Living model systems, ranging in complexity from bacterial culture to non-human primates, are a cornerstone in disease biology research. Despite their unquestionable usefulness, the disease modelling community remains acutely aware of the challenges and limitations of any individual model. To describe our collective predicament, we often (mis)use the quote by statistician George Box, ‘All models are wrong, but some are useful’.
Whilst no individual model is perfect, the sheer breadth of modelling systems available empowers us to answer increasingly complex questions in human disease research by using multiple models and leveraging their complementary advantages to mitigate individual weaknesses. Disease Models & Mechanisms (DMM) publishes articles that illustrate the success of such synergistic approaches across all fields of human disease biology (Fig. 1). A recent study by Gignac et al. (2023), maximised key strengths in complementary models to demonstrate the role of the adapter protein dishevelled 1 (DVL1) in skeletal phenotypes of the rare Robinow syndrome. The authors used chicken embryos to study limb skeletal development as it is highly conserved in humans, and they studied wing morphology in Drosophila as this genetically tractable model allows more precise control of the human DVL1 transgene. The authors also employed luciferase reporter assay in human embryonic kidney 293 (HEK293) cells, alongside in vivo readouts, to gain a more-mechanistic insight into aberrant DVL1 signalling in Robinow syndrome. Similar strategies in study design can be used to answer specific questions related to differing extent of biological complexity, from intricate molecular mechanisms to systemic processes. In a recent study, Almeida et al. (2023) discovered that the deubiquitinase OTULIN, which is involved in immune and inflammatory pathways, is key for neural tube formation in mouse embryos. To probe the mechanism of this, the authors then used Madin–Darby canine kidney cells, HEK293 cells and a human breast cancer cell line and found that OTULIN interacts with several key proteins involved in planar cell polarity, a process that is integral in several developmental and homeostatic processes.
Layering multiple disease models can also balance the level of biological complexity with the evolutionary distance from humans in the study design. As highlighted earlier, Drosophila are a valuable system to study human disease, but validating findings from flies in human cells can strengthen results. Likewise, Kervadec et al. (2023) developed a high throughput multi-model platform to study atrial fibrillation, which is associated with cardiac arrythmia. To screen the effects of atrial fibrillation-associated genes on arrhythmic phenotypes, the authors generated human induced pluripotent stem cell (iPSC)-derived atrial-like cardiomyocytes, alongside an assay to monitor cardiac contraction in flies. They then used computational models of human adult atrial myocytes and tissue to validate the top hit in this screen, and delved further into its mechanistic interactions. Similarly, Travaglio et al. (2023) used Drosophila and human iPSC-derived neural precursor cells (NPCs) to deepen our understanding of mitochondrial dysfunction in early-stage Parkinson's disease, and Faria et al. (2023) used Drosophila and a human mammary epithelial cell line to uncover the role of P-cadherin (CDH3) in the early stages of breast carcinogenesis.
Although mice are evolutionarily closer to humans than Drosophila, human cells also help validate findings in murine disease models. For instance, Zhang et al. (2023) induced iron-overload in mouse and human iPSC-derived retinal pigment epithelium (RPE) cells to better understand the role of reactive oxygen species in age-related macular degeneration. Furthermore, Guyer et al. (2023) used both human and mouse neuroblastoma cell lines to study gene expression and plasticity in distinct tumour cell states. By using the mouse neuroblastoma cell line, the authors were able to perform an isograft with these cells to study how the distinct tumour cell states induce tumorigenesis in vivo.
Directly integrating patient samples and data into preclinical research is a powerful strategy to enhance the translational potential of findings (Cheng et al., 2022). In one study, Stevenson et al. (2023) utilised multiple transcriptomic datasets derived from samples obtained from Wilms tumour patients to reveal upregulation of the WNT ligand WNT5A in SIX1/2 mutant tumours, and then verified the interaction between WNT5A and SIX1/2 in Wilms tumour samples and in HEK293 cells. Data from patients with the developmental disorder diphthamide-deficiency syndrome were also integral in the research by Ütkür et al. (2023), as the authors functionally assessed several known and previously uncharacterised patient-derived genetic variants in yeast and a human breast cancer cell line. A study by Lachgar-Ruiz et al. (2023) also identified novel variants of the CCDC50 gene in a Spanish family with non-syndromic sensorineural hearing loss. Interestingly, the authors did not find hearing impairment in two loss-of-function Ccdc50 mutant mouse lines but, when the specific CCDC50 variants identified in the patients were introduced into a mouse fibroblast cell line, they caused abnormal protein distribution. The functional characterisation of variants is essential to understand the mechanisms of genetic disorders and the above studies highlight how using multiple model systems can either strengthen findings or prevent key discoveries from being missed.
Research questions in biomedicine are becoming increasingly complex, but technological advances and enhanced crosstalk between fields are helping to advance our understanding for the benefit of patients (Caldwell, 2023; Devlin and Roberts, 2022; Sansom, 2022; Svenson et al., 2022; Verheyen, 2022; White and Patton, 2023). The examples discussed above demonstrate that integrating multiple animal model systems in the same study is becoming the norm.
Since its launch, DMM has aimed to publish multidisciplinary basic and translational research in disease biology. Beyond being a venue for cutting-edge research (Patton, 2021), we are committed to the disease modelling community by fostering open discussion, promoting new collaborations and supporting the next generation of biomedical scientists. To continue our support for the disease modelling community, we are sponsoring the ‘Host-pathogen and host-commensal interactions’ session at The Allied Genetics Conference in March 2024. We also actively encourage interdisciplinary collaborations by hosting journal meetings, like the upcoming ‘Pre-clinical Modelling of Human Genetic Disease and Therapy’, being held in Edinburgh, UK, in May 2024. We believe Open Access journals, such as DMM, have the potential and the responsibility to be community forums for exchanging ideas and for establishing and strengthening interdisciplinary collaborations in your area of biomedical research.
