Developing a phage therapy testing-tomanufacture platform for treating multidrug resistant Pseudomonas aeruginosa in Manchester hospitals.

Collaborators:

Dr. Rosanna Wright, Prof. Michael Bromley, Dr. Timothy Felton, Dr. Chris Kosmidis, Dr. Stephanie Thomas.

Every year in Manchester’s hospitals, the number of patients suffering from multi-drug resistant bacterial infections rises. Increasingly, multidrug resistant bacterial infections are seen which can no longer be treated with antibiotics, leading to worse patient outcomes and substantial cost to the healthcare system. Phage therapy is a promising alternative to antibiotic for treating multidrug resistant bacterial infections. Phage therapy has successfully been used as a last-resort treatment, however the UK lacks both testing infrastructure for determining which phages to use and manufacturing capability to safely produce phage therapeutics. Our project aims to bridge this gap by developing a pilot scale testing-to-manufacture pipeline at Wythenshawe Hospital.  

Phages are viruses that specifically target and kill bacteria whilst being harmless to humans. The high-specificity of phage therapy also means that it is less disruptive to the patient’s healthy microbiota. We have developed a new high-throughput screening technology enabling us to test susceptibility of multi-drug resistant P. aeruginosa from clinical infections against 100s of phages rapidly and in parallel. We will embed this technology at Wythenshawe Hospital to enable rapid susceptibility testing and phage therapeutic design. Using lab-based tests we will optimise dosing regimens to maximise therapeutic effectiveness whilst minimizing the evolution of resistance in treated bacteria. Finally, we will establish phage manufacturing in UK based on state-of-the-art approaches currently used in Belgium and Australia. This project will deliver the first testing-to-manufacture pipeline for phage therapy within the NHS.  

Determining tissue-of-origin from cfDNA methylation profiles for diagnosis in Cancers of Unknown Primary.

Doctors diagnose a patient with “Cancer of Unknown Primary” (CUP) when cancer has spread to another part of the body and doctors cannot tell where it has originated from using the usual tests including a biopsy. This is an important challenge to address as CUP is the 6^th most common cause of cancer death in the UK. For some patients with CUP, doctors are highly suspicious of a known cancer type and therefore manage and treat them as per that suspected primary cancer type. For this group of patients with CUP, access to treatments and therefore survival is similar to that of patients with a diagnosis of a specific cancer type. Unfortunately, this is not the case for most patients diagnosed with CUP leaving them with limited treatment options as we cannot determine where the cancer started.  

We have developed a blood test that uses DNA released by cancer cells into patient’s blood to predict where the cancer originally came from. This blood test will be less invasive than using a tumour biopsy to gain this information and we expect that it will also be quicker. Examining the tumour DNA in the blood may provide a better understanding of CUP and help doctors to select the best treatments.

The purpose of this project is to further improve our blood test to make it quicker, with high sample throughput, resulting in a reduced time to diagnosis of the original tumour, enabling access to new treatments options. We will then test our improved blood test to ensure that the results are reliable and reproducible. 

If we can prove that our new blood test is reliable, we will test it further in future clinical trials and ultimately hope to see it used across the NHS and beyond to maximise benefit for patients. 

CG-X: Explainable automated ECG interpretation for Long QT Syndrome – validation with further data.

Collaborators: Prof Rick Body, Dr Lukas Hughes-Noehrer, Dr Anthony Wilson.

Sudden cardiac death accounts for 100,000 UK deaths per annum. This research takes a completely new approach to ECG interpretation, allowing people to monitor ECGs easily, potentially saving thousands of lives every year. It automates the early detection of long QT syndrome (LQTS) – a symptomless heart condition caused by commonly prescribed medications that can lead to sudden death, even in young apparently healthy people.  LQTS can also be congenital or acquired (primary from pharmacological drugs, as well as secondary from cardiac and non-cardiac conditions). To do this, we use human-like AI that is intuitively ‘explainable’.   

Interpreting ECGs is extremely challenging. It requires years of training, and there are currently no computerised approaches reliable enough to use in clinical practice. Measuring the QT-interval is particularly difficult, due to the challenge of determining the start and the end of the ECG waves, which differ considerably in their characteristics across individuals.   

Our multidisciplinary team combined knowledge from cognitive psychology, medicine and computer science to produce an explainable algorithm that works with >90% accuracy and is easily understood by both clinicians and lay people, as it can be visualised using colour superimposed on the ECG signal. We are currently at D3 (proof of concept) on the translational research pathway, and in this study, we seek to move to D4 (optimisation).  The algorithm was tested with ECGs from a large clinical study examining four drugs that prolong the QT-interval. It is necessary to test with a wider variety of ECGs from different sources, including congenital and acquired LQTS to ensure its efficacy and generalizability.     

ECG signal can be affected differently by different genetic mutations, cardiac and non-cardiac conditions, and/or medications, so it is important to test and optimise the algorithm across a range of these (including different types of congenital and acquired LQTS, from drug-induced to cardiac conditions (e.g. myocardial infarction) and non-cardiac conditions (e.g. diabetes mellitus and metabolic causes (e.g. hypokalaemia)). Such diversity and complexity in the acquired form of LQTS, which is by far the most common cause of life-threating arrhythmia attacks in hospital settings, needs to be carefully considered and to have value in clinical practice the algorithm must be optimised to deal with this.  We will collaborate with the Clinical Data Science Unit at Manchester University NHS Foundation Trust to test the algorithm on all ECG data from MUSE MFT since its digitization in September 2022. We will use other data sources to optimise the algorithm including the UK-Biobank database.   

Following this validation and optimization, the next stage will be T1 (testing with patients in Phase 1 clinical trial).  This will trial an app that can take the live ECG feed from MUSE MFT database and provide an interpretation to the clinician (alongside the baseline of the current automated calculation and presentation of ECG data).  

HYPO-CHEAT: prospective data-driven hypoglycaemia reduction in the rare disease of Congenital Hyperinsulinism.

Collaborators: Prof Indi Banerjee, Prof Simon Harper, Dr Paul Nutter.

Congenital Hyperinsulinism (CHI) is the commonest cause of severe and recurrent low blood sugars (hypoglycaemia) in children. The severe and frequent hypoglycaemia experienced by children with CHI can lead to brain injury and developmental problems. Unfortunately, medical management of CHI is imperfect and patients and families have to rely on constant surveillance of their blood sugar levels in order to predict and prevent hypoglycaemia with food and medications. 

Current blood sugar monitoring is done by parents pricking children’s fingers multiple times a day to take a reading. However, this risks missing hypoglycaemia between tests and also offers no trend or prediction information. Continuous glucose monitoring (CGM) is an attractive alternative whereby a device attached to the skin provides 5-minutely values of glucose without pricking the skin. However, each measurement is prone to error and burden can be high for some families. 

We designed and tested an algorithm (HYPO-CHEAT) to use CGM data from children with CHI to spot weekly patterns of hypoglycaemia that are likely influenced by repetitive behaviours. Our algorithm highlights these areas for families and suggests areas of focus so they can change behaviour to prevent hypoglycaemia. Early tests of HYPO-CHEAT showed a reduction in the potentially damaging hypoglycaemia seen in these children. This grant will allow us to develop HYPO-CHEAT into a piece of software that clinicians and patients can use. We will trial this software with a small number of families to gather initial feedback while planning a bigger trial in the future.  

Biodistribution of Haematopoietic Stem Cell Gene Therapy in MPSI and MPSII Mice.

Collaborators: Dr Shaun Wood, Dr Stuart Ellison.

Mucopolysaccharidosis type I and type II (MPSI and MPSII respectively) are rare, inherited lysosomal diseases caused by mutations in enzymes breaking down long chain sugars. The resulting build-up of complex sugars causes problems including heart and lung disease, eye disease and, in severe forms, rapid progressive brain disease.  

Enzyme replacement therapy can alleviate most somatic problems, however brain correction is absent. Haematopoietic stem cell transplant offers a partial solution, as monocytes from the engrafted cells can traffic into the brain and release enzyme beyond the blood brain barrier, thus correcting brain cells. However, the levels of enzyme expressed from unaffected haematopoietic cells are relatively low, leading to poor correction especially in MPSII.  

Haematopoietic stem cell gene therapy (HSCGT) using a lentiviral vector to overexpress the missing enzyme in bone marrow cells from a patient with disease, has shown the ability to correct neurological disease in both mice and patients with at least MPSI, and partly with MPSII, but there is a significant disparity in the distribution of the therapy in the body with somatic organs receiving 50-100 times more enzyme than the brain.  

Redistributing this over-expressed somatic enzyme into the brain, during HSCGT could address the significant unmet need represented by severe MPSI and II. To do this, we have modified the missing lysosomal enzyme with a tag to improve uptake into neuronal cells.

In this study we will study the distribution of this new therapy throughout the body in multiple organs of mice treated with these therapies and assess its safety in preparation for a future clinical trial in patients. 

Validation and deployment of on-line monitoring of body composition of patients being treated with radiotherapy for head and neck cancer.

Collaborators: Dr Dónal McSweeney, Dr Eliana Vasquez Osorio, Dr James Price, Dr Cynthia Eccles, Dr Andrew McPartlin, Dr Tony Tadic.

Head and neck cancer is the 7th most common globally and, in the UK, there are 12,000 new cases each year. Treatment options include a combination of surgery, chemotherapy, and radiotherapy, with each carrying risks of side-effects. Radiotherapy targets high-energy x-rays at the tumour to kill the cancer cells, and is used in over half of cases. Radiotherapy is typically given over 6 weeks, with one of the most common side-effects being weight loss. Weight loss can be caused by difficulties in eating and swallowing due to treatment or directly by the cancer. Currently, patients are given nutritional support if they lose too much weight during treatment. However, this does not consider how they are losing weight, either from the loss of fat, loss of muscle, or both. In Manchester, we have found a group of patients where weight is stable during treatment, but who experience a large amount of muscle loss during radiotherapy. These patients have worse survival than other patients.  

This study proposes to use artificial intelligence (AI) for continuous monitoring of body composition during radiotherapy. We will work with Princess Margaret Hospital in Toronto, the largest cancer centre in Canada. We will validate our findings on their patients to show that patients who lose muscle mass during treatment have worse survival after radiotherapy. By validating our results in a second centre, with a different group of patients, this will strengthen our results. We will also place our AI solution directly in the clinical workflows to show we can provide information to the clinical teams quickly. This will allow interventions to be made to improve treatment for head and neck cancer patients in future studies.  

Novel use of bedside arm ergometry CardioPulmonary Exercise Tests for the riskstratification of patients with chronic limbthreatening ischaemia: A Feasibility Trial.

Collaborators: Professor Frank Bowling, Miss Stacie Hodge, Mr Liam Bagley.

Chronic limb-threatening ischaemia (CLTI) is the most severe clinical manifestation of peripheral arterial disease. It is a major cause of chronic pain, lower limb amputation and death. With >200 million people living with peripheral arterial disease in 2010 and a 23.5% increase since 2000, this is a growing global healthcare problem.

When diagnosed, urgent surgical management is critical, however given the high prevalence of medical conditions such as hypertension, ischaemic heart disease and diabetes amongst this cohort of patients, it’s unsurprising that surgery is associated with an alarmingly high risk of morbidity and mortality when compared to other types of surgery.

With almost half of the patients presenting as an emergency, preoperative risk assessment and stratification is challenging and currently there is no established method. Accurate preoperative risk assessment is of paramount importance to inform clinical decision making, reduce risk and improve individual patient outcomes.

Cardiopulmonary exercise testing (CPET) is a non-invasive clinical measurement tool established in risk stratification of patients undergoing elective surgery. It is used to assess a person’s exercise capacity and look at the response of the heart and lungs to physical stress. However traditional CPET requires the patient to cycle whilst on a fixed bicycle, thus limiting is applicability to patients unable to pedal, such as those with CLTI. Its current use is also almost exclusively used in the outpatient setting, largely due to the fixed and large size of the testing equipment.

In this Confidence for Translation proposal we aim to prove that CPET using an arm ergometer, rather than the traditional bicycle, is a safe, feasible and acceptable tool to use at the bedside in patients with CLTI undergoing emergency surgery. With future work demonstrating that values obtained from this testing can be used to predict post operative outcomes including major adverse cardiovascular events and mortality. 

Raman Spectroscopy and machine learning: A rapid and accurate diagnosis for meningitis

Collaborators: Derren Heyes

Meningitis can be a fatal disease, responsible for 236,000 deaths globally (WHO), if not diagnosed and treated rapidly for the specific infection. The highest incidence is in infants and accounts for the high death rates of infants in developing countries. Our project aims to create a novel diagnostic tool for meningitis, utilizing vibrational spectroscopy and machine learning. The device is targeted for point-of-care usage, particularly important in low-income countries and rural areas like the meningitis belt of sub-Saharan Africa. It will rapidly and accurately identify the type of meningitis, be it bacterial, viral, fungal, parasitic or amebic, for rapid and specific treatment. The technology significantly lowers the sample volume needed for meningitis diagnosis to just 10μl of cerebrospinal fluid, making the process less invasive for patients. It also eliminates the need for extensive lab infrastructure and perishable reagents. Thanks to its integrated AI-based diagnostic system, the device will reduce the reliance on specialized laboratory personnel. This makes it particularly suitable for low-income and remote areas, often vulnerable to epidemics, by offering a practical and efficient solution for managing meningitis in resource-limited settings. A key benefit of this device will be its speed, delivering results 24-72 times faster than the current gold standard methods in meningitis diagnostics. This swift turnaround time is crucial in managing and treating meningitis effectively. By providing prompt and accurate diagnoses, our tool has the potential to significantly improve treatment outcomes and decrease the incidence of mortality and morbidity, especially in regions where meningitis is most prevalent.  

Developing a liquid biopsy to detect minimal residual disease in women diagnosed with highgrade serous ovarian cancer.

Collaborators: Professor Stephen Taylor, Professor Gordon Jayson, Dr George Burghel, Mrs Catherine Ramsden.

In the UK, ovarian cancer leads to more deaths each year than any other gynaecological cancer. The most common type of ovarian cancer is high-grade serous carcinoma (HGSC), which accounts for between 60 and 70% of cases. HGSC is frequently treated with surgery followed by chemotherapy and targeted medicines, such as PARP-1/2 inhibitors and/or bevacizumab. For most women, these treatments are successful, leading to a period of remission, following first-line therapy, which may last for many months and/or a short number of years. During this period of remission, microscopic HGSC cells remain undetectable using standard blood tests (e.g., CA 125) and CT scans. However, sadly, for most women, remission is not permanent and recurrent disease occurs. Once HGSC has relapsed, the chance of long-term cure is very small.  

Currently, there is no blood test or scan that can detect microscopic HGSC cells during follow-up after first-line therapy. The fact that most women diagnosed with HGSC developed recurrent disease, despite cancer cells being undetectable using standard blood tests and scans, shows that microscopic disease exists. This microscopic disease is called ‘minimal residual disease’ (MRD). If women with MRD can be identified early-on during follow-up, new treatment strategies could be developed for these women to help prolong remission and/or get rid of the disease completely.  

We are developing a new blood test to detect molecules, called nucleic acids (called DNA), that are released into the blood stream from microscopic HGSC cells. We aim to use this test to identify women diagnosed with HGSC who have MRD at the end of first-line therapy. This new blood test detects defects (called mutations) in a gene called TP53. Defects in the gene TP53 occur in almost all cases of HGSC. In the C4T project, we will further develop the methods used in this TP53 MRD blood test using fresh samples donated from women treated for HGSC at The Christie Hospital. 

Assessment of thrombolytic efficacy of caADAMTS13 in ex vivo human stroke thrombi.

Collaborators: Professor Craig Smith, Dr Kieron South, Dr Josephine Thomas.

Ischaemic stroke is caused by a blood clot in the brain that leads to a reduced blood supply and the death of brain cells. It is a leading cause of disability and death worldwide. Current treatment involves giving a drug (called t-PA) shortly (within hours) after the stroke that leads to breakdown of the blood clot, allowing blood flow to return to the brain. While t-PA can provide benefit to some patients and lead to better survival and recovery from the stroke, it does not always work. This is in part because blood clots are not all the same and can be composed of different cells (e.g. white blood cells such as platelets and neutrophils) and chemicals. Certain types of blood clot are not broken up by t-PA treatment, meaning that patients do not get blood flow back to their brain and don’t recover well from their stroke.

We have developed a new drug that we think can break down blood clots that t-PA is unable to. This drug is known as caADAMTS13. It is an enzyme that breaks down a protein in the blood called Von Willebrand Factor, which is important in blood clotting. By creating blood clots in mice that are like those observed in humans we have been able to show that caADAMTS13 restores blood flow to the brain better than t-PA. We now want to test if the caADAMTS13 can break down human blood clots in the same way. To do this we will obtain from stroke patients the blood clots that are responsible for causing stroke. These clots are obtained by a process known as endovascular thrombectomy. This is a procedure used in some stroke patients where the clot blocking an artery in the brain is physically removed. We will obtain these clots and determine if they are broken down more efficiently by caADAMTS13 compared to t-PA. If this is the case, then we will be in a good position to further develop caADAMTS13 as a potential replacement for t-PA in the treatment of ischaemic stroke.