Function of murine adenosine deaminase in the gastrointestinal tract

Adenosine deaminase (ADA) deficiency in humans leads to a combined immunodeficiency characterized by severe T and B cell lymphopenia. ADA-deficient humans also display defective development of gut-associated lymphoid tissues (GALT). They lack lymphoid cells, and the Peyer's patches are without germinal centers. In mice, ADA-deficient fetuses die perinatally due to liver damage, but they also exhibit pathology in the thymus, spleen, and the small intestine. The GI phenotype associated with ADA-deficient humans prompted us to examine the effect of ADA-deficiency on mouse small intestine tissue. The work presented here focuses on understanding the physiological role of ADA in the GI tract, using ADA-deficient mice rescued from perinatal lethality by restoring Ada expression to trophoblast cells. Histologically and immunologically, the GALT was compromised at all sites in ADA-/- mice, with the most dramatic changes seen in the Peyer's patches. Profound disturbances in purine metabolism were detected in all the gastrointestinal tissues. In particular, adenosine and deoxyadenosine, the ADA substrates, increased markedly while the product inosine decreased. The activity of S-adenosylhomocysteine hydrolase decreased throughout the GI tract, indicating a possible disruption of cellular transmethylation and activation of apoptotic pathways. There were also disturbances in the purine metabolic pathway with a decrease in the production of downstream nucleosides hypoxanthine and xanthine.     

Transient expression of human adenosine deaminase cDNAs: identification of a nonfunctional clone resulting from a single amino acid substitution.

Human adenosine deaminase (ADA) is an important purine catabolic enzyme which irreversibly deaminates adenosine and deoxyadenosine. Severe genetic deficiency of ADA leads to an immunological deficiency state in which T-lymphoid cells are selectively destroyed by the accumulation of toxic levels of deoxyadenosine and deoxy-ATP. In preparation for transfer of ADA sequences into a variety of cell types, we explored expression of ADA cDNAs transfected into cultured cells within a simian virus 40-based expression vector. After transfection into monkey kidney (COS) cells, ADA cDNA encompassing the entire coding region of the protein generated human ADA activity. An unexpected finding, however, was the identification of a cDNA clone that failed to produce either human enzyme activity or immunoreactive ADA protein. As this pattern is typical of many naturally occurring mutant ADA alleles, we characterized the molecular defect in this clone. DNA sequence analysis revealed a single nucleotide substitution in amino acid position 50 (glycine-valine). Northern blotting with a unique 17-mer oligonucleotide demonstrated the absence of the mutant sequence in the mRNA from which the cDNA library giving rise to the mutant cDNA was constructed. Therefore, the substitution in the variant cDNA was created during cloning. These data define one critical region of the human ADA protein molecule and suggest a convenient strategy for characterization of the phenotypes associated with naturally occurring mutant alleles.     

Deoxyadenosine metabolism and cytotoxicity in cultured mouse T lymphoma cells: a model for immunodeficiency disease.

The inherited absence of either adenosine deaminase (ADA) or purine nucleoside phosphorylase is associated with severe immunological impairment. We have developed a cell culture model using a mouse T cell lymphoma to simulate ADA deficiency and to study the relationship between purine salvage enzymes and immune function. 2′-deoxyadenosine triphosphate (deoxyATP) levels have been shown to be greatly elevated in erythrocytes of immunodeficient, ADA-deficient patients, suggesting that deoxyadenosine is the potentially toxic substrate in ADA deficiency. Using a potent ADA inhibitor, we have demonstrated that deoxyadenosine is growth-inhibitory and cytotoxic to S49 cells, and that deoxyATP accumulates in these cells. Cell variants, unable to transport or phosphorylate deoxyadenosine, are much less sensitive to deoxyadenosine, indicating that intracellular phosphorylation of deoxyadenosine is required for the lethal effects.We have partially reversed the cytotoxic effects of deoxyadenosine with deoxycytidine in wild-type cells, but we cannot show any reversal in cell lines lacking deoxycytidine kinase. Adenosine (ado) kinase-deficient cells are extremely resistant to deoxyadenosine in the presence of deoxycytidine. This deoxycytidine reversal of deoxyadenosine toxicity is consistent with an inhibition of ribonucleotide reductase by deoxyATP, and we have shown that incubation of S49 cells with deoxyadenosine markedly reduces intracellular levels of deoxyCTP, deoxyGTP and TTP.Kinetics data in wild-type cells and in cell variants are consistent with the presence of two deoxyadenosine-phosphorylating activities — one associated with ado kinase and another associated with deoxycytidine kinase.The S49 cells appear to be a valid model for the simulation of ADA deficiency in cell culture, and from our results, we can suggest administration of deoxycytidine as a pharmacological regimen to circumvent the clinicopathologic symptoms in ADA deficiency.     

Adenosine kinase deficiency in tritiated deoxyadenosine-resistant mouse S49 lymphoma cell lines

Mutant sublines were derived of S49 mouse T-lymphoma cells that were resistant to tritiated deoxyadenosine. Twenty-five isolates that were selected in 1 microCi/ml of the nucleoside were cross-resistant to 6-thioguanine, were sensitive to HAT (hypoxanthine, aminopterin, and thymidine), and contained less than 1% of hypoxanthine phosphoribosyltransferase activity in wild-type cells. One of the mutant clones, S49-dA2, was further subjected to selection in a medium containing 2 microCi/ml tritiated deoxyadenosine and 1 microgram/ml deoxycoformycin, an inhibitor of adenosine deaminase. All resistant subclones were cross-resistant to tubercidin, 6-methylmercaptopurine riboside, and arabinosyladenine. One of the subclones, S49-12, was completely devoid of adenosine kinase and was partially deficient in deoxyadenosine kinase. This subclone, however, contained wild-type levels of deoxycytidine kinase. DEAE chromatography of the wild-type cell extracts revealed two deoxyadenosine phosphorylating activities, one of which coeluted with adenosine kinase and was the enzyme missing in S49-12. The other species phosphorylated both deoxyadenosine and deoxycytidine, of which deoxycytidine was the preferred substrate.     

The effects of an induced adenosine deaminase deficiency on T-cell differentiation in the rat

Inherited deficiency of the enzyme adenosine deaminase (ADA) has been found in a significant proportion of patients with severe combined immunodeficiency disease and inherited defect generally characterized by a deficiency of both B and T cells. Two questions are central to understanding the pathophysiology of this disease: (1) at what stage or stages in lymphocyte development are the effects of the enzyme deficiency manifested; (2) what are the biochemical mechanisms responsible for the selective pathogenicity of the lymphoid system. We have examined the stage or stages of rat T-cell development in vivo which are affected by an induced adenosine deaminase deficiency using the ADA inhibitors, erythro-9-(2-hydroxy-3-nonyl)adenine (EHNA) and 2'-deoxycoformycin (DCF). In normal rats given daily administration of an ADA inhibitor, cortical thymocytes were markedly depleted; peripheral lymphocytes and pluripotent hemopoietic stem cells (CFU-S) all were relatively unaffected. Since a deficiency of ADA affects lymphocyte development, the regeneration of cortical and medullary thymocytes and their precursors after sublethal irradiation was used as a model of lymphoid development. By Day 5 after irradiation the thymus was reduced to 0.10-0.5% of its normal size; whereas at Days 9 and 14 the thymus was 20-40% and 60-80% regenerated, respectively. When irradiated rats were given daily parenteral injections of the ADA inhibitor plus adenosine or deoxyadenosine, thymus regeneration at Days 9 and 14 was markedly inhibited, whereas the regeneration of thymocyte precursors was essentially unaffected. Thymus regeneration was at least 40-fold lower than in rats given adenosine or deoxyadenosine alone. Virtually identical results were obtained with both ADA inhibitors, EHNA and DCF. The majority of thymocytes present at Day 9 and at Day 14 in inhibitor-treated rats had the characteristics of subcapsular cortical thymocytes which are probably the most ancestral of the thymocytes. Thus, an induced ADA deficiency blocked the proliferation and differentiation of subcapsular cortical thymocytes which are the precursors of cortical and medullary thymocytes.     

Distribution of enzymes of purine metabolism in lymphocytes of horse, Equus caballus

A microassay requiring as few as 2 X 10(5) cells per assay was developed for systematic analysis of 9 purine enzymes in lymphocytes from equine peripheral blood, spleen, lymph node, thymus and bone marrow. The activities of adenosine deaminase (ADA), purine nucleoside phosphorylase (PNP), adenosine kinase (AK), deoxyadenosine kinase (dAK), deoxycytidine kinase (dCK), 5'-nucleotidase (5'-N), AMP deaminase, hypoxanthine-guanine phosphoribosyl transferase (HGPRT or HPRT), and adenine phosphoribosyl transferase (APRT) were measured by this microassay in lymphocytes from peripheral blood from four different breeds of horses (Arabian, Quarter Horse, Thoroughbred and Shetland Pony). There were no significant differences in the enzyme activities among the various breeds. Peripheral blood lymphocytes (PBL) from foals exhibited enzyme activities similar to those observed for adult animals. All lymphoid tissue contained similar levels of activity for each kinase (AK, dAK and dCK). Spleen had the highest activity for ADA, PNP, 5'-N, and HGPRT. The lowest activity for ADA, APRT, PNP and AMP deaminase was found in thymus. Enzymatic activities that varied the most among the tissue were 5'-N, ADA, APRT, HGPRT and AMP deaminase.     

Biosensor measurement of purine release from cerebellar cultures and slices

We have previously described an action-potential and Ca²⁺-dependent form of adenosine release in the molecular layer of cerebellar slices. The most likely source of the adenosine is the parallel fibres, the axons of granule cells. Using microelectrode biosensors, we have therefore investigated whether cultured granule cells (from postnatal day 7–8 rats) can release adenosine. Although no purine release could be detected in response to focal electrical stimulation, purine (adenosine, inosine or hypoxanthine) release occurred in response to an increase in extracellular K⁺ concentration from 3 to 25 mM coupled with addition of 1 mM glutamate. The mechanism of purine release was transport from the cytoplasm via an ENT transporter. This process did not require action-potential firing but was Ca²⁺dependent. The major purine released was not adenosine, but was either inosine or hypoxanthine. In order for inosine/hypoxanthine release to occur, cultures had to contain both granule cells and glial cells; neither cellular component was sufficient alone. Using the same stimulus in cerebellar slices (postnatal day 7–25), it was possible to release purines. The release however was not blocked by ENT blockers and there was a shift in the Ca²⁺ dependence during development. This data from cultures and slices further illustrates the complexities of purine release, which is dependent on cellular composition and developmental stage.     

Genetic heterogeneity in adenosine deaminase (ADA) deficiency: five different mutations in five new patients with partial ADA deficiency.

Complete genetic deficiency of adenosine deaminase (ADA) results in a fatal syndrome of severe combined immunodeficiency (SCID). Genetic partial deficiency of ADA, with no detectable enzyme activity in erythrocytes but with variable amounts of enzyme activity detectable in other cells, is usually associated with normal immunologic function but can give rise to a late-onset, cellular immunodeficiency syndrome. We have previously described four different mutant alleles in four such partially ADA-deficient children. We have now examined ADA in lymphoid cells from five additional newly ascertained children with partial ADA deficiency with respect to electrophoretic mobility in starch gel, isoelectric point, heat-stability, and apparent Km and Vmax. These techniques identify at least five different abnormal alleles in these five additional unrelated subjects. Three of these abnormal alleles result in expression of abnormal allelic isozymes (allozymes) different from those previously described. These are: (1) an acidic allozyme that is less acidic than the acidic allozyme we have previously reported; (2) an allozyme that is even less acidic than (1); and (3) an allozyme with apparently normal charge but which is so heat sensitive that the lability to heat can easily be detected at physiologic to febrile temperatures. Two abnormal alleles detected in these five children could correspond with previously reported mutants. These are (4) a basic allozyme that could (but probably doesn't) correspond to the basic allozyme we have previously reported and (5) a "null" allele that cannot be differentiated by these methods from any other "null" allele seen in complete ADA- -SCIDs. Three of the five new patients are genetic compounds, identified either by the presence of two electrophoretically distinguishable allozymes or by family studies that demonstrate presence of a "null" allele in addition to an electrophoretically abnormal allozyme. In three patients, one or both allozymes are phenotypically indistinguishable from an abnormal allozyme also seen in a different individual. Determination of the nucleotide sequence will be required to determine whether or not the phenotypically indistinguishable mutations are indeed genotypically identical. The newly ascertained individuals appear to share a common ethnic West Indian background, out of proportion to the frequency of this ethnic background in the newborn population from which they were ascertained, suggesting that partial ADA deficiency may confer a selective advantage to the homozygous or heterozygous phenotype.     

Structure-Activity Relationships of Adenosine Deaminase Inhibitors

Abstract

Adenosine deaminase (ADA) is an important catabolic enzyme which converts adenosine and deoxyadenosine to inosine and deoxyinosine, respectively. ADA exists in two different isoenzymes, namely ADA1 and ADA2, whose balance in monocytes-macrophages seems to guarantee the homeostasis of adenine nucleosides. Modifications of the purine moiety or/and substitution of the sugar moiety of adenosine with aliphatic chains led to derivatives which are good ADA inhibitors.     

Enzymes of the purine interconversion system in chronic lymphatic leukemia: decreased purine nucleoside phosphorylase and adenosine deaminase activity.

Activities of adenosine deaminase (ADA), adenosine kinase (AK), adenine phosphoribosyltransferase (APRT), hypoxanthine guanine phosphoribosyltransferase (HGPRT), and purine nucleoside phosphorylase (PNP), all enzymes of the purine interconversion system, were determined in lymphocytes of 25 patients with chronic lymphatic leukemia (CLL) and in 23 controls. A statistically significant decrease of PNP activities and a reduction of ADA activities at borderline levels were found in the patients, whereas for the other enzymes assayed no deviation from normal values was observed.     

Expression of human malaria parasite purine nucleoside phosphorylase in host enzyme-deficient erythrocyte culture. Enzyme characterization and identification of novel inhibitors

The intraerythrocytic human malaria parasite, Plasmodium falciparum, requires a source of hypoxanthine for nucleic acid synthesis and energy metabolism. Adenosine has been implicated as a major source for intraerythrocytic hypoxanthine production via deamination and phosphorolysis, utilizing adenosine deaminase and purine nucleoside phosphorylase, respectively. To study the expression and characteristics of human malaria purine nucleoside phosphorylase, P. falciparum was successfully cultured in purine nucleoside phosphorylase-deficient human erythrocytes to an 8% parasitemia level. Purine nucleoside phosphorylase activity was undetectable in the uninfected enzyme-deficient host red cells but after parasite infection rose to 1.5% of normal erythrocyte levels. The parasite purine nucleoside phosphorylase was not cross-reactive with antibody against human enzyme, exhibited a calculated native molecular weight of 147,000, and showed a single major electrophoretic form of pI 5.4 and substrate specificity for inosine, guanosine and deoxyguanosine but not xanthosine or adenosine. The Km values for substrates, inosine and guanosine, were 4-fold lower than that for the human erythrocyte enzyme. In these studies we have identified two novel potent inhibitors of both human erythrocyte and parasite purine nucleoside phosphorylase, 8-amino-5'-deoxy-5'-chloroguanosine and 8-amino-9-benzylguanine. These enzyme inhibitors may have some antimalarial potential by limiting hypoxanthine production in the parasite-infected erythrocyte.     

Differential inhibition of adenosine deaminase deficient peripheral blood lymphocytes and lymphoid line cells by deoxyadenosine and adenosine.

The response to mitogens of peripheral blood lymphocytes (PBL) from adenosine deaminase deficient (ADA^?) patients with Severe Combined Immunodeficiency (SCID), but not from normals, was more sensitive to inhibition by deoxyadenosine than by adenosine. In contrast, proliferation of long-term lymphoid line (B) cells from these patients was essentially equally inhibited by adenosine and deoxyadenosine. Deoxycytidine could “rescue” ADA^? PBLs from deoxyadenosine toxicity.     

Circadian variations of adenosine level in blood and liver and its possible physiological significance

The role of adenosine as a possible physiological modulator was explored by measuring its concentration in different tissues during a 24-hour period. Initially the circadian variations of adenosine and other purine compounds such as inosine, hypoxanthine, uric acid and adenine nucleotides were studied in the rat blood. A daily cyclic response was observed, with low levels of adenosine from 08.00 - 20.00 h, followed by an increase from this time on. Inosine and hypoxanthine levels were elevated during the day and low at night. The uric acid changes observed indicate that the decrease in purine catabolism coincides with a decrease in inosine and hypoxanthine levels and an increase in adenosine. The blood adenine nucleotides, energy charge and phosphorylation potential remained constant during the day and showed oscillatory changes during the night. Similar studies were made in the liver, a primary source of circulating purines. Liver adenosine was high during the night while inosine and hypoxanthine remained low along the 24 hours. The results suggest that liver purine metabolism might participate in the maintenance and renewal of the blood purine pool and in the energy state of erythrocytes in vivo.     

Development of a capillary electrophoresis method for analyzing adenosine deaminase and purine nucleoside phosphorylase and its application in inhibitor screening

A novel capillary electrophoresis (CE) method was developed for simultaneous analysis of adenosine deaminase (ADA) and purine nucleoside phosphorylase (PNP) in red blood cells (RBCs). The developed method considered and took advantage of the natural conversion from the ADA product, inosine to hypoxanthine. The transformation ratio was introduced for ADA and PNP analysis to obtain more reliable results. After optimizing the enzymatic incubation and electrophoresis separation conditions, the determined activities of ADA and PNP in 12 human RBCs were 0.237–0.833 U/ml and 9.013–10.453 U/ml packed cells, respectively. The analysis of ADA in mice RBCs indicated that there was an apparent activity difference between healthy and hepatoma mice. In addition, the proposed method was also successfully applied in the inhibitor screening from nine traditional Chinese medicines, and data showed that ADA activities were strongly inhibited by Rhizoma Chuanxiong and Angelica sinensis. The inhibition effect of Angelica sinensis on ADA is first reported here and could also inhibit PNP activity.     

Purine metabolism in microplasmodia of Physarum polycephalum.

The uptake and utilization of purine nucleosides and purines in microplasmodia of Physarum polycephalum were investigated. The results revealed a unique pattern, namely that exogenous purine nucleosides are readily taken up and metabolised, while free purine bases are hardly taken up. The pathways of incorporation have been elucidated in studies with whole cells and with cell-free extracts. The ribonucleosides (adenosine, inosine and guanosine) can be converted into ribonucleotides in two ways; either directly catalysed by a kinase or by a phosphorolytic cleavage to the free base (adenine, hypoxanthine and guanine respectively) which can then be activated by a purine phosphoribosyltransferase. Apparently the purine phosphoribosyltransferases do not react with exogenous purine bases. The deoxyribonucleosides (deoxyadenosine, deoxyinosine and deoxyguanosine) are also phosphorolysed by purine nucleoside phosphorylase to adenine, hypoxanthine and guanine respectively. A portion of deoxyadenosine is directly phosphorylated to dAMP. It appears that only a minor part of the soluble nucleotide pool can be synthesised from exogenous supplied nucleosides and that none of the deoxyribonucleosides specifically label DNA. There is no catabolism of the purine moiety. In agreement with the above findings, we have found that analoguees of purine nucleosides are more toxic than their corresponding purine base analogues.     

Inosine 5'-monophosphate vs inosine and hypoxanthine as substrates for purine salvage in human lymphoid cells

The ability of inosine 5'-monophosphate vs inosine or hypoxanthine to supply the total purine requirements of mitogen-stimulated human T cells or rapidly dividing human B lymphoblastoid cells was evaluated. Mitogen-stimulated human peripheral blood T cells were treated with aminopterin to inhibit purine synthesis de novo and make the cells dependent upon an exogenous purine source. Thymidine was added as a source of pyrimidines. Under these conditions, 25 microM inosine 5'-monophosphate, inosine, and hypoxanthine showed comparable abilities to support [3H]thymidine incorporation into DNA at rates equal to that of untreated control cultures. In parallel experiments with the rapidly dividing human B lymphoblastoid cell line, WI-L2, treatment with aminopterin (plus thymidine) inhibited the growth rate by greater than 95%. The normal growth rate was restored by the addition of 30 microM inosine 5'-monophosphate, inosine, or hypoxanthine to the medium. However, in similar experiments with cell line No. 1254, a derivative of WI-L2 which lacks detectable ecto-5'-nucleotidase activity, only inosine and hypoxanthine (plus thymidine), but not inosine 5'-monophosphate (and thymidine) were able to restore the growth inhibition due to aminopterin. These results show that the catalytic activity of ecto-5'-nucleotidase is sufficient to meet the total purine requirements of mitogen-stimulated human T cells or rapidly dividing human B lymphoblastoid cells and suggest that this enzyme may have functional significance when rates of purine synthesis de novo are limited and/or an extracellular source of purine nucleotides is available.     

Characterization of a cell culture model for the study of adenosine deaminase- and purine nucleoside phosphorylase-deficient immunologic disease.

The absence of erythrocytic adenosine deaminase (ADA) or purine nucleoside phosphorylase (PNP) has been associated with severe immunodeficiency disease in children. We have developed a cell culture model to study the possible relationships between purine salvage enzymes and immunologic function using an established T cell lymphosarcoma (S49) and a potent inhibitor of ADA, erythro-9(2-hydroxy-3-nonyl) adenine (EHNA). Wild-type S49 cells are killed by dexamethasone or dbc AMP, and adenosine (5 muM) in the presence of an ADA inhibitor (6 muM EHNA) also prevents the growth of and kills these S49 cells. It has been proposed that adenosine is toxic to lymphoid cells by virtue of its ability to increase the intracellular concentrations of cyclic AMP. We examined the sensitivity of three mutants of S49 cells, with distinctive defects in some component of cyclic AMP metabolism or action, to killing by adenosine and EHNA. All three mutants are resistant to killing by isoproterenol or cholera toxin and two are resistant to dbc AMP itself, but all are sensitive to killing by adenosine and EHNA. Similarly, two dexamethasone-resistant S49 mutants are as sensitive to adenosine and EHNA as are the wildtype cells. We have also simulated the purine nucleoside phosphorylase deficiency in S49 cells by adding inosine and adenosine to the growth medium. In the presence of EHNA or inosine, the toxic effects of adenosine can be partially reversed by addition of (10-20 muM) uridine, an observation suggesting that adenosine is toxic as the result of its inducing pyrimidine starvation.     

Establishment and characterization of adenosine deaminase-deficient human T cell lines

We have established long term cell lines from a patient with adenosine deaminase (ADA)-deficient severe combined immunodeficiency by stimulation of blood and bone marrow cells with PHA and IL-2 followed by transformation of the activated cells with the human retrovirus HTLV-I. Despite the absence of detectable T cells in the patients blood, cell lines grew that carried the phenotype of mature activated T cells. TJF-2, the line established from blood, was characterized in detail. The concentration of ADA in TJF-2 cells was less than 1% of normal (3.2 U vs 413.0 U). Studies with pharmacologic inhibitors of ADA suggest that the residual adenosine deaminating activity of TJF-2 is from an enzyme distinct from true ADA, a nonspecific aminohydrolyase. Growth of TJF-2 cells was hypersensitive to inhibition by 2'-deoxyadenosine compared to normal T cells (ID50, 55 microM vs greater than 1000 microM). Analysis of 2'-deoxyadenosine-challenged cells showed that TJF-2 cells accumulated significant levels of deoxyadenosine triphosphate, whereas normal T cells did not unless they were also incubated with the ADA inhibitor deoxycoformycin. Southern and Northern blot analysis of these cells revealed a grossly intact ADA gene that produced a normal size ADA mRNA. Yet, despite ADA deficiency, cells of the TJF-2 line were otherwise indistinguishable from HTLV-I-transformed T cells derived from normal donors with respect to dependence on exogenous IL-2 for growth, clonal rearrangement patterns of TCR beta-chain genes, response to PHA, and rapid restoration of cellular volume after hypotonic challenge. The TJF-2 line thus represents a unique HTLV-I-transformed human T cell line exhibiting ADA deficiency and its expected metabolic consequences.     

Ecto-5'-nucleotidase can provide the total purine requirements of mitogen-stimulated human T cells and rapidly dividing human B lymphoblastoid cells

The ability of mitogen-stimulated human T cells or rapidly dividing human B lymphoblastoid cells to drive their total purine requirements from inosine 5'-monophosphate, inosine, or hypoxanthine was compared. Inosine 5'-monophosphate first must be converted to inosine by the action of the enzyme ecto-5'-nucleotidase before it can be transported into the cell; inosine and hypoxanthine, however, can be transported directly. Mitogen-stimulated human peripheral blood T cells were treated with aminopterin to inhibit purine synthesis de novo and to make the cells dependent on an exogenous purine source. Thymidine was added as a source of pyrimidines. Under these conditions, 30 microM inosine 5'-monophosphate, inosine, and hypoxanthine showed comparable abilities to support [3H]thymidine incorporation into DNA or [3H]leucine incorporation into protein at rates equal to that of untreated control cultures. Similar results were found when azaserine was used to inhibit purine synthesis de novo, and thus DNA synthesis. In parallel experiments with the rapidly dividing human B lymphoblastoid cell line WI-L2, treatment with aminopterin (plus thymidine) inhibited the growth rate by greater than 95%. The normal growth rate was restored by the addition of 30 microM inosine 5'-monophosphate, inosine, or hypoxanthine to the medium. However, in similar experiments with cell line 1254, a derivative of WI-L2 which lacks detectable ecto-5'-nucleotidase activity, inosine and hypoxanthine (plus thymidine), but not inosine 5'-monophosphate (and thymidine) were able to restore the growth inhibition due to aminopterin. These results show that the catalytic activity of ecto-5'-nucleotidase is sufficient to meet the total purine requirements of mitogen-stimulated human T cells or rapidly dividing human B lymphoblastoid cells, and suggest that this enzyme may be important for purine salvage when rates of purine synthesis de novo are limited and/or an extracellular source of purine nucleotides is available.