Development of an Isotope Dilution Assay for Precise Determination of Insulin, C-peptide, and Proinsulin Levels in Non-diabetic and Type II Diabetic Individuals with Comparison to Immunoassay

Abstract

We describe the application of a stable isotope dilution assay (IDA) to determine precise insulin, C-peptide, and proinsulin levels in blood by extraction from serum and quantitation by mass spectrometry using analogues of each target protein labeled with stable isotopes. Insulin and C-peptide levels were also determined by immunoassay, which gave consistently higher results than by IDA, the relative difference being larger at low concentrations. Insulin, C-peptide, and proinsulin levels were all shown by IDA to be higher in type II diabetics than in non-diabetics, with mean values rising from 22 (± 2) to 92 (± 8), 335 (± 11) to 821 (± 24), and 6 (± 1) to 37 (± 3) pm, respectively. Interestingly, the ratio between IDA and immunoassay values for insulin levels increased from 1.3 in non-diabetics to 1.7 in type II diabetics. The ratio between proinsulin and insulin levels by IDA increased from 0.24 in non-diabetics to 0.36 in type II diabetics, whereas the ratio between C-peptide and insulin levels by IDA decreased from 17.6 to 10.7. This disproportionate change in protein levels between different types of individuals has implications for the metabolism of insulin in the diabetics studied (type II) and suggests that C-peptide levels are not always a reliable guide as to pancreatic insulin secretion. In addition, levels of the 33-residue C-peptide (partially trimmed form) were shown to be less than 10% that of the fully trimmed 31-residue C-peptide levels, and we tested IDA in a clinical context by two post-pancreatic graft studies. IDA was shown to give direct, positive identification of the target protein with unrivaled accuracy, avoiding many of the problems associated with present methodology for protein determination. We describe the application of a stable isotope dilution assay (IDA) to determine precise insulin, C-peptide, and proinsulin levels in blood by extraction from serum and quantitation by mass spectrometry using analogues of each target protein labeled with stable isotopes. Insulin and C-peptide levels were also determined by immunoassay, which gave consistently higher results than by IDA, the relative difference being larger at low concentrations. Insulin, C-peptide, and proinsulin levels were all shown by IDA to be higher in type II diabetics than in non-diabetics, with mean values rising from 22 (± 2) to 92 (± 8), 335 (± 11) to 821 (± 24), and 6 (± 1) to 37 (± 3) pm, respectively. Interestingly, the ratio between IDA and immunoassay values for insulin levels increased from 1.3 in non-diabetics to 1.7 in type II diabetics. The ratio between proinsulin and insulin levels by IDA increased from 0.24 in non-diabetics to 0.36 in type II diabetics, whereas the ratio between C-peptide and insulin levels by IDA decreased from 17.6 to 10.7. This disproportionate change in protein levels between different types of individuals has implications for the metabolism of insulin in the diabetics studied (type II) and suggests that C-peptide levels are not always a reliable guide as to pancreatic insulin secretion. In addition, levels of the 33-residue C-peptide (partially trimmed form) were shown to be less than 10% that of the fully trimmed 31-residue C-peptide levels, and we tested IDA in a clinical context by two post-pancreatic graft studies. IDA was shown to give direct, positive identification of the target protein with unrivaled accuracy, avoiding many of the problems associated with present methodology for protein determination. At present, protein levels in humans are measured by following radiolabeled material or by the use of immunological reagents. Radioiodination is the most common procedure of protein labeling and has been of great value. However, it has been recognized that chemical modification of a protein in this way can affect its biological and physicochemical behavior. Tissue dehalogenases, which are known to remove iodine from iodotyrosine, may invalidate concentration measurements, particularly of fragments of the target molecule (1Cockram C.S. Jones R.H. Boroujerdi M.A. Sonksen P.H. Diabetes. 1984; 33: 721-727Crossref PubMed Scopus (16) Google Scholar, 2Stumpo R.R. Llera A.S. Cardoso A.I. Poskus E. J. Immunol. Methods. 1994; 169: 241-249Crossref PubMed Scopus (19) Google Scholar). Tritiated proteins have been used since they escape these objections (3Davies J.G. Cerini F. Bradshaw C.G. Insulin and the Cell Membrane. Harwood Academic Publishers GmbH, Chur, Switzerland1990: 243-254Google Scholar), but the ethical issue of injecting radioactive proteins of any kind into human subjects remains under severe scrutiny. The alternative method involving the use of immunological reagents to estimate blood protein levels remains the best current approach. However, antibodies used in these measurements often cross-react with precursors of the target protein or with its smaller, degraded fragments, creating uncertainty in results (2Stumpo R.R. Llera A.S. Cardoso A.I. Poskus E. J. Immunol. Methods. 1994; 169: 241-249Crossref PubMed Scopus (19) Google Scholar, 4Temple R.C. Carrington C.A. Luzio S.D. Owens D.R. Schneider A.C. Sobey W.J. Hales C.N. Lancet. 1989; i: 293-295Abstract Scopus (348) Google Scholar, 5Temple R.C. Clark P.M. Nagi D.K. Schneider A.E. Yudkin J.S. Hales C. Clin. Endocrinol. 1990; 32: 684-693Crossref Scopus (126) Google Scholar, 6Saric T. Seitz H.J. Pavelic K. Mol. Cell. Endocrinol. 1994; 106: 23-29Crossref PubMed Scopus (12) Google Scholar). Although modern forms of radioimmunoassay, microparticle enzyme immunoassay (MEIA), 1The abbreviations used are: MEIA, microparticle enzyme immunoassay; CEI, chemiluminescent enzyme immunoassay; [2H14]C-peptide, [2H2]Gly7 human C-peptide analogue; [2H16]insulin, [2H8]PheB1-[2H8]ValB2human insulin analogue; ESI-MS, electrospray ionization mass spectrometry; IDA, isotope dilution assay; HPLC, high performance liquid chromatography; TPCK, l-1-tosylamido-2-phenylethyl chloromethyl ketone; IRMA, a monoclonal antibody-based two-site immunoradiometric assay. 1The abbreviations used are: MEIA, microparticle enzyme immunoassay; CEI, chemiluminescent enzyme immunoassay; [2H14]C-peptide, [2H2]Gly7 human C-peptide analogue; [2H16]insulin, [2H8]PheB1-[2H8]ValB2human insulin analogue; ESI-MS, electrospray ionization mass spectrometry; IDA, isotope dilution assay; HPLC, high performance liquid chromatography; TPCK, l-1-tosylamido-2-phenylethyl chloromethyl ketone; IRMA, a monoclonal antibody-based two-site immunoradiometric assay. and immunoradiometric assay (7Sobey W.J. Beer S.F. Carrington C.A. Clark P.M. Frank B.H. Gray I.P. Luzio S.D. Owens D.R. Schneider A.E. Siddle K. Temple R.C. Hales C.N. Biochem. J. 1989; 260: 535-541Crossref PubMed Scopus (339) Google Scholar, 8Clark P.M. Levy J.C. Cox L. Burnett M. Turner R.C. Hales C.N. Diabetologia. 1992; 35: 469-474Crossref PubMed Scopus (49) Google Scholar, 9Kjems L.L. Roder M.E. Dinesen B. Hartling S.G. Jorgensen P.N. Binder C. Clin. Chem. 1993; 39: 2146-2150Crossref PubMed Scopus (75) Google Scholar) reduce many of these cross-reactivities, the methodology still remains an indirect measure of the target molecule, where endogenous and administered exogenous protein cannot be distinguished. Also, pathological samples may contain endogenous antibodies that can interfere with conventional assays and invalidate the results obtained. Such difficulties, as shown by significant discrepancies between immunoassays performed at different centers on the same samples (10Robbins D.C. Andersen L. Bowsher R. Chance R. Dinesen B. Frank B. Gingerich R. Goldstein D. Widemeyer H.-M. Haffner S. Hales C.N. Jarett L. Polonsky K. Porte D. Skyler J. Webb G. Gallagher K. Diabetes. 1996; 45: 242-256Crossref PubMed Scopus (0) Google Scholar), have led to a feeling that measurements of insulin levels, for example, have been systematically overestimated by radioimmunoassay in type II (non-insulin-dependent) in being to insulin as to insulin R.C. Clark P.M. Nagi D.K. Schneider A.E. Yudkin J.S. Hales C. Clin. Endocrinol. 1990; 32: 684-693Crossref Scopus (126) Google Scholar). many type II and as or to a in the concentration of proteins as proinsulin and its that are with insulin T. Seitz H.J. Pavelic K. Mol. Cell. Endocrinol. 1994; 106: 23-29Crossref PubMed Scopus (12) Google Scholar, L. Pavelic K. S. 1984; PubMed Scopus Google Scholar, D.C. J. 1984; PubMed Scopus Google Scholar, J.S. J. 1993; PubMed Scopus Google Scholar). The clinical of results involving remains a measure of blood protein concentration be most in and be to fully the and of proteins in a is that can determine each of the a between endogenous and exogenous protein avoiding the of radioactive this we describe an alternative to as isotope dilution assay that precise quantitation of proteins at in In this target proteins are from a serum by or and by concentration is determined by mass spectrometry from the ratio between the of each target protein and that of an that has been to the serum in known protein is also in which of the proteins by and to protein by The and proteins are stable analogues of the target proteins that different IDA a method known as stable isotope dilution mass spectrometry that has for the of C. PubMed Scopus Google Scholar, 1992; PubMed Scopus Google Scholar, E. Chem. 1993; Scopus Google Scholar, C. 1993; PubMed Scopus Google Scholar). However, its use for as proteins is less to in the of the and the of by of insulin levels in serum R. L. L. Cerini F. K. Diabetes. PubMed Scopus Google Scholar), we the of an of this to measure in humans the protein proinsulin and insulin with the J.S. J. 1993; PubMed Scopus Google Scholar, S. PubMed Google Scholar, Clark S. PubMed Scopus Google Scholar, S. J. Chem. PubMed Google Scholar, Biochem. J. 1994; PubMed Scopus Google Scholar). The of insulin, is by the of the pancreatic of of the proinsulin is to insulin and the C-peptide known as but is from the Insulin is in the by the The relative levels of insulin, C-peptide, and the and are known to between different types of individuals R.C. Carrington C.A. Luzio S.D. Owens D.R. Schneider A.C. Sobey W.J. Hales C.N. Lancet. 1989; i: 293-295Abstract Scopus (348) Google Scholar, J.S. J. 1993; PubMed Scopus Google Scholar, M.E. J. PubMed Scopus Google Scholar, S. P.H. Porte D. J. Clin. Endocrinol. 1990; PubMed Scopus Google Scholar, S. Roder M.E. Hartling S.G. Binder C. J. PubMed Scopus Google Scholar, S. J. 1993; PubMed Scopus (16) Google Scholar). However, the precise difference between levels of these proteins is to since present methodology cannot between all of the of and of proteins in the is to as IDA, we insulin, and C-peptide levels in and type II individuals and results with by was from from human insulin in and serum and were from reagents the human C-peptide, were of and were from or The proinsulin in a a M. L. J. J. Immunol. Methods. 1993; PubMed Scopus Google Scholar), was and in by a J. T. Scholar) at a of of protein using a the was with of protein at and proinsulin was by a with a of in in in was labeled with by of the in a J. T. Scholar, D.C. 1989; PubMed Scopus Google Scholar), where the of was to be by mass proinsulin was with and by and where the of protein was to be proinsulin was by using of and of at to and insulin and C-peptide and 33-residue was as R. K. 1994; PubMed Scopus Google Scholar). [2H2]Gly7 and 33-residue were by using was performed by using as The concentration of proinsulin and insulin was determined from the at P.H. Biochem. 1989; PubMed Scopus Google Scholar), and the concentration of C-peptide was determined by HPLC, and electrospray ionization mass spectrometry a were at a concentration of and at protein any protein dilution was in with a of serum on the C-peptide to the 31-residue the 33-residue is dilution with of in the serum was a and of was a in serum that has a to known of each a concentration to that for the target and each and were to the was with and in in The serum was and the at using at of in and were the to remove proteins J. Frank B.H. M.A. Polonsky 35: PubMed Scopus Google Scholar). The target insulin proteins were from the with in and the was dilution with of the serum was a and of and a known of each and each were for The serum was an in the and the was with the The of an to which insulin antibodies and C-peptide antibodies known to also were at concentration was with in in and the of the of was the to the target proteins into the The was and with in The target insulin proteins were from the with in and the was extraction by or the samples were in in and a target protein from the serum was and with its and by with a of in at a of in in The concentration of the and of the proteins target its and its as a by The were each target with its and were in of The samples were by in positive ionization using at a of and a of or II at a of and a of The was to the most of the target protein and its whereas the was to be of this of the ratio between the of the target protein and the of the at known the concentration of the target protein in the serum was of the extraction and procedure not affect the samples from the were by Insulin concentration was determined with a MEIA, which monoclonal antibodies human C-peptide concentration was determined with a chemiluminescent enzyme immunoassay which antibodies The proinsulin concentration of these samples was not determined by serum samples from types of individuals were from the The concentration of insulin or C-peptide different was measured by samples most of the of blood concentration insulin and results by the two the insulin or C-peptide concentration in these samples was also measured by The of for to each was from to on the insulin concentration as determined by the of for was from to 6 on the C-peptide concentration as determined by The concentration of insulin, in from a serum was determined by IDA to the of This procedure was for different serum samples at insulin of and as determined by procedure was to measure by IDA the concentration of C-peptide in serum samples determined by to be and concentration was measured by IDA of serum samples at an insulin concentration of and and were at a concentration to that of the target protein determined by and for was at the insulin The 33-residue of the C-peptide an two and which at its from This is to be present in serum at a concentration than the fully trimmed C-peptide of The C-peptide concentration of serum samples was determined by The 33-residue C-peptide and the 31-residue C-peptide were in these samples by The of were of and of a serum samples were from types of and type II was from where individuals were as with a mass The precise of insulin, C-peptide, and proinsulin in each of these samples was by The concentration of insulin all and C-peptide in samples from was also measured by The of to each were of [2H16]insulin, of [2H14]C-peptide, and of This a with of the of the insulin and C-peptide levels were measured by immunoassay from blood samples for samples were by IDA to determine insulin, C-peptide, and proinsulin levels this in the of serum was from blood samples by The of to each were of [2H16]insulin, of [2H14]C-peptide, and of a to that for pancreatic was by an where of in of was a Insulin and C-peptide levels were measured by immunoassay from blood samples for insulin, C-peptide, and proinsulin levels were determined in these samples by The same of were to each as in the with protein the human and was by and all of these analogues to be used as and that as the same by as and in mass to the present in by that have the of target protein or for each of insulin, C-peptide, and proinsulin in each is of the by of each the with the was used for in each is used as used as used as used as used as used as used as used as the human and mass was determined by in each is the by of each the with the was used for in each is used as used as in a the human and mass was determined by Insulin, and the C-peptide were and from of serum by and extraction with of the target extraction most of the in the serum have the mass by This the to protein and is at low protein 1) a extraction of the target proteins than which also proteins of mass to the C-peptide as that at an between that of the C-peptide and extraction and an ratio was by of each target protein at its concentration in the of its the high of this method extraction of a of samples of each target protein or we that was by from endogenous in the serum extraction and that measurements of the target protein or its not samples from the were at for to by by IDA and this were and that was significant change in insulin, C-peptide, or proinsulin of of insulin, C-peptide, or proinsulin from the same serum that the of results by IDA is 10% at all However, we that this was at protein and estimate that the this is at concentration by immunoassay was using a of by the by IDA that the insulin or C-peptide concentration of each of these was in with the concentration of and [2H14]C-peptide, to 10% not The insulin and C-peptide concentration of serum samples by IDA from to and from to of these results with by immunoassay that measurements are higher by immunoassay for and of the samples and Although the difference between immunoassay and IDA values is larger at higher the relative difference is larger at and This may be to larger in at concentrations. is a between immunoassay and IDA values with and this values by IDA were from immunoassay of the relative between and IDA values is at all that the is not between and IDA of C-peptide serum samples were from types of results are higher for of the where the relative difference is larger at concentrations. that the is not from the between values values by IDA were from The relative between and IDA values that the is not The C-peptide concentration of the samples from to as determined by This not between the and 33-residue forms of the C-peptide that were shown to in the immunoassay not IDA, of each of the C-peptide since the 33-residue C-peptide the 31-residue C-peptide by and they a difference in The concentration of the 33-residue C-peptide determined in of the samples by IDA was and pm, where the 31-residue C-peptide was determined at a concentration of and pm, respectively. the 33-residue C-peptide concentration in these samples from to 10% of the 31-residue C-peptide The of the 33-residue C-peptide was low to be measured in the where the 31-residue C-peptide concentration was less than The insulin, C-peptide, and proinsulin of each as determined by IDA were with immunoassay results Insulin and C-peptide values by IDA are at the of the values determined by immunoassay (7Sobey W.J. Beer S.F. Carrington C.A. Clark P.M. Frank B.H. Gray I.P. Luzio S.D. Owens D.R. Schneider A.E. Siddle K. Temple R.C. Hales C.N. Biochem. J. 1989; 260: 535-541Crossref PubMed Scopus (339) Google Scholar, 8Clark P.M. Levy J.C. Cox L. Burnett M. Turner R.C. Hales C.N. Diabetologia. 1992; 35: 469-474Crossref PubMed Scopus (49) Google Scholar, A.S. Burnett M.A. Turner R.C. Lancet. Scopus Google Scholar). from individuals protein levels that were between from and type II levels in Insulin, C-peptide, and proinsulin levels were determined by IDA in serum samples from different individuals a were with insulin measurements by levels in type II Insulin, C-peptide, and proinsulin levels were determined by IDA in serum samples from different individuals a were with insulin measurements by and C-peptide measurements by In most insulin and C-peptide levels are higher by immunoassay than by Interestingly, the difference between and IDA results is for type II where insulin are higher Although results by IDA that insulin, C-peptide, and proinsulin levels are all increased in and type II individuals and 6 the ratio between proinsulin and insulin levels is whereas the ratio between C-peptide and insulin levels is decreased relative to individuals of insulin, C-peptide, and proinsulin levels in and type II concentration from samples are of the II ratio of determined from each not not ratio of determined from each ratio of determined from each measurements by IDA, insulin, C-peptide, and proinsulin were from the same serum using and samples from and type II are of the ratio of determined from each not in a measurements by IDA, insulin, C-peptide, and proinsulin were from the same serum using and samples from and type II respectively. The insulin and C-peptide concentration of samples the are by IDA than by immunoassay of results to values is a in concentration the levels are between and the the of to this an insulin and C-peptide levels a to in by a to by IDA are to by immunoassay the C-peptide concentration is by IDA the levels a but with a The in analogues of has often using dilution these analogues of insulin, C-peptide, and proinsulin were the target to have been to the of stable isotope dilution mass of the used in the of these are radioactive or in low each of the analogues used for in as an for in where exogenous and endogenous protein be by The protein concentration of serum is high for it to be to samples into a mass or an in the of a target proteins are to the of extraction methodology is for the quantitation of by is most the used with the target This is often not the to in on immunological Insulin antibodies used for this by IDA were shown to insulin with low C-peptide or proinsulin but the C-peptide antibodies C-peptide and proinsulin with high method of immunological by using a of antibodies insulin and C-peptide to all which are and by methodology is to the extraction of many different proteins since antibodies are not Although this extraction is less the identification of the target protein from the low of by of IDA immunoassay to methodology and were higher by immunoassay than by IDA for many of the samples between the two of results cannot be by or a difference in the concentration of used by each However, the difference between of the C-peptide by the of and 33-residue C-peptide levels and IDA we measure the 31-residue C-peptide levels may be by the significant of the 33-residue C-peptide present in is that insulin and C-peptide antibodies used in immunoassay can cross-react with proinsulin and its R.C. Carrington C.A. Luzio S.D. Owens D.R. Schneider A.C. Sobey W.J. Hales C.N. Lancet. 1989; i: 293-295Abstract Scopus (348) Google Scholar, 5Temple R.C. Clark P.M. Nagi D.K. Schneider A.E. Yudkin J.S. Hales C. Clin. Endocrinol. 1990; 32: 684-693Crossref Scopus (126) Google Scholar). Although we measured low proinsulin than in and CEI, these immunoassays were shown to cross-react with and of proinsulin in high to a of samples at known to not The of proinsulin be D. F. R. E. and K. proinsulin has been to be the present at to of the material in the R.C. Carrington C.A. Luzio S.D. Owens D.R. Schneider A.C. Sobey W.J. Hales C.N. Lancet. 1989; i: 293-295Abstract Scopus (348) Google Scholar, W.J. Beer S.F. Carrington C.A. Clark P.M. Frank B.H. Gray I.P. Luzio S.D. Owens D.R. Schneider A.E. Siddle K. Temple R.C. Hales C.N. Biochem. J. 1989; 260: 535-541Crossref PubMed Scopus (339) Google Scholar, S. L. S. Diabetes. 1993; PubMed Scopus Google Scholar), insulin and C-peptide levels may be overestimated by and in In type II high levels of proinsulin and its have been relative to insulin R.C. Carrington C.A. Luzio S.D. Owens D.R. Schneider A.C. Sobey W.J. Hales C.N. Lancet. 1989; i: 293-295Abstract Scopus (348) Google Scholar, J.S. J. 1993; PubMed Scopus Google Scholar, M.E. J. PubMed Scopus Google Scholar, S. P.H. Porte D. J. Clin. Endocrinol. 1990; PubMed Scopus Google Scholar, S. Roder M.E. Hartling S.G. Binder C. J. PubMed Scopus Google Scholar, S. J. 1993; PubMed Scopus (16) Google Scholar, P.M. L. Hales C.N. 1993; PubMed Scopus Google Scholar). However, this is a relative conventional cannot measure the change in concentration of each protein insulin levels are increased in these in the ratio of is by a in proinsulin levels This may be to of the pancreatic to the by insulin or decreased or an in or of the Diabetologia. PubMed Scopus Google Scholar, D. Diabetes. Scopus Google Scholar). using monoclonal antibody-based two-site immunoradiometric that proinsulin and proinsulin levels in type II subjects R.C. Carrington C.A. Luzio S.D. Owens D.R. Schneider A.C. Sobey W.J. Hales C.N. Lancet. 1989; i: 293-295Abstract Scopus (348) Google Scholar, W.J. Beer S.F. Carrington C.A. Clark P.M. Frank B.H. Gray I.P. Luzio S.D. Owens D.R. Schneider A.E. Siddle K. Temple R.C. Hales C.N. Biochem. J. 1989; 260: 535-541Crossref PubMed Scopus (339) Google P.M. Levy J.C. Cox L. Burnett M. Turner R.C. Hales C.N. Diabetologia. 1992; 35: 469-474Crossref PubMed Scopus (49) Google Scholar). Although is of the most immunoassays to the proinsulin cannot be fully distinguished. IDA not antibodies of high we that the of each be a of using this in the levels of proinsulin the of measured by may by since we a larger difference between and the IDA of insulin in samples from type II subjects the ratio between and IDA values for insulin levels in individuals was to be 1.3 This is in with of the two by of insulin levels in samples from types of individuals were where the ratio between and IDA results was by IDA that C-peptide levels in and type II individuals to a than insulin levels in a significant in the ratio between The of the C-peptide is not it biological of its with a on blood and in type D. Diabetologia. PubMed Scopus Google Scholar, B. J. Diabetologia. 1992; 35: PubMed Scopus Google Scholar). Insulin and the C-peptide are in an ratio the of insulin is to a extraction and Diabetes. 1984; 33: PubMed Scopus Google Scholar, E. F. J. Polonsky Diabetes. 1992; PubMed Scopus Google Scholar), C-peptide levels are used to measure insulin and However, this measure may not be since that the ratio between insulin and C-peptide levels is between different types of individuals to in the of of insulin and C-peptide from the The post-pancreatic graft of a clinical where the of protein levels is of in the of the by IDA with immunoassay The high and of proinsulin measured in the was to be to of the of the pancreatic following a in mass proinsulin levels to not This was not by IDA, since insulin and C-peptide levels are The in insulin and C-peptide concentration may be of the a to but the high in proinsulin levels by IDA have from of the as for We in results between immunoassay and IDA However, where conventional immunoassays as used forms of radioimmunoassay and are by as high or low concentration levels from or the of material of which cannot be larger are between the two of results the of a that the problems associated with present the of clinical results and these IDA be used the of insulin in measure of all proteins in proinsulin and insulin fragments, many of the to and a and in between insulin and proteins for example, be a of a pancreatic on the of the is are on of in this indirect method is not the are in the of of protein levels in humans is an a of many common and IDA is a and that be used to many different or Although for clinical IDA is not as a method as immunoassay since measurements to and is it into or in protein labeling and mass IDA to a for on protein behavior. We and for clinical serum of for human C-peptide and for and on the modification of the proinsulin and for