Weighing Viruses

When faced with limited access to coronavirus tests at the height of the pandemic, I was inspired to repurpose my laboratory – and mass spectrometers across the UK – to assist with national testing. Over the past six months I have acted as a Scientific Advisor to the UK Department of Health and Social Care. And now a network of eight academic labs in the UK have joined forces with twelve clinical laboratories based in the National Health Service (NHS), backed by government and industrial funding, to deliver a rapid and sensitive MS test for coronavirus. Here’s how we did it… 

Mobilizing MS
The first key challenge with our pilot was to develop a method that was robust enough to be readily adopted by the NHS. Thankfully, NHS staff already perform more than 750,000 MS-based tests per year on newborn babies alone, along with further tests for diagnosing other metabolic disorders and for therapeutic drug monitoring. We wanted to tap into this world-class capability that had already been established within our labs. 

When the pilot started, we had three clear criteria:

• The method should have clinical utility (in other words, it can be performed in a routine hospital laboratory).
• The assay could be widely adopted (which necessitated viral deactivation upon sampling).
• The lab-based test should be compared with RT-PCR assays in terms of its specificity and sensitivity.

Following a period of consultation with academic groups who had already signed up to the COVID Mass Spectrometry Coalition, eight laboratories with strengths in proteomic analysis and method development were selected to take part – we called these the “P1 labs.” They had to be willing to act, have strong links to hospital laboratories, and have experience with translation. Trials began with recombinant forms of viral proteins establishing limits of detection and quantification with proven methods. The work of Maarten Dhaenens at the University of Ghent, which involved some of those academic labs, was compelling. He had shown that quantitative proteomics performed on LC triple quad platforms was able to detect the presence of peptides from the spike and the nucleocapsidprotein at levels that were commensurate with PCR Ct values of 25 – the stage at which symptoms can be mild or non-existent. Also of interest were the MALDI methods developed by Ray Iles of MapSciences, and two academic labs were chosen accordingly for their expertise with MALDI instruments.

Handling live virus severely limits the type of laboratory and the staff that are available to perform such assays. Because of this, we started thinking about different methods of viral inactivation that were compatible with MS. We considered heat and UV exposure, but only organic solvents were suitable for the deployment of mass testing. Fortunately, robust tests performed at Public Health England’s testing labs in Porton Down helped us decide on the best approach, and volunteers at the Manchester Institute of Biotechnology (MIB) set about filling 30,000 falcon tubes with our deactivating solution. We decided to consider two different methods of sample acquisition. The first method was identical to the one used ahead of RT-PCR – an oral nasopharyngeal swab that could later be added to the deactivation solution. The second, which was also being used elsewhere in test development, collected saliva via a funnel that fitted into the sample tube. Each participant in our pilot would be asked to provide a saliva sample as well as being swabbed twice – one for the MS test and one for a comparator RT-PCR test.

Finding samples

Test packs at the ready and methods being developed, we needed a source of samples. But that proved much harder than we had anticipated. Eventually we found the indefatigable Rick Body – an emergency care consultant who was leading an NIHR research program called FALCON, which had been set up to enable trials of new tests. Rick applied for an ethical amendment to allow samples to be collected for MS, and, in early November 2020, we started to receive samples from patients who had been admitted to hospital with COVID-19-like symptoms. The number of hospital collecting sites grew and again the MIB volunteers, led by the marvelous Kat Hollywood, dug in and made up boxes of swabs, saliva funnels, and collecting tubes, as well as boxes for the hospitals to post samples on to the P1 academic labs. Students at the MIB and staff at Manchester Foundation Trust Hospitals also took part to boost sample numbers. As we neared the end of the P1 process at Christmas, we had received a total of 285 samples by this route; the positivity rate was just over 10 percent. 

Though several of the P1 labs had shown that LC-MS/MS could detect a few 100 attomoles of protein, 285 samples split 8 ways was not enough for the pilot labs to prove the validity of a new test. We were therefore very grateful for the suggestion from Ed Blanford (Head of Test Validation for the Pillar 2 test labs) that he could send the FALCON study coordinators a list of names of people who had tested positive and agreed to provide material for research into new tests. This was a lifeline; again, the MIB team swung into action; more packs were made up and sent out to people at home who had tested positive. With the new variant came a rise in people testing positive and by mid-January we had received enough samples. 

The final transfer
At this stage, we had shown that MALDI was not sensitive enough to compete with the LC-MS/MS methods; in fact, even the LC-MS/MS methods were struggling to reach the technical limits of detection with background from other proteins in the samples. We decided to try a capture method, where digested peptides specific to the nucleocapsid protein bind specifically to antibodies tethered to magnetic beads. The results from this were electrifying. The specificity of the assays shot up which in turn increased sensitivity. A harmonized method and SOP for samples from both saliva and swab was developed. The method, which is enabled by reagents from SISCAPA, can be automated; the inject-to-inject time on the MS is currently only 3 minutes – meaning at least 480 samples can be run in a day on each MS.

Swabs outperformed saliva, with sensitivity and specificity of 97 and 98 percent, respectively, for samples where the Ct value is 27 or less (600 samples). For saliva, the variation in the background protein levels means that “positivity” is not as clear cut, although the sensitivity is again above 80 percent.

The P1 labs have now handed the method over to 12 clinical MS laboratories, which are geographically spread across the UK. These labs will now seek accreditation for the method from UKAS and then be ready to test the population.

The role of MS as we learn to live with coronavirus
Now that an MS-based test has been developed and deployed in the NHS, what can we do with it? This new capacity could be used to monitor levels of infection in regional outbreaks, in hospital inpatients, and for surveillance studies in different populations, for example, at schools or workplaces. The fact that the virus is deactivated directly after sample collection makes the entire process much simpler; samples can be directly transported to labs for a diagnostic screen.

The method is robust to sample pooling (which enables many people to be screened in a single assay), and we are also able to detect mutated regions of the virus from the non-captured material. 

We all hope that the fantastic vaccine rollout will allow most people to have immunity from severe infection, but none of us know what the virus will do next. Therefore, we will continue to need to test people and to monitor vaccine resistance; MS could well have a role to play here. 

The investment into the MS test has enabled the deployment of 12 new mass spectrometers into the NHS for a six-month period. Existing staff are learning how to run this assay and I am sure they will continue to make improvements. The long-term effects of the pandemic will necessitate increased diagnostic capacity for COVID-19, long COVID-19, and many other diseases in a population who have avoided visiting their GPs or had test appointments delayed. The very same instruments and capable staff will be able to meet this increased need for accurate and robust analysis of infection and disease.

This government-funded program is an outstanding real-world example of an MS test being translated from academic labs to clinical laboratories across the country. And we have done it in four months. Necessity is indeed the mother of invention; this pandemic has driven so much collaborative and novel science. The entire process also demanded creativity, perseverance, and a willingness to work together to pool resources – as well as the need to compromise.

As for the future, I am optimistic that this will be the first of many such rapid deployment actions. The unique collaborative links that have been formed between academic and clinical labs, gratefully supported by industry, will continue to grow through this pandemic and for future diagnostic challenges.