the Johnson Research Animal Care and Use Committee and were conducted in compliance with the Guide for the Care and Use of Agricultural Animals in Research and Teaching.16 One hundred two sexually intact male Holstein calves were obtained from 1 dairy operation within 72 hours after birth. Calves were removed from their dams immediately after birth and were fed a commercial colostrum replacer,a which was prepared in accordance with the manufacturer's directions. Each calf received two 225-g bags of colostrum replacer within 2 hours after birth and again at 6 hours old. All calves were obtained from 1 dairy operation and fed colostrum replacer instead of colostrum to reduce variability in serum IgG titers against BHV1, M haemolytica, and P multocida among study calves.
All calves were individually housed in hutches at the source dairy until they were transported 69 km (43 miles) in sanitized individual hutches to the research facility, where each calf was individually housed in a sanitized hutch (1.22 × 1.22 × 1.22 m [4.0 × 4.0 × 4.0 feet]). Each hutch had an open roof and mesh floorb and was fully enclosed on all sides, with the exception of an opening (7 × 17 cm [2.8 × 6.7 inches]) on 1 side for animal observation. The hutches were on a concrete floor in an open-sided building, with 1.22 m between hutches, and the mesh floor of each hutch was elevated 0.5 m (1.6 feet) above the floor. The sides of the hutches were composed of a vinyl material to allow for easy disinfection. Calves received 1.9 L (2 quarts) of milk replacerc containing 28% protein and 25% fat twice daily and had ad libitum access to a calf starter grain mixd and water.
No pharmaceutical or biologic products were administered to any calf prior to study enrollment. Each calf was confirmed to not be persistently infected with bovine viral diarrhea virus on the basis of immunohistochemical evaluatione of a skin biopsy specimen obtained from an ear. On arrival at the research facility, each calf underwent a physical examination performed by a veterinarian, and a blood sample (approx 6 mL) was collected by jugular venipuncture for determination of serum TP concentration by means of a refractometer.
To be enrolled in the study, a calf had to be ≤ 14 days old on the day of vaccination (day 0), test negative for persistent infection with bovine viral diarrhea virus, have a serum TP concentration between 5.0 and 8.0 g/dL, and have no physical abnormalities or evidence of clinical illness. Twelve of the 102 calves were excluded from study enrollment for the following reasons: age exceeded 14 days on day 0 (n = 1), serum TP concentration was not within the specified range (3), neonatal diarrhea (1), fever (rectal temperature > 39.5°C [103.1°F]; 6 [all 6 calves were part of the same transport load from the source dairy]), or contact with febrile calves (1; this calf was part of the same transport load as the previously described 6 febrile calves). The 90 calves enrolled in the study ranged from 5 to 14 days old on day 0.
Study design
The study was conducted from May to September 2018 and had a randomized complete block design with 6 blocks. Within each block, each calf within a group of 3 age-matched calves was randomly assigned to receive vaccine A,f vaccine B,g or saline (0.9% NaCl) solution (control) by the IN route. Randomization was achieved by use of a randomization function within a computer spreadsheet program.h Briefly, a random number was generated for each calf, and the calves within each group of 3 age-matched calves were assigned to receive vaccine A, vaccine B, or the control treatment on the basis of ranking the assigned random numbers from highest to lowest. Initially, it was intended that each block would contain 15 calves (5 calves/treatment); however, because of variation in the availability of male calves from the source dairy, 1 block was limited to 9 calves (3 calves/treatment) and 2 blocks contained 18 calves (6 calves/treatment). Each treatment was equally represented in each block, and there was a total of 30 calves assigned to each treatment.
The day of vaccination was designated as day 0, and the calves were observed for 42 days after vaccination. At the research facility, calves were housed in individual hutches that were marked with a specific color, which denoted the treatment group assignment for each calf. However, the personnel responsible for assigning calves to the treatment groups, obtaining samples, and performing laboratory procedures remained unaware of (were blinded to) the color assigned to each treatment throughout the observation period. Following completion of the observation period, all calves were returned to the source dairy, except for 1 calf that was euthanized because it could not stand without assistance.
Treatment administration
Two different lots of both vaccine A and vaccine B were used for the study in accordance with the planned design. The same assigned vaccine lot was used to vaccinate all designated calves within a given block. All vaccines were stored and prepared in accordance with the label directions and were administered within 1 hour after vaccine preparation (rehydration). The recommended dose for each vaccine was 2 mL, and calves in the control group were also administered 2 mL of sterile saline solution.
For each calf, gentle manual restraint was used to elevate and maintain the nose at an approximately 45° angle relative to the ground. One milliliter of the assigned treatment was administered in each nostril through a sterile plastic cannula (a separate cannula was used for each calf). The nose was maintained in an elevated position for 30 seconds after treatment administration.
Sample collection
Study personnel wore latex gloves while handling each animal, and gloves were changed between calves. For each calf, body weight was recorded on days 0, 21, and 42 and rectal temperature was recorded on days 0, 3, 5, 7, 10, 14, 21, and 42. A blood sample (20 mL) was collected by jugular venipuncture into two 10-mL blood collection tubes with no additives on days 0, 7, 14, 21, and 42. Blood samples were either stored at 2°C to 7°C for 18 to 24 hours or incubated at 36°C for 1 to 2 hours after collection. The blood samples were then centrifuged at 1,000 × g until separation of serum from cells was achieved. From each sample, 1 to 2 mL of serum was transferred to each of 2 cryovials, and the duplicate serum samples were stored frozen at –20°C until analyzed.
Nasal secretion specimens were collected on study days 0, 3, 5, 7, 10, 14, 21, and 42 by use of the protocol described by Hill et al.8 Briefly, the nostril was moistened with sterile saline solution by use of a spray mister held approximately 12.7 mm from the external nares. A separate mister was used for each treatment group. A cotton tamponi was fully inserted into the ventral meatus of 1 nostril and left in place for 20 minutes. The tampon then was removed from the nostril and placed in a sterile 60-mL syringe. Fluid was expressed from the tampon by compressing the syringe plunger with a caulking gun and collected into a 15-mL sterile polypropylene tube maintained in ice. The NF samples were centrifuged at 1,000 × g for 5 minutes, and a 0.5-mL aliquot of each sample was transferred into a cryovial that contained 50 μL of 100× protease inhibitorj to minimize IFN degradation.17 Samples were stored at –20°C until analyzed. The alternate nostril was used for each successive specimen collection.
Sample analysis
All analyses were performed on single samples of serum and NF. The concentrations of IFN-α and IFN-γ in NF samples were determined by use of ELISAs as described.8 All NF samples were thawed at 4°C and analyzed on the same day. The ELISAs used to determine NF IFN concentrations have a minimum sensitivity of 8 pg/mL for both IFN-α and IFN-γ as determined by evaluation of serial dilutions of recombinant bovine IFN-α and IFN-γ.8 The IFN concentration in each NF sample was calculated by the use of commercial softwarek to extrapolate the ELISA optical density values from standard curves established with serial dilutions of recombinant bovine IFN-α and IFN-γ. The person who conducted the ELISAs and performed the data acquisition and analyses was blinded to the treatment group assignment for all calves.
Indirect ELISAs were used to quantitate serum IgG and NF IgA titers against BHV1 tgD as described.8,18 Briefly, 0.05 μg of purified BHV1 tgDl diluted with 0.05M carbonate-bicarbonate buffer solution (pH, 9.6) in a volume sufficient to yield 100 μL was added to each well of a 96-well polystyrene plate.m The plate was incubated at 4°C for 16 hours and then washed 3 times with TBS solution. Six serial 5-fold dilutions of each NF (starting dilution, 1:50) and serum (starting dilution, 1:100) sample were prepared with TBS solution, and 100 μL of each sample dilution was added to a well of the prepared 96-well plate. The plate was incubated at 20°C for 2 hours and then washed 3 times with TBS solution. For detection of anti-BHV1 tgD IgG, each sample was incubated with alkaline phosphatase–conjugated goat anti-bovine IgGn (dilution, 1:5,000) at 20°C for 60 minutes. For detection of anti-BHV1 tgD IgA, each sample was incubated with biotin-conjugated rabbit anti-bovine IgAo (dilution, 1:1,000) at 20°C for 60 minutes, washed 3 times with TBS solution, and then incubated with streptavidin–alkaline phosphatasep (dilution, 1:10,000) for another 60 minutes. For detection of both anti-BHV tgD IgG and IgA, the plate was washed 3 times with TBS solution and incubated with 200 μL of p-nitrophenyl phosphatej at 20°C for 60 minutes. The reaction was stopped by the addition of 3M EDTA to each well, and the optical density of the reaction product was measured at 405 nm with a microplate reader.q The antibody titer for each sample was reported as the reciprocal of the highest serum or NF dilution that generated an absorbance value that exceeded the mean ± 2 SD absorbance value calculated for the 3 negative serum or NF samples that were run on the same plate (ie, negative controls).
Indirect ELISAs to quantitate anti–M haemolytica and anti–P multocida IgA titers in NF samples were performed as described for the BHV1 tgD–specific IgA ELISA, except different coating antigens were used to capture IgA antibodies. Mannheimia haemolytica B122 and P multocida 1 isolatesl were used to produce bacterial lysates. Bacteria were grown in blood heart infusion brothj by incubation of 1-L cultures at 37°C for 16 hours on a shaker platform until the optical density measured at 600 nm reached 1.6. Cultures were stored at 4°C for 8 to 16 hours, after which the bacterial cells were pelleted by centrifugation at 5,000 × g and 4°C for 15 minutes. The supernatant was discarded. The bacterial cells were resuspended in 500 mL of sterile 5× TBS solution in a flask and stirred with a stir bar at 56°C for 20 minutes. The flask was cooled to 25°C, and cells were pelleted by centrifugation at 5,000 × g and 4°C for 15 minutes. The supernatant was collected, and the soluble proteins in the supernatant were pelleted by centrifugation at 100,000 × g and 4°C for 15 minutes. The extracted proteins were resuspended in sterile TBS solution. The protein concentration in the resulting solution was determinedr and adjusted to achieve a final protein concentration of 2 mg/mL. Coomasie blue–stained SDS-PAGE was used to confirm the presence of multiple bacterial proteins in both preparations. Aliquots of M haemolytica and P multocida protein solutions were stored frozen at –20°C until needed and were diluted with 0.05M carbonate-bicarbonate buffer solution (pH, 9.6) to a concentration of 0.1 μg/mL when used to coat the wells of a polystyrene plate.
Statistical analysis
Data were coded to ensure that the data analyst remained blinded to the treatment group assignment for all calves. All analyses were performed by the use of mixed models,s which included random intercept terms to account for the blocked study design and covariance parameters to account for repeated measures over time (a first-order autoregressive covariance matrix was used unless otherwise noted). The models included fixed effects for treatment, sample acquisition time (time), and the interaction between treatment and time. Final models were fitted with maximum likelihood estimation, Kenward-Roger degrees of freedom, and Newton-Raphson and Ridging optimization procedures. Model-adjusted means and SEMs were reported. The Tukey-Kramer method was used to calculate adjusted P values when multiple post hoc pairwise comparisons were necessary. Appropriate nonnormal distribution and link functions were used for the primary response variables. The IFN concentration data and NF IgA titers were modeled with a log link and Poisson distribution, and an unstructured covariance matrix was used for repeated measures. The IFN concentration data were also dichotomized (ie, did a calf have a measurable IFN response at a given time? [yes or no]) for analysis and were modeled with a logit link and binomial distribution. Values of P ≤ 0.05 were considered significant for all analyses.
Results
Calves
Four of the 90 calves had adverse health events recorded during the observation period. One calf that received vaccine B had diarrhea on days 5 through 7. It was treated orally with electrolytest and had an uneventful recovery. Another calf that received vaccine B appeared clinically normal until day 16, at which time it had difficulty standing. The calf could stand with assistance, maintained a normal attitude and appetite throughout the remainder of the observation period (ie, through day 42), did not receive any treatments, and was euthanized at the conclusion of the study. One calf in the control group was treated for undifferentiated BRD with ceftiofur crystalline free acidu in accordance with the research facility's protocol on day 22 and had an uneventful recovery. Another calf in the control group appeared clinically normal until it was found dead on day 27; necropsy revealed an intestinal volvulus. Overall, 89 of the 90 calves completed the 42-day observation period.
Clinical variables
The model-adjusted mean age (P = 0.85), serum TP concentration (P = 0.88), body weight (P = 0.28), and rectal temperature (P = 0.35) did not differ significantly among the 3 treatment groups on day 0 (Table 1). Body weight was significantly associated with time (P < 0.01) but was not associated with treatment (P = 0.21) or the treatment-time interaction (P = 0.19). For all study calves, the mean body weight increased from 47.2 kg (103.8 lb) on day 0 to 56.5 kg (124.3 lb) on day 21 and to 72.3 kg (159.1 lb) on day 42. Similarly, rectal temperature was significantly associated with time (P < 0.01) but was not associated with treatment (P = 0.32) or the treatment-time interaction (P = 0.79). The mean rectal temperature for all calves ranged from a high of 38.9°C (102.0°F) on day 3 to a low of 38.6°C (101.5°F) on day 42. A few calves had a rectal temperature outside the reference range (38.0° to 39.3°C [100.4° to 102.8°F]) during the observation period, but the abnormal rectal temperatures were always transient and resolved without treatment.
Model-adjusted mean (SEM) age, body weight, serum TP concentration, and rectal temperature for 90 neonatal sexually intact male Holstein calves immediately before (day 0) IN administration of 2 mL (1 mL/nostril) of vaccine A (n = 30), vaccine B (30), or saline (0.9% NaCl) solution (control; 30).
| Variable | Vaccine A | Vaccine B | Control | P value |
|---|---|---|---|---|
| Age (d) | 9.43 (0.45) | 9.37 (0.45) | 9.37 (0.45) | 0.85 |
| Body weight (kg) | 48.20 (0.87) | 46.50 (0.87) | 46.78 (0.87) | 0.28 |
| Serum TP (g/dL) | 5.48 (0.06) | 5.47 (0.06) | 5.51 (0.06) | 0.88 |
| Rectal temperature (°C) | 38.77 (0.06) | 38.71 (0.06) | 38.81 (0.06) | 0.35 |
IFN responses
The NF IFN-α concentration was significantly associated with treatment (P = 0.01), time (P = 0.03), and the treatment-time interaction (P = 0.01). The model-adjusted mean NF IFN-α concentration for the vaccine A–treated calves was significantly greater than that for the vaccine B–treated calves (P = 0.039) and control calves (P < 0.001) on day 5 (Figure 1). Additionally, on day 5, the proportion of calves with a detectable NF IFN-α concentration for the vaccine A group (20/30 [66%]) was significantly greater than that for the vaccine B group (10/30 [33%]; P = 0.034) and control group (2/30 [6.7%]; P < 0.001). On day 7, the model-adjusted mean NF IFN-α concentration for the vaccine B–treated calves was significantly greater than that for the control calves (P = 0.048) but did not differ significantly from that for the vaccine A–treated calves (P = 0.16). The proportion of calves with a detectable NF IFN-α concentration for the vaccine B group did not differ significantly from that for the control group on any study day.
Model-adjusted mean ± SEM values for NF IFN-α concentration (A) and proportion of calves with detectable NF IFN-α (B) at various times after IN administration of 2 mL (1 mL/nostril) of vaccine A (black bars; n = 30), vaccine B (gray bars; 30), or saline (0.9% NaCl) solution (control; white bars; 30) to neonatal (≤ 14 days old) sexually intact male Holstein calves. The assigned treatment was administered on day 0. Within a given day, model-adjusted mean values with different lowercase letters differ significantly (P ≤ 0.05).
Citation: Journal of the American Veterinary Medical Association 258, 10; 10.2460/javma.258.10.1119
Model-adjusted mean ± SEM values for NF IFN-α concentration (A) and proportion of calves with detectable NF IFN-α (B) at various times after IN administration of 2 mL (1 mL/nostril) of vaccine A (black bars; n = 30), vaccine B (gray bars; 30), or saline (0.9% NaCl) solution (control; white bars; 30) to neonatal (≤ 14 days old) sexually intact male Holstein calves. The assigned treatment was administered on day 0. Within a given day, model-adjusted mean values with different lowercase letters differ significantly (P ≤ 0.05).
Citation: Journal of the American Veterinary Medical Association 258, 10; 10.2460/javma.258.10.1119
Model-adjusted mean ± SEM values for NF IFN-α concentration (A) and proportion of calves with detectable NF IFN-α (B) at various times after IN administration of 2 mL (1 mL/nostril) of vaccine A (black bars; n = 30), vaccine B (gray bars; 30), or saline (0.9% NaCl) solution (control; white bars; 30) to neonatal (≤ 14 days old) sexually intact male Holstein calves. The assigned treatment was administered on day 0. Within a given day, model-adjusted mean values with different lowercase letters differ significantly (P ≤ 0.05).
Citation: Journal of the American Veterinary Medical Association 258, 10; 10.2460/javma.258.10.1119
Similar to NF IFN-α concentration, the NF IFN-γ concentration was significantly associated with treatment (P < 0.01), time (P < 0.01), and the treatment-time interaction (P < 0.01). The model-adjusted mean NF IFN-γ concentration for the vaccine A–treated calves was significantly greater than that for the control calves on days 3 (P = 0.021) and 5 (P < 0.001). The mean NF IFN-γ concentration for the vaccine B–treated calves was significantly greater than that for the control calves on days 5 (P = 0.001) and 7 (P = 0.043). The mean NF IFN-γ concentration for the vaccine A–treated calves did not differ significantly from that for the vaccine B–treated calves at any time, although it was approximately 1.5 times the NF IFN-γ concentration for the vaccine B–treated calves on day 5 (Figure 2). The mean NF IFN-γ concentration for vaccine B–treated calves on day 5 was skewed by 2 calves that had concentrations > 1,722 pg/mL, whereas the maximum NF IFN-γ concentration for vaccine A–treated calves was 1,446 pg/mL. The proportion of calves with a detectable NF IFN-γ concentration in the vaccine A group was significantly greater than that for the control group on days 3 (12/30 [40%] vs 3/30 [10%]; P = 0.016) and 5 (26/30 [87%] vs 3/30 [10%]; P < 0.001). The proportion of calves with a detectable NF IFN-γ concentration in the vaccine B group was significantly greater than that for the control group on days 5 (19/30 [63%] vs 3/30 [10%]; P < 0.001) and 7 (11/30 [37%] vs 3/30 [10%]; P = 0.033). The proportion of calves with a detectable NF IFN-γ concentration in the vaccine A group did not differ significantly from that for the vaccine B group at any time, although on day 5, the proportion of vaccine A–treated calves with a detectable NF IFN-γ concentration (26/30 [87%]) was approximately 1.4 times the proportion of vaccine B–treated calves with a detectable NF IFN-γ concentration (19/30 [63%]).
Model-adjusted mean ± SEM values for NF IFN-γ concentration (A) and proportion of calves with detectable NF IFN-γ (B) at various times after administration of the assigned treatments to the calves of Figure 1. See Figure 1 for remainder of key.
Citation: Journal of the American Veterinary Medical Association 258, 10; 10.2460/javma.258.10.1119
Model-adjusted mean ± SEM values for NF IFN-γ concentration (A) and proportion of calves with detectable NF IFN-γ (B) at various times after administration of the assigned treatments to the calves of Figure 1. See Figure 1 for remainder of key.
Citation: Journal of the American Veterinary Medical Association 258, 10; 10.2460/javma.258.10.1119
Model-adjusted mean ± SEM values for NF IFN-γ concentration (A) and proportion of calves with detectable NF IFN-γ (B) at various times after administration of the assigned treatments to the calves of Figure 1. See Figure 1 for remainder of key.
Citation: Journal of the American Veterinary Medical Association 258, 10; 10.2460/javma.258.10.1119
BHV1-specific antibody responses
Serum IgG titer against BHV1 (anti-BHV1 tgD IgG) was not significantly associated with treatment (P = 0.54) or the treatment-time interaction (P = 0.16) but was associated with time (P < 0.01). Therefore, the serum anti-BHV1 IgG titer data for all 3 treatment groups were combined and summarized (Figure 3). The model-adjusted mean IgG titer against BHV1 was greatest on day 0; this confirmed that the assigned treatments were administered to calves that were seropositive for anti-BHV1 IgG, which was most likely absorbed from the colostrum replacer administered on the day of birth (ie, in the presence of colostral antibodies). The mean serum IgG titer against BHV1 declined during the 42-day observation period. The fact that the mean serum anti-BHV1 IgG titer did not differ significantly among the 3 treatment groups suggested that IN vaccination did not alter the decay of colostral IgG titers.
Model-adjusted mean ± SEM serum anti-BHV1 tgD IgG titer at various times after administration of the assigned treatments to the calves of Figure 1. Neither treatment nor the treatment-time interaction had a significant effect on the anti-BHV1 tgD IgG titer; therefore, data for all 3 treatment groups were combined for presentation in this figure. One control calf died acutely from an intestinal volvulus on day 27; thus, only 89 of the 90 study calves contributed to the mean value on day 42. See Figure 1 for remainder of key.
Citation: Journal of the American Veterinary Medical Association 258, 10; 10.2460/javma.258.10.1119
Model-adjusted mean ± SEM serum anti-BHV1 tgD IgG titer at various times after administration of the assigned treatments to the calves of Figure 1. Neither treatment nor the treatment-time interaction had a significant effect on the anti-BHV1 tgD IgG titer; therefore, data for all 3 treatment groups were combined for presentation in this figure. One control calf died acutely from an intestinal volvulus on day 27; thus, only 89 of the 90 study calves contributed to the mean value on day 42. See Figure 1 for remainder of key.
Citation: Journal of the American Veterinary Medical Association 258, 10; 10.2460/javma.258.10.1119
Model-adjusted mean ± SEM serum anti-BHV1 tgD IgG titer at various times after administration of the assigned treatments to the calves of Figure 1. Neither treatment nor the treatment-time interaction had a significant effect on the anti-BHV1 tgD IgG titer; therefore, data for all 3 treatment groups were combined for presentation in this figure. One control calf died acutely from an intestinal volvulus on day 27; thus, only 89 of the 90 study calves contributed to the mean value on day 42. See Figure 1 for remainder of key.
Citation: Journal of the American Veterinary Medical Association 258, 10; 10.2460/javma.258.10.1119
The NF IgA titer against BHV1 (anti-BHV1 tgD IgA) was significantly associated with treatment (P < 0.01), time (P < 0.01), and the treatment-time interaction (P = 0.02). The NF IgA titer against BHV1 was nearly undetectable for all calves on day 0 (Figure 4). The model-adjusted mean NF anti-BHV1 IgA titer for vaccine A–treated calves was significantly greater than that for the control calves on days 14, 21, and 42. The mean NF anti-BHV1 IgA titer for vaccine B–treated calves was significantly greater than that for the control calves on days 10, 14, 21, and 42. On day 10, the mean NF anti-BHV1 IgA titer for the vaccine B–treated calves was significantly (P = 0.017) greater than that for the vaccine A–treated calves; it did not differ significantly between those 2 groups at any other time. However, on day 42, the mean NF anti-BHV1 IgA titer for the vaccine A–treated calves was approximately 1.5 times that for the vaccine B–treated calves.
Model-adjusted mean ± SEM values for NF IgA titers against BHV1 tgD (A), Pasteurella multocida (B), and Mannheimia haemolytica (C) at various times after administration of the assigned treatment to the calves of Figure 1. See Figure 1 for remainder of key.
Citation: Journal of the American Veterinary Medical Association 258, 10; 10.2460/javma.258.10.1119
Model-adjusted mean ± SEM values for NF IgA titers against BHV1 tgD (A), Pasteurella multocida (B), and Mannheimia haemolytica (C) at various times after administration of the assigned treatment to the calves of Figure 1. See Figure 1 for remainder of key.
Citation: Journal of the American Veterinary Medical Association 258, 10; 10.2460/javma.258.10.1119
Model-adjusted mean ± SEM values for NF IgA titers against BHV1 tgD (A), Pasteurella multocida (B), and Mannheimia haemolytica (C) at various times after administration of the assigned treatment to the calves of Figure 1. See Figure 1 for remainder of key.
Citation: Journal of the American Veterinary Medical Association 258, 10; 10.2460/javma.258.10.1119
IgA responses to P multocida and M haemolytica
The NF IgA titer against P multocida was significantly associated with treatment (P = 0.04), time (P < 0.01), and the treatment-time interaction (P < 0.01). For all 3 treatment groups, the model-adjusted mean NF anti–P multocida IgA titer was very low prior to vaccination on day 0 and increased throughout the 42-day observation period (Figure 4). The mean NF anti–P multocida IgA titer for the vaccine A–treated calves was greater than that for the control calves on days 21 (P = 0.005) and 42 (P < 0.001). The mean NF anti–P multocida IgA titer for the vaccine B–treated calves was significantly greater than that for the control calves on days 7 (P = 0.037) and 42 (P = 0.016). The mean NF anti–P multocida IgA titer was significantly (P = 0.007) greater for the vaccine B–treated calves than for the vaccine A–treated calves on day 7, but did not differ significantly between those 2 groups at any other time. However, the mean NF anti–P multocida IgA titer for the vaccine A–treated calves was approximately 1.4 times that for the vaccine B–treated calves on days 21 and 42.
The NF IgA titer against M haemolytica was significantly associated with time (P < 0.01) but was not associated with treatment (P = 0.42) or the treatment-time interaction (P = 0.71). For all 3 treatment groups, the mean NF anti–M haemolytica IgA titer was almost undetectable prior to vaccination on day 0 and gradually increased during the 42-day observation period (Figure 4). The mean NF anti–M haemolytica IgA titer did not differ significantly among the 3 groups at any time.
Discussion
Results of the present study confirmed that IN administration of MLV vaccines to neonatal calves seropositive for colostral antibodies induced both innate and acquired mucosal immune responses. Local IFN responses induced by IN administration of MLV vaccines containing a BHV1 component have been investigated in juvenile and adult cattle.3,4,5,6,7,9 To our knowledge, prior to the present study, only 1 other study8 has investigated IFN responses following IN administration of an MLV vaccine to newborn calves. Interferon-α and IFN-γ responses were not detected in the calves of that study.8 In the present study, IFN-α was undetectable in NF samples from 75 of 90 (83%) calves and IFN-γ was undetectable in NF samples from 79 of 90 (88%) calves prior to vaccination (day 0; baseline). Both vaccines A and B induced significant and transient increases in NF IFN-α and IFN-γ concentrations within 5 days after vaccination. The kinetics of the vaccine-induced IFN responses observed for the neonatal calves of this study were similar to those described for adult cattle,3,4 with increases in NF IFN-α and IFN-γ concentrations becoming detectable by 3 days after vaccination, peaking at approximately 5 days after vaccination, and returning to baseline concentrations by approximately 10 days after vaccination.
For the calves of the present study, the innate immune response induced by vaccine A differed from that induced by vaccine B. On day 5, the calves that received vaccine A had a significantly greater NF IFN-α concentration than did calves that received vaccine B, and the proportion of vaccine A–treated calves with a detectable NF IFN-α concentration was significantly greater than the proportion of vaccine B–treated calves with a detectable NF IFN-α concentration. Also on day 5, the mean NF IFN-γ concentration and proportion of calves with a detectable NF IFN-γ concentration for the vaccine A group were numerically, although not significantly, greater than those for the vaccine B group by approximately 1.5 and 1.4 times, respectively. The differences in the innate immune responses induced by vaccines A and B were likely associated with the biologic differences between the viral strains contained in the 2 vaccines, extent of vaccine virus replication within the calves, and specific location where the vaccine virus replicated in the upper respiratory tract. Vaccine B contained a BRSV component, whereas vaccine A did not. It is possible that the BRSV component of vaccine B might have affected the outcome variables observed. The primary objective of the present study was to compare the innate and acquired immune responses induced in the upper respiratory tract of neonatal calves by 2 commercially available vaccines approved for IN administration to cattle. Investigation of the specific effect of the BRSV strain contained in vaccine B on the upper respiratory mucosal immune response was beyond the scope of the present study.
In another study,9 IN administration of vaccine B to prepartum and postpartum dairy cows failed to induce significant increases of IFN-α, IFN-β, and IFN-γ concentrations in nasal secretions. Failure of the vaccine to induce a detectable innate immune response in the cows of that study9 might reflect the biologic attributes of vaccine B, which contains a temperature-sensitive strain of BHV1, or be a function of the physiologic status of the cows at the time of vaccination. Interestingly, the baseline (ie, prevaccination) NF IFN-α and IFN-γ concentrations for both the prepartum and postpartum cows of that study9 were remarkably high. In fact, the baseline mean NF IFN-α concentrations for the cows of that study9 were comparable to the peak mean NF IFN-α concentrations observed for the calves of the present study. Furthermore, the baseline mean NF IFN-γ concentrations for the cows of the other study9 were approximately 50- (prepartum cows) to 100-fold (postpartum cows) the peak mean NF IFN-γ concentrations observed for the vaccine B–treated calves of the present study. Some of the differences in IFN concentrations detected in the nasal secretions of the neonatal calves of the present study and the adult cows of that previous study9 might have been caused by differences in the sensitivity or specificity of the ELISAs used to quantitate IFN concentrations in each study. However, when the same ELISAs were used to quantify NF IFN concentrations, the magnitude of the IFN response induced by vaccine A in the calves of the present study was approximately a fortieth of the observed IFN response in 6-month-old BHV1-naïve calves after they were experimentally infected with a virulent BHV1 isolate.11 The marked difference in IFN production between the neonatal calves of the present study and the 6-month-old calves of that other study11 might be a reflection of the attenuated virulence and reduced replication of the BHV1 strain in vaccine A, compared with a virulent BHV1 field isolate, or a consequence of the immature immune system of the neonatal calves of the present study. Innate lymphoid cells, including natural killer cells, are an important source of IFN-γ and function as important effector cells during a primary BHV1 infection.19,20 However, the reduced IFN-γ response observed in calves of the present study following vaccination with an MLV vaccine containing BHV1 in relation to the IFN-γ response observed in older cattle following infection with a virulent field strain of BHV1 is inconsistent with results of another study,21 which indicate that natural killer cells are fully functional in newborn calves.
In some situations, the goal of IN vaccination is to induce an innate (IFN) response that does not cause harm to the animal but is physiologically adequate to protect against disease. Results of a dose titration study22 involving recombinant bovine IFN-γ indicate that high concentrations of IFN-γ can induce fever, diarrhea, and leukopenia. In the present study, the mean rectal temperature and body weight did not differ significantly among the 3 treatment groups at any time during the 42-day observation period, which suggested that the vaccine-induced IFN responses did not have adverse physiologic effects on the calves. Although the minimum IFN concentration necessary for an antiviral effect has not been established in cattle, the NF IFN-α and IFN-γ concentrations detected in the calves of the present study were similar to those that inhibited replication of an IFN-sensitive virus, such as vesicular stomatitis virus, by > 80% in a previous BHV1 challenge study.11
In the present study, the proportion of vaccine A–treated calves with detectable NF IFN concentrations in relation to the proportion of vaccine B–treated calves with detectable NF IFN concentrations was significantly greater on day 5 (20/30 [67%] vs 10/30 [33%]; P = 0.034) for IFN-α and numerically greater on day 5 (26/30 [87%] vs 19/30 [63%]; P = 0.058) for IFN-γ. This suggested that vaccine A may have a greater potential than vaccine B for inducing herd protection against viral pathogens. If a therapeutic goal for IN administration of an MLV vaccine with a BHV1 component is to stimulate IFN production and provide the vaccinated animals with protection against a broad range of viruses,12,23,24 then a greater number of animals responding to the vaccine should correspond to greater antiviral protection for the herd as a whole.25
Results of another study8 indicate that IN administration of a multivalent MLV vaccine to colostrum-fed neonatal (age, 3 to 8 days) calves induces production of BHV1-specific secretory IgA. However, results of other studies26,27 suggest that the presence of maternal antibodies in calves can interfere with the response to IN administration of MLV vaccines against other respiratory viruses (eg, BRSV). In the present study, both MLV vaccines induced an increase in anti-BHV1 IgA in nasal secretions, the titers of which progressively increased throughout the 42-day observation period. In a study8 in which colostrum-fed neonatal calves received an IN multivalent MLV vaccine different from those administered in the present study, the anti-BHV1 IgA titers in nasal secretions peaked 14 to 21 days after vaccination and then began to decline. In cattle (age not specified) that received an IN MLV vaccine that contained a temperature-sensitive strain of BHV1, the anti-BHV1 IgA concentration in nasal secretions likewise peaked 14 days after vaccination and progressively decreased at 21 and 28 days after vaccination.5 The magnitude and duration of the BHV1-specific IgA responses induced in the upper respiratory tract of cattle following IN administration of a multivalent MLV vaccine are likely dependent on the BHV1 isolate and other viral components contained in the vaccine and age of the animals at the time of vaccination. The mean NF anti-BHV1 IgA titers for the control calves of the present study gradually increased during the observation period despite the fact that only one of those calves developed signs of BRD and all calves were individually housed in hutches and did not have direct nose-to-nose contact with each other. Thus, natural exposure of calves to BHV1 early in life may also affect the magnitude and duration of the BHV1-specific IgA response to IN vaccine administration.5,8
In the present study, the serum anti-BHV1 IgG titers did not differ significantly among the 3 treatment groups at any time during the observation period and were highest on day 0. That finding confirmed that the vaccines were administered to calves seropositive for antibodies against BHV1. The findings of the present study were also consistent with results of another study8 in that IN administration of a multivalent MLV vaccine to neonatal calves that had received colostrum or colostrum replacer did not alter the decay of colostral IgG in the circulation. In calves, circulating colostral IgG may restrict IN vaccine BHV1 replication to mucosal surfaces. Intranasal administration of MLV vaccines with a BHV1 component to older cattle that are seronegative for anti-BHV1 antibodies results in seroconversion.3,4,5,7 Thus, IN administration of MLV vaccines can induce production of IgG in seronegative animals. In a BHV1 challenge study,28 IN administration of 1 dose of a multivalent MLV BRD vaccine to calves (age, 3 to 6 months) with circulating maternal antibodies induced a BHV1-specific immune response that persisted for 4 to 5 months and was sufficient to prevent secondary bacterial infection following experimental challenge with BHV1. Therefore, the observation period of the present study may not have been long enough to adequately evaluate whether either vaccine was able to induce a systemic IgG response when administered to calves with circulating material antibodies.
Commensal bacteria may influence the local immune response to IN vaccination.29 Results of the present study appeared to provide evidence of the converse. Specifically, IN administration of multivalent BRD virus vaccines to neonatal calves appeared to alter the immune response to at least 1 commensal bacterium that is a potential respiratory tract pathogen (P multocida). Interferon-γ is a potent immune modulatory molecule that regulates both innate and acquired immune responses.30 In cattle, IFN-γ functions as a potent adjuvant and enhances antibody responses.14 Therefore, we hypothesized that an increase in IFN-γ production by neonatal calves might be associated with changes in IgA responses to commensal bacteria, which colonize the upper respiratory tract during the first week of life. Mannheimia haemolytica and P multocida were of particular interest because they are abundant in the pioneer commensal community and are also opportunistic respiratory tract pathogens.13 For all 3 treatment groups of the present study, both the anti–M haemolytica and anti–P multocida IgA titers in nasal secretions increased significantly over time. However, the mean NF anti–P multocida IgA titer for the vaccine A–treated calves was significantly greater than that for the control calves at 21 days after vaccination, and the mean NF anti–P multocida IgA titers for both vaccine groups were significantly greater than that for the control calves at 42 days after vaccination. Further research is necessary to determine whether the IgA responses of cattle induced by IN administration of multivalent MLV vaccines are directly attributed to an increase in IFN-γ production. It is accepted that innate immune responses induced by IN administration of MLV vaccines with a BHV1 component can protect against other viruses that infect the respiratory tract of cattle.3,24 Results of the present study suggested that IN administration of multivalent MLV vaccines may also enhance acquired immune responses against bacterial pathogens of the respiratory tract. Further research is required to determine whether enhanced IgA responses against commensal bacteria will reduce the bacterial burden in the upper respiratory tract or decrease opportunistic bacterial infections in the lungs.
Findings of the present study provided evidence that IN administration of a multivalent MLV vaccine with a BHV1 component to neonatal calves with circulating colostral antibodies stimulates both innate (IFN-α and IFN-γ) and acquired (IgA) immune responses. Two commercially available IN vaccines were evaluated in this study, and significant differences were observed in both the magnitude of the IFN responses and the proportion of calves with detectable IFN responses between calves vaccinated with vaccine A and those vaccinated with vaccine B as well as between vaccinated calves and control calves.
Acknowledgments
Supported by Merck Animal Health, Madison, NJ.
The study was performed at Johnson Research LLC, Parma, Idaho. Johnson Research LLC owned the calves throughout the duration of the study.
Footnotes
Calf's Choice Total Gold, Saskatoon Colostrum Co, Saskatoon, SK, Canada.
Tenderfoot calf mesh, Tandem Products Inc, Minneapolis, Minn.
SWS 28-25 Dairy Herd and Beef Calf Milk Replacer, Simplot Western Stockmen's, Caldwell, Idaho.
Calf 20% Super Starter BOV CFLY, Simplot Western Stockmen's, Caldwell, Idaho.
Veterinary Diagnostic Center, University of Nebraska, Lincoln, Neb.
Nasalgen IP (lot Nos. 02241466 [expiration date, May 9, 2019] and 02241467 [expiration date, May 30, 2019]), Merck Animal Health, DeSoto, Kan.
Inforce 3 (lot Nos. 256970 [expiration date, Jan 8, 2019] and 277953 [expiration date, Apr 16, 2019]), Zoetis, Parsippany, NJ.
Excel, Microsoft Corp, Redmond, Wash.
o.b. tampon (regular size), Edgewell Personal Care Co, Shelton, Conn.
Sigma-Aldrich Corp, St Louis, Mo.
Microplate Manager Software, version 5.0.1, Bio-Rad Laboratories Inc, Hercules, Calif.
VIDO-Intervac, University of Saskatchewan, Saskatoon, SK, Canada.
Immulon 2, Dynex Technology Inc, Chantilly, Va.
Kirkegaard & Perry Laboratories, Gaithersburg, Md.
Bethyl Laboratories Inc, Montgomery, Tex.
Jackson ImmunoResearch Laboratories Inc, West Grove, Pa.
Benchmark, Bio-Rad Laboratories Inc, Hercules, Calif.
PierceTM BCA protein assay kit, ThermoFisher Scientific, Waltham, Mass.
Proc Glimmix, version 9.4, SAS Institute Inc, Cary, NC.
Hydralyte, AgriLabs LLC, St Joseph, Mo.
Excede, Zoetis Inc, Kalamazoo, Mich.
Abbreviations
| BHV1 | Bovine herpesvirus type 1 |
| BRD | Bovine respiratory disease |
| BRSV | Bovine respiratory syncytial virus |
| IFN | Interferon |
| IN | Intranasal |
| MLV | Modified-live virus |
| NF | Nasal fluid |
| TBS | Tris-buffered saline |
| tgD | Truncated form of recombinant BHV1 glycoprotein D |
| TP | Total protein |
References
- 2. ↑
Osman R, Malmuthuge N, Gonzalez-Cano P, et al. Development and function of the mucosal immune system in the upper respiratory tract of neonatal calves. Annu Rev Anim Biosci 2018;6:141–155.
- 3. ↑
Todd JD, Volenec FM, Paton IM. Intranasal vaccination against infectious bovine rhinotracheitis: studies on early onset of protection and use of the vaccine in pregnant cows. J Am Vet Med Assoc 1971;159:1370–1374.
- 4. ↑
Savan M, Angulo AB, Derbyshire JB. Interferon, antibody responses and protection induced by an intranasal infectious bovine rhinotracheitis vaccine. Can Vet J 1979;20:207–210.
- 5. ↑
Gerber JD, Marron AE, Kucera CJ. Local and systemic cellular and antibody immune responses of cattle to infectious bovine rhinotracheitis virus vaccines administered intranasally or intramuscularly. Am J Vet Res 1978;39:753–760.
- 6. ↑
Lucas MH, Roberts DH, Sands JJ, et al. The use of infectious bovine rhinotracheitis vaccine in a commercial veal unit: antibody response and spread of virus. Br Vet J 1982;138:23–28.
- 7. ↑
Mann DD, Buening GM, Thorne JG. Efficacy of aerosol, intranasal, and intramuscular vaccination against selected bovine viral diseases. Cornell Vet 1983;73:375–379.
- 8. ↑
Hill KL, Hunsaker BD, Townsend HG, et al. Mucosal immune response in newborn Holstein calves that had maternally derived antibodies and were vaccinated with an intranasal multivalent modified-live virus vaccine. J Am Vet Med Assoc 2012;240:1231–1240.
- 9. ↑
Cortese VS, Woolums A, Hurley DJ, et al. Comparison of interferon and bovine herpesvirus-1-specific IgA levels in nasal secretions of dairy cattle administered an intranasal modified live viral vaccine prior to calving or on the day of calving. Vet Immunol Immunopathol 2017;187:35–41.
- 10. ↑
Schroder K, Hertzog PJ, Ravasi T, et al. Interferon-γ: an overview of signals, mechanisms and functions. J Leukoc Biol 2004;75:163–189.
- 11. ↑
Osman R, Gonzalez-Cano P, Brownlie R, et al. Induction of interferon and interferon-induced antiviral effector genes following a primary bovine herpesvirus-1 (BoHV-1) respiratory infection. J Gen Virol 2017;98:1831–1842.
- 12. ↑
Babiuk LA, Sordillo LM, Campos M, et al. Application of interferons in the control of infectious diseases of cattle. J Dairy Sci 1991;74:4385–4398.
- 13. ↑
Lima SF, Teixeira AG, Higgins CH, et al. The upper respiratory tract microbiome and its potential role in bovine respiratory disease and otitis media. Sci Rep 2016;6:29050.
- 14. ↑
Gaertner FH, Babiuk LA, Mutwiri G, et al. Amended recombinant cells (ARCs) expressing bovine IFN-gamma: an economical and highly effective adjuvant system. Vaccine 2009;27:1377–1385.
- 15. ↑
Vandenbroeck K, Nauwynck H, Vanderpooten A, et al. Recombinant porcine IFN-γ potentiates the secondary IgG and IgA responses to an inactivated suid herpesvirus-1 vaccine and reduces postchallenge weight loss and fever in pigs. J Interferon Cytokine Res 1998;18:739–744.
- 16. ↑
Federation of Animal Science Societies. Guide for the care and use of agricultural animals in research and teaching. 3rd ed. Champaign, Ill: Federation of Animal Science Societies Inc, 2010.
- 17. ↑
Gaertner FH, Babiuk LA, Van Moorlehem EA, et al. Amended recombinant cells (ARCs): an efficient production and delivery vehicle for bovine IFN-γ. J Control Release 2005;107:189–202.
- 18. ↑
van Drunen Littel-van den Hurk S, Van Donkersgoed J, Kowalski J, et al. A subunit gIV vaccine, produced by transfected mammalian cells in culture, induces mucosal immunity against bovine herpesvirus-1 in cattle. Vaccine 1994;12:1295–1302.
- 19. ↑
Campos M, Griebel P, Bielefeldt-Ohmann H, et al. Cell-mediated cytotoxic responses in lungs following primary bovine herpesvirus-1 infection. Immunology 1992;75:47–52.
- 20. ↑
Osman RA, Griebel PJ. CD335 NKp46+ T cell recruitment to the bovine upper respiratory tract during a primary bovine herpesvirus-1 infection. Front Immunol 2017;8:1393.
- 21. ↑
Elhmouzi-Younes J, Storset AK, Boysen P, et al. Bovine neonate natural killer cells are fully functional and highly responsive to interleukin-15 and to NKp46 receptor stimulation. Vet Res 2009;40:54.
- 22. ↑
Bielefeldt Ohmann H, Babiuk LA. Alteration of some leukocyte functions following in vivo and in vitro exposure to recombinant bovine alpha- and gamma-interferon. J Interferon Res 1986;6:123–136.
- 23. ↑
Todd JD, Volence FJ, Faton IM. Interferon in nasal secretions and sera of calves after intranasal administration of avirulent infectious bovine rhinotracheitis virus: association of interferon in nasal secretions with early resistance to challenge with virulent virus. Infect Immun 1972;5:699–706.
- 24. ↑
Cummins JM, Rosenquist BD. Protection of calves against rhinovirus infection by nasal secretion interferon induced by infectious bovine rhinotracheitis virus. Am J Vet Res 1980;41:161–165.
- 25. ↑
Cortese VS, Seeger JT, Stokka GS, et al. Serologic response to Mannheimia haemolytica in calves concurrently inoculated with inactivated or modified-live preparations of M haemolytica and viral combination vaccines containing modified-live bovine herpesvirus type 1. Am J Vet Res 2011;72:1541–1549.
- 26. ↑
Kimman TG, Westenbrink F, Straver PJ. Priming for local and systemic antibody memory responses to bovine respiratory syncytial virus: effect of amount of virus, virus replication, route of administration and maternal antibodies. Vet Immunol Immunopathol 1989;22:145–160.
- 27. ↑
Ellis JA, Gow SP, Goji N. Response to experimentally induced infection with bovine respiratory syncytial virus following intranasal vaccination of seropositive and seronegative calves. J Am Vet Med Assoc 2010;236:991–999.
- 28. ↑
Hill K, Arsic N, Nordstrom S, et al. Immune memory induced by intranasal vaccination with a modified-live viral vaccine delivered to colostrum fed neonatal calves. Vaccine 2019;37:7455–7462.
- 29. ↑
Lynn DJ, Pulendran B. The potential of the microbiota to influence vaccine responses. J Leukoc Biol 2018;103:225–231.
