MALDI-TOF mass spectrometry-based quantification of C-peptide in diabetes patients

Abstract

Background:

Serum C-peptide concentrations reflect insulin secretion and beta cell function and can be used to diagnose and distinguish type-1 and type-2 diabetes. C-peptide is a more accurate indicator of insulin status than direct insulin measurement for monitoring patients with diabetes. However, the current methods available for C-peptide quantification exhibit poor reproducibility, are costly, and require highly trained laboratory personnel. Here, we have developed and evaluated a matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS)-based assay to standardize C-peptide measurements, providing highly accurate and comparable results across testing systems and laboratories.

Methods:

C-peptide from human serum was enriched using antibody-conjugated magnetic beads. The eluted isolates were further modified with 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate (AQC) to enhance the ionization of naturally acidic C-peptide. After desalting with ZipTips, the samples were subjected to MALDI-TOF MS analysis. Recombinant human C-peptide was used to develop the assay, and a heavy isotope labeled human C-peptide was used as an internal standard for quantification.

Results:

The MALDI-TOF MS method was validated in accordance with the restrictions of the device, with a limit of quantitation of 25 pmol/L. A correlation between the MAL-DI-TOF MS assay and a reference method was conducted using patient samples. The resulting regression revealed good agreement.

Conclusions:

A simple, high-throughput, cost effective and quantitative MALDI-TOF MS C-peptide assay has been successfully developed and validated in clinical serum samples.

Introduction

C-peptide, a component of the insulin pro-hormone secreted at equimolar concentrations with insulin, is used as a ‘gold standard’ clinical marker of endogenous insulin release and an indicator of β-cell activity.¹ It is also employed to diagnose and distinguish type-1 and type-2 diabetes, a distinction fundamental to diabetes management.² In addition, increased C-peptide levels, together with other parameters, have been found to be valuable indicators of the risk of colorectal cancer and other diseases.³

The reference range of C-peptide is 379 to 1631 pmol/L (90% confidence interval) in blood plasma.⁴ Numerous assays for C-peptide detection have been described. However, traditional methods, such as radioimmunoassays (RIAs) and enzyme-linked immunosorbent assays (ELISA), exhibit variable results that can be attributed to system calibration or factors related to inconsistencies with reagents. The latter must be applied within a specific concentration range. Outside of that range, a linear correction function cannot be applied.⁵ Electrochemiluminescence (ECL) is widely used for biomolecule detection in clinical practice due to its simplified optical setup, high sensitivity and labor-saving merits. This method has also been applied to C-peptide detection.^6,7 However, an increased sensitivity of the method is necessary, requiring both a reaction that stimulates compounds to increase the initial luminescence and an appropriate quenching reagent to decrease the final ECL signal.⁸ Newly developed methods for effectively overcoming the obstacles facing C-peptide measurements have been described. More specifically, high-performance liquid chromatography (HPLC) and high-resolution LC-MS/MS have both been applied to detect and quantify C-peptide in clinical samples. These assays involve standard technologies that are highly sensitive, identifying proteolytic fragments with high confidence;⁹ however, they usually include extremely complex processes and experienced laboratory personnel.

Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) is widely recognized as a high-throughput tool for the detection of various disease biomarkers in body fluids, with high sensitivity, accurate qualitative determination and rapid turnaround time. However, the ionization efficiency influences the utility of MALDI in peptide detection.¹⁰ Thus, it becomes necessary to increase the ionization potential of C-peptide prior to MS analysis to improve selective detection.

Derivatizing reagents, such as ammonia solution, ammonium bicarbonate, ammonium sulfate, dansyl chloride and 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate (AQC), are wildly used for labeling amines, amino acids, peptides, proteins and enhancing the MALDI detection sensitivity^11–14 by improving the ionization efficiency of target peptides.¹⁵ We chose AQC for chemical modification in C-peptide sample preparation prior to the MALDI analysis in our study, based on preliminary data (not shown) and the previous experience of other groups.¹⁶ AQC is a labeling reagent that preferentially binds to free amino groups, generating stable asymmetric urea derivatives. AQC can increase the ionization efficiency of C-peptide by adding a positive charge to the N-terminal amino group, and dramatically enhancing the signal in MALDI detection.¹⁶

In our study, we optimized an ionization procedure of the naturally acidic C-peptide by AQC derivatization and then applied a precise method for C-peptide detection by MALDI-TOF MS. The method we describe is efficient and robust, offering early C-peptide detection for clinical applications.

Materials and methods

Standards and reagents

Recombinant human C-peptide was purchased from Peptide International, Inc. (Louisville, KY, USA). Mouse monoclonal [7E10] anti-human C-peptide antibody was purchased from Abcam (Cambridge, MA, USA). Serum and plasma quality control samples were from a pool of human blood donors. Isotopically labeled C-peptide with the sequence EAEDLQ VGQVELeu(13C6,15N)Gly(15 N)Gly(15 N)Gly(15 N)-PGAGSLQPLALEGSLQ (mass shift, 10 Da) was synthesized by Shanghai Jier Bio-Chemical Ltd (Shanghai, China). The AccQ-Tag Ultra Derivatization Kit (consisting of AQC, acetonitrile for reconstituting the reagent powder and a borate buffer for ensuring the optimal pH) was obtained from Waters (Milford, MA, USA). ZipTip^™ pipette tips, which contain C18 silica (15 μm, 200 Å pore size) in a 0.6 μL bed volume, from EMD Millipore (Billerica, MA), were used for desalting and concentrating the peptides. All other chemicals were of the highest commercial purity. All other solvents were HPLC grade, purchased from Thermo Fisher Scientific Inc. (USA).

MALDI-TOF MS parameters

The measurements of mass spectra were acquired by a Bruker Microflex LRF (Bremen, Germany), equipped with a 60 Hz nitrogen laser. The spectra accumulated to 1000 shots (17 s) with three replicates were recorded in the positive linear mode with a mass range of 1000 to 4000 m/z at the 30% laser intensity. The raw data were analyzed with FlexAnalysis 3.0 (Bruker Daltonics).

Preparation of C-peptide standards

A stock solution of human C-peptide (recombinant human C-peptide) at 500 μmol/L was prepared in water and stored at −80°C until use. For use in experiments, the stock solution was immediately thawed on ice and diluted into C-peptide-free human serum to the desired final concentrations.

Antibody conjugation to magnetic beads

Mouse monoclonal anti-human C-peptide antibodies were conjugated to protein-A magnetic beads (Dynabeads^® Protein A, ThermoFisher Scientific, USA) according to the manufacturer’s instructions. The Ab-bead suspensions were prepared using 10 μg antibodies and 20 μL magnetic bead suspension for each sample.

Sample collection and preparation

Human blood samples were obtained from the National Institutes of Health (NIH) and stored at −80°C before use. Serum and plasma were immediately processed and stored at −80°C until analysis, at which time they were thawed on ice and mixed by vortexing before processing. One hundred microliters from each sample was mixed with 300 μL phosphate-buffered saline (PBS)/0.02% Tween-20 (PBST). The ratio of the sample to PBST was 1:3 (v/v). The samples were then incubated with the Ab-bead suspension using a rotating mixer at room temperature for 1 h. To separate the unbound material, the Ab-bead mixtures were placed on a magnetic stand and the liquid aspirated. The beads were then washed three times with PBS by repeatedly placing the samples on the magnetic stand, removing the buffer and resuspending the beads in fresh PBS. Five microliters of 0.1% trifluoroacetic acid were used to elute the bound antibody from the beads for 1 h and dried in a vacuum centrifuge.

Sample derivatization

Each sample pellet was dissolved in 20 μL of borate buffer. Five microliters of reconstituted AccQ-Tag Ultra reagent was added to each sample, which was then vortexed for several seconds. The sample were incubated for 1 min at room temperature and then heated in a heat block at 55°C for 10min. Different incubation times (10min, 20min, 30min, 40min, 50min and 60min) were compared. After the incubation, the samples were processed using ZipTips prior to MALDI-TOF analysis.

MALDI-TOF MS analysis

The prepared samples (2 μL), including 94C-peptide samples and two calibration standard solutions, were spotted onto the MALDI target plate for use with the MALDI (Bruker Daltonics Inc.) system. After the samples had dried, 1 μL of matrix solution (α-cyano-4-hydroxycinnamic acid, CHCA) was used to coat each sample well. Mass spectrometry was performed under a linear mode.

Method validation

Inter-day imprecision was determined by measuring the quality control samples on 20 different days. Intra-day imprecision was determined by measuring the concentrations of three replicates of the quality control samples on a single day. Recovery was performed by spiking in 2 μL of 500 nM recombinant C-peptide into 100 μL of C-peptide-free serum (label as 10 nM serum) and 1 μL of 500 nM into 100 μL of C-peptide-free serum (label as 5 nM serum), diluted with 300 μL of PBST. Antibody-conjugated Dynabeads (10 μg antibody and 20 μL beads per sample) were added into samples and incubated at 55°C. To calculate the recovery, the same amount of C-peptide (10nM 2 μL and 5 nM 1 μL) was included in this step. Thus, the same treatment was conducted on the eluted peptide and the peptide standards. AQC-based derivatization and ZipTip desalting were performed then. C-peptide isotopic standard was added into each sample before MALDI-TOF analysis. Limit of quantification (LOQ) is the lowest concentration at which an analyte can be assessed, with the bias and imprecision meeting a predefined standard.¹⁷ The LOQ was determined by measuring the intra-day CV% (n = 8) of samples with decreasing concentrations of C-peptide and the estimation of the lowest concentration that could be measured with an intra-day CV% < 10%.

Results

Workflow of C-peptide quantification using MALDI-TOF MS

The workflow for the optimized MS-based assay to measure C-peptide in serum is shown in Figure 1. Serum samples were prepared by diluting the sample to a ratio of 1:3 in 0.02% PBST before immunoaffinity purification. C-peptide was captured using C-peptide monoclonal antibodies conjugated to magnetic beads. Unbound sample components were washed away before the bound components were eluted. Bound C-peptide was dried in a Speed Vac and then modified by AQC. Finally, the C-peptide samples were processed on ZipTips before quantification by MALDI-TOF MS.

Figure 1.

Schematic of C-peptide detection work flow. C-peptide was captured using C-peptide monoclonal antibodies conjugated to magnetic beads. Unbound sample components were washed away before bound components were released. Bound C-peptide were dried, and then modified by AQC. Finally, C-peptide was quantified by MALDI-TOF MS.

MALDI-TOF MS: matrix-assisted laser desorption ionization time-of-flight mass spectrometry; PBS: phosphate-buffered saline; TFA: trifluoroacetic acid; PBST: phosphate-buffered saline 0.02% Tween-20; mAb: monoclonal antibody.

Immunoaffinity purification of C-peptide from serum

Nanotechnologies have been developed for novel applications in all areas of medicine.¹⁸ Magnetic beads, a class of nanoparticles, with immobilized antibodies can isolate proteins for subsequent detection systems. In contrast to most other immunoaffinity purification carriers, magnetic beads are readily applicable to small-scale samples.¹⁹

We initially tested two magnetic bead preparations (containing protein A or protein G) for the immune-pull down step. The intensity of C-peptide that was enriched with antibody-bound protein-A magnetic beads was higher than that using the same amount antibody-bound protein-G magnetic beads. The protein A magnetic beads were selected for use in this assay because of their better assay performance (Figure 2(a)). Sample mixtures were prepared with different amounts of PBST solution. We found that C-peptide intensity was maximally increased when diluting the serum samples to a ratio of 1:3 in 0.02% PBST before immunoaffinity purification (Figure 2(b)).

Figure 2.

Immunoaffinity purification of C-peptide from serum. (a) Two magnetic beads were compared about their affinity with antibody, the intensity of C-peptide that was enriched with antibody-banded protein-A magnetic beads is higher than those using same amount of antibody-banded protein-G magnetic beads. (b) Different volumes of PBST mixed with 100 μl serum samples before enrichment with magnetic beads. (c) The best ratio between the amounts of magnetic beads and C-peptide antibody are 20 μl beads to 10 μg antibody for each sample preparation.

PBST: phosphate-buffered saline 0.02% Tween-20.

Following the selection of magnetic beads, multiple parameters of the sample preparation and the amounts of antibody were optimized to allow for maximized C-peptide recovery. The minimum amounts of antibody required for optimal C-peptide recovery were determined to be 10 mg of antibody for each sample (Figure 2(c)).

Chemical modification and method optimization

Our results show that the intensities of the different C-peptide concentrations were increased in those samples which were modified with AQC without enrichment (Figure 3(a)). The 10-min incubation in the heat block resulted in the highest intensity among the different time intervals tested (p < 0.05) (Figure 3(b)). The mass spectrum of intensity of C-peptide with/without AQC modification is shown in Figure 3(c). Importantly, the intensity of C-peptide was further enhanced after enrichment with magnetic beads prior to the AQC modification (Figure 3(d)). Recovery was performed by spiking in 10 nM serum and 5 nM serum and the results are shown in Table 1.

Figure 3.

C-peptide signal is increased after AQC modification. (a) The intensity of C-peptide is increased in those samples which were modified with AQC without enrichment. (b) 10-min incubation in heat block exhibited the highest intention among different time interventions (p < 0.05). (c) Mass spectrum of intensity of C-peptide with/without AQC modification. (d) The intensity of C-peptide is increased after enrichment with magnetic beads and AQC modification.

AQC: 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate.

Table 1.

Spike and recovery of C-peptide in human serum.

C-peptide concentration (nM)	Observed concentration (nM)	Mean recovery (%)
0	0	0
5	0.2953/0.7589	0.39
10	0.4567/0.6965	0.66

Precision of the method

Precision was assessed by calculating the coefficient of variation (CV%) for three levels of C-peptide concentrations. The measurements were repeated six times over 20 days. The intra- and inter-day results were highly reproducible (CV < 10%) (Table 2).

Table 2.

Precision of serum in C-peptide measurements.

Concentration (observed, nM)	Relative standard deviation (RSD, %)
	Within-days	Between-days
0.1	3.40	5.48
1	2.82	2.97
5	4.98	4.98

Limit of detection, sample type, linearity and range of the method

The LOD of this method was 0.025 nmol/L (S/N = 10.4), while the LOQ of this method was estimated to be approximately the same (0.025 nmol/L (S/N = 10.4); Figure 4(a)). The clinical reference range of C-peptide was covered using our detection method. C-peptide concentrations in different sample types (serum and plasma) were compared with 10 clinical samples. The C-peptide signal in serum was higher than that in plasma. Therefore, we settled on serum as the preferred sample type in our study (Figure 4(b)).

Figure 4.

Limit of detection, sample tape, linearity and range of the method. (a) The limit of detection of this method was 0.025 nmol/L (S/N = 10.4), while the limit of quantification of this method was estimated to be approximately the same (0.025 nmol/L [S/N = 10.4]). (b) Two sample types (serum and plasma) were compared with 10 clinical samples. C-peptide signal in serum is higher than plasma. (c,d) Calibration curve for two different volumes (100 μL and 500 μL) of samples.

Linearity was examined by adding C-peptide into artificial serum after immunoaffinity purification and AQC modification. The artificial serum was commercially purchased and confirmed by MALDI-TOF detection to verify that it did not contain any detectable C-peptide. The area of the C-peptide signal exhibited a good linearity over the 0.025–5 nmol/L range with a linear regression coefficient (r²) of 0.9732 and 0.9841 according to different volumes of sample (Figure 4(c) and (d)).

Method comparison

For method comparison, we completed an analysis of correlations between the MALDI-TOF MS assay and a reference assay using seven identified patient samples across the C-peptide concentration range of 60–5760pmol/L. Regression analysis revealed a good agreement between the new MALDI-TOF MS method we describe here and the reference assay (y = 0.38x + 0.52), a comparison between both sets of measurements showed a good linear correlation (r² = 0.9409) (Figure 5).

Figure 5.

Correlation about C-peptide detection using two methods in clinical samples (including six patient samples and one standard sample). Correlation: r² = 0.9409.

Discussion

Diabetes is a chronic disease affecting millions of people and a significant contributor to the annual death rate.²⁰ Currently, C-peptide levels, rather than that of insulin, are often assessed in newly diagnosed diabetes patients because insulin concentrations in the portal vein can range from 2 to 10 times higher than in the peripheral circulation. The liver extracts about half of the insulin from plasma, while C-peptide has a threefold lower clearance rate than insulin and can be determined in insulin-treated patients.^21,22 Thus, C-peptide is a more comprehensive marker of insulin status than direct insulin measurement. Because of this, a fast and inexpensive alternative to traditional C-peptide testing and monitoring techniques is desirable for effective diabetes mellitus treatment. MALDI-TOF MS is a powerful technique that provides both qualitative and quantitative determination of low-molecular weight compounds. This technology is being embraced by laboratories worldwide for its rapid, accurate, and inexpensive qualities. MALDI-TOF MS is an automated, high-throughput and incontrovertibly beneficial technology for the clinical laboratory.^23,24

In our study, we developed an MALDI-TOF MS-based assay with sufficient LOD/LOQ and specificity to quantify C-peptide concentrations in clinical serum samples. The method is based on enrichment with magnetic beads, followed by AQC modification, ZipTip desalting and concentration and MALDI-TOF MS analysis of the C-peptide. The assay consisted of the following steps: (1) an addition of anti-C-peptide antibodies to the magnetic bead suspension, (2) isolation of the C-peptide in serum samples using the anti-C-peptide antibody-conjugated magnetic beads, (3) removal of non-specific peptides by washing the beads in PBS, (4) elution of C-peptide from the magnetic beads, (5) centrifugation in a Speed Vac at room temperature, (6) chemical modification with AQC, (7) ZipTip processing, (8) MALDI-TOF MS analysis. The C-peptide is difficult to ionize, most likely due to its highly acidic nature.²⁵ Thus, we developed a method to increasing the ionization efficiency via sample preparation and sample modification. In previous work, ammonium sulfate, ammonia solutions and AQC were used to increase peptide detection, peptide signal-to-noise ratio, as well as improve the sensitivity of ionization.^26–28 In our investigation, we also found that the C-peptide signal could be enhanced after modification with ammonia solution, ammonium sulfate and AQC. In considering these options, sample preparation performed with AQC modification is a rapid, one-step procedure. The reaction between AQC and C-peptide forms a stable, asymmetric urea derivative and is relatively stable. This result is consistent with a previous study by Kinumi et al., where the authors employed liquid chromatography-mass spectrometry/MS (LC-MS/MS) to examine AQC-treated C-peptide.¹⁶ The difference between that study and ours is that native C-peptide signals can be detected using our method, while the latter detected peptide fragment ions derived from AQC, possibly because the two methods are better suited to the detection of the different compounds. Compared to LC-MS/MS, MALDI-TOF MS provides a relatively faster analysis and is less expensive. We have also improved detection by our experimental procedure: (1) The C-peptide signal was increased in PBST-diluted serum samples (five-fold increase). (2) The C-peptide and AQC concentrations and incubation times were optimized. According to the AQC kit instructions, the reaction is to be performed in a heat block at 100°C for 10min. (3) For the sample type, serum provided improved results over plasma. (4) Protein A magnetic beads afforded a better performance in comparison with protein G magnetic beads.

The novel method we describe here has many advantages over traditional methods. Compared to RIA and ELISA, it demonstrated superior inter-laboratory concordance of results and reproducibility, resulted in a decreased false positive rate, avoids the use of hazardous isotopes and is less costly.²⁹ In comparison with other C-peptide detection methods (LC-MS/MS),³⁰ it is cost effective, has a higher throughput (96 samples could be performed at the same time), easier to perform and offers a faster turnaround time (<4h). A good consistency of results was observed between LC-MS/MS methods and our new strategy for C-peptide detection.¹⁶ In future studies, we will improve the method further to decrease the LOD and LOQ. Multiplex detection for clinical laboratory application will also be considered in future investigations.

Conclusions

We successfully developed a time- and money-saving, high-throughput, immunoaffinity-based mass spectrometry assay, using AQC as a derivatizing agent for serum C-peptide quantification. Shorter derivatization times and simple derivatization procedures for sample preparation efficiently reduced the contamination risk and made results more reliable. This assay has proved to be a tractable and simple system for C-peptide quantification and has the potential for application in clinical practice.