Extracellular Nucleotide Catabolism in Human B and T Lymphocytes THE SOURCE OF ADENOSINE PRODUCTION*

Extracellular nucleotide degradation was studied in intact human B and T lymphocyte subpopulations and in lymphoblastoid cell lines. Cells of B lymphocyte lineage showed high nucleotide degrading activity, whereas T lymphocytes were unable to degrade extracellular nucleotides. The external surface of B cells contained active sites of ecto-triphosphonucleotidase (ecto-ATPase), ecto-diphosphonucleotidase (ecto-AD- Pase), and ecto-monophosphonucleotidase (ecto-AM-Pase). The expression of all three ectoenzyme activities seemed closely associated with B cell development. ATPase and ADPase activities increase continuously during B cell maturation, ecto-AMPase activity, on the other hand, reaches maximal activity in late pre-B cells. These results combined with our previous studies of intracellular ATP catabolism (Barankiewicz, J., and Cohen, A. (1984) J. Biol. Chem. 259, 15178-15181) provide evidence that extracellular ATP catabolism may represent exclusive source for adenosine in lymphocytes. It is suggested that adenosine may serve as a means of communication between B and T cells in lymphoid organs, B lymphocytes being the sole produc-ers of adenosine and T lymphocytes being the recipi- ents of this signal. extracellular nucleotides was examined in cell pellets following three washes in medium. Nucleotides were extracted with 50 pl of 0.4 M perchloric acid for 5 min on ice, centrifuged, neutralized with Alamine-Freon mixture, and analyzed for radioactivity as described (26). Separation of Nucleotides, Nucleosides, and Bases-Separation of nucleotides was performed using one-dimensional chromatography on polyethyleneimine cellulose TLC in three steps of increasing sodium formate buffer concentrations (0.5, 2, 4 M) (27). Separation of nucleosides and bases was done on one-dimensional cellulose TLC with attached Whatman 3MM paper wick on top in 1-bu- tanol:methanol:water:ammonia solvent (6020201). This chromatography allowed the separation of adenosine, hypoxanthine, and ino- sine, whereas nucleotides remained at the base line. The radioactivity or separated nucleotides, nucleosides, and bases was measured in a Beckman LS 3801 scintillation counter.

In contrast to the established view, it has been found that nucleotides, such as ATP can be found in considerable amounts outside of cells, e.g. circulating in plasma (10). Moreover, a variety of cells have also been found to contain extracellular nucleotide degradation enzymes (10-13). Nucleotides which are released from some cells to the extracellular space can reach micromolar concentrations and can modulate many biological processes by acting via specific cell surface receptors.
Lymphocytes are among the cells which have been reported to express ecto-nucleotidase activities (14-16). Although little is known about extracellular enzyme expression and about the role of extracellular nucleotide metabolism in lymphocyte function, some effects of extracellular nucleotides have been suggested. At high concentrations ATP stimulates in vitro DNA synthesis in bone marrow and in thymus cells but it inhibits DNA synthesis in spleen, lymph node, and peripheral blood lymphocytes (17, 18). Extracellular ATP inhibits the activities of natural killer cells (19, 20). On the other hand, adenosine, the degradation product of ATP, acting via both intracellular and extracellular adenosine receptors (21) has been reported to regulate various lymphocyte functions through activation of adenylate cyclase (22-24).
Despite the reported regulatory roles of extracellular adenosine in lymphocyte function (5,24) the cellular source of adenosine production is not clear. In searching for the potential source of adenosine production in lymphocytes, we have previously shown that adenosine is not produced intracellularly from ATP degradation (25). In the present work we explore the possibility that adenosine may be produced by catabolism of extracellular nucleotides in lymphocyte subpopulations. Toward that end we have conducted a comprehensive study of extracellular nucleotide metabolism in different human B and T lymphocyte subpopulations and in lymphoblastoid cell lines representative of various differentiation stages. Cells-Fresh human lymphocytes were purified from tonsils and thymus. Mononuclear cells were prepared on Ficoll-Hypaque density gradients (S), resuspended in serum free RPMI 1640 medium, and immediately used for incubation with radioactive precursors. Fetal calf serum (FCS) was omitted in incubations with nucleotides because even heat-inactivated serum (68 "C, 30 min) contained considerable activities of enzymes participating in nucleotide degradation. When cells were incubated for several hours with other compounds, medium was supplemented with 10% heat-inactivated FCS, but before incubation with radioactive nucleotides cells were washed 3 times in serum free RPMI. E-rosetting T cells and non-E-rosetting B cells from peripheral blood or tonsils were enriched (>go%) by rosette depletion on Ficoll-Hypaque gradients (8). Briefly, nonseparated lymphocytes were mixed with neuramidase-treated (37 "C for 30 min) sheep red blood cells (37 "C for 5-10 min) and, after centrifugation at 4 "C (1000 rpm for 10 min), incubated at 4 "C for 30 min. After gentle resuspension, cells were centrifuged on Ficoll-Hypaque (2000 rpm for 30 min at 4 "C). Non-E-rosetting hand was then removed from gradient and washed 3 times with medium. E-rosetting cell pellet was resuspended with Tris-ammonium chloride buffer and immediately washed with medium. After separation, B and T cells were usually resuspended in RPMI medium containing 10% FCS for 1-2 h (37 "C) before they were used in these studies. B cell fraction contained at least 95% of surface Ig positive cells, whereas the T cell fraction contained at least 95% of OKT3 positive cells.

Materiuls
The presence of attached sheep red blood cell ghosts did not affect the assays, because even in their presence no ectonucleotidase activity was detected.
Lymphoblastoid cell lines of B or T lineage were maintained in logarithmic growth at 37 'C in 5% CO, in RPMI 1640 supplemented with 10% heat-inactivated FCS. Cells were washed three times in serum-free RPMI medium prior to incubations with radioactive precursors. Where indicated that nucleoside transport was blocked, cells were preincubated with 10 p~ dipyridamole (20 min, 37 "C).
Cell Integrity-Cell membrane integrity was determined by trypan blue dye exclusion. In addition, to estimate cell disruption and spill of intracellular enzymes, activities of the cytoplasmic enzymes adenylate deaminase and pyruvate kinase were measured in culture supernatants. No measurable extracellular activities of these enzymes were found during 2 h incubation of B or T lymphocytes indicative of lack of cell breakage. It was found that as far as 5% nonviable cells (by trypan blue dye exclusion) can produce significant spill of cytoplasmic enzymes, and at least 99% viability was required for the experiments reported. In cases where the viability was less than 99% contaminating dead cells and cytoplasmic enzymes were successfully removed by a second Ficoll-Hypaque gradient centrifugation (8).
Extracellular Nucleotide Metabolism-To measure the rate of extracellular degradation of nucleotides, intact cells (1 X IO6) were incubated with 1 pCi of the respective radioactive nucleotides (500 or 5 pM initial concentrations). Cells were then centrifuged (1500 rpm, 5 rnin), and supernatants were analyzed for radioactivity in nucleotides, nucleosides, and bases. To detect possible interference by nonspecific phosphatases, cells (1 X lo6) were incubated in RPMI containing 10 mM p-nitrophenyl phosphate for 60 min. Measurement of absorbance (A = 410 nm) showed no phosphatase activity outside the cells. The addition of 10 mM p-nitrophenyl phosphate to B cells did not affect ATP degradation.
Intracellular Nucleotide Incorporation-The intracellular incorporation of radioactivity from extracellular nucleotides was examined in cell pellets following three washes in medium. Nucleotides were extracted with 50 pl of 0.4 M perchloric acid for 5 min on ice, centrifuged, neutralized with Alamine-Freon mixture, and analyzed for radioactivity as described (26).
Separation of Nucleotides, Nucleosides, and Bases-Separation of nucleotides was performed using one-dimensional chromatography on polyethyleneimine cellulose TLC in three steps of increasing sodium formate buffer concentrations (0.5, 2, 4 M) (27). Separation of nucleosides and bases was done on one-dimensional cellulose TLC with attached Whatman 3MM paper wick on top in 1-butanol:methanol:water:ammonia solvent (6020201). This chromatography allowed the separation of adenosine, hypoxanthine, and inosine, whereas nucleotides remained at the base line. The radioactivity or separated nucleotides, nucleosides, and bases was measured in a Beckman LS 3801 scintillation counter.

7095
ible in at least three experiments. Variation of individual values from the mean was less than 10%. and T lymphocytes (-35%) actively degraded extracellular ATP, as well as other extracellular triphosphonucleosides, diphosphonucleosides, and monophosphonucleosides (results not shown). To determine the cellular source of these ectonucleotidases, B and T cells were separated and incubated with radioactive ATP. After 30 min of incubation of purified B cells with 500 p~ ATP, more than 80% of the extracellular ATP was degraded (Fig. lA), whereas T cells were unable to degrade any ATP (Fig. 1D). ATP degradation by B cells was accompanied by the accumulation of AMP, ADP (Fig. lA), and adenosine (Fig. 1B). After 20 min of incubation, the amount of extracellular ADP declined and AMP concentration reached a plateau. Adenosine was the main product of extracellular ATP degradation and its concentrations increased linearly in the medium. Small amounts of hypoxanthine and inosine were also found in the medium ( Fig. 1 B ) , but preincubation of tonsil cells with the nucleoside transport inhibitor dipyridamole completely prevented inosine and hypoxanthine accumulation (results not shown), indicating that inosine and hypoxanthine were produced exclusively inside the cells. Only a small amount (-1%) of radioactivity was incorporated into intracellular nucleotides, mainly ATP ( Fig.   1C).

Extracellular ATP Catabolism by Tonsil Cells-Mononuclear tonsil cells, containing mainly B cells (-65%)
At a lower initial concentration of extracellular ATP (5 p~) , 80% of the ATP added was degraded by lo6 purified B cells within 5 min of incubation, with a transient accumulation of a small amount of ADP and AMP ( Fig. 2A). Adenosine and inosine were formed transiently and then declined, whereas hypoxanthine was produced in a linear fashion (Fig.   2B). A significant incorporation of radioactivity into intracellular nucleotides was also observed (Fig. 2C). In the presence of dipyridamole, extracellular nucleotide degradation remained unchanged (Fig. 2D), but only adenosine accumu- lated in the extracellular medium (Fig. 2E). The inhibition of inosine and hypoxanthine formation in presence of dipyridamole was associated with inhibition of incorporation of radioactivity into intracellular nucleotide pools (Fig. 2 F ) indicating that inosine and hypoxanthine found in the extracellular medium (Fig. 2 B ) are produced exclusively by intracellular deamination of adenosine. Extracellular nucleotide catabolism seems to be of low specificity, with comparable degradation of dATP, GTP, dGTP, dCTP (Fig. 3). However, the activity toward ribonucleotides seemed to be higher than that for deoxyribonucleotides. M$+ was found to stimulate (400 pM MgClz optimal concentration) degradation of extracellular ATP. No  there was no release of any ectonucleotidase activities into the extracellular medium (results not shown), indicating that ectonucleotidases are integral cell surface membrane enzymes. Extracellular ATP and AMP Catabolism by Various B and T Lymphocytes-Extracellular degradation of ATP (Table I) and other nucleotides (results not shown), similar to that described above for B lymphocytes from tonsil, were also found in peripheral blood B lymphocytes, and in various B lymphoblastoid cell lines. In contrast, all cells of the T lymphocyte lineage such as thymocytes, peripheral blood T lymphocytes, and various T lymphoblastoid cell lines showed no or very little ability to degrade extracellular nucleotides (Table I). Also, extracellular AMP was fast degraded by different B lymphocyte subpopulations, whereas cells of T lymphocyte lineage showed practically no such abilities (Table 11). Human red blood cells showed no ability to degrade extracellular nucleotides (results not shown).
Extracellular Nucleotide Degradation during B Cell Maturution-To examine the expression of ecto-nucleotidase activities within the framework of human B cell development, a  Human lymphocytes or lymphoblastoid cells (1 X lo6) were preincubated for 20 min with 10 PM dipyridamole and then incubated with 1 FCi of radioactive AMP (500 W M initial concentration) for 60 min. Incorporation of radioactivity into extracellular nucleotides, nucleosides, and bases was measured. PBL, peripheral blood lymphocytes. Hyp, hypoxanthine. lymphocytes, showed at best little extracellular ATPase, AD-Pase, and AMPase activities (Fig. 4) AMP at 500 p~ initial concentration). After 30 min of incubation, radioactive nucleotides, nucleosides, and bases were analyzed in the medium and the sum of all radioactive products calculated.
somewhat lower in late pre-B cells as compared to the more mature B cell lines. Interestingly, the levels of ecto-AMPase expressed on mature B cells were very low and comparable to that found on the most immature clones in this series.

DISCUSSION
The importance of nucleotide degradation pathways to lymphocyte function has been evident from the association of severe immunodeficiencies with defects of nucleotide degradation enzymes, adenosine deaminase, and purine nucleoside phosphorylase (5). The relevance of ectoenzymes of the nucleotide degradation pathway expressed on the cell surface of lymphocytes and their potential significance to immune function has attracted attention only more recently (10-13, 28-30). Although one of these ectoenzymes, ecto-5'-nucleotidase (AMPase), has been studied more widely (31), understanding of its biological function awaits a more complete description of extracellular purine catabolism. Human B-lymphocytes and lymphoblastoid cell lines can actively hydrolyze exogenous triphosphonucleosides, diphosphonucleosides, and monophosphonucleosides, whereas T lymphocytes and T cell lines have at best little ability to do so (Tables I and 11). It was reported previously that certain T cell populations exhibit 5"nucleotidase activity (6-9). However our experience shows that (a) only T cell preparations with less than 99% viability and with detectable extracellular activity of cytoplasmic enzymes exhibit ecto-AMPase activity, and (b) complete elimination of dead cells removed ecto-AMPase activity in all T cells examined (Table I, see also "Experimental Procedures").
Within the B cells examined, the expression of ATPase/ ADPase ectoenzyme(s) appears to be closely associated with B cell differentiation, increasing continuously with maturation (Fig. 4) There are significant differences between extracellular and intracellular catabolism of ATP. Intracellular ATP catabolism proceeds exclusively via AMP deamination, and, therefore, under physiological conditions adenosine is not formed (25). Hypoxanthine, the end product of this intracellular catabolic pathway, can be salvaged back into intracellular nucleotide pools or excreted. In contrast, extracellular ATP catabolism proceeds exclusively via AMP dephosphorylation, resulting in the formation of adenosine as the end product (Scheme 1). Taken together, these results indicate that the exclusive source for adenosine in lymphoid tissues is extracellular ATP catabolism by B lymphocytes. Adenosine formed this way can act either as an extracellular physiological modulator through specific membrane receptors or, alternatively, it can enter the cells and participate in nucleotide synthesis. The results reported here (Table I), together with our previous observations (25), demonstrate that only B lymphocytes, and not T lymphocytes, can produce adenosine from ATP. On the other hand, only T lymphocytes, and not B lymphocytes, express the cell surface receptor for adenosine (33). A similar dichotomy has been observed in the interaction between endothelial cells (adenosine producers) and heart muscle cells (expressing adenosine receptors (34)). It has been suggested that this cellular dichotomy of adenosine production versus utilization may serve as a regulatory cycle controlling the interaction between adenosine producing endothelial cells and the recipient muscle cells (34). Similarly, in lymphoid tissues an adenosine regulatory cycle may operate between certain adenosine producing B cells and T cells expressing the adenosine receptor, but this needs more studies.