Homologous ELISA for detection of prednisolone in human serum

ABSTRACT An enzyme-linked immunosorbent assay to measure prednisolone have been developed. An antibody was raised in rabbits using prednisolone-21-hemisuccinate (PSL-21-HS) conjugated to bovine serum albumin. Similarly, PSL-21-HS was conjugated to horseradish peroxidase to prepare enzyme conjugate. The developed assay has been validated for sensitivity and effective displacement at 50% of the assay were 0.078 and 2.64 ng/mL, respectively. This assay showed cross-reaction with only 4 steroids – i.e. progesterone 1.76%, 17α-OH progesterone 5.89%, cortisol 7.69% and prednisone 1.13%, out of 55 analogous steroids. The percent recovery of prednisolone from the exogenously spiked human serum pools was in the range of 94.84–100.17%. The intra-assay and inter-assay coefficients of variation ranged from 5.79% to 8.00% and from 3.23% to 8.63%, respectively. The serum prednisolone values obtained by this method correlated well with the commercially available kit and found to be 0.93.


Introduction
The prednisolone is a synthetic analogue of endogenous cortisol, having potent glucocorticoid and low mineralocorticoid activity, and is used in the treatment and management of a broad range of ailments, including severe asthma, rheumatic arthritis, gastrointestinal and hematological disorders (Fiel & Vincken, 2006;Thrower, 2009). Although the actions of this agent is largely palliative rather than curative, it remain the first-line option for many physicians in the treatment of disease (Schimmer & Parker, 2001). According to Indian Pharmacopoeia, in India it is officially permitted to use in treatment of Addison's disease, in organ transplant and drug-resistant diseases (Cushing's syndrome) (Tripathi, 2008). Glucocorticoid belongs to steroid family, particularly of pregnane class containing C-21 derivatives. Glucocorticoids have important functions upon carbohydrate, protein and calcium metabolism; it also possesses potent anti-inflammatory and immunosuppressive activities (Beotra et al., 2009;Deventer, Thuyne, Mikulcikova, Van ELISA is very common as a biochemical and clinical analytical method. There is no literature available for direct estimation of prednisolone in serum by enzyme-linked immunosorbent assay (ELISA). Although commercial ELISA kits are available from different companies for the measurement of prednisolone in serum, but their know-how is not published. Enzyme immunoassays represent the most rapidly growing nonisotopic methods in the diagnostics industry. Hence, the present study was designed to develop an ELISA method for simple, rapid, direct, sensitive and accurate estimation of prednisolone in serum, using indigenously prepared antibody, PSL-21-HS-BSA-immunogen and PSL-21-HS-HRP-enzyme conjugate.

Materials and methods
The experimental protocols were approved by the Institutional Animal Ethics Committee (IAEC) of National Institute of Health and Family Welfare (NIHFW), New Delhi, India. All animal experimentation was performed in accordance with the guidelines of Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA), Government of India.

Preparation of PSL-21-HS-BSA immunogen
PSL-21-HS was coupled to BSA by using N-hydroxysuccinimide mediated carbodiimide reaction, previously reported with minor modification Wang et al., 2016). Five milligrams of PSL-21-HS was taken, and each of 200 μL of dimethyl formamide, dioxan and 100 μL of distilled water containing 10 mg of N-hydroxysuccinimide and 20 mg of 1-ethyl-3-(3-dimethyl-amino-propyl) carbodiimide-HCL were added into it. The reaction mixture was mixed and kept at 4°C for overnight. Following the overnight incubation, aqueous BSA solution (1 mg/0.5 mL) was added to the activated steroid reaction mixture and vortex-mixed and kept at 4°C for overnight. The resultant conjugate was dialyzed against distilled water for 3-4 changes at 4°C. Subsequently, the dialyzed PSL-21-HS-BSA was frozen at −20°C, lyophilized and stored at 4°C.

Immunization of rabbits and collection of antiserum
The polyclonal antibody was raised in New Zealand white rabbits against PSL-21-HS-BSA as per previously described procedure Wei, Zhao, Wang, & Wang, 2014). Briefly, PSL-21-HS-BSA (1 mg) was dissolved in 0.9% saline (0.5 mL) and emulsified with (0.5 mL) Freund's (complete) adjuvant. The prepared emulsion (250 μL) was injected intramuscularly in each limb of the rabbits. Primarily, the first five booster injections were prepared with Freund's complete adjuvant and given per week that was further followed by monthly booster injection prepared using Freund's incomplete adjuvant. Further, the rabbit was bled after 9-14 days after the booster dose injections. The blood was collected from rabbit and after incubation at room temperature, antiserum was collected by centrifugation at 1000 g for 15 min. The separated antiserum was stored at −30°C.
2.3.4. Collection of normal rabbit serum (NRS) and generation of anti-rabbit gamma globulin (ARGG) For the collection of normal rabbit serum, blood was withdrawn from the non-immunized New Zealand white rabbits. The collected blood was centrifuged and clear serum separated and stored as NRS at −30°C until use.
2.3.5. Preparation of PSL-21-HS-HRP-enzyme conjugate PSL-21-HS was conjugated to HRP using an active ester method with some modification (Shrivastav, 2003;Tao et al., 2014). Two hundred microlitres of each of Dioxan, DMF and water were added to 5 mg of PSL-21-HS, 10 mg NHS and 20 mg EDAC. The above mixture was vortex-mixed and kept for activation at 4°C for 24 h. Further, HRP solution (1 mg/mL) was added to activated PSL-21-HS solution and kept at 4°C for 24 h. The PSL-21-HS-HRP-enzyme conjugate was purified by passing through the pre-equilibrated (10 mM PBS containing 0.1% thimerosal) G-25 column. The purified brown-coloured fraction of the HRP-PSL-21-HS conjugate with enzyme activity was collected and then an equal volume of ethylene glycol was added along with 1% of ammonium sulphate, BSA and sucrose. The PSL-21-HS-HRP-enzyme conjugate was aliquoted and stored at −30°C for further use.
2.3.6. Checkerboard assay A checkerboard assay was performed to obtain the optimal dilutions of antibody and enzyme conjugate for the development of assay.
2.3.6.1. Coating of antibody to microtitre plates. The 96 well microtitre plates were coated using the immunobridge technique for primary antibody immobilization according to the method of Shrivastav et al. (Shrivastav, Chaube, Kariya, Kumari, et al., 2011;Shrivastav, Chaube, Kariya, Singh, et al., 2012;Shrivastav, Kariya, Prasad, Chaube, & Kumar, 2014). Briefly, 250 μL of normal rabbit serum (NRS) diluted (1:250) in water was dispensed into each well and incubated at 37°C overnight. Following incubation, the plate was washed two to three times using washing solution (buffer E). To the NRS coated wells, 250 μL of 1:1000 diluted goats anti-rabbit gamma globulin (ARGG) was added and incubated for 5 h at 37°C or overnight at 4°C. Thereafter the plate was washed thoroughly using washing solution (buffer E). The raised antiserum against immunogen (PSL-21-HS-BSA) was serially diluted in antibody dilution buffer to 1:500, 1:1000, 1:2000, 1:4000 and 150 μL was added per ARGG coated wells (single dilution per 8 well strip); for nonspecific binding (NSB), 150 μL of antibody dilution buffer was added in a separate ARGG coated 8-well strip and incubated for 2 h at 37°C. Unabsorbed antibody was then washed with buffer E and 250 μL of buffer "D" was added to block the unoccupied sites of the plate. The plate was kept at room temperature (RT) for 1 h. The contents were decanted and the plate was dried at RT; packed and stored at 4°C in zip-lock bags for further use.
2.3.6.2. Determination of optimal loading of primary antibody using prednisolone-21-HS-HRP conjugate. In order to determine the optimum dilution of antibody and enzyme conjugates required for assay development, 100 μL of serially diluted (1:500, 1:1000, 1:2000 and 1:4000) enzyme conjugate (PSL-21-HS-HRP) was added in the above coated wells (one dilution per two wells vertically) and kept at 37°C for 1 h. Thereafter, the plate was decanted and washed two to three times using washing solution (buffer E). Further, to measure the bound enzyme activity (which is a direct function of the antibody), 100 μL of substrate solution TMB/H 2 O 2 was added to each well and incubated for 15 min in dark at RT. 100 μL of 1 N HCL was used to stop the reaction and the developed colour was measured at 450 nm using Tecan Spectra micro-plate reader (TECAN, Austria). The dilutions of antiserum and enzyme conjugate combination showing maximum zero binding and least nonspecific binding were selected for assay development.

Preparation of prednisolone standard
Eight prednisolone working standards (0.0, 0.62, 1.25, 2.5, 5.0, 10.0, 20.0 and 40.0 ng/mL) were prepared in stripped pooled serum. Steroid stripped serum was prepared by charcoal treatment 100 mL of pooled human serum was taken and to it 5% charcoal was added and stirred for 2 h at RT. This solution was centrifuged at 10,000×g for 15 min to remove the charcoal and then filtered using 0.45 μM membrane filter. Add one pinch of thimerosal as a preservative.
2.3.8. Assay procedure for measurement of PSL in serum or PSL-21-HS-BSA conjugate One hundred microlitres of PSL standard of different concentrations were added followed by serum samples or diluted PSL-21-HS-BSA conjugate to the wells coated with optimal dilution of PSL-21-HS-BSA-antibody. To each well, 100 μL of optimal dilution of enzyme conjugate (PSL-21-HS-HRP) was added and incubated for 1 h at 37°C. Thereafter the plate was decanted, and washed with buffer E. Further, 100 μL of TMB/H 2 O 2 substrate was added to each wells and incubated for 15 min at RT, to estimate the bound enzyme activity. Finally, 100 μL of 1 N HCL was added to each well. The colour intensity was measured at 450 nm using a micro-plate reader. The concentration of PSL in serum sample or attached to BSA was obtained by interpolation of their respective OD from standard curve obtained by plotting OD at Yaxis and standard on X-axis.

Preparation of recovery pools
Six recovery pools were prepared in serum by spiking the serum with different known concentrations of prednisolone, viz., 0.0, 0.62, 2.5, 5.0, 15.0 and 25.0 ng/mL.

Preparation of standard curve and determination of affinity constant and sensitivity
Standard curve was plotted by software Graph Pad Prism 6.0 and Microsoft Excel programme. The concentration and the percent ratio of mean absorbance of standard and mean absorbance at zero dose (A/A 0 × 100) were plotted on X-axis (log-scale) and Yaxis, respectively. A QBASIC language-based program developed in house has been used to compute the values of unknown samples that uses the logit-log regression method . The affinity constant of PSL antibody and sensitivity of the developed assay were calculated by the Feldman and Rodbard method (Shrivastav, Chaube, Kariya, Singh, et al., 2012).

Recovery
Recovery is often confused with accuracy although these two concepts are quite separate. Recovery is an indirect assessment of accuracy. Recovery assesses more than just whether or not an assay is correctly calibrated. It is the ability of a test to recover, or measure, a known incremental amount of an analyte from a sample matrix. The experimental protocol for a study includes adding a known amount of analyte (A) to a base (B) and measuring the concentration (C) (Shrivastav et al., 2014). The percentage of recovery can be calculated by using the formula: (C − B)/A * 100.

Precision
Precision is probably the most important technical aspect of an immunoassay performance. Precision, also sometimes referred to as reproducibility, is a statistical measure of the variation between repeated determinations on the same sample, either within the same run or from day to day, i.e. between runs. Imprecision is the opposite of precision.
Serum pools containing various prednisolone concentrations (i.e. high, high-medium, medium, low-medium and low) were used for measuring the level of imprecision in the standardized assay by quantitating prednisolone concentrations, in each serum pool by repeating the assay eight times and in eight different assays. The statistics conventionally used to express the precision profile of an assay are the mean (X ), standard deviation (SD) and the coefficient of variation, CV. The % coefficient of variation is calculated as % CV = (SD/X ) * 100 at a particular analyte level (Shrivastav et al., 2014).

Method comparison (correlation)
A newly developed immunoassay method should produce results similar to those of other reliable, clinically validated immunoassays or methods of assay. A graph of the two methods plotted against one another with identical axes, along with the bisecting line of perfect agreement, allows a good visual assessment of the degree of agreement between the two methods. Simple linear regression analysis is widely used to estimate the Pearson correlation coefficient (R²) (Chen et al., 2017).

Statistical analysis
Statistical analysis, including average, standard deviations (SD), logit-log transformation and correlation coefficient, was calculated using Microsoft Excel and graph was plotted by OriginPro 2016 (64) Bit software.

Characterizations of immunogen by UV and visible spectroscopic methods
The UV spectrum of BSA gave single peak at 280 nm, which is a characteristic peak of proteins due to amino acid tryptophan and tyrosine. However, PSL, PSL-21-HS and PSL-21-HS-BSA (after conjugation) showed three peaks at 242, 243 and 254 nm respectively as given in Figure 1, which represents the characteristic peak of hapten (PSL, PSL-21-HS) and hapten-protein (PSL-21-HS-BSA) conjugate.

Analysis of the hapten density in immunogen by ELISA
The amount of PSL-21-HS-BSA conjugated in immunogen was found to be 16 μg of prednisolone per mg of immunogen by ELISA.

Standard curve
The composite dose-response curve of prednisolone ELISA using PSL-21-HS-BSA antiserum and PSL-21-HS-HRP-enzyme conjugate has been shown in Figure 2. The standard curve obtained after repetition of several assays remained precise and stable. The CV % of the developed ELISA, for the A/A 0 ratio of each standard ranged from 2.56% to 9.93%. The logit-log transformation of standard curve in which y = −1.845x + 0.759 ng/mL and R² = 1. The affinity constant (Ka) of the PSL-21-HS-BSA antibody for prednisolone antigen was found to be 6.29 × 10 −8 (L/mol) calculated by a Scatchard plot. The slope and intercept of the curves were calculated by logit-log transformation of standard curve data, as shown in Table 1.

Sensitivity
Sensitivity is the capability to measure the minimum quantity of target analyte under defined standard condition (Shrivastav et al., 2014). The sensitivity of the assay is expressed in terms of its lowest detection dose (LDD) and the effective dose at 50% (ED 50 ). The LDD is the lowest concentration of analyte (A) that responds to statistically different from that detected in the lack of analyte (A 0 ). It is calculated as A 0 − 2× SD after 32-times determination of A 0 . The ED 50 is the effective concentration at which 50% of inhibition in the binding of enzyme conjugates occurs in an assay in the presence of analyte. It is calculated as ED 50 ± SD, after eight times determination of ED 50 . The values obtained were interpolated from the standard curve, and the concentration found in the low detection dose and the ED 50 of the developed assay were 0.078 and 2.64 ng/mL, respectively.

Recovery
The capability of the assay to accurately measure PSL in serum samples was verified. Six allocates of pooled serum were prepared by externally adding different concentrations of PSL. After spiking, the concentration of PSL was measured and recovery was calculated for each pool. The recovery ranged from 94.84% to 100.17%, as shown in Table 3.

Precision
The serum samples having roughly the similar concentrations of PSL were pooled and five different pools were prepared containing different concentrations of PSL. Each pool was examined eight times within the assay (intra-assay) and between separate assays (interassay) ( Table 4). The intra-assay and inter-assay CV% (n = 8, replicate of each pool) was <7.22%. The intra-assay coefficients of variation range from 1.39% to 6.15%. The inter-assay coefficients of variation range from 1.58% to 7.22%.

Correlation coefficient
The newly developed technique must yield results analogous to those of the already available conventional method. The correlation coefficient for values of prednisolone in serum samples (n = 64) measured by developed ELISA and commercial prednisolone ELISA kit catalogue no. E13651458 (purchased from Sincere Biotech Co. Ltd., Beijing, China) was found to be r 2 = 0.93. The regression analysis has been performed in which both X and Y was subject for measurement error. Figure 3 showed that method fit a straight line to a two-dimensional data where both the variables, X and Y, are measured with errors that can accommodate differences in measurement error between the test and the reference method. The linear regression curve of the correlated data was plotted by Graph Pad Prism version 6.0 for MS windows.

Estimation of PSL in patient treated for different diseases
We determined the PSL in serum of patient taking treatment through oral, local or intramuscular routes in different diseases such as chronic asthma (17 male and 16 female), Table 2. Cross-reactivity of steroid compounds with Prednisolone in homologous and bridge heterologous assays of Prednisolone using PSL-21-HS-BSA-antibody with PSL-21-HS-HRP-enzyme conjugates by ELISA. rheumatoid arthritis (18 male and 14 female), allergic (22 male and 13 female) and in healthy (38 male, 22 female) age group between 10 and 72 years (Table 5).

Discussion
In a homologous immunoassay, haptens used to generate the antibody and those used to prepare enzyme conjugate are the same. In the present study, the PSL-21-HS was used to couple to BSA and HRP so as to prepare immunogen and enzyme conjugate, respectively.   Syedanaglbashl et al. (1980) developed RIA for PSL estimation using the antiserum generated in rabbits to prednisolone-21-hemisuccinate-bovine serum albumin and [ 3 H]-prednisolone as tracer, where the detection limits were 2 ng/mL (Syedanaglbashl et al., 1980). Although, RIA has been the standard method for the measurement of steroids and other hormones in biological fluid and tissues but it has a number of serious disadvantages that restrict its applicability. Apart from this, the availability of radiolabeled steroids from commercial sources is very limited and severely restricts the range of steroids that can be measured.
Nawab Sher et al. (2016) utilized the HPLC method using a C18 analytical column as stationary phase. The mobile phase was 30:70 methanol: pH 2.5 phosphate buffer at a flow rate of 1.0 mL/min with absorbance detection at 235 nm. The method was linear for concentrations ranged from 40 to 10,000 ng/mL. However, the retention time of prednisolone was 10 min. The LOD, LOQ and coefficient of correlation (r 2 ) were 2.11, 6.39 ng/mL and 0.9995, respectively, for PSL in human serum (Sher et al., 2016). Jusko et al. (1994) quantified prednisolone in human plasma by high-performance liquid chromatography (HPLC) with ultraviolet (UV) at 254 nm, where the detection limit was 10 ng/mL (Jusko et al., 1994). Similarly Doppenschmitt et al. (1995) determined prednisolone levels using HPLC-UV in blood (serum) and the LOD was 2 ng/mL, where samples were extracted from 1.0 mL serum with 3 mL (1:1 v/v) ethyl acetate/tert-methyl butyl ether and 0.1 mL phosphoric acid (Doppenschmitt et al., 1995). Cho et al. (2003) used reversed-phase high-performance liquid chromatography (RP-HPLC-UV) for estimation of prednisolone in blood (plasma) samples. The intra and inter-assay variabilities ranged from 1.8% to 10.5% and from 0.7% to 9.5%, respectively, for PSL. The LOD, LOQ and coefficient of correlation (r 2 ) were 0.5, 2 ng/mL and 0.998, respectively, for PSL (Cho et al., 2003). In addition, Kurakula et al. (2011) also developed RP-HPLC method for detection of prednisolone in proliposomal formulation. However, the retention time of prednisolone was 6.4 min. The LOD and LOQ were 35 and 62.5 ng/mL, respectively. The linearity of the drug was in the range of 1-5 μg/mL with coefficient of correlation r 2 = 0.999. The % recovery of the PSL was 100.64% (Kurakula et al., 2011). Brett et al. (2010 describe an ultra-high-performance liquid chromatography-tandem mass spectrometry (UHPLC-MS/MS) method suitable for a laboratory to determine endogenous and exogenous glucocorticoids (PSL) in plasma, plasma ultra-filtrate, urine and saliva in a single analytical run. The limit of quantitation and CV% for PSL were 5.0 nmol/L and 7.6%, respectively (McWhinney et al., 2010). Shobha et al. (2012) estimated prednisolone using HPLC (+) Electrospray ionization (ESI) -MS/MS method. The concentrations of calibration standard were 15-120 ng/mL. The inter CV was within ±15% of the actual value. The coefficient of correlation was r 2 = 0.99 and LOD was 15 ng/ mL (Shobha et al., 2012). Leporati et al. (2013) and Vincenti et al. (2012) expanded and revalidated existing quantitative LC-MS/MS method for the detection of PSL in beef cattle (Leporati et al., 2013) and cow urine (Vincenti et al., 2012). The LOD of PSL in beef cattle and in cow urine was 0.42 and 0.5 ng/mL, respectively. Shibasakia et al. (2008) estimated plasma concentrations of prednisolone, prednisone, cortisol and cortisone simultaneously using gas chromatography-mass spectrometry (GC-MS). The LOD was 114.0 ng/mL for prednisolone (Shibasakia et al., 2008). Amendola et al. (2003) also determined prednisolone in urine by (GC-MS) with electron impact ionization, where limit of detection was 4 ng/mL, but it also requires additional preparation such as extraction, outgassing and derivatization, etc. and therefore it is time-consuming (Amendola et al., 2003).
The present study showed that PSL-21-HS-based homologous assay was more sensitive and very simple than others available system for PSL detection counterpart. The sensitivity and the ED 50 of the developed homologous assay was 0.078 and 2.64 ng/mL, respectively. This developed assay is specific for PSL estimation and it's showed less cross-reaction with commercially available fifty five steroids such as C 27 , C 22 , C 21 , C 19 and C 18 with analogous structure. This assay showed cross-reaction with only four steroids; i.e. progesterone 1.76%, 17α-OH progesterone 5.89%, cortisol 7.69% and prednisone 1.13% out of 55 analogous C18, C19, C21 and C27 steroids. The percent recovery of prednisolone from the exogenously spiked human serum pools was in the range of 94.84-100.17%. The intra-assay and inter-assay CV% was <7.22%. The correlation coefficient for values of prednisolone in serum samples (n = 64) measured by developed ELISA and commercially available assay was found to be r 2 = 0.93. Also we determined the PSL in patient receiving PSL medication under different diseases by using developed ELISA. Serum samples from healthy volunteers (38 male, 22 female) and in patient taking treatment with prednisolone drug through oral, locally or injection in diseases such as chronic asthma (17 male and 16 female), acute rheumatic (18 male and 14 female) and allergic (22 male and 13 female) having age group between 10 and 72 years. The developed assay is sensitive, specific, precise and requires an assay time of 1 h and 15 min only to complete; so it is rapid when compared to others methods.

Conclusion
A sensitive, specific, reproducible direct homologous assay for prednisolone quantitation has been developed, which require low volume of serum sample as well as less time to perform the assay. The analytical tool developed for detection of prednisolone may be extended for its measurement as drug of abuse in an athlete sports person, horse and meat of veal calves, bulls and old cows.