Modulation of the rat hepatic cytochrome P4501A subfamily using biotin supplementation.

Studies have found that biotin favors glucose and lipid metabolism, and medications containing biotin have been developed. Despite the use of biotin as a pharmacological agent, few studies have addressed toxicity aspects including the possible interaction with cytochrome P450 enzyme family. This study analyzed the effects of pharmacological doses of biotin on the expression and activity of the cytochrome P4501A subfamily involved in the metabolism of xenobiotics. Wistar rats were treated daily with biotin (2 mg/kg, i.p.), while the control groups were treated with saline. All of the rats were sacrificed by cervical dislocation after 1, 3, 5, or 7 days of treatment. CYP1A1 and CYP1A2 mRNAs were modified by biotin while enzyme activity and protein concentration were not affected. The lack of an effect of biotin on CYP1A activity was confirmed using other experimental strategies, including (i) cotreatment of the animals with biotin and a known CYP1A inducer; (ii) the addition of biotin to the reaction mixtures for the measurement of CYP1A1 and CYP1A2 activities; and (iii) the use of an S9 mixture that was prepared from control and biotin-treated rats to analyze the activation of benzo[a]pyrene (BaP) into mutagenic metabolites using the Ames test. The results suggest that biotin does not influence the CYP1A-mediated metabolism of xenobiotics.

Effects of biotin supplementation in the diet on insulin secretion, islet gene expression, glucose homeostasis and beta-cell proportion.

Besides its role as a carboxylase cofactor, biotin has a wide repertoire of effects on gene expression, development and metabolism. Pharmacological concentrations of biotin enhance insulin secretion and the expression of genes and signaling pathways that favor islet function in vitro. However, the in vivo effects of biotin supplementation on pancreatic islet function are largely unknown. In the present study, we investigated whether in vivo biotin supplementation in the diet has positive effects in rodent pancreatic islets. Male BALB/cAnN Hsd mice were fed a control or a biotin-supplemented diet over 8 weeks postweaning and tested for glucose homeostasis, insulin secretion, islet gene expression and pancreatic morphometry. Insulin secretion increased from the islets of biotin-supplemented mice, together with the messenger RNA (mRNA) expression of several transcription factors regulating insulin expression and secretion, including forkhead box A2, pancreatic and duodenal homeobox 1 and hepatocyte nuclear factor 4α. The mRNA abundance of glucokinase, Cacna1d, acetyl-CoA carboxylase, and insulin also increased. Consistent with these effects, glucose tolerance improved, and glucose-stimulated serum insulin levels increased in biotin-supplemented mice, without changes in fasting glucose levels or insulin tolerance. Biotin supplementation augmented the proportion of beta cells by enlarging islet size and, unexpectedly, also increased the percentage of islets with alpha cells at the islet core. mRNA expression of neural cell adhesion molecule 1, an adhesion protein participating in the maintenance of islet architecture, decreased in biotin-supplemented islets. These findings provide, for the first time, insight into how biotin supplementation exerts its effects on function and proportion of beta cells, suggesting a role for biotin in the prevention and treatment of diabetes.

Effect of retinoic acid on glucokinase activity and gene expression in neonatal and adult cultured hepatocytes.

It has been shown that all-trans retinoic acid induces prematurely hepatic glucokinase mRNA in ten days-old neonatal rat hepatocytes, however, this effect could be related to the capacity of the retinoid to promote a more differentiated state of the hepatocyte. In this report we demonstrate that physiological concentrations of all-trans retinoic acid stimulate glucokinase activity in both mature fully differentiated hepatocytes and at the onset of the induction of the enzyme in 15 to 17 days-old neonatal hepatocytes. The effects produced by the retinoid were similar both in magnitude and in time, to those elicited by insulin, a well-known stimulator of hepatic glucokinase expression. No additive effect was observed when insulin and retinoic acid were tested together. Using the branched DNA assay, a sensitive signal amplification technique, we detected relative increases in glucokinase mRNA levels of about 70% at 3 and 24 h after the treatment with 10(-6) M all-trans retinoic acid, in both neonatal and adult hepatocytes. These data show that retinoic acid exerts a stimulatory effect on hepatic glucokinase independent of the hepatocyte stage of maturity and suggest a physiological role of retinoic acid on glucose metabolism. The action of retinoic acid on hepatic glucokinase might explain previous observations on the relationship between vitamin A status and liver glycogen synthesis. These findings may serve as basis for further investigations on the biological functions of retinoic acid derivatives on hepatic glucose metabolism.

Testosterone effect on insulin content, messenger ribonucleic acid levels, promoter activity, and secretion in the rat.

Coexistence of hyperinsulinemia and hyperandrogenism in women has been frequently described. Most of the studies addressing this issue have focused on the mechanisms by which insulin produces hyperandrogenism. In the present study, we analyzed the effects of testosterone in vivo and in vitro upon insulin gene expression and release in the rat. Our studies demonstrate that testosterone increases insulin messenger RNA (mRNA) levels in vitro as well as in vivo. In both prepuberal and intact adult rats, serum testosterone concentrations were positively correlated with insulin mRNA levels and insulin concentration in serum. Testosterone deprivation after gonadectomy decreased both insulin gene expression and serum insulin concentration. Insulin mRNA levels were partially restored after 3 days of testosterone administration and serum insulin was 80% and 27% above baseline values at 5 and 7 days posttreatment. Primary cultured pancreatic islets treated with the sexual steroid increased about 80% insulin mRNA, as well as protein, and release. In transfected islets, testosterone increased the activity of the -410 bp rat insulin promoter I by 154%. These data demonstrate that testosterone has a direct effect upon pancreatic islet function by favoring insulin gene expression and release.

Cyclic adenosine 3',5'-monophosphate increases pancreatic glucokinase activity and gene expression.

Comparison of the pancreatic and hepatic glucokinase gene transcripts reveals tissue-specific control of expression and the existence of two distinct promoters in a single glucokinase gene. The existence of alternate promoters suggests that separate factors regulate glucokinase transcription in the two tissues. Hepatic glucokinase expression has been shown to be repressed by cAMP; however, in the pancreatic beta-cell it is unlikely that cAMP represses glucokinase activity, as cAMP is known to positively affect glucose-induced insulin secretion, a process that in mature islets requires pancreatic glucokinase activity. In this work we demonstrate that cAMP indeed has a stimulatory effect on pancreatic glucokinase. The cyclic nucleotide stimulates pancreatic glucokinase activity after 3-h incubation, and maximal effects are observed after 6 and 12 h of treatment. Using the bDNA assay, a sensitive signal amplification technique, we detected relative increases in glucokinase messenger RNA levels of 40.5 +/- 7.5% after 3-h incubation with cAMP. This stimulatory effect was increased to 106.3 +/- 22% after 6-h incubation and sustained up to 12 h of incubation. Inhibition of gene transcription by actinomycin D abolishes cAMP-induced glucokinase activity. In transfected fetal islets, cAMP increased the activity of the -1000 bp rat glucokinase promoter by 60 +/- 6%. These data demonstrate that cAMP has a stimulatory effect on pancreatic glucokinase gene expression and that the nucleotide has opposite effects on pancreatic and hepatic glucokinase, supporting the concept that glucokinase transcription in the liver and that in the beta-cell differ.

Biotin regulation of pancreatic glucokinase and insulin in primary cultured rat islets and in biotin-deficient rats.

Biotin has been reported to affect glucose homeostasis; however, its role on pancreatic islets of Langerhans has not been assessed. In this report, we demonstrate that physiologic concentrations of biotin stimulate glucokinase activity in rat islets in culture. Using the branched DNA (bDNA) assay, a sensitive signal amplification technique, we detected relative increases in glucokinase mRNA levels of 41.5 +/- 13% and 81.3 +/- 19% at 12 and 24 h respectively in islets treated with [10(-6) M] biotin. Because glucokinase activity controls insulin secretion, we also investigated the effect of biotin on insulin release. Treatment with [10(-6) M] biotin for 24 h increased insulin secretion. We extended our studies by analyzing the effect of biotin deficiency on pancreatic islet glucokinase expression and activity, as well as insulin secretion. Our results show that islet glucokinase activity and mRNA are reduced by 50% in the biotin deficient rat. Insulin secretion in response to glucose was also impaired in islets isolated from the deficient rat. These data show that biotin affects pancreatic islet glucokinase activity and expression and insulin secretion in cultured islets.

Effect of retinoic acid on glucokinase activity and gene expression and on insulin secretion in primary cultures of pancreatic islets.

Retinoic acid has manifold effects on pancreatic beta-cells. Previously we reported that retinoic acid increases glucokinase activity and messenger RNA (mRNA) levels in the insulinoma cell line RIN-m5F; however, we could not rule out the possibility that the effect of retinoic acid on RIN-m5F glucokinase was inherent to the cell line or related to its differentiating capacity. In this report, we demonstrate that physiologic concentrations of retinoic acid stimulate glucokinase activity in both fetal islets and differentiated adult islets in culture. In the adult tissue, the response to the retinoid was less pronounced, achieving about half of the maximal effect produced on the fetal tissue. Using the branched DNA (bDNA) assay, a sensitive signal amplification technique, we detected relative increases in glucokinase mRNA levels of 51.8+/-13.3% and 62.8+/-16.1% at 12 and 24 h, respectively, in adult islets treated with] 10(-6) M retinoic acid. In fetal islets, increases of 55+/-14.9% and 107+/-30.5% at 12 and 24 h, respectively, were observed. In transfected fetal islets, retinoic acid increased the activity of the -1000 kb rat glucokinase promoter by 51.3%. Because glucokinase activity controls insulin secretion, we also investigated the effect of retinoic acid on insulin secretion. Treatment with 10(-6) M retinoic acid for 24 h increased insulin secretion in both fetal and adult islets; however, the increases on insulin secretion were more pronounced in the mature islets; in contrast, retinoic acid produced higher levels of insulin mRNA in the fetal islets. These data show that retinoic acid increases pancreatic glucokinase in cultured islets and that the mechanism may involve a stimulatory effect on the glucokinase promoter.

The human insulin gene contains multiple transcriptional elements that respond to glucocorticoids.

Glucocorticoids inhibit insulin expression in cultured pancreatic islet cells. In this study, we provide evidence that transcriptional downregulation of insulin gene expression by glucocorticoids is the result of synergistic interaction between various elements of the insulin promoter. Similar synergistic effects on insulin gene transcription were previously reported for other key insulin regulators, cyclic adenosine monophosphate (cAMP) and glucose. Transfection of CAT constructs containing different segments of the insulin promoter into the pancreatic cell line, HIT T-15 2.2.2, demonstrated that dexamethasone decreased CAT activity in all constructs tested. However, differences were found in the relative sensitivities of the various constructs. Glucocorticoid inhibition of expression from plasmids containing A elements may result from decreased expression of the pancreatic homeodomain protein STF-1. However, a different mechanism must be invoked for insulin promoter constructs lacking A sites, which nevertheless still demonstrated negative regulation. Glucocorticoid-induced inhibition of one of these regions (-882 to -342) was seen to require pancreas-specific factors, because inhibition was observed in HIT T-15 2.2.2 cells but not in the nonpancreatic COS-1 cells. We conclude, therefore, that the human insulin gene contains multiple transcriptional elements that respond to glucocorticoids, some of which require beta cell-specific factors.

Identification of the human insulin negative regulatory element as a negative glucocorticoid response element.

Insulin gene transcription in adults is restricted to pancreatic beta cells. Studies with both transgenic mice and islet cell lines have demonstrated that beta cell specific expression is conferred by the 5' flanking region of the insulin gene. Transfection analysis has shown that cell specific expression involved an interaction between both positive and negative promoter cis elements. An upstream region (between -258 and -279) of the human insulin promoter served as a site of negative regulation. Transfection analysis in the pancreatic cell line HIT T-15 M 2.2.2 revealed that a DNA fragment containing this region causes a 45% reduction in promoter activity when linked to the native insulin promoter and a 72% reduction when linked to a heterologous tk promoter. Electrophoretic mobility shift analysis of this negative regulatory region (NRE) reveals a complex pattern of binding, wherein two major and several minor complexes are observed. Competition experiments demonstrated that formation of the fastest mobility complex is completely inhibited with excess cold glucocorticoid responsive element (GRE) consensus oligonucleotide. Purified glucocorticoid receptor binding domain (T7X556) demonstrated binding to the NRE oligonucleotide. Functional studies showed that dexamethasone treatment of HIT T-15 M 2.2.2 cells containing an NRE-tk CAT plasmid decreased CAT gene expression by 48%. Analysis of the NRE revealed 73% homology with the negative GRE consensus sequence. These data show that the human insulin NRE is a negative glucocorticoid response element.

The hexosamine biosynthetic pathway and glucose-induced down regulation of glucose transport in L6 myotubes.

Based on experiments in cultured adipocytes, it has been proposed that glucose-induced down regulation of glucose transport is mediated by the conversion of fructose-6-phosphate to glucosamine-6-phosphate via the first and rate-determining enzyme of the hexosamine biosynthetic pathway, glutamine: fructose-6-phosphate amidotransferase (glutamine hexosephosphate aminotransferase). Evidence for this assertion was: (a) L-glutamine, the provider group for the aminotransferase was essential; (b) two inhibitors of glutamine hexosephosphate aminotransferase, 6-diazo-5-oxonorleucine (L form) and azaserine, blocked glucose-induced down regulation of glucose transport; (c) azaserine inhibited the activity of the aminotransferase, (d) glucosamine, which enters the hexosamine pathway distal to this enzyme was 40-times more potent than glucose; and (e) azaserine was unable to block the effect of glucosamine. Since muscle is quantitatively much more important than adipose tissue for whole body glucose utilization, we sought to determine if the hexosamine pathway was involved in glucose-induced down regulation of glucose transport in L6 myotubes. Glucose was effective, both in the presence and absence of glutamine in the incubation media. Glucosamine was also effective but was as equipotent as glucose. Small amounts of glutamine hexosephosphate aminotransferase were present in the L6 myotubes and although the leucine derivative (20 microM) inhibited the enzyme, it did not impair glucose-induced down regulation of glucose transport. Total GLUT-1 levels were similar when the cells were incubated in the absence or presence of 5 mM glucose or glucosamine although glucosamine was associated with a marked increase in a lower molecular weight band. These results do not suggest that the hexosamine biosynthetic pathway is involved in glucose-induced down regulation of glucose transport in L6 myotubes. Thus, this phenomenon is regulated differently in muscle and fat.

Effect of sodium butyrate on glucose transport and glucose-phosphorylating enzymes in RIN-m5F cells.

Sodium butyrate is widely used to differentiate insulinoma cell lines. However, sodium has been shown to decrease glucose phosphorylation in the liver and heart and decrease the expression of glucose transporter. Since these mechanisms are essential for glucose-induced insulin secretion, the ultimate function of the pancreatic beta-cell, we investigated the effect of sodium butyrate on both glucose-phosphorylating enzymes as well as glucose transport in the pancreatic cell line RIN-m5F. Treatment of RIN-m5F cells with 2.5 mM sodium butyrate for 72 h increased by twofold both hexokinase and glucokinase (GK) activities, as well as the gene expression of GK. Sodium butyrate treatment had no effect on GLUT-1 mRNA levels but increased the GLUT-2 mRNA 3.7-fold. Kinetic analysis of 2-deoxyglucose transport displayed a single curve with Km = 1.2 mM and Vmax = 10.9 pmol/micrograms protein/min in the untreated cells, values similar to the low Km glucose transport reported in the pancreatic beta-cells. This low Km transport component markedly decreased with sodium butyrate treatment, and interestingly a second component with a higher Km appeared, consistent with the increase in GLUT-2 mRNA. We conclude that the differentiating action of sodium butyrate involves increases in GK and GLUT-2 gene expression, which characterizes the differentiated state of the pancreatic beta-cell. However, the inhibitory effect of sodium butyrate on low Km glucose transport needs to be considered in the use of this compound to promote differentiation.

Analysis of purine metabolic enzymes in human CD4 Leu 8- and CD4 Leu 8+ lymphocyte subpopulations.

1. Specific activities of adenosine deaminase, purine nucleoside phosphorylase, adenosine kinase, 5'-nucleotidase, S-adenosyl-L-homocysteine hydrolase, AMP deaminase, adenine phosphoribosyl transferase, and hypoxanthine phosphoribosyl transferase were analyzed in human CD4 T-lymphocyte subsets. 2. CD4 Leu 8- (helper/inducer) and CD4 Leu 8+ (suppressor/inducer) subpopulations were obtained by panning or fluorescence activated cell sorting techniques using specific monoclonal antibodies. 3. A 45% decrease of 5'-NT AMP activity in the CD4 Leu 8- cells (suppressor/inducer) compared with CD4 total cell population. 4. No statistical significant differences in enzyme activity were found between the subsets analyzed in other purine enzymes. 5. These results suggest that the distribution of purine metabolic enzymes is homogeneous in CD4 Leu 8- and CD4 Leu 8+ T-lymphocyte subpopulations.

Regulation of glucokinase and proinsulin gene expression and insulin secretion in RIN-m5F cells by dexamethasone, retinoic acid, and thyroid hormone.

In RIN-5mF cells, treatment with either dexamethasone or retinoic acid resulted in increased glucokinase activity. Treatment with dexamethasone (50 and 500 nM) increased glucokinase activity 140% and 260%, respectively, whereas treatment with retinoic acid (100 and 1000 nM) increased glucokinase activity by 50% and 80%, respectively. An additive effect was observed when 1000 nM retinoic acid was added to either 50 or 500 nM dexamethasone. Treatment with either of these factors resulted in increased glucokinase mRNA levels. In contrast, treatment with thyroid hormone (T3; 0.2 and 10 nM) had no effect on glucokinase activity, but decreased glucokinase mRNA levels. T3 did not affect the response of either activity or mRNA levels to retinoic acid. Binding experiments revealed the presence of a single class of T3 nuclear receptors, with a Kd of 0.46 x 10(-10) and a binding capacity of 3.84 pmol/mg protein in these cells. This suggests that the changes produced by T3 may be due to a physiological effect mediated via nuclear receptors. Proinsulin mRNA levels were increased by retinoic acid (10-100 nM) and dexamethasone (5-500 nM), but decreased by T3 (0.04-10 nM). Insulin secretion in the absence or presence of glucose and potassium was increased about 50% in cells previously treated with 1000 nM retinoic acid. T3 (0.2 nM) did not affect insulin secretion in the absence or presence of glucose or in the presence of secretagogues (potassium and glyceraldehyde). Dexamethasone (50 nM) did not significantly increase insulin secretion in the absence or presence of glucose or potassium. In contrast, dexamethasone decreased glyceraldehyde-induced insulin secretion by 60% (P less than 0.05). We conclude that the glucokinase and proinsulin genes respond in parallel to dexamethasone and retinoic acid (both increased) and to T3 (both decreased).

Pituitary insulin-like growth factor-I gene expression: regulation by triiodothyronine and growth hormone.

Most of the growth-promoting effects of GH are mediated through insulin-like growth factor-I (IGF-I). Pituitary GH gene expression is, in turn, inhibited by IGF-I. Since rat pituitary tissue and GH3 pituitary tumor cells express both the GH and the IGF-I genes, we have attempted to clarify their potential interactions in the somatotroph by examining hormonal factors involved in the regulation of pituitary IGF-I gene expression. IGF-I mRNA was measured in GH3 cells by a solution hybridization/RNase protection assay, using riboprobes to differentially protect the IGF-I variant mRNAs arising by alternative splicing at both the 5' untranslated (UT) and 3' ends of the primary transcript. GH3 cells contained both class A and class C 5' UT variant mRNAs, with a relative abundance similar to that found in the liver. Sixty-five percent of the total IGF-I mRNA in GH3 cells was processed at the 3' end to IGF-Ia, and 35% to IGF-Ib mRNAs, whereas in the liver the proportions were 85% and 15%, respectively. GH3 cells grown in thyroid hormone-depleted medium for 4 days contained low levels of IGF-I mRNA. T3 and human (h) GH induced total IGF-I mRNA content in thyroid hormone-depleted cells, with both 5' and 3' alternative transcripts regulated coordinately, an effect that was maximal at 48-72 h. T3 stimulation of GH3 IGF-I mRNA over 48 h was dose dependent (0.01-5 nM). Similarly, hGH (0.5-10 micrograms/ml) evoked a dose-dependent induction of IGF-I mRNA in the thyroid hormone-deficient GH3 cells. The effects of T3 (5 nM) and hGH (10 micrograms/ml) on IGF-I mRNA were not additive. Furthermore, the effects of both T3 and hGH were selective for IGF-I mRNA, as neither of these treatments stimulated PRL mRNA, and treatment with hGH decreased GH3 cell GH mRNA content. This model does not discriminate whether T3 has an independent effect on IGF-I gene expression or if its action is mediated solely through induction of GH. In conclusion, IGF-I mRNA transcripts are present in GH3 cells and are modulated by T3 and GH. Local paracrine or autocrine interactions may, therefore, be involved in the feedback control of GH secretion.

Retinoic acid selectively stimulates growth hormone secretion and messenger ribonucleic acid levels in rat pituitary cells.

Hydrocortisone (HCT) is known to stimulate rat GH gene expression. As the retinoic acid receptor cDNA bears a 15% homology to the hydrocortisone receptor cDNA, the effects of retinoic acid were tested in GH3 rat pituitary cells. Retinoic acid selectively stimulated GH secretion after 24 h. Maximal induction of basal GH secretion (220% of controls) was achieved with 1 microM retinoic acid after 72 h. HCT (50 nM) alone stimulated GH secretion 3.5-fold during 72-h incubation, and retinoic acid (1 microM) with HCT caused a further 7-fold increase in GH secretion. The stimulating effects of retinoic acid on GH were observed when cells were incubated in defined serum-free medium, 4% fetal calf serum, or Dulbecco's Modified Eagle's Medium containing HCT. Northern gel hybridization showed that steady state levels of GH mRNA transcripts in these cells were also selectively induced by retinoic acid. PRL secretion and mRNA levels, however, were not appreciably altered by retinoic acid. The results show that retinoic acid selectively stimulates basal and HCT-induced GH secretion and mRNA levels in these cells in a dose- and time-dependent manner. This selective action of retinoic acid appears to occur independently of the presence of serum in the culture medium.

Adenosine deaminase deficiency with late onset of recurrent infections: response to treatment with polyethylene glycol-modified adenosine deaminase.

We report a 5-year-old girl with adenosine deaminase (ADA) deficiency who was asymptomatic during the first years of life. At 3 years of age, she developed chronic and recurrent sinopulmonary infections, and at 4 1/2 years of age she had one major infection with Streptococcus pneumoniae (bacteremia and septic arthritis of the hip). Immunologic evaluation at 5 years of age revealed persistent lymphopenia, decreased helper-suppressor T cell ratios, and low proliferative responses to mitogens. The IgG, IgM, and IgA levels were normal; the IgG2 level was low normal or below normal. The patient had specific antibodies against toxoids and viral antigens but failed to produce antibodies against Haemophilus influenzae type b and pneumococcal polysaccharides. Although no symptoms of allergy were present, she had persistent eosinophilia and elevated IgE levels. The patient had 0.6% of normal ADA activity in erythrocytes and approximately 1% of normal ADA activity in peripheral blood mononuclear cells. Beginning at 6 years of age, she was treated with weekly injections of polyethylene glycol-modified bovine ADA. This treatment was well tolerated and effectively reversed the biochemical consequence of ADA deficiency. Concomitantly, she improved clinically and her T lymphocyte numbers and blastogenic responses to mitogens in vitro became normal. The late onset of clinical symptoms and relatively benign clinical course in this patient emphasize the need to consider ADA deficiency in a broad spectrum of immunodeficient children.

Adenosine-resistant Chinese hamster fibroblast variants with hyperactive adenosine-deaminase: an analysis of the protection against exogenous adenosine afforded by increased activity of the deamination pathway.

The activity of purine salvage and interconversion enzymes was examined in two sublines of Chinese hamster cells--RA11 and RA41--isolated on the basis of their resistance to adenosine concentrations toxic to wild-type CCL39 cells. Adenosine deaminase (ADA) activity was found to be two times higher in RA11 and three times higher in RA41 than in CCL39. Inhibition of ADA activity by coformycin reduced the level of adenosine resistance but did not restore wild-type sensitivity, indicating that a second defect contributes to the adenosine-resistant phenotype of these variants; evidence was indeed obtained for the presence in both lines of additional alterations protecting them against the lethal depletion of phosphoribosylpyrophosphate (Ishii and Green, 1973) imposed by adenosine to wild-type cells. To gain better insight into the influence of ADA hyperactivity on adenosine resistance, a procedure was developed for the specific isolation of variants with increased levels of ADA activity. Cell lines with 3-5 times and then 100-500 times the wild-type ADA activity were stepwise recovered. These investigations confirmed that amplification of ADA can efficiently contribute in protecting cells against high concentrations of exogenous adenosine. The variants isolated by this procedure again manifested, in addition to amplification of ADA activity, another alteration decreasing their sensitivity to adenosine. A possible mechanism accounting for the frequent isolation of variants that coexpress ADA hyperactivity and a second defect contributing protection against adenosine toxicity are considered.