Islet Amyloid Polypeptide and Insulin Expression Are Controlled Differently in Primary and Transformed Islet Cells
Ole D. Madsen, Jens H. Nielsen, Birgitte Michelsen, Per Westermark, Christer Betsholtz, Masahiro Nishi, and Donald F. Steiner Hagedorn Research Laboratory DK-2820 Gentofte, Denmark Department of Pathology University Hospital (P.W.) Linkoping, Sweden Department of Pathology University Hospital (C.B.) Uppsala, Sweden Department of Biochemistry and Molecular Biology Howard Hughes Medical Institute University of Chicago (M.N., D.F.S.) Chicago, Illinois 60637
clones, which are able to activate insulin gene transcription during passage in vivo. (Molecular Endocrinology 5: 143-148, 1991)
The pancreatic /S-cell is a major site of islet amyloid polypeptide (IAPP) biosynthesis, and the peptide is coreleased with insulin. We have analyzed the expression of IAPP (mRNA and protein) in various cell types in normal and transformed murine islet cell cultures by Northern blot analyses and immunocytochemistry. IAPP is primarily coexpressed with insulin in the /?-cell of GH-promoted primary rat islet cell cultures. Additionally, a small population of non/0-cells exhibited a prominent IAPP expression, and double staining experiments showed colocalization with glucagon or somatostatin in some of these cells. IAPP mRNA was confined to the /?-cell phenotype when analyzing the phenotypically stable in vivo tumor lines, MSL-G2-IN (insulinoma) and MSLG-AN (glucagonoma), and the transgenic mouse islet cell lines, /?-Tc and a-Tc. However, IAPP and insulin expression were completely uncoupled in unstable heterogeneous clones such as NHI-6F. This clone is composed of primarily glucagon-producing cells in vitro, but insulin gene expression becomes dominant after passage in vivo. Interestingly, IAPP was hyperexpressed with glucagon under in vitro conditions in this clone. We conclude that the tissue specificity of expressions of IAPP and insulin are controlled differently, and that coexpression of IAPP with hormones different from insulin may be a marker for pluripotent transformed rat islet cell
INTRODUCTION Islet amyloid polypeptide (IAPP or amylin) is the major component in islet amyloid deposits and has been implicated in the development of type II diabetes (1-4). In the adult pancreas the presence of IAPP seems to be restricted to the pancreatic /3-cell, as determined by immunocytochemistry (5). Although small amounts of extrapancreatic IAPP mRNA (6) and protein (7) have been reported, the islet #-cell seems to be the major site of IAPP synthesis (5, 7, 8), and IAPP is cosecreted with insulin (9, 10). The strict /?-cell-specific expression of insulin is controlled by 5' flanking insulin gene enhancer and promoter elements interacting with putative /?-cell-specific transacting factors, as recently confirmed in a series of transgenic experiments (11). Since there is no obvious homology between the 5' flanking sequences of the human IAPP gene and insulin genes (12,13), expression of the IAPP gene may be regulated by separate factors. The aim of the present study was to compare expression of IAPP and insulin in normal neonatal rat islet cells stimulated by GH (14), in a pluripotent rat islet tumor cell line transfected with the human insulin gene (15, 16), and in SV-40 T-antigen-transformed mouse islet cell lines (17,18).
0888-8809/91/0143-0148$02.00/0 Molecular Endocrinology Copyright© 1991 by The Endocrine Society
143
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 22:23 For personal use only. No other uses without permission. . All rights reserved.
Vol 5 No. 1
MOL ENDO-1991 144
RESULTS Nontransformed Islet Cell Cultures Monolayers of newborn rat islet cells stimulated to proliferate by GH showed highly intense staining for insulin (Fig. 1A). A few scattered cells were strongly stained for glucagon and somatostatin (14). IAPP staining showed a heterogeneous pattern of one or two strongly positive cells scattered among a majority of weakly stained cells (Fig. 1B) in most colonies. Interestingly, the brightly stained IAPP cells were consistently negative for insulin, as visualized by double exposure (Fig. 1C). These cells could be characterized by double immunostaining as either a-cells (not shown) or 5-cells (Fig. 1, E and F). Stable Islet Tumor Phenotypes When analyzing IAPP expression in stable rat insulinomas (15) or glucagonomas of common monoclonal origin (our unpublished data) as well as the /3 (17)- and a (18)-cell lines from transgenic mice, we found a good overall correlation of the presence of IAPP mRNA with the insulin-producing phenotype (Fig. 2). However, on the cellular level the a-Tc culture was heterogeneous, as revealed by discrete colonies producing insulin (Fig. 1H). Only some of these colonies were shown to coexpress IAPP (Fig. 1G). Unstable Islet Tumor Phenotypes The glucagon-producing MSL-G2 cells and human insulin gene-transfected subclone NHI-6F can be induced to produce insulin in vivo. Repassage in vitro caused a reversal in phenotype, with a gradual loss of insulin gene expression (15, 16). Cultures representing these stages were analyzed for insulin, glucagon, and IAPP mRNA distribution (Fig. 3). Surprisingly, the overall IAPP mRNA profile followed that of glucagon, but not that of insulin. These data were confirmed by immunocytochemistry, where the noninsulin-producing primary cells of NHI-6F were highly positive for IAPP immunoreactivity (Fig. 1J). No obvious coupling of IAPP and insulin expression was seen in NHI-6F cultures during the gradual loss of the insulin-expressing phenotype, as visualized in Fig. 1K.
DISCUSSION
The primary site of IAPP expression in the adult seems to be the pancreatic /3-cell (5, 8-10). The /3-cell specificity is possibly less restricted than that of insulin, since IAPP mRNA has been reported outside of the pancreas (6,7). Our goal was to study IAPP expression in various cell types in normal and transformed islet cells by Northern blot analyses and immunocytochemistry. IAPP expression relative to that of insulin in these abnormal
cells may tell us whether two respective genes are influenced by similar /3-cell-specific control mechanisms. In general, we can conclude that this may not be the case. Firstly, nontransformed newborn islet cell cultures are clearly heterogeneous with respect to IAPP expression. While IAPP is weakly identified in most /3-cells, a very small subpopulation of a- as well as 5-cells were highly immunoreactive with IAPP antisera. Whether the phenomenon observed is due to the effect of GH, which is known to induce /3-cell proliferation (14), remains to be clarified. We have not been able to identify a similar non-/3-cell population with IAPP expression on sections of normal adult pancreas. Secondly, although IAPP expression was generally /8-cell specific in phenotypically stable islet tumor lines (representing homogeneous /3-cell tumors), IAPP and insulin expressions were not coupled in unstable phenotypes, which are characterized by a high level of heterogeneity. Such cultures could probably represent transformed islet stem cells, which are able to express several hormone genes simultaneously (19). Similar phenotypes are transiently present during islet ontogeny in the mouse (20), where they further mature to select and elaborate a particular hormone. Heterogeneous MSL cell clones are able to form either stable transplantable insulinomas (15) or glucagonomas (our unpublished data) when passaged in vivo. Such tumors are highly homogeneous, and Northern blot analyses revealed an insulinoma-specific expression of IAPP. Whether the transition from heterogeneity to homogeneity within a clone reflects a process of normal islet cell maturation is only speculative. Another likely explanation could be that in vitro culture conditions in general favor cellular heterogeneity, which may allow uncoupled expression of IAPP and insulin in transformed as well as normal islet cultures. A series of MSL clones was shown by immunocytochemistry to have a low frequency of lAPP-producing cells in vitro. However, a high frequency of strongly stained cells was only found in cultures that were able to induce insulinomas by in vivo passage (15,16). Thus, a high frequency of lAPP-expressing cells was found in the noninsulin-producing culture of NHI-6F. These cells form stable insulinomas in vivo, where the transfected human insulin gene is coactivated with the endogeneous insulin genes (15). During repassage in vitro, these cells may lose insulin production (16), but retain active IAPP genes. We conclude that the tissue specificity of expression of IAPP and insulin is controlled differently by the fact that lAP-P-expression is not necessarily /3-cell specific in cultures of normal or transformed islet cells. This may be due in part to an apparently different nature of the tissue-specific IAPP gene enhancer, which was recently characterized (13). Additionally, coexpression of IAPP with hormones different from insulin is a clonal marker for pluripotent rat islet tumor cells, which are able to form stable insulinomas by passage in vivo. MATERIALS AND METHODS Cells Cultures of the human insulin gene-transfected subclone (MSL-G2), called NHI-6F, were passaged in vivo to form
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 22:23 For personal use only. No other uses without permission. . All rights reserved.
IAPP and Insulin Expression in Islet Cells
145
Fig. 1. Immunocytochemical Characterization of Various Murine Islet Cell Cultures GH-promoted (14) monolayers of newborn rat islet cells (A-F) were stained for insulin (A; red), IAPP (B; green), insulin and IAPP (C; double exposure). D is a phase contrast photo micrograph of the same field (A-C); note an insulin-negative, but lAPP-positive, cell (arrowhead, A-D). E, IAPP (green); F, somatostatin [note one cell (arrowhead) coexpressing the two hormones]. The glucagonproducing culture a-Tc (18) is heterogeneous (G and H) with insulin (red)-producing colonies (H). Only some of these colonies coexpress IAPP, as in G (green). J, Glucagon producing NHI-6F cells have a high frequency of lAPP-expressing cells (indirect immunoperoxidase staining), while insulin-producing NHI-6F secondary cultures (derived from NHI-6F insulinoma; see Results) show uncoupling with IAPP (red is insulin, green is IAPP, and yellow is coexpression) in the double exposure shown in K.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 22:23 For personal use only. No other uses without permission. . All rights reserved.
Vol 5 No. 1
MOL ENDO-1991 146
1 2
12
11 21
3 4 INSULIN
INSULIN
IAPP
IAPP
GLUCAGON GLUCAGON Fig. 2. Northern Blot Analyses of Stable Tumor Phenotypes Lanes 1 and 2 represent total RNA from the insulinoma MSL-G2-IN (15) and the glucagonoma MSL-G-AN (unpublished), respectively. Lanes 1' and 2' represent total RNA from /3-Tc (17) and a-Tc (18), respectively. Note that IAPP mRNA expression follows the j3-cell phenotype. That a-Tc is heterogeneous at the cellular level with a small fraction (
Ole D. Madsen, Jens H. Nielsen, Birgitte Michelsen, Per Westermark, Christer Betsholtz, Masahiro Nishi, and Donald F. Steiner Hagedorn Research Laboratory DK-2820 Gentofte, Denmark Department of Pathology University Hospital (P.W.) Linkoping, Sweden Department of Pathology University Hospital (C.B.) Uppsala, Sweden Department of Biochemistry and Molecular Biology Howard Hughes Medical Institute University of Chicago (M.N., D.F.S.) Chicago, Illinois 60637
clones, which are able to activate insulin gene transcription during passage in vivo. (Molecular Endocrinology 5: 143-148, 1991)
The pancreatic /S-cell is a major site of islet amyloid polypeptide (IAPP) biosynthesis, and the peptide is coreleased with insulin. We have analyzed the expression of IAPP (mRNA and protein) in various cell types in normal and transformed murine islet cell cultures by Northern blot analyses and immunocytochemistry. IAPP is primarily coexpressed with insulin in the /?-cell of GH-promoted primary rat islet cell cultures. Additionally, a small population of non/0-cells exhibited a prominent IAPP expression, and double staining experiments showed colocalization with glucagon or somatostatin in some of these cells. IAPP mRNA was confined to the /?-cell phenotype when analyzing the phenotypically stable in vivo tumor lines, MSL-G2-IN (insulinoma) and MSLG-AN (glucagonoma), and the transgenic mouse islet cell lines, /?-Tc and a-Tc. However, IAPP and insulin expression were completely uncoupled in unstable heterogeneous clones such as NHI-6F. This clone is composed of primarily glucagon-producing cells in vitro, but insulin gene expression becomes dominant after passage in vivo. Interestingly, IAPP was hyperexpressed with glucagon under in vitro conditions in this clone. We conclude that the tissue specificity of expressions of IAPP and insulin are controlled differently, and that coexpression of IAPP with hormones different from insulin may be a marker for pluripotent transformed rat islet cell
INTRODUCTION Islet amyloid polypeptide (IAPP or amylin) is the major component in islet amyloid deposits and has been implicated in the development of type II diabetes (1-4). In the adult pancreas the presence of IAPP seems to be restricted to the pancreatic /3-cell, as determined by immunocytochemistry (5). Although small amounts of extrapancreatic IAPP mRNA (6) and protein (7) have been reported, the islet #-cell seems to be the major site of IAPP synthesis (5, 7, 8), and IAPP is cosecreted with insulin (9, 10). The strict /?-cell-specific expression of insulin is controlled by 5' flanking insulin gene enhancer and promoter elements interacting with putative /?-cell-specific transacting factors, as recently confirmed in a series of transgenic experiments (11). Since there is no obvious homology between the 5' flanking sequences of the human IAPP gene and insulin genes (12,13), expression of the IAPP gene may be regulated by separate factors. The aim of the present study was to compare expression of IAPP and insulin in normal neonatal rat islet cells stimulated by GH (14), in a pluripotent rat islet tumor cell line transfected with the human insulin gene (15, 16), and in SV-40 T-antigen-transformed mouse islet cell lines (17,18).
0888-8809/91/0143-0148$02.00/0 Molecular Endocrinology Copyright© 1991 by The Endocrine Society
143
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 22:23 For personal use only. No other uses without permission. . All rights reserved.
Vol 5 No. 1
MOL ENDO-1991 144
RESULTS Nontransformed Islet Cell Cultures Monolayers of newborn rat islet cells stimulated to proliferate by GH showed highly intense staining for insulin (Fig. 1A). A few scattered cells were strongly stained for glucagon and somatostatin (14). IAPP staining showed a heterogeneous pattern of one or two strongly positive cells scattered among a majority of weakly stained cells (Fig. 1B) in most colonies. Interestingly, the brightly stained IAPP cells were consistently negative for insulin, as visualized by double exposure (Fig. 1C). These cells could be characterized by double immunostaining as either a-cells (not shown) or 5-cells (Fig. 1, E and F). Stable Islet Tumor Phenotypes When analyzing IAPP expression in stable rat insulinomas (15) or glucagonomas of common monoclonal origin (our unpublished data) as well as the /3 (17)- and a (18)-cell lines from transgenic mice, we found a good overall correlation of the presence of IAPP mRNA with the insulin-producing phenotype (Fig. 2). However, on the cellular level the a-Tc culture was heterogeneous, as revealed by discrete colonies producing insulin (Fig. 1H). Only some of these colonies were shown to coexpress IAPP (Fig. 1G). Unstable Islet Tumor Phenotypes The glucagon-producing MSL-G2 cells and human insulin gene-transfected subclone NHI-6F can be induced to produce insulin in vivo. Repassage in vitro caused a reversal in phenotype, with a gradual loss of insulin gene expression (15, 16). Cultures representing these stages were analyzed for insulin, glucagon, and IAPP mRNA distribution (Fig. 3). Surprisingly, the overall IAPP mRNA profile followed that of glucagon, but not that of insulin. These data were confirmed by immunocytochemistry, where the noninsulin-producing primary cells of NHI-6F were highly positive for IAPP immunoreactivity (Fig. 1J). No obvious coupling of IAPP and insulin expression was seen in NHI-6F cultures during the gradual loss of the insulin-expressing phenotype, as visualized in Fig. 1K.
DISCUSSION
The primary site of IAPP expression in the adult seems to be the pancreatic /3-cell (5, 8-10). The /3-cell specificity is possibly less restricted than that of insulin, since IAPP mRNA has been reported outside of the pancreas (6,7). Our goal was to study IAPP expression in various cell types in normal and transformed islet cells by Northern blot analyses and immunocytochemistry. IAPP expression relative to that of insulin in these abnormal
cells may tell us whether two respective genes are influenced by similar /3-cell-specific control mechanisms. In general, we can conclude that this may not be the case. Firstly, nontransformed newborn islet cell cultures are clearly heterogeneous with respect to IAPP expression. While IAPP is weakly identified in most /3-cells, a very small subpopulation of a- as well as 5-cells were highly immunoreactive with IAPP antisera. Whether the phenomenon observed is due to the effect of GH, which is known to induce /3-cell proliferation (14), remains to be clarified. We have not been able to identify a similar non-/3-cell population with IAPP expression on sections of normal adult pancreas. Secondly, although IAPP expression was generally /8-cell specific in phenotypically stable islet tumor lines (representing homogeneous /3-cell tumors), IAPP and insulin expressions were not coupled in unstable phenotypes, which are characterized by a high level of heterogeneity. Such cultures could probably represent transformed islet stem cells, which are able to express several hormone genes simultaneously (19). Similar phenotypes are transiently present during islet ontogeny in the mouse (20), where they further mature to select and elaborate a particular hormone. Heterogeneous MSL cell clones are able to form either stable transplantable insulinomas (15) or glucagonomas (our unpublished data) when passaged in vivo. Such tumors are highly homogeneous, and Northern blot analyses revealed an insulinoma-specific expression of IAPP. Whether the transition from heterogeneity to homogeneity within a clone reflects a process of normal islet cell maturation is only speculative. Another likely explanation could be that in vitro culture conditions in general favor cellular heterogeneity, which may allow uncoupled expression of IAPP and insulin in transformed as well as normal islet cultures. A series of MSL clones was shown by immunocytochemistry to have a low frequency of lAPP-producing cells in vitro. However, a high frequency of strongly stained cells was only found in cultures that were able to induce insulinomas by in vivo passage (15,16). Thus, a high frequency of lAPP-expressing cells was found in the noninsulin-producing culture of NHI-6F. These cells form stable insulinomas in vivo, where the transfected human insulin gene is coactivated with the endogeneous insulin genes (15). During repassage in vitro, these cells may lose insulin production (16), but retain active IAPP genes. We conclude that the tissue specificity of expression of IAPP and insulin is controlled differently by the fact that lAP-P-expression is not necessarily /3-cell specific in cultures of normal or transformed islet cells. This may be due in part to an apparently different nature of the tissue-specific IAPP gene enhancer, which was recently characterized (13). Additionally, coexpression of IAPP with hormones different from insulin is a clonal marker for pluripotent rat islet tumor cells, which are able to form stable insulinomas by passage in vivo. MATERIALS AND METHODS Cells Cultures of the human insulin gene-transfected subclone (MSL-G2), called NHI-6F, were passaged in vivo to form
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 22:23 For personal use only. No other uses without permission. . All rights reserved.
IAPP and Insulin Expression in Islet Cells
145
Fig. 1. Immunocytochemical Characterization of Various Murine Islet Cell Cultures GH-promoted (14) monolayers of newborn rat islet cells (A-F) were stained for insulin (A; red), IAPP (B; green), insulin and IAPP (C; double exposure). D is a phase contrast photo micrograph of the same field (A-C); note an insulin-negative, but lAPP-positive, cell (arrowhead, A-D). E, IAPP (green); F, somatostatin [note one cell (arrowhead) coexpressing the two hormones]. The glucagonproducing culture a-Tc (18) is heterogeneous (G and H) with insulin (red)-producing colonies (H). Only some of these colonies coexpress IAPP, as in G (green). J, Glucagon producing NHI-6F cells have a high frequency of lAPP-expressing cells (indirect immunoperoxidase staining), while insulin-producing NHI-6F secondary cultures (derived from NHI-6F insulinoma; see Results) show uncoupling with IAPP (red is insulin, green is IAPP, and yellow is coexpression) in the double exposure shown in K.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 17 November 2015. at 22:23 For personal use only. No other uses without permission. . All rights reserved.
Vol 5 No. 1
MOL ENDO-1991 146
1 2
12
11 21
3 4 INSULIN
INSULIN
IAPP
IAPP
GLUCAGON GLUCAGON Fig. 2. Northern Blot Analyses of Stable Tumor Phenotypes Lanes 1 and 2 represent total RNA from the insulinoma MSL-G2-IN (15) and the glucagonoma MSL-G-AN (unpublished), respectively. Lanes 1' and 2' represent total RNA from /3-Tc (17) and a-Tc (18), respectively. Note that IAPP mRNA expression follows the j3-cell phenotype. That a-Tc is heterogeneous at the cellular level with a small fraction (
