Experimental site

The study was conducted in the College of Veterinary Medicine and Agriculture, Addis Ababa University, Bishoftu, Oromia Regional State, Ethiopia. The College is located 45 km southeast of Addis Ababa at an altitude of 1900 m above sea level. The average annual rainfall is 851 mm and the minimum and maximum temperature is 8.9 °C and 26.2 °C, respectively. The average humidity level is 58.6 %.

Preparation of experimental house

An experimental house with an area of 16 m × 5 m was constructed. The ceiling, walls and floor of the house were disinfected using 1 % formalin. Clean, disinfected teff straw was spread over the floor for bedding. Equipment including drinker, feeder and buckets were cleaned, disinfected and introduced to the houses. The house was kept close for 40 days before the chickens were introduced.

Experimental management

A total of 400 fertilized Bovans brown chicken eggs were obtained from Genesis Farms PLC and hatched at National Veterinary Institute mini-hatchery room. Eggs were cleaned, fumigated and incubated. Finally, 300 chicks were hatched and collected at the 21st and 22nd day of incubation. Of these, the experimental study utilized 225 chicks.

The 225 experimental chicks were brooded together until 14 days in a pen with infrared bulbs for heating and teff straw for bedding. On the 14th day the chicks were randomly split into the 15 treatment groups as described under study design below (Fig. 1). The chickens were fed on purchased starter commercial ration for 2 months, grower ration the next 3–5 months, and layer ration from 5 months onwards. Water was given ad libitum. Antibiotic (oxytetracycline), minerals and vitamins mix in a sachet (i.e. Vytlet) was purchased and supplied for 3 days after each bleeding (Fig. 1). Chickens showing signs of disease (suspected infectious coryza and coccidiosis) were given 20 % oxytetracycline and amprolium. Mortality was recorded daily.

Experimental design and sample size

Chicken care and experimental procedures were performed under approval from the Animal Ethical Committee (AEC permit No. 6–2010) of the College of Veterinary Medicine and Agriculture of the Addis Ababa University.

A complete randomised design (CRD) was employed. Each chick was identified using wing tag and randomly assigned to one of the 15 pens. Accordingly, a total of 225 chickens were randomly assigned to 15 pens with 15 chicks per pen. The study involved 15 different treatment groups: mock vaccinated (naïve), NDI₂ vaccine administered in drinking water (conventional delivery as positive control) and NDI₂ vaccine administered with each of the 13 different carrier grains. The 15 different treatment groups were randomly assigned to the 15 pens as shown on Fig. 1. The sample size per group (n = 15 chickens) was calculated by setting type I error at 5 % and type II error at 20 % (80 % power); assumed survival for vaccinated (80 %) and for non-vaccinated chickens (20 %) as described by Chan [16].

Preparation of carrier grains

All carrier grains were purchased, washed, sun dried and stored at room temperature until use. About 2Kg of each grain was cracked to small size to improve the swallowing by the chicks. About 2 kg of each grain was parboiled grains according to Cumming [17]. The grains were added to boiling water at ratio of 1 kg per 3 l and boiled for two minutes, rinsed with cool distilled water, drained and sun dried.

Vaccine and coating of grains with the vaccine

Vials of freeze dried NDI₂ vaccine (300 doses per vial) were purchased from National Veterinary Institute (NVI) located in Bishoftu. Vials were reconstituted in 150 ml of clean, sterile, non-chlorinated distilled water (manufacturer’s instruction). About 48.5 ml of clean, non-chlorinated water was added to the grain first to wet it. Subsequently, for each treatment group (n = 15 chickens) 7.5 ml of vaccine suspension (0.5 ml per chick) was mixed with 150 g of carrier grain (10 g per chick) to deliver one dose (10⁷ EID₅₀) as described by Wambura et al. [18]. The final suspension of vaccine was stored at room temperature for 6 h. Prior to vaccination, feed was withheld for 7 h, after which the chickens were given the grains coated with the vaccine. Commercial feed was not provided until the chickens consumed the coated grains completely.

NDI₂ vaccination interval and sera collection

Vaccination of the chickens was administered on day 14, 35, and 105 as shown on Fig. 1. Sera were collected for haemagglutination inhibition (HI) assay. At every bleeding, sera were collected on day 0, 14, 28, 49, 77, 105 and on day 119 or 126 from the 225 chicken’s wing vein aseptically. Bleeding of maize and barley treatment group transferred to day 126 from 119 due to shortage of syringes and Eppendorf tube for sera storage.

Haemagglutination inhibition (HI) assay

HI assay was conducted within serology laboratory of NVI. Serum prepared from sequential blood collections (Fig. 1) was heat inactivated at 56 °C for 30 min and stored at −20 °C. The level of ND virus antibodies in serum samples were determined using the HI test as described by OIE [5]. The HI test has 98 % specificity and 69–98 % sensitivity [19]. HI titration was made to determine the right HI concentration via 2-fold serial dilution of 25 μl sera in 25 μl PBS followed by 25 μl loading of viral antigen per well. Then, after 30 min 25 μl of 1 % RBC per well was loaded and kept for 45 min to determine the end point of haemagglutination. The antibody level for each serum sample was expressed as a log to the base two and recorded. For convenience, the titer was recorded as just the log index. For example, the titer of log₂ ² was recorded as two. The geometric mean titers (GM) were calculated. In this study we used the published cut off value for the protective HI antibody titer (HI titer ≥ log₂ ³ i.e. GM ≥ 3) for ND vaccination in chickens [5, 13, 20].

Experimental challenge by virulent ND virus

All the vaccinated and non-vaccinated chickens were challenged by lethal dose (0.5 × 10^6.5 ELD₅₀ based on viral titration). Local virulent ND Alamya strain was obtained from NVI and inoculated via intramuscular route into the breast muscle. The Alamya strain has a mean embryonic death time of 51.1 h, an intracerebral pathogenicity index of 1.84 and an intravenous pathogenicity index of 2.51 [15]. The challenge time was at the age of 129 which was 3 weeks after the 3rd vaccination.

Post challenge, the chickens were examined daily for 4 weeks until the age of 160 day for clinical signs and death due to ND.

Statistical analysis

The mean value and standard deviation (SD) of HI antibody titers were determined and classified according to treatment groups. The post vaccination mean HI antibody titers were compared by General Linear model of SPSS version 15. Where the HI results were significant, least square difference (LSD) was used to compare antibody response via pair-wise treatment comparisons. Proportion of chickens with HI antibody titer ≥ log₂ ³ between treatment groups was used as a cut-off value to compare and decide the protective level [5, 13, 20]. Serology (HI) was to predict level of protection. However, real level of protection was evaluated using challenge experiment. Subsequently, the time to death among treatment groups was compared using Kaplan-Meier survival curve with log rank test used to assess equality of survival distribution among the groups. Chickens were censored in Kaplan-Meier survival analysis whose death was not related to events of interest (i.e. death due to any other causes than ND challenge). Significant differences set at 5 % alpha and at 95 % confidence interval. The relationship between chicken mortality (%) and % of chicken with HI titer above cut-off value were compared to the mean HI titer and statistically tested by Pearson correlation.

Serological analysis

The base line antibody titer

The eggs were collected from the same ND vaccinated parents of Bovans brown. The mean maternal antibody (geometric mean ± SD) titer of the 225 study chickens at day old age was 3.3 ± 0.5 and reduced over time during 14 days to 1.5 ± 0.6 at time of vaccination (Fig. 2).

Monitoring the survival rate of chickens following artificial challenge

The mean HI titer and mortality (%) were inversely correlated as shown on Fig. 3 (r = −0.938, p < 0.000). The mean HI titer above log₂ ⁵ corresponds to a 100 % survival rate as shown in chickens vaccinated by conventional (water), cracked maize and parboiled wheat coated vaccine. The HI titer between log₂ ^(3–5) corresponded to survival rates between 70–80 %. The HI titer between log₂ ² – log₂ ³ corresponded to survival rates of 35–60 % as shown in chickens vaccinated by cracked sorghum, untreated millet, untreated sorghum, cracked millet and cracked barley coated vaccine (Table 4, Fig. 3, left y-axis).

Vaccine carrier type	Chicks N†	Mean HI titer	# morbidity (%)	# death total	#mortalitya (%)	Lower 95 % CI	Upper 95 % CI	# Survival (%)
NDI2 in water	15	7.1	0(0.0)	0	0(0.0)	0	0	15(100)
Cracked maize	15	6.8	0 (0)	0	0(0.0)	0	0	15(100)
Parboiled barley	15	5.1	0(0.0)	0	0(0.0)	0	0	15(100)
Parboiled wheat	14	4.5	5(30.0)	2	2(14.3)	0	32.6	12(85.7)
Untreated wheat	15	3.3	6 (40.0)	3	3(20.0)	0	40.2	12(80.0)
Parboiled sorghum	15	4.4	6(26.7)	3	3(20.0)	0	40.2	12 (80.0)
Cracked wheat	14	3.3	8(53.3)	4	4(28.6)	4.9	52.2	10(71.4)
Parboiled millet	14	3.2	7(46.7)	4	4(28.6)	4.9	52.2	10(71.4)
Cracked millet	15	2.5	10(66.7)	6	6(40.0)	15.2	64.8	9(60.0)
Untreated barley	15	2.8	8(53.3)	6	6(40.0)	15.2	64.8	9(60.0)
Cracked barley	15	2.5	7(46.7)	6	6(40.0)	15.2	64.8	9(60.0)
Untreated sorghum	15	2.4	10(66.7)	7	7(46.7)	21.4	71.9	8(53.3)
Untreated millet	12	2.4	8(53.3)	6	6 (50)	21.7	78.3	6 (50.0)
Cracked sorghum	14	1.9	11(73.3)	9	9(64.3)	39.2	89.4	5(35.7)
Naïve	15	0.0	14 (93.3)	12	12(80.0)	59.8	100	3 (20.0)

N^† represented number of chickens present in each group at age of 129 day (onset of the challenge). Figures outside the brackets indicated number of chickens in each group whereas figures in brackets indicated percentages. ^a Chickens in the treatment groups listed up to parboiled millet had significantly (p < 0.0001) higher survival rate compared to the naïve group. The lower and upper 95 % confidence interval (CI) of mortality of chickens per each treatment group was shown in the table

Figure 3 (right y-axis) indicated that when more than 80 % of the chickens in a treatment group had HI titer above cut-off value, the mortality (%) in that treatment group was nearly 0 %. In parboiled wheat, the mortality (%) was nearly 15 % while 78.5 % of the chickens in this treatment group had HI titre above cut-off value for protection. In this study, more than 50 % of the chickens in majority of the treatment groups had HI titre above cut-off value for protection that that subsequently resulted in mortality below 50 %. However, in cracked sorghum group mortality was higher (i.e. 64.3 %) as only 20 % of the chickens had HI titer above the cut-off value.

Survival analysis

The overall differences in mortality (%) were statistically significant (Log-rank = 77.35, D.F. = 14, p < 0.0000) among the treatment groups (Fig. 4). A small number of chickens from untreated (n = 3) and parboiled millet (n = 1), cracked sorghum (n = 1), parboiled (n = 1) and cracked wheat (n = 1) died due to other disease than ND at different days during the study. Infectious cryza and coccidiosis were suspected and treated long ago (about 2 months) before the inoculation of the velogenic ND viral challenge. The dead chickens were censored in Kaplan-Meier survival analysis as their death was happened before the inoculation of the virulent ND virus challenge. Post-challenge with ND virus, mortality occurred only during the first 14 days; no mortality was observed during days 15 to 30 of the follow up period (Table 4; Fig. 4). The death of chickens due to the challenge started at fourth days post challenge and persisted up to 13th days with the median survival time of 7 days among the dead chickens by the challenge.

Maternal HI antibody titer

At day old, chicks included in this study have HI antibody titer above log₂ ³. Such high maternal antibody titer in the baby chicks is deleterious to vaccination [5]. Thus, we waited until it declined to log₂ ^1.5 titer at 14 days to overcome the risk of its interference with the vaccine. In line with this, it has been well established that chicks from immunized parents possess high level of maternal antibody which protects the chicks against virulent virus and interferes with vaccine antigens [23, 24].

Monitoring serological response using HI antibody titer

An HI titer of ≥ log₂ ³ following vaccination has been considered protective against virulent ND virus. HI titers lower than log₂ ³ have been associated with lower levels of protection [13, 20]. The results of the current study showed that administration of 2nd and 3rd booster vaccination significantly and progressively increased HI antibody titer in all treatment groups except the naïve control. Furthermore, different grains induced different level of HI antibody titer. This implies the presence of inherent variation in virus carrying capacity of different grains. This is an opportunity to screen grains of different species and varieties. Interestingly, treating grains (either cracking or parboiling) increased their efficacy as vaccine carrier, evident by induction of higher HI antibody titer than that was induced by untreated form. Similar results have been reported in Nigeria by Olabode [25] as to the efficacy of treated grain particularly maize compared to untreated grain. Grains have been known to contain tannins, anthraquinone, cardiac glycosides and alkaloids. Some of these chemicals have been shown to have antiviral properties [22, 26]. The higher HI titer induced by treated grains than untreated ones could be due to the fact that cracking grains increase the surface area of the grains to adsorb the vaccine virus [10, 13, 14, 18, 22, 25, 26]. Likewise, parboiling might destroy the antiviral factors from the seed of the grains [22, 26]. Hence, both cracking and parboiling grains would induce better HI antibody titer via adsorbing and releasing live virus in the gut. However, inconsistent results were observed for sorghum in which cracked, untreated and parboiled sorghum induced HI titer of ≥ log₂ ³ in 20 %, 46.7 % and 60 % of chickens, respectively, after the 3rd booster vaccination.

This implies that repeated vaccination induces progressively higher HI titer that could correspond to high levels of protection. Moreover, different grain species as well as their preparations in different forms have induced different level of HI antibody titers. Treatment of grains (parboiling or cracking) renders the grains suitable for vaccine carrier for oral vaccination as evident from their corresponding improved antibody response. Higher HI antibody titer corresponded to higher protection level for most grain carriers in the order of parboiled, cracked and untreated grains. This, however, is not the case for sorghum particularly untreated form was found to be better than cracked form in HI titer and post challenge survival. Exceptionally, the parboiled and untreated sorghum induced higher HI titer than the cracked sorghum. Cracking in sorghum might have released anti-viral factors (tannins) and enzymes that probably inactivated the virus. To this end anthraquinone, alkaloids and cardiac glycosides were reported to be abundant in sorghum [26].

Despite low HI antibody titer induced by some grains (sorghum), unexpectedly higher survival % was observed following the challenge. This perhaps highlights the weak correlation between low HI titer and protection for sorghum and the difficulties of using serum antibodies to determine protection to respiratory pathogens [27, 28]. In addition to serum antibody, secretory antibody (IgA) at mucosal surfaces and cell mediated immunity are thought to play a role in resistance to challenge [29]. In general, however, this study showed a negative correlation between mean HI titer and mortality % at the group level (Fig. 3).

Other than the treatment (grain type for vaccine delivery) that of the pen effect has been handled by grouping of the chickens by randomization. However, pens are not replicated due to shortage of space in the experimental house and finance to account for the pen effect on immune response. We believe that the variance among pens is generally less than the variance of chickens within pens due to treatment effect.

Protection assessment following artificial viral challenge

In the current study, intramuscular route of the breast muscle was used to challenge the chickens. Intramuscular route was preferred as it allows birds to receive equal doses of the challenge virus. Water, cracked maize and parboiled barley were found to be carriers of NDI₂ vaccine than the others. They protected 100 % (15/15) of the chickens against virulent challenge with ND strain compared to 20 % protection in the naïve group (3/15). In agreement with this results of full protection (100 %) using the NDI₂ vaccination via drinking water and parboiled barley was previously reported in Ethiopia [15]. Another similar heat stable vaccine (NDV₄₎ coated cracked maize has also been reported to induce high levels of protection in Nigeria [25] which is similar to our NDI₂ results. The 80 % protection achieved using parboiled sorghum currently is, however, in disagreement with previous results report in Ethiopia [15], Nigeria [30] and Tanzania [31].

It is widely accepted that the recommended protection level of ND vaccination is 80 % [32]. However, we have achieved 100 % protection using NDI₂ in water, cracked maize and parboiled barley at on-station condition.

The survival rate for the chickens vaccinated with untreated and cracked millet is 60 %, but 73.3 % for the parboiled millet. Our current finding is comparable to the 70 % survival rate reported elsewhere among chicken vaccinated with NDI₂ using parboiledand broken millet [33].

The chicken survival rate is 71.4 % for cracked wheat, 80 % for untreated wheat and 85.7 % for parboiled wheat group. The untreated barley as carrier for NDI₂ in the current study had 60 % survival rate. This is far lower than the previous work in Ethiopia [15] and what was reported elsewhere with 100 % protection [34]. This variation in serological response between and within grain forms could be explained by (i) differences in the contents of anti-viral factors in the grains, (ii) variability in vaccine carrying capacity among different grains, (iii) difference in agro-ecology and soil characteristics that can have different effect on physico- chemical characteristics of the grains. Oakeley [22] suggested that grains grown in different agro-ecology and on different soil characteristics tend to vary in their vaccine virus carrying capacity due to variation in the grains’ physico- chemical characteristics, especially their surface properties and chlorine content [22].

In general, parboiled grains, followed by cracked ones, induced higher serological response and protection level than intact (untreated) grains. Heating, soaking, washing and cracking grains might be useful in developing a successful vaccine carrier feed. Similar findings have been reported from other countries [17, 35]. Cracked maize and parboiled barley are found to be better vaccine carriers under Ethiopian context. Our promising finding on wheat was consistent with Spradbrow [13]. Thus, cracked maize, parboiled barley or parboiled wheat should be the base for a large scale grain screening and oral based ND vaccination program. They can be used widely as carrier for oral NDI₂ vaccination. However, it should be noted that the protection level of the grain based NDI₂ vaccine varies under laboratory conditions i.e. > 90 % protection [36] and under real village conditions i.e. < 60 % [37] and with vaccine delivered by farmers [36]. This signals the necessity of pilot field trial at village level to evaluate the results of the current on-station study at real village conditions.

Durability of the virus in the grains and at room temperature

In the current study, the vaccine virus coated on grains was highly immunogenic after 6 h of exposure to room temperature; hence it could be used to vaccinate village chicken against ND. We haven’t measured the upper limit of the time when it is still efficacious in order to enable central vaccine production and then distribution to rural villages. A very important condition for successful development and use of any chosen feed as vaccine carrier is the ability to allow firm binding or adherence of the coated vaccine virus without interfering with the survival of the vaccine virus. In this regard different results have been reported from different countries. According to Tu et al. [12], the NDI₂ in grains has substantial infectivity and induction of immunity in chickens under laboratory conditionsafter storage for 17 days and in village conditions after storage for 21 days. Echeonwu et al. [30] reported that the virus coated feed without additive remained stable and immunogenic for 3 weeks (millet); 3.5 weeks (sorghum) and 5 weeks for maize at room temperature. Nassir et al. [15] recovered viable vaccine virus after 14 h at room temperature on parboiled barley.