New technology trends: mass spectrometry

In order to diagnose disease, physicians rely on clinical laboratories to deliver rapid, robust, sensitive and specific results on a vast number of analytes, with more than 70% of the diagnoses in the NHS relying on pathology testing.

The techniques to measure these compounds are typically based on colorimetric or immunoassay techniques, which are prone to a number of interferences, including heterophilic antibodies, cross reactivity with similar compounds and the hook effect, which in some cases may lead to a delayed or incorrect diagnosis. Consequently there has been a drive by physicians and scientists alike to develop tests that minimise or eliminate these interferences and therefore lead to improved patient care. This, coupled with the reduced cost and increased robustness of the technology, has seen the use of liquid chromatography-tandem mass spectrometry (LC-MS/MS) increase massively in routine clinical diagnostics. The high speed, sensitivity and specificity of analysis, combined with the relative ease of sample preparation and ability to multiplex analysis, has made LC-MS/MS the method of choice for a number of routine applications. These include steroid analysis (for example, 25-hydroxyvitamin D and testosterone), newborn screening, immunosuppressive monitoring (for example, ciclosporin and tacrolimus) and drugs of abuse screening. The high demand for this type of analysis has led to a number of manufacturers to produce in vitro diagnostic (IVD) MS analysers and reagent kits to allow laboratories with little or no experience of LC-MS/MS to offer these assays without having to develop and validate these tests in-house.

So what next for the use of LC-MS/MS in clinical diagnostics? Excluding the large amount of promising biomarker data being generated by metabolomic and proteomic studies, there are four areas in which clinical laboratory professionals are currently focusing on. These are:

1. Methods and technologies to increase throughput for routine LC-MS/MS analysis
2. Sensitive routine quantification of small molecules at concentrations previously out of reach for LC-MS/MS due to increasing sensitivity and selectivity of LC-MS/MS systems and associated technologies
3. The use of high resolution mass spectrometers for the screening of drugs and other compounds
4. The analysis of proteins by LC-MS/MS

With healthcare services across the UK facing increased pressure to reduce costs and improve efficiency, recent reviews of laboratory services in England have estimated that around 20% of the pathology budget could be saved, despite an average annual increase in workload of 8–10%.¹ This has led to increasing strain on specialised services such as LC-MS/MS centres, with pressure for laboratories to move back towards inferior immunoassay-based techniques due to the increased throughput these techniques can offer. As an example, our laboratory has seen an exponential increase in 25-hydroxyvitamin D testing from around 5000 tests per year in 2008 to more than 90,000 in 2014. Clearly, workloads of this size are impossible to deliver using conventional manual preparation techniques (for example, liquid-liquid extraction) and a single LC-MS/MS system. One solution to this is to maximise the use of MS/MS system time by using multiple HPLC systems to feed a single mass spectrometer, with applications showing analysis time for complex amino acids halved.² Manual preparation of samples can be avoided by a number of automated sample preparation techniques such as online solid phase extraction³ or automated liquid handling.⁴ However, such techniques can prove to be expensive and may not appear cost effective when compared to automated immunoassay. Our laboratory’s solution to this has been to utilise a simple protein precipitation step performed by a sample handling robot, with the supernatant subjected to a two-step chromatographic separation prior to MS/MS analysis. This consists of a trap-and-elute preparative step using a C-18 online extraction column followed by analytical separation by 2.6µm 30 x 3mm C-18 column. Total sample run time is four minutes. High analytical sensitivity is achieved (2nmol/l) with excellent column life (>7000 injections) and very few samples rejected. To further minimise hands-on staff time and reduce transcription errors, all sample barcodes are automatically scanned at the start of analysis with results reported straight back into our LIMS from the LC-MS/MS software using software developed by C.Sols. The difference in service provision can be seen in Figure 1.

Figure 1. High throughput sample preparation for the analysis of 25-hydroxyvitamin D. Due to increases in requests for the measurement of 25-hydroxyvitamin D, the original sample preparation in use at BHH (left) was too complicated and lengthy (one plate took around 4–5 hours) to cope with the demand. Switching to a ‘trap-and-elute’ online extraction system meant that sample preparation could be reduced considerably (one plate takes around 45 minutes) and throughput increased. Column life, sample rejection rate and instrument downtime have vastly improved with the online extraction method.
Image used with kind permission of Joanne Duffy, Birmingham Heartlands Hospital.

With a number of small molecules now analysed rapidly and routinely by LC-MS/MS, what about the more specialised, low throughput investigations required in clinical biochemistry? The determination of certain endogenous steroid hormones in human serum illustrates a common analytical problem. These hormones are present at low circulating concentrations, with some in the pmol/l range, are prone to matrix interferences and are also structurally very similar to each other, with an order of magnitude concentration difference sometimes present between isobaric compounds. Sensitive detection of these hormones have historically been achieved using radioimmunoassay (RIA) based methodologies with this technology still used in many centres for analytes such as aldosterone and 1,25-dihydroxyvitamin D. Like 25-hydroxyvitamin D, the logical solution to this problem would be to use LC-MS/MS. However, until recently, MS/MS systems lacked the requisite sensitivity to reliably quantify these hormones without the use of a large amount of sample (that is, 1ml of serum) and extensive sample preparation not suitable for routine clinical diagnostics. The latest generation of highly sensitive triple quadrupole mass spectrometers have now permitted these low level steroids to now be measured by LC-MS/MS. Our laboratory has recently moved from a RIA method to a LC-MS/MS method which utilises 200µl of sample, a simple liquid-liquid extraction and a SCIEX 6500 system. This method has a low limit of quantification (69pmol/l, Figure 2) and a total analytical imprecision within the aldosterone reference range (12.0% at 354pmol/l) which are both equivalent to the RIA methodology. The switch to LC-MS/MS methodology has reduced the cost of the assay since specialised radiation precautions are no longer required and the staff time needed to perform the analysis has decreased from two days to half a day, freeing up staff to work on other tests, hence improving patient care.

Figure 2. Highly sensitive aldosterone analysis by LC-MS/MS. The chromatogram on the left (A) is the lowest aldosterone standard, with an assigned value of 69pmol/l. The internal standard, aldosterone-d7 is shown on the right hand side (B). Aldosterone retention time is 4.57 minutes.
Image used with kind permission of Briony Johnson, Birmingham Heartlands Hospital.

Further improvements in sensitivity and selectivity have been achieved by the addition of ion mobility spectrometers to the source of the MS/MS system (for example, the SelexION system).  This allows chemical background and chromatographic interferences to be decreased, thereby increasing the signal/noise ratio of target compounds even when simple sample preparation and fast chromatographic run times are used. Figure 3 demonstrates the improvement that this technique can achieve.

Figure 3. Improvement in analysis of steroids via the use of Differential Mobility Spectrometry (DMS). This image shows an example of the analysis of testosterone in a female subject, using liquid-liquid extraction as the method of sample preparation. Note the improvement in separation when the DMS is turned on.
Used with kind permission of SCIEX.⁵

So advances in throughput and sensitivity had led to improved analysis of steroids and other small molecules in the field of clinical biochemistry. What about clinical toxicology? Certainly the major challenge in the detection of drugs over the past 10 years has been the large increase in novel psychoactive substances (NPS) or ‘legal highs’. Since 2005, the number of new psychoactive substances reported to the Early Warning system in Europe has been increasing year on year with the number now exceeding 80 in 2013 alone, with the majority of these new drugs belonging to the phenethylamine or cannabinoid class.⁶ This has represented a great challenge for toxicology laboratories as these drugs will not be detected by conventional immunoassay-based screens and will also be missed using targeted LC-MS/MS drug screening assays, which have replaced immunoassay and GC-MS in many centres. Broad spectrum ‘unknown’ drug screens will also have difficulty in detecting these compounds as they may be at concentrations below the limit of detection for some older methodologies such as GC-MS and HPLC-DAD, with commercial (for example, NIST) and in-house libraries struggling to cope with the large number of new compounds on the market. A solution to this problem is the use of high resolution mass spectrometers such as quadrupole time-of-flight or orbitrap enabled instruments. These combine the high sensitivity and easy sample preparation of a traditional triple quadrupole instrument with the broad spectrum screening capabilities of a GC-MS instrument. Crucially however, the high resolution accurate mass measurements, together with structural information obtained via MS/MS fragmentation with subsequent comparison to online databases (for example, ChemSpider) allow the potential for unknown peaks on chromatograms to be identified without the need for commercial libraries or certified reference material. Also, since all data is acquired, all of the time, these systems allow retrospective interrogation without the need to re-acquire data. Despite the clear advantages that these techniques offer to the toxicology community, CRM must still be used for forensic level confirmation and quantitative analysis.

It has previously been mentioned that mass spectrometry has improved on immunoassay for the measurement of small molecules, with some laboratories now using LC-MS/MS for the routine analysis of steroid hormones, which have been routinely available on automated platforms for the past 15 years.⁷ So can LC-MS/MS help with the analysis of clinically relevant proteins? These are subject to a number of the same interferences that small molecules are and, like steroid hormones, their measurement is crucial to the diagnosis and monitoring of diseases, with a number of documented cases of unnecessary medical and surgical treatment following false positive results.⁸ So why not use LC-MS/MS for measurement of proteins and peptides? Large proteins such as thyroglobulin are outside the mass range of conventional LC-MS/MS systems, which typically stop around 2000Da. This remains true even if multiply charged forms of the intact protein are used. Therefore an additional step must be introduced into the workflow for measurement of proteins by LC-MS/MS: digestion of the protein by trypsin, which yields peptides that are small enough to analyse by LC-MS/MS. This process contains a number of steps (denaturation, reduction, alkylation and digestion), which must be tightly controlled in order to yield reliable results, with heavy labelled proteins or ‘winged’ labelled peptides ideally used as internal standards, increasing the cost and complexity of workflow when compared to LC-MS/MS analysis of small molecules. With 250,000 different peptides produced via a tryptic digest of serum, it is essesntial to use computational programmes to enable laboratories to select the most specific peptides for the target protein. This complexity alone may deter some laboratories but until recently the more important issue has been sensitivity. Although LC-MS/MS assays for proteins are quite specific, they have lacked sensitivity when compared to immunoassays, with measurement limited to a few abundant proteins above around 1mg/l (for example, albumin, apolipoproteins and caeruloplasmin). This is because of the large degree of ion suppression due to co-elution of a large number of peptides from the tryptic digest, limiting the detection of low abundance proteins. Over the past five years solutions to these problems have made analysis of clinically relevant proteins by LC-MS/MS a reality. The complexity of peptide selection has been made easier with the development of easy, free to use software such as Skyline with digestion workflows controlled by automated liquid handling instruments. The sensitivity of these assays has been overcome with the use of immunoaffinity enrichment, in which a column containing antibodies for the target peptides is used post-digestion to concentrate up target peptides, which are then eluted and analysed by LC-MS/MS. This enriches target peptides, reduces non-target peptides, thereby reducing ion suppression and greatly improving the sensitivity of the method. A conservative estimate is that this process can enrich peptides by 1000-fold.⁹ This technique has been used for the analysis of parathyroid hormone (PTH),¹⁰ thyroglobulin,⁹ various forms of human chorionic gonadotropin¹¹ and vitamin D binding protein¹²  with thyroglobulin LC-MS/MS analysis in routine use in some US laboratories.

Technologies to increase throughput for LC-MS/MS analysis, together with more sensitive, specific and accurate mass spectrometers are allowing laboratories to screen and measure more clinically relevant compounds than ever before, with more certainty that ever before thereby providing a greater degree of clinical diagnostic support and improving patient management

References

1. Fryer et al. Managing demand for laboratory tests: a laboratory toolkit. J Clin Pathol 2013;66:62–72.
2. Taylor. Increased Throughput for Analysis of Physiological Amino Acids Using Multiplexed LC/MS/MS. AB SCIEX Application Note 1930211-01, 2011.
3. Van Faassen et al. An LC-MS/MS Clinical Research Method for the Measurement of 25-OH Vitamin D2 and D3 Metabolites. Waters Technology Brief, 2015.
4. Couchman et al. Automated, High-Throughput LC-MS/MS Workflow for the Analysis of 25-Hydroxyvitamin D2/3 and 3-epi-25-Hydroxyvitamin D3. Thermo Application Note 584, 2013.
5. Liu et al. Highly Selective Analysis of Steroid Biomarkers using SelexION™ Ion Mobility Technology. AB SCIEX Application Note 6740112-01, 2012.
6. New Psychoactive Substances Review, Report of the Expert Panel. September 2014.
7. Owen et al. Development of a rapid assay for the analysis of serum cortisol and its implementation into a routine service laboratory. Ann Clin Biochem. 2013;50(4):345–52.
8. Cole et al. False-Positive hCG Assay Results Leading to Unnecessary Surgery and Chemotherapy and Needless Occurrences of Diabetes and Coma. Clin Chem. 1999;45(2) 313–4.
9. Hoofnagle, Roth. Clinical review: improving the measurement of serum thyroglobulin with mass spectrometry. J Clin Endocrinol Metab. 2013;98(4):1343–52.
10. Lopez et al. Selected Reaction Monitoring–Mass Spectrometric Immunoassay Responsive to Parathyroid Hormone and Related Variants. Clin Chem 2010;56(2):281–90.
11. Woldemariam, Butch. Immunoextraction-tandem mass spectrometry method for measuring intact human chorionic gonadotropin, free ?-subunit, and ?-subunit core fragment in urine. Clin Chem 2014;60(8):1089–97.
12. Hoofnagle et al. Vitamin D–Binding Protein Concentrations Quantified by Mass Spectrometry. N Engl J Med 2015;373:1480–2.