Introduction

Human C-peptide (3021 Daltons), composed of 31 amino acids, a by-product of insulin biosynthesis is produced by a series of enzymatic cleavages of the precursor molecules preproinsulin and proinsulin [1]. Thereby, insulin and C-peptide are released in equimolar amounts (1:1 ratio) from β-cells of pancreas into the portal circulation [2]. Soon after their entry, both insulin and C-peptide are carried to the liver (prime target organ) wherein insulin is extracted and degraded [3]. Another vital property is that the half-life of C-peptide is two to five times higher as compared to insulin. As a result, C-peptide levels may be used as a stable indicator of insulin secretion and important marker of β-cell function [4, 5] and under conditions of beta cell culture (in vitro) [6]. Insulin measurement in diabetic patients treated with insulin may be complicated by anti-insulin antibodies and cross reactivity with partially degraded insulin, proinsulin and split forms of proinsulin [7]. Extensive research on C-peptide has been carried out over the past few decades and it was found that measurement of C-peptide alone is adequate to make accurate clinical predictions and also to change treatment modality. By measuring C-peptide levels, a wide gamut of beneficial properties has been reported include; differentiation of Type 1 and Type 2 diabetes. A study by Bakhtadze et al. [8] has shown that patients diagnosed with type 1 diabetes mellitus had significantly lower median fasting plasma C-peptide than patients with type 2 diabetes mellitus or diabetes mellitus that was unclassified. The median values were 0.73 ng/ml (0.24 nmol/l) for type 1 diabetes mellitus, 2.24 ng/ml (0.74 nmol/l) for type 2 diabetes mellitus, and 1.45 ng/ml (0.48 nmol/l) for unclassified diabetes mellitus. C-Peptide levels can also predict the occurrence of autoantibodies. Torn et al. [9] have shown that C-peptide levels predict the presence of autoantibodies like islet cell antibodies, glutamic acid decarboxylase antibodies or tyrosine phosphatase antibodies. C-peptide levels also aid in diagnosis of various hypoglycemic disorders including islet cell tumors and factitious hypoglycemia [10]. Levels of C-peptide were found to be elevated in some of the obese individuals with Type 2 diabetes mellitus indicating insulin resistance [11]. Immunoassays based on the solid phase separation are widely preferred due to their ease of separation despite of heavy antibody consumption. In sandwich assays, like immunoradiometric assay (IRMA) solid phase separation is essential.

IRMA usually makes use of two monoclonal antibodies raised against the same analyte, but directed towards two different epitopes. A polyclonal antibody can also replace one of these monoclonal antibodies. One of these antibody will be immobilized on the solid support extracts or picks up the antigen (analyte) molecule from the sample and hence called capture antibody. In order to detect and quantitate the amount of the analyte captured, the other antibody labelled with 125I, which is called as a detector antibody or tracer, is used. The development of IRMA further involves proper selection of capture and detector antibody pair, preparation of antibody coupled to solid support (immunosorbent), labelling of the antibody and assay optimization for desired sensitivity and range [12].

Among the available immunosorbents, single piece solid support like antibody coated plastic tubes are the most convenient, but large scale production of these coated tubes require sophisticated technology with automatic gadgets to ensure uniform coating of the antibody. Besides, this is expensive as large quantities of antibody along with good quality tubes are required. On the other hand, separation methods based on antibody coupled to magnetizable particles as solid phase (magnetizable immunosorbents) are relatively inexpensive and can serve as an attractive alternative to coated tubes, without compromising much upon simplicity of the assay system. Among the variety of magnetizable particles, magnetizable cellulose particles are often preferred as they are associated with low non specific binding (NSB) and higher stability of resulting immunosorbent besides being inexpensive [13]. The objective of the present study was to develop and validate a simple one step inclusive IRMA for C-peptide in human serum samples using magnetizable cellulose particles as solid support with special emphasis on cross matching of monoclonal antibody pairs and selection of a suitable pair for standardization of C-peptide IRMA.

Experimental

Materials

Bovine serum albumin (BSA; RIA Grade), C-peptide (Purity ≥ 95 % (HPLC)), 1,1′-Carbonyldiimidazole (CDI) and Sephadex G-75 were from Sigma Chemical Company, USA. Bovine serum was obtained from Himedia, India. A pair of monoclonal antibodies designated as A & B were obtained from Medix Biochemica, Finland, C & D from Fitzgerald Industries International, USA and E & F from Abcam, USA. Carrier free 125I as Sodium Iodide (Specific activity ~17 mCi/µg, radioactive concentration 100 mCi/ml) was obtained from BARC, India. Magnetizable cellulose particles were prepared in-house at BRIT Laboratories, Mumbai, India (Indian Patent No: 193445). Multiwell RIA gamma counter was from Stratec Biomedicals Systems, Germany and Single well gamma counter NaI(T) was from Electronic Corporation of India (ECIL), India. Paper electrophoresis system was from Biotech R&D Laboratories, India.

Preparation of detector antibody

Radioiodination of monoclonal antibody was performed with 125I–NaI using Chloramine-T oxidation method as described by Greenwood et al. [14] with minor modifications. Radiolabeling procedures were performed with attention to good radiation safety practice. In particular, radioiodinations were performed in a well-ventilated fume hood in a glass test tube (75 × 12 mm) as detailed in Table 1. At the termination of the iodination reactions aliquots were immediately removed for trichloroacetic acid (TCA) precipitation/paper electrophoresis and the reaction mixture was immediately purified by gel filtration. Purification of radiolabelled anti-C-peptide monoclonal antibody from free radioiodide was carried out by using Sephadex G-75 column (30 × 1 cm) pre-equilibrated with 0.05 Μ phosphate buffer and BSA. 1 ml fractions were collected in 1 ml of 1 % BSA in pre-marked tubes at a flow rate of 8 drops/min, and their radioactivity was monitored in a single well manual gamma scintillation counter calibrated for 125I with changed geometry. All the six monoclonal antibodies A–F were radioiodinated using the same protocol and radioiodinated antibodies are designated as A*–F*. 125I- anti-C-peptide monoclonal antibody fractions, were pooled and diluted in 0.05 M phosphate buffer, pH 7.4 containing 0.5 % BSA and stored in aliquots at −20 °C for evaluation in the assay.

Table 1 Radioiodination protocol for anti-C-peptide monoclonal antibody
Full size table

The radioiodination yield was estimated by paper electrophoresis and TCA precipitation of the reaction mixture prior to the purification. The radiochemical purity (RCP) of the purified and pooled fractions of radioiodinated anti-C-peptide monoclonal antibody was determined using paper electrophoresis and TCA precipitation. An aliquot of the reaction mixture and peak fractions obtained from the column were spotted on a 30 cm Whatman 3 mm chromatography paper. The electrophoresis was run for 1 h at 240 V using 0.025 M phosphate buffer, pH 7.4. At the end of the run, the paper was removed, dried and cut into 1 cm segments and measured for radioactivity. The electrophoresis pattern of the reaction mixture was used for calculating the percentage radioiodination yield while those of the purified fractions were used as a measure of the RCP. Radiochromatogram was plotted by counts against strip number. Radioiodination yield and RCP of tracer was evaluated. The free iodide moves towards the anode and labelled molecule remains at the point of application. TCA precipitation method involves addition of 50 μl of the reaction mixture/peak fractions and 50 μl of C-peptide free bovine serum followed by1 ml of 10 % TCA. The contents were centrifuged and radioactivity associated with pellet is measured to calculate percentage radioiodination in the case of reaction mixture, whereas purified fraction for the RC purity [15].

Preparation of capture antibody

Anti-C-peptide monoclonal antibody was coupled to in-house prepared magnetizable cellulose particles using CDI activation method using 900 µg of anti-C-peptide monoclonal antibody per g of magnetizable cellulose particles. 1 g of in-house prepared magnetizable cellulose particles were washed with 100 ml of double distilled water (DDW) followed by five washes with ~10 ml of acetone. The washed particles were suspended in 10 ml acetone containing 500 mg of CDI for activation with gentle shaking for 1 h at room temperature. After activation, particles were washed with acetone followed by 0.1 M bicarbonate buffer, pH 8.6 (bicarbonate buffer) and suspended in 10 ml of bicarbonate buffer, pH 8.6 containing 900 µg of anti-C-peptide monoclonal antibody. The total reaction volume was maintained at 20 ml. The reaction was allowed to proceed overnight with gentle shaking. Anti-C-peptide monoclonal antibody coupled magnetizable cellulose particles were washed with bicarbonate buffer, pH 8.6 and 0.1 M acetate buffer, pH 4 alternatively to block the residual free binding sites. Anti-C-peptide monoclonal antibody coupled particles was saturated with 0.05 M phosphate buffer, pH 7.4 containing 2 M glycine and 1 % BSA. Finally, the coupled magnetizable cellulose particles were washed and suspended in 0.05 M phosphate buffer, pH 7.4 containing 0.5 % BSA and 0.1 % Tween-20 (assay buffer) and stored at 4 °C until further use. All the six monoclonal antibodies A-F were coupled to magnetizable cellulose particles using the same protocol and monoclonal antibodies coupled to antibodies magnetizable cellulose particles are designated as M–A, M–B, M–C, M–D, M–E and M–F.

Preparation of standards

A total of six C-peptide standards 0.3, 1.2, 6, 12, 18 and 30 ng/ml containing 0.1 % sodium azide as a preservative were prepared in C-peptide free bovine serum using highly purified synthetic C-peptide in aseptic conditions. The standards were later calibrated against WHO Ref IRP 84/510 and stored at 4 °C after lyophilization.

Compatibility studies for identification of capture and detector antibody

All the 125I-anti-C-peptide monoclonal antibodies (A*–F*: detector antibodies or tracer) and all the six monoclonal antibodies coupled to magnetizable cellulose particles (M–A to M–F: capture antibodies) were used for this study. The strategy applied for the selection of process was to identify a detector antibody to match capture antibody or vice versa.

100 µl of capture antibodies in the form of anti-C-peptide monoclonal antibodies coupled to magnetizable cellulose particles were incubated with 100 µl C-peptide standards (0, 0.3 and 30 ng/ml) along with 100 µl of detector antibodies (~1.5 × 105 CPM) overnight with gentle stirring. At the end of the incubation, particles were washed with 1 ml of 0.05 M phosphate buffer, pH 7.4 containing 0.1 % Tween-20 (wash buffer) twice and measured for radioactivity in a multiwell automatic RIA gamma counter. Using best possible combination of antibodies in terms of NSB (binding with C-peptide standard of 0 ng/ml) and binding with C-peptide standard with highest concentration (30 ng/ml) IRMA procedure for C-peptide was developed and validated. Optimized assay procedure consists of simultaneous addition of all the reagents. 200 µl of sample or standards (0.3–30 ng/ml) in bovine serum were incubated with 100 µl of capture antibody coupled to magnetizable cellulose particles and 100 µl of tracer on a shaker for overnight at room temperature. At the end of incubation, the particles were washed twice with wash buffer using a magnetic separator and counted for 1 min in a multiwell automatic RIA gamma counter. A standard curve was constructed by plotting %B/T (Y axis) against C-peptide standard concentration (X axis).

Results and discussion

Radioiodination yield and RCP were determined by paper electrophoresis as well as by TCA precipitation. Radioiodination yield/specific activity observed in the case all monoclonal antibodies were almost similar with an average of specific activity of 10.65 µCi/µg. This corresponds to 0.71 125I atom per every antibody molecule. The average RCP of the fraction collected was found to be 92 % (Table 2). Specific activity and RC purity was also confirmed by TCA precipitation method with comparable results.

Table 2 Specific activity and RCP of 125I labelled monoclonal antibodies
Full size table

The paramagnetic magnetizable cellulose particles used in this work were prepared in-house by a novel method based on wet grinding of iron oxide and microcrystalline cellulose in a stirred ball [16]. Particle size distribution of the particles is depicted in Fig. 1. Average particle size was found to be 1.89 µm with cutoff size of 7.7 µm. Percentage on 2 µm size particle was found to be 55.4 and the specific surface area as measured by the same equipment was found to be 57,128 cm2/cm3. The process of wet grinding not only firmly binds magnetite and cellulose physically but also results in the formation of magnetizable cellulose with smaller particle size. High surface area of the particles resulted in uptake of antibody of >80 % which was determined during the preparation of magnetizable cellulose particles coupled antibody particles by measuring protein (antibody) concentration before and after coupling reaction. Antibody coupled to magnetizable cellulose particles stored at 4 °C retained immunoreactivity not less than 2 years. In-house prepared magnetizable particles also have been used for other radioimmunoassay (RIA) and IRMA procedures [17–20].

Fig. 1
figure 1

Particle size distribution of magnetizable cellulose particles

Full size image

As the development of highly sensitive and specific sandwich IRMA systems can be hampered by the unpredictability and inflexibility of monoclonal antibodies as sandwich partners all the 30 possible combinations of capture—detector antibody pairs were tried (Table 3). Out of 30 possible combinations, the performance of the each pair of the antibody in terms of NSB and maximum binding (binding with 30 ng/ml standard) differed considerably. Even though similar amount of antibodies were immobilized on magnetizable cellulose particles (900 µg of antibody per 1 g of particles) and using radioiodinated monoclonal antibodies of specific activity of around 11 µCi/µg, yet there was a significant difference in the performance of these antibodies as a pair. In the case of the best possible combination (M–C and D*), NSB and maximum binding observed were 0.2 and 34 % respectively. Results obtained in the case of rest of other combinations were poor with maximum binding as low as 0.26 % (M–B and C*). NSB as high as 1.01 % and maximum binding as low as 0.46 % was also observed in one of the combinations (M–E and F*). C-peptide standard of 0.3 ng/ml was also included in the study in order to assess the feasibility of sensitive assays. Results indicated that proper screening and careful selection of detector and capture antibody is necessary and the combination M–C and D* was used for further optimization and validation of the assay system. Different dilutions (1:2, 1:3, 1:4 and 1:5) of the selected capture antibody (M–C) using assay buffer were tested with its complementary detector antibody (D*) with maximum binding of 33, 33, 29 and 20 % respectively. Further drastic reduction in binding was observed when the particles were diluted beyond 1:5.

Table 3 Cross matching of different capture antibody diluted (1:1) with detector antibody
Full size table

The other important requirement for developing any assay is the production of high quality tracer. Addition of a single iodine atom to the biological molecules does not alter the function and property of the labeled compounds. The substituted iodine does not cause any conformational change to the structure of biomolecules [21]. Tracer with specific activity of around 11 µCi/µg was used throughout the experiments and in order to find out the optimal concentration of the tracer required, effect of concentration of tracer in terms of CPM was studied. Total CPM were varied from 5 × 104 to 3 × 105 and NSB, % B/T for the highest standard were calculated. Total CPM of 1.5 × 105 was found to be ideal with minimum NSB and high maximum binding. Radioiodination of anti-C-peptide monoclonal antibody used in this technique is simple than performing radioiodination for tyrosylated-C-peptide as required in case of RIA as C-peptide is devoid of tyrosine or histidine residue which is essential for radioiodination. Typical standard curve using optimized assay parameters are depicted in Fig. 2.

Fig. 2
figure 2

Typical standard curve

Full size image

In C-peptide immunoassays, limiting factors like interference of even minor cross reactivity with proinsulin may significantly interfere with the C-peptide immunoassay. Purification of C-peptide by HPLC [22] and precipitation by 25 % polyethylene glycol [23] prior to radioimmunoassay was also carried out by some research groups to remove cross reactivity with proinsulin. Treatment for diabetes by insulin therapy may produce antibodies against insulin which will lead to the retention of proinsulin resulting in concentrations high enough to interfere significantly in a C-peptide immunoassay [24]. To assess the extent of C-peptide specificity, we measured the cross reactivity with other related peptides/proteins like human Proinsulin, Insulin, Glucagon, Somatostatin and Pancreatic polypeptide in the assay at concentrations of 2 ng/ml, 200 µIU/ml, 200, 100 and 1,000 pg/ml respectively. However, apparent concentration of added cross-reactants were less than 0.01 % indicating the absence of cross reactivity (Table 4). Such low levels of cross reactivity mainly with proinsulin are essential to ensure accurate and reproducible C-peptide measurements [25].

Table 4 Specificity of anti-C-peptide monoclonal antibody
Full size table

High dose hook effect is always associated with sandwich assays and usually observed when capture antibody sites are saturated with the antigen as in the case of samples with high antigen concentration. This problem can be tackled to some extent by increasing the concentration of capture antibody. Use of capture antibody prepared with magnetizable cellulose particles with high surface area facilitated the increase of capture antibody concentration in the assay system hook effect was not observed up to 50 ng/ml (Fig. 3). In addition, unlike antibody coated tubes or microliter plates, capture antibody prepared using magnetizable cellulose particles have the flexibility in varying their concentration as per the requirement.

Fig. 3
figure 3

High dose hook effect

Full size image

Developed assay system was validated using standard assay parameters. Sensitivity of the assay determined by assaying 15 replicates of the zero standard and calculating the concentration corresponding to the sum of the mean CPM and its two standard deviation was found to be 0.16 ng/ml. Sensitivity of the developed method is quite higher compared to radioimmunoassay reported in other studies [26, 27]. The intra-assay variation was determined by analyzing two serum samples 15 times in the same assay. Similarly, inter-assay variation was determined by analyzing the same two serum samples in ten different assays (each sample in duplicate) run over a period of a time (1 month) (Table 5). For estimation of recovery, serum sample containing C-peptide were spiked with a known amount of C-peptide standards, both the native sample and spiked sample were analyzed in the same assay. % Recovery ranged from 88 to 105 % which was well within acceptable limits of 80–110 % in immunoassays (Table 6). C-peptide samples were diluted twofold, fourfold and eightfold with zero standard. Both native samples as well as diluted samples were analyzed in the assay. Percentage recovery for dilution parallelism ranged from 93 to 116 % (Table 7).

Table 5 Assay variation
Full size table
Table 6 Recovery test
Full size table
Table 7 Linearity test
Full size table

Around 55 clinical samples were analyzed using developed method. C-peptide values obtained were compared with those values obtained by an established commercial kit (IZOTOP, Hungary) with acceptable correlation of (Y = 0.9242X + 0.81 ng/ml; r = 0.95) (Fig. 4).

Fig. 4
figure 4

Clinical sample analysis

Full size image

Conclusion

Proper screening and careful selection of detector and capture antibody is necessary in development of an IRMA system as shown in the present study as different antibody pairs behaved differently as sandwich partners affecting overall assay parameters. By using magnetizable immunosorbent prepared with in-house magnetizable cellulose particles an assay procedure with a sensitivity of 0.16 ng/ml can be used to assess the course, prognosis and management of diabetic disorders. Assay procedure can be compared with assays based on antibody coated tubes with an added advantage of economy and flexibility in the use of solid phase reagent.