Fish Physiology and Biochemistry vol. 11 no. 1-6 pp 411-420 (1993) Kugler Publications, Amsterdam/New York
Binding and bioactivity of insulin in primary cultures of carp (Cyprinus carpio) hepatocytes 2 Helmut Segner l , Ralf B6hm and Werner Kloas Zoology II, University of Karlsruhe, Kaiserstr. 12 D- W- 7500 Karlsruhe, Germany
Keywords: carp, liver cell culture, insulin, insulin binding, insulin bioactivity
Resum6 Des h6patocytes isoles de carpes, cultiv6s dans un milieu d6fini chimiquement et ne contenant pas de s6rum, ont t6 utilis6s pour tudier les caract6ristiques de liaison de l'insuline ainsi que ses effets m6taboliques. Des experiences de cin6tiques, de saturation et de d6placement ont et6 effectuees pour caract6riser la liaison de 1'12 5I-insuline. Un effet de la collagenase sur la liaison a t6 d6montr6e. La liaison de l'insuline d6crolt durant les douze premieres heures de culture, puis reste constante durant les douze heures suivantes. La liaison de 1' 12 5I-insuline atteint un tat d'6quilibre apres 20-30 min d'incubation. L'analyse math6matique des courbes de saturation rvele l'existence de deux sites de liaison, I'un de haute affinity (Kdl = 5.5 pM) avec une faible capacity (Bmax = 0.14 fmol/mg de prot6ine soit 77 sites de liaison par cellule) et l'autre de plus faible affinity (Kd 2 = 2.4 nM) et de plus grande capacity (Bmax 2 = 17.6 fmol/mg de prot6ine soit 9623 sites/cellule). Une 6tude de competition a montr6 que la moiti6 de liaison de 312 pM d' 12 5I-insuline est d6plac6e par 2.2 nM d'insuline froide, 7.9 nM d'IGF-I et 20.3 nM d'IGF-II. Le glucagon est sans effet. La liaison de l'insuline aux hepatocytes de carpes induit une rduction significative de la scretion de glucose et une augmentation significative de la synthese prot6ique et de la synthese de novo des acides gras. Abstract Isolated carp hepatocytes cultured in serum-free, chemically defined medium were used to investigate within the same cell preparation characteristics of the binding of insulin as well as effects of insulin on cellular metabolism. The binding of human [125I]-insulin to carp hepatocytes was studied in kinetic, saturation and displacement experiments. A dependency of insulin binding on the collagenase used for cell isolation was demonstrated. Insulin binding decreased during the first 12h of culture but remained constant during the following 12h. The kinetic experiments revealed that [125I]-insulin binding reached a steady state within 20-30 min of incubation. The mathematical analysis of the saturation experiments demonstrated the existence of two populations of binding sites, one with high affinity (Kdl = 5.5 pM) and low capacity (Bmaxl = 0.14 fmol/mg protein or 77 binding sites/cell) and one with low affinity (Kd 2 = 2.4 nM) and high capacity (Bmax 2 = 17.6 fmol/mg protein or 9623 binding sites/cell). In competition experiments, 312 pM [125 I]-insulin was displaced by cold insulin, IGF-I and IGF-II with IC 50 values of 2.2, 7.9 and 20.3 nM, respectively. Glucagon was without effect. Binding of insulin to carp hepatocytes resulted in a significant reduction of glucose release and a significant increase of protein synthesis as of de novo fatty acid synthesis. 1Present address: Environmental Research Center, Section for Environmental Chemistry and Ecotoxicology, Permoserstr. 15, D-0-7050 Leipzig, Germany; 2 dedicated to Prof. Dr. W. Hanke on the occasion of his 6 5th birthday.
412 Introduction The binding of insulin to its cell surface receptor is the first step to initiate physiological effects in target cells. The receptor plays a critical role in both directing the hormone to a specific target tissue and programming the biological responses of the tissue to the hormone. The liver receives its blood through the portal system which drains the endocrine pancreas. Therefore, the liver is the first organ to see any changes in insulin secretion (Plisetskaya and Sullivan 1989) and represents a primary target for insulin action. The insulin receptor of liver cells has been extensively studied in mammals (e.g., Czech 1985). For teleost fish, on the other hand, only scarce and partly contradictory information on insulin receptors is available (Plisetskaya 1990). Muggeo et al. (1979), when comparing insulin receptors from a variety of vertebrate species, described characteristics of insulin binding to erythrocytes of rainbow trout (Oncorhynchus mykiss). With respect to the liver of teleosts, Ablett et al. (1983) were the first to study the specific interaction of insulin with hepatocytes. Kinetic data on insulin binding sites in teleostean liver including Scatchard plot analysis, saturation and displacement experiments, have been published to date only for the two closely related carnivorous species, coho salmon, (Oncorhynchus kisutch, Gutierrez and Plisetskaya 1991) and rainbow trout (Gutierrez et al. 1991) as well as for the scorpionfish (Scorpaena porcus, Leibush and Bondareva 1991, cited by Gutierrez and Plisetskaya 1991). The small data base is surprising, particularly when considering the possible involvement of defects at the insulin binding level in the well-documented glucose intolerance of fish (Mommsen and Plisetskaya 1991; Gutierrez et al. 1991). The present communication reports data on insulin binding to liver cells of the omnivorous carp, Cyprinus carpio. Instead of isolated liver membranes, as employed in the binding studies with trout and coho salmon (Gutierrez et al. 1990, Gutierrez and Plisetskaya 1991), we use isolated, intact liver cells cultured in serum-free, chemically defined medium. This enables us to investigate within the same preparation the characteristics of in-
sulin binding as well as metabolic responses of the cells to insulin (see also Varandani et al. 1982).
Materials and methods Animals Two-year-old Cyprinus carpio of 200-400 g body weight were used for cell isolation. The fishes were maintained under a 12: 12 dark: light cycle at a constant water temperature of 20°C in aerated tanks equipped with a flow-through water system. Once per day (17.00h) the fishes were fed adlibitum with a commercial pelleted feed containing 40% crude protein and 7% crude fat. The carp were acclimated to those conditions for at least one year. The binding studies were performed from September to October 1991. Cell isolation After the fish was anaesthesized, the aorta coeliaca coming from the dorsal aorta to the liver was cannulated. The liver was cleared of blood by perfusion for 10 min with an aerated Ca 2 + - and Mg2 +-free Hank's salt solution containing 5 mmol/l EDTA and 15 mM HEPES. After all the blood had been cleared, a solution containing Mg2 +-free Hank's salts, 15 mM HEPES and 0.026% collagenase (Collagenase D, Boehringer Mannheim, 0,026%) was perfused for 20 min. This was followed by a final 5-min perfusion with a solution made of Ca 2 + - and Mg2 + - free Hank's salts, 2 mM EDTA and 15 mM HEPES. The liver was then dissected, minced into small pieces and passed through nylon sieves of 250, 100 and 50 zm mesh size. Cells were collected and repeatedly washed by centrifugation at 80 x g. The cell pellet was resuspended in culture medium and the cell number was counted in a hemocytometer. Cell viability according to trypan blue exclusion was usually higher than 950,'o. Final cell density was adjusted to 1.5-2 x 106 cells/ml. Cell culture The culture system was similar to that described for rainbow trout hepatocytes by Blair et al. (1990). The cells were seeded as monolayers in tissue cham-
413 ber slides (Nunc, Mannheim) for the binding studies (200 Al/chamber), or in 35 mm Petri dishes (Falcon Primaria, Heidelberg) for metabolic studies (1.5 ml/dish). The culture medium consisted of Hank's salts (138 mM NaCl, 5.4 mM KCI, 0.8 mM MgSO 4, 1.86 mM CaCI 2 , 0.34 mM Na 2 HPO 4, 0.44 mM KH 2PO 4 ), 0.05 mM NaHCO 3, 20 mM HEPES, 20 ml/l medium of MEM amino acid concentrate (50 x stock; Serva, Heidelberg), 10 ml/l medium of BME vitamin concentrate (100 x stock; Serva, Heidelberg), 3.3 ml/l medium of 0.5% sodium phenol-red (Serva, Heidelberg), 100 mg/l medium of streptomycin (Sigma, Deisenhofen), 100,000 U/I medium of penicillin (Sigma, Deissenhofen) and 2 mM L-glutamine (Serva, Heidelberg). To minimize insulin destruction, 25 mg/100 ml of bacitracin (Sigma, Deissenhofen) was added: this chemical does not affect insulin binding processes (Juul and Jones 1982). The cells were cultured in a humified air atmosphere at 20°C. The medium was changed every 12h. Influence of collagenase on cellular insulin binding In preliminary experiments, the influence of different collagenases on cellular insulin binding was investigated. Samples were taken immediately after isolation (0 h), and after 12 and 24h of culture. The tested collagenases included a preparation from PBH Biotechnologie (Hannover: 0.1-0.15 PZunits/mg lyophilisate. One PZ-unit is the amount of enzyme which hydrolyses 1 /mole of the peptide PZ-pro-leu-gly-pro-D-arg per min at 25°C and pH 7.1), type IV from Sigma (Deissenhofen; 1.1 FALGPA-units/mg lyophilisate. One FALGPAunit is the amount of enzyme which hydrolyses 1 /mole of the peptide furylacryloyl-leu-gly-pro-ala per min at 25 °C at pH 7.5 in the presence of calcium ions) and type D from Boehringer (Mannheim, 0.59 Wiinsch-units/mg lyophilisate. One Wiinsch-unit is the amount of enzyme which hydrolyses 1 Armole of the peptide 4-phenyl-azo-benzol-oxycarbonyl-Lprolyl-L-leucyl- L-glycyl-L-prolyl-D-arginine per min at 25C). For all further studies, only cells isolated with collagenase D and precultured for at least 12 h were used. The treatments were always done in triplicates.
Binding studies To determine insulin binding of freshly isolated cells (Oh), 200 1lof cell suspension were added to Eppendorf vials and cells were collected by centrifugation at 80 x g. The supernatant was carefully removed and cells were resuspended in 200 Al culture medium containing 70 pM [125 I]-insulin (human, iodinated at tyrosine A14, Amersham, Braunschweig; specific activity 1960 Ci/mmol) for the detection of total binding (TB). Nonspecific binding (NB) was assayed at the same concentration of [125 I]-insulin but in the presence of a 1000-fold excess of unlabeled insulin (human, recombinant; Serva, Heidelberg). After 30 min of incubation at 20°C cells were collected and washed three times with ice-cold culture medium. Their bound radioactivity was measured in a gamma counter. The insulin binding of cultured cells was assayed using cell monolayers seeded in tissue chamber slides (Nunc, Mannheim). For TB, culture medium was removed and replaced by 200 ttl medium containing [12 5I]-insulin at a concentration of 70 pM and cells were incubated for 30 min at 200 C. NB was assayed in the presence of a 1000-fold excess of unlabeled insulin. The incubation was terminated by removing the medium, washing the monolayers 3 x with 200 /1 ice-cold medium and finally fixing them by paraformaldehyde solution (10%). The fixed cells were dried, and the chamber slides were cut into pieces containing the monolayer of one chamber. The radioactivity of these pieces was measured in the gamma counter. (i) Kinetic experiments (a) Association The time course of binding was determined between 5 and 60 min at a concentration of 80 pM [125I]insulin. (b) Dissociation After association for 30 min at 80 pM [125I]insulin, the ligand-containing medium was removed and replaced by normal medium. The time course of dissociation for TB and NB was determined between 5 and 60 min.
414 [
I]-INSULIN BINDING after CELLDISPERSION with 3 different collagenases
1.00
zZ
f
z
-
w
_
0.75
E
0.50
0
total nonspecific
m z
1)
o0
z4
N'__
E
0.25
a,
0.00
C1 C2 C3 Oh
C1 C2 C3 12 h
C1 C2 C3 24 h
Fig. 1. [ 25 1]-insulin binding (mean of three independent experiments) to freshly isolated cells (0 h) and after 12 and 24h of culture. Three different collagenases were used for cell isolation: C = Biotechnologie collagenase, C2 = Sigma type IV. C3 = Boehringer, 25 type D. Incubations for TB were performed with 70 pM [ 1]-insulin and for NB at the same concentration but in the presence of a 1000-fold excess of unlabeled insulin. The incubation time was 30 min at 20°C. Note that Figs. 2-4, contrary to Figs. 1 and 5, display results from one particular experiment out of three independent incubations; mean values + S.D. of the three experiments are provided in Table 1 as well as in the text.
(ii) Saturation experiments To determine TB, cell monolayers were incubated for 30 min with culture medium containing [125I]insulin at 20 concentrations ranging from 2.8 to 5335 pM. NB was assayed in parallel at the same concentrations but in the presence of a 1000-fold excess of unlabeled insulin. (iii) Displacement experiments Cell monolayers were incubated with 312 pM [125I]-insulin for 30 min, both in the presence or absence of unlabeled peptides. Unlabeled insulin (human, recombinant), insulin-like growth factor IGF-I (human, recombinant), IGF-II (human, recombinant) and glucagon (bovine) (all peptides: Serva, Heidelberg, Germany) were present at concentrations from 10- Il to 10- 5 M. Analysis of binding data TB and NB were expressed as [12 5I]-insulin bound/mg cell protein. Protein was measured in chambers incubated in parallel to the chambers
receiving [125I]-insulin, using the method of Lowry et al. (1951). Protein contents per assay were between 0.2-0.3 mg. Specific insulin binding was calculated as the difference between TB and NB. Binding characteristics were analysed following Scatchard (1949) and Hill (1910). The analyses for a one-site or a two-site model of [12 5I]-insulin binding were performed with the LIGAND computer program (Munson 1983) calculating dissociation constants (Kd) and maximal binding capacities (Bmax). The resulting Bma values of the two-site model (sum of Bmal = high affinity binding site, and of Bmax2 = low affinity binding site) were used for Hill plot analyses resulting in nH (Hill coefficient) and EC 50 (effective concentration) where half maximal binding occurs. In the displacement experiments, the ligand bound specifically was expressed as percentage of specific [ 25I]-insulin binding. The specific affinities of the unlabeled ligands, their inhibiting concentration (IC 50, where half of the [125 1]-insulin binding is competitively displaced by the unlabeled
415 -
ASSOCIATION
1 .UU
o z
o
Z LJi 0
Z
r
D
E
z
o
i
E
m
ON
0.75
0.50
0
) total I specific v nonspecific
0.25
0.00 0
15
30
45
60
75
90
TIME [ min ] 5 Fig. 2. Time course of [125I]-insulin binding for TB, NB and the resulting specific binding at 80 pM [12 1]-insulin. Dissociation was assayed after 30 minutes of association, with the medium containing the labeled ligand being replaced by normal culture medium.
ligand) and EC 50 of the competitive ligand were calculated. The IC 50 was calculated by transforming the displacement curve in a logit-log plot. The intersection of the linear regression and the zero line gives the IC 50 value. The corresponding EC50 of the competitive ligand was calculated following Cheng and Prusoff (1973) as EC50 = IC 50 /(1 + F/EC 50 ), where F is the concentration of free labeled ligand and EC50 the value for the labeled ligand obtained in saturation experiments. Insulin bioactivity Metabolic experiments were conducted with cells from the same preparation which were used for saturation and displacement studies. Freshly isolated cells were first cultured in serum-free, defined medium for 12h. After change of the medium, the cells were incubated under control conditions or in the presence of 10 - 7 M human recombinant insulin. For analysis of protein synthesis, 1.5 Ci U-14 C-leucine (Amersham, Braunschweig; 307 mCi/mmol) was added to each dish and the cells were incubated for another 12h. After this period, medium was removed and the cells were washed 3 times with ice-cold medium. After adding 1 ml of ice-cold 20% trichloroacetic acid, the cells were transferred to centrifuge tubes and washed 3 times
with trichloroacetic acid (check of washing solution for radioactivity). The cell pellets were solubilized by incubation for 2h at 37°C in 1N NaOH, mixed with scintillation fluid and radioactivity was counted. For studies of lipogenesis, 1.5 C/Ci U-1 4 Cacetate (sodium salt, Amersham, Braunschweig; 53.7 mCi/mmol) was added to each dish. After a 3, 12 or 24h incubation period, medium was removed and the cells were washed and extracted for lipids by the method of Bligh and Dyer (1959). Lipid extracts were dried under nitrogen, redissolved in 100 /1l chloroform, mixed with scintillation fluid and radioactivity was counted. For studies of cellular glucose release into the medium, a 150 /d aliquot of the medium was stopped by addition of 10 1 60% PCA. Glucose was analysed by the glucose oxidase method using the GOD kit of Boehringer (Mannheim).
Results Influence of collagenase The preliminary experiments demonstrated the existence of specific [125I]-insulin binding (TB-NB = specific binding) to carp hepatocytes directly after cell isolation and after 12 and 24h of culture (Fig. 1).
416 5
4 m
cific
a
Q_
E
i
T E
3C
2
1
K - 4.32.10 1
20
M
Bml,= 0.09 fmtomol
15 0
0l
.
Kd2- 1.58*10 Bm - 12.40 femto
10
ool
m 5
i
0 0.0
, 2.5
5.0
7.5
10.0
12.5
25
[ I]-INSULIN BOUND [ femtomol/mg PROTEIN ]
m E m
i52 pM
0
o~
4 log [
25
which even declined during subsequent culture. The Sigma and Boehringer collagenases resulted in higher capacities of [125I]-insulin binding, both initially and after 12 or 24h of culture. These findings are in accordance to cytological examinations of freshly isolated hepatocytes which revealed more severe cell damage after isolation with Biotechnologie collagenase than with the two other items (Segner, unpublished). Accordingly, reduction of cellular glucose release in response to insulin treatment was only observed with carp liver cells isolated with Boehringer or Sigma collagenase (data not shown).
I]-INSULIN FREE [ pM ]
Fig. 3. Saturation experiments. a. Saturation curve for TB, NB and resulting specific binding after incubation with various concentrations of [125 1]-insulin ranging from 2.8 to 5335.0 pM. NB appeared nearly linear at higher concentrations. b. Scatchard plot analysis of the data of Fig. 3b. The curvilinear plot indicates the presence of two populations of [12 51]-insulin binding sites: a high affinity and low capacity site and a low affinity, high capacity site. c. Hill plot corresponding to Fig. 3c. The slope of the linear regression resulted in nH = 0.87. The intersection between regression line and x-axis resulted in EC 50 = 1.8 nM, in this individual incubation.
Specific binding was less than 1 of the total [125 I]-insulin added, but the capability of the cells to bind [1251I]-insulin was strongly dependent on the collagenase used for cell dispersion. The use of Biotechnologie collagenase yielded only a small capacity of [125I]-insulin binding in freshly isolated cells
Binding studies (i) Kinetic experiments The time course of TB and NB for association (Fig. 2) resulted in a steady state of specific insulin binding after 25 to 30 min. The dissociation curve showed that the binding is almost completely reversible within 60 min. (ii) Saturation experiments Specific [125I]-insulin binding to cultured carp hepatocytes is saturable while NB appears linear at higher concentrations of free [125I]-insulin (Fig. 3a). The analyses of specific insulin binding using the LIGAND program indicated that the two-site model (existence of two populations of binding sites) was a better fit than the one-site model. These findings were confirmed by the corresponding Scatchard plots (Fig. 3b) revealing a high affinity, low capacity insulin binding site (average value from three independent experiments: Kdl = 5.5 2.8 pM, Bml = 0.14 0.06 fmol/mg protein) and a low affinity, high capacity binding site (Kd 2 = 2.4 + 1.1 nM, Bmax2 = 17.6
9.8 fmol/mg
protein). The conversion of the Bma values from [12 5I]-insulin bound/mg protein to [125 I]-insulin bound/cell resulted in 77 ± 31 high affinity binding sites/cell, and 9623 5335 low affinity binding sites/cell. The Hill analyses (Fig. 3c) confirmed by a nH < 1 the existence of two populations of binding sites and resulted in EC 50 = 3.0 ± 1.7 nM. The binding characteristics for human [12 5I]-insulin to carp hepatocytes are summarized in Table 1.
417 125
CD z 0 m z .J Z, z
?
N
100
75
50
25
0 0
-11
-10
-9
-8
-7
COMPETITIVE LIGAND
log [
-6
-5
M ]
Fig. 4. Displacement curves of [1251]-insulin binding (percentage of [125 1]-insulin bound specifically at a concentration of 312 pM), in the presence of insulin. IGF-I, IGF-II or glucagon ranging from 10-ll to 10-5.
Table 1. Parameters of [12 51]-insulin binding to monolayers of carp hepatocytes in primary cell culture Analyses by the LIGAND (Munson 1983) computer program: Bma:x Kd: High affinity binding sites: 5.5 + 2.8 [pM] 0.14 + 0.06 [fmol/mg protein] 77 + 31 [binding sites/cell]
Low affinity binding sites: 2.4
9.8 [fmol/mg protein] 1.1 [nM] 17.6 9623 5335 [binding sites/cell]
Analyses by the Hill (1910) equation:
2.0 nM was in the same range as the EC50c for [125I]-insulin (3.0 1.7 nM), followed by IGF-I (EC50c = 7.1 nM) and IGF-II (EC 50 c = 18.4 nM). Insulin bioactivity Insulin effects on the synthesis of proteins and lipids and on the cellular glucose release could be readily demonstrated. Incorporation of 14Cleucine into proteins of carp liver cells (Fig. 5a) and incorporation of 4C-acetate into cellular lipids (Fig. 5b) was clearly stimulated by insulin treatment, whereas hepatocellular glucose release (Fig. 5c) was reduced in insulin-treated cells.
nH (Hill coefficient): 0.84 + 0.03
EC 5 0:
3.0
+ 1.7 [nM]
Values are given as mean + SD of 3 independent experiments. Within each experiment, the incubations were done in triplicate. The variation within the triplicates was less than 15%70.
(iii) Displacement experiments [125I]-insulin was competitively displaced by unlabeled insulin. IGF-I and IGF-II, with IC 50-values of 2.2, 7.9 and 20.3 nM. respectively (Fig. 4). Glucagon had no effect on insulin binding. The affinity for unlabeled insulin with an EC 50 c value of
Discussion The understanding both of insulin-receptor interaction and of the metabolic effects of insulin in teleostean tissues is still insufficient (e.g., Plisetskaya 1990). Carp hepatocytes in primary culture are viable for several days and carry out numerous functions characteristic of the liver in vivo (Bohm and Segner, unpublished). Using this system, we wanted to investigate (a) characteristics and specificity of insulin binding in carp liver cells, and (b) the cellular metabolic response to the binding of insulin.
418 .S C c 0
a Po =ID 0 Un o
°
of Sa 0
E a
U
4 hours of incubation
0 C
0
a
0
0
~
o
a- -"
o C
o
a
a) a,
=
aj -,
o
u
M
hours of incubation
-I
0
o oa
;3
E hours of incubation
Fig. 5. Effects of insulin on cell metabolism (mean values of three independent experiments). After 12 h of preincubation, cells were exposed either to control conditions or 10 - 7 M insulin. Following effects were measured during the further exposure period: a) incorporation of 14 C-leucine into cellular protein; b) incorporation of 4C-sodium acetate into lipids; c) glucose released into the medium.
Human insulin instead of carp insulin could be employed for the binding study since the insulin receptor of vertebrates has been functionally well conserved during evolution. Hepatic insulin recep-
tors of different species have remarkably similar affinities for human insulin (Muggeo et al. 1979; Stuart 1986). In addition, since iodination of insulin at (A14 Tyr) seems to be decisive for its biological potency (cf., Mommsen and Plisetskaya 1991; Gutierrez and Plisetskaya 1991) and since carp insulin can not be iodinated at (A 14 Tyr) due to an amino acid substitution (Makower et al. 1982), labeled carp insulin may show less potency in receptor binding than human insulin. Insulin binding to cultured carp hepatocytes shows the features reported from other vertebrates. First, insulin binding is saturable. A steady state was reached within 20-30 min of incubation which agrees closely with data from rat hepatocytes (Terris and Steiner 1975). Second, insulin binding produces a curvilinear Scatchard plot indicating two populations of binding sites (cf., Muggeo et al. 1979). Accordingly, a curvilinear Scatchard plot of insulin binding was found for isolated hepatocytes of rainbow trout and Atlantic salmon Salmo salar (Plisetskaya et al. 1993). However, these authors using a competition binding assay and regression analysis obtained Kd values for the high affinity binding sites (3.2 0.22 nM and 6.8 + 0.72 nM, respectively) which were clearly higher than the data found in our study with carp hepatocytes. Whether such differences in affinities and capacities of hepatic insulin binding are due to various experimental and/or computation procedures, to the physiological condition of the test fish or whether they are due to real species differences awaits further studies. The fact that varying Kd values are also common in studies with rat hepatocytes (e.g., Morin et al. 1982: Kdl = 0.5 nM, Kd 2 = 17.0 nM; Varandani et al. 1982: Kdl = 0.76 nM, Kd2 = 190 nM) suggests that a great part of the variability is due to experimental protocols and to the condition of the experimental fish. A third aspect in which the results on insulin binding on carp hepatocytes is comparable to data from other vertebrates is the specifity of binding. In competitive displacement experiments, unlabeled insulin was the most effective competitor, followed by IGF-I and IGF-II, whereas glucagon was without effect. The same sequence of peptides was described for coho salmon (Gutierrez and Plisetskaya 1991).
419 Human insulin evoked significant alterations in the metabolism of carp hepatocytes, i.e., reduced glucose release and increased protein synthesis and lipogenesis from acetate. Dose-response experiments (data not shown) revealed that optimum concentrations were between 10- 7 and 10-8 M insulin. Stimulation of protein synthesis is one of the best documented effects of insulin in fish hepatocytes (Mommsen and Plisetskaya 1991). On the other hand, insulin effects on fish liver carbohydrate metabolism are somewhat equivocal (Foster and Moon 1989). Likewise, an influence of insulin on hepatic lipogenesis of teleosts is controversial (Mommsen and Plisetskaya 1991), although recent reports for rainbow trout (Sheridan 1993; Segner et al. 1993) as well as our data on carp hepatocytes are in favor of a stimulatory action of insulin. The results of this study provide basic data on the parameters of insulin binding and insulin action in liver cells of an omnivorous teleost, the common carp. Further studies will have to evaluate on the one hand factors regulating the dynamics of the insulin receptor in carp liver cells and, on the other hand, the molecular events occurring in cells after insulin binding to its receptor. Acknowledgements This study was financially supported by German Research Foundation DFG Se 466/2-1. References cited Ablett, R.F., Taylor, M.J. and Selivonchik, D.P. 1983. The effects on high protein and high carbohydrate diets on (1-125) iodoinsulin binding in skeletal muscle membranes and isolated hepatocytes of rainbow trout (Salmo gairdneri). Brit. J. Nutr. 50: 129-140. Blair, J.B., Miller, M.R., Pack, D., Barnes, R., Teh, S.J. and Hinton, D.E. 1990. Isolated trout liver cells: establishing short-term primary cultures exhibiting cell-to-cell interactions. In Vitro Cell. Dev. Biol. 26: 237-249. Cheng, Y.C. and Prusoff, W.H. 1973. Relationship between the inhibition constant (Ki) and the concentration of inhibitor which causes 50 percent inhibition (IC 50 ) of an enzymatic reaction. Biochem. Pharm. 22: 3099-4010. Cowley, D.J. and Sheridan, M.A. 1993. Insulin stimulates hepatic lipogenesis in rainbow trout, Oncorhynchus mykiss. Fish Physiol. Biochem. Vol. 11, pp. 421-428.
Czech, M.P. 1985. The nature and regulation of the insulin receptor: structure and function. Ann. Rev. Physiol. 47: 357-397. Foster, G.D. and Moon, T.W. 1989. Insulin and the regulation of glycogen metabolism and gluconeogenesis in American eel hepatocytes. Gen. comp. Endocrinol. 73: 374-380. Gutierrez, J. and Plisetskaya, E.M. 1991. Insulin binding to liver plasma membranes of coho salmon during smoltification. Gen. comp. Endocrinol. 82: 466-475. Gutierrez, J., Asgard, T., Fabbri, E. and Plisetskaya, E.M. 1991. Insulin binding in skeletal muscle of trout fed carbohydrate-supplmented diet. Fish Physiol. Biochem. 9: 351360. Hill, A.W. 1910. The possible effects of the aggregation of the molecules of hemoglobin on its dissociation curves. J. Physiol. 40: IV-VII. Juul, S.M. and Jones, R.H. 1982. Evidence for a direct effect of bacitracin on cell-mediated insulin degradation in isolated hepatocytes. Biochem. J. 206: 295-299. Leibush, B.N. and Bondareva, V.M. 1981. Insulin receptors in the liver membranes of the scorpionfish, Scorpaena porcus, as compared to mammalian receptor. Zh. Evol. Biokhim. Fiziol. 17: 141-147. (in Russian). Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randell, R.J. 1951. Protein measurement with the folin phenol reagent. J. Biol. Chem. 193: 265-275. Makower, A., Dettmer, R., Rapoport, T.A., Knospe, S., Behlke, J., Prehn, S., Franke, P., Etzold, G. and Rosenthal, S. 1982. Carp insulin: amino acid sequence, biological activity and structural properties. Eur. J. Biochem. 122: 339-345. Mommsen, T.P. and Plisetskaya, E.M. 1991. Insulin in fish and agnathans: history, structure and metabolic regulation. Rev. Aquat. Sci. 4: 225-259. Muggeo, M., Ginsberg, B.H., Roth, J., Neville, D.M., de Meyts, P. and Kahn, C.R. 1979. The insulin receptor in vertebrates is functionally more conserved during evolution than insulin itself. Endocrinology 104: 1393-1402. Munson, P.J. 1983. A computerized analysis of ligand binding data. Meth. Enzymol. 92: 543-576. Morin, O., Fehlmann, M. and Freychet, P. 1982. Binding and action of insulin and glucagon in monolayer cultures and fresh suspensions of rat hepatocytes. Mol. Cell. Endocrinol. 25: 339-352. Plisetskaya, E.M. 1990. Recent studies on fish pancreatic hormones: selected topics. Zool. Sci. 7: 335-353. Plisetskaya, E.M., Fabbri, E., Moon, T.W., Gutierrez, J. and Ottolenghi, C. 1993. Insulin binding to isolated hepatocytes of Atlantic salmon and rainbow trout. Fish Physiol. Biochem. Vol. 11, pp. 401-409. Plisetskaya, E.M. and Sullivan, C.V. 1989. Pancreatic and thyroid hormones in rainbow trout (Salmo gairdner): what concentrations does the liver see? Gen. comp. Endocrinol. 75: 310-315. Scatchard, G. 1949. The attractions of proteins for small molecules and ions. Ann. N.Y. Acad. Sci. 51: 660-672. Segner, H., Blair, J.B., Wirtz, G. and Miller, M.R. 1993. Cultured trout liver cells: utilization of substrates and response to hormones. In Vitro Cell. Dev. Biol. (In press).
420 Stuart, C.A. 1986. Phylogenetic distance from man correlates with immunologic cross-reactivity among liver-insulin receptors. Comp. Biochem. Physiol. 84B: 167-172. Scapin, S. and Incerpi, S. 1992. Annual variations in the binding of insulin to hepatic membranes of the frog Rana esculenta. Gen. comp. Endocrinol. 86: 128-137. Terris, S. and Steiner, D.F. 1975. Binding and degradation
of [1 25 1]-insulin by rat hepatocytes. J. Biol. Chem. 250: 8389-8398. Varandani, P.T., Darrow, R.M., Nafz, M.A. and Norris, G.L. 1982. Binding, degradation, and bioactivity of insulin in primary cultures of rat hepatocytes. Am. J. Physiol. 243: E132-E139.
Binding and bioactivity of insulin in primary cultures of carp (Cyprinus carpio) hepatocytes 2 Helmut Segner l , Ralf B6hm and Werner Kloas Zoology II, University of Karlsruhe, Kaiserstr. 12 D- W- 7500 Karlsruhe, Germany
Keywords: carp, liver cell culture, insulin, insulin binding, insulin bioactivity
Resum6 Des h6patocytes isoles de carpes, cultiv6s dans un milieu d6fini chimiquement et ne contenant pas de s6rum, ont t6 utilis6s pour tudier les caract6ristiques de liaison de l'insuline ainsi que ses effets m6taboliques. Des experiences de cin6tiques, de saturation et de d6placement ont et6 effectuees pour caract6riser la liaison de 1'12 5I-insuline. Un effet de la collagenase sur la liaison a t6 d6montr6e. La liaison de l'insuline d6crolt durant les douze premieres heures de culture, puis reste constante durant les douze heures suivantes. La liaison de 1' 12 5I-insuline atteint un tat d'6quilibre apres 20-30 min d'incubation. L'analyse math6matique des courbes de saturation rvele l'existence de deux sites de liaison, I'un de haute affinity (Kdl = 5.5 pM) avec une faible capacity (Bmax = 0.14 fmol/mg de prot6ine soit 77 sites de liaison par cellule) et l'autre de plus faible affinity (Kd 2 = 2.4 nM) et de plus grande capacity (Bmax 2 = 17.6 fmol/mg de prot6ine soit 9623 sites/cellule). Une 6tude de competition a montr6 que la moiti6 de liaison de 312 pM d' 12 5I-insuline est d6plac6e par 2.2 nM d'insuline froide, 7.9 nM d'IGF-I et 20.3 nM d'IGF-II. Le glucagon est sans effet. La liaison de l'insuline aux hepatocytes de carpes induit une rduction significative de la scretion de glucose et une augmentation significative de la synthese prot6ique et de la synthese de novo des acides gras. Abstract Isolated carp hepatocytes cultured in serum-free, chemically defined medium were used to investigate within the same cell preparation characteristics of the binding of insulin as well as effects of insulin on cellular metabolism. The binding of human [125I]-insulin to carp hepatocytes was studied in kinetic, saturation and displacement experiments. A dependency of insulin binding on the collagenase used for cell isolation was demonstrated. Insulin binding decreased during the first 12h of culture but remained constant during the following 12h. The kinetic experiments revealed that [125I]-insulin binding reached a steady state within 20-30 min of incubation. The mathematical analysis of the saturation experiments demonstrated the existence of two populations of binding sites, one with high affinity (Kdl = 5.5 pM) and low capacity (Bmaxl = 0.14 fmol/mg protein or 77 binding sites/cell) and one with low affinity (Kd 2 = 2.4 nM) and high capacity (Bmax 2 = 17.6 fmol/mg protein or 9623 binding sites/cell). In competition experiments, 312 pM [125 I]-insulin was displaced by cold insulin, IGF-I and IGF-II with IC 50 values of 2.2, 7.9 and 20.3 nM, respectively. Glucagon was without effect. Binding of insulin to carp hepatocytes resulted in a significant reduction of glucose release and a significant increase of protein synthesis as of de novo fatty acid synthesis. 1Present address: Environmental Research Center, Section for Environmental Chemistry and Ecotoxicology, Permoserstr. 15, D-0-7050 Leipzig, Germany; 2 dedicated to Prof. Dr. W. Hanke on the occasion of his 6 5th birthday.
412 Introduction The binding of insulin to its cell surface receptor is the first step to initiate physiological effects in target cells. The receptor plays a critical role in both directing the hormone to a specific target tissue and programming the biological responses of the tissue to the hormone. The liver receives its blood through the portal system which drains the endocrine pancreas. Therefore, the liver is the first organ to see any changes in insulin secretion (Plisetskaya and Sullivan 1989) and represents a primary target for insulin action. The insulin receptor of liver cells has been extensively studied in mammals (e.g., Czech 1985). For teleost fish, on the other hand, only scarce and partly contradictory information on insulin receptors is available (Plisetskaya 1990). Muggeo et al. (1979), when comparing insulin receptors from a variety of vertebrate species, described characteristics of insulin binding to erythrocytes of rainbow trout (Oncorhynchus mykiss). With respect to the liver of teleosts, Ablett et al. (1983) were the first to study the specific interaction of insulin with hepatocytes. Kinetic data on insulin binding sites in teleostean liver including Scatchard plot analysis, saturation and displacement experiments, have been published to date only for the two closely related carnivorous species, coho salmon, (Oncorhynchus kisutch, Gutierrez and Plisetskaya 1991) and rainbow trout (Gutierrez et al. 1991) as well as for the scorpionfish (Scorpaena porcus, Leibush and Bondareva 1991, cited by Gutierrez and Plisetskaya 1991). The small data base is surprising, particularly when considering the possible involvement of defects at the insulin binding level in the well-documented glucose intolerance of fish (Mommsen and Plisetskaya 1991; Gutierrez et al. 1991). The present communication reports data on insulin binding to liver cells of the omnivorous carp, Cyprinus carpio. Instead of isolated liver membranes, as employed in the binding studies with trout and coho salmon (Gutierrez et al. 1990, Gutierrez and Plisetskaya 1991), we use isolated, intact liver cells cultured in serum-free, chemically defined medium. This enables us to investigate within the same preparation the characteristics of in-
sulin binding as well as metabolic responses of the cells to insulin (see also Varandani et al. 1982).
Materials and methods Animals Two-year-old Cyprinus carpio of 200-400 g body weight were used for cell isolation. The fishes were maintained under a 12: 12 dark: light cycle at a constant water temperature of 20°C in aerated tanks equipped with a flow-through water system. Once per day (17.00h) the fishes were fed adlibitum with a commercial pelleted feed containing 40% crude protein and 7% crude fat. The carp were acclimated to those conditions for at least one year. The binding studies were performed from September to October 1991. Cell isolation After the fish was anaesthesized, the aorta coeliaca coming from the dorsal aorta to the liver was cannulated. The liver was cleared of blood by perfusion for 10 min with an aerated Ca 2 + - and Mg2 +-free Hank's salt solution containing 5 mmol/l EDTA and 15 mM HEPES. After all the blood had been cleared, a solution containing Mg2 +-free Hank's salts, 15 mM HEPES and 0.026% collagenase (Collagenase D, Boehringer Mannheim, 0,026%) was perfused for 20 min. This was followed by a final 5-min perfusion with a solution made of Ca 2 + - and Mg2 + - free Hank's salts, 2 mM EDTA and 15 mM HEPES. The liver was then dissected, minced into small pieces and passed through nylon sieves of 250, 100 and 50 zm mesh size. Cells were collected and repeatedly washed by centrifugation at 80 x g. The cell pellet was resuspended in culture medium and the cell number was counted in a hemocytometer. Cell viability according to trypan blue exclusion was usually higher than 950,'o. Final cell density was adjusted to 1.5-2 x 106 cells/ml. Cell culture The culture system was similar to that described for rainbow trout hepatocytes by Blair et al. (1990). The cells were seeded as monolayers in tissue cham-
413 ber slides (Nunc, Mannheim) for the binding studies (200 Al/chamber), or in 35 mm Petri dishes (Falcon Primaria, Heidelberg) for metabolic studies (1.5 ml/dish). The culture medium consisted of Hank's salts (138 mM NaCl, 5.4 mM KCI, 0.8 mM MgSO 4, 1.86 mM CaCI 2 , 0.34 mM Na 2 HPO 4, 0.44 mM KH 2PO 4 ), 0.05 mM NaHCO 3, 20 mM HEPES, 20 ml/l medium of MEM amino acid concentrate (50 x stock; Serva, Heidelberg), 10 ml/l medium of BME vitamin concentrate (100 x stock; Serva, Heidelberg), 3.3 ml/l medium of 0.5% sodium phenol-red (Serva, Heidelberg), 100 mg/l medium of streptomycin (Sigma, Deisenhofen), 100,000 U/I medium of penicillin (Sigma, Deissenhofen) and 2 mM L-glutamine (Serva, Heidelberg). To minimize insulin destruction, 25 mg/100 ml of bacitracin (Sigma, Deissenhofen) was added: this chemical does not affect insulin binding processes (Juul and Jones 1982). The cells were cultured in a humified air atmosphere at 20°C. The medium was changed every 12h. Influence of collagenase on cellular insulin binding In preliminary experiments, the influence of different collagenases on cellular insulin binding was investigated. Samples were taken immediately after isolation (0 h), and after 12 and 24h of culture. The tested collagenases included a preparation from PBH Biotechnologie (Hannover: 0.1-0.15 PZunits/mg lyophilisate. One PZ-unit is the amount of enzyme which hydrolyses 1 /mole of the peptide PZ-pro-leu-gly-pro-D-arg per min at 25°C and pH 7.1), type IV from Sigma (Deissenhofen; 1.1 FALGPA-units/mg lyophilisate. One FALGPAunit is the amount of enzyme which hydrolyses 1 /mole of the peptide furylacryloyl-leu-gly-pro-ala per min at 25 °C at pH 7.5 in the presence of calcium ions) and type D from Boehringer (Mannheim, 0.59 Wiinsch-units/mg lyophilisate. One Wiinsch-unit is the amount of enzyme which hydrolyses 1 Armole of the peptide 4-phenyl-azo-benzol-oxycarbonyl-Lprolyl-L-leucyl- L-glycyl-L-prolyl-D-arginine per min at 25C). For all further studies, only cells isolated with collagenase D and precultured for at least 12 h were used. The treatments were always done in triplicates.
Binding studies To determine insulin binding of freshly isolated cells (Oh), 200 1lof cell suspension were added to Eppendorf vials and cells were collected by centrifugation at 80 x g. The supernatant was carefully removed and cells were resuspended in 200 Al culture medium containing 70 pM [125 I]-insulin (human, iodinated at tyrosine A14, Amersham, Braunschweig; specific activity 1960 Ci/mmol) for the detection of total binding (TB). Nonspecific binding (NB) was assayed at the same concentration of [125 I]-insulin but in the presence of a 1000-fold excess of unlabeled insulin (human, recombinant; Serva, Heidelberg). After 30 min of incubation at 20°C cells were collected and washed three times with ice-cold culture medium. Their bound radioactivity was measured in a gamma counter. The insulin binding of cultured cells was assayed using cell monolayers seeded in tissue chamber slides (Nunc, Mannheim). For TB, culture medium was removed and replaced by 200 ttl medium containing [12 5I]-insulin at a concentration of 70 pM and cells were incubated for 30 min at 200 C. NB was assayed in the presence of a 1000-fold excess of unlabeled insulin. The incubation was terminated by removing the medium, washing the monolayers 3 x with 200 /1 ice-cold medium and finally fixing them by paraformaldehyde solution (10%). The fixed cells were dried, and the chamber slides were cut into pieces containing the monolayer of one chamber. The radioactivity of these pieces was measured in the gamma counter. (i) Kinetic experiments (a) Association The time course of binding was determined between 5 and 60 min at a concentration of 80 pM [125I]insulin. (b) Dissociation After association for 30 min at 80 pM [125I]insulin, the ligand-containing medium was removed and replaced by normal medium. The time course of dissociation for TB and NB was determined between 5 and 60 min.
414 [
I]-INSULIN BINDING after CELLDISPERSION with 3 different collagenases
1.00
zZ
f
z
-
w
_
0.75
E
0.50
0
total nonspecific
m z
1)
o0
z4
N'__
E
0.25
a,
0.00
C1 C2 C3 Oh
C1 C2 C3 12 h
C1 C2 C3 24 h
Fig. 1. [ 25 1]-insulin binding (mean of three independent experiments) to freshly isolated cells (0 h) and after 12 and 24h of culture. Three different collagenases were used for cell isolation: C = Biotechnologie collagenase, C2 = Sigma type IV. C3 = Boehringer, 25 type D. Incubations for TB were performed with 70 pM [ 1]-insulin and for NB at the same concentration but in the presence of a 1000-fold excess of unlabeled insulin. The incubation time was 30 min at 20°C. Note that Figs. 2-4, contrary to Figs. 1 and 5, display results from one particular experiment out of three independent incubations; mean values + S.D. of the three experiments are provided in Table 1 as well as in the text.
(ii) Saturation experiments To determine TB, cell monolayers were incubated for 30 min with culture medium containing [125I]insulin at 20 concentrations ranging from 2.8 to 5335 pM. NB was assayed in parallel at the same concentrations but in the presence of a 1000-fold excess of unlabeled insulin. (iii) Displacement experiments Cell monolayers were incubated with 312 pM [125I]-insulin for 30 min, both in the presence or absence of unlabeled peptides. Unlabeled insulin (human, recombinant), insulin-like growth factor IGF-I (human, recombinant), IGF-II (human, recombinant) and glucagon (bovine) (all peptides: Serva, Heidelberg, Germany) were present at concentrations from 10- Il to 10- 5 M. Analysis of binding data TB and NB were expressed as [12 5I]-insulin bound/mg cell protein. Protein was measured in chambers incubated in parallel to the chambers
receiving [125I]-insulin, using the method of Lowry et al. (1951). Protein contents per assay were between 0.2-0.3 mg. Specific insulin binding was calculated as the difference between TB and NB. Binding characteristics were analysed following Scatchard (1949) and Hill (1910). The analyses for a one-site or a two-site model of [12 5I]-insulin binding were performed with the LIGAND computer program (Munson 1983) calculating dissociation constants (Kd) and maximal binding capacities (Bmax). The resulting Bma values of the two-site model (sum of Bmal = high affinity binding site, and of Bmax2 = low affinity binding site) were used for Hill plot analyses resulting in nH (Hill coefficient) and EC 50 (effective concentration) where half maximal binding occurs. In the displacement experiments, the ligand bound specifically was expressed as percentage of specific [ 25I]-insulin binding. The specific affinities of the unlabeled ligands, their inhibiting concentration (IC 50, where half of the [125 1]-insulin binding is competitively displaced by the unlabeled
415 -
ASSOCIATION
1 .UU
o z
o
Z LJi 0
Z
r
D
E
z
o
i
E
m
ON
0.75
0.50
0
) total I specific v nonspecific
0.25
0.00 0
15
30
45
60
75
90
TIME [ min ] 5 Fig. 2. Time course of [125I]-insulin binding for TB, NB and the resulting specific binding at 80 pM [12 1]-insulin. Dissociation was assayed after 30 minutes of association, with the medium containing the labeled ligand being replaced by normal culture medium.
ligand) and EC 50 of the competitive ligand were calculated. The IC 50 was calculated by transforming the displacement curve in a logit-log plot. The intersection of the linear regression and the zero line gives the IC 50 value. The corresponding EC50 of the competitive ligand was calculated following Cheng and Prusoff (1973) as EC50 = IC 50 /(1 + F/EC 50 ), where F is the concentration of free labeled ligand and EC50 the value for the labeled ligand obtained in saturation experiments. Insulin bioactivity Metabolic experiments were conducted with cells from the same preparation which were used for saturation and displacement studies. Freshly isolated cells were first cultured in serum-free, defined medium for 12h. After change of the medium, the cells were incubated under control conditions or in the presence of 10 - 7 M human recombinant insulin. For analysis of protein synthesis, 1.5 Ci U-14 C-leucine (Amersham, Braunschweig; 307 mCi/mmol) was added to each dish and the cells were incubated for another 12h. After this period, medium was removed and the cells were washed 3 times with ice-cold medium. After adding 1 ml of ice-cold 20% trichloroacetic acid, the cells were transferred to centrifuge tubes and washed 3 times
with trichloroacetic acid (check of washing solution for radioactivity). The cell pellets were solubilized by incubation for 2h at 37°C in 1N NaOH, mixed with scintillation fluid and radioactivity was counted. For studies of lipogenesis, 1.5 C/Ci U-1 4 Cacetate (sodium salt, Amersham, Braunschweig; 53.7 mCi/mmol) was added to each dish. After a 3, 12 or 24h incubation period, medium was removed and the cells were washed and extracted for lipids by the method of Bligh and Dyer (1959). Lipid extracts were dried under nitrogen, redissolved in 100 /1l chloroform, mixed with scintillation fluid and radioactivity was counted. For studies of cellular glucose release into the medium, a 150 /d aliquot of the medium was stopped by addition of 10 1 60% PCA. Glucose was analysed by the glucose oxidase method using the GOD kit of Boehringer (Mannheim).
Results Influence of collagenase The preliminary experiments demonstrated the existence of specific [125I]-insulin binding (TB-NB = specific binding) to carp hepatocytes directly after cell isolation and after 12 and 24h of culture (Fig. 1).
416 5
4 m
cific
a
Q_
E
i
T E
3C
2
1
K - 4.32.10 1
20
M
Bml,= 0.09 fmtomol
15 0
0l
.
Kd2- 1.58*10 Bm - 12.40 femto
10
ool
m 5
i
0 0.0
, 2.5
5.0
7.5
10.0
12.5
25
[ I]-INSULIN BOUND [ femtomol/mg PROTEIN ]
m E m
i52 pM
0
o~
4 log [
25
which even declined during subsequent culture. The Sigma and Boehringer collagenases resulted in higher capacities of [125I]-insulin binding, both initially and after 12 or 24h of culture. These findings are in accordance to cytological examinations of freshly isolated hepatocytes which revealed more severe cell damage after isolation with Biotechnologie collagenase than with the two other items (Segner, unpublished). Accordingly, reduction of cellular glucose release in response to insulin treatment was only observed with carp liver cells isolated with Boehringer or Sigma collagenase (data not shown).
I]-INSULIN FREE [ pM ]
Fig. 3. Saturation experiments. a. Saturation curve for TB, NB and resulting specific binding after incubation with various concentrations of [125 1]-insulin ranging from 2.8 to 5335.0 pM. NB appeared nearly linear at higher concentrations. b. Scatchard plot analysis of the data of Fig. 3b. The curvilinear plot indicates the presence of two populations of [12 51]-insulin binding sites: a high affinity and low capacity site and a low affinity, high capacity site. c. Hill plot corresponding to Fig. 3c. The slope of the linear regression resulted in nH = 0.87. The intersection between regression line and x-axis resulted in EC 50 = 1.8 nM, in this individual incubation.
Specific binding was less than 1 of the total [125 I]-insulin added, but the capability of the cells to bind [1251I]-insulin was strongly dependent on the collagenase used for cell dispersion. The use of Biotechnologie collagenase yielded only a small capacity of [125I]-insulin binding in freshly isolated cells
Binding studies (i) Kinetic experiments The time course of TB and NB for association (Fig. 2) resulted in a steady state of specific insulin binding after 25 to 30 min. The dissociation curve showed that the binding is almost completely reversible within 60 min. (ii) Saturation experiments Specific [125I]-insulin binding to cultured carp hepatocytes is saturable while NB appears linear at higher concentrations of free [125I]-insulin (Fig. 3a). The analyses of specific insulin binding using the LIGAND program indicated that the two-site model (existence of two populations of binding sites) was a better fit than the one-site model. These findings were confirmed by the corresponding Scatchard plots (Fig. 3b) revealing a high affinity, low capacity insulin binding site (average value from three independent experiments: Kdl = 5.5 2.8 pM, Bml = 0.14 0.06 fmol/mg protein) and a low affinity, high capacity binding site (Kd 2 = 2.4 + 1.1 nM, Bmax2 = 17.6
9.8 fmol/mg
protein). The conversion of the Bma values from [12 5I]-insulin bound/mg protein to [125 I]-insulin bound/cell resulted in 77 ± 31 high affinity binding sites/cell, and 9623 5335 low affinity binding sites/cell. The Hill analyses (Fig. 3c) confirmed by a nH < 1 the existence of two populations of binding sites and resulted in EC 50 = 3.0 ± 1.7 nM. The binding characteristics for human [12 5I]-insulin to carp hepatocytes are summarized in Table 1.
417 125
CD z 0 m z .J Z, z
?
N
100
75
50
25
0 0
-11
-10
-9
-8
-7
COMPETITIVE LIGAND
log [
-6
-5
M ]
Fig. 4. Displacement curves of [1251]-insulin binding (percentage of [125 1]-insulin bound specifically at a concentration of 312 pM), in the presence of insulin. IGF-I, IGF-II or glucagon ranging from 10-ll to 10-5.
Table 1. Parameters of [12 51]-insulin binding to monolayers of carp hepatocytes in primary cell culture Analyses by the LIGAND (Munson 1983) computer program: Bma:x Kd: High affinity binding sites: 5.5 + 2.8 [pM] 0.14 + 0.06 [fmol/mg protein] 77 + 31 [binding sites/cell]
Low affinity binding sites: 2.4
9.8 [fmol/mg protein] 1.1 [nM] 17.6 9623 5335 [binding sites/cell]
Analyses by the Hill (1910) equation:
2.0 nM was in the same range as the EC50c for [125I]-insulin (3.0 1.7 nM), followed by IGF-I (EC50c = 7.1 nM) and IGF-II (EC 50 c = 18.4 nM). Insulin bioactivity Insulin effects on the synthesis of proteins and lipids and on the cellular glucose release could be readily demonstrated. Incorporation of 14Cleucine into proteins of carp liver cells (Fig. 5a) and incorporation of 4C-acetate into cellular lipids (Fig. 5b) was clearly stimulated by insulin treatment, whereas hepatocellular glucose release (Fig. 5c) was reduced in insulin-treated cells.
nH (Hill coefficient): 0.84 + 0.03
EC 5 0:
3.0
+ 1.7 [nM]
Values are given as mean + SD of 3 independent experiments. Within each experiment, the incubations were done in triplicate. The variation within the triplicates was less than 15%70.
(iii) Displacement experiments [125I]-insulin was competitively displaced by unlabeled insulin. IGF-I and IGF-II, with IC 50-values of 2.2, 7.9 and 20.3 nM. respectively (Fig. 4). Glucagon had no effect on insulin binding. The affinity for unlabeled insulin with an EC 50 c value of
Discussion The understanding both of insulin-receptor interaction and of the metabolic effects of insulin in teleostean tissues is still insufficient (e.g., Plisetskaya 1990). Carp hepatocytes in primary culture are viable for several days and carry out numerous functions characteristic of the liver in vivo (Bohm and Segner, unpublished). Using this system, we wanted to investigate (a) characteristics and specificity of insulin binding in carp liver cells, and (b) the cellular metabolic response to the binding of insulin.
418 .S C c 0
a Po =ID 0 Un o
°
of Sa 0
E a
U
4 hours of incubation
0 C
0
a
0
0
~
o
a- -"
o C
o
a
a) a,
=
aj -,
o
u
M
hours of incubation
-I
0
o oa
;3
E hours of incubation
Fig. 5. Effects of insulin on cell metabolism (mean values of three independent experiments). After 12 h of preincubation, cells were exposed either to control conditions or 10 - 7 M insulin. Following effects were measured during the further exposure period: a) incorporation of 14 C-leucine into cellular protein; b) incorporation of 4C-sodium acetate into lipids; c) glucose released into the medium.
Human insulin instead of carp insulin could be employed for the binding study since the insulin receptor of vertebrates has been functionally well conserved during evolution. Hepatic insulin recep-
tors of different species have remarkably similar affinities for human insulin (Muggeo et al. 1979; Stuart 1986). In addition, since iodination of insulin at (A14 Tyr) seems to be decisive for its biological potency (cf., Mommsen and Plisetskaya 1991; Gutierrez and Plisetskaya 1991) and since carp insulin can not be iodinated at (A 14 Tyr) due to an amino acid substitution (Makower et al. 1982), labeled carp insulin may show less potency in receptor binding than human insulin. Insulin binding to cultured carp hepatocytes shows the features reported from other vertebrates. First, insulin binding is saturable. A steady state was reached within 20-30 min of incubation which agrees closely with data from rat hepatocytes (Terris and Steiner 1975). Second, insulin binding produces a curvilinear Scatchard plot indicating two populations of binding sites (cf., Muggeo et al. 1979). Accordingly, a curvilinear Scatchard plot of insulin binding was found for isolated hepatocytes of rainbow trout and Atlantic salmon Salmo salar (Plisetskaya et al. 1993). However, these authors using a competition binding assay and regression analysis obtained Kd values for the high affinity binding sites (3.2 0.22 nM and 6.8 + 0.72 nM, respectively) which were clearly higher than the data found in our study with carp hepatocytes. Whether such differences in affinities and capacities of hepatic insulin binding are due to various experimental and/or computation procedures, to the physiological condition of the test fish or whether they are due to real species differences awaits further studies. The fact that varying Kd values are also common in studies with rat hepatocytes (e.g., Morin et al. 1982: Kdl = 0.5 nM, Kd 2 = 17.0 nM; Varandani et al. 1982: Kdl = 0.76 nM, Kd2 = 190 nM) suggests that a great part of the variability is due to experimental protocols and to the condition of the experimental fish. A third aspect in which the results on insulin binding on carp hepatocytes is comparable to data from other vertebrates is the specifity of binding. In competitive displacement experiments, unlabeled insulin was the most effective competitor, followed by IGF-I and IGF-II, whereas glucagon was without effect. The same sequence of peptides was described for coho salmon (Gutierrez and Plisetskaya 1991).
419 Human insulin evoked significant alterations in the metabolism of carp hepatocytes, i.e., reduced glucose release and increased protein synthesis and lipogenesis from acetate. Dose-response experiments (data not shown) revealed that optimum concentrations were between 10- 7 and 10-8 M insulin. Stimulation of protein synthesis is one of the best documented effects of insulin in fish hepatocytes (Mommsen and Plisetskaya 1991). On the other hand, insulin effects on fish liver carbohydrate metabolism are somewhat equivocal (Foster and Moon 1989). Likewise, an influence of insulin on hepatic lipogenesis of teleosts is controversial (Mommsen and Plisetskaya 1991), although recent reports for rainbow trout (Sheridan 1993; Segner et al. 1993) as well as our data on carp hepatocytes are in favor of a stimulatory action of insulin. The results of this study provide basic data on the parameters of insulin binding and insulin action in liver cells of an omnivorous teleost, the common carp. Further studies will have to evaluate on the one hand factors regulating the dynamics of the insulin receptor in carp liver cells and, on the other hand, the molecular events occurring in cells after insulin binding to its receptor. Acknowledgements This study was financially supported by German Research Foundation DFG Se 466/2-1. References cited Ablett, R.F., Taylor, M.J. and Selivonchik, D.P. 1983. The effects on high protein and high carbohydrate diets on (1-125) iodoinsulin binding in skeletal muscle membranes and isolated hepatocytes of rainbow trout (Salmo gairdneri). Brit. J. Nutr. 50: 129-140. Blair, J.B., Miller, M.R., Pack, D., Barnes, R., Teh, S.J. and Hinton, D.E. 1990. Isolated trout liver cells: establishing short-term primary cultures exhibiting cell-to-cell interactions. In Vitro Cell. Dev. Biol. 26: 237-249. Cheng, Y.C. and Prusoff, W.H. 1973. Relationship between the inhibition constant (Ki) and the concentration of inhibitor which causes 50 percent inhibition (IC 50 ) of an enzymatic reaction. Biochem. Pharm. 22: 3099-4010. Cowley, D.J. and Sheridan, M.A. 1993. Insulin stimulates hepatic lipogenesis in rainbow trout, Oncorhynchus mykiss. Fish Physiol. Biochem. Vol. 11, pp. 421-428.
Czech, M.P. 1985. The nature and regulation of the insulin receptor: structure and function. Ann. Rev. Physiol. 47: 357-397. Foster, G.D. and Moon, T.W. 1989. Insulin and the regulation of glycogen metabolism and gluconeogenesis in American eel hepatocytes. Gen. comp. Endocrinol. 73: 374-380. Gutierrez, J. and Plisetskaya, E.M. 1991. Insulin binding to liver plasma membranes of coho salmon during smoltification. Gen. comp. Endocrinol. 82: 466-475. Gutierrez, J., Asgard, T., Fabbri, E. and Plisetskaya, E.M. 1991. Insulin binding in skeletal muscle of trout fed carbohydrate-supplmented diet. Fish Physiol. Biochem. 9: 351360. Hill, A.W. 1910. The possible effects of the aggregation of the molecules of hemoglobin on its dissociation curves. J. Physiol. 40: IV-VII. Juul, S.M. and Jones, R.H. 1982. Evidence for a direct effect of bacitracin on cell-mediated insulin degradation in isolated hepatocytes. Biochem. J. 206: 295-299. Leibush, B.N. and Bondareva, V.M. 1981. Insulin receptors in the liver membranes of the scorpionfish, Scorpaena porcus, as compared to mammalian receptor. Zh. Evol. Biokhim. Fiziol. 17: 141-147. (in Russian). Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randell, R.J. 1951. Protein measurement with the folin phenol reagent. J. Biol. Chem. 193: 265-275. Makower, A., Dettmer, R., Rapoport, T.A., Knospe, S., Behlke, J., Prehn, S., Franke, P., Etzold, G. and Rosenthal, S. 1982. Carp insulin: amino acid sequence, biological activity and structural properties. Eur. J. Biochem. 122: 339-345. Mommsen, T.P. and Plisetskaya, E.M. 1991. Insulin in fish and agnathans: history, structure and metabolic regulation. Rev. Aquat. Sci. 4: 225-259. Muggeo, M., Ginsberg, B.H., Roth, J., Neville, D.M., de Meyts, P. and Kahn, C.R. 1979. The insulin receptor in vertebrates is functionally more conserved during evolution than insulin itself. Endocrinology 104: 1393-1402. Munson, P.J. 1983. A computerized analysis of ligand binding data. Meth. Enzymol. 92: 543-576. Morin, O., Fehlmann, M. and Freychet, P. 1982. Binding and action of insulin and glucagon in monolayer cultures and fresh suspensions of rat hepatocytes. Mol. Cell. Endocrinol. 25: 339-352. Plisetskaya, E.M. 1990. Recent studies on fish pancreatic hormones: selected topics. Zool. Sci. 7: 335-353. Plisetskaya, E.M., Fabbri, E., Moon, T.W., Gutierrez, J. and Ottolenghi, C. 1993. Insulin binding to isolated hepatocytes of Atlantic salmon and rainbow trout. Fish Physiol. Biochem. Vol. 11, pp. 401-409. Plisetskaya, E.M. and Sullivan, C.V. 1989. Pancreatic and thyroid hormones in rainbow trout (Salmo gairdner): what concentrations does the liver see? Gen. comp. Endocrinol. 75: 310-315. Scatchard, G. 1949. The attractions of proteins for small molecules and ions. Ann. N.Y. Acad. Sci. 51: 660-672. Segner, H., Blair, J.B., Wirtz, G. and Miller, M.R. 1993. Cultured trout liver cells: utilization of substrates and response to hormones. In Vitro Cell. Dev. Biol. (In press).
420 Stuart, C.A. 1986. Phylogenetic distance from man correlates with immunologic cross-reactivity among liver-insulin receptors. Comp. Biochem. Physiol. 84B: 167-172. Scapin, S. and Incerpi, S. 1992. Annual variations in the binding of insulin to hepatic membranes of the frog Rana esculenta. Gen. comp. Endocrinol. 86: 128-137. Terris, S. and Steiner, D.F. 1975. Binding and degradation
of [1 25 1]-insulin by rat hepatocytes. J. Biol. Chem. 250: 8389-8398. Varandani, P.T., Darrow, R.M., Nafz, M.A. and Norris, G.L. 1982. Binding, degradation, and bioactivity of insulin in primary cultures of rat hepatocytes. Am. J. Physiol. 243: E132-E139.
