`
`THE NEW ENGLAND JOURNAL OF MEDICINE
`
`May 14, 1992
`
`ANTIDIABETOGENIC EFFECT OF GLUCAGON-LIKE PEPTIDE-I (7-36)AMIDE IN NORMAL
`SUBJECTS AND PATIENTS WITH DIABETES MELLITUS
`
`MARK GuTNIAK, M.D., CATHRINE 0RsKov, M.D., PH.D.,JENsJ. HoLsT, M.D., Ptt.D.,
`Bo AHREN, M.D., Ptt.D., AND SuAo EFENDIC, M.D., PH.D.
`
`Abstract Background. Glucagon-like peptide-1 (7-
`36)amide (glucagon-like insulinotropic peptide, or GLIP) is
`a gastrointestinal peptide that potentiates the release of
`insulin in physiologic concentrations. Its effects in patients
`with diabetes mellitus are not known.
`Methods. We compared the effect of an infusion of
`GLIP that raised plasma concentrations of GLIP twofold
`with the effect of an infusion of saline, on the meal-related
`release of insulin, glucagon, and somatostatin in eight
`normal subjects, nine obese patients with non-insulin(cid:173)
`dependent diabetes mellitus (NIDDM), and eight patients
`with insulin-dependent diabetes mellitus (IDDM). The
`blood glucose concentrations in the patients with diabetes
`were controlled by a closed-loop insulin-infusion system
`(artificial pancreas) during the infusion of each agent, al(cid:173)
`lowing measurement of the meal-related requirement for
`exogenous insulin. In the patients with IDDM, normoglyce(cid:173)
`mic-clamp studies were performed during the infusions of
`GLIP and saline to determine the effect of GLIP on insulin
`sensitivity.
`In the normal subjects, the infusion of GLIP
`Results.
`significantly lowered the meal-related increases in the
`
`T HE existence of a chemical excitant of the endo(cid:173)
`
`crine pancreas was suggested as early as 1906. 1
`This idea gained support in 1930, when it was demon(cid:173)
`strated that the intravenous injection of crude secretin
`caused hypoglycemia in dogs by stimulating the endo(cid:173)
`crine pancreas.2 This implied that the crude prepara(cid:173)
`tion of secretin contained an intestinal factor -
`"in(cid:173)
`cretin" -
`that was able to stimulate the endocrine
`pancreas. The concept of incretin was clearly outlined
`by Creutzfeldt in 1979.3 lncretin was defined as an
`endocrine transmitter that is produced in the gastroin(cid:173)
`testinal tract, is released by food intake ( especially of
`carbohydrates), and stimulates insulin secretion in the
`presence of plasma peptide concentrations not exceed(cid:173)
`ing those reached after meals. Glucose-dependent in(cid:173)
`sulinotropic polypeptide is believed to have an impor(cid:173)
`tant role in the mediation of this signal between the
`intestine and the pancreatic B cells after eating,4 and it
`has therefore become a strong candidate for an incre(cid:173)
`tin.3 However, supraphysiologic levels of glucose(cid:173)
`dependent insulinotropic polypeptide are needed to
`potentiate insulin secretion,5 and amplification of the
`insulin response to oral glucose is partially preserved
`in the presence of antibodies to glucose-dependent in(cid:173)
`sulinotropic polypeptide.6 Therefore, it seems that
`
`From the Departments of Endocrinology, Karolinska Institute, Stockholm,
`Sweden (M.G., S.E.); the Institute of Medical Physiology C, Panum Institute,
`University of Copenhagen, Copenhagen, Denmark (C.0., J.J.H.); and the De(cid:173)
`partment of Pharmacology, Lund University, Lund, Sweden (B.A.). Address
`reprint requests to Dr. Gutniak at the Karolinska Institute, Karolinska sjukhuset,
`Box 60500, S-104 01 Stockholm, Sweden.
`Supported by a grant (0034) from the Swedish Medical Research Council and
`by grants from the Nordic Insulin Foundation, the Swedish Diabetes Association,
`and Svenska Hoechst AB, Stockholm, Sweden.
`
`blood glucose concentration (P<0.01) and the plasma
`concentrations of insulin and glucagon (P<0.05 for both
`comparisons). The insulinogenic index (the ratio of insulin
`to glucose) increased almost 10-fold, indicating that GLIP
`had an insulinotropic effect. In the patients with NIDDM,
`the infusion of GLIP reduced the mean (±SE) calculat(cid:173)
`ed isoglycemic meal-related requirement for insulin from
`17.4±2.8 to 2.0±0.5 U (P<0.001 ), so that the integrated
`area under the curve for plasma free insulin was de(cid:173)
`creased (P<0.05) in spite of the stimulation of insulin
`release. In the patients with IDDM, the GLIP infusion
`decreased the calculated isoglycemic meal-related insulin
`requirement from 9.4±1.5 to 4.7±1.4 U. The peptide
`decreased glucagon and somatostatin release in both
`groups of patients. In the normoglycemic-clamp studies in
`the patients with IDDM, the GLIP infusion significantly in(cid:173)
`creased glucose utilization (saline vs. GLIP, 7.2±0.5 vs.
`8.6±0.4 mg per kilogram of body weight per minute;
`P<0.01).
`Conclusions. GLIP has an antidiabetogenic effect,
`and it may therefore be useful in the treatment of patients
`with NIDDM. (N Engl J Med 1992;326:1316-22.)
`
`there must be incretins other than glucose-dependent
`insulinotropic polypeptide. 7•8
`Glucagon-like peptide- I ( G LP-1) is a fragment of
`the proglucagon molecule.9 This peptide has no meta(cid:173)
`bolic effect in mammals. However, two shorter forms
`of GLP-l - GLP-l (7-37) and GLP-1 (7-36)amide
`exert strong insulinotropic effects in vitro10·11 and
`-
`in vivo. 12·13 Since GLP-l (7-36)amide, the natural(cid:173)
`ly occurring form in humans, 14 is released during
`a meal 12·15 and after oral glucose administration16
`and potentiates glucose-induced insulin release, 12 this
`truncated form of GLP-1 may be an important incre(cid:173)
`tin. 8• 11 The peptide could thus be of potential value in
`the treatment of diabetes. Therefore, we investigated
`the effect of GLP-1 (7-36)amide, hereafter referred
`to as GLIP (glucagon-like insulinotropic peptide), on
`the release of hormones from islet cells in normal sub(cid:173)
`jects and patients with non-insulin-dependent diabe(cid:173)
`tes mellitus (NIDDM). In addition, we determined
`the effect of GLIP on the need for insulin after a stand(cid:173)
`ard meal in these patients. Since the peptide markedly
`decreased the insulin requirement and inhibited glu(cid:173)
`cagon secretion in the patients with NIDDM, we ex(cid:173)
`tended the study to patients with insulin-dependent
`diabetes mellitus (IDDM).
`
`Study Subjects
`
`METHODS
`
`The study protocols were approved by the Ethics Committee of
`the Karolinska Hospital, and all subjects gave written informed
`consent. Twenty-five subjects (eight normal subjects, nine patients
`with NIDDM, and eight patients with IDDM) participated in the
`study; their characteristics are shown in Table I. The patients with
`diabetes were recruited from among those attending an outpatient
`
`MPI EXHIBIT 1054 PAGE 1
`
`MPI EXHIBIT 1054 PAGE 1
`
`
`
`Vol. 326 No. 20
`
`ANTIDIABETOGENIC EFFECT OF GLIP IN DIABETES MELLITUS - GUTNIAK ET AL.
`
`1317
`
`Table 1. Characteristics of the Three Study Groups.*
`
`after short-term normalization of the blood glucose concentration. 19
`Insulin (0.8 mU per kilogram per minute) was infused for four
`hours, with or without GLIP (0.75 pmol per kilogram per minute).
`These experiments were conducted in random order, 14 to 28 days
`apart. The blood glucose concentration was kept at 4.7 mmol per
`liter. Glucose utilization was calculated during the last three hours
`of the insulin infusion.
`
`Assays
`
`Blood samples were collected in plastic tubes containing EDT A
`(0.048 ml, 0.34 M) and aprotinin (Trasylol containing 1000 IU of
`kallikrein inhibitor; Bayer, Leverkusen, Germany) and immediately
`placed on ice. The samples were centrifuged at 4°C, and the plasma
`was frozen at -20°c.
`Blood glucose concentrations were measured according to the
`glucose oxidase method. 20 Hemoglobin A1c was measured by iso(cid:173)
`electric focusing. 21 Plasma C-peptide concentrations were deter(cid:173)
`mined by radioimmunoassay with commercially available kits
`(Novo Research Institute, Bagsv~rd, Denmark). The intraassay
`coefficient of variation was 6 percent, and the interassay coefficient
`of variation was 7 percent; the cross-reactivity of proinsulin in this
`assay was 75 percent. In the normal subjects, plasma insulin was
`measured by radioimmunoassay in which the intraassay coefficient
`of variation was 5 percent, the interassay variation was 10 percent,
`and the cross-reactivity of proinsulin was 80 percent. There was no
`cross-reactivity between C peptide and insulin in these two assays.
`In the patients, plasma free insulin was measured after insulin(cid:173)
`antibody-insulin complexes were precipitated with polyethylene
`glycol. 22 Plasma glucagon was measured by radioimmunoassay
`with the antibody 30K. 23 The intraassay coefficient of variation was
`5 percent, and the interassay variation was 14 percent; the lower
`limit of detection was 50 ng per liter. Somatostatin was measured in
`acid ethanol extracts of plasma24•25 by radioimmi.moassay with tyro(cid:173)
`sine-I somatostatin (kindly provided by Dr. A. Arimura, Tulane
`University, New Orleans) labeled with iodine-125, synthetic so(cid:173)
`matostatin as the assay standard, and somatostatin-14 antibody
`produced in our laboratory. 26 The limit of detection was 0.32 pmol
`per assay tube. The interassay coefficient of variation was 7 per(cid:173)
`cent, the intraassay variation 5 percent, and the recovery 84 to 91
`percent.
`•
`GLIP was measured in extracts of plasma by radioimmunoassay
`with synthetic GLP-1 (PG (78-107) amide, code 7168; Peninsula
`Laboratories) as the assay standard, antiserum 2135 (final dilution,
`1: 150,000), and synthetic GLP-1 labeled with iodine-125 according
`to the stoichiometric chloramine-T method and purified by reverse(cid:173)
`phase high-performance liquid chromatography on a Vydac C-18
`column (Separations Group, London) for 100 minutes with a 30 to
`50 percent gradient of acetonitrile in water (Grade S. Rathburn
`Chemicals, Walkerburn, United Kingdom). The antiserum used
`cross-reacts with equal strength with all peptides containing the
`GLP-1 sequence, regardless of the presence or absence of amino(cid:173)
`terminal or carboxy-terminal extensions. The antiserum against
`GLP-1 did not cross-react with glucagon or secretin. The limit of
`detection of the assay in plasma was 5 pmol per liter. The intraassay
`coefficient of variation was 8 percent, and the interassay coefficient
`16 percent. GLIP was extracted from plasma with the use of 70
`percent ethanol (vol/vol, final concentration). The supernatant was
`dried in a vacuum centrifuge (Heto, Hillerod, Denmark) and redis(cid:173)
`solved in veronal buffer (20 mM, pH 8.4) containing 0.1 percent
`bovine serum albumin (A-7034, Sigma Chemical, St. Louis) and
`thimerosal (0.6 mM). All plasma extracts were assayed in duplicate.
`The mean (±SD) recovery of GLIP added to plasma before extrac(cid:173)
`tion was 75±8 percent. 16
`The results of the assays for somatostatin and GLIP were correct(cid:173)
`ed for losses that occurred during extractions. All samples from each.
`participant were analyzed at the same time.
`
`Statistical Analysis
`
`Results are expressed as means ±SE unless otherwise indicated.
`Testing for significant differences was carried out with Student's
`t-test for paired data. Comparisons between groups of subjects were
`performed with Student's unpaired I-test or the Mann-Whitney
`
`CHARACTERISTIC
`
`Sex (M/F)
`Age (yr)
`Mean
`Range
`Body-mass indext
`Mean
`Range
`Duration of diabetes (yr)
`Mean
`Range
`Hemoglobin A1c (%)+
`Mean
`Range
`
`NORMAL
`SUBJECTS
`(N ~ 8)
`
`5/3
`
`52±12
`29-63
`
`6/3
`
`57±8
`44-67
`
`25.4±2.3
`22.3-28.4
`
`32.9±8.4
`22.8-44.8
`
`9.5±4.8
`5-21
`
`8.5±3.3
`6.1-12.0
`
`PATIENTS
`WITH NIDDM
`(N ~ 9)
`
`PATIENTS
`WITH IDDM
`(N ~ 8)
`
`6/2
`
`36±14
`21-49
`
`23.7±2.0
`21.7-29.0
`
`23.0± 13.0
`5-35
`
`6.2±0.9
`5.6-7.0
`
`*Plus-minus values are means ±SD.
`tThe ratio of the weight in kilograms to the square of the height in meters.
`tNormal value, <5.7 percent.
`
`clinic, and their illness fulfilled the criteria for NIDDM and IDDM
`described by the National Diabetes Data Group. 17 None of the
`patients had impaired renal function, autonomic neuropathy, or
`proliferative retinopathy, and all had normal liver function. The
`patients with IDDM had undetectable plasma C-peptide concen(cid:173)
`trations at base-line evaluation and after oral glucose administra(cid:173)
`tion. All 17 patients were being treated with NPH insulin and regu(cid:173)
`lar insulin. They were instructed to follow a standard diet for
`patients with diabetes for at least two weeks before the study and
`during the study. The normal subjects continued to follow their
`usual diet. The injections of NPH insulin were stopped 24 hours
`before the studies, and blood glucose concentrations were controlled
`with subcutaneous injections of regular insulin.
`After each subject had fasted overnight, three cannulas were in(cid:173)
`serted at 7:30 a.m. on the day of each study. One cannula was
`placed in an antecubital vein and used to sample blood intermittent(cid:173)
`ly for hormone assays. The cannula was flushed with saline after
`each sampling. A second cannula, inserted retrogradely in a dorsal
`vein of the hand, was used for continuous monitoring of blood
`glucose concentrations. The venous blood was arterialized by heat(cid:173)
`ing the forearm and hand in a thermoregulated sleeve (Kanthal
`Medical Heating, Stockholm) at 45°C. 18 The third cannula was
`inserted in the contralateral antecubital vein and was used for all
`infusions. From approximately 8 a.m. to the end of the study, the
`patients (but not the normal subjects) were connected to a Biostator
`(Miles, Diagnostic Division, Elkhart, Ind.), a closed-loop insulin(cid:173)
`infusion system (artificial pancreas), and received insulin intrave(cid:173)
`nously to keep their basal and postprandial blood glucose concen(cid:173)
`trations normal. The target range for blood glucose concentrations
`was 4 to 5 mmol per liter under basal conditions and 6 to 7 mmol per
`liter after the study meal. The experiments were started 30 minutes
`after normoglycemia was achieved, which was a mean (±SE) of
`58±5 minutes after connection to the Biostator in the patients with
`IDDM and 109± 16 minutes in the patients with NIDDM, or 60
`minutes after insertion of the cannulas in the normal subjects. The
`latter were not connected to the Biostator and received no insulin,
`but were otherwise studied in the same way. An infusion of saline or
`GLIP (Peninsula Laboratories, St. Helens, Merseyside, United
`Kingdom) at a rate of 0.75 pmol per kilogram of body weight per
`minute was then started and continued for 3V2 hours. The two
`studies were performed in a random order 6 to 28 days apart. At
`time zero all participants were given a standard lunch, which they
`ate within 15 minutes while sitting in bed. The meal consisted of
`boiled potatoes, boiled beef, cooked carrots, a glass of milk co?tain(cid:173)
`ing 0.5 percent butterfat, and a slice of bread made from a mixture
`of wheat and rye flours; 28, 26, and 46 percent of the energy from
`this lunch were derived from protein, fat, and carbohydrates, re(cid:173)
`spectively. Blood samples were obtained at -30, 0, 15, 30, 90, 120,
`150, and 180 minutes.
`In the patients with IDDM, insulin sensitivity was measured
`during hyperinsulinemic-normoglycemic-clamp studies performed
`
`MPI EXHIBIT 1054 PAGE 2
`
`MPI EXHIBIT 1054 PAGE 2
`
`
`
`
`
`
`
`
`
`
`
`1322
`
`THE NEW ENGLAND JOURNAL OF MEDICINE
`
`May 14, 1992
`
`7. Creutzfeldt W, Ebert RF. New developments in the incretin concept. Diabe(cid:173)
`tologia 1985;28:565-73.
`8. Conlon JM. Proglucagon-derived peptides: nomenclature, biosynthetic rela(cid:173)
`tionships and physiological roles. Diabetologia 1988;31:563-6.
`9. Drucker DJ. Glucagon and the glucagon-like peptides. Pancreas 1990;5:
`484-8.
`IO. Mojsov S, Weir GC, Habener JF. Insulinotropin: glucagon-like peptide
`I (7-37) co-encoded in the glucagon gene is a potent stimulator of
`insulin release in the perfused rat pancreas. J Clin Invest 1987;79:616-
`9.
`11. Holst JJ, 0rskov C, Vagn Nielsen 0, Schwartz TW. Truncated glucagon(cid:173)
`like peptide I, an insulin-releasing hormone from the distal gut. FEBS Lett
`!987;211:169-74.
`12. Kreymann B, Williams G, Ghatei MA, Bloom SR. Glucagon-like peptide-I
`7-36: a physiological incretin in man. Lancet 1987;2: 1300-4.
`13. Takahashi H. Manaka H. Suda K, et al. Radioimmunoassay for glucagon(cid:173)
`like peptide-I in human plasma using N-terminal and C-terminal directed
`antibodies: a physiologic insulinotropic role ofGLP-1(7-36 amide). Biomed
`Res 1990; I I :99-108.
`14. 0rskov C, Bersani M, Johnsen AH, Hojrup P, Holst JJ. Complete se(cid:173)
`quences of glucagon-like peptide-I from human and pig small intestine.
`J Biol Chem 1989;264:12826-9.
`15. 0rskov C, Holst JJ, Seier Poulsen S, Kirkegaard P. Pancreatic and intestinal
`processing of proglucagon in man. Diabetologia 1987;30:874-81.
`16. 0rskov C, Jeppesen J, Madsbad S, Holst JJ. Proglucagon products in plas(cid:173)
`ma of noninsulin-dependent diabetics and nondiabetic controls in the fasting
`state and after oral glucose and intravenous arginine. J Clin Invest 1991 ;87:
`415-23.
`17. National Diabetes Data Group. Classification and diagnosis of diabetes
`mellitus and other categories of glucose intolerance. Diabetes 1979;28:
`1039-57.
`18. Gutniak M, Blomqvist G, Widen L, Stone-Elander S, Hamberger B, Grill
`V. D-[U-1 IC]glucose uptake and metabolism in the brain of insulin-depend(cid:173)
`ent diabetic subjects. Am J Physiol 1990;258:E805-E812.
`19. Verdonk CA, Rizza RA, Westland RE, Nelson RL, Gerich JE, Service FJ.
`Glucose clamping using the Biostator GCIIS. Horm Metab Res 1980;12:
`133-5.
`20. Huggett ASG, Nixon DA. Use of glucose oxidase, peroxidase, and
`o-dianisidine in determination of blood and urinary glucose. Lancet 1957;2:
`368-70.
`2 I. Simon M, Cuan J. Hemoglobin A,c by isoelectric focusing. Clin Chem
`1982;28:9-12.
`22. Gennaro WD, Van Norman JD. Quantitation of free, total and anti(cid:173)
`body-bound insulin in insulin-treated diabetics. Clin Chem 1975;2 I :873-
`9.
`23. Faloona GR, Unger RH. Glucagon. In: Jaffe BM, Behrman HE, eds. Meth(cid:173)
`ods of hormone radioimmunoassay. New York: Academic Press, 1974:317-
`30.
`24. Penman E, Wass JAH, Lund A, et al. Development and validation of a
`specific radioimmunoassay for somatostatin in human plasma. Ann Clin
`Biochem 1979;16: 15-25.
`25. Arimura A, Lundqvist G, Rothman J, et al. Radioimmunoassay of somato(cid:173)
`statin. Metabolism 1978;27:Suppl: 1139-44.
`26. Efendic S, Nylen A, Roovete A, Uvniis-Wallenstein K. Effects of glucose
`and arginine on the release of immunoreactive somatostatin from the isolat(cid:173)
`ed perfused rat pancreas. FEBS Lett 1978;92:33-5.
`
`27. Gutniak MK, Karlander SG, Efendic S. Glyburide decreases insulin require(cid:173)
`ment, increases beta-cell response to mixed meal, and does not affect insulin
`sensitivity: effects of short- and long-term combined treatment in secondary
`failure to sulfonylurea. Diabetes Care 1987;10:545-54.
`28. Sutherland EW, Rall TW. The relation of adenosine-3' ,5' -phosphate and
`phosphorylase to the actions of catecholamines and other hormones. Phar(cid:173)
`macol Rev 1960;12:265-99.
`29. Miller LL. Glucagon: a protein catabolic hormone in the isolated perfused
`rat liver. Nature 1960;185:248.
`30. Efendic S, Lins PE, Luft R, Uvniis-Wallensten K, Szecowka J. Somatostat(cid:173)
`paracrine or endocrine substance? In: Creutzfeldt W, ed. The entero(cid:173)
`in -
`insular axis. Vol. 7 of Frontiers of hormone research. Basel, Switzerland:
`S. Karger, 1980:41-51.
`31. Johansson C, Wisen 0, Efendic S, Uvniis-Wallensten K. Effects of somato(cid:173)
`statin on gastrointestinal propagation and absorption of oral glucose in man.
`Digestion 1981;22:126-37.
`32. Krejs GJ, Browne R, Raskin P. Effect of intravenous somatostatin on je(cid:173)
`juna! absorption of glucose, amino acids, water, and electrolytes. Gastroen(cid:173)
`terology l 980;78:26-31.
`33. 0rskov C, Holst JJ, Nielsen OV. Effect of truncated glucagon-like peptide- I
`[proglucagon (78-107) amide] on endocrine secretion from pig pancreas,
`antrum, and nonantral stomach. Endocrinology 1988;123:2009-13.
`34. Suzuki S, Kawai K, Ohashi S, Mukai H, Yamashita K. Comparison of the
`effects of various C-terminal and N-terminal fragment peptides of glucagon(cid:173)
`like peptide- I on insulin and glucagon release from the isolated perfused rat
`pancreas. Endocrinology 1989;125:3109-14.
`35. Fehmann HC, Habener JF. Functional receptors for the insulinotropic hor(cid:173)
`mone glucagon-like peptide-1(7-37) on somatostatin secreting cell line.
`FEBS Lett 1991;279:335-40.
`36. Gutniak MK, Grill V, Wiechel KL, Efendic S. Basal and meal-induced
`somatostatin-like immunoreactivity in healthy subjects and in IDDM and
`totally pancreatectomized patients: effects of acute blood glucose normaliza(cid:173)
`tion. Diabetes 1987;36:802-7.
`37. Wettergren A, Schjoldager B, Mortensen PE, Myhre J, Christiansen J,
`Holst JJ. Truncated GLP-1 (proglucagon 78- 107 amide) from distal gut: role
`in regulation of gastric and pancreatic functions. Digestion 1990;46:
`Suppl:120. abstract.
`38. Hansen IL, Cryer PE, Rizza RA. Comparison of insulin-mediated and glu(cid:173)
`cose-mediated glucose disposal in patients with insulin-dependent diabetes
`mellitus and in nondiabetic subjects. Diabetes l985;34:751-5.
`39. DeFronzo RA, Simonson D, Ferrannini E. Hepatic and peripheral insulin
`resistance: a common feature of type 2 (non-insulin-dependent) and type I
`(insulin-dependent) diabetes mellitus. Diabetologia 1982;23:313-9.
`40. Finegood DT, Bergman RN, Vranic M. Modeling error and apparent iso(cid:173)
`tope discrimination confound estimation of endogenous glucose production
`during euglycemic glucose clamps. Diabetes .,988;37:1025-34.
`41. Polonsky KS, Rubenstein AH. C-peptide as a measure of the secretion and
`hepatic extraction of insulin: pitfalls and limitations. Diabetes l 984;33:486-
`94.
`42. Lebovitz HE, Feinglos MN. Sulfonylurea drugs: mechanism of antidiabetic
`action and therapeutic usefulness. Diabetes Care l978;1:189-98.
`43. Robertson DA, Hale PJ, Nattrass M. Macrovascular disease and hyperinsu(cid:173)
`linaemia. Baillieres Clin Endocrinol Metab 1988;2:407-24.
`44. Scheen AJ, Lefebvre PJ. Insulin versus insulin plus sulfonylureas in type 2
`diabetic patients with secondary failure to sulfonylureas. Diabetes Res Clin
`Pract l 989;6:S33-S42.
`
`MPI EXHIBIT 1054 PAGE 7
`
`MPI EXHIBIT 1054 PAGE 7
`
`