- Email: [email protected]
Polychlorinated biphenyls release insulin from RINm5F cells
Life Sciencu, Vd. 59, No. 24, pp. 2041~2049,19!X chpyri&t~1996-sciclreInc. Printed in the USA. Au rights revned om4-3m/96 slma + .I0
ELSEVIER
PII SOO24-3205(96)00557-I
POLYCHLORINATED BIPHENYIS RELEASE INSULIN FROM RINmSF CELLS
Lawrence J. Fischer, Hui-Ren Zhou and Margaret A. Wagner Institute for Environmental Toxicology and Department of Pharmacology and Toxicology, Michigan State University, East Lansing, Ml 48824 (Received in tinal form October 3,19%)
Summary Polychlorinated biphenyls (PCBs) possess a variety of biological effects, including alterations in growth, development and metabolism, that may be dependent on insulin. However, no reports on the action of PCBs on cells which produce and secrete insulin are available. The current study examined the ability of a commercial mixture of PCBs (Aroclor 1254) and three specific PCB congeners, to alter the release of insulin using the hormone producing cell line RINm5F. Exposure of cells to Aroclor 1254 (A1254) produced a concentration-dependent increase in media insulin reaching a peak, when expressed as percent of control, at 30 min. In spite of continued exposure, media insulin relative to control declined and no treatment-related difference was observed at 48 hrs. Cellular levels of the hormone declined as much as 50% by that time. The insulin releasing action of A-1254 was mimicked by each of the non-coplanar congeners 2,2’,4,4’-tetrachlorobiphenyl (KS) and 2,2’,4,4’,5,5’-hexachlorobiphenyl (HCB) but the coplanar congener 3,3’,4,4’-TCB showed no significant activity. These results indicate that PCBs are capable of producing a release of insulin from RINmSF cells, an effect that is unlikely to be associated with coplanar congeners that initiate their action by binding to the Ah-receptor. Kq, WO&: polychlorinated biphenyls, insulin, RINmSF Polychlorinated biphenyls (PCBs) are ubiquitous environmental contaminants, exhibiting resistance to degradation and a propensity to accumulate in the food supply. They were used commercially as mixtures of individual congeners, each component of the mixture having chlorine atoms substituted in different positions in the biphenyl nucleus. Eiiher as individual congeners or mixtures, PCBs have a variety of adverse biological actions in laboratory animals (1). This has led to estimates of health risks from environmental exposures based upon the assumption that the toxic effects of PCBs are initiated by the
Corresponding Author: Lawrence J. Fischer, Institute for Environmental Toxicology, C231 Holden Hall, Michigan State University, East Lansing, Ml 48824, USA. Tel: 517 3536469, Fax: 517 355-4603, E-mail: [email protected]
2042
PCBs and Insulin Release
Vol. 59, No. 24, 1996
binding of structurally coplanar PCB congeners to the Ah-receptor (1). Recently,biological actions of non-coplanar congeners of PCBs have been demonstrated (2,3) indicating that the toxicity of PCBs does not reside entirely within the Ah-receptor based mechanism as these congeners lack affinity for the receptor (4). The toxicity of PCBs involves a variety of organs including endocrine systems important in reproduction and development (5). Actions of PCBs on signal transduction systems (6,7,8) provide the possibility that critical functions of highly differentiated cells, such as those secreting hormones, may be directly impacted by these chemicals. There have been no reports concerning PCB-induced alterations in the function of insulin producing cells, even though a characteristic of the acute toxicity of these agents includes disruption of mitochondrial function and intermediary metabolism (1,9). In the present study, we have examined the function of insulin-producing RINm5F cells exposed to Aroclor 1254 (A-1254) and to a coplanar and two non-coplanar PCB congeners. Aroclor was used to study a commercial mixture of the pollutants and pure congeners were examined to determine the potential involvement of the Ah-receptor in the actions of PCBs on insulin producing cells. These cells, derived from a rat insulinoma, have been extensively used to investigate mechanisms of insulin synthesis and release (eg. 10,11,12). In addition, RINm5F cells have been successfully employed as a model system to investigate the actions of diabetogenic chemicals and drugs (13). The results obtained from this well characterized system and presented in this report are the first to document direct alterations by PCBs on the function of insulin producing cells.
Insulin Release Experiments The procedures used in this study were similar to those utilized to investigate the effects of the drug cyproheptadine on insulin secretion and synthesis in RINm5F cells (13). The sources of cells and chemicals used in experiments, unless given here are identical to those described previously (13). Briefly, cells were allowed to grow in 35 mm plastic dishes in RPM11640 media containing 10% fetal bovine serum for 4 days prior to initiating experiments. These were started by replacement of media with the fresh media and allowing the cells to accommodate under culture conditions for 30 min. The media was then removed and replaced with media containing treatments, either DMSO (control) or A-1254 or individual PCB congeners (ChemSetvice, West Chester, PA) dissolved in DMSO. The highest concentration of DMSO used in control and treatment media was 2 PI/ml (0.1%). Over a 48 hour exposure period, this concentration of vehicle produced no alteration in cellular or media insulin in preliminary experiments. Also, no effects of DMSO were observed on basal insulin release over a 30 min. period (data not shown). For studies of insulin release, cells undergoing PCB treatment and control cells were kept under culture conditions for various lengths of time up to 48 hr. At specified times, media (25 ~1)was removed, centrifuged briefly to pellet any detached cells and the supernatant frozen until analyzed for immunoreactive insulin. In most experiments, cells were collected for DNA and insulin analysis. After removal of media, cells were harvested with gentle trypsinization followed by two washes with calcium, magnesium free PBS containing 0.45 mM EDTA. The cells were resuspended in 1.0 ml of Krebs Ringer bicarbonate buffer containing 10 mM HEPES and separate 0.45 ml aliquots taken for analysis of DNA and insulin as previously described (13).
Vol. 59, No. 24,1996
PCBs and Inmdin Release
2043
Experiments in which the effect of 10 pg/ml A-1254 on 20 mM KCI-induced insulin secretion was investigated were also conducted as previously described (13). In these studies, culture media was removed from cells and the cells washed with a Krebs Ringer bicarbonate buffer containing 5 mg/ml BSA and 10 mM HEPES (Sigma, St. Louis, MO) and preincubated under culture conditions in the same buffer for 30 min. Thereafter, the preincubation buffer was replaced by 1 ml of the same buffer containing treatments (A1254 and/or KCI) and cells incubated for a 45 min period. The buffer was then removed from the cells and analyzed for insulin concentration by radioimmunoassay. Cells were removed by gentle trypsinization, washed and analyzed for DNA. Insulin Synthesis Investigation of the effect of A-1254 on insulin and total protein biosynthesis was performed using incorporation of 3H-leucine into immunoprecipitated and TCA-precipitated protein using previously reported procedures (13). Media containing 0,1,5 or 10 pg/ml of A-1254 was placed on the cells and culture continued for 24 hr. The media was discarded, cells washed with KRB buffer, removed from the plates by gentle trypsinization and the cells collected by centrifugation. Insulin-like proteins were extracted using acetic acid and precipitated using guinea pig anti-insulin serum and Protein ASepharose (Pharmacia LKB, Piscataway, NJ) separation. Total protein was precipitated with 5% trichloroacetic acid (TCA). Radioactivity associated in immuno- and TCAprecipitated proteins was measured by scintillation counting. Results were calculated as dpm/petri dish and converted to % of control (no A-1254 treatment). Data Analysis Results of measurements of insulin in cells and media in each petri dish were expressed as ng immunoreactive insulin/pg DNA. No significant change, due to PCB treatment, in the amount of DNA/dish was detected and this indicated that the treatments did not alter cell growth. In the time course of insulin release experiments, values are reported as a percent of insulin released by control cells in each experiment to normalize data for possible daily variations in control insulin release. RINm5F cells are known to exhibit a reduced capacity to secrete insulin when the passage number becomes large but this characteristic was largely eliminated by using cells that represented passages 40-64. The results are presented as mean ‘- S.E. for the number of plates per treatment and represent at least three experiments performed on different occasions. Statistical differences were determined by ANOVA and a multiple comparisons test or Students t test (Ps 0.05).
Culture of RINmSF cells with A-1254 for 48 hours resulted in a concentration dependent decline in cellular content of insulin as shown in Table I. This effect was detected using 1 vg/ml of A-1254 and a 50% depletion of the hormone occurred using 10 pg/ml of the PCB mixture. There was no significant difference from control in media insulin measured at the end of the treatment period. The concentrations of A-1254 used produced no decline in DNA relative to control cells and in a separate experiment, no decrease in trypan blue exclusion was observed (data not shown). These results indicated that the decline in cellular insulin observed in A-1254 treated cells was not due to cell death.
PCBa amdInsulin R&me
2044
Vol. 59, No. 24, 1996
TABLE I Cellular and Media insulin After a 48 hr. Exposure of RINmSF Cells to A-1254. INSULIN (ng/pg DNA) IN: CELLS MEDIA
A-l 254 (u&ml/ 0
43.6 f 2.3
78.5 + 5.0
1
*
36.8 2 2.2
93.9 f 6.6
5
*
29.122.4
87.7 +- 2.3
10
*
21.8 rt 2.1
92.2 f 9.9
’ Values are mean 2 S.E., N=4. Asterisk (*) denotes significant difference from control (no A-1254 treatment). The more immediate effects of A-1254 (10 pg/ml) were investigated by examining media insulin during a 2 hr. exposure followed by a recovery period during which the previously exposed cells were cultured in fresh media. The results in Fig. 1 show that media insulin increased during the treatment period and that the rate and extent of increase was dependent on the concentration of A-1254. When expressed as a percent of control, 70
-
60
-
0
Control
0
5ug/ml
V
lOug/ml
a Aroclor
1254
Aroclor
,
,
,
,
I 2
-i
1254
,
,
,
.
,
I
,
/
,
,
,
1
/P
30
‘illii.,
~~
0
0 I
I
1
I
30
60
90
Treatment
Media insulin concentrations are mean + corresponding
I 120 I
I 0 I
Minutes
_. a1 &-30
__ -la 60
1 90
I 120 I
Recovery
Fig. 1 during a 2 hr. treatment of RINmSF cells with various A-1254 and during recovery of cells in fresh media. Values S.E., N=6. “a” indicates a significant difference from control value.
Vol. 59, No. 24,19%
PCBs and Insulin Releasa
2045
media insulin from cells treated with 10 pg/ml of the PCB mixture was highest at 30 min after initiation of treatment reaching a value nearly 3.5 times that of control. The relative increase in released hormone declined thereafter and was approximately 2.5 times the control value after two hrs of treatment. A similar time course of hormone accumulation in media, although lesser in amount was observed with the 5 pg/ml concentration of A1254. During the recovery phase (no A-1254 exposure), media insulin from cells previously treated with 5 pg/ml A-1254 increased in a manner identical to control cells (Fig. 1). However, lower amounts of insulin were released during recovery by cells previously treated with 10 pg/ml of the PCB mixture. The results obtained from analysis of media insulin as shown in Fig. 1 are consistent with results obtained from an analysis of cellular insulin after two hours of treatment with A1254 and after a two hour recovery period. The data shown in Table II indicate that 10 pg/ml A-l 254 significantly lowered cellular insulin whereas treatment with 5 pgjml did not significantly reduce insulin levels in the cells. After two hours of recovery, insulin in cells previously treated with either concentration of A-1254 was not different from control.
TABLE II Cellular Insulin After Two Hours of A-1254 Treatment and After a Two Hour Recovery Period
A-1254 (Da/ml)
CELLULAR INSULIN (ng/pg DNA) AFTER: TREATMENT RECOVERY
0
49.5 + 1.9
43.9 + 1.0
5
45.1 f 2.9
42.3 2 3.1
10
* 31.4 + 1.1
39.5 f 2.4
Values are mean + S.E., N =4, (*) significant difference from control (no A-l 254 treatment). The ability of A-1254 to acutely after depolarization-induced insulin release was investigated. The results indicated that the presence of 10 pg/ml of A-1254 did not increase the amount of insulin released by a depolarizing concentration (20 mM) of KCI, 17.2 kO.9 versus 19.6 21.2 ng insulin/kg DNA/45 min (mean + S.E., N=3). Basal insulin release was 5.3 20.4 in those experiments. Table Ill presents results of experiments investigating effects of the PCB mixture on insulin and total protein synthesis. The presence of A-1254 significantly reduced, in a concentration dependent manner, the incorporation of 3H-leucine into immunoreactive insulin. The extent of this effect was identical to that observed when incorporation of the labeled amino acid into total protein was measured in the same experiment.
F’CBs and Insulir~Reksse
2046
Vol. 59, No. 24, 1%
TABLE III Inhibition by A-1254 of Insulin and Total Protein Synthesis A-1254 0
INSULIN (% CONTROL)
TOTAL PROTEIN (% CONTROL)
1
98.3 + 2.1
99.7 + 2.0
5
* 84.1 + 1.8
* 87.1 + 2.9
* 70.5 +7.0
* 73.1 + 2.4
10
’ Measured as 3H-leucine incorporation into insulin-immunoprecipitable and TCA-precipitable protein. Values are mean f S.E., N=3. (*) indicate significant difference from no A-1254 treatment control. The insulin-releasing activities of a non-coplanar (2,2’,4,4’-TCB) and coplanar PC6 congener (3,3’,4,4’-TCB) were compared utilizing exposure of cells to 10 pg/ml(3.5 PM) of each congener over a four hour period. The results shown in Fig. 2 indicate that the coplanar congener exhibited no activity relative to control whereas the non-coplanar congener produced a rapid release of insulin similar to that observed with A-1254. This
V .
a
I
100
0:
lOug/ml lOvg/ml
2.2’.4.4’-TCB 3,3’,4.4’-7CB
a
t
0
60
120
160
240
Minutes
Fg. 2 Time course of insulin release from RINm5F cells during treatment with 10 kg/ml of a non-coplanar PCB congener (2,2’,4,4’-TCB) or the same concentration of a coplanar congener (3,3’,4,4’-TCB). Media insulin concentrations are shown as a percent of no treatment controls. Values represent mean f S.E., N =8. “a” indicates significant difference from control value at the same time point.
Vol. 59, No. 24, 1996
2047
PCBS and Insulin Release
result led to an assessment of the concentration-response characteristics of another noncoplanar congener 2,2’,4,4’,5,5’-HCB. Data presented in Fig. 3 indicate that this congener also elicits a concentration dependent release of insulin. No activity was observed using a concentration of 1 pg/ml (0.28 PM) and increasing insulin releasing activity which reached a maximum at 30-60 min with 5 pg/ml (1.4 (LM) and 10 pg/ml (2.8 PM). 600
500
r:I a
I ’
I
I
c = .- 0 iL 2 5 - 0 50 0, t
s
400
-
‘.
r
/i 4 f
200
-
a
OL 0
-.
A
lOug/ml
2.2’4,4’5,5’-HCB
‘. ‘.
i *-
=c
a
.q__
a
)--r--C
2,2’,4,4’5,5’-HCB 2,2’,4,4’5,5’-HCB
*..i
a>
300
1 ug/ml 5ug,‘ml
‘.
,
-
n A
a
---__ - - _ _----__ -_ _---‘;_
4 a
------i
n
, 60
a _ --;-__ .a
120
180
, 240
Miff utes
Fig. 3 Concentration and time dependent release of insulin from RINm5F cells produced by 2,2’,4,4’,5,5’-HCB. Media insulin is expressed as percent of the vehicle control value at each time point. Values are mean f SE, N =56. “a” indicates significant difference from control.
Discussion The PCB mixture utilized in these experiments exhibited an ability to release insulin from RINm5F cells equivalent to the release observed from depolarization of the cells with KCI. Depolarization-induced insulin release which opens voltage-dependent calcium channels is a major component of glucose-stimulated insulin release in normal pancreatic beta cells (14). Glucose does not stimulate insulin release from RINm5F ceils due to a deficiency in glucose metabolism in the insulinoma-derived cells (15). However, glucose metabolites such as glyceraldehyde as well as amino acids and depolarizing concentrations of KCI can release insulin from the cells indicating a functional insulin release mechanism (11,13). The process by which A-1254 induces insulin release could be related to PCB effects on the signal transduction systems known to be involved in insulin release (11,I 4) or due to a direct action on insulin granule exocytosis, the final stage of insulin secretion. The inability of A-1254 to augment KCI-induced insulin release, as shown in this study would be consistent with an effect of PCBs on the secretion process triggered by depolarization-induced entry of extracellular calcium into RINmSF cells (11). While calcium
2048
PCBs
andInwlin Release
Vol. 59, No. 24, 19%
influx plays an important role in insulin secretion, there is evidence for the contribution of calcium-independent processes in stimulated release of the hormone and PCBs may also function via these processes (10). PCBs have been reported to increase intracellular calcium and activate calcium-dependent PKC in cerebellar granule cells (616). These effects, if occurring in RINm5F cells, could be expected to result in insulin release. The role of calcium and other signal transduction mechanisms in PCB-induced insulin secretion will be evaluated in future experiments. The major portion of the PCB effect on insulin release appears to be terminated within four hours of continuous exposure. The transient nature of the effect may be due, in part, to the observed decline in cellular insulin produced during A-1254 exposure. A decline in cellular insulin undoubtedly results from stimulated insulin release and possibly an inability of the exposed cells to adequately replace released hormone with newly synthesized insulin. Results presented here indicate that A-1254 can inhibit insulin and total protein synthesis and this effect may contribute to the decline in PCB-related release of insulin during continuous exposure. Another possible reason for a transient effect of PCBs is that tolerance or exhaustion develops within a signal transduction system involved in the chemical-induced insulin release. The instability of insulin at the pH of media would also contribute to its decline over a sufficiently prolonged period. Any or all of these factors would allow media insulin to return to control levels during continuous PCB exposure. The results obtained using three different PCB congeners indicate that the insulin releasing action of A-1254 can probably be attributed to the non-coplanar congeners present in the mixture. The two non-coplanar congeners used in this study were found to release insulin whereas the coplanar congener 3,3’,4,4’-TCB exhibited no activity. Because the estimated affinity for binding to the Ah-receptor of congeners active in releasing insulin is three or four orders of magnitude lower than the inactive coplanar congener (4), it appears unlikely that insulin release by PCBs involves an interaction with that receptor. It is becoming increasingly apparent that non-coplanar congeners, those having chlorine atoms at ortho positions on each of the biphenyl aromatic rings have the capability of producing important biological effects. This includes, in neuronal cells, alterations in dopamine levels (17), phorbol ester binding (3) and PKC translocation (6). Other reports indicate alterations by these congeners in calcium transport by ryanodine sensitive channels in a clonal cell system (18) and alterations in neutrophil function (19). It is possible, but seems unlikely, that the effect of PCBs is due to a general membrane leakage of insulin caused by a direct action on the membrane which altered its physical properties. The two isomers 2,2’4,4’-TCB and 3,3’,4,4’-TCB were active and inactive, respectively, but could be expected to have similar direct effects, if any, on the physical characteristics of the membranes limiting insulin release. It seems more likely that an enzyme or other receptor is involved in the insulin releasing action of non- coplanar PCBs. The toxicological significance of the results presented in this report remain to be established. Because PCBs are present in the environment and produce adverse biological effects in laboratory animals and in wildlife, their risk to humans must be understood. The non-coplanar congener 2,2’,4,4’,5,5’-HCB (PCB congener #153) shown in this study to release insulin from RINm5F cells has been found in relatively high levels in human tissue (1,x1,21). Risk evaluations based only on toxicity resulting from coplanar congeners acting through the Ah-receptor do not take into account possible effects by this and other active, non-coplanar constituents of environmental PCB exposures (1). Current results indicate that additional attention to the potentially toxic effects of non-
Vol. 59,No. 24,19%
coplanar
PCBS in the environment
PCBs and Insulin Release
2049
is warranted.
Acknowfedaements Support for this research was provided by USPHS-NIH Grant ES0491 1.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.
S. SAFE, Crit. Rev. Toxicol. 24 l-63 (1994). W. SHAIN, B. BUSH, and R. SEEGAL, Toxicol. Appl. Pharmacol. 11133-42 (1991). P.R.S. KODAVANTI, T.R. WARD, J.D. McKENNEY AND H.A. TILSON, Toxicol. Appl. Pharmacol. X$Q 140-148 (1995). S.A. KAFAFI, H.Y. AFEEFY, A.H. ALI, H.K. SAID and A.G. KAFAFI, Environ. Health Perspectives 1Q1422-428 (1993). J.D. McKlNNEY and C.L. WALLER, Environ. Health Perspectives 120 290297 (1994). P.R.S. KODAVANTI, T.J. SHAFER,T.R.WARD, T. FREUDENRICH, G.J.HARRY and H.A. TILSON, Brain Res. 6@ 75-82 (1994). P.K. TITHOF, M.L. CONTRERAS and P.E. GANEY, Toxicol. Appl. Pharmacol. m 136-143 (1995). P.K. TITHOF, M.PETERS-GOLDEN, L. SCHIAMBURG and P.E. GANEY, Environ. Health Perspectives m 52-58 (1996). Y. NISHIHARA, and K. UTSUMI, Biochem. Pharmacol. % 3335-3339 (1986). G. LI, D. MIIANI, M.J. DUNNE, W. PRALONG, J. THELERT, O.H.PETERSEN and C.B. WOLLHEIM, J. Biol. Chem. s 3449-3457 (1991). T.YADA, L.L. RUSSO and G.W. SHARP, J. Biol. Chem. % 2455-2462 (1989). A. SJCLOM, N. WELSH, V. HOFTIEZER, P.W.BANKSTON and C. HELLERSTRCM, Biochem. J. 277 533-540 (1991). C.P. MILLER and L.J.FISCHER, Biochem. Pharmacol. s 1983-1990 (1999). A.E. BOYD, J. Cellular Biochem. a 234-241 (1992). P.A. HALBAN, G.A. PRAZ and C.B. WOLLHEIM, Biochem. J. 212 439-443 (1983). P.R.S. KODAVANTI, D. SHIN, H.A. TILSON and G.J. HARRY, Toxicol. Appl. Pharmacol. 123 97-106 (1993). W. SHAIN, B.BUSH and R. SEEGAL, Toxicol. Appl. Pharmacol. 11133-42 (1991). P.W. WONG and I.N. PESSAH, The Toxicologist 3Q 226-227 (1996). A.P. Brown and P.E. Ganey, Toxicol. Appl. Pharmacol. 1;11 198-205 (1995). J. FALANDYSZ, N.YAMASHITA, S. TANABE, R. TATSUKAWA, Science of the Total Environment -113-l 19 (1994). D.G. PATERSON, G.D. TODD, W.E. TURNER, V. MAGGIO, L.R. ALEXANDER and L.L. NEEDHAM, Environmental Health Perspectives 102 Suppl. 1:195-204 (1994).
ELSEVIER
PII SOO24-3205(96)00557-I
POLYCHLORINATED BIPHENYIS RELEASE INSULIN FROM RINmSF CELLS
Lawrence J. Fischer, Hui-Ren Zhou and Margaret A. Wagner Institute for Environmental Toxicology and Department of Pharmacology and Toxicology, Michigan State University, East Lansing, Ml 48824 (Received in tinal form October 3,19%)
Summary Polychlorinated biphenyls (PCBs) possess a variety of biological effects, including alterations in growth, development and metabolism, that may be dependent on insulin. However, no reports on the action of PCBs on cells which produce and secrete insulin are available. The current study examined the ability of a commercial mixture of PCBs (Aroclor 1254) and three specific PCB congeners, to alter the release of insulin using the hormone producing cell line RINm5F. Exposure of cells to Aroclor 1254 (A1254) produced a concentration-dependent increase in media insulin reaching a peak, when expressed as percent of control, at 30 min. In spite of continued exposure, media insulin relative to control declined and no treatment-related difference was observed at 48 hrs. Cellular levels of the hormone declined as much as 50% by that time. The insulin releasing action of A-1254 was mimicked by each of the non-coplanar congeners 2,2’,4,4’-tetrachlorobiphenyl (KS) and 2,2’,4,4’,5,5’-hexachlorobiphenyl (HCB) but the coplanar congener 3,3’,4,4’-TCB showed no significant activity. These results indicate that PCBs are capable of producing a release of insulin from RINmSF cells, an effect that is unlikely to be associated with coplanar congeners that initiate their action by binding to the Ah-receptor. Kq, WO&: polychlorinated biphenyls, insulin, RINmSF Polychlorinated biphenyls (PCBs) are ubiquitous environmental contaminants, exhibiting resistance to degradation and a propensity to accumulate in the food supply. They were used commercially as mixtures of individual congeners, each component of the mixture having chlorine atoms substituted in different positions in the biphenyl nucleus. Eiiher as individual congeners or mixtures, PCBs have a variety of adverse biological actions in laboratory animals (1). This has led to estimates of health risks from environmental exposures based upon the assumption that the toxic effects of PCBs are initiated by the
Corresponding Author: Lawrence J. Fischer, Institute for Environmental Toxicology, C231 Holden Hall, Michigan State University, East Lansing, Ml 48824, USA. Tel: 517 3536469, Fax: 517 355-4603, E-mail: [email protected]
2042
PCBs and Insulin Release
Vol. 59, No. 24, 1996
binding of structurally coplanar PCB congeners to the Ah-receptor (1). Recently,biological actions of non-coplanar congeners of PCBs have been demonstrated (2,3) indicating that the toxicity of PCBs does not reside entirely within the Ah-receptor based mechanism as these congeners lack affinity for the receptor (4). The toxicity of PCBs involves a variety of organs including endocrine systems important in reproduction and development (5). Actions of PCBs on signal transduction systems (6,7,8) provide the possibility that critical functions of highly differentiated cells, such as those secreting hormones, may be directly impacted by these chemicals. There have been no reports concerning PCB-induced alterations in the function of insulin producing cells, even though a characteristic of the acute toxicity of these agents includes disruption of mitochondrial function and intermediary metabolism (1,9). In the present study, we have examined the function of insulin-producing RINm5F cells exposed to Aroclor 1254 (A-1254) and to a coplanar and two non-coplanar PCB congeners. Aroclor was used to study a commercial mixture of the pollutants and pure congeners were examined to determine the potential involvement of the Ah-receptor in the actions of PCBs on insulin producing cells. These cells, derived from a rat insulinoma, have been extensively used to investigate mechanisms of insulin synthesis and release (eg. 10,11,12). In addition, RINm5F cells have been successfully employed as a model system to investigate the actions of diabetogenic chemicals and drugs (13). The results obtained from this well characterized system and presented in this report are the first to document direct alterations by PCBs on the function of insulin producing cells.
Insulin Release Experiments The procedures used in this study were similar to those utilized to investigate the effects of the drug cyproheptadine on insulin secretion and synthesis in RINm5F cells (13). The sources of cells and chemicals used in experiments, unless given here are identical to those described previously (13). Briefly, cells were allowed to grow in 35 mm plastic dishes in RPM11640 media containing 10% fetal bovine serum for 4 days prior to initiating experiments. These were started by replacement of media with the fresh media and allowing the cells to accommodate under culture conditions for 30 min. The media was then removed and replaced with media containing treatments, either DMSO (control) or A-1254 or individual PCB congeners (ChemSetvice, West Chester, PA) dissolved in DMSO. The highest concentration of DMSO used in control and treatment media was 2 PI/ml (0.1%). Over a 48 hour exposure period, this concentration of vehicle produced no alteration in cellular or media insulin in preliminary experiments. Also, no effects of DMSO were observed on basal insulin release over a 30 min. period (data not shown). For studies of insulin release, cells undergoing PCB treatment and control cells were kept under culture conditions for various lengths of time up to 48 hr. At specified times, media (25 ~1)was removed, centrifuged briefly to pellet any detached cells and the supernatant frozen until analyzed for immunoreactive insulin. In most experiments, cells were collected for DNA and insulin analysis. After removal of media, cells were harvested with gentle trypsinization followed by two washes with calcium, magnesium free PBS containing 0.45 mM EDTA. The cells were resuspended in 1.0 ml of Krebs Ringer bicarbonate buffer containing 10 mM HEPES and separate 0.45 ml aliquots taken for analysis of DNA and insulin as previously described (13).
Vol. 59, No. 24,1996
PCBs and Inmdin Release
2043
Experiments in which the effect of 10 pg/ml A-1254 on 20 mM KCI-induced insulin secretion was investigated were also conducted as previously described (13). In these studies, culture media was removed from cells and the cells washed with a Krebs Ringer bicarbonate buffer containing 5 mg/ml BSA and 10 mM HEPES (Sigma, St. Louis, MO) and preincubated under culture conditions in the same buffer for 30 min. Thereafter, the preincubation buffer was replaced by 1 ml of the same buffer containing treatments (A1254 and/or KCI) and cells incubated for a 45 min period. The buffer was then removed from the cells and analyzed for insulin concentration by radioimmunoassay. Cells were removed by gentle trypsinization, washed and analyzed for DNA. Insulin Synthesis Investigation of the effect of A-1254 on insulin and total protein biosynthesis was performed using incorporation of 3H-leucine into immunoprecipitated and TCA-precipitated protein using previously reported procedures (13). Media containing 0,1,5 or 10 pg/ml of A-1254 was placed on the cells and culture continued for 24 hr. The media was discarded, cells washed with KRB buffer, removed from the plates by gentle trypsinization and the cells collected by centrifugation. Insulin-like proteins were extracted using acetic acid and precipitated using guinea pig anti-insulin serum and Protein ASepharose (Pharmacia LKB, Piscataway, NJ) separation. Total protein was precipitated with 5% trichloroacetic acid (TCA). Radioactivity associated in immuno- and TCAprecipitated proteins was measured by scintillation counting. Results were calculated as dpm/petri dish and converted to % of control (no A-1254 treatment). Data Analysis Results of measurements of insulin in cells and media in each petri dish were expressed as ng immunoreactive insulin/pg DNA. No significant change, due to PCB treatment, in the amount of DNA/dish was detected and this indicated that the treatments did not alter cell growth. In the time course of insulin release experiments, values are reported as a percent of insulin released by control cells in each experiment to normalize data for possible daily variations in control insulin release. RINm5F cells are known to exhibit a reduced capacity to secrete insulin when the passage number becomes large but this characteristic was largely eliminated by using cells that represented passages 40-64. The results are presented as mean ‘- S.E. for the number of plates per treatment and represent at least three experiments performed on different occasions. Statistical differences were determined by ANOVA and a multiple comparisons test or Students t test (Ps 0.05).
Culture of RINmSF cells with A-1254 for 48 hours resulted in a concentration dependent decline in cellular content of insulin as shown in Table I. This effect was detected using 1 vg/ml of A-1254 and a 50% depletion of the hormone occurred using 10 pg/ml of the PCB mixture. There was no significant difference from control in media insulin measured at the end of the treatment period. The concentrations of A-1254 used produced no decline in DNA relative to control cells and in a separate experiment, no decrease in trypan blue exclusion was observed (data not shown). These results indicated that the decline in cellular insulin observed in A-1254 treated cells was not due to cell death.
PCBa amdInsulin R&me
2044
Vol. 59, No. 24, 1996
TABLE I Cellular and Media insulin After a 48 hr. Exposure of RINmSF Cells to A-1254. INSULIN (ng/pg DNA) IN: CELLS MEDIA
A-l 254 (u&ml/ 0
43.6 f 2.3
78.5 + 5.0
1
*
36.8 2 2.2
93.9 f 6.6
5
*
29.122.4
87.7 +- 2.3
10
*
21.8 rt 2.1
92.2 f 9.9
’ Values are mean 2 S.E., N=4. Asterisk (*) denotes significant difference from control (no A-1254 treatment). The more immediate effects of A-1254 (10 pg/ml) were investigated by examining media insulin during a 2 hr. exposure followed by a recovery period during which the previously exposed cells were cultured in fresh media. The results in Fig. 1 show that media insulin increased during the treatment period and that the rate and extent of increase was dependent on the concentration of A-1254. When expressed as a percent of control, 70
-
60
-
0
Control
0
5ug/ml
V
lOug/ml
a Aroclor
1254
Aroclor
,
,
,
,
I 2
-i
1254
,
,
,
.
,
I
,
/
,
,
,
1
/P
30
‘illii.,
~~
0
0 I
I
1
I
30
60
90
Treatment
Media insulin concentrations are mean + corresponding
I 120 I
I 0 I
Minutes
_. a1 &-30
__ -la 60
1 90
I 120 I
Recovery
Fig. 1 during a 2 hr. treatment of RINmSF cells with various A-1254 and during recovery of cells in fresh media. Values S.E., N=6. “a” indicates a significant difference from control value.
Vol. 59, No. 24,19%
PCBs and Insulin Releasa
2045
media insulin from cells treated with 10 pg/ml of the PCB mixture was highest at 30 min after initiation of treatment reaching a value nearly 3.5 times that of control. The relative increase in released hormone declined thereafter and was approximately 2.5 times the control value after two hrs of treatment. A similar time course of hormone accumulation in media, although lesser in amount was observed with the 5 pg/ml concentration of A1254. During the recovery phase (no A-1254 exposure), media insulin from cells previously treated with 5 pg/ml A-1254 increased in a manner identical to control cells (Fig. 1). However, lower amounts of insulin were released during recovery by cells previously treated with 10 pg/ml of the PCB mixture. The results obtained from analysis of media insulin as shown in Fig. 1 are consistent with results obtained from an analysis of cellular insulin after two hours of treatment with A1254 and after a two hour recovery period. The data shown in Table II indicate that 10 pg/ml A-l 254 significantly lowered cellular insulin whereas treatment with 5 pgjml did not significantly reduce insulin levels in the cells. After two hours of recovery, insulin in cells previously treated with either concentration of A-1254 was not different from control.
TABLE II Cellular Insulin After Two Hours of A-1254 Treatment and After a Two Hour Recovery Period
A-1254 (Da/ml)
CELLULAR INSULIN (ng/pg DNA) AFTER: TREATMENT RECOVERY
0
49.5 + 1.9
43.9 + 1.0
5
45.1 f 2.9
42.3 2 3.1
10
* 31.4 + 1.1
39.5 f 2.4
Values are mean + S.E., N =4, (*) significant difference from control (no A-l 254 treatment). The ability of A-1254 to acutely after depolarization-induced insulin release was investigated. The results indicated that the presence of 10 pg/ml of A-1254 did not increase the amount of insulin released by a depolarizing concentration (20 mM) of KCI, 17.2 kO.9 versus 19.6 21.2 ng insulin/kg DNA/45 min (mean + S.E., N=3). Basal insulin release was 5.3 20.4 in those experiments. Table Ill presents results of experiments investigating effects of the PCB mixture on insulin and total protein synthesis. The presence of A-1254 significantly reduced, in a concentration dependent manner, the incorporation of 3H-leucine into immunoreactive insulin. The extent of this effect was identical to that observed when incorporation of the labeled amino acid into total protein was measured in the same experiment.
F’CBs and Insulir~Reksse
2046
Vol. 59, No. 24, 1%
TABLE III Inhibition by A-1254 of Insulin and Total Protein Synthesis A-1254 0
INSULIN (% CONTROL)
TOTAL PROTEIN (% CONTROL)
1
98.3 + 2.1
99.7 + 2.0
5
* 84.1 + 1.8
* 87.1 + 2.9
* 70.5 +7.0
* 73.1 + 2.4
10
’ Measured as 3H-leucine incorporation into insulin-immunoprecipitable and TCA-precipitable protein. Values are mean f S.E., N=3. (*) indicate significant difference from no A-1254 treatment control. The insulin-releasing activities of a non-coplanar (2,2’,4,4’-TCB) and coplanar PC6 congener (3,3’,4,4’-TCB) were compared utilizing exposure of cells to 10 pg/ml(3.5 PM) of each congener over a four hour period. The results shown in Fig. 2 indicate that the coplanar congener exhibited no activity relative to control whereas the non-coplanar congener produced a rapid release of insulin similar to that observed with A-1254. This
V .
a
I
100
0:
lOug/ml lOvg/ml
2.2’.4.4’-TCB 3,3’,4.4’-7CB
a
t
0
60
120
160
240
Minutes
Fg. 2 Time course of insulin release from RINm5F cells during treatment with 10 kg/ml of a non-coplanar PCB congener (2,2’,4,4’-TCB) or the same concentration of a coplanar congener (3,3’,4,4’-TCB). Media insulin concentrations are shown as a percent of no treatment controls. Values represent mean f S.E., N =8. “a” indicates significant difference from control value at the same time point.
Vol. 59, No. 24, 1996
2047
PCBS and Insulin Release
result led to an assessment of the concentration-response characteristics of another noncoplanar congener 2,2’,4,4’,5,5’-HCB. Data presented in Fig. 3 indicate that this congener also elicits a concentration dependent release of insulin. No activity was observed using a concentration of 1 pg/ml (0.28 PM) and increasing insulin releasing activity which reached a maximum at 30-60 min with 5 pg/ml (1.4 (LM) and 10 pg/ml (2.8 PM). 600
500
r:I a
I ’
I
I
c = .- 0 iL 2 5 - 0 50 0, t
s
400
-
‘.
r
/i 4 f
200
-
a
OL 0
-.
A
lOug/ml
2.2’4,4’5,5’-HCB
‘. ‘.
i *-
=c
a
.q__
a
)--r--C
2,2’,4,4’5,5’-HCB 2,2’,4,4’5,5’-HCB
*..i
a>
300
1 ug/ml 5ug,‘ml
‘.
,
-
n A
a
---__ - - _ _----__ -_ _---‘;_
4 a
------i
n
, 60
a _ --;-__ .a
120
180
, 240
Miff utes
Fig. 3 Concentration and time dependent release of insulin from RINm5F cells produced by 2,2’,4,4’,5,5’-HCB. Media insulin is expressed as percent of the vehicle control value at each time point. Values are mean f SE, N =56. “a” indicates significant difference from control.
Discussion The PCB mixture utilized in these experiments exhibited an ability to release insulin from RINm5F cells equivalent to the release observed from depolarization of the cells with KCI. Depolarization-induced insulin release which opens voltage-dependent calcium channels is a major component of glucose-stimulated insulin release in normal pancreatic beta cells (14). Glucose does not stimulate insulin release from RINm5F ceils due to a deficiency in glucose metabolism in the insulinoma-derived cells (15). However, glucose metabolites such as glyceraldehyde as well as amino acids and depolarizing concentrations of KCI can release insulin from the cells indicating a functional insulin release mechanism (11,13). The process by which A-1254 induces insulin release could be related to PCB effects on the signal transduction systems known to be involved in insulin release (11,I 4) or due to a direct action on insulin granule exocytosis, the final stage of insulin secretion. The inability of A-1254 to augment KCI-induced insulin release, as shown in this study would be consistent with an effect of PCBs on the secretion process triggered by depolarization-induced entry of extracellular calcium into RINmSF cells (11). While calcium
2048
PCBs
andInwlin Release
Vol. 59, No. 24, 19%
influx plays an important role in insulin secretion, there is evidence for the contribution of calcium-independent processes in stimulated release of the hormone and PCBs may also function via these processes (10). PCBs have been reported to increase intracellular calcium and activate calcium-dependent PKC in cerebellar granule cells (616). These effects, if occurring in RINm5F cells, could be expected to result in insulin release. The role of calcium and other signal transduction mechanisms in PCB-induced insulin secretion will be evaluated in future experiments. The major portion of the PCB effect on insulin release appears to be terminated within four hours of continuous exposure. The transient nature of the effect may be due, in part, to the observed decline in cellular insulin produced during A-1254 exposure. A decline in cellular insulin undoubtedly results from stimulated insulin release and possibly an inability of the exposed cells to adequately replace released hormone with newly synthesized insulin. Results presented here indicate that A-1254 can inhibit insulin and total protein synthesis and this effect may contribute to the decline in PCB-related release of insulin during continuous exposure. Another possible reason for a transient effect of PCBs is that tolerance or exhaustion develops within a signal transduction system involved in the chemical-induced insulin release. The instability of insulin at the pH of media would also contribute to its decline over a sufficiently prolonged period. Any or all of these factors would allow media insulin to return to control levels during continuous PCB exposure. The results obtained using three different PCB congeners indicate that the insulin releasing action of A-1254 can probably be attributed to the non-coplanar congeners present in the mixture. The two non-coplanar congeners used in this study were found to release insulin whereas the coplanar congener 3,3’,4,4’-TCB exhibited no activity. Because the estimated affinity for binding to the Ah-receptor of congeners active in releasing insulin is three or four orders of magnitude lower than the inactive coplanar congener (4), it appears unlikely that insulin release by PCBs involves an interaction with that receptor. It is becoming increasingly apparent that non-coplanar congeners, those having chlorine atoms at ortho positions on each of the biphenyl aromatic rings have the capability of producing important biological effects. This includes, in neuronal cells, alterations in dopamine levels (17), phorbol ester binding (3) and PKC translocation (6). Other reports indicate alterations by these congeners in calcium transport by ryanodine sensitive channels in a clonal cell system (18) and alterations in neutrophil function (19). It is possible, but seems unlikely, that the effect of PCBs is due to a general membrane leakage of insulin caused by a direct action on the membrane which altered its physical properties. The two isomers 2,2’4,4’-TCB and 3,3’,4,4’-TCB were active and inactive, respectively, but could be expected to have similar direct effects, if any, on the physical characteristics of the membranes limiting insulin release. It seems more likely that an enzyme or other receptor is involved in the insulin releasing action of non- coplanar PCBs. The toxicological significance of the results presented in this report remain to be established. Because PCBs are present in the environment and produce adverse biological effects in laboratory animals and in wildlife, their risk to humans must be understood. The non-coplanar congener 2,2’,4,4’,5,5’-HCB (PCB congener #153) shown in this study to release insulin from RINm5F cells has been found in relatively high levels in human tissue (1,x1,21). Risk evaluations based only on toxicity resulting from coplanar congeners acting through the Ah-receptor do not take into account possible effects by this and other active, non-coplanar constituents of environmental PCB exposures (1). Current results indicate that additional attention to the potentially toxic effects of non-
Vol. 59,No. 24,19%
coplanar
PCBS in the environment
PCBs and Insulin Release
2049
is warranted.
Acknowfedaements Support for this research was provided by USPHS-NIH Grant ES0491 1.
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.
S. SAFE, Crit. Rev. Toxicol. 24 l-63 (1994). W. SHAIN, B. BUSH, and R. SEEGAL, Toxicol. Appl. Pharmacol. 11133-42 (1991). P.R.S. KODAVANTI, T.R. WARD, J.D. McKENNEY AND H.A. TILSON, Toxicol. Appl. Pharmacol. X$Q 140-148 (1995). S.A. KAFAFI, H.Y. AFEEFY, A.H. ALI, H.K. SAID and A.G. KAFAFI, Environ. Health Perspectives 1Q1422-428 (1993). J.D. McKlNNEY and C.L. WALLER, Environ. Health Perspectives 120 290297 (1994). P.R.S. KODAVANTI, T.J. SHAFER,T.R.WARD, T. FREUDENRICH, G.J.HARRY and H.A. TILSON, Brain Res. 6@ 75-82 (1994). P.K. TITHOF, M.L. CONTRERAS and P.E. GANEY, Toxicol. Appl. Pharmacol. m 136-143 (1995). P.K. TITHOF, M.PETERS-GOLDEN, L. SCHIAMBURG and P.E. GANEY, Environ. Health Perspectives m 52-58 (1996). Y. NISHIHARA, and K. UTSUMI, Biochem. Pharmacol. % 3335-3339 (1986). G. LI, D. MIIANI, M.J. DUNNE, W. PRALONG, J. THELERT, O.H.PETERSEN and C.B. WOLLHEIM, J. Biol. Chem. s 3449-3457 (1991). T.YADA, L.L. RUSSO and G.W. SHARP, J. Biol. Chem. % 2455-2462 (1989). A. SJCLOM, N. WELSH, V. HOFTIEZER, P.W.BANKSTON and C. HELLERSTRCM, Biochem. J. 277 533-540 (1991). C.P. MILLER and L.J.FISCHER, Biochem. Pharmacol. s 1983-1990 (1999). A.E. BOYD, J. Cellular Biochem. a 234-241 (1992). P.A. HALBAN, G.A. PRAZ and C.B. WOLLHEIM, Biochem. J. 212 439-443 (1983). P.R.S. KODAVANTI, D. SHIN, H.A. TILSON and G.J. HARRY, Toxicol. Appl. Pharmacol. 123 97-106 (1993). W. SHAIN, B.BUSH and R. SEEGAL, Toxicol. Appl. Pharmacol. 11133-42 (1991). P.W. WONG and I.N. PESSAH, The Toxicologist 3Q 226-227 (1996). A.P. Brown and P.E. Ganey, Toxicol. Appl. Pharmacol. 1;11 198-205 (1995). J. FALANDYSZ, N.YAMASHITA, S. TANABE, R. TATSUKAWA, Science of the Total Environment -113-l 19 (1994). D.G. PATERSON, G.D. TODD, W.E. TURNER, V. MAGGIO, L.R. ALEXANDER and L.L. NEEDHAM, Environmental Health Perspectives 102 Suppl. 1:195-204 (1994).