Metformin ameliorates extreme insulin resistance in a patient with anti-insulin receptor antibodies: description of insulin receptor and postreceptor effects in vivo and in vitro Salvatore Di Paolo Policlinico, Università di Bari, Italy
Di Paolo S. Metformin ameliorates extreme insulin resistance in a patient with anti-insulin receptor antibodies: description of insulin receptor and postreceptor effect in vivo and in vitro. Acta Endocrinol 1992:126:117-23. ISSN 0001-5598
The effect of metformin on insulin binding and insulin action in the presence of anti-insulin receptor antibodies was investigated in a case of type B extreme insulin resistance. Oral administration of metformin (1 500 mg/d) for 10 days significantly decreased plasma blood glucose and insulin levels and enhanced the hypoglycemic response to exogenous insulin. In vitro preincubation of normal erythrocytes with insulin receptor antibody from the patient plus 4\m=x\ 10-5 mol/l metformin markedly enhanced insulin binding to receptors, compared to cells incubated with antibody alone. This effect was apparent after 2 h, was maximal after 4 h and did not change up to 24 h. Closely similar results were found when human adipocytes were studied. Analysis of binding data confirmed the increase in both receptor number and affinity. One hour exposure of control adipocytes to metformin enhanced basal lipogenesis by more than 30%. Acute exposure of fat cells to the patient's receptor antibodies resulted in a stimulation of glucose transport and a state of severe insulin resistance. The addition of metformin to antibody in preincubation buffer strongly enhanced basal glucose incorporation into lipids, but did not prevent insulin unresponsiveness. It is suggested that metformin increases, possibly through a change in the spatial conformation of insulin receptor within the plasma membrane, the availability of preexisting receptors to insulin binding and/or decreases the availability of specific epitopes to antibody anchoring. Further, in the model of insulin resistance described here, metformin enhanced the basal rate of glucose transport through a direct insulin-mimicking activity and/or a potentiation of the sensitivity of glucose transport to the antibody.
S Di Paolo, Via
Calefati,
7, 70121 Bari,
Italy
is an anti-dia¬ used and Canada for the in Europe widely management of non-insulin-dependent diabetes melli-
Patient and methods
tus.
Mono-[12,I]-(Tyr Ai4)-human
N'.N'-dimethyl-biguanide (metformin)
betic agent
Clinical studies demonstrated that it does not stimu¬ late insulin secretion; in fact, treatment with metformin may result in lowered insulin levels (1). Moreover, the biguanide has been reported to have little or no effect on intestinal glucose transport in humans (2, 3). Therefore, the major effects of metformin are to either potentiate or mimic the actions of insulin (2-6). Recent reports have demonstrated its effectiveness also in the treatment of insulin resistance caused by antiinsulin antibodies (7) and of Iipoatrophic diabetes (8). In the present investigation, I have further explored the mode of action of metformin in a patient with type B extreme insulin resistance. I also investigated the effect of metformin on anti-insulin receptor antibody-induced insulin resistance in isolated human adipocytes.
Materials insulin (HPLC purified, receptor grade) was from Sorin (Turin, Italy) (spec, activity: 2208 Ci/mmol). D-[U-14C]-glucose was from NEN (Boston, MA) (spec, activity: 258.5 mCi/mmol). Metformin was kindly given
by Lipha Farmaceutici (Florence, Italy). Protein ASepharose was purchased from Pharmacia Fine Chemi¬ cals, (Uppsala, Sweden). All other reagents were of the best grade commercially available and were from Sigma (St Louis, MO).
In vivo studies
58-year-old white woman with established systemic lupus erythematosus and extreme insulin resistance due A
118
S Di Paolo
to anti-insulin receptor antibodies was studied (9). At the time of study, the patient had intermittent arthral-
gia, leukopenia (2.9 x 109 cells/1), low levels of C5 and C4
fractions of complement, an increased erythrocyte sedi¬ mentation rate, high levels of antinuclear antibodies, anti-double-stranded DNA antibodies were positive and IgG was 41.4 g/1. BMI was 16 kg/m2, HbAic was 18.2% (normal range: 4.6-6.1 ). Anti-insulin antibodies were in the normal range, anti-insulin receptor antibody titer was 1.0 g/1. Regular human insulin had little effect on blood glucose level in doses up to 1000 IlJ/day. The patient, however, remained free of ketoacidosis even when insulin was completely omitted for several weeks. Insulin therapy was withheld at least 10 days before starting the study protocol. The patient took metformin (500 mg three times daily) for three periods of 10 days each, with 30 days of wash-out between each consecu¬ tive period. No other medication, besides metformin, was allowed throughout the study period.
ACTA F.NDOCRINOLOGICA 1992. 126
HEPES buffer, pH 7.4. plus 5% human albumin and 1 mmol/l glucose, in a final volume of 300 yl (10). Non¬ specific binding, determined in the presence of 10 /tmol/1 unlabelled insulin (3.8±1.3% of total binding), was subtracted to determine specific [12,I]-insuIin binding. Preincubation with
healthy donors
Adipocyte isolation After informed consent had been obtained, subcuta¬ neous adipose tissue was taken from the lower abdomi¬ nal wall of patients undergoing abdominal surgery. None of the patients, age 30-48 years, suffered from endocrine disorders, isolated human adipocytes were prepared by the method of Pedersen et al. (10), with minor modifications (11). Cell suspension was adjusted to a lipocrit of 5-10% (v/v). The cv of the lipocrit determination was 2.1%. The concentration of cells in adipocyte suspensions was calculated as adipocyte volume fraction divided by mean adipocyte volume, exactly as described (10). Adipose tissue from a single donor was used for each couple of experimental conditions in all experiments.
Insulin Insulin
binding studies
diluted 10-fold with Buffer G (50
mmol/l HEPES, 50 mmol/l TRIS, 10 mmol/l MgCl2, 10 mmol/l CaCl2, 50 mmol/l NaCl, 5 mmol/l KC1, 2 mmol/l EDTA) (12), supplemented with 57 ttmol/l phenyl methyl sulfonyl fluoride, 105 U/l penicillin, 100 mg/1 streptomycin, 1% human serum albumin, 1 mmol/l adenosine and 5.6 mmol/l glucose and adjusted to a pH of 7.5 (13). Immunoglobulin G (IgG), isolated from patient's serum by Protein A-Sepharose column chromatography (14), was added to the erythrocyte suspension at a final concentration of 1.0 g/1. Preincuba¬ tion took
f12,I]-insulin binding to circulating erythrocytes was shaking,
measured the day before starting oral therapy and, again, after 10 days of treatment. Insulin tolerance tests were performed on the days of insulin binding studies, 1 h after erythrocytes were taken out. At 09.00 h. after an overnight fast, the patient received 1 IU/kg iv of human insulin. Blood glucose concentration was determined immediately before and 1 5, 30 and 60 min after insulin injection. Anti-insulin receptor antibody titer did not vary throughout the study.
metformin. Erythrocytes from normal
were
place in a water bath at 37°C with very gentle in the presence
or
absence of metformin
(4 10-5 mol/l: effective serum concentrations (5)) and x
lasted up to 24 h. At the indicated times, the cells were sedimented at 400 x g at 4°C, washed three times with Buffer G, pH 8.0, supplemented with 5.6 mmol/l glucose and 0.1% human serum albumin, and resuspended in the same buffer at a concentration of 4.5 x 1012 cells/1. Thereafter, insulin receptor assay was performed. Isolated human adipocytes were resuspended (1:10 ratio) in Medium-199 buffered with 2 5 mmol/l HEPES (Sigma, St Louis MO), containing 5% human albumin, KP U/l penicillin, 100 mg/1 streptomycin and 1.0 g/1 IgG from the patient. Then cells were incubated at 3 7°C, pH 7.4 (5% CO2 95% air) in the presence or absence of 4 x 10~5 mol/l metformin. At the end of incubation, cells were washed twice in Medium-199, buffered as above and supplemented with 5% human albumin, then once in HEPES buffer, pH 7.4, and finally used for further studies. In some experiments, erythrocytes preincubated with patient's IgG were treated with an acidic buffer, pH 3.5, to dissociate immunoglobulins from cell binding prior to use in insulin binding assay (9).
Lipogenesis assay The procedure was a modification ( 15) of the method of Pedersen et al. (16). [14C]-radioactivity averaged 10 ±3% of basal lipo¬ genesis when incubations were performed in the absence of fat cells (blank values). The conversion rate of D-glucose to total lipids was measured at tracer glucose concentration ( 5 /¿mol/l) to yield an indirect estimation of glucose transport rates
binding to erythrocytes was determined as previously described (9). Non-specific binding, estimated by the addition of 10 /¿mol/l unlabelled insulin, averaged 11.1 ± 3.6% of total binding. (17). Freshly isolated adipocytes at 5% lipocrit (i.e. about Long-term effect of metformin on human fat cell 1 x 10s cells/1) were incubated with [125I]-insulin (final lipogenesis was studied by incubating adipocytes with or concentration: 30 pmol/1) in the absence and presence of without 4 x 10 "5 mol/l metformin for up to 16 h at unlabelled insulin (0-17 nmol/1) for 1 h at 37°C in 37°C, in Medium-199 containing 0.1 g/1 IgG from =
Metformin and insulin receptor antibodies
ACTA KND0CR1N0I.0GICA 1992. 12ft
119
patient's serum. At the end of preincubation, cells were washed twice in Medium-199, once in a glucose-free HEPES buffer, pH 7.4, and then incubated with or without different amounts of insulin for 60 min, before the addition of 0.4 yCi of D-[U-14C]-glucose.
Statistics All data was
by
are presented as mean±SD. Statistical analysis performed using Student's f-test for paired data and
ANOVA 2-way test. Values with considered to be significant.
p
Di Paolo S. Metformin ameliorates extreme insulin resistance in a patient with anti-insulin receptor antibodies: description of insulin receptor and postreceptor effect in vivo and in vitro. Acta Endocrinol 1992:126:117-23. ISSN 0001-5598
The effect of metformin on insulin binding and insulin action in the presence of anti-insulin receptor antibodies was investigated in a case of type B extreme insulin resistance. Oral administration of metformin (1 500 mg/d) for 10 days significantly decreased plasma blood glucose and insulin levels and enhanced the hypoglycemic response to exogenous insulin. In vitro preincubation of normal erythrocytes with insulin receptor antibody from the patient plus 4\m=x\ 10-5 mol/l metformin markedly enhanced insulin binding to receptors, compared to cells incubated with antibody alone. This effect was apparent after 2 h, was maximal after 4 h and did not change up to 24 h. Closely similar results were found when human adipocytes were studied. Analysis of binding data confirmed the increase in both receptor number and affinity. One hour exposure of control adipocytes to metformin enhanced basal lipogenesis by more than 30%. Acute exposure of fat cells to the patient's receptor antibodies resulted in a stimulation of glucose transport and a state of severe insulin resistance. The addition of metformin to antibody in preincubation buffer strongly enhanced basal glucose incorporation into lipids, but did not prevent insulin unresponsiveness. It is suggested that metformin increases, possibly through a change in the spatial conformation of insulin receptor within the plasma membrane, the availability of preexisting receptors to insulin binding and/or decreases the availability of specific epitopes to antibody anchoring. Further, in the model of insulin resistance described here, metformin enhanced the basal rate of glucose transport through a direct insulin-mimicking activity and/or a potentiation of the sensitivity of glucose transport to the antibody.
S Di Paolo, Via
Calefati,
7, 70121 Bari,
Italy
is an anti-dia¬ used and Canada for the in Europe widely management of non-insulin-dependent diabetes melli-
Patient and methods
tus.
Mono-[12,I]-(Tyr Ai4)-human
N'.N'-dimethyl-biguanide (metformin)
betic agent
Clinical studies demonstrated that it does not stimu¬ late insulin secretion; in fact, treatment with metformin may result in lowered insulin levels (1). Moreover, the biguanide has been reported to have little or no effect on intestinal glucose transport in humans (2, 3). Therefore, the major effects of metformin are to either potentiate or mimic the actions of insulin (2-6). Recent reports have demonstrated its effectiveness also in the treatment of insulin resistance caused by antiinsulin antibodies (7) and of Iipoatrophic diabetes (8). In the present investigation, I have further explored the mode of action of metformin in a patient with type B extreme insulin resistance. I also investigated the effect of metformin on anti-insulin receptor antibody-induced insulin resistance in isolated human adipocytes.
Materials insulin (HPLC purified, receptor grade) was from Sorin (Turin, Italy) (spec, activity: 2208 Ci/mmol). D-[U-14C]-glucose was from NEN (Boston, MA) (spec, activity: 258.5 mCi/mmol). Metformin was kindly given
by Lipha Farmaceutici (Florence, Italy). Protein ASepharose was purchased from Pharmacia Fine Chemi¬ cals, (Uppsala, Sweden). All other reagents were of the best grade commercially available and were from Sigma (St Louis, MO).
In vivo studies
58-year-old white woman with established systemic lupus erythematosus and extreme insulin resistance due A
118
S Di Paolo
to anti-insulin receptor antibodies was studied (9). At the time of study, the patient had intermittent arthral-
gia, leukopenia (2.9 x 109 cells/1), low levels of C5 and C4
fractions of complement, an increased erythrocyte sedi¬ mentation rate, high levels of antinuclear antibodies, anti-double-stranded DNA antibodies were positive and IgG was 41.4 g/1. BMI was 16 kg/m2, HbAic was 18.2% (normal range: 4.6-6.1 ). Anti-insulin antibodies were in the normal range, anti-insulin receptor antibody titer was 1.0 g/1. Regular human insulin had little effect on blood glucose level in doses up to 1000 IlJ/day. The patient, however, remained free of ketoacidosis even when insulin was completely omitted for several weeks. Insulin therapy was withheld at least 10 days before starting the study protocol. The patient took metformin (500 mg three times daily) for three periods of 10 days each, with 30 days of wash-out between each consecu¬ tive period. No other medication, besides metformin, was allowed throughout the study period.
ACTA F.NDOCRINOLOGICA 1992. 126
HEPES buffer, pH 7.4. plus 5% human albumin and 1 mmol/l glucose, in a final volume of 300 yl (10). Non¬ specific binding, determined in the presence of 10 /tmol/1 unlabelled insulin (3.8±1.3% of total binding), was subtracted to determine specific [12,I]-insuIin binding. Preincubation with
healthy donors
Adipocyte isolation After informed consent had been obtained, subcuta¬ neous adipose tissue was taken from the lower abdomi¬ nal wall of patients undergoing abdominal surgery. None of the patients, age 30-48 years, suffered from endocrine disorders, isolated human adipocytes were prepared by the method of Pedersen et al. (10), with minor modifications (11). Cell suspension was adjusted to a lipocrit of 5-10% (v/v). The cv of the lipocrit determination was 2.1%. The concentration of cells in adipocyte suspensions was calculated as adipocyte volume fraction divided by mean adipocyte volume, exactly as described (10). Adipose tissue from a single donor was used for each couple of experimental conditions in all experiments.
Insulin Insulin
binding studies
diluted 10-fold with Buffer G (50
mmol/l HEPES, 50 mmol/l TRIS, 10 mmol/l MgCl2, 10 mmol/l CaCl2, 50 mmol/l NaCl, 5 mmol/l KC1, 2 mmol/l EDTA) (12), supplemented with 57 ttmol/l phenyl methyl sulfonyl fluoride, 105 U/l penicillin, 100 mg/1 streptomycin, 1% human serum albumin, 1 mmol/l adenosine and 5.6 mmol/l glucose and adjusted to a pH of 7.5 (13). Immunoglobulin G (IgG), isolated from patient's serum by Protein A-Sepharose column chromatography (14), was added to the erythrocyte suspension at a final concentration of 1.0 g/1. Preincuba¬ tion took
f12,I]-insulin binding to circulating erythrocytes was shaking,
measured the day before starting oral therapy and, again, after 10 days of treatment. Insulin tolerance tests were performed on the days of insulin binding studies, 1 h after erythrocytes were taken out. At 09.00 h. after an overnight fast, the patient received 1 IU/kg iv of human insulin. Blood glucose concentration was determined immediately before and 1 5, 30 and 60 min after insulin injection. Anti-insulin receptor antibody titer did not vary throughout the study.
metformin. Erythrocytes from normal
were
place in a water bath at 37°C with very gentle in the presence
or
absence of metformin
(4 10-5 mol/l: effective serum concentrations (5)) and x
lasted up to 24 h. At the indicated times, the cells were sedimented at 400 x g at 4°C, washed three times with Buffer G, pH 8.0, supplemented with 5.6 mmol/l glucose and 0.1% human serum albumin, and resuspended in the same buffer at a concentration of 4.5 x 1012 cells/1. Thereafter, insulin receptor assay was performed. Isolated human adipocytes were resuspended (1:10 ratio) in Medium-199 buffered with 2 5 mmol/l HEPES (Sigma, St Louis MO), containing 5% human albumin, KP U/l penicillin, 100 mg/1 streptomycin and 1.0 g/1 IgG from the patient. Then cells were incubated at 3 7°C, pH 7.4 (5% CO2 95% air) in the presence or absence of 4 x 10~5 mol/l metformin. At the end of incubation, cells were washed twice in Medium-199, buffered as above and supplemented with 5% human albumin, then once in HEPES buffer, pH 7.4, and finally used for further studies. In some experiments, erythrocytes preincubated with patient's IgG were treated with an acidic buffer, pH 3.5, to dissociate immunoglobulins from cell binding prior to use in insulin binding assay (9).
Lipogenesis assay The procedure was a modification ( 15) of the method of Pedersen et al. (16). [14C]-radioactivity averaged 10 ±3% of basal lipo¬ genesis when incubations were performed in the absence of fat cells (blank values). The conversion rate of D-glucose to total lipids was measured at tracer glucose concentration ( 5 /¿mol/l) to yield an indirect estimation of glucose transport rates
binding to erythrocytes was determined as previously described (9). Non-specific binding, estimated by the addition of 10 /¿mol/l unlabelled insulin, averaged 11.1 ± 3.6% of total binding. (17). Freshly isolated adipocytes at 5% lipocrit (i.e. about Long-term effect of metformin on human fat cell 1 x 10s cells/1) were incubated with [125I]-insulin (final lipogenesis was studied by incubating adipocytes with or concentration: 30 pmol/1) in the absence and presence of without 4 x 10 "5 mol/l metformin for up to 16 h at unlabelled insulin (0-17 nmol/1) for 1 h at 37°C in 37°C, in Medium-199 containing 0.1 g/1 IgG from =
Metformin and insulin receptor antibodies
ACTA KND0CR1N0I.0GICA 1992. 12ft
119
patient's serum. At the end of preincubation, cells were washed twice in Medium-199, once in a glucose-free HEPES buffer, pH 7.4, and then incubated with or without different amounts of insulin for 60 min, before the addition of 0.4 yCi of D-[U-14C]-glucose.
Statistics All data was
by
are presented as mean±SD. Statistical analysis performed using Student's f-test for paired data and
ANOVA 2-way test. Values with considered to be significant.
p
