The measurement of KRAS G12 mutants using multiplexed selected reaction monitoring and ion mobility mass spectrometry

Rationale There is a considerable clinical demand to determine key mutations in genes involved with cancer which necessitates the deployment of highly specific and robust analytical methods. Multiplex liquid chromatography with selected reaction monitoring (LC/SRM) assays offer the ability to achieve quantitation down to levels expected to be present in clinical samples. Ion mobility mass spectrometry (IMS/MS) assays can provide increased peak capacity and hence separation in an extremely short time frame, and in addition provide physicochemical data regarding the collision cross‐section of an analyte which can be used in conjunction with the m/z value of an ion to increase detection specificity. Methods For LC/SRM, unlabelled peptides and corresponding stable‐isotope‐labelled standards were spiked into digested human plasma and analysed using ultrahigh‐performance liquid chromatography (UHPLC) coupled to a triple quadrupole mass spectrometer to enable the generation of analyte‐specific calibration lines. Synthetic unlabelled peptides were infused into a Synapt G2 mass spectrometer for travelling wave ion mobility separation and TWCCSN2 values were derived from comparison with previously generated TWCCSN2 calibration values. Results Linear calibration lines (0.125 to 25 fmol/μL) were established for each of the KRAS peptides. UHPLC separated the peptides and hence enabled them to be split into different retention time functions/windows. This separation enabled detection of three or four transitions for each light and heavy peptide with at least 10 points per peak for accurate quantitation. All six KRAS G12 peptides were separated using IMS/MS, enabling precise TWCCSN2 values to be determined. Although some of the G12 peptides chromatographically co‐eluted, all the peptides were distinguished by m/z, retention time and/or drift time. Conclusions This study advocates that LC/SRM and IMS/MS could both be used to identify single amino acid substitutions in KRAS as an alternative to commonly used methods such as circulating tumour DNA analysis.

Rationale: There is a considerable clinical demand to determine key mutations in genes involved with cancer which necessitates the deployment of highly specific and robust analytical methods. Multiplex liquid chromatography with selected reaction monitoring (LC/SRM) assays offer the ability to achieve quantitation down to levels expected to be present in clinical samples. Ion mobility mass spectrometry (IMS/MS) assays can provide increased peak capacity and hence separation in an extremely short time frame, and in addition provide physicochemical data regarding the collision cross-section of an analyte which can be used in conjunction with the m/z value of an ion to increase detection specificity.
Methods: For LC/SRM, unlabelled peptides and corresponding stable-isotopelabelled standards were spiked into digested human plasma and analysed using ultrahigh-performance liquid chromatography (UHPLC) coupled to a triple quadrupole mass spectrometer to enable the generation of analyte-specific calibration lines. Synthetic unlabelled peptides were infused into a Synapt G2 mass spectrometer for travelling wave ion mobility separation and TW CCS N2 values were derived from comparison with previously generated TW CCS N2 calibration values.
Results: Linear calibration lines (0.125 to 25 fmol/μL) were established for each of the KRAS peptides. UHPLC separated the peptides and hence enabled them to be split into different retention time functions/windows. This separation enabled detection of three or four transitions for each light and heavy peptide with at least 10 points per peak for accurate quantitation. All six KRAS G12 peptides were separated using IMS/MS, enabling precise TW CCS N2 values to be determined.
Although some of the G12 peptides chromatographically co-eluted, all the peptides were distinguished by m/z, retention time and/or drift time.
Conclusions: This study advocates that LC/SRM and IMS/MS could both be used to identify single amino acid substitutions in KRAS as an alternative to commonly used methods such as circulating tumour DNA analysis.

| INTRODUCTION
Targeted liquid chromatography with selected reaction monitoring (LC/SRM) assays have great potential for use in the clinic to detect protein biomarkers for the diagnosis and monitoring of a wide variety of diseases. Despite offering many advantages, significant challenges still exist and a very limited number of LC/SRM-based assays have been adopted in the clinical setting to date. 1 Implementing this technique in clinical laboratories necessitates the development of robust, quantitative assays for clinically relevant biomarkers for which reliable detection methods do not currently exist. 2 In addition, if the assay is unable to provide a financial operation cost advantage, the information derived from the assay has to be of superior quality to that provided by other methods.
KRAS is a protein first identified in the Kirsten rat sarcoma virus 3 and is involved in cell signalling which activates cell growth in response to growth factor binding. It contains an intrinsic GTPase which switches off the pathway; however, single amino acid substitutions can inactivate the GTPase. This leads to constitutive activation of KRAS and aberrant growth. KRAS is implicated in many cancers including colorectal cancer, pancreatic cancer and non-small cell lung cancer. [4][5][6] Amino acid substitutions are driver mutations, occurring early in the carcinogenesis process. 7 Thus, methods that identify KRAS mutations could enable detection of early-stage cancer and subsequently inform the clinician of preferential treatment regimens. KRAS mutations occur at positions 12, 13 and 61 of the KRAS gene with substitutions of the glycine at position 12 the most common. 8,9 KRAS has an extremely low copy number and has not previously been detected in human plasma. Consequently, determination of KRAS status is primarily undertaken using genetic testing. 10 However, genetic testing relies on invasive, costly and timeconsuming (typically around 7 days) methodologies which rely on high-fidelity amplification to achieve genetic determination.
Travelling wave ion mobility separation (TWIMS) enables gasphase electrophoretic separation by creating a dynamic pulse of alternating voltages that results in a "travelling wave" of ions within a dense gas-filled chamber (in this case nitrogen). 11 Ions are separated by their size, shape or charge state. Since mutated peptides differ by a single amino acid and some of them co-elute using reversed-phase LC/SRM, we also investigated the use of Peptides unique to each RAS protein were selected to distinguish between the RAS isoforms, NRAS, HRAS and KRAS, because the tryptic peptide containing the mutation is found in all isoforms.
Stable-isotope-labelled (heavy) standards were incorporated into the LC/SRM assay to enable absolute quantification of the peptides using isotope dilution and multipoint calibration lines.

| Materials
All chemicals and reagents were purchased from Sigma Aldrich (Poole, UK) or Fisher Scientific (Loughborough, UK) unless otherwise stated.

| Data analysis
where t D is the measured drift time of the precursor ion, t 0 D is the corrected drift time of the precursor ion, m/z is the mass-to-charge ratio of the precursor ion and c is the enhanced duty cycle delay coefficient. A corrected DT was determined for BSA peptides and these times were plotted against known Ω 0 N2 values to establish the calibration line from the Bush Laboratory. 16 Values of Ω 0 N2 for KRAS mutant peptide standards were then derived using the coefficients from the calibration line. The corresponding Ω N2 values (absolute TW CCS N2 ) were obtained by factoring in the calculated reduced mass.
IMS analysis for BSA and KRAS standards was ultimately performed using an IMS wave velocity of 1330 m/s.

| RESULTS AND DISCUSSION
A method was required to identify the various KRAS mutations which are commonly found in different tumours including pancreatic, colorectal and lung cancers. Using LC/SRM, the wild-type and five F I G U R E 1 Representative calibration line from wild-type 6-16 unlabelled peptide injected in triplicate and normalized to the heavy ( 13 C 6 , 15 N 2 ) internal standard, and a representative Skyline chromatogram from 25 fmol/μL wild-type unlabelled peptide standard most common mutant peptides were separated using a 60 min LC gradient and accurately quantified using stable-isotope-labelled standards; the RAS isoforms were also detected. A representative chromatogram and a calibration line constructed in digested human plasma are shown in Figure 1, demonstrating the linearity of quantification of the wild-type peptide. The separation of the RAS peptides using LC/SRM is shown in Figure 2. Four discrete retention time windows were used to enable the detection of 3-4 transitions for each peptide without compromising sensitivity. For each peptide the linear range was 0.125 to 25 fmol/μL, ensuring a sensitive assay for KRAS detection. Table 2 Figure 3A shows the separation of the peptides spiked into digested human plasma in a three-dimensional ion mobility spectrum. Due to the infusion aspect of the IMS experiment, we can rapidly detect the different mutations of KRAS. Figure 3B shows the same analysis of digested human plasma without the peptides and it should be noted that the scale is 37.5 times greater in the KRAS peptide mobilogram. The MS spectra and mobilograms thus demonstrate that the ion mobility F I G U R E 2 Chromatograms showing the separation of all nine analytes spiked in digested human plasma (WT, G12S, G12D, G12C, G12A and G12V 6-16 peptides and KRAS-, NRAS-and HRAS-specific peptides) with a 60 min reversed-phase LC gradient using a Waters nanoACQUITY UHPLC system coupled to a Waters Xevo TQ mass spectrometer using LC/SRM T A B L E 2 Calibration data from peptide calibration lines established in human plasma. Lower limit of detection (LLOD) and lower limit of quantitation (LLOQ) values for each peptide when measured using LC/SRM a