Introduction

The measurement of non-protein bound concentration of thyroid and steroid hormones have emerged as a standard diagnostic procedure in many clinical laboratories. It is being widely accepted that the free hormone concentration measured under equilibrium conditions in vitro determines the hormones physiological activity in-vivo [1].

Development of methods to measure free hormones started in 1970 s mainly by the pioneering work of Ekins and colleagues [2, 3]. Determination of free hormone is particularly challenging as the assay must detect very low concentrations of free hormone relative to the vast excess of protein bound fraction [4]. Indirect methods combine total hormone concentration and free hormone fraction while direct methods yield a free hormone estimate from a single determination without the need for the total hormone concentration [5].

Absolute Direct free T4 methods currently serve mainly as reference methods. It involves physical separation of the free hormone from the bound by equilibrium dialysis or ultrafiltration and quantification of the free fraction by radioimmunoassay or chromatography. These assays are technically demanding, relatively expensive, time consuming and hence unsuited for routine laboratory practices [6]. Novel high performance liquid chromatography - tandem mass spectrometry includes small sample size, simultaneous measurement of many analytes, and enhanced specificity compared to immunoassay methods. Mass spectrometric methods are still fairly labor intensive, and certainly require a higher level of laboratory expertise than do immunoassays [7].

Comparative methods include two step immunoassay and one step analogue immunoassays. To avoid the inconvenience of two separate steps for the measurement of free hormones one step analogue assays were developed [8]. In these assays chemically modified hormone analogues were employed which can avidly bind to the specific antibodies but much less avidly to the serum binding proteins than the native hormone. However, interaction of thyroid hormone analogues with albumin is the source of bias encountered. Free hormone assays are designed such that the equilibrium existing between the hormone and the binding proteins is not disturbed. Hence the presence of factors affecting the equilibrium will also affect the hormone measurement, which include presence of heparin or displacing agents, anti-iodothyronine antibodies, heterophilic antibodies and variant thyroid hormone binding proteins. The two-step assay method with a washing step prior to tracer addition may reduce such interferences due to serum components [9].

Currently, many laboratories worldwide measure FT4 and FT3 levels directly in serum but several diagnostics avoid this direct measurement due to the limitation associated with accuracy and validity of the methods [10]. Use of adequate affinity antibody, controlling the concentration of antibody, avoiding the use of BSA and carrying out the assay in two steps under physiological temperature and pH can yield reliable estimates of free hormones [11].

Even though two-step approach is valid in comparison with the one step analogue assays there are two major limitations associated with it. Firstly, it necessitates removal the traces of serum proteins before the second incubation. Even very minute amounts of serum proteins can sequester the labelled hormone probe added in the second step due to its high affinity. And secondly the possibility of back displacement of the hormone bound to the antibody during the first incubation by labelled probe in the second incubation [12]. These two aspects have been taken care without modifying the labelled probe as in some analogue assays [13]. No additives such as ANS is incorporated to prevent the binding of tracer to binding proteins and also BSA was avoided from the system as it affects the assay performance [14, 15].

Materials and methods

Star bottom polystyrene tubes were procured from Greiner Bio One, Germany; normal rabbit gamma globulin from SIGMA; goat anti-rabbit antibody procured from Genei, Bangalore, India; rabbit anti T3 antibody were raised in-house; bovine serum albumin from SIGMA; thiomersol from John Baker Inc., USA; tris salt form S.D. Fine Chemicals, India; radioactive 125I from BARC, India; gelatin from Himedia, India; sephadex G-25 from G.E Healthcare, Sweden; triiodothyronine from Aldrich Chem. Co, USA.

Buffers used for the immobilisation of antibody are as follows:

  • Coating buffer: 0.05 M carbonate buffer pH 9.2, coupling buffer: 0.15 M tris-HCl buffer pH 8.25 with 0.1% BSA and 0.02% thiomersol, wash buffer: 0.01 M tris-HCl buffer pH 8.25, saturation buffer: 0.1 M carbonate buffer with 0.05% sodium azide and 0.2 M glycine.

  • Assay buffer: 0.14 M tris-saline buffer pH 7.4 with 0.1% azide.

  • Wash buffer used in the assay system: 0.035 M Tris-saline buffer pH 7.4 with 0.125% tween-20.

125I labelled triiodothyronine: 125I labelled triiodothyronine of high specific activity ~ 3000µCi/ µg was prepared by radioiodination of diiodothyronine by chloramine-T oxidation method and purified over Sephadex G-25 column [16,17,18]. Purified fraction was diluted to yield around 1.25 kBq /0.5mL in tris gelatin buffer pH 7.4 corresponds to 11–12 pg of labelled T3 /tube.

Free triiodothyronine standards: Gravimetrically weighted T3 was appropriately diluted in hormone free human serum to obtain standards of T3 concentration ranging from 0 to 9.6ng/mL. The free T3 concentrations of the resultant standards were assigned using CIS BIO FT3 RIA kit. The values estimated were 0, 1.0, 2.1, 4.1, 7.6, 14.2 and 29.8pg/mL for a typical batch analysed in replicates of five in three different assays with CV of measurement less than 5%.

Anti T3 antibody coated tubes:

Star bottom polystyrene tubes from Greiner Bio One was used for immobilizing anti-T3 antibody using the pre-optimized [19, 20] following general protocol.

Step 1: In each tube, 0.5 mL of 1 mg/mL concentration of rabbit gamma globulin (RIgG) in 0.05 M carbonate buffer was added. The tubes were incubated at room temperature for 22 h.

Step 2: The tubes were washed twice with 4 mL wash buffer, 0.5 mL of optimised (1:500 dilution) concentration of goat anti rabbit antibody (secondary antibody) in coupling buffer was added in each tube and incubated at room temperature for 22 h.

Step 3: The tubes were washed once with 4 mL wash buffer, 0.5 mL of optimised concentration of anti T3 antibody (1:175,000 dilution) in coupling buffer was added in each tube and incubated at room temperature for 22 h.

Step 4: The tubes were washed twice with 4mL wash buffer. 1.25 mL of saturation buffer was added to each tube and incubated for 1.5 h, the tubes were then washed twice with wash buffer, drip dried and stored at 4 °C.

Critical parameters studied in Optimisation of assay protocol.

The important aspect of < 1% extraction efficiency (%EE) by the antibody coated tubes from normal human serum was optimised by adjusting the antibody concentration, sample volume, reaction volume and reaction time using normal human serum spiked with 125I labelled T3.

Effect of antibody concentration on %EE: Tubes were coated with various dilutions of anti T3 antibody while keeping the RIgG and secondary antibody concentrations fixed. These tubes of various concentrations were incubated with 100 µL of spiked normal human serum and 900 µL of assay buffer for 30 min at 37 °C. At the end of the reaction contents decanted, tubes washed twice with 1mL wash buffer and bound activity measured. Table 1 shows the effect of varying dilutions of antibody on %EE.

Effect of sample volume on %EE was studied by incubating the antibody coated tubes with 1:1,75,000 dilution at 37 °C for 30 min. with varying volumes of spiked normal human serum and reaction volume adjusted to 1mL. After 30 min. Incubation contents decanted, tubes were washed twice with wash buffer and radioactivity was measured. Table 2 shows the effect of varying sample volumes on %EE.

Effect of reaction time on %EE was studied by incubating antibody coated tubes of 1:175,000 titre with 200 µL spiked normal human serum with reaction voume of 1mL at 37 °C. Reaction was stopped by decanting and washing twice with 1mL wash buffer at various time intervals and radioactivity measured in Gamma counter as shown in Table 3.

Effect of reaction volume on %EE was assessed by incubating antibody coated tubes with 200 µL spiked serum and volume adjusted to 500, 1000 and 1200 µL using assay buffer, incubated for 20 min at 37 °C, contents decanted, washed twice with wash buffer and activity measured. Table 4 shows the change in %EE with different reaction volumes.

Optimisation of second incubation condition was done by studying the reaction kinetics of antibody coated tubes at room temperature and at 4 °C. Based on the above study 2 h incubation with tracer was selected and assays were carried out at room temperature and at 4 °C. Figure 1 shows the difference in the slope of the curve at two different temperatures.

Optimised assay procedure: Based on the above study following assay protocol is optimised. 200µL sample or standard with 1mL assay buffer in anti-T3 antibody coated tubes were incubated for 20 min at 37 °C, contents decanted washed twice with 1mL wash buffer each time, decanted thoroughly, incubated with 0.5 mL radiolabelled triiodothyronine at 4 °C for 2 h. At the end of the second incubation 1mL wash buffer was added, contents vortexed, decanted and radioactivity measured in a gamma counter. Figure 2 shows a typical standard curve using average %B/Bo of 20 different assay runs.

Validation of the assay system: Assay parameters such as sensitivity, recovery, intra and inter assay precision were studied. FT3 measurements made were checked using External Quality Control sera obtained from Bio Rad, USA. Dilution test was carried out using hypo, hyper and euthyroid patient serum samples. The above samples were diluted serially from 1:2 to 1:16 using assay buffer and FT3 concentration of the diluted samples were measured along with undiluted serum samples.

Effect of varying the sample volume in the optimised assay: Varying volumes of two samples were analysed repeatedly with volumes ranging from 100 to 400µL as shown in Fig. 3.

Establishment of normal range: FT3 concentration defining the central 95% reference range was established using 74 healthy volunteer’s samples. Samples were classified as normal by measuring total T3, T4 and TSH concentrations.

Comparison with the commercial kit: FT3 concentrations of 55 clinical samples in the range of 0 to 4.2 pg/mL measured using the developed system was compared with an established commercial kit as shown in Fig. 4.

Results and Discussion:

An important aspect of antibody concentration for a valid free hormone measurement was optimised by selecting the minimum concentration of immobilised antibody which can extract ~ 1% of TT3 from normal human serum.

Table 1 Effect of varying antibody concentration on %EE
Full size table

It was observed that 1: 175,000 dilution of T3 antibody showed 1.7% EE and 25–30% assay binding. % EE was further reduced by adjusting the sample volume as shown in Table 2. Varying Sample volumes from 50 to 300 µL resulted in EE ranging from 3.8 to 0.76%. Incubation with 200 µL of spiked serum resulted in 1% EE.

Table 2 Effect of varying sample volume %EE
Full size table

Study on incubation time of serum with antibody coated tubes from 10 to 40 min showed %EE increasing from 0.6 to 1.8.

Table 3 Effect of reaction time on %EE
Full size table

1.2% at 20 min. was selected and further decreased by increasing the incubation volume to 1200 µL as shown in Table 4.

Table 4 Effect of reaction volume on %EE
Full size table

Based on the above results the first step of the assay was optimised. 200µL standard or sample with 1mL assay buffer was incubated at 37℃ for 20 min. contents decanted and washed twice with 1mL wash buffer ensuring the complete removal of serum components [20]. It is a critical step in two step RIA which removes serum binding proteins as well as interfering factors present in the serum, because of which it scores over one step analogue assays. Second incubation with 0.5mL labelled T3 was carried out at 4℃ and at ambient temperature. Dose response curves obtained at 4 °C showed significant improvement in sensitivity due to two reasons, (i) slow rate of dissociation at 4 °C hence the free fraction picked up by the antibody in the first incubation had not dissociated (ii) due to the slow rate of reaction tracer added in the second step occupied only the unbound sites on limited antibody which otherwise competes with the free fraction already bound to the antibody as shown in Fig. 1.

Fig. 1
figure 1

Effect of Second Incubation Temperature on Dose Response Curve

Full size image

The average minimum detection limit of the optimised assay obtained from 5 different assays was found to be 0.4 ± 0.16 pg/mL.

Analytical recoveries at three clinically useful region of 1.3, 2.8 and 4.4 pg/mL were 90.3, 102 and 92.5% respectively. Intra assay variation for 10 replicates at 2.8, 5.6 and 8.1 pg/mL were estimated to be 10.4%, 4.5 and 5.4% respectively. Inter assay variation for 8 different assays at FT3 concentrations 3.0, 6.8 and 10.5 pg/mL were estimated to be 7.3, 7.1 and 9.5% respectively.

Assessment of the assay system using external quality control serum (Bio-Rad controls) at two different levels in five different assay runs were estimated to be 2.3 ± 0.23 pg/mL and 6.7 ± 0.48 pg/mL for which the specified concentrations were 2.06 ± 0.64 pg/mL and 6.34 ± 2.9 pg/mL respectively.

Figure 2 shows %B/Bo at standard concentrations 1.5, 2.9, 5.5, 9.6, 20 and 37.6pg/mL for 20 different assay runs with one standard deviation of 4.2, 4.3, 3.8, 3.3, 2.7, and 2.5 respectively. The %CV for 20 different runs were estimated to be 4.8, 5.6, 5.9, 6.3, 6.2 and 6.7% respectively.

Fig. 2
figure 2

Average Dose Response Curve [n = 20]

Full size image

Dilution of hypo, normal and hyper serum samples to 1:2, 1:4, 1:8 and 1:16 with assay buffer resulted in actual dilution of 1:12, 1:24, 1:48, 1:96 of the samples during the reaction. It was observed that dilution of hypo and two normal samples of concentrations 1.5, 4.07, 4.7pg/mL yielded relatively same values with the %CV of measurement of 16.5%, 9.8%, 6.2%. However hyper sample with FT3 concentration of 9.9pg/mL showed a gradual decrease in concentration with dilution and was only 63% of the undiluted serum value at 1:16 times dilution owing to the decreased binding protein concentration associated with the hyper serum samples as shown in Table 5.

Table 5 Effect of sample dilution on FT3 measurement
Full size table

Repeated measurements of two samples of concentration around 3.1 and 3.35 pg/mL were estimated with increasing samples volumes ranging from 100 to 400µL with CV of measurement 5.5 and 5.2% respectively as shown in Fig. 3 further validates the optimised assay.

Fig. 3
figure 3

Effect of varying sample volume on FT3 measurement

Full size image

The central 95% euthyroid reference range established using 74 normal healthy volunteers was observed to be 1.6–4 pg/mL.

FT3 concentration of 55 clinical samples measured using the developed system was regressed against the values obtained from an established commercial kit as shown in Fig. 4. The following relation was obtained for FT3 values from the developed system(Y) regressed against the commercial kit values(X) for 55 clinical samples: Y = 0.9471X + 0.1241 (r = 0.94), SE = 0.285, slope = 0.947.

Fig. 4
figure 4

Correlation with the established commercial kit

Full size image

Conclusions

Optimised a reliable FT3 RIA procedure suitable for the routine laboratory analysis. Adequate attention was paid in selecting the concentration of the antibody, minimising the dilution of the sample during the assay, maintaining the extraction efficiency around 1%, avoiding the use of BSA in the system and carrying out the measurement at physiological conditions has resulted in a valid assay system.