Detecting Antibiotic Resistance Proteins in Clinical Samples Using Proteomics

Approximately 40,000 people die in the UK every year as a result of Sepsis, which is a medical condition usually triggered by the body's reaction to bacteria in the blood. When bacteria are present in the blood this is called bacteraemia. The bacteria can come from all sorts of places and can be of many different species. So a diagnosis of Sepsis doesn't tell a doctor what bacterium is responsible. Antibiotics kill bacteria, and so antibiotic therapy is absolutely critical to treating Sepsis. Without removing the underlying cause - the bacteraemia - treatment of Sepsis is unlikely to succeed. The dilemma that doctors face is that because they don't know the identity of the bacterium they want to kill, they are not certain what antibiotics to use. The rise of antibiotic resistance in bacteria makes this choice even more difficult. One way of dealing with this is to use "empiric therapy": to try a particular antibiotic, wait to see if the patient improves and if they don't, try another. But in the meantime, the patient may be getting more and more ill. The alternative approach is that the doctor may start treatment with the latest, most broad acting antibiotic they can find to give them the best chance of killing the bacteria. This means that this "last resort" drug might have been used when it wasn't really needed. Inappropriate use of a last resort drug is the primary driver for antibiotic resistance and will inevitably shorten its useful life.

What we really need is to give doctors information about the identity of the bacterium infecting a patient's blood and, more importantly, what antibiotics it is susceptible to. Then they can make informed antibiotic prescribing choices. At the moment, from the time a blood sample is taken from a patient where bacteraemia is suspected it can take 48 hours just to prove there are any bacteria present. Using new MALDI-TOF machines it is possible to identify the bacterium a few hours later, but that doesn't tell you anything about antibiotic susceptibility. It may take another 24 h to find out what antibiotics can be used. This means that patients can be on the wrong antibiotic for up to 72 hours. If that's a non-effective antibiotic, the patient's life is in danger, if it is an inappropriately used last resort antibiotic, the antibiotic's useful life is being shortened.

Everyone agrees that reducing the time it takes to get antibiotic susceptibility data to doctors is the key, not just for the treatment of patients, but also to better protect our dwindling supply of useful antibiotics. We feel that it may be possible to achieve this by identifying antibiotic resistance proteins - the tools bacteria employ to resist antibiotics - directly in bacteria isolated from patients' blood. If a particular resistance protein is present, the doctor would know not to use a particular drug. To test our hypothesis we want to test whether we can identify resistance proteins in bacteria in blood samples that have been cultured and processed exactly as they would be in hospital diagnostic labs. We will find out whether it is possible to use existing MALDI-TOF machines to identify at least some antibiotic resistance proteins 24 h earlier than is currently the case. We will also test whether it is possible to use more specialised LC-MS/MS machines to reduce the time to get antibiotic sensitivity data by up to 60 hours, giving a positive indication of antibiotic susceptibility about 12-15 h after sampling. It is not necessary to provide a diagnostic test that works minutes after sampling to have real clinical benefit. For severe Sepsis, each hour without working antibiotics gives a 6% increase in patient mortality, so even shaving tens of hours off the current minimum time it takes to predict antibiotic susceptibility would transform patient care.

Currently it takes 18-48 h to confirm clinical bacteraemia and 2-3 h to identify the bacterium responsible using MALDI-TOF after processing the blood culture. Determining antibiotic susceptibility requires additional plating techniques, taking another 24 h. Using MALDI-TOF to identify antimicrobial drug resistance (AMR) proteins would allow clinicians to exclude certain antibiotics up to 24 h earlier than they can now, benefiting patients and prudent antibiotic use. Preliminary data lead us to hypothesise that AMR proteins are highly abundant, and so may be identifiable using the MALDI-TOF machines currently available in diagnostic microbiology labs. Objective 1 is to move towards MALDI-TOF identification of AMR proteins in culture positive blood samples. We shall (1A) create a set of transformants of key pathogens expressing key AMR genes at levels typical of clinical strains. (1B) Quantify AMR proteins in these transformants relative to ribosomal proteins using LC-MS/MS. (1C) Identify unique MALDI-TOF peaks representing these AMR proteins using protocols minimally adapted from those currently used in diagnostic microbiology labs to identify bacteria. (1D) Test whether the MALDI-TOF methodology developed in 1C can predict antibiotic susceptibility in 50 well characterised clinical isolates. (1E) Test whether antibiotic susceptibility can be predicted in 50 processed culture positive blood samples from a clinical diagnostic lab. Objective 2 is to move towards the use of LC-MS/MS to identify AMR proteins in blood cultures. This would allow a wider range of possible AMR proteins to be identified than MALDI-TOF, and it might be performed on blood cultures before they currently flag positive. To do this we shall (2) test whether AMR proteins can be identified using LC-MS/MS in 10 bloodstream isolates following growth in human blood for 12 h (i.e. less than the expected time of culture positivity) with starting inocula representative of clinical bacteraemia.

In addition to the Academic Beneficiaries, outlined separately, we have identified the following groups who will benefit from the research and in the following ways.

The UK Population
Sepsis is one of the most common life-threatening medical conditions seen in the UK. Around 100,000 patients suffer from this condition each year. Many will die - around 40,000 people per year - but many more will recover and the rate and extent of their recovery will affect their futures. There has recently been significant publicity of this problem, and the advocacy group the UK Sepsis Trust, have publicised the "Sepsis Six" steps to treatment. The research project presented in this application will address steps 2 and 3 (take blood cultures, give broad spectrum antibiotics). The use of even a broad spectrum antibiotic may be inappropriate if the target organism is resistant, but it is not possible for a clinician to know this without antibiotic susceptibility data. Successful completion of this project will pave the way to use proteomics to shorten (by tens of hours) the time it takes to obtain antibiotic susceptibility data for blood stream isolates. Given the estimation that one hour without a working antibiotic increases mortality by 6% for severe sepsis, an advance such as this will potentially save hundreds of lives per year. Moreover, it will reduce time to cure, which will reduce morbidity and long term complications. All of which will directly benefit the UK population.

The UK Health Service and Wider Society
According to NHS estimates (http://www.england.nhs.uk/wp-content/uploads/2013/12/spesis-brief.pdf) the cost to the NHS of acute sepsis care is around £2.5 billion per year. The more seriously ill a patient gets and the longer they take to get better, the greater the cost. The more long term complications they have - a result of prolonged and more serious disease - the more long term cost there will be for outpatient and GP care, social care, and potential reductions in the working lifetime of the patient. Any intervention that will reduce the time it takes to prescribe a working antibiotic will shorten time to cure, reduce disease severity and will ultimately cost society less money. The step changes that might result from completion of this project could considerably reduce the cost of Sepsis to society.

Industry/UK Economic Activity
Sepsis is a global problem and the UK population is not unusually burdened by it. If the UK government, through its research councils, spearheads research such as this, then there is potential for the UK to directly benefit economically from the research, development and sale of patented diagnostic equipment and test kits to healthcare systems from other countries. There is also potential for job creation in the UK and economic flow from Europe, since at least one European company that currently makes diagnostic microbiology proteomics equipment (Bruker) already has R&D facilities in the UK.

Protection of the Global Antibiotic Resources
Antimicrobial drug resistance is a massive global problem, and a real threat to humanity. Not just because of its impact on the ability to treat infectious diseases in humans, but also because of its impact on farming. There is a dearth of new antibiotics in development, and recent initiatives to stimulate this will take many years to come to fruition. In the meantime, we need to protect our existing supply of antibiotics as a global resource. This means using narrow spectrum drugs when they will suffice to protect broad spectrum, last resort drugs from the inevitable rise of resistance that comes from their inappropriate use. Obtaining reliable antibiotic susceptibility data is the keystone to appropriate antibiotic prescribing. Completion of this project will provide a means to obtaining susceptibility data sooner, allowing a switch to appropriate therapy earlier than is currently possible. This will protect existing antibiotics.