My focus has been leukemia research since 1984. I began working with proteomics in 1999, and then moved into biomarker discovery, simply through wanting to do the best possible research for the most clinical impact. We soon started using our proteomics platform outside leukemia – for example, biofluids analysis – to identify markers relevant to cancer. Consequently, we applied for £13 million Medical Research Council (MRC) funding to develop the Stoller Biomarker Discovery Centre, which is now the largest clinical proteomics center in Europe.
It relied on teamwork. First, we pulled together stratified medicine groups with expertise in psoriasis, lupus, arthritis, and cancer – everything starts with clinical need. Then we co-operated to show the MRC that we had the ideal basis for development of a transformational biomarker discovery platform. But it took 4–5 iterations before we got the award, and each one was more difficult than the last. I still find it hard to believe that we succeeded!
It’s the best technology for discovery proteomics, and the instruments get better and more sensitive every year; for example, it’s clear that SWATH-MS markedly increases efficiency. So we concluded that the industrialization of proteomics would be best-served by adopting mass spectrometry (MS) technology. There are other approaches to proteomics – for example, antibody arrays – but we think MS is best for biomarker discovery.
We’ve always sought collaborations to identify disease-associated biomarkers, and we continue to invite collaboration on biofluid and tissue sets from other centres. One collaboration with Tessa Holyaoke at the University of Glasgow has led to a potentially curative strategy for chronic myeloid leukaemia (1) which I hope will soon be in clinical trials. That development could not have happened in either institute alone.
Too long – about 10 years. One problem is that samples aren’t collected early enough in the clinical study, due to issues such as cost, time and availability of research nurses. But another point is that you need a very well-regulated laboratory environment, including a recording regime which passes regulatory scrutiny. At the Stoller Centre, we work to the highest relevant standards as early as possible. This will compress the period between biomarker discovery and clinical application – our target is to bring four biomarkers to the clinic in the next five years. Our quality regime is aimed at meeting international standards – including MHRA, FDA and EMA requirements – and covers MS, antibody-based assays, sample storage, laboratory information management systems and informatics.
No – the rate of biomarker regulatory approvals isn’t changing, perhaps because biomarker development is so hard. You must address issues including statistical surety, sensitivity, specificity – which must be very high to support a biomarker launch. You cannot afford to get it wrong. It’s just not acceptable to have a high false positive or negative rate.
There are obvious clinical and health economic benefits to using a biomarker test to assess responsiveness before or just after initiation of treatment. For example, if you treat patients with expensive biologics, at £10,000/ year, for months before you know whether they respond to it, that’s not good for the health service, the taxpayer or the patient. But biomarkers help to get the right drug to the right patient at the right time. There are good exemplars of this in leukemia – for example, in CML you test for the chromosomal translocation encoding BCR-ABL, which indicates treatment with tyrosine kinase inhibitors.
Informatics tools support digitization of the proteomic map for any individual, and association of the map with an electronic healthcare record. And that allows you to identify false positives and false negatives and develop the test surety. Furthermore, future CDx will not rely on single markers, but will use algorithms to derive information from multiple measured factors, and this too will require informatics capabilities. Applying informatics to proteomic data will allow accommodation of co-morbidities, concomitant medications, and other confounding factors, thus enabling development of algorithms with high sensitivity and surety.
It’s a fantastic asset! We absolutely need pathology expertise to move from the lab to clinical usage as swiftly as possible. That’s why we applied for and obtained another £3 million from the MRC to build the pathology node, led by Tony Freemont, which takes our discoveries and translates them into something of clinical value. This integration of expertise will allow us, ultimately, to improve healthcare outcomes for everyone.
Newsletters
Receive the latest pathology news, personalities, education, and career development – weekly to your inbox.
References
- SA Abraham et al., “Dual targeting of p53 and c-MYC selectively eliminates leukaemic stem cells”, Nature, 8, 534, 341–346 (2016). PMID: 27281222.
