The FAdV-1 strain used in this experimental study was isolated from an outbreak of gizzard erosion in Germany . This virus was grouped together with European “pathogenic” FAdV-1, according to nucleic acid sequences of the long and short fiber. These European “pathogenic” strains differ in their PCR - restriction fragment length polymorphism (RFLP) patterns from previously reported and investigated Japanese “pathogenic” FAdV-1 as described by Marek et al. . Experimental oral infection of day-old SPF broilers as well as commercial broilers with maternal FAdV-1 antibodies with this European FAdV-1 isolate resulted in clinical signs, macroscopical and histological gizzard lesions comparable to those described from the natural outbreak, demonstrating the pathogenicity of this strain.
In the present study, experimentally infected broilers with adenoviral gizzard erosion showed clinical signs of growth retardation, reflecting production data collected from the natural outbreak. Similar to this, decreased weight gain was described by Okuda et al.  in an experimental study of 1-week-old and 3-week-old commercial broilers. The generated data as well as information collected from field outbreaks [6, 8, 9] confirm that growth retardation in the course of FAdV-1 induced gizzard erosion may present a serious problem for the economic success of broiler production. Similar to the majority of other experimental studies set up to reproduce gizzard erosion, no other clinical signs were observed in the present investigation. This is in direct contrast to the experimental work reported by Domanska-Blicharz et al. , who noticed 100% mortality in day-old birds infected with a Polish FAdV-1 strain by nasal and ocular routes.
Although experimental reproductions of FAdV-1 induced gizzard erosion have been described [10, 11], this is the first study aiming to link viral load in gizzards beside other tissue samples and cloacal swabs with the appearance and severity of gizzard lesions. In order to achieve this, a recently developed real-time PCR was applied , offering certain advantages in comparison to reported cell-culture based titrations methods and conventional PCR to detect FAdV-1. Quantifying the viral load in sample material by tissue culture requires special facilities and is less suitable for processing a large number of samples. Furthermore, these investigations may present an incomplete picture, considering the high sensitivity of the real-time PCR, which was demonstrated in a previous study investigating viral shedding of all FAdV species in experimentally infected birds . In the present study viral DNA was detected in nearly all samples investigated throughout the experiment, whereas virus isolation on cell culture detected fewer numbers of positive samples. This result underlines the applicability of real-time PCR for investigating organ samples. However, variations in sample preparation limits the comparison between data obtained for organ samples and those recorded for cloacal swabs.
In the present study the highest viral genome copies per reaction were found in the gizzard. Furthermore, in accordance to previous experimental studies of FAdV-1 induced gizzard erosion , successful virus re-isolations from organ samples were shown predominantly from the gizzards. Comparing results obtained by real-time PCR with those from virus isolation, it can be concluded that virus isolation may succeed more easily from samples with higher amounts of viral DNA as they represent higher amounts of viable virus. Overall, the calculated viral load from the gizzards was followed by similar quantitative and chronological distribution patterns in other investigated samples. Nevertheless, even the highest genome copy numbers in other investigated organs did not reach the numbers calculated for gizzard samples. Pathological changes and detection of the virus in the gastrointestinal system, particularly in the gizzard, in the course of a FAdV-1 infection are contrary to other documented FAdV infections, such as inclusion body hepatitis and hydropericardium/hepatitis syndrome, both targeting the liver with corresponding tissue destruction . A putative FAdV-1 cell receptor in gizzard mucosal cells, described by Taharaguchi et al. , may explain this tissue tropism of virulent FAdV-1.
Overall, the levels of viral DNA in liver and gizzard achieved their highest numbers in groups SPFB at 7 dpi and CB at 10 dpi. At the same time, onset of clinical signs was observed and macroscopical lesions in the gizzard mucosa appeared most severe. Additionally, histolopathological changes, in particular numbers of intranuclear inclusion bodies in the gizzard epithelial cells, were assessed concurrently with the highest scores. However, erosion of the gizzard koilin layer was found to be most distinct at later dates, when viral load was already declining and intranuclear inclusion bodies were found far less frequent or absent. Based on these data, the development of adenoviral gizzard erosion reflects a cascade of reactions in which infection and subsequent necrosis of mucosal epithelial cells play a crucial role , even though the precise mechanism needs to be resolved.
In accordance with a previous experimental study , our results underline that the development of gizzard erosion is largely due to a localized FAdV-1 infection in the gizzard, with the inability of maternal antibodies to prevent the disease. Nonetheless, the presence of maternal antibodies in experimentally infected commercial broilers delayed the formation of gizzard erosion for a short time compared to the progress of infection in SPF broilers, suggesting that viremia may play a role in gizzard erosion pathogenesis as well. In contrast to previous experimental studies [7, 10, 11], the systemic distribution of live virus to the liver was confirmed in SPF and commercial broilers. In the spleen viable FAdV-1 could only be detected in SPF broilers which might reflect an influence of maternal antibodies. The systemic distribution of the virus to the liver and spleen was confirmed in the majority of infected birds by real-time PCR, with a tendency of lower numbers of viral DNA in samples of commercial broilers. Nevertheless, similar to previous experimental studies [11, 12, 20], neither macroscopical nor histopathological changes were documented in the liver or spleen, no adenoviral inclusion bodies were observed and no viral DNA was demonstrated by in-situ hybridization, suggesting that replication may not be very efficient in these organs, if it takes place at all.
The actual study also demonstrates that diagnosis of adenoviral gizzard erosion based on macroscopical lesions of the koilin layer, virus isolation or the appearance of typical adenoviral inclusion bodies alone may be inconclusive. Ono et al.  postulated that cellular infiltration in the lamina propria as well as mononuclear infiltration in the perivascular tissue of submucosa and muscle layer, both of them were also observed in the course of our experimental study, may aid in histological diagnostics of adenoviral gizzard erosion. Furthermore, the actual study demonstrates that the highly sensitive real-time PCR is a useful tool to complete the range of diagnostics in the course of adenoviral gizzard erosion.
To summarize, for the first time experimental infection with a European FAdV-1 succeeded in reproducing typical clinical signs and pathological changes of adenoviral gizzard erosion. Both, SPF broilers and commercial broilers with homologous maternal antibodies, developed typical gizzard lesions as well as clinical signs of decreased weight gain, confirming that maternal antibodies have no protective effect in the course of an infection with a virulent FAdV-1. Investigations of samples from experimentally infected broilers over time with a recently developed real-time PCR showed that maximum viral load in the gizzard and other investigated samples coincided with the onset of clinical signs as well as maximum macroscopical and histological lesions, altogether helpful to elucidate the pathogenesis of the disease.