0021-972X/91/7305-1044$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright © 1991 by The Endocrine Society
Vol. 73, No. 5 Printed in U.S.A.
Role of Tumor Necrosis Factor-a and Interferon-7 as Growth Factors to the Human Fetal /?-Cell* BERNARD E. TUCHf, ANN M. SIMPSON, AND IAIN L. CAMPBELL Department of Medicine, University of Sydney, New South Wales, and the Burnet Clinical Research Unit, Walter and Eliza Hall Institute of Medical Research (I.L.C.), Parkville, Victoria, Australia
ABSTRACT. The effects of the cytokines tumor necrosis factor-a and interferon-7 ° n the adult /8-cell have been well described: a reduction of insulin secretion and content and death of the cell. For this reason and because these cytokines may be released from activated lymphocytes and macrophages that infiltrate islets in insulin-dependent diabetes, they have been implicated in the pathophysiology of this form of diabetes. As to whether the human fetal /3-cell, which differs from the adult /3-cell in not releasing insulin in response to the nutrient glucose and not being adversely affected by the toxin streptozotocin, is similarly affected is unknown. To examine this question we cultured monolayers of a single cell suspension of human fetal pancreas in the presence or absence of 1000 U/mL of these cytokines for 7 days. Chronic insulin release was enhanced for the first 2 days of culture, but unchanged thereafter. Acute insulin release in response to the secretagogue theophylline (10
I
mM) was enhanced on day 7, but not earlier. There was an increase in the insulin content of the cells by the fourth day, probably due to an increase in the number of /3-cells present (45 ± 5% vs. 22 ± 3%). Microscopically, non-/3-cells also seemed to increase in number; there was an increase in both DNA and cell number by the seventh day. In contrast to these beneficial effects on the human fetal ^S-cell, treatment of adult rat insulinoma cells, represented by RIN-m5F cells, resulted in inhibition of insulin secretion during the first day of culture and subsequent death of 86% of the cells by the sixth day of culture. It is hypothesized that the functional immaturity and lack of normal (adult) metabolic activity of the human fetal /3-cell somehow confers protection on these cells from the cytotoxic effects of tumor necrosis factor-a and interferon-y. Indeed, our findings suggest that these cytokines may be trophic for the developing /3-cell. (J Clin Endocrinol Metab 73:1044-1050,1991)
T HAS previously been reported that the cytokines tumor necrosis factor-a and interferon-7 have deleterious effects on adult murine /3-cells in regard to their ability to secrete insulin, their insulin content, and their morphological appearance (1). Interferon-7 has been shown to inhibit insulin synthesis in /3-cells from the rat insulinoma cell line RIN-m5F (2). Both of these cytokines independently are capable of inducing expression of class I major histocompatibility proteins of adult /3cells in humans (3) and mice (4, 5). When combined, they also induce expression of class II antigens (5-7). Because of these pleiotropic effects, both tumor necrosis factor-a and interferon-7, which are secreted by activated lymphocytes (8, 9), have been implicated in the destruction of adult /3-cells (insulitis) that results in insulin-dependent diabetes (10,11).
is relatively unknown. There is one report of the induction of class I antigens on the human fetal /8-cell (3), but no reports of their effect on the functioning of these cells. There are sufficient differences between the human fetal /3-cell and its adult counterpart for it not to be assumed that any effect of tumor necrosis factor-a and interferon-7 on the adult is the same in the fetus. For example, glucose exerts a major effect on insulin release from the adult /3-cell, but only a minimal (12), and in some cases an inhibitory (13), effect on the fetal /3-cell. Further, the adult /3-cell toxin streptozotocin has no deleterious effect on the human fetal /3-cell (14). It was for these reasons that the effects of tumor necrosis factor-a and interferon-7 on the ability of human fetal /3-cells to store and secrete insulin were examined.
While the effects of these agents on the adult /3-cell have been well described, their effect on the fetal /3-cell
Materials and Methods Source of human fetal tissue
Received February 25,1991. Address all correspondence and requests for reprints to: Bernard E. Tuch, M.D., Ph.D., Department of Endocrinology, Prince of Wales Hospital, Randwick, New South Wales 2031, Australia. * This work was supported by research grants from the Juvenile Diabetes Foundation International (JDFI), the Juvenile Diabetes Foundation Australia, and the Hoechst Australia Diabetes Foundation. t Recipient of a Career Development Award from the JDFI.
Four human fetal pancreata were obtained from the therapeutic termination of pregnancies carried out between 17-20 weeks gestation. Permission for the organ culture of this tissue was successfully sought from the mother, while approval for carrying out the experiments was obtained from the Ethics Committee of the University of Sydney.
1044
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 12:05 For personal use only. No other uses without permission. . All rights reserved.
CYTOKINES AND HUMAN FETAL PANCREAS Tissue culture 3
The pancreata were diced into 1-mm explants, then digested into a single cell suspension with 20 mg collagenase (Boehringer Mannheim, Mannheim, Germany) in 5 mL Dulbecco's phosphate-buffered saline containing sodium bicarbonate (26 mM) and HEPES (20 mM) in a 37 C incubator; mixing was performed with a sterile magnetic stirrer. Termination of the digestion after 10-15 min was carried out by the addition of 15 mL PBS at room temperature. The cells were washed twice with PBS and centrifuged for 5 min at 1000 rpm between each wash. Cells were then resuspended in 12 mL culture medium before being evenly dispersed in 12 flat bottom wells of a multiwell plate (Linbro, Flow Laboratories, McLean, VA) and incubated at 37 C in a humidified atmosphere of 5% CO2 in air. The culture medium used was RPMI-1640 buffered with 20 mM HEPES and enriched with amino acids, 7% fetal calf serum, and antibiotics (15). After 3 days when the cells had attached, half of the wells were exposed to the combination of tumor necrosis factor-a (1000 U/mL) and interferon-7 (1000 U/mL) made up in medium; the other half acted as controls. Highly purified recombinant human tumor necrosis factor-a and interferon-7 were kindly provided by Dr. H. Michael Shepard, Genentech, Inc. (South San Francisco, CA). The specific activities of the cytokines used were: tumor necrosis factor-a, 2.87 X 107 U/mg, and interferon-7, 2.2 X 107 U/mg. Medium was generally changed on a daily basis; the conditioned medium was stored at —20 C until assayed for insulin. Experiments were conducted for 1-7 days after addition of the cytokines; the gross effects of these cytokines on the fetal pancreatic cells were observed daily with an inverted phase contrast microscope. Acute insulin release On the third day in one experiment and the seventh in another, the ability of the /3-cells to release insulin in response to 20 mM glucose or 10 mM theophylline was examined. No response to glucose would normally have been expected (12), while theophylline should have acted as an excellent stimulus (12). The design of the experiment was to remove the culture medium and expose the cells to two 30-min basal periods of phosphate-buffered saline containing 2.8 mM glucose (basal I and basal II) before adding 20 mM glucose for 30 min. This was followed by a third basal period (basal III), exposure to 10 mM theophylline, and then a fourth basal period (basal IV) to ensure that insulin release returned to the baseline. All periods lasted 30 min. Conditioned medium was collected after each change and assayed for insulin using a porcine insulin standard and guinea pig antiinsulin antibody (kindly donated by Dr. D. Yue, Sydney, Australia). The lower limit of detection of this assay was 1.6 jiU/mL. Stimulation indices were derived by dividing the level of insulin released during the stimulation phases with that secreted in the immediately preceding basal period. These indices allowed each well of cells to act as its own control. Cell number and insulin and DNA contents At the conclusion of the cultures (days 1, 4, and 7), the cells were trypsinized off each well, and an aliquot was counted in a
1045
hemocytometer chamber. Of the remaining cells half were extracted overnight at 4 C, in acid-ethanol, and the insulin content of the supernatant was determined. The rest of the cells were pelleted and solubilized in 1% sodium dodecyl sulfate, and their DNA content was measured fluometrically (16). Histology
After 3 days of culture, the cells from one experiment were trypsinized off the bottom of the culture wells, cytospun, and stained for insulin using a guinea pig antiinsulin antibody (Dako, Santa Barbara, CA) and the chromogen alkaline phosphatase. A mean ± SEM of 1023 ± 55 cells were counted from each well, with the observer blinded to the identity of the well. RIN-m5F These cells were subcultured onto 24-well plates (Linbro, McLean, VA) and exposed to tumour necrosis factor-a and interferon-7, at a concentration of 1000 U/mL for six days. The medium used, RPMI-1640 containing 10% fetal calf serum, was changed daily for 3 days, and all conditioned media were assayed for insulin content by RIA against porcine standards, as for cultured human fetal /3-cells. At the termination of culture, the cells were trypsinized off and counted, and the DNA content was determined. The insulin content of the cells, both controls and treated, was also measured after extraction with acid-ethanol, but was negligible, a finding noted by others previously (Hulinsky, I., personal communication). Statistics The effects of the cytokines on chronic insulin release were examined by repeated measures analysis of variance, using the 4V Program on BMDP (17). The difference between the groups in regard to acute insulin release, stimulation indices, cell number, and insulin and DNA contents was examined using Student's t test (18).
Results Chronic insulin release The release of insulin from the monolayers of human fetal pancreatic cells remained relatively constant for the 7 days of culture (Fig. 1). Exposure of these cells to the combination of tumor necrosis factor-a and interferon7, each at a concentration of 1000 U/mL, caused enhancement of insulin release during the first 2 days of exposure to these cytokines (P = 0.041), but not thereafter (Fig. 1). Acute insulin release As expected, fetal j8-cells in the monolayer were responsive to the secretagogue theophylline, with a 3.9fold enhancement of insulin release on both the third and seventh days of culture (Fig. 2 and Table 1). The response of these cells to the physiological stimulus of
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 12:05 For personal use only. No other uses without permission. . All rights reserved.
TUCH, SIMPSON, AND CAMPBELL
1046
JCE&M«1991 Vol73«NoB
2.0 • • I
(IB)
£
1.6-1
D
E
Vol. 73, No. 5 Printed in U.S.A.
Role of Tumor Necrosis Factor-a and Interferon-7 as Growth Factors to the Human Fetal /?-Cell* BERNARD E. TUCHf, ANN M. SIMPSON, AND IAIN L. CAMPBELL Department of Medicine, University of Sydney, New South Wales, and the Burnet Clinical Research Unit, Walter and Eliza Hall Institute of Medical Research (I.L.C.), Parkville, Victoria, Australia
ABSTRACT. The effects of the cytokines tumor necrosis factor-a and interferon-7 ° n the adult /8-cell have been well described: a reduction of insulin secretion and content and death of the cell. For this reason and because these cytokines may be released from activated lymphocytes and macrophages that infiltrate islets in insulin-dependent diabetes, they have been implicated in the pathophysiology of this form of diabetes. As to whether the human fetal /3-cell, which differs from the adult /3-cell in not releasing insulin in response to the nutrient glucose and not being adversely affected by the toxin streptozotocin, is similarly affected is unknown. To examine this question we cultured monolayers of a single cell suspension of human fetal pancreas in the presence or absence of 1000 U/mL of these cytokines for 7 days. Chronic insulin release was enhanced for the first 2 days of culture, but unchanged thereafter. Acute insulin release in response to the secretagogue theophylline (10
I
mM) was enhanced on day 7, but not earlier. There was an increase in the insulin content of the cells by the fourth day, probably due to an increase in the number of /3-cells present (45 ± 5% vs. 22 ± 3%). Microscopically, non-/3-cells also seemed to increase in number; there was an increase in both DNA and cell number by the seventh day. In contrast to these beneficial effects on the human fetal ^S-cell, treatment of adult rat insulinoma cells, represented by RIN-m5F cells, resulted in inhibition of insulin secretion during the first day of culture and subsequent death of 86% of the cells by the sixth day of culture. It is hypothesized that the functional immaturity and lack of normal (adult) metabolic activity of the human fetal /3-cell somehow confers protection on these cells from the cytotoxic effects of tumor necrosis factor-a and interferon-y. Indeed, our findings suggest that these cytokines may be trophic for the developing /3-cell. (J Clin Endocrinol Metab 73:1044-1050,1991)
T HAS previously been reported that the cytokines tumor necrosis factor-a and interferon-7 have deleterious effects on adult murine /3-cells in regard to their ability to secrete insulin, their insulin content, and their morphological appearance (1). Interferon-7 has been shown to inhibit insulin synthesis in /3-cells from the rat insulinoma cell line RIN-m5F (2). Both of these cytokines independently are capable of inducing expression of class I major histocompatibility proteins of adult /3cells in humans (3) and mice (4, 5). When combined, they also induce expression of class II antigens (5-7). Because of these pleiotropic effects, both tumor necrosis factor-a and interferon-7, which are secreted by activated lymphocytes (8, 9), have been implicated in the destruction of adult /3-cells (insulitis) that results in insulin-dependent diabetes (10,11).
is relatively unknown. There is one report of the induction of class I antigens on the human fetal /8-cell (3), but no reports of their effect on the functioning of these cells. There are sufficient differences between the human fetal /3-cell and its adult counterpart for it not to be assumed that any effect of tumor necrosis factor-a and interferon-7 on the adult is the same in the fetus. For example, glucose exerts a major effect on insulin release from the adult /3-cell, but only a minimal (12), and in some cases an inhibitory (13), effect on the fetal /3-cell. Further, the adult /3-cell toxin streptozotocin has no deleterious effect on the human fetal /3-cell (14). It was for these reasons that the effects of tumor necrosis factor-a and interferon-7 on the ability of human fetal /3-cells to store and secrete insulin were examined.
While the effects of these agents on the adult /3-cell have been well described, their effect on the fetal /3-cell
Materials and Methods Source of human fetal tissue
Received February 25,1991. Address all correspondence and requests for reprints to: Bernard E. Tuch, M.D., Ph.D., Department of Endocrinology, Prince of Wales Hospital, Randwick, New South Wales 2031, Australia. * This work was supported by research grants from the Juvenile Diabetes Foundation International (JDFI), the Juvenile Diabetes Foundation Australia, and the Hoechst Australia Diabetes Foundation. t Recipient of a Career Development Award from the JDFI.
Four human fetal pancreata were obtained from the therapeutic termination of pregnancies carried out between 17-20 weeks gestation. Permission for the organ culture of this tissue was successfully sought from the mother, while approval for carrying out the experiments was obtained from the Ethics Committee of the University of Sydney.
1044
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 12:05 For personal use only. No other uses without permission. . All rights reserved.
CYTOKINES AND HUMAN FETAL PANCREAS Tissue culture 3
The pancreata were diced into 1-mm explants, then digested into a single cell suspension with 20 mg collagenase (Boehringer Mannheim, Mannheim, Germany) in 5 mL Dulbecco's phosphate-buffered saline containing sodium bicarbonate (26 mM) and HEPES (20 mM) in a 37 C incubator; mixing was performed with a sterile magnetic stirrer. Termination of the digestion after 10-15 min was carried out by the addition of 15 mL PBS at room temperature. The cells were washed twice with PBS and centrifuged for 5 min at 1000 rpm between each wash. Cells were then resuspended in 12 mL culture medium before being evenly dispersed in 12 flat bottom wells of a multiwell plate (Linbro, Flow Laboratories, McLean, VA) and incubated at 37 C in a humidified atmosphere of 5% CO2 in air. The culture medium used was RPMI-1640 buffered with 20 mM HEPES and enriched with amino acids, 7% fetal calf serum, and antibiotics (15). After 3 days when the cells had attached, half of the wells were exposed to the combination of tumor necrosis factor-a (1000 U/mL) and interferon-7 (1000 U/mL) made up in medium; the other half acted as controls. Highly purified recombinant human tumor necrosis factor-a and interferon-7 were kindly provided by Dr. H. Michael Shepard, Genentech, Inc. (South San Francisco, CA). The specific activities of the cytokines used were: tumor necrosis factor-a, 2.87 X 107 U/mg, and interferon-7, 2.2 X 107 U/mg. Medium was generally changed on a daily basis; the conditioned medium was stored at —20 C until assayed for insulin. Experiments were conducted for 1-7 days after addition of the cytokines; the gross effects of these cytokines on the fetal pancreatic cells were observed daily with an inverted phase contrast microscope. Acute insulin release On the third day in one experiment and the seventh in another, the ability of the /3-cells to release insulin in response to 20 mM glucose or 10 mM theophylline was examined. No response to glucose would normally have been expected (12), while theophylline should have acted as an excellent stimulus (12). The design of the experiment was to remove the culture medium and expose the cells to two 30-min basal periods of phosphate-buffered saline containing 2.8 mM glucose (basal I and basal II) before adding 20 mM glucose for 30 min. This was followed by a third basal period (basal III), exposure to 10 mM theophylline, and then a fourth basal period (basal IV) to ensure that insulin release returned to the baseline. All periods lasted 30 min. Conditioned medium was collected after each change and assayed for insulin using a porcine insulin standard and guinea pig antiinsulin antibody (kindly donated by Dr. D. Yue, Sydney, Australia). The lower limit of detection of this assay was 1.6 jiU/mL. Stimulation indices were derived by dividing the level of insulin released during the stimulation phases with that secreted in the immediately preceding basal period. These indices allowed each well of cells to act as its own control. Cell number and insulin and DNA contents At the conclusion of the cultures (days 1, 4, and 7), the cells were trypsinized off each well, and an aliquot was counted in a
1045
hemocytometer chamber. Of the remaining cells half were extracted overnight at 4 C, in acid-ethanol, and the insulin content of the supernatant was determined. The rest of the cells were pelleted and solubilized in 1% sodium dodecyl sulfate, and their DNA content was measured fluometrically (16). Histology
After 3 days of culture, the cells from one experiment were trypsinized off the bottom of the culture wells, cytospun, and stained for insulin using a guinea pig antiinsulin antibody (Dako, Santa Barbara, CA) and the chromogen alkaline phosphatase. A mean ± SEM of 1023 ± 55 cells were counted from each well, with the observer blinded to the identity of the well. RIN-m5F These cells were subcultured onto 24-well plates (Linbro, McLean, VA) and exposed to tumour necrosis factor-a and interferon-7, at a concentration of 1000 U/mL for six days. The medium used, RPMI-1640 containing 10% fetal calf serum, was changed daily for 3 days, and all conditioned media were assayed for insulin content by RIA against porcine standards, as for cultured human fetal /3-cells. At the termination of culture, the cells were trypsinized off and counted, and the DNA content was determined. The insulin content of the cells, both controls and treated, was also measured after extraction with acid-ethanol, but was negligible, a finding noted by others previously (Hulinsky, I., personal communication). Statistics The effects of the cytokines on chronic insulin release were examined by repeated measures analysis of variance, using the 4V Program on BMDP (17). The difference between the groups in regard to acute insulin release, stimulation indices, cell number, and insulin and DNA contents was examined using Student's t test (18).
Results Chronic insulin release The release of insulin from the monolayers of human fetal pancreatic cells remained relatively constant for the 7 days of culture (Fig. 1). Exposure of these cells to the combination of tumor necrosis factor-a and interferon7, each at a concentration of 1000 U/mL, caused enhancement of insulin release during the first 2 days of exposure to these cytokines (P = 0.041), but not thereafter (Fig. 1). Acute insulin release As expected, fetal j8-cells in the monolayer were responsive to the secretagogue theophylline, with a 3.9fold enhancement of insulin release on both the third and seventh days of culture (Fig. 2 and Table 1). The response of these cells to the physiological stimulus of
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 November 2015. at 12:05 For personal use only. No other uses without permission. . All rights reserved.
TUCH, SIMPSON, AND CAMPBELL
1046
JCE&M«1991 Vol73«NoB
2.0 • • I
(IB)
£
1.6-1
D
E
