Article pubs.acs.org/jpr
Using Mass Spectrometry to Monitor Monoclonal Immunoglobulins in Patients with a Monoclonal Gammopathy David R. Barnidge,† Surendra Dasari,‡ Chad M. Botz,† Danelle H. Murray,† Melissa R. Snyder,† Jerry A. Katzmann,† Angela Dispenzieri,† and David L. Murray*,† †
Department of Laboratory Medicine and Pathology and ‡Biomedical Statistics and Informatics, Mayo Clinic, Rochester, Minnesota 55905, United States S Supporting Information *
ABSTRACT: A monoclonal gammopathy is defined by the detection a monoclonal immunoglobulin (M-protein). In clinical practice, the M-protein is detected by protein gel electrophoresis (PEL) and immunofixation electrophoresis (IFE). We theorized that molecular mass could be used instead of electrophoretic patterns to identify and quantify the M-protein because each light and heavy chain has a unique amino acid sequence and thus a unique molecular mass whose increased concentration could be distinguished from the normal polyclonal background. In addition, we surmised that top-down MS could be used to isotype the M-protein because each immunoglobulin has a constant region with an amino acid sequence unique to each isotype. Our method first enriches serum for immunoglobulins followed by reduction using DTT to separate light chains from heavy chains and then by microflow LC-ESI-Q-TOF MS. The multiply charged light and heavy chain ions are converted to their molecular masses, and reconstructed peak area calculations for light chains are used for quantification. Using this method, we demonstrate how the light chain portion of an M-protein can be monitored by molecular mass, and we also show that in sequential samples from a patient with multiple myeloma the light chain portion of the M-protein was detected in all samples, even those negative by PEL, IFE, and quantitative FLC. We also present top-down MS isotyping of M-protein light chains using a unique isotype-specific fragmentation pattern allowing for quantification and isotype identification in the same run. Our results show that microLC-ESI-Q-TOF MS provides superior sensitivity and specificity compared to conventional methods and shows promise as a viable method of detecting and isotyping an M-protein. KEYWORDS: Monoclonal gammopathy, immunoglobulin, multiple myeloma, mass spectrometry, top-down MS, isotype
■
INTRODUCTION Plasma cell proliferative disorders are characterized by a clonal expansion of plasma cells. These disorders are referred to as monoclonal gammopathy, paraproteinemia, or plasma cell dyscrasia. The plasma cell clones often secrete elevated levels of a monoclonal immunoglobulin clinically referred to as an Mprotein and commonly referred to as a monoclonal antibody (mAb). The M-protein may consist of the intact immunoglobulin in combination with a free light chain that has not paired with a heavy chain (FLC). When there is clinical suspicion of a plasma cell proliferative disorder, a patient is tested for the presence of the M-protein in serum and urine using a combination of agarose gel serum protein electrophoresis (SPEP), immunofixation electrophoresis (IFE), and free light chain (FLC)-specific immunoassays. If the patient exhibits clinical symptoms suggesting multiple myeloma, then a bone marrow biopsy is performed to quantitate the plasma cells.1 The change in the peak area for the M-spike is normally used to monitor a patient’s response to treatment.2 If treatment eliminates M-protein-producing malignant plasma cells, then the M-spike peak observed by SPEP will fade into the normal © 2014 American Chemical Society
polyclonal immunoglobulin background below the limit of quantification (500 mg/L or 3 μM). IFE, FLC, and bone marrow biopsy can then be used to monitor treatment because they have a lower limit of quantification. The role of minimal residual disease (MRD) by higher sensitivity methods is still being investigated. Currently, new multiparameter flow cytometric assays and PCR-based assays are currently being investigated for their ability to determine minimal residual disease (MRD).3−5 Recent publications have pointed out the major role mass spectrometry has played in the quality control of therapeutic monoclonal antibodies (mAb), including the use of top-down mass spectrometry.6−11 We hypothesized that even without knowing the amino acid sequence of the M-protein, as compared to therapeutic mAb quality control, accurate molecular mass could be used to monitor the M-protein above the polyclonal background. We also surmised that it would be possible to determine the isotype of an M-protein Received: September 27, 2013 Published: January 27, 2014 1419
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Journal of Proteome Research
Article
Samples were then acidified with 1 μL of formic acid prior to injection.
light chain or heavy chain using top-down MS provided the fragment ions gave a sequence tag from the isotype-specific constant region. Currently, there are clinical assays that use mass spectrometry to monitor the molecular mass of intact proteins to diagnose various disorders such as hemoglobinopathies,12 familial amyloidosis,13−16 and congenital disorders of glycosylation.17 However, top-down mass spectrometry has had limited implementation in the clinical laboratory18−20 despite advancements in bioinformatics tools21 and innovative instrument design.22,23 Here, we describe the application of a mass spectrometry-based platform routinely used by pharma in mAb quality control, microLC-ESI-Q-TOF MS, for the rapid detection, quantification, and isotyping of monoclonal immunoglobulins from patients with a monoclonal gammopathy. This methodology represents a major shift in the characterization of monoclonal immunoglobulins. Rather than using gel electrophoresis and anti-isotypic antibody reagents, we combine molecular mass and constant region fragment ions to create a novel clinical laboratory technique for detecting and monitoring monoclonal immunoglobulins.
■
LC Conditions
An Eksigent Ekspert 200 microLC (Dublin, CA) was used for separation; mobile phase A was water + 0.1% FA, and mobile phase B was 80% acetonitrile + 10% 2-propanol + 0.1% FA. A 2 μL injection was made onto a 1.0 × 75 mm Poroshell 300SBC3, 5 μm column flowing at 25 μL/min. A 15 min gradient was started at 80% A/20% B, held for 0.5 min, ramped to 70% A/ 30% B over 1 min, then ramped to 60% A/40% B over 4 min, then ramped to 5% A/95% B over 5 min, held for 2.5 min, then ramped to 80% A/20% B over 1 min, and then equilibrated at 80% A/20% B for 1 min. ESI-Q-TOF MS
Spectra were collected on an ABSciex TripleTOF 5600 quadrupole time-of-flight mass spectrometer (ABSciex, Vaughan, ON, Canada) in ESI positive mode with a Turbo V dual-ion source with an automated calibrant delivery system (CDS). Source conditions were IS, 5500; temp, 500; CUR, 45; GS1, 35; GS2, 30; and CE, 50 ± 5. TOF MS scans were acquired from m/z 600−2500 with an acquisition time of 100 ms. Fragment ion scans were acquired from m/z 350−2000 with an acquisition time of 100 ms. The instrument was calibrated every five injections through the CDS using calibration solution supplied by the manufacturer.
EXPERIMENTAL SECTION
Materials
Serum and Urine Samples. Serum and urine were waste samples in the clinical lab. The paired serum/urine samples and banked sequential samples were from a multiple myeloma patient consented under IRB protocol 521-93. Reagents. Ammonium bicarbonate, dithiothreitol (DTT), and formic acid were purchased from Sigma-Aldrich (St. Louis, MO). Water and acetonitrile were purchased from Honeywell Burdick and Jackson (Muskegon, MI). Monoclonal Immunoglobulins. The therapeutic monoclonal immunoglobulin adalimumab (Humira) was purchased from Abbott Laboratories (Chicago, IL). The amino acid sequence for adalimumab was kindly provided by L. J. Dekker.24 A lambda immunoglobulin light chain standard was purchased from Bethyl Laboratories (Montgomery, TX) and was purified from the urine of a patient with multiple myeloma who consented to have their urine used as a source for lambda light chain immunoglobulin.
Data Analysis
Analyst TF v1.6 was used for instrument control. Data were viewed using Analyst TF v1.6 and PeakView v1.2.0.3. Multiply charged ion peak centroids were used to calculate average molecular mass and the peak area value used for quantification through BioAnalyst software provided with Analyst TF. Multiple ion deconvolution was done between 20 000 and 28 000 Da as a protonated molecule, a step size of 1, a S/N of 20, and 20 iterations for light chain molecular mass calculations. Multiple ion deconvolution was done between 45 000 and 60 000 Da for the heavy chain. MultiQuan v2.1 was used to calculate peak areas for standards using multiple extracted ions. Top-down fragmentation spectra for lambda light chains were searched using ProSight software (Thermo-Fisher Scientific, Waltham, MA) against a small protein sequence database containing human immunoglobulin constant chains and common contaminants. Generated sequence tags were matched against the protein database, and significant p-value scores were found for a C-terminal portion of the lambda light chain.
Serum
A volume of 50 μL of serum was enriched for immunoglobulins using Melon Gel (Thermo-Fisher Scientific, Waltham, MA) following the manufacturer’s instructions. The serum sample was diluted 10-fold with sample purification buffer (50 μL of serum + 450 μL of buffer) and was then added to 500 μL of gel slurry in a sample collection microcentrifuge tube containing a frit. The sample was mixed end-over-end for 5 min and then centrifuged with the purified immunoglobulins being collected in the flow through. After immunoglobulin enrichment, 25 μL of sample was reduced by adding 25 μL of 100 mM DTT and 25 μL of 50 mM ammonium bicarbonate followed by incubation at 55 °C for 30 min. Samples were then acidified with 1 μL of formic acid prior to injection.
Protein Electrophoresis
All assays were performed according to protocols in the Clinical Immunology Laboratory, Department of Laboratory Medicine and Pathology, Mayo Clinic. PEL was performed on a SPIFE SPE system (Helena Laboratories, Beaumont, TX), and IFE, on Sebia 9IF gels (Sebia, Norcross, GA). The total protein concentration was determined by colorimetric assay using biuret reagents from Roche and a Roche Hitachi 912 chemistry analyzer system (Roche Diagnostics, Indianapolis, IN).
Urine
Immunoglobulins were analyzed directly from urine without purification. A volume of 25 μL of urine was reduced by adding 25 μL of 100 mM DTT and 25 μL of 50 mM ammonium bicarbonate followed by incubation at 55 °C for 30 min.
PEL, IFE, and FLC
All assays were performed according to protocols in the Clinical Immunology Laboratory and have been previously described.1 1420
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Journal of Proteome Research
Article
Figure 1. Mass spectra from (A) a normal serum sample and (B) normal serum spiked with 0.5 g/dL of the IgG kappa recombinant mAb adalimumab. The normal serum mass spectrum displays a broad range of unresolved peaks, whereas the normal serum spiked with 0.5 g/dL of adalimumab shows a clearly defined multiply charged protein ions. (C) Converted spectrum for the normal serum sample displaying a broad range of unresolved peaks. (D) Converted mass spectrum for the normal serum spiked with 0.5 g/dL of adalimumab showing a single peak at an average molecular mass of 23 412.19 Da. This mass is in excellent agreement with the calculated average molecular mass of 23 412.13 Da for the kappa light chain of adalimumab.
■
RESULTS
broad distribution of masses with no single mass higher in abundance than the background. This is in sharp contrast to the converted molecular mass for the normal serum spiked with 0.5 g/dL of adalimumab in panel D, which displays a single peak with the observed molecular mass of 23 412.19 Da. This molecular mass agrees with the average molecular mass calculated from the known amino acid sequence for the kappa light chain of adalimumab (23 412.13 Da). To assess microLC-ESI-Q-TOF MS for quantification, adalimumab was spiked into 50 mM ammonium bicarbonate buffer, normal serum, and normal urine (see Supporting Information Table 1). Ten different standard concentrations were used ranging from 0.005 to 5.0 g/dL. Standard curves made in serum used Melon Gel to enrich for immunoglobulins, whereas curves made in urine and buffer were reduced and analyzed without Melon Gel purification (see Supporting Information Figure 6 for raw data). Linearity and linear dynamic range values in the table are split according to the two quantification techniques. The first approach labeled “deconvolution peak area” uses the peak area found after deconvolution of the multiply charged ions to molecular mass, whereas the second approach labeled “extracted ion peak area” refers to using the peak areas
Adalimumab in Normal Serum as a Model System
Adalimumab is an anti-TNF therapeutic monoclonal immunoglobulin that is widely prescribed for downregulating the inflammatory response in patients with autoimmune disorders. Therapeutic monoclonal immunoglobulins such as adalimumab are ideal surrogate standards for simulating a monoclonal immunoglobulin in serum because they are readily available in high purity and typically have a large body of literature on their structural properties. Figure 1 shows the mass spectra for normal serum and serum spiked with 0.5 g/dL (30 μM) of adalimumab. Each mass spectrum represents the spectra summed together over the adalimumab light chain elution time (see Supporting Information Figure 1). The mass spectrum from normal serum in section A shows a broad unresolved peak with a maximum relative abundance of 300 counts per second (cps). Alternatively the mass spectrum from the serum spiked with adalimumab in section B shows a distinct series of peaks from multiply charged protein ions with a maximum relative abundance of 6000 cps. The converted molecular mass for the normal serum in panel C shows a set of 1421
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Journal of Proteome Research
Article
Figure 2. Results from the analysis of a serum sample from a patient with multiple myeloma. (A) Mass spectrum clearly showing a set of multiply charged ions that are (B) converted to a molecular mass of 23 452.64 Da, which is within the expected mass range for a light chain. (C) Another set of multiply charged ions that are (D) converted to a molecular mass of 51 596.07 and 51 758.27 Da, which is within the expected mass range for a heavy chain. The difference between the molecular mass of series 2 and series 3 is 162.20 Da, which closely matches the mass of a hexose subunit.
for this series shown in panel D was found to have two components, one at 51 596.07 Da and another at 51 758.27 Da, both of which are consistent with IgG heavy chain. The difference of 162.20 Da between series 2 and 3 may be due to two heavy chain proteoforms differing by the number of hexose units (average MW, 162.14 Da) in the carbohydrate chain.25 Additional samples taken over a 7 year period were available for testing from this patient. Figure 3 shows the result using mass spectrometry for a sample taken after the patient had been treated for multiple myeloma and was found to be negative by PEL, IFE, and the quantitative FLC immunoassay. However, multiply charged ions from the light chain are clearly evident in the mass spectrum shown in panel A. After conversion to molecular mass, a distinct peak at 23 452.17 Da is observed that differs by 0.47 Da compared to the value calculated in the spectrum from the initial diagnosis sample taken over 6 years earlier. Table 1 lists a summary of the results of monitoring the M-protein in serum by PEL, IFE, and microLC-ESI-Q-TOF MS and shows that the light chain is observed throughout the sampling dates, including all of the dates where PEL and IFE were negative. Also, the molecular mass of the light chain remains consistent with an average value of 23 452.54 Da and a standard deviation of 0.86 Da for molecular mass calculations over the 7 year sample period. The heavy chain was observed in
obtained from a selected set of extracted ions. The table demonstrates that the standard curves have a linear dynamic range within the concentration range needed in clinical practice. We examined the interassay precision of 10 replicate Melon Gel preparations of adalimumab spiked into normal serum at 0.1 g/dL and found the CV for the peak area of the light chain to be 6.2%, whereas the CV for the heavy chain was 11%. The limit of quantification as defined by a CV < 20% for 10 replicates using the deconvolution peak areas was 0.005 g/ dL for the light chain and 0.025 g/dL for the heavy chain of adalimumab spiked into normal serum. Monitoring a Monoclonal Immunoglobulin in a Patient with Multiple Myeloma
We next examined a series of samples from a patient diagnosed with IgG kappa multiple myeloma. The mass spectrum from a serum sample is shown in Figure 2. The spectrum in panel A represents a portion of the summed mass spectra across the immunoglobulin LC peak and shows a series of multiply charged ions. The converted molecular mass is shown in panel B and was calculated to be 23 452.64 Da, representing the proposed molecular mass of the kappa light chain portion of the M-protein. The spectrum in panel C shows another portion of the summed immunoglobulin LC peak and displays a different series of multiply charged ions. The molecular mass 1422
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Journal of Proteome Research
Article
Figure 3. Results from the analysis of a serum sample from the same patient shown in Figure 2 after treatment for multiple myeloma. The sample was negative by PEL, IFE, and the FLC assay. (A) Mass spectrum clearly showing a series of multiply charged ions. (B) Calculated molecular mass for the ions in panel A, which is only 0.47 Da different than the molecular mass observed in Figure 2 for the light chain. (C) Molecular mass for the proposed heavy chain ions can no longer be calculated because the ions are below the level of detection.
Table 1. Comparison of M-protein PEL and IFE Results with the Peak Areas and Molecular Masses Observed for the M-protein by microLC-ESI-Q-TOF MSa light chain
heavy chain
sampling date
M-spike (g/dL)
IFE
peak area
molecular mass (Da)
2/23/2005b 3/29/2006 4/24/2007 10/11/2007 4/23/2008 5/7/2009 7/27/2010 8/22/2011c 3/5/2012
4.35 0.26 0 0 0.54 0.43 3.24 0 0.79
pos pos neg neg pos pos pos neg pos
3 010 899 34 839 9 301 11 496 152 021 322 375 3 121 072 2112 600 281
23 452.64 23 452.10 23 451.78 23 452.31 23 452.20 23 452.34 23 452.50 23 452.17 23 452.50
molecular mass (Da) 51 595.07
51 758.27
51 595.46 51 596.66 51 596.56
51 757.84 51 758.52 51 758.91
51 596.44
51 758.74
a
Results are from serum samples obtained from a patient diagnosed with multiple myeloma taken over a 7 year period. bSample date 2/23/2005 was used in Figure 2. cSample date 8/22/2011 was used in Figure 3.
was r2 = 0.9455, whereas the two heavy chain proteoforms had r2 = 0.9205 and 0.9222, respectively. Additional experiments were performed using matched urine and serum samples taken from a patient with a known monoclonal gammopathy to determine if the molecular mass of the monoclonal light chain would remain constant after being excreted through the kidney into the urine. We examined two
the PEL and IFE positive samples, and in those samples, the molecular mass calculations were consistent over the 7 year period. This supports the assumption that M-protein molecular mass is a highly sensitive marker of the plasma cell clone. In addition, a linear regression analysis was done to evaluate the correlation in response between the M-spike value and the peak areas from deconvolution. The correlation for the light chain 1423
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Journal of Proteome Research
Article
Figure 4. Top-down MS of adalimumab spiked into normal serum. The ion at m/z = 1233 in the top spectrum matches the +19 charge state ion from the kappa light chain of adalimumab and was selected for top-down MS. The arrow points to the fragment ion mass spectrum. The labeled fragment ions match the expected masses for fragment ions from the C-terminal portion of the kappa light chain that contains the constant region. The calculated y-ion masses for the kappa light chain constant region-specific amino acid sequence are shown in the table.
light chains, we prepared a set of 20 IgG kappa patients and performed top-down MS analysis on the light chains. (See Supporting Information Figure 3A−C for a comparison of topdown spectra from multiple patients. Top-down MS data from the patient whose results are shown in Figures 2 and 3 are shown in Supporting Information Figure 7.) The fragment ion mass spectrum for each patient was generated from a different multiply charged precursor ion because of an individual’s different variable region amino acid sequence. However, regardless of the patient-specific precursor ion, the same y ions matching the kappa constant region were identified. All 20 of the patients tested showed the same kappa-specific fragment ions. LC-ESI-Q-TOF MS experiments were also performed on a commercially available lambda light chain. The mass spectrum on the top of Figure 5 shows the multiply charged lambda light chain with the fragment ion spectrum on the bottom of Figure 5. Initial comparison between the monoisotopic masses for the observed fragment ions and the potential y ions from the 5′ lambda constant region did not produce matches. The topdown MS protein database search engine ProSight was used to search the fragment ions to find a possible match within the constant region. The table on the right of Figure 5 lists the lambda light chain constant region sequence tag found by ProSight along with the monoisotopic masses for b ions from
patients (one IgA kappa and one IgA lambda) who had previously been identified as having a monoclonal gammopathy. However, the samples analyzed by microLC-ESI-Q-TOF MS were negative by PEL and IFE for a monoclonal immunoglublin in both serum and urine. Our findings showed that the molecular mass of the light chain did not change between serum and urine within the expected mass error of the experiment (see Supporting Information Figure 2). These results reinforce the principle that molecular mass alone can be used to monitor a monoclonal immunoglobulin regardless of the sample type and that microLC-ESI-Q-TOF MS is a method for identifying a monoclonal immunoglobulin in urine. Identification of Light Chain Isotype by Top-Down MS
Top-down MS was done on a multiply charged ion from the light chain of adalimumab, which has a kappa isotype. The results from a top-down analysis using adalimumab spiked into normal serum are shown in Figure 4. Panel A shows the multiply charged ions from the kappa light chain along with an arrow to the fragment ion mass spectrum produced from the CID of the precursor at m/z = 1233 shown in panel B. Fragment ions are labeled with their monoisotopic masses, which closely match the calculated monoisotopic masses for y ions from the constant region of the kappa light chain. To determine if this would hold true for other patients with kappa 1424
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Journal of Proteome Research
Article
Figure 5. Top-down MS of a lambda immunoglobulin light chain standard. The light chain ion at m/z = 1193 in the top spectrum was selected for MS/MS, and the fragment ion mass spectrum is shown below. Fragment ions that match a portion of the lambda light chain constant region are labeled with their respective monoisotopic masses. The calculated b-ion monoisotopic masses for the lambda constant region-specific sequence are shown in the table.
protein biomarker value that can be used to monitor clonal plasma cells over long periods of time. Compared to PEL, this approach is analytically more specific because of the precision of the mass measurements and is more sensitive because of the resolution of different molecular masses within polyclonal backgrounds. The postanalytical process of converting multiply charged ions to their molecular mass (also referred to as deconvolution)26 is done routinely in our laboratories and is used daily in reporting the molecular mass of proteins. The observation that light chain ions from different patients produce the same fragment ions by top-down MS is significant because it allows M-protein monitoring as well as light chain isotyping to be performed in a single assay. It also illustrates the utility of top-down MS for identifying conserved amino acid sequence information in patient-specific immunoglobulin light chains even though the N-terminal half of the protein is randomly rearranged from patient to patient. Evaluation of topdown MS began by first running the sample in LC−MS mode, reviewing the data, and then manually selecting the multiply charged ion for top-down MS. To accelerate the process, we investigated using the automated precursor ion-selection algorithm to select a multiply charged light chain ion. Initial experiments established that the algorithm could properly select light chain ions and generate fragment ion spectra that could be
the sequence. Fragment ions that match the monoisotopic masses in the table are labeled in the spectrum. Although the intensity of the b-ion series for lambda may be less pronounced than those observed for kappa, the fragment ions observed are unique to the lambda light chain isotype. We analyzed a set of 20 patients positive for an IgG lambda light chain by top-down MS fragment, and the b ions matching the N-terminal portion of the lambda constant region were observed in each patient (see Supporting Information Figure 4A−C for a comparison of top-down spectra from multiple patients). In addition, we ran IFE positive urine samples by top-down MS and found that lambda-positive samples had lambda-specific fragment ions and kappa-positive samples had kappa-specific fragment ions. These findings led us to conclude that top-down MS could be used to isotype kappa and lambda light chains.
■
CONCLUSIONS The concept behind monitoring monoclonal immunoglobulins by mass spectrometry is based upon the hypothesis that a clonal plasma cell will secrete an immunoglobulin with a unique amino acid sequence that will therefore have a unique molecular mass. The mass spectrometer can determine the molecular mass of the M-protein light chain and heavy chain with high precision and accuracy, creating an individualized M1425
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Journal of Proteome Research
■
used to determine the light chain isotype (see Supporting Information Figure 5).27 The use of an information-dependent acquisition (IDA) in a top-down MS experiment allows for the expansion of top-down MS into a high-throughput clinical environment. These findings establish that microLC-ESI-QTOF MS can be used to determine M-protein light and heavy chain molecular mass along with light chain isotype in the same run. To date, all automated top-down MS isotyping runs have matched the isotype determined by immunofixation, and in certain cases, top-down MS has been able to isotype a light chain below the IFE level of detection. We are also exploring the use of full-scan microLC-ESI-Q-TOF MS to monitor free light chains in nonreduced serum, and we have observed free kappa, kappa dimers, free lambda, and lambda dimers in both patients and model systems. The results shown here provide the empirical evidence to substantiate the utility of mass spectrometry as a tool to monitor an M-protein in patients with a monoclonal gammopathy. The molecular mass of the monoclonal immunoglobulin, whether it is the light chain, heavy chain, or the intact molecule, represents a sensitive and specific marker of immunoglobulin-secreting plasma cell clones. The methodology can readily identify a monoclonal immunoglobulin above the polyclonal background, providing exceptionally detailed information about the status of patient-specific plasma cell clones. We anticipate that in the future mass spectrometry will play an important role in the quantitation and monitoring of immunoglobulins in human health and disease.
■
Article
AUTHOR INFORMATION
Corresponding Author
*Phone: (507) 293-1321. Fax: (507) 266-4088. E-mail: murray. [email protected]. Notes
The authors declare the following competing financial interest(s): A provisional patent has been filed on the technology.
■
REFERENCES
(1) Kyle, R. A.; Buadi, F.; Rajkumar, S. V. Management of monoclonal gammopathy of undetermined significance (MGUS) and smoldering multiple myeloma (SMM). Oncology 2011, 25, 578−586. (2) Katzmann, J. A. Screening panels for monoclonal gammopathies: Time to change. Clin. Biochem. Rev. 2009, 30, 105−111. (3) Korthals, M.; Sehnke, N.; Kronenwett, R.; Schroeder, T.; Strapatsas, T.; Kobbe, G.; Haas, R.; Fenk, R. Molecular monitoring of minimal residual disease in the peripheral blood of patients with multiple myeloma. Biol. Blood Marrow Transplant. 2013, 19, 1109− 1115. (4) Korthals, M.; Sehnke, N.; Kronenwett, R.; Bruns, I.; Mau, J.; Zohren, F.; Haas, R.; Kobbe, G.; Fenk, R. The level of minimal residual disease in the bone marrow of patients with multiple myeloma before high-dose therapy and autologous blood stem cell transplantation is an independent predictive parameter. Biol. Blood Marrow Transplant. 2012, 18, 423−431. (5) Bianchi, G.; Kyle, R. A.; Larson, D. R.; Witzig, T. E.; Kumar, S.; Dispenzieri, A.; Morice, W. G.; Rajkumar, S. V. High levels of peripheral blood circulating plasma cells as a specific risk factor for progression of smoldering multiple myeloma. Leukemia 2013, 27, 680−685. (6) Bondarenko, P. V.; Second, T. P.; Zabrouskov, V.; Makarov, A. A.; Zhang, Z. Mass measurement and top-down HPLC/MS analysis of intact monoclonal antibodies on a hybrid linear quadrupole ion trapOrbitrap mass spectrometer. J. Am. Soc. Mass Spectrom. 2009, 20, 1415−1424. (7) Ren, D.; Pipes, G. D.; Hambly, D.; Bondarenko, P. V.; Treuheit, M. J.; Gadgil, H. S. Top-down N-terminal sequencing of immunoglobulin subunits with electrospray ionization time of flight mass spectrometry. Anal. Biochem. 2009, 384, 42−48. (8) Zhang, Z.; Pan, H.; Chen, X. Mass spectrometry for structural characterization of therapeutic antibodies. Mass Spectrom. Rev. 2009, 28, 147−176. (9) Zhang, Z.; Shah, B. Characterization of variable regions of monoclonal antibodies by top-down mass spectrometry. Anal. Chem. 2007, 79, 5723−5729. (10) Tsybin, Y. O.; Fornelli, L.; Stoermer, C.; Luebeck, M.; Parra, J.; Nallet, S.; Wurm, F. M.; Hartmer, R. Structural analysis of intact monoclonal antibodies by electron transfer dissociation mass spectrometry. Anal. Chem. 2011, 83, 8919−8927. (11) Fornelli, L.; Damoc, E.; Thomas, P. M.; Kelleher, N. L.; Aizikov, K.; Denisov, E.; Makarov, A.; Tsybin, Y. O. Analysis of intact monoclonal antibody IgG1 by electron transfer dissociation Orbitrap FTMS. Mol. Cell. Proteomics 2012, 11, 1758−1767. (12) Falick, A. M.; Shackleton, C. H.; Green, B. N.; Witkowska, H. E. Tandem mass spectrometry in the clinical analysis of variant hemoglobins. Rapid Commun. Mass Spectrom. 1990, 4, 396−400. (13) Ando, Y.; Ohlsson, P. I.; Suhr, O.; Nyhlin, N.; Yamashita, T.; Holmgren, G.; Danielsson, A.; Sandgren, O.; Uchino, M.; Ando, M. A new simple and rapid screening method for variant transthyretinrelated amyloidosis. Biochem. Biophys. Res. Commun. 1996, 228, 480− 483. (14) Kishikawa, M.; Nakanishi, T.; Miyazaki, A.; Shimizu, A.; Nakazato, M.; Kangawa, K.; Matsuo, H. Simple detection of abnormal serum transthyretin from patients with familial amyloidotic polyneuropathy by high-performance liquid chromatography/electrospray
ASSOCIATED CONTENT
S Supporting Information *
List of linear regression r2 values and reportable ranges for light chain and heavy chain portions of adalimumab spiked into buffer, serum, and urine. LC-MS total-ion chromatogram of adalimumab spiked into serum and analyzed by microLC-ESIQ-TOF MS. Mass spectra from matched serum and urine samples from two patients with multiple myeloma analyzed by microLC-ESI-Q-TOF MS demonstrating that the light chain portion of the IgA M-protein has the same molecular mass in serum and in urine. Top-down fragment ion mass spectrum from three patients with a kappa light chain M-protein out of the 20 analyzed. Top-down fragment ion mass spectra from 10 patients out of the 20 analyzed with a kappa light chain Mprotein shown in an overlaid format as well as close-up view of the monoisotopic y-7 ion from the kappa constant region from the 10 overlaid patients demonstrating that different patients give the same constant region y ions that can be used to isotype the kappa light chain. Top-down fragment ion mass spectrum from three patients with a lambda light chain M-protein out of the 20 analyzed. Top-down fragment ion mass spectra from 10 patients out of the 20 analyzed with a lambda light chain Mprotein shown in an overlaid format as well as close-up view of the monoisotopic b-20 ion from the lambda constant region from the 10 overlaid patients demonstrating that different patients give the same constant region b ions that can be used to isotype the lambda light chain. Example of an automated top-down MS precursor ion selection for the isotyping of an Mprotein light chain. Raw data and linear regression analysis of standard curves shown in Table 1. LC−MS and top-down MS analysis of a positive and negative sample from the patient whose M-protein was monitored over 7 years with the results presented in Table 1. This material is available free of charge via the Internet at http://pubs.acs.org. 1426
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Journal of Proteome Research
Article
ionization mass spectrometry using material precipitated with specific antiserum. J. Mass Spectrom. 1996, 31, 112−114. (15) Theberge, R.; Connors, L.; Skinner, M.; Skare, J.; Costello, C. E. Characterization of transthyretin mutants from serum using immunoprecipitation, HPLC/electrospray ionization and matrixassisted laser desorption/ionization mass spectrometry. Anal. Chem. 1999, 71, 452−459. (16) Bergen, H. R., 3rd; Zeldenrust, S. R.; Naylor, S. An on-line assay for clinical detection of amyloidogenic transthyretin variants directly from serum. Amyloid 2003, 10, 190−197. (17) Lacey, J. M.; Bergen, H. R.; Magera, M. J.; Naylor, S.; O’Brien, J. F. Rapid determination of transferrin isoforms by immunoaffinity liquid chromatography and electrospray mass spectrometry. Clin. Chem. 2001, 47, 513−518. (18) Coelho Graca, D.; Lescuyer, P.; Clerici, L.; Tsybin, Y. O.; Hartmer, R.; Meyer, M.; Samii, K.; Hochstrasser, D. F.; Scherl, A. Electron transfer dissociation mass spectrometry of hemoglobin on clinical samples. J. Am. Soc. Mass Spectrom. 2012, 23, 1750−1756. (19) Edwards, R. L.; Griffiths, P.; Bunch, J.; Cooper, H. J. Top-down proteomics and direct surface sampling of neonatal dried blood spots: Diagnosis of unknown hemoglobin variants. J. Am. Soc. Mass Spectrom. 2012, 23, 1921−1930. (20) Theberge, R.; Infusini, G.; Tong, W.; McComb, M. E.; Costello, C. E. Top-down analysis of small plasma proteins using an LTQorbitrap. Potential for mass spectrometry-based clinical assays for transthyretin and hemoglobin. Int. J. Mass Spectrom. 2011, 300, 130− 142. (21) Mortz, E.; O’Connor, P. B.; Roepstorff, P.; Kelleher, N. L.; Wood, T. D.; McLafferty, F. W.; Mann, M. Sequence tag identification of intact proteins by matching tanden mass spectral data against sequence data bases. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 8264−8267. (22) Zubarev, R. A.; Horn, D. M.; Fridriksson, E. K.; Kelleher, N. L.; Kruger, N. A.; Lewis, M. A.; Carpenter, B. K.; McLafferty, F. W. Electron capture dissociation for structural characterization of multiply charged protein cations. Anal. Chem. 2000, 72, 563−573. (23) Cargile, B. J.; McLuckey, S. A.; Stephenson, J. L., Jr. Identification of bacteriophage MS2 coat protein from E. coli lysates via ion trap collisional activation of intact protein ions. Anal. Chem. 2001, 73, 1277−1285. (24) Dekker, L. J.; Zeneyedpour, L.; Brouwer, E.; van Duijn, M. M.; Sillevis Smitt, P. A.; Luider, T. M. An antibody-based biomarker discovery method by mass spectrometry sequencing of complementarity determining regions. Anal. Bioanal. Chem. 2011, 399, 1081− 1091. (25) Beck, A.; Sanglier-Cianferani, S.; Van Dorsselaer, A. Biosimilar, biobetter, and next generation antibody characterization by mass spectrometry. Anal. Chem. 2012, 84, 4637−4646. (26) Mann, M.; Meng, C. K.; Fenn, J. B. Interpreting mass-spectra of multiply charged ions. Anal. Chem. 1989, 61, 1702−1708. (27) Barnidge, D. R.; Botz, C. M.; Dasari, S.; Snyder, M. R.; Katzmann, J. A.; Murray, D. L. In Diagnosing monoclonal gammopathies using top-down analysis of immunoglobulin light chains in serum and urine by LC-ESI-QTOF mass spectrometry, 61st Annual ASMS Conference on Mass Spectrometry and Allied Topics, Minneapolis, MN, June 9−13; Wiley: Minneapolis, MN, 2013.
1427
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Using Mass Spectrometry to Monitor Monoclonal Immunoglobulins in Patients with a Monoclonal Gammopathy David R. Barnidge,† Surendra Dasari,‡ Chad M. Botz,† Danelle H. Murray,† Melissa R. Snyder,† Jerry A. Katzmann,† Angela Dispenzieri,† and David L. Murray*,† †
Department of Laboratory Medicine and Pathology and ‡Biomedical Statistics and Informatics, Mayo Clinic, Rochester, Minnesota 55905, United States S Supporting Information *
ABSTRACT: A monoclonal gammopathy is defined by the detection a monoclonal immunoglobulin (M-protein). In clinical practice, the M-protein is detected by protein gel electrophoresis (PEL) and immunofixation electrophoresis (IFE). We theorized that molecular mass could be used instead of electrophoretic patterns to identify and quantify the M-protein because each light and heavy chain has a unique amino acid sequence and thus a unique molecular mass whose increased concentration could be distinguished from the normal polyclonal background. In addition, we surmised that top-down MS could be used to isotype the M-protein because each immunoglobulin has a constant region with an amino acid sequence unique to each isotype. Our method first enriches serum for immunoglobulins followed by reduction using DTT to separate light chains from heavy chains and then by microflow LC-ESI-Q-TOF MS. The multiply charged light and heavy chain ions are converted to their molecular masses, and reconstructed peak area calculations for light chains are used for quantification. Using this method, we demonstrate how the light chain portion of an M-protein can be monitored by molecular mass, and we also show that in sequential samples from a patient with multiple myeloma the light chain portion of the M-protein was detected in all samples, even those negative by PEL, IFE, and quantitative FLC. We also present top-down MS isotyping of M-protein light chains using a unique isotype-specific fragmentation pattern allowing for quantification and isotype identification in the same run. Our results show that microLC-ESI-Q-TOF MS provides superior sensitivity and specificity compared to conventional methods and shows promise as a viable method of detecting and isotyping an M-protein. KEYWORDS: Monoclonal gammopathy, immunoglobulin, multiple myeloma, mass spectrometry, top-down MS, isotype
■
INTRODUCTION Plasma cell proliferative disorders are characterized by a clonal expansion of plasma cells. These disorders are referred to as monoclonal gammopathy, paraproteinemia, or plasma cell dyscrasia. The plasma cell clones often secrete elevated levels of a monoclonal immunoglobulin clinically referred to as an Mprotein and commonly referred to as a monoclonal antibody (mAb). The M-protein may consist of the intact immunoglobulin in combination with a free light chain that has not paired with a heavy chain (FLC). When there is clinical suspicion of a plasma cell proliferative disorder, a patient is tested for the presence of the M-protein in serum and urine using a combination of agarose gel serum protein electrophoresis (SPEP), immunofixation electrophoresis (IFE), and free light chain (FLC)-specific immunoassays. If the patient exhibits clinical symptoms suggesting multiple myeloma, then a bone marrow biopsy is performed to quantitate the plasma cells.1 The change in the peak area for the M-spike is normally used to monitor a patient’s response to treatment.2 If treatment eliminates M-protein-producing malignant plasma cells, then the M-spike peak observed by SPEP will fade into the normal © 2014 American Chemical Society
polyclonal immunoglobulin background below the limit of quantification (500 mg/L or 3 μM). IFE, FLC, and bone marrow biopsy can then be used to monitor treatment because they have a lower limit of quantification. The role of minimal residual disease (MRD) by higher sensitivity methods is still being investigated. Currently, new multiparameter flow cytometric assays and PCR-based assays are currently being investigated for their ability to determine minimal residual disease (MRD).3−5 Recent publications have pointed out the major role mass spectrometry has played in the quality control of therapeutic monoclonal antibodies (mAb), including the use of top-down mass spectrometry.6−11 We hypothesized that even without knowing the amino acid sequence of the M-protein, as compared to therapeutic mAb quality control, accurate molecular mass could be used to monitor the M-protein above the polyclonal background. We also surmised that it would be possible to determine the isotype of an M-protein Received: September 27, 2013 Published: January 27, 2014 1419
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Journal of Proteome Research
Article
Samples were then acidified with 1 μL of formic acid prior to injection.
light chain or heavy chain using top-down MS provided the fragment ions gave a sequence tag from the isotype-specific constant region. Currently, there are clinical assays that use mass spectrometry to monitor the molecular mass of intact proteins to diagnose various disorders such as hemoglobinopathies,12 familial amyloidosis,13−16 and congenital disorders of glycosylation.17 However, top-down mass spectrometry has had limited implementation in the clinical laboratory18−20 despite advancements in bioinformatics tools21 and innovative instrument design.22,23 Here, we describe the application of a mass spectrometry-based platform routinely used by pharma in mAb quality control, microLC-ESI-Q-TOF MS, for the rapid detection, quantification, and isotyping of monoclonal immunoglobulins from patients with a monoclonal gammopathy. This methodology represents a major shift in the characterization of monoclonal immunoglobulins. Rather than using gel electrophoresis and anti-isotypic antibody reagents, we combine molecular mass and constant region fragment ions to create a novel clinical laboratory technique for detecting and monitoring monoclonal immunoglobulins.
■
LC Conditions
An Eksigent Ekspert 200 microLC (Dublin, CA) was used for separation; mobile phase A was water + 0.1% FA, and mobile phase B was 80% acetonitrile + 10% 2-propanol + 0.1% FA. A 2 μL injection was made onto a 1.0 × 75 mm Poroshell 300SBC3, 5 μm column flowing at 25 μL/min. A 15 min gradient was started at 80% A/20% B, held for 0.5 min, ramped to 70% A/ 30% B over 1 min, then ramped to 60% A/40% B over 4 min, then ramped to 5% A/95% B over 5 min, held for 2.5 min, then ramped to 80% A/20% B over 1 min, and then equilibrated at 80% A/20% B for 1 min. ESI-Q-TOF MS
Spectra were collected on an ABSciex TripleTOF 5600 quadrupole time-of-flight mass spectrometer (ABSciex, Vaughan, ON, Canada) in ESI positive mode with a Turbo V dual-ion source with an automated calibrant delivery system (CDS). Source conditions were IS, 5500; temp, 500; CUR, 45; GS1, 35; GS2, 30; and CE, 50 ± 5. TOF MS scans were acquired from m/z 600−2500 with an acquisition time of 100 ms. Fragment ion scans were acquired from m/z 350−2000 with an acquisition time of 100 ms. The instrument was calibrated every five injections through the CDS using calibration solution supplied by the manufacturer.
EXPERIMENTAL SECTION
Materials
Serum and Urine Samples. Serum and urine were waste samples in the clinical lab. The paired serum/urine samples and banked sequential samples were from a multiple myeloma patient consented under IRB protocol 521-93. Reagents. Ammonium bicarbonate, dithiothreitol (DTT), and formic acid were purchased from Sigma-Aldrich (St. Louis, MO). Water and acetonitrile were purchased from Honeywell Burdick and Jackson (Muskegon, MI). Monoclonal Immunoglobulins. The therapeutic monoclonal immunoglobulin adalimumab (Humira) was purchased from Abbott Laboratories (Chicago, IL). The amino acid sequence for adalimumab was kindly provided by L. J. Dekker.24 A lambda immunoglobulin light chain standard was purchased from Bethyl Laboratories (Montgomery, TX) and was purified from the urine of a patient with multiple myeloma who consented to have their urine used as a source for lambda light chain immunoglobulin.
Data Analysis
Analyst TF v1.6 was used for instrument control. Data were viewed using Analyst TF v1.6 and PeakView v1.2.0.3. Multiply charged ion peak centroids were used to calculate average molecular mass and the peak area value used for quantification through BioAnalyst software provided with Analyst TF. Multiple ion deconvolution was done between 20 000 and 28 000 Da as a protonated molecule, a step size of 1, a S/N of 20, and 20 iterations for light chain molecular mass calculations. Multiple ion deconvolution was done between 45 000 and 60 000 Da for the heavy chain. MultiQuan v2.1 was used to calculate peak areas for standards using multiple extracted ions. Top-down fragmentation spectra for lambda light chains were searched using ProSight software (Thermo-Fisher Scientific, Waltham, MA) against a small protein sequence database containing human immunoglobulin constant chains and common contaminants. Generated sequence tags were matched against the protein database, and significant p-value scores were found for a C-terminal portion of the lambda light chain.
Serum
A volume of 50 μL of serum was enriched for immunoglobulins using Melon Gel (Thermo-Fisher Scientific, Waltham, MA) following the manufacturer’s instructions. The serum sample was diluted 10-fold with sample purification buffer (50 μL of serum + 450 μL of buffer) and was then added to 500 μL of gel slurry in a sample collection microcentrifuge tube containing a frit. The sample was mixed end-over-end for 5 min and then centrifuged with the purified immunoglobulins being collected in the flow through. After immunoglobulin enrichment, 25 μL of sample was reduced by adding 25 μL of 100 mM DTT and 25 μL of 50 mM ammonium bicarbonate followed by incubation at 55 °C for 30 min. Samples were then acidified with 1 μL of formic acid prior to injection.
Protein Electrophoresis
All assays were performed according to protocols in the Clinical Immunology Laboratory, Department of Laboratory Medicine and Pathology, Mayo Clinic. PEL was performed on a SPIFE SPE system (Helena Laboratories, Beaumont, TX), and IFE, on Sebia 9IF gels (Sebia, Norcross, GA). The total protein concentration was determined by colorimetric assay using biuret reagents from Roche and a Roche Hitachi 912 chemistry analyzer system (Roche Diagnostics, Indianapolis, IN).
Urine
Immunoglobulins were analyzed directly from urine without purification. A volume of 25 μL of urine was reduced by adding 25 μL of 100 mM DTT and 25 μL of 50 mM ammonium bicarbonate followed by incubation at 55 °C for 30 min.
PEL, IFE, and FLC
All assays were performed according to protocols in the Clinical Immunology Laboratory and have been previously described.1 1420
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Journal of Proteome Research
Article
Figure 1. Mass spectra from (A) a normal serum sample and (B) normal serum spiked with 0.5 g/dL of the IgG kappa recombinant mAb adalimumab. The normal serum mass spectrum displays a broad range of unresolved peaks, whereas the normal serum spiked with 0.5 g/dL of adalimumab shows a clearly defined multiply charged protein ions. (C) Converted spectrum for the normal serum sample displaying a broad range of unresolved peaks. (D) Converted mass spectrum for the normal serum spiked with 0.5 g/dL of adalimumab showing a single peak at an average molecular mass of 23 412.19 Da. This mass is in excellent agreement with the calculated average molecular mass of 23 412.13 Da for the kappa light chain of adalimumab.
■
RESULTS
broad distribution of masses with no single mass higher in abundance than the background. This is in sharp contrast to the converted molecular mass for the normal serum spiked with 0.5 g/dL of adalimumab in panel D, which displays a single peak with the observed molecular mass of 23 412.19 Da. This molecular mass agrees with the average molecular mass calculated from the known amino acid sequence for the kappa light chain of adalimumab (23 412.13 Da). To assess microLC-ESI-Q-TOF MS for quantification, adalimumab was spiked into 50 mM ammonium bicarbonate buffer, normal serum, and normal urine (see Supporting Information Table 1). Ten different standard concentrations were used ranging from 0.005 to 5.0 g/dL. Standard curves made in serum used Melon Gel to enrich for immunoglobulins, whereas curves made in urine and buffer were reduced and analyzed without Melon Gel purification (see Supporting Information Figure 6 for raw data). Linearity and linear dynamic range values in the table are split according to the two quantification techniques. The first approach labeled “deconvolution peak area” uses the peak area found after deconvolution of the multiply charged ions to molecular mass, whereas the second approach labeled “extracted ion peak area” refers to using the peak areas
Adalimumab in Normal Serum as a Model System
Adalimumab is an anti-TNF therapeutic monoclonal immunoglobulin that is widely prescribed for downregulating the inflammatory response in patients with autoimmune disorders. Therapeutic monoclonal immunoglobulins such as adalimumab are ideal surrogate standards for simulating a monoclonal immunoglobulin in serum because they are readily available in high purity and typically have a large body of literature on their structural properties. Figure 1 shows the mass spectra for normal serum and serum spiked with 0.5 g/dL (30 μM) of adalimumab. Each mass spectrum represents the spectra summed together over the adalimumab light chain elution time (see Supporting Information Figure 1). The mass spectrum from normal serum in section A shows a broad unresolved peak with a maximum relative abundance of 300 counts per second (cps). Alternatively the mass spectrum from the serum spiked with adalimumab in section B shows a distinct series of peaks from multiply charged protein ions with a maximum relative abundance of 6000 cps. The converted molecular mass for the normal serum in panel C shows a set of 1421
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Journal of Proteome Research
Article
Figure 2. Results from the analysis of a serum sample from a patient with multiple myeloma. (A) Mass spectrum clearly showing a set of multiply charged ions that are (B) converted to a molecular mass of 23 452.64 Da, which is within the expected mass range for a light chain. (C) Another set of multiply charged ions that are (D) converted to a molecular mass of 51 596.07 and 51 758.27 Da, which is within the expected mass range for a heavy chain. The difference between the molecular mass of series 2 and series 3 is 162.20 Da, which closely matches the mass of a hexose subunit.
for this series shown in panel D was found to have two components, one at 51 596.07 Da and another at 51 758.27 Da, both of which are consistent with IgG heavy chain. The difference of 162.20 Da between series 2 and 3 may be due to two heavy chain proteoforms differing by the number of hexose units (average MW, 162.14 Da) in the carbohydrate chain.25 Additional samples taken over a 7 year period were available for testing from this patient. Figure 3 shows the result using mass spectrometry for a sample taken after the patient had been treated for multiple myeloma and was found to be negative by PEL, IFE, and the quantitative FLC immunoassay. However, multiply charged ions from the light chain are clearly evident in the mass spectrum shown in panel A. After conversion to molecular mass, a distinct peak at 23 452.17 Da is observed that differs by 0.47 Da compared to the value calculated in the spectrum from the initial diagnosis sample taken over 6 years earlier. Table 1 lists a summary of the results of monitoring the M-protein in serum by PEL, IFE, and microLC-ESI-Q-TOF MS and shows that the light chain is observed throughout the sampling dates, including all of the dates where PEL and IFE were negative. Also, the molecular mass of the light chain remains consistent with an average value of 23 452.54 Da and a standard deviation of 0.86 Da for molecular mass calculations over the 7 year sample period. The heavy chain was observed in
obtained from a selected set of extracted ions. The table demonstrates that the standard curves have a linear dynamic range within the concentration range needed in clinical practice. We examined the interassay precision of 10 replicate Melon Gel preparations of adalimumab spiked into normal serum at 0.1 g/dL and found the CV for the peak area of the light chain to be 6.2%, whereas the CV for the heavy chain was 11%. The limit of quantification as defined by a CV < 20% for 10 replicates using the deconvolution peak areas was 0.005 g/ dL for the light chain and 0.025 g/dL for the heavy chain of adalimumab spiked into normal serum. Monitoring a Monoclonal Immunoglobulin in a Patient with Multiple Myeloma
We next examined a series of samples from a patient diagnosed with IgG kappa multiple myeloma. The mass spectrum from a serum sample is shown in Figure 2. The spectrum in panel A represents a portion of the summed mass spectra across the immunoglobulin LC peak and shows a series of multiply charged ions. The converted molecular mass is shown in panel B and was calculated to be 23 452.64 Da, representing the proposed molecular mass of the kappa light chain portion of the M-protein. The spectrum in panel C shows another portion of the summed immunoglobulin LC peak and displays a different series of multiply charged ions. The molecular mass 1422
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Journal of Proteome Research
Article
Figure 3. Results from the analysis of a serum sample from the same patient shown in Figure 2 after treatment for multiple myeloma. The sample was negative by PEL, IFE, and the FLC assay. (A) Mass spectrum clearly showing a series of multiply charged ions. (B) Calculated molecular mass for the ions in panel A, which is only 0.47 Da different than the molecular mass observed in Figure 2 for the light chain. (C) Molecular mass for the proposed heavy chain ions can no longer be calculated because the ions are below the level of detection.
Table 1. Comparison of M-protein PEL and IFE Results with the Peak Areas and Molecular Masses Observed for the M-protein by microLC-ESI-Q-TOF MSa light chain
heavy chain
sampling date
M-spike (g/dL)
IFE
peak area
molecular mass (Da)
2/23/2005b 3/29/2006 4/24/2007 10/11/2007 4/23/2008 5/7/2009 7/27/2010 8/22/2011c 3/5/2012
4.35 0.26 0 0 0.54 0.43 3.24 0 0.79
pos pos neg neg pos pos pos neg pos
3 010 899 34 839 9 301 11 496 152 021 322 375 3 121 072 2112 600 281
23 452.64 23 452.10 23 451.78 23 452.31 23 452.20 23 452.34 23 452.50 23 452.17 23 452.50
molecular mass (Da) 51 595.07
51 758.27
51 595.46 51 596.66 51 596.56
51 757.84 51 758.52 51 758.91
51 596.44
51 758.74
a
Results are from serum samples obtained from a patient diagnosed with multiple myeloma taken over a 7 year period. bSample date 2/23/2005 was used in Figure 2. cSample date 8/22/2011 was used in Figure 3.
was r2 = 0.9455, whereas the two heavy chain proteoforms had r2 = 0.9205 and 0.9222, respectively. Additional experiments were performed using matched urine and serum samples taken from a patient with a known monoclonal gammopathy to determine if the molecular mass of the monoclonal light chain would remain constant after being excreted through the kidney into the urine. We examined two
the PEL and IFE positive samples, and in those samples, the molecular mass calculations were consistent over the 7 year period. This supports the assumption that M-protein molecular mass is a highly sensitive marker of the plasma cell clone. In addition, a linear regression analysis was done to evaluate the correlation in response between the M-spike value and the peak areas from deconvolution. The correlation for the light chain 1423
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Journal of Proteome Research
Article
Figure 4. Top-down MS of adalimumab spiked into normal serum. The ion at m/z = 1233 in the top spectrum matches the +19 charge state ion from the kappa light chain of adalimumab and was selected for top-down MS. The arrow points to the fragment ion mass spectrum. The labeled fragment ions match the expected masses for fragment ions from the C-terminal portion of the kappa light chain that contains the constant region. The calculated y-ion masses for the kappa light chain constant region-specific amino acid sequence are shown in the table.
light chains, we prepared a set of 20 IgG kappa patients and performed top-down MS analysis on the light chains. (See Supporting Information Figure 3A−C for a comparison of topdown spectra from multiple patients. Top-down MS data from the patient whose results are shown in Figures 2 and 3 are shown in Supporting Information Figure 7.) The fragment ion mass spectrum for each patient was generated from a different multiply charged precursor ion because of an individual’s different variable region amino acid sequence. However, regardless of the patient-specific precursor ion, the same y ions matching the kappa constant region were identified. All 20 of the patients tested showed the same kappa-specific fragment ions. LC-ESI-Q-TOF MS experiments were also performed on a commercially available lambda light chain. The mass spectrum on the top of Figure 5 shows the multiply charged lambda light chain with the fragment ion spectrum on the bottom of Figure 5. Initial comparison between the monoisotopic masses for the observed fragment ions and the potential y ions from the 5′ lambda constant region did not produce matches. The topdown MS protein database search engine ProSight was used to search the fragment ions to find a possible match within the constant region. The table on the right of Figure 5 lists the lambda light chain constant region sequence tag found by ProSight along with the monoisotopic masses for b ions from
patients (one IgA kappa and one IgA lambda) who had previously been identified as having a monoclonal gammopathy. However, the samples analyzed by microLC-ESI-Q-TOF MS were negative by PEL and IFE for a monoclonal immunoglublin in both serum and urine. Our findings showed that the molecular mass of the light chain did not change between serum and urine within the expected mass error of the experiment (see Supporting Information Figure 2). These results reinforce the principle that molecular mass alone can be used to monitor a monoclonal immunoglobulin regardless of the sample type and that microLC-ESI-Q-TOF MS is a method for identifying a monoclonal immunoglobulin in urine. Identification of Light Chain Isotype by Top-Down MS
Top-down MS was done on a multiply charged ion from the light chain of adalimumab, which has a kappa isotype. The results from a top-down analysis using adalimumab spiked into normal serum are shown in Figure 4. Panel A shows the multiply charged ions from the kappa light chain along with an arrow to the fragment ion mass spectrum produced from the CID of the precursor at m/z = 1233 shown in panel B. Fragment ions are labeled with their monoisotopic masses, which closely match the calculated monoisotopic masses for y ions from the constant region of the kappa light chain. To determine if this would hold true for other patients with kappa 1424
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Journal of Proteome Research
Article
Figure 5. Top-down MS of a lambda immunoglobulin light chain standard. The light chain ion at m/z = 1193 in the top spectrum was selected for MS/MS, and the fragment ion mass spectrum is shown below. Fragment ions that match a portion of the lambda light chain constant region are labeled with their respective monoisotopic masses. The calculated b-ion monoisotopic masses for the lambda constant region-specific sequence are shown in the table.
protein biomarker value that can be used to monitor clonal plasma cells over long periods of time. Compared to PEL, this approach is analytically more specific because of the precision of the mass measurements and is more sensitive because of the resolution of different molecular masses within polyclonal backgrounds. The postanalytical process of converting multiply charged ions to their molecular mass (also referred to as deconvolution)26 is done routinely in our laboratories and is used daily in reporting the molecular mass of proteins. The observation that light chain ions from different patients produce the same fragment ions by top-down MS is significant because it allows M-protein monitoring as well as light chain isotyping to be performed in a single assay. It also illustrates the utility of top-down MS for identifying conserved amino acid sequence information in patient-specific immunoglobulin light chains even though the N-terminal half of the protein is randomly rearranged from patient to patient. Evaluation of topdown MS began by first running the sample in LC−MS mode, reviewing the data, and then manually selecting the multiply charged ion for top-down MS. To accelerate the process, we investigated using the automated precursor ion-selection algorithm to select a multiply charged light chain ion. Initial experiments established that the algorithm could properly select light chain ions and generate fragment ion spectra that could be
the sequence. Fragment ions that match the monoisotopic masses in the table are labeled in the spectrum. Although the intensity of the b-ion series for lambda may be less pronounced than those observed for kappa, the fragment ions observed are unique to the lambda light chain isotype. We analyzed a set of 20 patients positive for an IgG lambda light chain by top-down MS fragment, and the b ions matching the N-terminal portion of the lambda constant region were observed in each patient (see Supporting Information Figure 4A−C for a comparison of top-down spectra from multiple patients). In addition, we ran IFE positive urine samples by top-down MS and found that lambda-positive samples had lambda-specific fragment ions and kappa-positive samples had kappa-specific fragment ions. These findings led us to conclude that top-down MS could be used to isotype kappa and lambda light chains.
■
CONCLUSIONS The concept behind monitoring monoclonal immunoglobulins by mass spectrometry is based upon the hypothesis that a clonal plasma cell will secrete an immunoglobulin with a unique amino acid sequence that will therefore have a unique molecular mass. The mass spectrometer can determine the molecular mass of the M-protein light chain and heavy chain with high precision and accuracy, creating an individualized M1425
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Journal of Proteome Research
■
used to determine the light chain isotype (see Supporting Information Figure 5).27 The use of an information-dependent acquisition (IDA) in a top-down MS experiment allows for the expansion of top-down MS into a high-throughput clinical environment. These findings establish that microLC-ESI-QTOF MS can be used to determine M-protein light and heavy chain molecular mass along with light chain isotype in the same run. To date, all automated top-down MS isotyping runs have matched the isotype determined by immunofixation, and in certain cases, top-down MS has been able to isotype a light chain below the IFE level of detection. We are also exploring the use of full-scan microLC-ESI-Q-TOF MS to monitor free light chains in nonreduced serum, and we have observed free kappa, kappa dimers, free lambda, and lambda dimers in both patients and model systems. The results shown here provide the empirical evidence to substantiate the utility of mass spectrometry as a tool to monitor an M-protein in patients with a monoclonal gammopathy. The molecular mass of the monoclonal immunoglobulin, whether it is the light chain, heavy chain, or the intact molecule, represents a sensitive and specific marker of immunoglobulin-secreting plasma cell clones. The methodology can readily identify a monoclonal immunoglobulin above the polyclonal background, providing exceptionally detailed information about the status of patient-specific plasma cell clones. We anticipate that in the future mass spectrometry will play an important role in the quantitation and monitoring of immunoglobulins in human health and disease.
■
Article
AUTHOR INFORMATION
Corresponding Author
*Phone: (507) 293-1321. Fax: (507) 266-4088. E-mail: murray. [email protected]. Notes
The authors declare the following competing financial interest(s): A provisional patent has been filed on the technology.
■
REFERENCES
(1) Kyle, R. A.; Buadi, F.; Rajkumar, S. V. Management of monoclonal gammopathy of undetermined significance (MGUS) and smoldering multiple myeloma (SMM). Oncology 2011, 25, 578−586. (2) Katzmann, J. A. Screening panels for monoclonal gammopathies: Time to change. Clin. Biochem. Rev. 2009, 30, 105−111. (3) Korthals, M.; Sehnke, N.; Kronenwett, R.; Schroeder, T.; Strapatsas, T.; Kobbe, G.; Haas, R.; Fenk, R. Molecular monitoring of minimal residual disease in the peripheral blood of patients with multiple myeloma. Biol. Blood Marrow Transplant. 2013, 19, 1109− 1115. (4) Korthals, M.; Sehnke, N.; Kronenwett, R.; Bruns, I.; Mau, J.; Zohren, F.; Haas, R.; Kobbe, G.; Fenk, R. The level of minimal residual disease in the bone marrow of patients with multiple myeloma before high-dose therapy and autologous blood stem cell transplantation is an independent predictive parameter. Biol. Blood Marrow Transplant. 2012, 18, 423−431. (5) Bianchi, G.; Kyle, R. A.; Larson, D. R.; Witzig, T. E.; Kumar, S.; Dispenzieri, A.; Morice, W. G.; Rajkumar, S. V. High levels of peripheral blood circulating plasma cells as a specific risk factor for progression of smoldering multiple myeloma. Leukemia 2013, 27, 680−685. (6) Bondarenko, P. V.; Second, T. P.; Zabrouskov, V.; Makarov, A. A.; Zhang, Z. Mass measurement and top-down HPLC/MS analysis of intact monoclonal antibodies on a hybrid linear quadrupole ion trapOrbitrap mass spectrometer. J. Am. Soc. Mass Spectrom. 2009, 20, 1415−1424. (7) Ren, D.; Pipes, G. D.; Hambly, D.; Bondarenko, P. V.; Treuheit, M. J.; Gadgil, H. S. Top-down N-terminal sequencing of immunoglobulin subunits with electrospray ionization time of flight mass spectrometry. Anal. Biochem. 2009, 384, 42−48. (8) Zhang, Z.; Pan, H.; Chen, X. Mass spectrometry for structural characterization of therapeutic antibodies. Mass Spectrom. Rev. 2009, 28, 147−176. (9) Zhang, Z.; Shah, B. Characterization of variable regions of monoclonal antibodies by top-down mass spectrometry. Anal. Chem. 2007, 79, 5723−5729. (10) Tsybin, Y. O.; Fornelli, L.; Stoermer, C.; Luebeck, M.; Parra, J.; Nallet, S.; Wurm, F. M.; Hartmer, R. Structural analysis of intact monoclonal antibodies by electron transfer dissociation mass spectrometry. Anal. Chem. 2011, 83, 8919−8927. (11) Fornelli, L.; Damoc, E.; Thomas, P. M.; Kelleher, N. L.; Aizikov, K.; Denisov, E.; Makarov, A.; Tsybin, Y. O. Analysis of intact monoclonal antibody IgG1 by electron transfer dissociation Orbitrap FTMS. Mol. Cell. Proteomics 2012, 11, 1758−1767. (12) Falick, A. M.; Shackleton, C. H.; Green, B. N.; Witkowska, H. E. Tandem mass spectrometry in the clinical analysis of variant hemoglobins. Rapid Commun. Mass Spectrom. 1990, 4, 396−400. (13) Ando, Y.; Ohlsson, P. I.; Suhr, O.; Nyhlin, N.; Yamashita, T.; Holmgren, G.; Danielsson, A.; Sandgren, O.; Uchino, M.; Ando, M. A new simple and rapid screening method for variant transthyretinrelated amyloidosis. Biochem. Biophys. Res. Commun. 1996, 228, 480− 483. (14) Kishikawa, M.; Nakanishi, T.; Miyazaki, A.; Shimizu, A.; Nakazato, M.; Kangawa, K.; Matsuo, H. Simple detection of abnormal serum transthyretin from patients with familial amyloidotic polyneuropathy by high-performance liquid chromatography/electrospray
ASSOCIATED CONTENT
S Supporting Information *
List of linear regression r2 values and reportable ranges for light chain and heavy chain portions of adalimumab spiked into buffer, serum, and urine. LC-MS total-ion chromatogram of adalimumab spiked into serum and analyzed by microLC-ESIQ-TOF MS. Mass spectra from matched serum and urine samples from two patients with multiple myeloma analyzed by microLC-ESI-Q-TOF MS demonstrating that the light chain portion of the IgA M-protein has the same molecular mass in serum and in urine. Top-down fragment ion mass spectrum from three patients with a kappa light chain M-protein out of the 20 analyzed. Top-down fragment ion mass spectra from 10 patients out of the 20 analyzed with a kappa light chain Mprotein shown in an overlaid format as well as close-up view of the monoisotopic y-7 ion from the kappa constant region from the 10 overlaid patients demonstrating that different patients give the same constant region y ions that can be used to isotype the kappa light chain. Top-down fragment ion mass spectrum from three patients with a lambda light chain M-protein out of the 20 analyzed. Top-down fragment ion mass spectra from 10 patients out of the 20 analyzed with a lambda light chain Mprotein shown in an overlaid format as well as close-up view of the monoisotopic b-20 ion from the lambda constant region from the 10 overlaid patients demonstrating that different patients give the same constant region b ions that can be used to isotype the lambda light chain. Example of an automated top-down MS precursor ion selection for the isotyping of an Mprotein light chain. Raw data and linear regression analysis of standard curves shown in Table 1. LC−MS and top-down MS analysis of a positive and negative sample from the patient whose M-protein was monitored over 7 years with the results presented in Table 1. This material is available free of charge via the Internet at http://pubs.acs.org. 1426
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
Journal of Proteome Research
Article
ionization mass spectrometry using material precipitated with specific antiserum. J. Mass Spectrom. 1996, 31, 112−114. (15) Theberge, R.; Connors, L.; Skinner, M.; Skare, J.; Costello, C. E. Characterization of transthyretin mutants from serum using immunoprecipitation, HPLC/electrospray ionization and matrixassisted laser desorption/ionization mass spectrometry. Anal. Chem. 1999, 71, 452−459. (16) Bergen, H. R., 3rd; Zeldenrust, S. R.; Naylor, S. An on-line assay for clinical detection of amyloidogenic transthyretin variants directly from serum. Amyloid 2003, 10, 190−197. (17) Lacey, J. M.; Bergen, H. R.; Magera, M. J.; Naylor, S.; O’Brien, J. F. Rapid determination of transferrin isoforms by immunoaffinity liquid chromatography and electrospray mass spectrometry. Clin. Chem. 2001, 47, 513−518. (18) Coelho Graca, D.; Lescuyer, P.; Clerici, L.; Tsybin, Y. O.; Hartmer, R.; Meyer, M.; Samii, K.; Hochstrasser, D. F.; Scherl, A. Electron transfer dissociation mass spectrometry of hemoglobin on clinical samples. J. Am. Soc. Mass Spectrom. 2012, 23, 1750−1756. (19) Edwards, R. L.; Griffiths, P.; Bunch, J.; Cooper, H. J. Top-down proteomics and direct surface sampling of neonatal dried blood spots: Diagnosis of unknown hemoglobin variants. J. Am. Soc. Mass Spectrom. 2012, 23, 1921−1930. (20) Theberge, R.; Infusini, G.; Tong, W.; McComb, M. E.; Costello, C. E. Top-down analysis of small plasma proteins using an LTQorbitrap. Potential for mass spectrometry-based clinical assays for transthyretin and hemoglobin. Int. J. Mass Spectrom. 2011, 300, 130− 142. (21) Mortz, E.; O’Connor, P. B.; Roepstorff, P.; Kelleher, N. L.; Wood, T. D.; McLafferty, F. W.; Mann, M. Sequence tag identification of intact proteins by matching tanden mass spectral data against sequence data bases. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 8264−8267. (22) Zubarev, R. A.; Horn, D. M.; Fridriksson, E. K.; Kelleher, N. L.; Kruger, N. A.; Lewis, M. A.; Carpenter, B. K.; McLafferty, F. W. Electron capture dissociation for structural characterization of multiply charged protein cations. Anal. Chem. 2000, 72, 563−573. (23) Cargile, B. J.; McLuckey, S. A.; Stephenson, J. L., Jr. Identification of bacteriophage MS2 coat protein from E. coli lysates via ion trap collisional activation of intact protein ions. Anal. Chem. 2001, 73, 1277−1285. (24) Dekker, L. J.; Zeneyedpour, L.; Brouwer, E.; van Duijn, M. M.; Sillevis Smitt, P. A.; Luider, T. M. An antibody-based biomarker discovery method by mass spectrometry sequencing of complementarity determining regions. Anal. Bioanal. Chem. 2011, 399, 1081− 1091. (25) Beck, A.; Sanglier-Cianferani, S.; Van Dorsselaer, A. Biosimilar, biobetter, and next generation antibody characterization by mass spectrometry. Anal. Chem. 2012, 84, 4637−4646. (26) Mann, M.; Meng, C. K.; Fenn, J. B. Interpreting mass-spectra of multiply charged ions. Anal. Chem. 1989, 61, 1702−1708. (27) Barnidge, D. R.; Botz, C. M.; Dasari, S.; Snyder, M. R.; Katzmann, J. A.; Murray, D. L. In Diagnosing monoclonal gammopathies using top-down analysis of immunoglobulin light chains in serum and urine by LC-ESI-QTOF mass spectrometry, 61st Annual ASMS Conference on Mass Spectrometry and Allied Topics, Minneapolis, MN, June 9−13; Wiley: Minneapolis, MN, 2013.
1427
dx.doi.org/10.1021/pr400985k | J. Proteome Res. 2014, 13, 1419−1427
