Escherichia coli represents a significant challenge to the poultry industry due to compromised animal welfare, vast productivity losses, elevated mortality, and increased use of antimicrobial compounds. Therefore, effective preventive strategies and insight into the pathogenesis and disease mechanisms of colibacillosis are essential to secure a healthy poultry production. Consequently, discriminative in vivo models of colibacillosis are prerequisite tools for evaluating e.g., preventive measures, exploring novel treatments and understanding disease development. Numerous models of colibacillosis are applied for experimental studies in poultry. Yet, few studies provide a proper characterisation of the model enabling other authors to reproduce experiments or use the model in general. The present paper provides a literature review on avian in vivo models of primary colibacillosis.
In poultry, infection with avian pathogenic Escherichia coli (APEC) constitutes a significant health challenge compromising both animal welfare, productivity and results in the usage of antimicrobial drugs [1, 2].
As one of the commonest diseases in the most abundant domestic livestock species worldwide , colibacillosis is a major contributor to vast economic losses and widespread animal suffering. Thus, means to control colibacillosis, e.g., effective vaccines, are highly warranted, and in-depth insight into pathogenesis and disease mechanisms is vital.
To ensure effective control, whether therapeutic or preventive, discriminative animal models are essential. Likewise, there is a need for valid in vivo models in the quest to study the mechanisms behind the disease .
When evaluating in vivo models the concepts of validity must be considered. In models used to assess preventive strategies or therapeutic effectiveness, it is essential that the model holds proper predictive validity—i.e., are the results obtained using the model similar to the outcome within the spontaneous infections? In other words, how well does the model predict what would truly happen within the natural host of the disease [4, 5]. Failure of a model to exhibit proper predictive validity can result in a false sense of security if, e.g., a vaccine is ineffective under field conditions, treatment fails, or it could, conversely, lead to faulty abandonment of, e.g., preventive strategies, which could have hindered animal suffering and economic losses. Construct validity is particularly important when investigating disease development as this concept applies to the similarity between the mechanisms resulting in disease within the natural host and the model . Thus, are the pathogenesis and mechanisms of disease within the model similar to the naturally occurring condition? Or could it be, e.g., unnecessarily invasive and lack important steps of disease development. Face validity concerns the model’s ability to mirror the actual condition. For example, if the clinical signs and pathomorphology of lesions are similar to those of spontaneous disease .
In vivo models exhibiting high validity on these essential concepts exemplifies a discriminative animal model. A discriminative infection model would, therefore, mimic the natural route of infection, pathogenic agent, disease progression, clinical signs, gross and histopathological changes, and the immunocompetence should be equal to that of the natural host succumbing to disease .
Another vital concept in animal studies is the appropriate usage of a proper study design, e.g., randomisation, blinding and adequate controls . Likewise, proper reporting of methods and results is essential. The methods should be described in sufficient detail, allowing others to evaluate and use the model, whilst adequate reporting on outcomes and results enables researchers to assess whether the model fits their study.
In the present review, the primary aim was to examine the available literature, describing the development or characterisation of in vivo models of colibacillosis in poultry as their main purpose. This was done to establish an overview of inoculation methods, inoculum characteristics, animal characteristics, housing conditions and the methods applied to assess the outcome of the study. Evaluation of the experimental study design constituted a secondary aim, seeking to elucidate the use of lege artis principles, including randomisation and blinding, in in vivo studies conducted in poultry.
A systematic review was conducted through a literature search of the electronic databases PubMed  and Web of Science , with “All databases” set as default in the latter. The search terms were as illustrated in Table 1, and only peer-reviewed papers eliciting original research utilising avian animal models of E. coli infections were included in the current review. Inclusion criteria were studies stating the development and/or characterisation of an in vivo model of colibacillosis as a main study scope, purpose or aim. Exclusion criteria were as follows: all non-original research articles (e.g., meeting or conference abstract), models without application of viable E. coli bacteria (e.g., lipopolysaccharides instead of live bacteria) and models applying dual/co-infections with other pathogens or pharmaceutical immunosuppression. In addition, studies utilising only avian embryos were excluded as well as colonisation- and, e.g., transmission studies not aiming to cause or evaluate infection. Publications written in other languages than English were also excluded. Abstracts of all the papers identified through the two databases were thoroughly assessed for their relevance to decide upon either inclusion or exclusion. The included papers were reviewed in detail, and systematic registrations were made, including, but not limited to, the use of controls, randomisation, and blinding, group sizes, animal information (e.g., breed gender, age), methods utilised in the assessment of the model and the use of statistics for evaluation.
Abstracts from PubMed (n = 561) and Web of Science (n = 583) were identified through the search terms and subsequently examined. Of these, a total of 14 papers met the inclusion criteria without succumbing to any of the criteria of exclusion. A single study not explicitly stating model development as a main goal was included due to the application of a route of inoculation (intra-navel) not described by the other studies . Although the authors have made considerable efforts to identify relevant publications, some might have been missed due to, e.g., unclear study aims, not being obtainable online or identified by the included search terms.
Amongst the described models, the majority were conducted in the species Gallus gallus domesticus of both laying and meat-production types and of various ages (Table 2). The reports on the breed or line of animals, as well as research animal provider, varied, whereas the age of the birds and group sizes were readily provided. Other information related to the animals, such as housing conditions, temperature within the facilities and the feed provided, was often sparse (Table 2).
A wide span of inoculation routes was exploited to establish colibacillosis (Table 3). Those considered invasive were based on direct inoculation into the vagina, uterus, oviduct, trachea, peritoneum, air sac, or subcutis (Table 3) [10,11,12,13,14,15,16,17,18,19]. Among these, intratracheal (IT), oviduct, intraperitoneal (IP), intra-air sac (IAS) and subcutaneous (SC) inoculation would be considered most invasive, and thus, those having the lowest construct validity, as they bypass important steps of the initial pathogenesis and natural defence mechanisms. Yet, utilising these inoculation routes might still provide valuable results for some studies. Contrary to invasive methods of inoculation, a number of models sought to mimic the natural infection closely by applying, e.g., aerosols [5, 16, 20]. Also, several studies reported models of cellulitis, which were based on superficial scratching of the skin followed by housing on litter inoculated with E. coli [21, 22]. Though the pathogenesis of cellulitis in broilers is not completely clear, invasion through a compromised skin barrier is definitely likely to be a common route of infection [22, 23], however, it could also be a result of E. coli septicaemia. The models listed in this review were capable of reproducing characteristic lesions through skin scratches [21, 22]. Coliform cellulitis is an important and high-cost manifestation in broilers due to condemnations, and, therefore, it is not surprising that multiple models have been developed [14, 18, 21].
The intravaginal (IVAG) and intrauterine (IU) models reproduced peritonitis variably [13, 16], and one study reported infection highly aided by concurrent IP administration of sterile egg yolk . A study, depositing E. coli directly into the oviduct during a surgical procedure, reported peritonitis but not signs of salpingitis though cultures from the oviduct yielded E. coli growth . An IAS model described subjected the birds to transportation stress following inoculation to aid induction of disease . In this study, they reported an increased susceptibility to infection in males and fast-growing lines of turkey, hence underlining the effect of animal characteristics and, thus, the necessity to properly state such information within publications.
Reports on the inoculum were generally comprehensive, with details on isolate origin, dose, vehicle used etc. (Table 3). However, the importance of information on dose (colony forming units (CFU) received by each animal), and the volume the bacteria is administered within, should still be emphasised, as it has previously been reported that, e.g., the concentration can impact the severity of disease  or that the sheer volume might act as an irritant per se .
Evaluation of the disease outcome by histopathology was not particularly common among the identified studies, whereas gross pathology was readily reported (Table 4). Also, numerous studies provided a detailed scoring or recording scheme for gross pathology [11,12,13, 15, 19, 20]. Likewise, microbiology and statistics were widely applied to assess the models (Table 4).
A randomised order of animal assessment and blinding was practically absent throughout the examined studies. Particularly, when comparing subjective parameters or subtle changes, blinding is crucial to eliminate bias and should always be prioritised in in vivo research studies, whilst randomisation is a key component in controlling variation . Both these elements are also crucial for the 3R principle of “Reduction”, as valid research results are unlikely to be generated without proper control of variation and blinding .
As pointed out by Piercy and West in 1976, comparison of results regarding, e.g., pathogenesis and treatments are limited due to a lack of uniformity in experimental techniques highlighting variation in culture media, stage of growth, concentration etc. . In the current review, the authors acknowledge such limitations but wish to emphasise that the lack of sufficient information serves as the main problem as this prohibits studies from being evaluated, compared and/or reproduced. Also, whilst an almost unlimited number of studies apply animal models of colibacillosis, only very few focus on the description and development of these models, thereby failing to provide usable references and share experiences with other authors.
In the present literature review, an overview of avian models of colibacillosis is given with a comprehensive presentation of details related to the animals, inoculum, inoculation, housing, and model assessment. Numerous portals of infection have been successfully applied to reproduce colibacillosis in poultry with varying degrees of invasiveness. Thus, not all the models would be considered particularly discriminative as essential steps of the natural pathogenesis were often bypassed. Randomisation and blinding during outcome assessment were rarely performed in the studies and should be included as a standard in the future.
Availability of data and materials
A list of the reviewed literature can be obtained from the corresponding author.
Nolan LK, Vaillancourt J-P, Barbieri NL, Logue CM. Colibacillosis. In: Swayne DE, Boulianne M, editors. Diseases of poultry. Hoboken: Wiley; 2020. p. 770–830.
Festing MF, Overend P, Das RG, Borja MC, Berdoy M. The Design of animal experiments: reducing the use of animals in research through better experimental design. 2nd ed. Thousand Oaks: SAGE Publications; 2016.
Allan B, Wheler C, Koster W, Sarfraz M, Potter A, Gerdts V, et al. In ovo administration of innate immune stimulants and protection from early chick mortalities due to yolk sac infection. Avian Dis. 2018;62:316–21.
Antao E-M, Glodde S, Li G, Sharifi R, Homeier T, Laturnus C, et al. The chicken as a natural model for extraintestinal infections caused by avian pathogenic Escherichia coli (APEC). Microb Pathog. 2008;45:361–9.
Alber A, Costa T, Chintoan-Uta C, Bryson KJ, Kaiser P, Stevens MP, et al. Dose-dependent differential resistance of inbred chicken lines to avian pathogenic Escherichia coli challenge. Avian Pathol. 2019;48:157–67.
Macklin KS, Norton RA, McMurtrey BL. Scratches as a component in the pathogenesis of avian cellulitis in broiler chickens exposed to cellulitis origin Escherichia coli isolates collected from different regions of the US. Avian Pathol. 1999;28:573–8.
Soerensen KE, Skovgaard K, Heegaard PMH, Jensen HE, Nielsen OL, Leifsson PS, et al. The impact of Staphylococcus aureus concentration on the development of pulmonary lesions and cytokine expression after intravenous inoculation of pigs. Vet Pathol. 2012;49:950–62.
HEJ and SK fostered the idea of the review and discussed the results. SK drafted the initial manuscript. All authors read and approved the final manuscript.
SK is a veterinarian and PhD who conducted her PhD project in the field of avian colibacillosis. HEJ is a professor of Veterinary Pathology who has worked and published extensively within the area of infectious disease, including experimental animal models.
This review did not require official or institutional ethical approval.
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.