Attempts to improve on a challenge model for subtype C avian pneumovirus

Abstract

Respiratory disease caused by avian pneumovirus (APV) has a strong negative impact on the economy of the turkey industry in many countries. Progress in developing vaccines against this infection in the US has been slow partly because of the lack of a consistent challenge model to conduct vaccine efficacy studies. This study was designed to determine whether in vivo passages of a US isolate of APV, designated subtype C (APV-C), would increase virus virulence, leading to consistent clinical signs in turkeys. Three different experiments were performed. In experiments 1 and 2, a cell culture adapted APV was passaged four times in vivo in turkeys. Following each passage, clinical signs were found to increase in severity. In addition, inoculated birds were found to shed both APV RNA (by reverse transcriptase-polymerase chain reaction) and live virus (by virus isolation) at each passage. The mean antibody titres also increased with each passage. The results of the second experiment were not in complete agreement with those of experiment 1. In the third experiment, APV grown in three different cell lines was inoculated into three groups of turkeys. Clinical signs were observed in inoculated birds and virus could be isolated from all three groups. The results of this preliminary study indicate that in vivo passage of APV-C in birds may increase virus virulence, but the results obtained in experiment 2 suggest that further studies are needed to confirm this.

Essais d'amélioration d'un modèle d'épreuve pour le pneumovirus aviaire de sous type C

La maladie respiratoire causée par le pneumovirus aviaire (APV) a un impact fortement négatif sur l'économie de l'industrie de la dinde dans de nombreux pays. Le progrès dans le développement des vaccins contre cette infection aux USA a été lent, en partie à cause de l'absence de modèle d'épreuve performante pour réaliser les études d'efficacité de vaccin. Cette étude a été programmée pour déterminer si les passages in vivo d'une souche américaine d'APV, désignée sous type C (APV-C), pouvait augmenter la virulence du virus, en entraînant des symptômes évidents chez les dindes. Trois expérimentations différentes ont été réalisées. Pour les expérimentations 1 et 2, un APV adapté à la culture cellulaire a été passé quatre fois in vivo sur des dindes. Après chaque passage, les symptômes ont été plus graves. De plus, les animaux inoculés ont diffusé à la fois l'ARN de L'APV détecté par RT-PCR et le virus vivant isolé par culture à chaque passage. Les titres moyens des anticorps ont également augmenté à chaque passage. Les résultats de la deuxième expérimentation n'ont pas été en total accord avec ceux de la première expérimentation. Dans la troisième expérimentation, l'APV, multiplié sur trois lignées cellulaires différentes, a été inoculé à trois groupes de dindes. Des symptômes ont été observés chez les animaux inoculés et le virus a pu être isolé à partir des trois groupes. Les résultats de cette étude préliminaire indiquent que le passage in vivo de l'APV-C chez les oiseaux peut augmenter la virulence du virus mais les résultats obtenus dans l'expérimentation 2, suggèrent que des études complémentaires sont nécessaires pour le confirmer.

Versuche zur Verbesserung eines Challengemodels für aviäre Pneumoviren des Subtyps C

Durch aviäre Pneumoviren verursachte Respirationserkrankungen führen in vielen Ländern zu starken wirtschaftlichen Einbußen in der Putenindustrie. Die Fortschritte in der Entwicklung von Vakzinen gegen diese Erkrankung waren in den USA nur gering, teilweise wegen des Fehlens eines beständigen Challengemodels zur Durchführung der Vakzine-Wirksamkeitsstudien. Diese Studie sollte klären, ob in vivo-Passagen eines US-APV-Isolats, bezeichnet als Subtyp C (APV-C), die Virusvirulenz erhöhen und damit gleichbleibende klinischen Symptome in Puten hervorrufen würden. Es wurden drei verschiedene Versuche durchgeführt. In Experiment 1 und 2 wurden Zellkultur-adaptierte APV viermal in vivo in Puten passagiert. Nach jeder Passage wurde ein Anstieg der Schwere der klinischen Symptome festgestellt. Außerdem wurde in jeder Passage nachgewiesen, dass die Tiere sowohl APV-RNS (durch RT-PCR) als auch Lebendvirus (durch Virusisolierung) ausschieden. Der mittlere Antikörpertiter stieg ebenfalls mit jeder Passage an. Die Ergebnisse des zweiten Versuchs stimmten mit denen des ersten Versuchs nicht vollständig überein. In einem dritten Experiment wurde in drei verschiedenen Zelllinien vermehrtes APV in drei Putengruppen inokuliert. Bei den inokulierten Tieren wurden klinische Symptome beobachtet und bei allen drei Gruppen konnte Virus isoliert werden. Die Ergebnisse dieser Vorversuche zeigen, dass in vivo-Passagen mit APV-C die Virusvirulenz im Tier erhöhen können, die Resultate aus Experiment 2 legen nahe, dass weitere Versuche notwendig sind, um dies zu bestätigen.

Ensayos para mejorar el modelo de infección experimental con el subtipo C de pneumovirus aviar

La enfermedad respiratoria causada por el pneumovirus aviar (APV) tiene un importante impacto económico negativo en el sector del pavo en muchos países. Los progresos en el desarrollo de vacunas frente a esta enfermedad realizados en U.S. avanzan lentamente, en parte debido a la falta de un modelo de infección experimental reproducible para llevar a cabo los estudios de eficacia de las vacunas. Este estudio se diseñó para determinar si los pases in vivo de un aislamiento de APV de US, conocido como subtipo C (APV-C), incrementarían la virulencia vírica, produciendo signos clínicos claros en pavos. Se realizaron tres experimentos distintos. En los experimentos 1 y 2, se realizaron 4 pases in vivo en pavos de un APV adaptado a cultivo celular. Después de cada pase, se observó un incremento en la gravedad de los signos clínicos. Además, se detectó que las aves infectadas experimentalmente excretaron tanto RNA de APV (mediante RT-PCR) así como virus vivo (mediante aislamiento vírico) en cada pase. La media de los títulos de anticuerpos también aumentó en cada pase. No hubo correlación absoluta entre los resultados del primer y del segundo experimento. En el tercer experimento, se inoculó APV replicado a partir de tres líneas celulares diferentes en tres grupos de pavos. Se observaron signos clínicos y se aisló virus en los tres grupos de aves infectadas. Los resultados de este estudio preliminar indican que el pase in vivo en aves de APV-C puede aumentar la virulencia vírica, aunque los resultados obtenidos en el experimento 2 sugieren la necesidad de llevar a cabo más estudios para confirmar este hallazgo.

Introduction

Avian pneumovirus (APV) causes acute respiratory tract infection in turkeys of all ages. The disease was first reported from South Africa in 1980 (Buys & Du Preez, 1980) and later from many European countries (Naylor & Jones, 1993; Jones, 1996). In the US, APV was first isolated from affected turkeys in Colorado (Kleven, 1997) and Minnesota (Goyal et al., 2000). APV is a Metapneumovirus in the subfamily Pneumovirinae, family Paramyxoviridae (Pringle, 1998). The disease is characterized by depression, coughing, nasal and ocular discharge, and swollen infraorbital sinuses with morbidity of up to 100% (Gulati et al., 2001; Jirjis et al., 2002). Mortality may reach 30% in cases complicated by secondary bacterial infections. Economic losses are due to mortality, a sharp drop in egg production, and increased carcass condemnation rate at slaughter because of air sacculitis (Gulati et al., 2001; Jirjis et al., 2002). On the basis of antigenicity and genetic characterization, four subtypes of APV have been recognized, designated A, B, C, and D. Of these, subtypes A, B, and D are present in Europe (Juhasz & Easton, 1994; Bayon-Auboyer et al., 2000; Cook & Cavanagh, 2002), but only subtype C (APV-C) is prevalent in the US (Seal, 1998; Seal et al., 2000; Dar et al., 2001).

Live attenuated vaccines have been developed in Europe against subtypes A and B of APV (Cook et al., 1989; Williams et al., 1991). In the US we have recently developed two live vaccines against APV by serial propagation of APV-C in cell cultures (Patnayak et al., 2002) and by cold adaptation (Patnayak et al., 2003a), and have conducted many experimental and field trials with these vaccines. One problem that we have encountered is the lack of a consistent challenge against APV-C. In other words, the challenge virus that produces severe clinical signs in one study may produce negligible clinical signs in the next. Other workers have also observed that disease induced by APV-C is not reproducible experimentally (Panigrahy et al., 2000; Alkhalaf et al., 2002; Turpin et al., 2002). In Europe, however, challenge against subtype A using virus passaged in tracheal organ culture has been reported to be successful, without loss of virulence even after passaging it 98 times (Williams et al., 1991). Various approaches have been tried to produce consistent clinical signs under experimental conditions, including co-infection of birds with APV and other bacterial (Cook et al., 1991 1999; Alkhalaf et al., 2002) or viral (Cook et al., 1999; Turpin et al., 2002) agents. However, an ideal challenge would be described as one producing clinical signs with the challenge virus alone in a dose-dependent manner. The first objective of this study was to develop a challenge model so that clinical disease is consistently reproduced in virus-challenged birds. To accomplish this, APV-C was passaged in vivo for four passages to determine whether in vivo passaging in birds increased virus virulence.

In previous studies, we used APV-C grown in Vero cells for challenging birds. Another objective of the present study was to determine whether APV grown in different cell lines would produce consistent clinical signs. For this purpose, APV grown in primary chicken embryo fibroblasts (CEF), a continuous cell line of CEF, namely DF-1, (Himly et al., 1998) and Vero cell-grown virus were compared. The belief was that the virus grown in cells of avian origin (homologous host) may be more virulent in turkeys.

Materials and Methods

Virus

A Minnesota strain of APV-C (APV/MN-2a) at passage level 7 was used. This virus was isolated in CEF cells from a respiratory disease outbreak in turkeys.

Cell culture

CEF cells prepared from 9-day-old to 11-day-old specific pathogen free embryonated chicken eggs (Charles River, North Franklin, Connecticut, USA) were used in this study. The cells were grown in minimum essential medium supplemented with penicillin (150 IU/ml), streptomycin (150 µg/ml), neomycin (50 µg/ml), fungizone (1 µg/ml), edamin-S (2%), and foetal bovine serum (8%).

Virus cultivation

The virus at passage 7 (APV/CEF7) was passaged three more times in CEF cells. Passage 10 virus (APV/CEF10) was harvested by two cycles of freezing and thawing, clarified by centrifugation at 1000×g for 10 min, aliquoted in small amounts and stored at −70°C. The stock virus had a titre of 10^5.63 median tissue culture infectious doses (TCID₅₀) per millilitre in CEF cells.

Experiment 1

APV/CEF10 was passaged four times in vivo in 4-week-old turkey poults. One-day-old turkey poults (BUTA, n = 25) were obtained from a commercial hatchery that has been consistently free of APV and antibodies against APV. Birds were procured at 5-day intervals at four different times in such a way that all birds at the time of inoculation were 4 weeks old. Thus, there were four groups of birds labelled 1 to 4. All birds were reared in the isolation facility at the College of Veterinary Medicine, University of Minnesota. At 4 weeks of age, 10 birds from each of the four groups were randomly bled to ensure that they were free of APV antibody. Birds in group 1 (n=15) were inoculated with 200 µl APV/CEF10 (50 µl into each eye and nostril) and were observed daily for the appearance of clinical signs until 2 weeks post infection (p.i.) (6 weeks of age).

On day 5 p.i., choanal clefts from all birds in group 1 were swabbed and each swab was placed in 0.5 ml minimum essential medium with 1% foetal bovine serum and antibiotics as already described. Then, 300 µl each swab fluid was pooled together. The remaining swab fluids were saved individually at −70°C for further use. During inoculation, the individual and pooled swab fluids were kept on ice. An aliquot of the pooled swabs was saved separately for testing for APV RNA by a reverse transcriptase-polymerase chain reaction (RT-PCR) (Shin et al., 2000), modified for use in a Taqman format, and for determining virus titres by titration in CEF cells. The remaining pooled material (labelled as APV/CEF10/T-1) was inoculated into birds (n=15) in group 2 (200 µl into each bird). These birds were also observed for the development of clinical signs for 2 weeks p.i. Following 5 days of infection, choanal swabs from birds in group 2 were collected and pooled in a similar manner. This pool (APV/CEF10/T-2) was used to infect birds in group 3 (n=15). Five days after infection, choanal swabs were collected and pooled from group 3 birds (APV/CEF10/T-3) and used to infect birds in group 4. The pooled swabs from groups 2, 3, and 4 were also tested for the presence of viral RNA by RT-PCR and for virus titre by titration in CEF cells.

Clinical signs

After each infection, all birds were observed daily for the development of clinical signs. Briefly, unilateral or bilateral nasal discharge, watery eyes and sinus swelling were noted and scored. Clinical scores were assigned to birds in all groups individually as described previously (Patnayak et al., 2002). After an observation period of 2 weeks, clinical scores of each individual bird were added together and the mean clinical score/group was calculated. The clinical sign scores in all four passages were statistically analysed using non-parametric analysis (Kruskall–Wallis test, Systat, Version 6).

Virus isolation and titration

The material from swab pools was diluted serially and inoculated into 96-well plates containing CEF cells for virus isolation and titration. The end points for virus replication were determined by indirect immunofluorescence (Jirjis et al., 2001). The titres were calculated by the Reed & Muench (1938) method and were expressed as TCID₅₀/ml.

Seroconversion

Birds in all groups were bled before infection and 21 days p.i. and the sera tested by enzyme-linked immunosorbent assay (ELISA) (Chiang et al., 2000) for APV antibodies. The birds were killed 3 weeks p.i. (7 weeks of age).

Experiment 2

The already described experiment 1 was repeated in a similar manner, with 10 poults in each group.

Experiment 3

Virus grown in three different cell lines was used to inoculate three different groups of turkeys. For this purpose, 45 1-day-old birds were procured from a commercial APV-free hatchery and reared in isolation facilities. Five birds were bled for baseline serology. At 4 weeks of age, 40 birds were divided into four groups (A1 to A4) of 10 birds each. Birds in group A1 were inoculated with 200 µl (50 µl into each eye and nostril) of APV/CEF7/Vero 5 (APV/CEF7 passed five times in Vero cells). Birds in groups A2 and A3 were inoculated with APV/CEF7/Vero5/DF5 (APV/CEF7/Vero5 passaged five times in DF-1 cells) and APV/CEF10, respectively. Birds in group A4 served as the negative control group. Clinical signs were recorded in birds from all groups following inoculation. The clinical scores were analysed statistically using non-parametric analysis (Kruskall–Wallis test, Systat, Version 6). Five days p.i., choanal swabs were collected individually from birds in all groups as described in experiment 1. An aliquot of pooled swabs was tested for viral RNA by RT-PCR and for the presence of live virus by passage in CEF cells.

Results

Experiment 1

Clinical signs

Following inoculation with APV/CEF10/T-1, nine of 14 birds showed clinical signs with a mean clinical score of 9.2/bird (Table 1). One bird in this group died due to non-specific causes. The severity of clinical signs increased with passage. Following inoculation with APV/CEF10/T-2, 14 of 15 birds showed clinical signs (mean clinical score 17.6/bird). Birds from in vivo passage 3 (APV/CEF10/T-3) exhibited more severe clinical signs (mean clinical score 22.3/bird). The severity of clinical signs further increased in group 4 (APV/CEF10/T-4) with a mean score of 28.8/bird. The peak signs in all groups were observed on days 5 and 6 p.i. The difference in mean clinical scores in all four passages was statistically highly significant (P < 0.001). Birds in group 5 (negative control) remained free of clinical signs throughout the study.

Table 1.  The effect of in vivo passage of subtype C avian pneumovirus on virulence

Inoculum^a	Mean clinical score^b	Virus titre (log10 TCID50/ml)^c	Virus detection by RT-PCR^e	Geometric mean antibody titre (ELISA)^f
Experiment 1
APV/CEF10/T-1	9.2 (9/14)	2.0	+	30 (9/12)
APV/CEF10/T-2	17.6 (14/15)	2.33	+	28 (11/15)
APV/CEF10/T-3	22.3 (14/15)	2.33	+	33 (11/15)
APV/CEF10/T-4	28.8 (15/15)	1.66	+	76 (14/15)
Negative control	0 (0/10)	<10^d	–	10 (0/10)
Experiment 2
APV/CEF10/T-1	11.5 (8/10)	1.66	+	27 (7/9)
APV/CEF10/T-2	15.4 (9/10)	2.5	+	35 (7/10)
APV/CEF10/T-3	7.6 (10/10)	0.0	+	40 (9/10)
APV/CEF10/T-4	9.6 (8/10)	1.66	+	70 (10/10)
Negative control	0 (0/5)	<10^d	–	10 (0/5)
^aFour in vivo passages (T1 to T4) were given to APV grown initially in CEF for 10 passages.
^bFigures in parentheses denote the number of sick birds/total birds. The birds were observed and scored for clinical signs daily.
^cVirus titres were calculated by the method of Reed & Muench (1938).
^dLimit of virus detection.
^eAt 5 days p.i., choanal swabs from all birds in a group were pooled and tested for APV RNA by RT-PCR.
^fA geometric mean titre of 10 or below is considered negative. Figures in parentheses indicate the number of birds positive/total birds.

Virus isolation

Virus titres following each in vivo passage are presented in Table 1. Live virus was isolated from all four in vivo passages, but not from the negative control group.

Detection of APV RNA

Pools of swabs from all in vivo passages were positive for APV RNA by RT-PCR (Table 1), while no viral RNA was detected in group 5.

Seroconversion

The geometric mean antibody titres of birds in all in vivo passage groups are presented in Table 1. The mean titres were found to increase at passage 4. Birds in the negative control group remained free of APV antibodies.

Experiment 2

This experiment was a replicate of experiment 1 with 10 birds in each group at each in vivo passage.

Clinical signs

In group 1, eight of 10 birds showed clinical signs (mean clinical score = 11.5/bird) (Table 1). In passage 2 (T-2), nine of 10 birds were positive for clinical signs with a mean score of 15.4/bird. In the further two in vivo passages (passages 3 and 4), 10 of 10 and eight of 10 birds showed clinical signs, respectively, but the mean clinical scores were lower than those observed in the T1 and T2 groups. The statistical analysis of the data showed that the clinical scores in the four passages were not significantly different. No clinical signs were observed in group 5 (negative control group).

Virus isolation

Pools of choanal swabs from birds in groups 1, 2, and 3 were positive for virus isolation. No virus was isolated from groups 3 and 5 (Table 1).

Detection of APV RNA

Choanal swab pools were positive for viral RNA as detected by RT-PCR (Table 1). No viral RNA was detected in group 5.

Seroconversion

The birds from all in vivo passages were positive for APV antibodies. The mean antibody titres increased at the fourth in vivo passage (Table 1). No APV antibodies were detected in group 5.

Experiment 3

Clinical signs

In this experiment, virus grown in three different cell types was inoculated into three different groups of birds. The maximum clinical score was observed in group A1 (APV/CEF7/Vero5). Eight of 10 birds showed clinical signs with a mean clinical score of 18.6/bird (Table 2). All birds showed clinical signs (mean clinical score of 15.5/bird) in group A2, inoculated with virus grown in DF-1 cells. Only six of 10 birds in-group A3, in which CEF-grown virus was inoculated, showed clinical signs, with a score of 6.01/bird. The difference in clinical scores in the three groups was statistically significant (P < 0.05). Birds in group A4 (negative control group) remained free of clinical signs throughout the study.

Table 2.  Pathogenicity of subtype C avian pneumovirus grown in three different cell types

Group^a	Inoculum^b	Clinical score^c	Virus titre (log10 TCID50/ml)^d	RT-PCR^e	Geometric mean antibody titre (ELISA)^f
A1	APV/CEF7/Vero5	18.6 (8/10)	2.5	+	46 (7/10)
A2	APV/CEF7/Vero5/DF5	15.5 (10/10)	2.5	+	98 (10/10)
A3	APV/CEF10	6.01 (6/10)	2.25	+	80 (10/10)
A4	Negative control	0.0 (0/10)	0	–	10 (0/10)
^aBirds in groups A1, A2 and A3 were inoculated with virus grown in Vero, DF-1 and CEF cells, respectively. Birds in group A4 served as the negative control group.
^bIn group A1, virus at passage 12 in Vero cells was used. Vero cell passaged virus was further passaged in DF-1 cells for five passages and inoculated in group A2 birds. In group A3, virus grown in CEF cells at passage 10 was inoculated. The initial titres of viruses were 10^4.8 TCID50/ml, 10^5.0 TCID50/ml and 10^5.6 TCID50/ml in Vero, DF and CEF cells, respectively.
^cFigures in parentheses denote the number of sick birds/total birds.
^dVirus titres were calculated by the method of Reed & Muench (1938).
^eAt 5 days p.i., choanal swabs from all birds were pooled and tested for APV RNA by RT-PCR.
^fSerum samples were tested for APV antibodies by ELISA. Figures in parentheses indicate the number of birds positive for APV antibodies/total birds. A geometric mean titre of 10 or below is considered negative.

Virus isolation

Pools of choanal swab fluid from all inoculated groups yielded live virus (Table 2). Samples from the negative control group were negative.

Detection of APV RNA

Swab fluids from groups A1, A2 and A3 were positive for APV RNA. Samples from group A4 were negative (Table 2).

Seroconversion

Birds in groups A1, A2 and A3 were positive for APV antibodies at day 21 as detected by ELISA. The mean geometric titres are presented in Table 2. The highest seroconversion (mean geometric titre=98) was observed in group A2 birds, which were inoculated with DF-1 propagated virus. The birds in the negative control group remained free of APV antibodies.

Discussion

Experimental APV infections are typically less severe than those observed under field conditions (Alkhalaf et al., 2002; Van de Zande et al., 2001; Turpin et al., 2002). To reproduce the disease experimentally, dual infection with APV and bacteria or APV and other viral agents has been tried. For example, Bordetella avium (Cook et al., 1991; Alkhalaf et al., 2002), Pasteurella spp. (Cook et al., 1999; Jirjis et al., 2001), Mycoplasma synoviae (Khera et al., 1999), Mycoplasma imitans (Ganapathy et al., 1998), infectious bronchitis virus (Cook et al., 2001), turkey herpes virus (Van de Zande et al., 2001), and Newcastle disease virus (Turpin et al., 2002) have been used along with APV. The results of these studies have been variable. In addition, to ascertain the efficacy of a vaccine, it is important to use the causative agent alone as a challenge rather than complicate matters with dual infections.

The results of experiment 1 showed that successive in vivo passages of APV-C resulted in increases in clinical signs and antibody response, indicating that virus virulence was enhanced by in vivo passages. This conclusion was further strengthened by the virus isolation and viral RNA shedding results. In the present study, a single pool of choanal swabs was tested at each in vivo passage to determine the presence of viral RNA following inoculation. In previous studies, we have used three different criteria to assess protection; namely, development of clinical signs, seroconversion, and a reduction in virus shedding (Patnayak et al., 2002 2003a b). Although, testing of individual choanal swabs would have given an indication of amount of virus shed by each individual bird, it was not the primary focus of the present study.

In experiment 2, 80 to 100% of birds showed clinical signs at each in vivo passage, but the clinical scores were not as high as those observed in experiment 1 and the severity of clinical signs declined with the last two passages. This discrepancy can possibly be explained by time of year in which these two experiments were performed. Experiment 1 was carried out in March and April while experiment 2 was performed in July and August. A seasonal trend of APV outbreaks has previously been reported in Minnesota (Shin et al., 2000; Goyal et al., 2003), APV outbreaks being observed between March and May and in October and November (Shin et al., 2000; Nezworski & Halvorson, 2002). Although our experiments were conducted in isolation units, variation in the temperature of the isolation building between seasons cannot be ruled out. In our previous studies also, we have empirically observed this seasonal effect. However, further studies are required to clarify this. Another reason for this difference could be experimental variation. In the first experiment, the clinical signs were enhanced dramatically following in vivo passages. In the repeated experiment, the extent of clinical signs was lower with in vivo passages but there was still a difference between inoculated and negative control groups. The absence of virus isolation in T-3 in experiment 2 is difficult to explain. One of the reasons could be the use of a single passage for the virus isolation attempts, but for definitive conclusions more experiments need to be performed.

In the third experiment, APV-C grown in Vero and DF-1 cells produced extensive clinical signs in 60 to 100% of birds. However, the virus grown in CEF cells produced only mild signs—although birds inoculated with all the three viruses produced humoral antibodies and shed live virus along with viral RNA. All these viruses were titrated in CEF cells to maintain uniformity. These results are surprising in that virus adapted to mammalian cells (Vero cells) was more virulent than that grown in avian cells (CEF cells). The Vero cell grown virus was 10-fold lower in titre than the CEF-grown virus but it produced the maximum clinical signs. When data on clinical signs obtained with CEF-grown virus in all three experiments were compared, not all birds showed clinical signs after the first inoculation. However, clinical signs did increase following in vivo passages in experiments 1 and 2. Since this observation is based on the results of a single experiment, it should be interpreted with caution.

Acknowledgments

This research was supported in part by research grants from Minnesota Turkey Growers Association and Midwest Poultry Consortium. The authors are thankful to Claudia Munoz-Zanzi for helpful suggestions.