Rapid Publication
`
`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
`Svetlana Mojsov,* Gordon C. Weir,* and Joel F. Haboner*
`*Laboratory ofMolecular Endocrinology, Massachusetts General Hospital and Howard Hughes Medical Institute, Harvard Medical
`School, Boston, Massachusetts 02114; and tJoslin Diabetes Center and New England Deaconess Hospital, and Harvard Medical School,
`Boston, Massachusetts 02115
`
`Abstract
`
`Insulin secretion is controlled by a complex set of factors that
`include not only glucose but amino acids, catecholamines, and
`intestinal hormones. We report that a novel glucagon-like pep-
`tide, co-encoded with glucagon in the glucagon gene is a potent
`insulinotropic factor. The glucagon gene encodes a proglucagon
`that contains in its sequence glucagon and additional glucagon-
`like peptides (GLPs). These GLPs are liberated from proglu-
`cagon in both the pancreas and intestines. GLP-I exists in at
`least two forms: 37 amino acids GLP-I(1-37), and 31 amino
`acids, GLP-I(7-37). We studied the effects of synthetic GLP-Is
`on insulin secretion in the isolated perfused rat pancreas. In the
`presence of 6.6 mM glucose, GLP-I(7-37) is a potent stimulator
`of insulin secretion at concentrations as low as 5 X 10-11 M (3-
`to 10-fold increases over basal). GLP-I(1-37) had no effect on
`insulin secretion even at concentrations as high as 5 X 10-7 M.
`The earlier demonstration of specific liberation of GLP-I(7-37)
`in the intestine and pancreas, and the magnitude of the insuli-
`notropic effect at such low concentrations, suggest that GLP-
`I(7-37) participates in the physiological regulation of insulin se-
`cretion.
`
`Introduction
`
`Pancreatic glucagon and intestinal glicentin are synthesized in
`the form of a 180-residue protein, preproglucagon encoded in
`a single gene (1). The precursor contains in addition to glicentin
`and glucagon the sequences of two glucagon-like peptides
`(GLPs)', GLP-I and GLP-II, separated by an intervening peptide
`(IP-I) (2-5). The posttranslational processing of preproglucagon
`differs in pancreas and intestine (1, 6). In the pancreas the pre-
`cursor is processed to glucagon and GLP-I, and in both large
`and small intestines glicentin, GLP-I, GLP-II and IP-II-leucine-
`amide are found. Both pancreas and intestine contain GLP-I in
`at least two forms-31 and 37 residues long (1).
`The close similarity of the amino acid sequence of GLP-Is
`and GLP-II with glucagon and the other peptides related in
`structure to glucagon (secretin, vasoactive intestinal peptide,
`gastric inhibitory peptide, growth hormone-releasing hormone)
`
`Receivedfor publication 28 October 1986.
`
`1. Abbreviation used in this paper: GLP, glucagon-like peptide.
`
`J. Clin. Invest.
`© The American Society for Clinical Investigation, Inc.
`$1.00
`0021-9738/87/02/0616/04
`Volume 79, February 1987, 616-619
`
`616
`
`S. Mojsov, G. C. Weir, and J. F. Habener
`
`suggests that the GLPs might have a role in metabolic regulation.
`The specific liberation of GLP-I and GLP-II in the intestine
`indicates that these peptides may be components of the entero-
`insular axis (7), which comprises multiple intestinal factors in-
`fluencing the release of hormones produced in the pancreatic
`islets. Further, they may be incretins, endocrine transmitters
`produced in the gastrointestinal tract that are released by nu-
`trients and stimulate insulin secretion in the presence ofelevated
`glucose ifexogeneously infused in amounts not exceeding blood
`levels achieved after food ingestion (8). Detection ofboth GLP-
`I(1-37) and GLP-I(7-37) in pancreas and intestines raises the
`possibility that GLP-I( 1-37) is itselfa prohormone that undergoes
`a proteolytic cleavage at the single arginine residue at position
`6 to release the biologically active GLP-I(7-37). In these studies
`we used synthetic GLP-I(1-37) and GLP-I(7-37) to investigate
`their effects on insulin secretion in the perfused rat pancreas
`and find that GLP-I(7-37) has uniquely potent insulinotropic
`actions.
`
`Methods
`
`Synthesis of peptides. Glucagon and GLP-Is were synthesized by the
`stepwise solid-phase method (9). Because the assembly of the peptide
`chain proceeds in the carboxyl- to the amino-terminal direction, GLP-
`I( 1-37) and GLP-I(7-37) were prepared in the same synthesis by separating
`the peptide resin after incorporation of a protected histidyl residue at
`position 7 and continuing the assembly of amino acids on the other
`aliquot of the peptide resin to obtain protected GLP-I(1-37) peptide resin.
`Peptides were purified by preparative reverse-phase C- 18 chromatography.
`Purified peptides were shown to be homogeneous by amino acid analysis,
`preview-sequence analysis, and high performance liquid chromatography
`(HPLC) on reverse-phase C-18 and ion-exchange DEAE-52 columns.
`Radioimmunoassays. Development of the antisera and competitive
`binding radioimmunoassays for glucagon and GLP-I are described else-
`where (1). In brief, samples were incubated with the antisera in borate
`buffer (pH 8.1) for 24 h at 0°C, followed by addition of '25I-labeled
`peptide for an additional 24 h in a total volume of 0.5 ml. Separation
`of the antibody bound from the free peptide was accomplished with
`dextran-coated charcoal. Assay sensitivity with all three antisera was 10
`pg/ml. The antiserum against GLP-I was obtained by immunization
`with GLP-I(l-37) and is directed against both the amino-terminal (1-6)
`part of the molecule and to 7-37 determinants. Therefore, the amount
`ofGLP-I(7-37) may be over or underestimated with respect to GLP-I(1-
`37) in the assay. The assay for insulin (10) used charcoal separation and
`rat insulin standards (Novo Research Institute, Copenhagen, Denmark).
`Rat-perfused pancreas experiments. The preparation of the in situ
`isolated rat pancreas has been described previously (11, 12). The perfusate
`contained bicarbonate buffer (pH 7.4) and 120 mg/dl glucose, 4% dextran
`T-70, and 0.2% bovine serum albumin, and was equilibrated with 95%
`02 and 5% CO2. The first 20 min ofeach perfusion was an equilibration
`period and is not represented in the data graphs.
`
`MPI EXHIBIT 1036 PAGE 1
`
`MPI EXHIBIT 1036 PAGE 1
`
`MPI EXHIBIT 1036 PAGE 1
`
`

`

`After the initial 20-min equilibration period, aliquots of perfusate
`were removed every 2-4 min for additional 20 min, thus allowing the
`system to equilibrate for a total of 40 min. The perfusion, with GLP-
`I(1-37) or GLP-I(7-37), was for 6 min and samples were collected at 1-
`min intervals. The peptide perfusions were followed by equilibration
`periods of 20 min, during which four samples 5 min apart were collected.
`A second 6-min perfusion followed with the same peptide as the first
`perfusion only at 100 times higher concentration ofpeptide. Again, sam-
`ples 1 min apart were collected. The entire perfusion time was between
`70 and 85 min.
`In each aliquot of perfusate obtained, insulin was determined by
`radioimmunoassay. In addition the efficiency of delivery of the GLP-Is
`was confirmed by radioimmunoassay of corresponding aliquots of per-
`fusate in which insulin was measured (1).
`
`Results
`
`To optimally study the effects of GLP-I(7-37) and GLP-I( 1-37)
`on insulin secretion we used separate perfusions with each pep-
`tide, perfusing twice at two different concentrations of peptides
`and allowing 20-min time intervals between the two perfusions.
`In perfusions of two separate pancreases using this protocol,
`GLP-I(7-37) was a potent stimulator of insulin secretion, giving
`about a 20-fold stimulation at 5 X 10-9 M and a sixfold stim-
`ulation at 5 X 10" M (Fig. 1). In comparson, GLP-I(1-37),
`also studied in two pancreases, showed no effect on insulin se-
`cretion at either 5 X 10-9 or 5 X l-7 M (Fig. 2). At the latter
`concentration no effect was observed even during a 15-min per-
`fusion period (Fig. 2 B).
`Using a slightly different perfusion protocol than that de-
`scribed above (Figs. 1 and 2) we gave alternate 5-min infusions
`of the peptides at concentrations ranging from 5 X lo-7 to 5
`X 10-12 M to five additional individual pancreases. We repro-
`ducibly observed insulin release in response to GLP-I(7-37) at
`concentrations as low as 5 X 10-" M, and little if any insulin
`responses to GLP-I(1-37) at concentrations as high as 5 X l0-7
`M. Thus, the potent insulinotropic actions ofGLP-I(7-37) have
`been observed in studies of seven separate pancreases.
`Effects of glucagon on insulin secretion in the perfused pan-
`creas have been established previously (13). We also compared
`the effects of glucagon to that of the GLP-Is. We used synthetic
`
`glucagon in the concentration range of I0- -10-7 M and found
`it to be less potent than GLP-I(7-37).
`
`Discussion
`
`The results ofthese studies clearly indicate that GLP-I(7-37) has
`potent insulinotropin activity. The liberation ofthis peptide from
`proglucagon in the intestine, and to a lesser extent in the pancreas
`(1), raises the possibility that GLP-I(7-37) has a role in endocrine
`regulation in the entero-insular axis (7). Our data, taken together
`with earlier observations that glucagon-like immunoreactivity
`in crude gut extracts released insulin after ingestion of glucose
`and fat, (8) suggest that GLP-I(7-37) could potentially be an
`incretin. Of all the known intestinal hormones tested for their
`insulin-releasing potency in the past, gastric inhibitory peptide
`has been considered as a possible incretin (14, 15). However,
`the concentrations ofgastric inhibitory peptide required to stim-
`ulate insulin secretion exceed the physiologic levels ofthe peptide
`achieved after a meal. In the rat-perfused pancreas in the presence
`of 8.9 mM glucose, gastric inhibitory peptide (I0-' M) increased
`insulin secretion sixfold (16). We find a comparable increase in
`insulin secretion with GLP-I(7-37) at concentrations 100-fold
`lower than those required for an insulinotropic response to gastric
`inhibitory peptide.. By radioimmunoassay we have measured
`both GLP-I(1-37) and GLP-I(7-37) levels of - 150 pg/ml (50
`pM) in rat portal blood and 50 pg/ml (15 pM) in peripheral
`blood (S. Mojsov, unpublished results). Therefore, the insuli-
`notropic effect that we have observed at concentrations ofGLP-
`I(7-37) ofbetween 5 and 50 pM are well within the physiological
`levels of GLP-I(7-37) found in the circulation.
`There has been considerable interest in the potential intra-
`islet relationships which might occur between A, B, and D cells,
`such that the secretory product of one cell type might influence
`the function of a neighboring cell (17). Interaction could take
`place via a paracrine mechanism or through a local intra-islet
`portal system. Glucagon can stimulate both insulin and soma-
`tostatin secretion (13, 18), but because there appears to be a
`functional compartmentalization between islet cells, it is unclear
`whether glucagon can actually reach B and D cells (19). Taking
`into account the vascular arrangement of the rat islet, the glu-
`
`C-93Figure 1. The effects of sepa-
`rate perfusions in two represen-
`tative pancreas GLP-I(7-37) at
`two concentrations, 5 X 10-"
`and 5 x 10-9 M. Solid lines,
`insulin values determined by
`radioimmunoassay. Dashed
`lines, amount of peptide per-
`fused as determined in a com-
`petitive binding radioimmu-
`noassay with antisera against
`GLP-I(1-37). The amount of
`GLP-I(7-37) at 5 X 10" M is
`beyond the detection sensitivity
`of the radioimmunoassay. Each
`graph represents a perfusion of
`a separate pancreas with a
`given peptide.
`
`i-8 I
`tn
`Iii
`3
`-.4-L
`+-2
`
`60
`
`Insulinotropin
`
`617
`
`Bn
`
`16
`
`-17
`
`2
`
`K
`
`o
`
`A
`
`2 2 2
`
`?4
`
`E1
`
`16
`,I
`ol
`'S
`C:
`12,-
`-I
`CD
`C:
`
`Time (min)
`
`20
`
`40
`Time (min)
`
`MPI EXHIBIT 1036 PAGE 2
`
`MPI EXHIBIT 1036 PAGE 2
`
`MPI EXHIBIT 1036 PAGE 2
`
`

`

`B
`
`5 x 10-9
`
`- 7
`
`-i250
`
`-2IC
`
`- 7C
`
`-13C r
`
`-90
`
`-5C 3
`TI C
`
`1
`1%
`Ii
`
`I0
`
`1
`
`11
`
`I
`
`6C
`
`A
`
`5
`
`I---9
`
`~~ 1
`1
`~
`x
`E ~ I.
`
`0
`
`20
`
`40
`Time (min)
`
`I90
`
`- 150;
`
`270
`j30
`lI 5
`t;10
`5
`60
`
`Figure 2. The effects of separate
`perfusions in two representative
`pancreas with GLP-I(1-37) at
`two concentrations, 5 X 10-9 and
`5 X 10-7 M. Details of the exper-
`iment and explanation of sym-
`bols are described in legend to
`Fig. 1.
`
`1JC1
`
`Time (min)
`
`and radioimmunoassay, respectively. We thank Esther Hoomis for typing
`the manuscript.
`The studies were supported in part by Public Health Service grants
`AM 30834 and AM 20349.
`
`Note added in proof In studies ofisolated perfused pig ileum and pancreas,
`Orskov et al. (25) recently reported finding secretion of a GLP-l peptide
`from ileum, but the pancreas secreted a large peptide with both GLP-I
`and GLP-2 immunoreactivity.
`
`References
`
`1. Mojsov, S., G. Heinrich, I. B. Wilson, M. Ravazzola, L. Orci, and
`J. F. Habener. 1986. Preproglucagon gene expression in pancreas and
`intestine diversifies at the level ofpost-transcriptional processing. J. Biol.
`Chem. 261:11880-11889.
`2. Heinrich, G., P. Gros, and J. F. Habener. 1984. Glucagon gene
`sequence: four of six exons encode separate functional domains of rat
`preproglucagon. J. Biol. Chem. 259:14082-14084.
`3. Bell, G. I., R. F. Santerre, and G. T. Mullenbach. 1983. Hamster
`preproglucagon contains the sequence of glucagon and two related pep-
`tides. Nature (Lond.). 302:716-718.
`4. Bell, G. I., R. Sanchez-Pescador, P. J. Laybourn, and R. C. Najarian.
`1983. Exon duplication and divergence in the human preproglucagon
`gene. Nature (Lond.). 304:368-371.
`5. Lopez, L. C., M. L. Frazier, C. J. Su, A. Humar, and G. F. Sanders.
`1983. Mammalian pancreatic preproglucagon contains three glucagon-
`related peptides. Proc. Nail. Acad. Sci. USA. 80:5485-5489.
`6. George, S. K., L. D. Uttenthal, M. Ghiglione, and S. R. Bloom.
`1985. Molecular forms of glucagon-like peptides in man. FEBS (Fed.
`Eur. Biochem. Soc.) Lett. 192:275-278.
`7. Unger, R. H., and A. M. Eisentraut. 1969. Entero-insular axis.
`Arch. Intern. Med. 123:261-266.
`8. Creutzfelt, W. 1979. The incretin concept today. Diabetologia.
`16:75-85.
`9. Merrifield, R. B. 1963. Solid phase peptide synthesis. J. Am. Chem.
`Soc. 95:2149-2154.
`10. Albano, J. D. M., R. P. Ekins, G. Maritz, and R. C. Turner.
`1972. A sensitive, precise radioimmunoassay of serum insulin, relying
`on charcoal separation of bound and free hormone moities. Acta En-
`docrinol. 70:487-509.
`11. Weir, G. C., S. D. Knowlton, and D. B. Martin. 1974. Glucagon
`secretion from the perfused rat pancreas. Studies with glucose and cat-
`echolamines. J. Clin. Invest. 54:1403-1412.
`
`cagon-containing A cells ofthe mantle appear to be downstream
`from the B cells of the core, and therefore glucagon may not
`reach the B cells in high enough concentration to exert a sig-
`nificant influence (20). The mantle ofA and D cells are, however,
`adjacent and this makes the possibility of paracrine interaction
`more feasible, although experimental support for such an inter-
`action is not available. The finding that GLP-I(7-37) is a more
`potent insulin secretagogue than glucagon raises important
`questions about its potential intra-islet role.
`Amino and carboxyl-termini of glucagon, GLP-I(7-37) and
`GLP-II are closely related to each other in their amino acid
`sequences and to vasointestinal peptide that possibly exerts a
`neuronal stimulation of insulin secretion (21). A most striking
`similarity among them is the conservation of a histidine residue
`at position 1. It is noteworthy that gastric inhibitory peptide,
`also closely related in its structure to the GLPs, has a tyrosine
`residue at position 1 instead of histidine (22). Inasmuch as a
`histidine residue at this position is essential for adenylate cyclase
`stimulation in various systems, the greater insulinotropic potency
`of GLP-1(7-37) compared with GIP may in part be accounted
`for by the histidine substitution for tyrosine (23).
`Additional evidence in support ofthe concept that GLP-I(7-
`37) is a potent insulinotropic peptide is provided by our recent
`observation that GLP-I(7-37), and not GLP-I(1-37) or GLP-II,
`is a potent activator of adenylate cyclase at concentrations as
`low as 5 X 10-11 M and also stimulates cellular levels of insulin
`mRNA and insulin release in a rat insulinoma cell line (RIN-
`38) (Drucker, D. J., J. Philippe, S. Mojsov, W. L. Chick, and
`J. F. Habener, manuscript in preparation). Further, studies by
`Schmidt and co-workers showed that in isolated precultured is-
`lets, 10-' to 10-8 M concentrations of the peptide GLP-I( 1-36
`des Gly-Arg amide) were required to release insulin (24).
`Determining whether GLP-I(7-37) is the hormone whose
`primary function is to stimulate insulin secretion in response to
`feeding, or is one of a complex group of hormones involved in
`maintaining glucose homeostasis, will require further investi-
`gation.
`
`Acknowledaments
`
`We thank Henrietta Cooper, Deanna Deery, and Adacie Allen for their
`expert experimental assistance in carrying out the pancreatic perfusions
`
`618
`
`S. Mojsov, G. C. Weir, and J. F. Habener
`
`MPI EXHIBIT 1036 PAGE 3
`
`MPI EXHIBIT 1036 PAGE 3
`
`MPI EXHIBIT 1036 PAGE 3
`
`

`

`12. Penkos, J. C., C.-H. Wu, J. C. Basabe, N. Lopez, and F. W. Wolf.
`1969. A rat pancreas-small gut preparation for the study of intestinal
`factor(s) and insulin release. Diabetes. 18:733-738.
`13. Samols, E., G. Marri, and V. Marks. 1966. Interrelationship of
`glucagon, insulin, and glucose: the insulinogenic effect of glucagon. Di-
`abetes. 15:855-865.
`14. Dupre, J., S. A. Ross, D. Watson, and J. C. Brown. 1973. Stim-
`ulation of insulin secretion by gastric inhibitory polypeptide in man. J.
`Clin. Endocrinol. Metab. 37:826-828.
`15. Andersen, D. K., D. Elahi, J. C. Brown, J. D. Tobin, and R.
`Andres. 1978. Oral glucose augmentation ofinsulin secretion. Interactions
`ofgastric inhibitory polypeptide with ambient glucose and insulin levels.
`J. Clin. Invest. 62:152-161.
`16. Brown, J. C., and R. A. Federson. 1976. The insulinotropic action
`of gastric inhibitory polypeptide in the perfused isolated rat pancreas.
`Endocrinology. 99:780-785.
`17. Samols, E., G. C. Weir, and S. Bonner-Weir. 1983. Intraislet
`insulin-glucagon-somatostatin relationships. In P. J. Lefebvre, editor.
`Handbook of Experimental Pharmacology. Vol. 66. Springer-Verlag,
`Berlin. 133-173.
`18. Unger, R. H., and L. Orci. 1977. Hypothesis: the possible role
`of the pancreatic D-cell in the normal and diabetic states. Diabetes. 26:
`241-244.
`
`19. Kawai, K., E. Ipp, L. Orci, A. Pezzelet, and R. H. Unger. 1982.
`Circulating somatostatin acts on the islets of Langerhans by way of so-
`matostatin-poor compartment. Science (Wash. DC). 218:477-478.
`20. Bonner-Weir, S., and L. Orci. 1982. New perspectives on the
`microvasculature of the islets of Langerhans in the rat. Diabetes. 31:
`883-889.
`21. Schebalin, M., S. I. Said, and G. M. Makhlouf. 1977. Stimulation
`of insulin and glucagon secretion by vasoactive intestinal peptide. Am.
`J. Physiol. 232:E197-E200.
`22. Moody, A. J., L. Thim, and I. Valverde. 1984. The isolation and
`sequence of human gastric inhibitory peptide (GIP). FEBS (Fed. Eur.
`Biochem. Soc.) Lett. 172:142-148.
`23. Pandol, S. J., H. Seifert, M. W. Thomas, J. Rivier, and W. Vale.
`1984. Growth hormone-releasing factor stimulates pancreatic enzyme
`secretion. Science (Wash. DC). 225:326-328.
`24. Schmidt, W. E., E. G. Siegel, and W. Creutzfeld. 1985. Glucagon-
`like peptide- I but not glucagon-like peptide-2 stimulates insulin release
`from isolated rat pancreatic islets. Diabetologia. 28:704-707.
`25. Orskov, C., J. J. Holst, S. Knuhtsen, F. G. A. Baldissera, S. S.
`Poulsen, and 0. V. Nielsen. 1986. Glucagon-like peptides GLP-1 and
`GLP-2, predicted products of the glucagon gene, are secreted separately
`from pig small intestine but not pancreas. Endocrinology. 119:1467-
`1475.
`
`Insulinotropin
`
`619
`
`MPI EXHIBIT 1036 PAGE 4
`
`MPI EXHIBIT 1036 PAGE 4
`
`MPI EXHIBIT 1036 PAGE 4
`
`