Background: Despite individual low prevalence, rare diseases (>7,000) collectively affect a substantial population, estimated to be between 263 million and 446 million worldwide. Most (95%) rare diseases do not have approved treatment and approximately 45% cause neurological disorders. We initiated the “Treatment for those who have none” platform for Western Australian children with a rare disease in 2021, with support from a Channel 7 Telethon Trust research grant. The purpose of the platform is to accelerate treatments from bench to bedside for rare diseases using antisense oligomer-mediated modulation of gene expression.
Methods: Through the Rare Care Centre, Perth Children Hospital, we recruited patients diagnosed with a rare disease, and analysed the consequences of the mutation at the RNA levels. Based on these observations and known pathogenic mechanisms, we selected three patients, two with neurological disorders and one with kidney disease, for ASO design and assessment. We designed ASOs to either increase levels of gene expression for the children with neurological problems or redirect pre-mRNA splicing to induce a truncated protein with some functionality for child with kidney disease.
Results: Within 18 months, we have designed ASOs and performed proof-of-concept studies in patient-derived cells for three rare diseases Kleefstra syndrome, Birk-Landau-Perez Syndrome and Alport syndrome.
Conclusion: With this experience, we are now extending this platform nationwide through collaborations with Royal Brisbane Hospital, Queensland University Technology, University of New South Wales, Garvan Institute, Sydney Children Hospital network, Murdoch Children Research Institute, and Royal Hobert Hospital. We aim to integrate this program into the healthcare system as a treatment route for rare diseases.

Background: Retinitis pigmentosa (RP) and Leber congenital amaurosis (LCA) are some of the most common forms of inherited retinal dystrophy (IRD) leading to photoreceptor dysfunction and vision loss. Autosomal recessive forms potentially suitable for gene replacement therapy as used for treatment of RPE65 IRD, are caused by variants in genes including RPGR, AIPL1, RLBP1, PDE6B and RPGRIP1. Due to high clinical and genetic heterogeneity, there is a need to ascertain relevant biomarkers to assist genetic variant classification and provide patients with a diagnosis and opportunities for therapy. In this study, we aim to establish human models for elucidating disease mechanisms and testing novel AAV gene replacement therapies for variants in suitable RP and LCA causing-genes using induced pluripotent stem cells (iPSCs) differentiated to retinal organoids.
Methods: We generated patient-derived and CRISPR/Cas9 engineered iPSC lines carrying different variants, including pathogenic and variants of uncertain significance (VUS), in two genes modelling RP and LCA, respectively. All iPSC lines were differentiated to 32-week-old retinal organoids for immunohistochemistry, western blotting and transcriptomic analysis. For therapy testing, 21-week-old variant retinal organoids were transduced with AAV carrying wild-type cDNA of the retinal gene.
Results: Compared with controls, all variant organoids had diminished/reduced expression along with abnormal photoreceptor cell staining. Furthermore, these photoreceptors displayed inadequate function, either due to mislocalisation, aberrant binding characteristics or accumulation of enzyme substrate due to inactivity. At the transcript level, gene set enrichment analysis showed enrichment of abnormal transcriptomic profiles in all forms of mutant organoids compared with controls. Two months post AAV therapy of variant organoids, we observed induction of transgene protein expression and improved photoreceptor morphology compared to equivalent organoids that did not receive therapy.
Conclusion: This study demonstrates utility of iPSC-retinal organoids for functional genomics and therapy testing in retinitis pigmentosa and Leber congenital amaurosis.

Background: Gene therapy is a promising treatment for Hypomyelination with Brainstem and Spinal cord involvement and Leg Spasticity (HBSL), a potentially fatal disease caused by DARS1 gene mutations impacting brain and spinal cord white matter. We developed gene augmentation vectors (AAV.PHP.eB-hDARS1) for targeted CNS expression of human DARS1. In our existing set of transgenic mouse models, these vectors improved HBSL symptoms when delivered systemically. However, earlier intervention is needed to achieve a complete rescue. Here, we established a novel HBSL mouse model (‘Massi’) carrying two HBSL patient variants (Dars1A274V/D367Y), which suffers pre-natal mortality, to explore the potential of in-utero gene therapy. For minimal-invasive gene delivery, we tested the safety and efficacy of ultrasound-guided pre-natal AAV administration.
Methods: Dars1A274V/D367Y Massi mice were characterised using molecular, behavioural, histological and biochemical measurements. Using AAV.PHP.eB-eGFP, we assessed the safety and CNS-targeting of UltraSound-guided, PerCutaneous, Foetal IntraPeritoneal (USPCFIP) injections at gestation day 14.
Results: Dars1A274V/D367Y Massi mice show mild to moderate adult symptoms but high pre-natal mortality with a lower live birth rate (~20%) than expected from the Mendelian pattern of inheritance (50%). Preliminary results exploring USPCFIP delivery of AAV.PHP.eB-eGFP led to a 75% postnatal death rate in the eGFP-treated group, compared to 25% in controls. Vector copies were found in 2/17 newborn brains, while eGFP transgene expression was observed in 3/6 livers but not in brains.
Conclusion: Massi (Dars1A274V/D367Y) mice model moderate HBSL disease but display high pre-natal mortality, making them a valuable tool to study AAV-mediated in-utero gene delivery. USPCFIP injection allows CNS transduction in foetuses, but challenges include target accuracy and potential eGFP toxicity. Furthermore, rapid cell division of the developing CNS, which might obscure transgene expression and detection, is an important consideration for the future development of appropriate in-utero gene therapy strategies.

Background: Brittle bone disease (also known as osteogenesis imperfecta or OI) is a congenital bone fragility condition that leads to frequent fractures. Current pharmacotherapies do not address the underlying genetic cause and we have conceptualised gene repair strategies to correct disease-causing mutations. As a proof-of-concept, we have developed a novel approach for treating a patient with a severe OI caused by a 20 bp deletion (∆20) in COL1A1.
Methods: To model the disease, a human cell line was created possessing the COL1A1∆20 mutation that produces a frameshift and significant C-terminal missense readthrough. In addition, we generated a Col1a1∆20/+ mutant mouse with an analogous mutation. One challenge with CRISPR/Cas9 gene editing is efficiency, however we identified that this poison protein allele could be susceptible to gene disruption; a targeted SaCas9 gene disruption (non-homologous end joining) approach was trialled. Preclinical testing in the Col1a1∆20/+ mouse will be performed using a single-vector AAV approach, making use of our prior research targeting gene constructs to the bone.
Results: MicroCT analysis of bones isolated from the Col1a1∆20/+ mouse demonstrated reduced bone mass in the axial and appendicular skeleton. 4-point-bending and compression testing showed reduced biomechanical strength in the tibiae and vertebrae. SaCas9 targeting of the COL1A1∆20 mutant showed >70% disruption of the problematic readthrough allele with 0% effect in the wild type COL1A1 locus. Data from cell tracking studies using fluorescent reporter mice showed high efficiency, selectivity, and persistence in the bone compartment with an AAV8 approach that will be used for subsequent rescue studies.
Conclusion: We have generated novel cell and mouse models of a patient OI mutation. Our gene therapy approach has the potential to be transformative for OI patients and could also be adapted to treat to other conditions caused by dominant negative gene products.

Background: The rapid rise of antimicrobial resistance (AMR) poses a critical global health challenge, with projections indicating up to 10 million deaths annually by 2050 if current trends continue. Antimicrobial peptides (AMPs) offer a promising alternative to conventional antibiotics, with broad-spectrum activity, rapid action, and reduced resistance development. However, discovering natural AMPs is a complex, time-consuming, and costly process.
Methods: This study leveraged pattern recognition (PR) and artificial intelligence (AI) to discover novel encrypted AMPs within phage biological information. We built a PR and AI ensemble approach, integrating various machine learning techniques to improve prediction accuracy. Our PR algorithm, inspired by Ma et al. (2022) and Melo et al. (2021), detects AMP-like sequences based on physicochemical properties such as charge and hydrophobicity. Deep-learning models were then applied to identify sequences with low similarity to known AMPs and predict activity based on the amino acid sequence context.
We analysed the entire phage and viral pan-proteome, comprising 1.5 million proteins from UniProt. Peptides were grouped based on sequence similarity, charge, and hydrophobicity, selecting the top predictions for in vitro testing.
Results: Our ensemble approach identified hundreds of millions of peptides, narrowing down to 23 high-scoring AMPs for empirical testing. Initial screenings confirmed the antimicrobial activity of this subset, significantly reducing survival rates of multi-drug or pan-resistant pathogens like Pseudomonas aeruginosa and Staphylococcus aureus. These peptides demonstrated broad-spectrum activity and lower toxicity toward mammalian cells, and based on our preliminary results, are among the most potent AMPs reported to date. making them viable candidates for further pre-clinical development.
Conclusion: Our AI ensemble approach effectively identified novel AMPs within phage proteomes, showcasing potential as next-generation antimicrobials. These findings support the feasibility of using advanced computational tools to accelerate AMP discovery, providing a valuable framework for the inexpensive development of bioinspired antimicrobials to combat AMR.

Introduction: Disease-causing variants in PROM1 lead to inherited retinal dystrophies (IRD), characterized by both cone and rod dysfunction with various retinal sub-phenotypes reported. The aim of this study was to investigate disease mechanism in mouse and patient-derived retinal organoid model systems and to aid pre-clinical assessment of novel therapy approaches.
Methods: To study PROM1 disease, we used two models: a mouse model and patient-derived induced pluripotent stem cells (iPSCs) differentiated into retinal organoids (ROs). Serial electroretinograms (ERG) and optical coherence tomography (OCT) were conducted from P21 to P180 and compared to wild-type mice. Histological analyses of mouse retinal sections were performed at postnatal days (P) 14 to P90.
Immunohistochemistry (IHC) assays targeted cone and rod-specific markers, as well as glial markers. Retinal organoids matured until day 210 (D210) were collected for morphological assessment, IHC and RNA extraction. Transcriptomics was performed using RNA-Seq to investigate differentially expressed genes (DEGs).
Results: Prom1 mutant mice showed retinal degeneration starting from P14 over a 6-month period compared to controls, based on OCT, ERG, histological, immunohistochemical studies. OCT quantification showed a significant difference in retinal thickness at P30, and by P180, the mean retinal thickness difference was 65% of wild type. Electrophysiology indicated cone dysfunction at P21, shown by reduced flicker function amplitude. Histology revealed reduced outer nuclear cell density at P14. ROs with pathogenic bi-allelic PROM1 variants exhibited an absence of huPROM1 and a reduction in cone numbers compared to controls. Transcriptomics revealed significant downregulation in pathways related to angiogenesis, eye morphogenesis, and bleb assembly.
Conclusion: Autosomal recessive PROM1 IRD causes both rod and cone dysfunction, as demonstrated in a mouse model and patient-derived ROs. Evidence from this study suggests that cone dysfunction may precede rod loss, providing insights into the varied phenotypic and genotypic presentations of the disease and biomarkers for therapy assessment.

Background: Healthcare professionals (HCPs) are responsible for communicating with families of children with hard-to-treat cancers on issues surrounding experimental treatments and paediatric precision medicine. Co-designing with HCPs enables the simultaneous development of targeted strategies and optimal uptake of these resources. The Theoretical Domains Framework (TDF) is an innovative implementation science framework that facilitates understanding of the underlying determinants of behaviour, such that we can find the best approaches to facilitate change. We used the TDF to identify determinants of behaviour relevant to HCP communication skills in complex paediatric cancer treatment, and to develop accessible strategies to address implementation problems and support resource dissemination.
Methods: We interviewed Australian HCPs who had direct responsibilities in managing children/adolescents with a hard-to-treat cancer within the past 24 months. Interview transcripts were qualitatively coded to domains of the TDF, and behaviour change techniques were identified. Based on these interviews, we developed a video resource for HCPs as well as recognised strategies to achieve optimal resource dissemination.
Results: We interviewed 10 oncologists, seven nurses, and three social workers who identified several challenges for communication with families including: balancing information provision while maintaining realistic hope; managing their own uncertainty; and nurses and allied health workers being under-utilised during conversations with families. Key determinants of behaviour were environmental context and resources, knowledge, and skills. Behaviour change techniques included seeking support from colleagues, role-playing difficult conversations, and understanding the potential impact these discussions have on families.
Conclusion: Resources are needed to provide inclusive support for all types of HCPs in communication with families of children with hard-to-treat cancers, particularly when using precision medicine in paediatric cancer. By identifying implementation strategies, our research may innovate the integration of future resources and make them accessible in HCP training, such that best-practice person-centred care is achieved with a wider range of families.