Exp. Clin. Endocrinol. Endocrinol. Vol. Vol.95, 95,No. No.2,2,1990, 1990,pp. pp.229-238 229-238
J. A. Barth, Leipzig
Institute of Clinical Chemistry and Laboratory Diagnosis (Director: MR Prof. Dr. Sc. med. se. med. W. W. Rotzsch) Rotzsch) and and Department Department of of Pediatrics Pediatrics (Director: (Director: OMR OMR Prof. Prof. Dr. Dr. Sc. Sc. med. W. Braun), Karl-Marx-University Medical School Leipzig, Central Institute of Diabetes 'Gerhardt Katsch" (Director: OMR 0MB Prof. Prof.Dr. Dr.Sc. sc. med. H. Bibergeil) Karlsburg/GDR
for Measurement of Insulin on Microtiter Plates J. KRATZSCH, KRATZSCH,W. W.ACKERMAn.T, ACKRMAnÇ HHKILACKIR, KInLACJcJm, W.W. BESCH BEsdiland and E. E. KELLER KELLER With 5 Figures
Summary. A sensitive enzyme immunoassay for the measurement of insulin in human sera on microtiter plates was established. The assay is based on the sandwich technique with guinea pig anti-insulin IgG adsorbed adsorbed at at microtiter microtiter plate plate wells, wells, human humaninsulin insulinas asstandard standardand andthe thesame sameanti-insuanti-insulin TgG IgG labeled labeled with with horseradish horseradish peroxidase. peroxidase. Standards used cover a range from O to 1200 1200 pmol/1 pmol/l with a detection limit of 10 pmol/l. Coefficients of variation between 3-7% for intraassay precision and 5-11% for interassay precision were obtained over the concentration range of 80-1000 pmol/l. The correlation of ETA-data ETA-data with with those those of of aa commercially commerciallyavailable availabledouble doubleantibody antibodyradioimmunoradioimmunoassay (r = 0.98) could be expressed expressed by by the the equationS equation: ETA = 0.97 RTA RIA -- 57 57 pmol/l. pmol/1. Normal Normal fasting fasting serum insulin concentrations concentrations in in healthy healthy subjects subjects ranged rangedfrom from11-165 11-165pmol/1. pmol/l. In In subjects subjects with with poPo-
tentially diminished basal values concentrations of 10-79 pmol/l were determined. The insulin response response in in oral oral glucose glucose tolerance tolerance tests tests of of children children was was discussed, discussed, who who had had aa constitutional constitutional tall tall stature stature or Turner's syndrom, respectively.
Key K e ywords: words:Insulin Insulindetermination determination - Enzyme immunoassay - Microtiter plates - Serum insulin concentrations
Introduction
The determination of insulin secretion is a helpful parameter in characterizing a defect carbohydrate metabolism in patients with various metabolic disorders including diabetes mellitus (Rausch-Stroohmann, 1986). Since the first description of radioimmunoassay (RIA) for the measurement of insulin by Berson and Yalow (1959), RIA has been used most widely to assay that hor-
mone. In the last years significant improvements in reliability and practicability of the RIA measurement have been reported (Besch et al., 1987). However, in order to avoid disadvantages inherent with the application of radioisotopes, enzyme (Ishikawa, 1973) and fluorescence labels (Yamaguchi et al., 1982) have been introduced. Competitive assays are not capable of measuring the whole range of clinically encountered insulin concentrations with a single dilution of sample. The use of sandwichtype tests provides a wider working range, a better sensitivity and a lowered number of pitfalls because of matrix effects (Imagawa et al., 1983; Hoffmann, 1985).
Downloaded by: University of Pittsburgh. Copyrighted material.
A Sensitive Sandwich Enzyme Immunoassay
230
Exp. Clin. Endocrinol. 95 (1990) 2
To ensure a high degree of effectivity and automation in the laboratory, microtiter plates are a suitable bound/free separation system allowing the determination of hundreds of samples in a very short time. This paper describes a simple simple and and sensitive sensitive sandwich sandwich enzyme enzymeimmunoassay immunoassay(ETA) (EIA) for the determination of insulin in human sera in the absence of serum insulin-binding antibodies. The antigen capturing antibody is immobilized to the wells of microtiter plates, the second (detecting) antibody is labeled with horseradish peroxidase. The clinical
reliability of the assay is tested in the discrimination of insulin concentrations from fasting healthy adults and children with potentially diminished basal values (neonates, patients with Turner's syndrom). Besides, the insulin response is compared in oral glucose tolerance tests of children with constitutional tall stature or Turner's syndrom, Material and Methods Matørial Insulin-free Insulin - freeserum. serum.AAnormal normalserum serumpool poolwas wastreated treatedby bycharcoal charcoalextraction extractionand andfiltration, filtration, it was supplied by the VEB Schsisches Serumwerk Dresden (GDR).
Buffers and solutions. Coating Coating buffer buffer was was aa carbonate carbonatebuffer buffer(0.1 (0.1Mfl) M/l)pH pH9.6. 9.6.The Theincubation incubation buffer buffer contained contained 0.01 0.01 M M phosphate/I, phosphate/I, 0.15 0.15 M M NaC1/l NaCIfJ and and 1% 1% insulin-free insulin-free serum. serum. As As washing washing buffer buffer 0.01 Mfl MR phosphate-buffered isotonic saline with 0.05% Tween 80 was used. Substrate solution contained 0.1 M citrate/phosphate citiateJphosphate (pH 5.0), 0.8 g o-phenylendiamine (Fluka-AG, FRG) and 3.2 ml of 3% H202 per litre. Microtiter plates with 96 flat-bottomed wells were purchased from VEB Polyplast Halberstadt (GDR).
Insulin standards. Insulin-free Insulin-free serum serum was was loaded loaded with withhuman humaninsulin insulinstandard standardHT-8 III-8 (Kohnert et al., 1973) adjusted to WHO human insulin standard (lot 66/304). Standard concentrations ranged from from 1200 1200 to to 18.75 18.75 pmol pmol insulin insulin per per litre litre as as aa series series of of 1/2 1/2 dilutions. dilutions. Highly Highly purified purified porcine porcine (S (5 23267), rat rat (998 (998 611) 611) and and mouse mouse (M (M 20169) 20169) insulins; insulins; and anUporcine porcineproinsulin proinsulin(1670) (1670)were wereobtained obtainedfrom fromNOVO NOVO (Denmark). Bovine insulin insulin (lot (lot 104F-0740) 104F-0740) was was supplied suppliedby bySigma SigmaChem. Chem.Corp. Corp.(St. (St.Louis, Louis,USA). USA). Ovine insulin (1f7J74) (1/7/74) was a gift froni from "Central Institute of Diabetes". Diabetes".
Production and purification of antiserum. Antiserum 432 was generated in guinea pig with crystalline porcine insulin (Bosch (Besch et al., 1987). The antibody was purifieU purified by by means means of of precipita precipitation with 18% Na2SO4 Na2504 and and affinity affinity chromatography chromatography on Protein A-Sepharose (Pharmacia Fine Fine CheChemicals Uppsala, Sweden) to receive the IgG-fraction. Protein concentration was determined by the
method of Lowry, and aliquotes °C. aliquotes were were stored stored at at 20 20°C.
Preparation of the IgG-peroxidase conjugate. The purified IgG-fraction as IgG.fraction obtained a9 described above was coupled via the periodate-method, reported by Nakane and Kawaoi (1974) with horseradish-peroxidase (RZ 2.7; VEB Arzneimittelwerk Dresden, GDR). The original method was modified using a molar ratio IgG to peroxidase of 1:2. The reacting mixture was fractionated on a 30 30x1 x 1cm cmSephadex SephadexG-200 G-200(Pharmacia) (Pharmacia)column. column.Aliquotes Aliquotesof ofthe the conjugate conjugate were were stored stored at at
2020°C. °C. T es tprocedure. Test proc e du re. Microtiter Microtiterplates plateswere werecoated coatedovernight overnight with 4.6>< 4.6 x 10 10 mg guinea pig antiporcine insulin IgG per ml coating buffer (0.3 mlfwell). mI/well). After repeated rinsing with washing buffer the plate is ready to use and may be stored at 4°C after drying and sealing. To perform the assay, 0.2 ml of incubation buffer were deposited to each well. In addition, 0.05 ml samples or standards were pipetted. After two hours of incubation the plates were rinsed three times. 0.25 ml anti-insulin IgG-peroxidase in IgG-peroxid.ase in incubation incubation buffer buffer (1:1000) were added and allowed to incubate for for another another two two hours. Plates were rinsed again, 0.25 ml substrate solution was added and, after a 30 min incubation
Downloaded by: University of Pittsburgh. Copyrighted material.
respectively.
231
for Measurement of Insulin J. KRATZSCH et al., Sandwich-ETA Sandwich-EIA for
carried out in the dark, the reaction was stopped with 0.05 ml 5M H2SO4. All incubation steps were carried (NTJNC Roskilde, Roskilde, Denmark) was Immunoreader (NTJNC was used used on MTP-shakers at room room temperature. temperature. An An Immunoreader to measure the absorbance at 492 nm.
Insulin IA. Serum In s u lin- R - RIA. Seruminsulin insulinconcentrations concentrationswere wereradioimmunoassayed radioimmunoassayed with with aa commercially available kit for human insulin described by Besch et al. (1987). The kit contained 125-I-labeled Bound/free separation porcine insulin, human insulin as standard and anti-insulin guinea pig serum. Bound/free was performed by the double antibody technique. Calculation and statistics. Insulin concentrations, intraassay and interassay precision in human sera were calculated and expressed as mean values (i), standard deviations (SD) or standard deviation of variation variation (CV). (CV). Welsh's Welsh's heterogenic heterogenic t-test t-test was emdeviation of the mean (SEM) and coefficients of
Results
Homogeneity of the microtiter plates. The assay of one concentration in all
homogeneity thicrotiter plate 96-wells of the ii'iicrotiter plate provides provides a quantitative expression for the homogeneity of the solid phase. Testing two different human serum poois, pools, the variation coefficients were 5.9% at an absorbance of 0.42 ± 0.025 and 3.8% at an absorbance of 1.32 ± 0.05
(x+SD). (x±SD). Standards and sensitivity. Fig. 1 demonstrates the standard curve of thespresent method. The coefficients of variation of the absorbance values at five consecutive tests Nonspecificbinding binding as as measured measured on different different days days were werecalculated calculatedbetween between7-10%. 7-10%.Nonspecific in the insulin free serum was was 0.057 0.057 0.01 (x SD) absorbance units for eleven consecutive assays. Sensitivity (lower detection limit) was defined as the content of insulin pmol/1, when detected in pmo1!1, in the the 2SD-range 25D-rangeof ofthe theinsulin insulinfree freeserum. serum.ItItwas wasfound foundtotobebe1010 determined. 21 replicates were determined !
Precision and accuracy. Results of intra- and interassay precision are shown in Table 1. 1. The The CV CV for for the the intraassay intraassay precision precisionwas wascomputed computedfrom from21 21 replicates of 5 Table
50
i2
2W icXl Insulin prrsiII(l Insulin pnAI(]
so
ic
Fig.11 Standard Standardcurve curveof ofthe theinsulin insulin ETA: ETA: Each Each point point represents represents the mean of five consecutive Fig. assays (duplicates) at five different aifferent days. Vertical lines indicate ± SD
Downloaded by: University of Pittsburgh. Copyrighted material.
ployed for checking significance.
232
Exp. Clin. Endocrinol. 95 (1990) 2
human pooi pool sera containing insulin concentrations within the standard range. Interassay precision is calculated calculated by by coefficients coefficients of of variation variation of of four fourcontrol controlsera serainintwelve twelve assay runs over three months. Table 1 Intra- and interassay precision of insulin concentrations in human sera
Intraassay variation (n = 21)
53 ± 5
Mean ± SD
97±6
217±10
505 ± 18
1020 ± 63
[pmol/lJ [pmol/1J 9.4
Cv [%] cv [%]
6.2
4.6
3.6
6.2
83 ± 9
Mean ± SD
159 159 ±± 14 14
715 ± 56
957 957 ± 51
[pmol1l]
10.8
CV [%]
8.8
5.4
7.8
In recovery experiments a serum pooi pool of blood donors (n = 50) was loaded with 60, 120, 240,. .360, 480, 720, and 960 pmol/l human insulin. Between 90% to 111% of the expected insulin concentrations was found. Dilution tests were carried out using a serum pooi pool from eight healthy persons after insulin infusion. The sample containing 980 pmol/1 pmol/l insulin insulinwas wasassayed assayedafter afterserial serial
dilution steps with insulin-free serum or incubation buffer up to 1/16 of the initial concentration. Comparing measured values with those expected a recovery of 90-114% was obtained.
Spe cifity. The Specifity. The cross-reactivity cross-reactivity of of related related peptides peptides with the antiserum 432 is demonstrated in Fig. 2. Complete crossreactivity was found found with with porcine porcine insulin, insulin, aa nearly nearly complete with bovine, ovine and rat insulin. Porcine proinsulin as well as mouse insulin seem to be nearly three fold less reactive than human insulin. 2.1.2.4
*
2.0 1.6
d 1.2o 1.2
o
0.8 0.4 0.4-
0oi';'
20
5b iÔO 2O0 2Ô0 160 Antigen concentration AntigGn concentrationEpmol/!] pmol/!]
5Ó0
lobo iobo
20b0 20'OO
Fig. 2 Cross-reactivity of human human insulin insulin and and related related peptides peptidesin inthe theinsulin insulinETA EIA system:
. human human insulin, insulin, * - *-* * porcine porcineinsulin, insulin,QQ Q-0bovine bovineinsulin, insulin,xx- -xxovine ovineinsulin, insulin, mouseinsulin insulin ¿ mouse A - A rat insulin, + - + porcine proinsulin,
Downloaded by: University of Pittsburgh. Copyrighted material.
Interassay variation (n = 12)
233
J. KBATZSCH et al., Sandwich-ETA Sandwich-EIA for for Measurement of Insulin
ETA-RIA E IA - RIAcorre'ation. correlation.AAcorre'ation correlationstudy studytotocompare comparethe theresWts resultsof ofthe thepresent present insulin enzyme immunoassay immunoassay with with the the commercially commercially available availabk radioimmunoassay was human sera. sera. Fig. Fig. 33 shows shows aa highly high'y significant significant correlation corre'ation with with somesomeperformed with 132 human what lower tower values va'ues in the ETA EIA (r(r==0.98, 0.98,yy==0.97x 0.97x--57 57pmo1!!). pmoL/). 1200
1000 800 Q-
C
yr 0.97x-57 y0.97x-57
oo
r = 0.98
w 200
200
800 1000 ¿00 00 600 RIA InsuLin Insulin [pmol/I] [pmot/I]
1200
(y) versus versus RIA RIA results results (x) for insulin determination in 132 132 human sera EIA (y) Fig. 3 Comparison of ETA
4Jlinica Applications Clinical Measurement of fasting insulin insulin concentrations concentrations in in human human sera: sera:In InTable Tabk 2 mean values va'ues
and the range of fasting insulin concentrations were compared in sera from healthy heakhy aduks and chi'dren adults children with potentially diminished basal basai values va'ues like neonates (Besch et al., aL, respective'y. In that way 1987) or patients with Turner's syndrom (Amendt, 1985), respectively. the ability of the present ETA to discriminate diminished insulin concentrations from the lower tower detectable detectabk limit limit was was examined examined.In Inon'y onlyone onecase caseofofaaneonate neonatethe themeasured measured concentration was found to to be be tower lower than than 10 10 pmoVl. pmol/l. Table Meanfasting fastinginsulin insulinconcentrations concentrationsin insera seraof of healthy healthy adults adults and and children children with potentially Table 22 Mean diminished insulin basal concentration Subjects
n
Age (years)
Insulin concentration concentration [pmol/1] [pmol/l] Range Mean ± SD
Adults Neonates Turner's syndrom
29 13 14
18-60 5-10days
60 ± 39 47 ± 22 33 ± 16
5-15
11-165
10 79*
14 67
* One sample was found found to to be be lower lower than than 10 10 pmol/l pmol/l
Insulin response to oral ora' g'ucose glucose under under varioñs variois metabolic conditions: Fig. 4 shows the insulin response to oral glucose (1.75 (1.75 g/kg g/kg idea' idealbody bodyweight weightup uptotoaamaximum maximumof of ora' g'ucose compari75 75 g) g) of of children children with with constitutional constitutiona' tall tall stature stature (mean (mean age age 12 12years, years,nn 9) in compari14). Their Their carbohydrate son to children with Turner's syndrom (mean age 11 years, n = 14). tokrance was tolerance wascontrolled controlledprevious previousto toaatherapy therapywith withestrogens estrogensor orrecombinant recombinantgrowth growth
Downloaded by: University of Pittsburgh. Copyrighted material.
600
234
Exp. Clin. Endoirino1. 95 (1990) 2
hormone, respectively. respective'y. Basal Basai values va'ues and and the the response after 30 min of patients with Turner' syndrom ner'a syndromwere werefound foundto tobe be significant significant towered lowered (p
J. A. Barth, Leipzig
Institute of Clinical Chemistry and Laboratory Diagnosis (Director: MR Prof. Dr. Sc. med. se. med. W. W. Rotzsch) Rotzsch) and and Department Department of of Pediatrics Pediatrics (Director: (Director: OMR OMR Prof. Prof. Dr. Dr. Sc. Sc. med. W. Braun), Karl-Marx-University Medical School Leipzig, Central Institute of Diabetes 'Gerhardt Katsch" (Director: OMR 0MB Prof. Prof.Dr. Dr.Sc. sc. med. H. Bibergeil) Karlsburg/GDR
for Measurement of Insulin on Microtiter Plates J. KRATZSCH, KRATZSCH,W. W.ACKERMAn.T, ACKRMAnÇ HHKILACKIR, KInLACJcJm, W.W. BESCH BEsdiland and E. E. KELLER KELLER With 5 Figures
Summary. A sensitive enzyme immunoassay for the measurement of insulin in human sera on microtiter plates was established. The assay is based on the sandwich technique with guinea pig anti-insulin IgG adsorbed adsorbed at at microtiter microtiter plate plate wells, wells, human humaninsulin insulinas asstandard standardand andthe thesame sameanti-insuanti-insulin TgG IgG labeled labeled with with horseradish horseradish peroxidase. peroxidase. Standards used cover a range from O to 1200 1200 pmol/1 pmol/l with a detection limit of 10 pmol/l. Coefficients of variation between 3-7% for intraassay precision and 5-11% for interassay precision were obtained over the concentration range of 80-1000 pmol/l. The correlation of ETA-data ETA-data with with those those of of aa commercially commerciallyavailable availabledouble doubleantibody antibodyradioimmunoradioimmunoassay (r = 0.98) could be expressed expressed by by the the equationS equation: ETA = 0.97 RTA RIA -- 57 57 pmol/l. pmol/1. Normal Normal fasting fasting serum insulin concentrations concentrations in in healthy healthy subjects subjects ranged rangedfrom from11-165 11-165pmol/1. pmol/l. In In subjects subjects with with poPo-
tentially diminished basal values concentrations of 10-79 pmol/l were determined. The insulin response response in in oral oral glucose glucose tolerance tolerance tests tests of of children children was was discussed, discussed, who who had had aa constitutional constitutional tall tall stature stature or Turner's syndrom, respectively.
Key K e ywords: words:Insulin Insulindetermination determination - Enzyme immunoassay - Microtiter plates - Serum insulin concentrations
Introduction
The determination of insulin secretion is a helpful parameter in characterizing a defect carbohydrate metabolism in patients with various metabolic disorders including diabetes mellitus (Rausch-Stroohmann, 1986). Since the first description of radioimmunoassay (RIA) for the measurement of insulin by Berson and Yalow (1959), RIA has been used most widely to assay that hor-
mone. In the last years significant improvements in reliability and practicability of the RIA measurement have been reported (Besch et al., 1987). However, in order to avoid disadvantages inherent with the application of radioisotopes, enzyme (Ishikawa, 1973) and fluorescence labels (Yamaguchi et al., 1982) have been introduced. Competitive assays are not capable of measuring the whole range of clinically encountered insulin concentrations with a single dilution of sample. The use of sandwichtype tests provides a wider working range, a better sensitivity and a lowered number of pitfalls because of matrix effects (Imagawa et al., 1983; Hoffmann, 1985).
Downloaded by: University of Pittsburgh. Copyrighted material.
A Sensitive Sandwich Enzyme Immunoassay
230
Exp. Clin. Endocrinol. 95 (1990) 2
To ensure a high degree of effectivity and automation in the laboratory, microtiter plates are a suitable bound/free separation system allowing the determination of hundreds of samples in a very short time. This paper describes a simple simple and and sensitive sensitive sandwich sandwich enzyme enzymeimmunoassay immunoassay(ETA) (EIA) for the determination of insulin in human sera in the absence of serum insulin-binding antibodies. The antigen capturing antibody is immobilized to the wells of microtiter plates, the second (detecting) antibody is labeled with horseradish peroxidase. The clinical
reliability of the assay is tested in the discrimination of insulin concentrations from fasting healthy adults and children with potentially diminished basal values (neonates, patients with Turner's syndrom). Besides, the insulin response is compared in oral glucose tolerance tests of children with constitutional tall stature or Turner's syndrom, Material and Methods Matørial Insulin-free Insulin - freeserum. serum.AAnormal normalserum serumpool poolwas wastreated treatedby bycharcoal charcoalextraction extractionand andfiltration, filtration, it was supplied by the VEB Schsisches Serumwerk Dresden (GDR).
Buffers and solutions. Coating Coating buffer buffer was was aa carbonate carbonatebuffer buffer(0.1 (0.1Mfl) M/l)pH pH9.6. 9.6.The Theincubation incubation buffer buffer contained contained 0.01 0.01 M M phosphate/I, phosphate/I, 0.15 0.15 M M NaC1/l NaCIfJ and and 1% 1% insulin-free insulin-free serum. serum. As As washing washing buffer buffer 0.01 Mfl MR phosphate-buffered isotonic saline with 0.05% Tween 80 was used. Substrate solution contained 0.1 M citrate/phosphate citiateJphosphate (pH 5.0), 0.8 g o-phenylendiamine (Fluka-AG, FRG) and 3.2 ml of 3% H202 per litre. Microtiter plates with 96 flat-bottomed wells were purchased from VEB Polyplast Halberstadt (GDR).
Insulin standards. Insulin-free Insulin-free serum serum was was loaded loaded with withhuman humaninsulin insulinstandard standardHT-8 III-8 (Kohnert et al., 1973) adjusted to WHO human insulin standard (lot 66/304). Standard concentrations ranged from from 1200 1200 to to 18.75 18.75 pmol pmol insulin insulin per per litre litre as as aa series series of of 1/2 1/2 dilutions. dilutions. Highly Highly purified purified porcine porcine (S (5 23267), rat rat (998 (998 611) 611) and and mouse mouse (M (M 20169) 20169) insulins; insulins; and anUporcine porcineproinsulin proinsulin(1670) (1670)were wereobtained obtainedfrom fromNOVO NOVO (Denmark). Bovine insulin insulin (lot (lot 104F-0740) 104F-0740) was was supplied suppliedby bySigma SigmaChem. Chem.Corp. Corp.(St. (St.Louis, Louis,USA). USA). Ovine insulin (1f7J74) (1/7/74) was a gift froni from "Central Institute of Diabetes". Diabetes".
Production and purification of antiserum. Antiserum 432 was generated in guinea pig with crystalline porcine insulin (Bosch (Besch et al., 1987). The antibody was purifieU purified by by means means of of precipita precipitation with 18% Na2SO4 Na2504 and and affinity affinity chromatography chromatography on Protein A-Sepharose (Pharmacia Fine Fine CheChemicals Uppsala, Sweden) to receive the IgG-fraction. Protein concentration was determined by the
method of Lowry, and aliquotes °C. aliquotes were were stored stored at at 20 20°C.
Preparation of the IgG-peroxidase conjugate. The purified IgG-fraction as IgG.fraction obtained a9 described above was coupled via the periodate-method, reported by Nakane and Kawaoi (1974) with horseradish-peroxidase (RZ 2.7; VEB Arzneimittelwerk Dresden, GDR). The original method was modified using a molar ratio IgG to peroxidase of 1:2. The reacting mixture was fractionated on a 30 30x1 x 1cm cmSephadex SephadexG-200 G-200(Pharmacia) (Pharmacia)column. column.Aliquotes Aliquotesof ofthe the conjugate conjugate were were stored stored at at
2020°C. °C. T es tprocedure. Test proc e du re. Microtiter Microtiterplates plateswere werecoated coatedovernight overnight with 4.6>< 4.6 x 10 10 mg guinea pig antiporcine insulin IgG per ml coating buffer (0.3 mlfwell). mI/well). After repeated rinsing with washing buffer the plate is ready to use and may be stored at 4°C after drying and sealing. To perform the assay, 0.2 ml of incubation buffer were deposited to each well. In addition, 0.05 ml samples or standards were pipetted. After two hours of incubation the plates were rinsed three times. 0.25 ml anti-insulin IgG-peroxidase in IgG-peroxid.ase in incubation incubation buffer buffer (1:1000) were added and allowed to incubate for for another another two two hours. Plates were rinsed again, 0.25 ml substrate solution was added and, after a 30 min incubation
Downloaded by: University of Pittsburgh. Copyrighted material.
respectively.
231
for Measurement of Insulin J. KRATZSCH et al., Sandwich-ETA Sandwich-EIA for
carried out in the dark, the reaction was stopped with 0.05 ml 5M H2SO4. All incubation steps were carried (NTJNC Roskilde, Roskilde, Denmark) was Immunoreader (NTJNC was used used on MTP-shakers at room room temperature. temperature. An An Immunoreader to measure the absorbance at 492 nm.
Insulin IA. Serum In s u lin- R - RIA. Seruminsulin insulinconcentrations concentrationswere wereradioimmunoassayed radioimmunoassayed with with aa commercially available kit for human insulin described by Besch et al. (1987). The kit contained 125-I-labeled Bound/free separation porcine insulin, human insulin as standard and anti-insulin guinea pig serum. Bound/free was performed by the double antibody technique. Calculation and statistics. Insulin concentrations, intraassay and interassay precision in human sera were calculated and expressed as mean values (i), standard deviations (SD) or standard deviation of variation variation (CV). (CV). Welsh's Welsh's heterogenic heterogenic t-test t-test was emdeviation of the mean (SEM) and coefficients of
Results
Homogeneity of the microtiter plates. The assay of one concentration in all
homogeneity thicrotiter plate 96-wells of the ii'iicrotiter plate provides provides a quantitative expression for the homogeneity of the solid phase. Testing two different human serum poois, pools, the variation coefficients were 5.9% at an absorbance of 0.42 ± 0.025 and 3.8% at an absorbance of 1.32 ± 0.05
(x+SD). (x±SD). Standards and sensitivity. Fig. 1 demonstrates the standard curve of thespresent method. The coefficients of variation of the absorbance values at five consecutive tests Nonspecificbinding binding as as measured measured on different different days days were werecalculated calculatedbetween between7-10%. 7-10%.Nonspecific in the insulin free serum was was 0.057 0.057 0.01 (x SD) absorbance units for eleven consecutive assays. Sensitivity (lower detection limit) was defined as the content of insulin pmol/1, when detected in pmo1!1, in the the 2SD-range 25D-rangeof ofthe theinsulin insulinfree freeserum. serum.ItItwas wasfound foundtotobebe1010 determined. 21 replicates were determined !
Precision and accuracy. Results of intra- and interassay precision are shown in Table 1. 1. The The CV CV for for the the intraassay intraassay precision precisionwas wascomputed computedfrom from21 21 replicates of 5 Table
50
i2
2W icXl Insulin prrsiII(l Insulin pnAI(]
so
ic
Fig.11 Standard Standardcurve curveof ofthe theinsulin insulin ETA: ETA: Each Each point point represents represents the mean of five consecutive Fig. assays (duplicates) at five different aifferent days. Vertical lines indicate ± SD
Downloaded by: University of Pittsburgh. Copyrighted material.
ployed for checking significance.
232
Exp. Clin. Endocrinol. 95 (1990) 2
human pooi pool sera containing insulin concentrations within the standard range. Interassay precision is calculated calculated by by coefficients coefficients of of variation variation of of four fourcontrol controlsera serainintwelve twelve assay runs over three months. Table 1 Intra- and interassay precision of insulin concentrations in human sera
Intraassay variation (n = 21)
53 ± 5
Mean ± SD
97±6
217±10
505 ± 18
1020 ± 63
[pmol/lJ [pmol/1J 9.4
Cv [%] cv [%]
6.2
4.6
3.6
6.2
83 ± 9
Mean ± SD
159 159 ±± 14 14
715 ± 56
957 957 ± 51
[pmol1l]
10.8
CV [%]
8.8
5.4
7.8
In recovery experiments a serum pooi pool of blood donors (n = 50) was loaded with 60, 120, 240,. .360, 480, 720, and 960 pmol/l human insulin. Between 90% to 111% of the expected insulin concentrations was found. Dilution tests were carried out using a serum pooi pool from eight healthy persons after insulin infusion. The sample containing 980 pmol/1 pmol/l insulin insulinwas wasassayed assayedafter afterserial serial
dilution steps with insulin-free serum or incubation buffer up to 1/16 of the initial concentration. Comparing measured values with those expected a recovery of 90-114% was obtained.
Spe cifity. The Specifity. The cross-reactivity cross-reactivity of of related related peptides peptides with the antiserum 432 is demonstrated in Fig. 2. Complete crossreactivity was found found with with porcine porcine insulin, insulin, aa nearly nearly complete with bovine, ovine and rat insulin. Porcine proinsulin as well as mouse insulin seem to be nearly three fold less reactive than human insulin. 2.1.2.4
*
2.0 1.6
d 1.2o 1.2
o
0.8 0.4 0.4-
0oi';'
20
5b iÔO 2O0 2Ô0 160 Antigen concentration AntigGn concentrationEpmol/!] pmol/!]
5Ó0
lobo iobo
20b0 20'OO
Fig. 2 Cross-reactivity of human human insulin insulin and and related related peptides peptidesin inthe theinsulin insulinETA EIA system:
. human human insulin, insulin, * - *-* * porcine porcineinsulin, insulin,QQ Q-0bovine bovineinsulin, insulin,xx- -xxovine ovineinsulin, insulin, mouseinsulin insulin ¿ mouse A - A rat insulin, + - + porcine proinsulin,
Downloaded by: University of Pittsburgh. Copyrighted material.
Interassay variation (n = 12)
233
J. KBATZSCH et al., Sandwich-ETA Sandwich-EIA for for Measurement of Insulin
ETA-RIA E IA - RIAcorre'ation. correlation.AAcorre'ation correlationstudy studytotocompare comparethe theresWts resultsof ofthe thepresent present insulin enzyme immunoassay immunoassay with with the the commercially commercially available availabk radioimmunoassay was human sera. sera. Fig. Fig. 33 shows shows aa highly high'y significant significant correlation corre'ation with with somesomeperformed with 132 human what lower tower values va'ues in the ETA EIA (r(r==0.98, 0.98,yy==0.97x 0.97x--57 57pmo1!!). pmoL/). 1200
1000 800 Q-
C
yr 0.97x-57 y0.97x-57
oo
r = 0.98
w 200
200
800 1000 ¿00 00 600 RIA InsuLin Insulin [pmol/I] [pmot/I]
1200
(y) versus versus RIA RIA results results (x) for insulin determination in 132 132 human sera EIA (y) Fig. 3 Comparison of ETA
4Jlinica Applications Clinical Measurement of fasting insulin insulin concentrations concentrations in in human human sera: sera:In InTable Tabk 2 mean values va'ues
and the range of fasting insulin concentrations were compared in sera from healthy heakhy aduks and chi'dren adults children with potentially diminished basal basai values va'ues like neonates (Besch et al., aL, respective'y. In that way 1987) or patients with Turner's syndrom (Amendt, 1985), respectively. the ability of the present ETA to discriminate diminished insulin concentrations from the lower tower detectable detectabk limit limit was was examined examined.In Inon'y onlyone onecase caseofofaaneonate neonatethe themeasured measured concentration was found to to be be tower lower than than 10 10 pmoVl. pmol/l. Table Meanfasting fastinginsulin insulinconcentrations concentrationsin insera seraof of healthy healthy adults adults and and children children with potentially Table 22 Mean diminished insulin basal concentration Subjects
n
Age (years)
Insulin concentration concentration [pmol/1] [pmol/l] Range Mean ± SD
Adults Neonates Turner's syndrom
29 13 14
18-60 5-10days
60 ± 39 47 ± 22 33 ± 16
5-15
11-165
10 79*
14 67
* One sample was found found to to be be lower lower than than 10 10 pmol/l pmol/l
Insulin response to oral ora' g'ucose glucose under under varioñs variois metabolic conditions: Fig. 4 shows the insulin response to oral glucose (1.75 (1.75 g/kg g/kg idea' idealbody bodyweight weightup uptotoaamaximum maximumof of ora' g'ucose compari75 75 g) g) of of children children with with constitutional constitutiona' tall tall stature stature (mean (mean age age 12 12years, years,nn 9) in compari14). Their Their carbohydrate son to children with Turner's syndrom (mean age 11 years, n = 14). tokrance was tolerance wascontrolled controlledprevious previousto toaatherapy therapywith withestrogens estrogensor orrecombinant recombinantgrowth growth
Downloaded by: University of Pittsburgh. Copyrighted material.
600
234
Exp. Clin. Endoirino1. 95 (1990) 2
hormone, respectively. respective'y. Basal Basai values va'ues and and the the response after 30 min of patients with Turner' syndrom ner'a syndromwere werefound foundto tobe be significant significant towered lowered (p
