A Radioimmunoassay for LY315902 an analog of Glucagon-like Insulinotropic Peptide, and Its Application in the Study of Canine Pharmacokinetics

RESEARCH ARTICLES

A Radioimmunoassay for LY315902, an analog of Glucagon-like Insulinotropic Peptide, and Its Application in the Study of Canine Pharmacokinetics JAMES Z. CHOUX, GARY D. PLACE, DAVID G. WATERS, JEFFREY A. KIRKWOOD,

AND

RONALD R. BOWSHER

Received February 6, 1997, from the Department of Drug Disposition, Lilly Research Laboratories, Eli Lilly & Company, Indianapolis, IN 46285. Final revised manuscript received April 23, 1997. Accepted for publication April 28, 1997X. Abstract 0 Glucagon-like insulinotropic peptide (GLP-1) and its analogs are of interest because of their therapeutic potential in type II diabetes. LY315902 is a GLP-1-(7−37)-OH analog with a modified N-terminus (IP7), an octanoic acid (C8) acylated on the lysine residue at position 34, and a substitution with arginine at position 26. We developed a sensitive and specific radioimmunoassay (RIA) for the determination of immunoreactive LY315902 in the plasma of animals. A homobifunctional crosslinker was used to couple the nonacylated form of LY315902 [IP7-R26GLP-1-(7−37)-OH] to carrier proteins to enhance its immunogenicity. Following immunization, animal antisera were screened by RIA for the presence of LY315902 antibodies. One rabbit produced a high-affinity antiserum that display insignificant cross-reactivity against two forms of native GLP-1 and possible major metabolites of LY315902. In this RIA method, plasma samples were combined with radioiodinated LY315902 and rabbit anti-IP7-R26-GLP-1-(7−37)-OH serum, and then incubated overnight at room temperature. The bound forms of LY315902 were separated by polyethylene glycol assisted second antibody precipitation. The sensitivity of the assay was estimated to be 19 pM. Inter-assay precision (%CV) and accuracy (recovery) for quality control samples in dog plasma ranged from 8.0% to 14.7% and 92.8% to 107.3%, respectively. By applying this assay to measure plasma concentrations of immunoreactive LY315902 in dogs following twice daily subcutaneous injections of LY315902, we determined that the plasma half-life of LY315902 is significantly longer than that of native GLP-1-(7−37)-OH. We concluded that the structural modifications which were made to produce LY315902 prolonged its plasma half-life. The extended plasma half-life of LY315902 correlated well with its prolonged pharmacology in dogs.

Introduction Glucagon-like insulinotropic peptide (GLP-1) is secreted by the L-cells of the intestine and is liberated by proteolytic processing of a precursor protein, preproglucagon.1 Numerous investigations have reported that GLP-1 possesses many useful properties, such as stimulation of insulin secretion, only during periods of hyperglycemia and that the amount of insulin secreted is proportional to the magnitude of hyperglycemia, its suppression of glucagon secretion, and its effects on slowing down the gastric emptying.1,2 This peptide may have therapeutic potential as an alternative to insulin for treatment of type II diabetics, especially for patients who have failed on oral sulfonylurea therapy. However, the short plasma half-life of GLP-1 may limit its therapeutic potential. Compound LY315902 is a novel analog of native GLP-1 with three chemical modifications compared to GLP-1-(7-37)-OH. X

Abstract published in Advance ACS Abstracts, June 1, 1997.

768 / Journal of Pharmaceutical Sciences Vol. 86, No. 7, July 1997

This analog was modified systematically to extend its plasma half-life relative to native GLP-1. The modifications included the removal of the amino group at the N-terminal histidine residue, replacement of lysine at position 26 with an arginine, and the addition of an aliphatic octanoic acid moiety, abbreviated as C8 at position 34. These modifications allowed LY315902 to retain similar pharmacological properties of native GLP-1, while significantly extending the time action profiles as demonstrated by its pharmacodynamic response (e.g. insulin secretion) in animal models.3 In this report, we describe the development of a sensitive and specific radioimmunoassay (RIA) for the determination of LY315902 in canine plasma. This method was used to measure the plasma concentrations of LY315902 in beagle dogs following administration of subcutaneous doses (100, 300, or 1000 µg/kg) of LY315902 twice daily (b.i.d.) for 10 days. Pharmacokinetic parameters were estimated to provide information concerning the therapeutic potential of LY315902.

Materials and Methods MaterialssRabbit serum albumin (RSA), Na2EDTA, sodium azide, and polyethylene glycol (PEG, average molecular weight 8000) were purchased from Sigma. Keyhole Limpet Hemocyanin (KLH), disuccinimidyl glutarate (DSG), and Tween-20 were purchased from Pierce. Glycine was purchased from Bio-Rad. Phosphate buffer (PBS) was from Kemp Biotechnologies and sodium phosphate was from Mallinckrodt. Concentrated sodium hydroxide (5 N) was from Curtin Matheson Scientific. Bovine serum albumin (BSA) was from Boehringer Mannheim. Aprotinin (Trasylol) was from Miles. Normal rabbit serum (lyophilized, 5 g/dL total protein) and goat anti-rabbit gamma globulin (125 units) were from Calbiochem. GLP-1-(7-37)OH was from BaChem. Canine plasma with EDTA was purchased from Harlan Bioproducts. LY315902, radioiodinated LY315902, the nonacylated analog of LY315902 [IP7-R26-GLP-1-(7-37)-OH], GLP1-(9-36)-R26-NH2, and GLP-1-(9-37)-OH were prepared at Lilly Research Laboratories. Preparation of the ImmunogensCarrier proteins used in the conjugation included BSA, RSA, and KLH. The nonacylated analog of LY315902 [IP7-R26-GLP-1-(7-37)-OH] was used to prepare protein conjugate for immunization. The molar ratio for peptide to carrier protein was 5:1 for all three carrier proteins. The peptide was coupled to carrier proteins via the homobifunctional cross-linker, DSG, at a ratio of 5:1 (DSG:peptide). Typically, a mixture of several milliliters of both 10 mg/mL nonacylated LY315902 and 12.5 mg/mL DSG was incubated with continuous stirring for several minutes at room temperature for activation. Then, several milliliters of the carrier protein (10 mg/mL) was added with continuous stirring for about 1 h at room temperature. As a reaction quencher, 1.0 M glycine was added to the mixture to yield a final concentration of ∼20 mM to terminate the reaction. Purification of the conjugates was achieved by dialyzing the reaction mixture using Spectra/Por 6 molecular weight cuttoff dialysis tubing with PBS. ImmunizationsAntiserum production was carried out at HRP Inc. (Denver, PA). The immunogen solution was emulsified with an

S0022-3549(97)00056-7 CCC: $14.00

© 1997, American Chemical Society and American Pharmaceutical Association

Table 1sCross-Reactivity of the Antisera Used in the RIA with Some LY315902-Related Peptides and Metabolites of Native GLP-1 Peptides (fragments) GLP-1-(7−36)-NH2 GLP-1-(7−37)-OH GLP-1-(9−36)-R26-NH2 GLP-1-(9−37)-OH

Concentration (pM)a

Cross-reactivityb

up to 10000 up to 1000 up to 10000 up to 10000 up to 10000

BLQc BLQ <70 pM BLQ BLQ

a The concentration to which each peptide was spiked into the dog plasma. Expressed in terms of concentration measured by this assay as if it were LY315902. c Below the limit of quantitation of 50 pM. b

Figure 1sRelative binding of the antisera against the native GLP-1(7−37) (b), the nonacylated form of LY315902 (O), and LY315902 (1) in a competitive RIA. The antisera was collected from the sixth bleed after the administration of BSA conjugate to a rabbit. Relative binding is expressed as B/B0, the binding of radiolabeled LY315902 (B) relative to maximum binding (B0), where B0 ) the level of binding in the absence of unlabeled LY315902. equal volume of Freund’s complete adjuvant. Six New Zealand white rabbits were immunized with 0.5 mg of immunogen by multiple intramuscular injections. For booster injections, the immunogen was emulsified with an equal volume of Freund’s incomplete adjuvant. The animals received intramuscular booster injections with 0.1 mg of the immunogen at days 14 and 28 and subsequently at 28 day intervals. Antisera from different host animals were screened for their antibody response to LY315902 by using radioiodinated LY315902. For the RIA, bleeds 5-8 collected from rabbit EL-1007 were pooled and diluted 1:2 with assay buffer and stored in aliquots at approximately -70 °C. RIA ReagentssThe assay buffer consisted of 0.1 M sodium phosphate, 1% Na2EDTA, 0.1% sodium azide, 0.1% Tween-20 (10% solution), and 0.1% BSA. The pH of the buffer was adjusted to approximately 7.5 with concentrated NaOH solution. Assay buffer was used to dilute the rabbit antisera and [125I]LY315902. All working solutions were stored at ∼4 °C and were usable for at least 8 weeks. A stock LY315902 solution (in assay buffer) of 200 µg/mL was stable for at least 78 days at approximately -70 °C. The samples used to generate the standard curve, diluted in canine plasma with EDTA, were stable for at least 4 days at ∼ 4 °C. Canine plasma with ∼2% aprotinin (∼200 KIU/mL) was used as standard curve diluent. The standards and controls were considered stable if their recovery is within 20% of the theoretical concentrations. RIA ProceduresThe assay involved combining 200 µL of plasma samples, standards, or controls with 100 µL of rabbit anti-IP7-R26GLP-1-(7-37)-OH serum (diluted 1:7500) and 100 µL of [125I]LY315902 (∼50 pg/tube). After incubation for 18-24 h at room temperature, the bound/free forms were separated by adding 100 µL of 0.25% normal rabbit serum followed by 1 mL of 1% goat anti-rabbit γ-globulin containing 6% polyethylene glycol in assay buffer. The mixture was mixed thoroughly by vortexing and incubating for 1 h at room temperature. After centrifugation at approximately 2500g for ∼15 min at ∼4 °C, the liquid phase was decanted and the radioactivity in the precipitate was measured in a γ-counter. We used a weighted four-parameter logistic algorithm to analyze the binding data.4 The plasma concentrations of immunoreactive LY315902 in the samples were determined from a standard curve of reference standard LY315902 ranging in concentration from 0.0098 to 100 nM. Prior to the analysis, plasma samples were diluted 1:10, 1:100, or 1:1000 with EDTA/canine plasma containing ∼2% aprotinin. The resulting concentrations from different dilutions or neat sample analysis were averaged using a weighted-average algorithm which takes into account a different amount of error estimated statistically on different parts of the standard curve.5,6 RIA ValidationsThe limit of detection was estimated by the statistical estimation of the minimal detectable concentration.7 The accuracy was determined by measuring the recovery of the controls which were made by adding LY315902 to EDTA/canine plasma at

concentrations of 50, 500, and 2000 pM. Precision was assessed by calculating the percent coefficient of variance (%CV). Plasma controls were stored at approximately -70 °C, and they were stable for at least 35 days. The specificity of the antibodies used in the assay was investigated by measuring the cross-reactivity to native GLP-1-(737)-OH, the nonacylated form of LY315902, GLP-1-(9-36)-R26-NH2, and GLP-1-(9-37)-OH. Animal StudiessMale and female beagle dogs were dosed subcutaneously (b.i.d.) with 100, 300, and 1000 µg/kg for 10 days. LY315902 (10 mg/vial, Lilly Research Laboratories) was reconstituted in 5 mM dibasic sodium phosphate buffer with 16 mg/mL glycerin. At the start of the study, male dogs were 19-47 months old, with an average weight of 13.2 ( 0.96 kg (mean ( SE) while female dogs were 19-32 months old, with an average weight of 11.1 ( 0.66 kg. Dogs were from Marshall Farms, North Rose, NY. These subcutaneous administrations were given twice a day with approximately 6 h between doses. Following the first dose on each day, blood samples (∼3 mL) were collected from one animal/sex/treatment group on study days 1 and 10 at the following time points: predose and 0.5, 1, 1.5, 3, and 6 h postdose. Samples were collected using collection tubes with EDTA as an anticoagulant, with an additional ∼80 µL (∼800 kIU) of aprotinin as an enzyme inhibitor. The plasma portion of each sample was separated and stored at approximately -70 °C until analysis.

Results Antibody Screening and SelectionsAll rabbits produced antibodies against LY315902. The largest antibody titers were obtained with BSA or RSA conjugates. However, the antiserum from one rabbit immunized with BSA-IP7-R26GLP-1-(7-37)-OH exhibited a consistently high titer starting around the fifth bleed. Thus, the serum from bleeds 5-8 of this rabbit were pooled and used in the development of the RIA method. The specificity of the pool of rabbit antiserum was evaluated in competitive binding experiments. The results suggested that the antiserum was specific for LY315902 (and the nonacylated form of LY315902) while native GLP-1-(7-37)OH displayed an insignificant level of cross-reactivity (Figure 1). The specificity of the antiserum pool was investigated further for cross-reactivity with structurally related peptides (Table 1). The results indicated that this antiserum pool had an insignificant amount of recognition of the two fragments that are closely related to possible metabolites from native GLP-1 and LY315902. GLP-1-(9-36)-R26-NH2 and GLP-1(9-37)-OH represents the class of fragments of possible metabolites that are presumably inactive because the loss of N-terminal residues. RIA ValidationsA typical standard curve for the LY315902 RIA is shown in Figure 2. Although the sensitivity of the assay is approximately 19 pM averaged over six assays, estimated by the method of Rodbard,7 50 pM was used as the routine lower limit of quantitation to decrease the potential of false positives resulting from differences in sample matrices. The optimal working range of the assay was 50-2000 pM. For six validation assays, the level of nonspecific binding was 1.4 ( 0.3% (mean ( SD) with a maximum binding of 39 ( 1.4%. The slope and the 50% effective concentration (EC50,

Journal of Pharmaceutical Sciences / 769 Vol. 86, No. 7, July 1997

Figure 2sTypical standard curve of LY315902 (b) in this RIA. B/B0, binding of radiolabeled LY315902 (B) relative to maximum binding (B0), where B0 ) the level of binding in the absence of unlabeled LY315902. Table 2sInter-Assay Precision Data for Plasma Controlsa Assay

0 pM

50 pM

500 pM

2000 pM

GLIP0008 GLIP0009 GLIP0014 GLIP0015 GLIP0018 GLIP0021

BLQb

62.4 52.0 48.3 62.1 42.2 54.9

488.9 498.5 471.0 568.3 423.2 475.2

1700.1 1899.0 1766.5 2125.1 1799.5 1851.8

53.6 7.9 14.7 107.3

487.5 47.3 9.7 97.5

1857.0 148.2 8.0 92.8

Mean SDc % CVd % Recoverye

BLQ BLQ BLQ BLQ BLQ BLQ

a Controls are made by adding 50, 500, or 2000 pM of LY315902 into canine plasma. The mean concentration of at least duplicate measurements at the same control level are listed in this table. b Below the limit of quantitation of 50 pM. c Standard deviation of the mean. d Percent coefficient of variation (SD/mean × 100). e Measured concentration/expected concentration × 100.

Table 3sIntra-Assay Precision Data for Plasma Assay Set 1 Set 2 Set 3 Set 4 Set 5 Set 6 Mean SDc % CVd % Recoverye

Controlsa

0 pM

50 pM

500 pM

2000 pM

BLQb

46.0 42.1 44.0 39.9 44.0 37.3

433.1 398.6 410.4 446.2 422.3 428.4

1770.0 1867.0 1797.3 1900.5 1749.8 1712.2

42.2 3.2 7.5 84.4

423.2 16.9 4.0 84.6

1799.5 71.7 4.0 90.0

BLQ BLQ BLQ BLQ BLQ BLQ

a All measurements were performed in duplicate. b Below the limit of quantitation of 50 pM. c Standard deviation of the mean. d Percent coefficient of variation (SD/ mean × 100). e Measured concentration/expected concentration × 100.

the concentration of analyte necessary to displace 50% of the bound [125I]LY315902) of the standard curve were 1.14 ( 0.05 and 353 ( 46 pM, respectively. The validation data of the RIA in measuring immunoreactive LY315902 in dog plasma are reported in Tables 2 and 3. The accuracy was estimated by measuring the recovery of the LY315902 in quality control samples in canine plasma at concentrations of the lower (50 pM) and upper (2000 pM) limits of quantitation and near the EC50 (500 pM). The inter-assay precision was evaluated by measuring the %CV at each level of the controls. The interassay recovery and %CV ranged from 92.8% to 107.3% and 770 / Journal of Pharmaceutical Sciences Vol. 86, No. 7, July 1997

Figure 3sImmunoreactive LY315902 concentrations in plasma samples of the dogs receiving 100 (b), 300 (O), or 1000 (1) µg/kg of LY315902 daily (b.i.d.) subcutaneously on study days (a) 1 and (b) 10. The average concentrations from male and female dogs are presented at each time point. A value of zero was used for plasma concentrations determined to be below the limit of quantitation of 0.05 nM.

8.0% to 14.7%, respectively (Table 2). The intra-assay recovery and %CV ranged from 84.4% to 90.0% and 4.0% to 7.5%, respectively (Table 3). Because of the high doses used in the dog study, dilutions of 1:10, 1:100, or even 1:1000 were sometimes necessary to yield concentrations of LY315902 in the optimal range of the assay. Dilution linearity was confirmed by making certain that the relative error was within several percent of each dilution for samples requiring multiple dilutions. Pharmacokinetic Study in DogssThe validated RIA was used to determine the concentrations of immunoreactive LY315902 in dog plasma. Male and female dogs were averaged within each treatment group, because no substantial differences in plasma LY315902 concentration or pharmacokinetics between genders were found (Table 4). The predose plasma concentrations of immunoreative LY315902 in these dogs on study day 1 were either low or below the limit of quantitation of 0.05 nM. However, on study day 10, the predose plasma concentrations of immunoreactive LY315902 were substantially higher than the plasma levels on day 1 for a given dose group (Table 4). In addition, on study day 10, the predose LY315902 plasma levels appeared to be doserelated, i.e. the magnitude of the predose levels increased with

Table 4sPharmacokinetic Parameters of LY315902 on Study Days 1 and 10 in Male and Female Dogs following Daily (b.i.d.) Subcutaneous Doses of 100, 300, or 1000 µg/kg Study Day

Dose Level (µg/kg)

1

100 300 1000 100 300 1000 100 300 1000 100 300 1000 100 300 1000 100 300 1000

10

Sex M F M+F M F M + Fc

AUC0-ta (nM‚h)

AUC/Dose (nM‚h‚kg/µg)

Cmax (nM)

Cmax/Dose (nM‚kg/µg)

Predose Plasma Concn (nM)

Tmax (h)

Half-Life (h)

62.2 73.1 87.7 51.3 86.2 132.9 56.7 79.6 110.3 61.5 110.4 232.3 58.2 122.6 186.5 59.9 116.5 209.4

0.622 0.244 0.088 0.513 0.287 0.133 0.567 0.265 0.110 0.615 0.368 0.232 0.582 0.409 0.187 0.599 0.388 0.209

21.7 23.4 20.7 20.3 23.3 38.1 21.0 23.4 29.4 15.4 24.6 58.6 17.4 38.0 82.1 16.4 30.9 70.3

0.217 0.078 0.021 0.203 0.078 0.038 0.210 0.078 0.029 0.154 0.082 0.059 0.174 0.127 0.082 0.164 0.103 0.070

BLQb 0.05 0.40 BLQ 0.67 1.79 BLQ 0.36 1.10 0.76 4.78 7.14 1.60 4.01 8.08 1.18 4.40 7.61

1.5 1.0 1.5 1.0 1.5 1.0 1.3 1.3 1.3 1.5 3.0 1.0 1.5 0.5 1.0 1.5 1.8 1.0

1.1 3.1 3.9 1.4 2.3 3.7 1.2 2.6 3.8 2.9 3.2 8.3 2.7 3.9 5.2 2.8 3.5 6.4

a Area under the curve from time ) 0 to t, where t was the last time point whose concentration is higher than the limit of quantitation (0.05 nM). The zero-point (time ) 0, concentration ) 0) was added for predose value that is below the limit of quantitation. b Below the limit of quantitation of 50 pM. c Average values from male and female dogs.

dose. For example, at 100 µg/kg dose, the predose LY315902 level on day 10 was approximately 1 nM, while for dogs receiving 1000 µg/kg, the predose LY315902 level on day 10 was close to 8 nM (Table 4). Following subcutaneous injections of LY315902, the plasma concentrations increased rather rapidly and achieved a maximum level relatively early (Tmax ) 0.5-1.5 h postdose) for most of the dogs (Figure 3). Cmax values remained mostly at approximately 20 nM even with increasing doses on study day 1, but Cmax values increased with dose on day 10 (Figure 4). Cmax values also appeared to increase from day 1 to day 10 within a dose group (Figure 4). Pharmacokinetic parameters were estimated following subcutaneous doses of 100, 300, or 1000 µg/kg in three different male and female beagle dogs (Table 4). The same dogs were dosed for study days 1 and 10. The AUC (area under the plasma concentration curve) values were obtained from 0 to t, where t is the last time point with its plasma concentration of immunoreative LY315902 higher than the limit of quantitation (50 pM). For predose concentrations that were below the limit of quantitation, the zero-point (time ) 0, concentration ) 0) was used in AUC determination. For dogs receiving 100-1000 µg/kg of LY315902, AUC values were increased with dose on both study days 1 and 10 (Figure 4). In addition, the AUC values appeared to increase as a function of repeated subcutaneous administration for 10 days within a dose group, especially at higher doses, which was consistent with the increasing Cmax values observed. Half-life (T1/2) values were estimated mostly from the last three time points of plasma concentrations with the assumption that the declining portion of the plasma concentration time profile follows a model independent decay. Overall, halflife values range from approximately 1 h at 100 µg/kg to approximately 8 h at 1000 µg/kg (Table 4). The T1/2 values appeared to increase with increasing dose on both day 1 and day 10. The T1/2 values also became longer as a function of repeated subcutaneous administration within a dose group in 10 days (Table 4).

Discussion RadioimmunoassaysImmunoassays have been widely used to determine the plasma concentration of immunoreac-

tive glucagon-like peptides. Generally, extractions of plasma samples are required before the RIA can be performed.8-10 For example, a solid phase Sep-Pak cartridge with ∼60% acetonitrile for elution10 or a liquid-liquid extraction with ∼70% ethanol8,9 has been employed to remove most of the plasma proteins for problems of interference and/or stability. Recently, plasma GLP-1 levels are reported to be measured by RIA and ELISA without extraction.11 In our RIA, we assessed the in vitro stability of LY315902 by performing overnight incubations with aliquots of plasma. LY315902 was sufficiently stable to allow measurement of the immunoreactive LY315902 concentration directly using dog plasma without any extraction procedure. The pooled antisera used in this assay had an adequate affinity and selectivity for the quantitation of LY315902 in plasma. An insignificant amount of cross-reactivity was found with GLP-1-(7-37)-OH and some known inactive metabolites of native GLP-1 (Table 1). The antisera recognized native GLP-1-(7-37)-OH very slightly at high concentrations (10000 pM) at least several hundred times higher than the plasma circulating concentration of native GLP-1 for fasted animals, which is also believed to be similar to that in human.10,12,13 As expected, the nonacylated form of LY315902 displayed full cross-reactivity (Figure 1), as this peptide fragment was used to synthesize the immunogen. Cross-reactivity with this nonacylated peptide is probably not problematic, as in vitro experiments of LY315902 in serum suggested that cleavage of octanoic acid (C8) from LY315902 is probably not a major route of LY315902 metabolism. These studies reported an undetectable amount of IP7-R26-GLP-1-(7-37)-OH by matrixassisted laser desorption mass spectrometry, when LY315902 was incubated with rat or human serum at 37 °C for ∼24 h.14 The cross-reactivity data (Table 1) also suggests that the principle epitopes probably reside at or near the N-terminus of the peptide. The only other difference between the nonacylated form of LY315902 [IP7-R26-GLP-1-(7-37)-OH] and native GLP-1-(7-37)-OH is the substitution of arginine at position 26. Thus, the insignificant cross-reactivity of the assay antibodies to GLP-1-(7-37)-OH could also come from additional antibody recognition from the residues close to the portion that includes the modified Arg residue (position 26). However, the limited amount of recognition from the antibodies to GLP-1(9-37)-R26-NH2 suggests that this contribution,

Journal of Pharmaceutical Sciences / 771 Vol. 86, No. 7, July 1997

Figure 4sAverage values (male + female) of (a) Cmax and (b) AUC0-t as a function of doses in male and female dogs following daily (b.i.d.) subcutaneous administration of 100, 300, or 1000 µg/kg on study days 1 (b) and 10 (O). AUC0-t is calculated from time ) 0 to t, where t was the last time point whose concentration is higher than the limit of quantitation (0.05 nM). The zero-point (time ) 0, concentration ) 0) was added for the predose value that is below the limit of quantitation.

if any, is probably a minor one. The amino terminus is critical for the biological activity of GLP-1 to function.1 In the case of native GLP-1, the dipeptidases cleave the histidine and alanine residues very rapidly and generate inactive metabolites.15,16 Protection of the amino-terminus slows down this process substantially, which is achieved by the removal of the N-terminal amine for LY315902. GLP-1(9-37)-R26-NH2 and GLP-1-(9-37)-OH were chosen to confirm that the antibodies have a insignificant amount of cross-reactivity to the fragments of GLP-1 analogs with loss of amino-terminus residues, assuming that LY315902 has a similar metabolism pathway as native GLP-1. Thus, these cross-reactivity data suggested that this RIA measures principally the active form (no loss of amino-terminal residues) of the LY315902, even though more in vivo metabolite could possibly be present in plasma samples. Blood Level Data and PharmacokineticssThe relatively high doses (100, 300, and 1000 µg/kg) of LY315902 selected for this study were chosen to assess the toxicology of LY315902 in dogs. However, these studies still provided blood level data for pharmacokinetic evaluation. Plasma concentrations were readily detected by this method after dose admin772 / Journal of Pharmaceutical Sciences Vol. 86, No. 7, July 1997

istration (Figure 3). At the lowest dose group of 100 µg/kg in this study, a Cmax of approximately 20 nM was determined, which is about 400-times higher than the lower limit of quantitation for this assay. Thus, this assay certainly has sufficient sensitivity for examining the blood levels in toxicology studies. Although the potential therapeutic dose may be substantially lower than 100 µg/kg, this RIA should still be sensitive enough to measure the plasma levels in clinical trials since the native GLP-1 plasma level on the order of a few hundred picomolar are believed to be necessary for its pharmacological action, even though the endogenous level of GLP-1 were reported to increase from basal level of 20-50 pM to 60-100 pM postmeal.1,9,15 The blood level and pharmacokinetic data suggest the possibility of accumulation of LY315902 upon repeated daily administration in dogs. The predose LY315902 concentration increased both as a function of repeated dosing for 10 days (Table 4). Both the AUC and Cmax values increased as a function of repeated dosing over 10 days (Table 4). The Cmax value did not increase substantially on day 1 as a function of dose; however, the AUC values increased with increasing dose (Table 4). The additional amounts in AUC values for the highdose groups came from the higher blood levels at longer time points, suggesting that LY315902 remained in circulation for a longer period of time, at least at higher doses. This is consistent with the observation that half-life values increased as a function of escalating dose. LY315902 was dosed b.i.d. with 6 h separation between the two doses. When the blood samples were taken following the first dose on day 10, LY315902 plasma level had not returned to baseline level completely by 16 h postdose from the second dose on day 9. However, the predose plasma levels of LY315902 were also dose-related. Thus, while these data suggested an accumulation of LY315902 at the current doses, the consequence of the possible accumulation should be evaluated carefully in clinical settings, because when the dose is lowered, the amount of accumulation would probably decrease depending on the therapeutic dose (which is expected to be much lower), and may be even negligible. The half-life values were determined from plasma concentrations in samples collected <6 h postdose. It is possible that the actual elimination phase of LY315902 lies beyond the time points collected in this study since the concentration of immunoreactive LY315902 has not returned to baseline by 6 h postdose. Thus, it is possible that the true “elimination” half-life of LY315902 maybe longer than what was measured here. However, a half-life of several hours does correlate well with the pharmacodynamic response of LY315902 in dogs.3 This extended half-life results mainly from the modified N-terminal histidine residue, which reduced the degradation of the peptide into inactive forms by dipeptidyl peptidase iv,16,17 and from the acylation near the C-terminal, which increased binding of LY315902 to serum albumin, similar to strategies previously reported with acylated forms of insulin.18 This extended half-life was at least several times longer than the half-life of unmodified native GLP-1.1 With modification in formulation and/or the addition of novel protracting agents, it would not be surprising if the plasma half-life of LY315902 as well as its pharmacological effect could be extended even more.

References and Notes 1. Holst, J. J. Glucagonlike Peptide 1: A Newly Discovered Gastrointestinal Hormone. Gastroenterology 1994, 107, 18481855. 2. Ørskov, C. Glucagon-like peptide-1, a new hormone of the enteroinsular axis. Diabetologia 1992, 35,701-711. 3. Myers, S. Personal communication.

4. Dudley, R. A.; Edwards, P.; Ekins, R. P.; Finney, D. J.; McKenzie, I. G. M.; Raab, G. M.; Rodbard, D.; Rodgers, R. P. C. Guidelines for immunoassay data processing. Clin. Chem. 1985, 31, 12641271. 5. Carroll, R. J.; Ruppert, D. Transformation and Weighting in Regression; Chapman & Hall: New York, 1988; Chapter 3. 6. Davidian, M.; Carroll R. J.; Smith, W. Variance functions and the minimum detectable concentration in assays. Biometrika 1988, 75, 549-556. 7. Rodbard, D. Statistical estimation of the minimal detectable concentration (“sensitivity”) for radioligand assays. Anal. Biochem. 1978, 90, 1-12. 8. Ørskov, C.; Holst, J. J. Radioimmunoassays for glucagon-like peptides 1 and 2 (GLP-1 and GLP-2). Scand. J. Clin. Lab. Invest. 1987, 47, 165-174. 9. Ørskov, C.; Jeppesen, J.; Madsbad, S.; Holst, J. J. Proglucagon products in plasma of noninsulin-dependent diabetics and nondiabetic controls in the fasting state and after oral glucose and intravenous arginine. J. Clin. Invest. 1991, 87, 415-423. 10. Kreymann, B.; Williams, G.; Ghatei, M. A.; Bloom, S. R. Glucagon-like peptide-1 7-36: A physiological incretin in man. Lancet 1987, 2, 1300-1304. 11. Pridal, L.; Ingwersen, S. H.; Larsen, F. S.; Holst, J. J.; Adelhorst, K.; Kirk, O. Comparison of sandwich enzyme-linked immunoadsorbent assay and radioimmunoassay for determination of exogenous glucagon-like peptide-1(7-36) amide in plasma. J. Pharm. Biomed. Anal. 1995, 13, 841-850. 12. Elliott, R. M.; Morgan, L. M.; Tredger, J. A.; Deacon, S.; Wright, J.; Marks, V. Glucagon-like peptide-1(7-36) amide and glucosedependent insulinotropic polypeptide secretion in response to nutrient ingestion in man: acute post-prandial and 24-h secretion patterns. J. Endocrinol. 1993, 138, 159-166. 13. Ørskov, C.; Rabenhøj, L.; Wettergren, A.; Kofod, H.; Holst, J. J. Tissue and plasma concentrations of amidated and glycineextended glucagon-like peptide I in humans. Diabetes 1994, 43, 535-539.

14. Chou, J. Z. Personal communication. 15. Deacon, C. F.; Nauck, M. A.; Toft-Nielen, M.; Pridal, L.; Willms, B.; Holst, J. J. Both subcutaneously and intravenously administered glucagon-like peptide I are rapidly degraded from the NH2-terminus in type II diabetic patients and in healthy subjects. Diabetes 1995, 44, 1126-1131. 16. Deacon, C. F.; Johnsen, A. H.; Holst, J. J. Degradation of glucagon-like peptide-1 by human plasma in vitro yields an N-terminally truncated peptide that is a major endogenous metabolite in vivo. J. Clin. Endocrinol. Metab. 1995, 80, 952957. 17. Kiefffer, T. J.; McIntosh, C. H. S.; Pederson, R. A. Degradation of glucose-dependent insulinotrophic polypeptide and truncated glucagon-like peptide 1 in vitro and in vivo by dipeptidyl peptidase-IV. Endocrinology 1995, 136, 3585-3596. 18. Markussen, J.; Havelund, S.; Kurtzhals, P.; Andersen, A. S.; Halstrom, J; Hasselager, E; Larsen, U. D.; Ribel, U.; Schaffer, L.; Vad, K; Jonassen, I. Soluble, fatty acid acylated insulins bind to albumin and show protracted action in pigs. Diabetologia 1996, 39, 281-288.

Acknowledgments The authors would like to thank G. G. Bricker, D. J. Fletcher, Dr. J. R. Means and D. L. Wilson for their excellent support in conducting the live phase of animal experiment. The authors would also like to thank W. E. Legan for providing us with [125I]LY315902, D. L. Smiley for providing us with various analogs of GLP-1/LY315902, and J. M. Snow and J. F. Fears for their assistance in the counting room.

JS9700560

Journal of Pharmaceutical Sciences / 773 Vol. 86, No. 7, July 1997