Concentrative Influx of Functionally Active Cyclic ADP-ribose in Dimethyl Sulfoxide-differentiated HL-60 Cells*

Native human HL-60 cells do not express CD38, a multifunctional ectoenzyme, which generates cyclic ADP-ribose (cADPR), a potent calcium mobilizer. However, when HL-60 cells are induced to differentiate to granulocytes by treatment with retinoic acid (RA), they express CD38 and accumulate cADPR. Both processes play a causal role in RA-induced differentiation. Other granulocyte differentiation-inducers, including dimethyl sulfoxide (Me2SO), fail to induce CD38 expression. We investigated whether treatment of HL-60 cells with Me2SO involves any changes in the cADPR/intracellular calcium ([Ca2+]i) signaling system and, specifically, whether Me2SO affects those nucleoside transporters (NT) (both equilibrative (ENT) and concentrative (CNT)) that mediate influx of extracellular cADPR. Semiquantitative polymerase chain reaction analysis of transcripts, binding of [3H]nitrobenzylthioinosine (NBMPR) to intact cells, and influx experiments of extracellular cADPR (with selective inhibitors of NT as NBMPR or in specific conditions) were performed in native and Me2SO-differentiated HL-60 cells. The native cells showed uptake of cADPR across ENT2, whereas influx of cADPR into the Me2SO-differentiated cells occurred mostly by concentrative processes mediated by CNT3 and by an NBMPR-inhibitable concentrative NT designated cs-csg. Me2SO-differentiated, but not native HL-60 cells, accumulated cADPR and showed increased [Ca2+]i levels when grown in a transwell co-culture setting over CD38-transfected 3T3 fibroblasts where nanomolar cADPR concentrations are present in the medium. NBMPR inhibited both responses of Me2SO-induced cells. Thus, concentrative influx of extracellular cADPR across CNT3 and cs-csg NT could substitute in the absence of CD38 in eliciting cADPR-dependent [Ca2+]i increases in granulocyte-differentiated HL-60 cells, as well as in other CD38– cells.

mide and cyclic ADP-ribose (cADPR) and NADP ϩ to NAADP ϩ (1)(2)(3)(4). Both cADPR and NAADP ϩ are powerful, universal calcium-mobilizing metabolites and behave as the second messengers of many agonists, thereby modulating a wide number of calciummediated cell processes (reviewed in Refs. [5][6][7][8][9]. In mammalian cells, the best known ADP-ribosyl cyclases are CD38, a type II transmembrane 48 kDa glycoprotein (10,11) and BST-1, a glycosylphosphatidylinositol-anchored protein (12,13). Both CD38 and BST-1, sharing significant sequence identity, feature two important properties (9). (i) They are multifunctional enzymes able to generate cADPR and to hydrolyze cADPR to ADP-ribose. (ii) Both are ectoenzymes. The latter property has prompted studies aimed at defining two major topological inconsistencies of the CD38/cADPR system. (i) The apparent unavailability of the substrate NAD ϩ to the ectocellular active site of CD38 and (ii) generation of cADPR at sites opposite to cADPR-responsive Ca 2ϩ stores and gated by the ryanodine receptors (RyR) (either the extracellular environment or the intravesicular space of intracellular membrane vesicles, Ref. 14). These investigations led to the discovery that the CD38/cADPR system is a remarkably compartmentalized one and that a number of previously unrecognized transporters can translocate NAD ϩ and cADPR, thereby circumventing both topological restrictions (15). Specifically, NAD ϩ proved to be transported in a calcium-regulated manner, according to an equilibrative mechanism by hexameric hemichannels of connexin 43 (Cx43) that allow it to access the active site of CD38 (16,17). In contrast, cADPR showed a surprising redundancy of transporters. Native, transmembrane CD38 itself is a homodimer acting as a catalytically active transporter of cADPR, i.e. as an enzyme that couples cADPR generation at its ectocellular active site with subsequent concentrative channeling of this product across the plasma membrane (18). Recently, cADPR was also demonstrated to cross the plasma membrane of a number of CD38 Ϫ cells through some members of the wide family of nucleoside transporters, both equilibrative (ENT) and concentrative (CNT). Specifically, among these promiscuous transporters, ENT2, CNT2, and a not yet molecularly characterized nitrobenzylthioinosine (NBMPR)-inhibitable nucleoside transporting system proved to mediate translocation of cADPR in CD38 Ϫ cells, independent of the concentrative transport catalyzed by CD38 (19).
The functional interplay among some of these nucleotide transporters, notably Cx43 hemichannels and CD38, can increase the levels of [Ca 2ϩ ] i , as shown by recent uncoupling experiments on 3T3 fibroblasts lacking individually CD38 or Cx43 (17). These and other results demonstrated that an enhanced subcellular trafficking of NAD ϩ and cADPR can upregulate [Ca 2ϩ ] i levels and accordingly some calcium-mediated functions (15). However, the same transporters, including the cADPR-translocating CNT, are also responsible for an extracellular trafficking, whereas some cADPR-generating cells can target neighboring, even CD38 Ϫ but RyR ϩ cells, so as to induce into them significant cADPR-related increases of [Ca 2ϩ ] i and of calcium-mediated cell functions. These include increased contractility of smooth myocytes (20), increased calcium responses in neurons (21), enhanced proliferation of 3T3 fibroblasts (22), and expansion of human hemopoietic progenitors (23).
Recently, Munshi et al. (24) demonstrated that CD38 expression in the constitutive CD38 Ϫ human promyelocytic leukemia cell line HL-60 and the consequent intracellular accumulation of cADPR play a causal role in mediating its granulocytic differentiation induced by RA. These results and our recent observation that extracellular cADPR, even at extremely low concentrations, triggers expansion of human hemopoietic progenitors (23) led us to investigate whether cADPR can be internalized by HL-60 cells across some NT. The striking differences between native and differentiated HL-60 cells suggested an investigation of both cell populations in terms of NT expression and of cADPR influx. The results obtained in this study provide evidence that the undifferentiated and the dimethyl sulfoxide (Me 2 SO)-differentiated HL-60 cells differ significantly in both parameters. The comparatively enhanced influx of extracellular cADPR that occurs in differentiated cells is accounted for by overexpression of CNT3 and of the NBMPRinhibitable CNT system designated as cs-csg (25)(26)(27).
Cell Cultures-HL-60 cells, obtained from ATCC (Manassas, VA), were maintained in suspension in RPMI 1640 medium supplemented with 10% fetal calf bovine serum and kept at 37°C in a 5% CO 2 atmosphere. COS-7 cells were cultured as described (19). Transfected NIH CD38 ϩ and CD38 Ϫ 3T3 fibroblasts were obtained and cultured as described (28) Differentiation of HL-60 Cells by Various Inducers-Native HL-60 cells were harvested during exponential growth and seeded at a density of 0.2 ϫ 10 6 cells/ml in RPMI 1640 medium supplemented with 10% fetal calf bovine serum, in the presence of 1.2% Me 2 SO for 5 days (29), or of 1 M RA for 4 days (24), or of 200 ng/ml phorbol 12-myristate 13-acetate (PMA) for 2 days (30). Differentiation was measured as described by Munshi et al. (24). Briefly, HL-60 cells (0.4 -2 ϫ 10 6 ) were incubated for 60 min at 37°C in PBS containing 2 mg/ml of NBT and 20 ng/ml of PMA. The reaction was terminated by the addition of 0.4 ml of 2 N HCl and cooled on ice for 30 min. After centrifugation for 5 min at 700 ϫ g, the black formazan deposits in the pellets were dissolved in 1 ml of Me 2 SO, and the absorbance at 590 nm was measured. Data are expressed as absorbance units/mg protein.
Assays of CD38 Enzymatic Activities-GDP-ribosyl cyclase and cADPR-hydrolase activities were assayed on intact native and differentiated HL-60 cells as described (22).
cADPR Influx into Intact HL-60 cells and COS-7 Cells-Native and Me 2 SO-differentiated HL-60 cells, as well as trypsinized COS-7 cells, were resuspended at 10 7 cells/ml in Na ϩ buffer (135 mM NaCl, 6.3 mM K 2 HPO 4 , 2.7 mM KCl, 1.5 mM KH 2 PO 4 , 0.5 mM MgCl 2 , 0.9 mM CaCl 2 , 10 mM glucose, pH 7.4) or Li ϩ buffer (same as Na ϩ buffer, but with 135 mM LiCl instead of NaCl) in the presence of different cADPR concentrations at 22°C (19). At the times indicated after cADPR addition (from 0 to 10 min), 50-l aliquots were withdrawn and centrifuged at 5,000 ϫ g for 15 s. Pellets were washed with 1.5 ml of ice-cold appropriate buffer (Na ϩ or Li ϩ buffer), containing 10 mM uridine, in order to inhibit equilibrative efflux of cADPR (19), and submitted to two consecutive centrifuga-tions as above in order to completely remove the supernatants. Pellets were resuspended in 300 l of water, and the suspensions were sonicated for 30 s at 3 watts in ice. Aliquots of 280 l were deproteinized with 0.6 M perchloric acid (final concentration). cADPR was then measured by the enzymatic cycling assay recently described by Graeff and Lee (31), whose sensitivity is 10 fmol of cADPR. Protein content was determined for 20-l aliquots according to Bradford (32).
[ 14 C]Inosine Influx into COS-7 Cells-Intact pcDNA/hCNT3 transfected COS-7 cells (10 7 cells/ml, see below) were incubated at 22°C in either Na ϩ or Li ϩ buffer containing 10 M [ 14 C]inosine (117,000 cpm/ nmol). The specific inosine influx, defined as the total uptake subtracted of the influx observed in the presence of 10 mM cold inosine, was determined as described (19).
NBMPR Binding Studies-Native and Me 2 SO-differentiated HL-60 cells were collected, washed twice in 5.0 ml of PBS, and resuspended at 6 ϫ 10 6 cells/ml in Na ϩ buffer, with or without 20 M cold NBMPR, to determine nonspecific binding. Binding assays were performed on 0.5-ml cell suspensions containing graded concentrations of [ 3 H]-NBMPR (0.05-2.5 nM). After 90 min of incubation at 22°C, 400-l aliquots were collected and rapidly filtered on a prewashed glass fiber filter (Whatman, Maidenstone, UK). Filters containing the retained cells were washed with 5.0 ml of ice-cold Na ϩ buffer and analyzed to determine [ 3 H]NBMPR content by liquid scintillation counting. Free NBMPR was determined from the radioactivity present in 100 l of the incubation buffer at the end of the binding assay. Specific binding was determined as the difference between the total binding of [ 3 H]NBMPR to the cells and the [ 3 H]NBMPR binding in the presence of excess non-isotopic NBMPR (nonspecific binding). Parameters of NBMPR binding were obtained from Scatchard plots of bound/free versus bound NBMPR (33). Non-linear plots were resolved graphically by the method of Rosenthal (34).
Reverse Transcriptase (RT)-Polymerase Chain Reaction Analyses-Total RNA from both native and Me 2 SO-differentiated HL-60 cells was isolated using the RNase Mini Kit (Qiagen, Milan, Italy) and treated with RNase-free DNase-set (Qiagen) according to the manufacturer's instructions. RT reactions were performed using Thermoscript RT-PCR System (Invitrogen); 4 g of total RNA were denatured in the presence of 1 mM dNTP and 50 pmol of oligo(dT) 20 for 5 min at 65°C, then incubated in the appropriate buffer (1ϫ cDNA synthesis buffer) with 5 mM dithiothreitol, 20 units of RNase inhibitor (RNaseOUT TM ) and 15 units of ThermoScript TM at 50°C for 1 h. The reactions were terminated by incubation for 5 min at 85°C, and the cDNA was treated with 2 units of RNase H (Invitrogen). The resulting total cDNA was then amplified by PCR to analyze the mRNA expression of the five cloned nucleoside transporters (ENT1, ENT2, CNT1, CNT2, CNT3), using the genespecific primer pairs reported previously (19). PCR amplifications were run in 25-l reaction volumes containing 1.5 mM MgCl 2 , 200 M dNTP, 2.5 units of TaqDNA polymerase (Promega Italia, Milan, Italy) and using 10 pmol of the specific primers (TibMolBiol, Genova, Italy). The PCR conditions for all cDNAs were as follows: an initial denaturation at 94°C for 1 min, followed by 94°C for 30 s, 58°C for 30 s, 72°C for 1 min, for 35 cycles, and then a final elongation step was performed at 72°C for 7 min. The results of PCR amplification, analyzed by agarose gel electrophoresis and ethidium bromide staining, were visualized using the Chemi Doc System (Bio-Rad). Gene Ruler TM 100-bp DNA Ladder (MBI Fermentas, Hanover, MD) was used as a standard marker.
Semiquantitative PCR Analyses of CNT3 and ENT2 Transcripts-The cDNA equivalent of 50 ng of starting RNA from undifferentiated and Me 2 SO -differentiated HL-60 cells was amplified by PCR using gene-specific primers (19) for human CNT3 (hCNT3), ENT2 (hENT2), ENT1 (hENT1), and ␤-actin as internal control. The PCR incubation was performed as described above, and PCR conditions for cDNAs were: 30 s, 94°C; 30 s, 58°C; 1 min, 72°C, for a number of cycles optimized for each gene (30 cycles for CNT3, 28 cycles for ENT2 and 22 cycles for ENT1 and ␤-actin). The abundance of PCR products was semiquantified by densitometric scanning of the ethidium bromide-stained agarose gels using Chemi Doc System, and the cDNA fragments corresponding to each amplified gene was compared between native and Me 2 SOdifferentiated HL-60 cells.
Transient Transfection of COS-7 Cells with Human CNT3 cDNA-Total RNA from Me 2 SO-differentiated HL-60 cells was purified, and the reverse transcription reaction was performed as described above (see reverse transcriptase-PCR analyses). The following primers were used to amplify full-length human cDNA CNT3 (GenBank TM accession number AF305210): sense, 5Ј-CACCAAGAGCATGGAGCT-GAGGAG-3Ј (nucleotide positions 87-106); antisense, 5Ј-TCAGAGTTC-CACTGGAGAAGTG-3Ј (nucleotide positions 2197-2176). PCR amplifi-cation was performed in a 25-l reaction volume containing 1 mM MgSO 4 , 300 M dNTP, 15 pmol of primers, and using 1.25 units of AccuPrime TM Pfx DNA polymerase (Invitrogen). The PCR conditions were as follows: an initial denaturation at 95°C for 2 min, followed by 95°C for 15 s, 56°C for 30 s, 68°C for 3 min, for 35 cycles. The 2.11-kb PCR product was cloned in the eukaryotic expression plasmid pcDNA3.1D/V5-His-TOPO using the pcDNA3.1 Directional TOPO expression kit (Invitrogen). The plasmid obtained was sequenced on both strands of cDNA to verify the presence of correct insert. The pcDNA3.1 plasmid and the pcDNA3.1/CNT3 construct were transfected into COS-7 cells by LipofectAMINE 2000 (Invitrogen) according to the manufacturer's instructions. After 48 h, cells were collected, and the expression of CNT3 and ␤-actin (as internal control) was monitored by RT-PCR, performed as described above.
Co-cultures of Native and Me 2 SO-differentiated HL-60 Cells over 3T3 CD38 Ϫ/ϩ Feeder Cells-Both native and Me 2 SO-differentiated HL-60 cells (2.0 ml), after 4 days of treatment, were seeded in transwell plates (2.5-cm diameter; 0.4-M pore size) on pre-established (1.0 ml) confluent feeder layers of CD38 ϩ/Ϫ 3T3 cells (transfected with the sense or antisense cDNA for human CD38 Ϫ , Ref. 28) and incubated for 48 h. Co-cultures were grown at 37°C in the absence or in the presence of 1 M NBMPR. HL-60 cells were recovered, washed once with 5.0 ml of PBS, resuspended in 300 l of water, and sonicated for 30 s at 3 watts in ice. A 20-l aliquot was used for protein determination. Perchloric acid was then added at a final 0.6 M concentration to cell lysates. cADPR content of the samples was determined by the enzymatic cycling assay (31).
Fluorimetric Determination of Intracellular Calcium [Ca 2ϩ ] i Concentrations-HL-60 cells co-cultured 48 h over 3T3 CD38 ϩ/Ϫ feeder cells were recovered from the transwells (about 10 6 cells for undifferentiated samples and 0.5 ϫ 10 6 for Me 2 SO-differentiated samples), washed twice in 5.0 ml of phosphate-buffered saline, resuspended in 1.0 ml of fresh complete medium, and incubated in the presence of Fura2-AM (10 M) for 45 min at 37°C. Fura2-loaded cells were washed twice in 3.0 ml of zero calcium solution (35). Calcium measures were performed in a 2.0-ml cuvette under continuous stirring in zero calcium solution as described (35). Statistical analysis of different calcium (values) was performed using one-way analysis of variance. p values were considered statistically significant when Ͻ 0.05.

Distinctive Expression of Equilibrative and Concentrative
Nucleoside Transporters in Native and Me 2 SO-differentiated HL-60 Cells-Native, CD38 Ϫ (29, 36) HL-60 cells can be induced to terminally differentiate into granulocytes, monocytes, or macrophages in response to various inducers (29). Their treatment with RA, which induces differentiation into granulocytes (37), results in a dramatic extent of CD38 expression (24,29,36). This increase was recently demonstrated to play a causal role in the process of differentiation (24) and is RAspecific. Other differentiation-inducing agents, e.g. Me 2 SO (granulocyte differentiation-inducing) and PMA, a macrophage differentiation inducer, Ref. 29), failed to induce CD38 expression in HL-60 cells. Conversely, HL-60 cells induced to differentiate by treatment with both Me 2 SO and PMA exhibit a decrease of equilibrative nucleoside transport (30,38), accompanied by an increase in the rate of concentrative Na ϩ -dependent transport of nucleosides, the extent of which correlates with enhanced CNT3 expression in PMA-differentiated HL-60 cells (30).
Our recent observation of translocation of cADPR across selected NT (19) suggested the use of various inducers of HL-60 cell differentiation in order to comparatively investigate the uptake of extracellular cADPR by native and differentiated HL-60 cells, respectively.
We first analyzed the mRNA expression of the five molecularly defined human NT (hNT), both equilibrative (ENT1 and ENT2) and concentrative (CNT1, CNT2, and CNT3), in native HL-60 cells and, by comparison, in the same cells induced to differentiate with Me 2 SO, RA, and PMA, respectively. PCR analysis demonstrated the occurrence of mRNA for ENT1, ENT2, and CNT3 in native undifferentiated HL-60 cells (Fig.  1A). None of the three inducers determined de novo expression of any other hNT (CNT1 and CNT2, not shown). We then comparatively determined by semiquantitative RT-PCR analysis the relative levels of ENT2 and CNT3 transcripts in undifferentiated HL-60 cells and in the same cells induced to differentiate by Me 2 SO, RA, and PMA, respectively. The PCR cycle numbers were optimized for each gene to ensure that comparison of the level of expression of each gene was within the linear phase of amplification. Furthermore, to ensure that an equal amount of RNA was used for each RT-PCR reaction, ␤-actin was used as an internal control. The observed PCR product intensity of CNT3 and ENT2 was normalized with the PCR product intensity observed for ␤-actin. The results obtained indicate that RA does not affect mRNA expression of CNT3 and ENT2 (Fig. 1B). On the contrary, the CNT3 mRNA is significantly up-regulated (ϳ6-fold) in the Me 2 SO-differentiated as well as in the PMA-differentiated HL-60 cells as compared with the native cells (Fig. 1B). No significant variations were recorded in the ENT1 transcript (ENT1 is not competent for cADPR uptake, Ref. 19) following differentiation with either Me 2 SO or PMA (not shown). In contrast, the expression of  4, and 6), respectively. Thereafter, RNA was reverse-transcribed and cDNA was subjected to PCR using gene-specific primers for CNT3 (lanes 1 and 2), ENT2 (lanes 3 and 4), and ␤-actin (lanes 5 and 6). Semiquantitative PCR reactions were performed as described under "Experimental Procedures." The intensity of PCR products was semiquantified by densitometric scanning analysis of the ethidium bromide-stained agarose gels using Chemi Doc System. Results are representative of three different experiments.
Cyclase Activity and cADPR Influx into Native and HL-60 Cells Differentiated with Me 2 SO, RA, and PMA-We next measured the uptake of extracellular cADPR in native HL-60 cells and in the cells differentiated with either Me 2 SO or RA or PMA (Fig. 2C). The levels of cell differentiation ( Fig. 2A) and of ectocellular GPD-ribosyl cyclase activity (as a measure of CD38 expression, Fig. 2B) were also analyzed. Influx of cADPR was markedly enhanced in the Me 2 SO-and PMA-differentiated, but not in the RA-differentiated HL-60 cells compared with the native cells. The slightly lower cADPR influx in the RA-treated HL-60 cells than in the native ones might be caused by partial hydrolysis of extracellular cADPR by the ectocellular cADPR hydrolase activity of de novo expressed CD38.
Together, the results shown in Figs. 1 and 2 demonstrate that differentiation of HL-60 cells by Me 2 SO and PMA, but not by RA, is paralleled by increased expression of CNT3 and that the underlying changes of NT expression are accompanied by enhanced cADPR uptake. The causal role of CD38 in granulocytic differentiation of RA-treated HL-60 cells (24) prompted us to investigate cADPR influx further in the Me 2 SO-differentiated rather than in the PMA-treated, macrophage-differentiated, HL-60 cells.
Characterization of cADPR Influx into Native and Me 2 SOdifferentiated HL-60 Cells- Fig. 3 shows the uptake of extracellular cADPR (50 M) by the native HL-60 cells and, by comparison, the Me 2 SO-differentiated cells. A fast influx of cADPR was observed in both cell populations, yet to a markedly greater extent in the differentiated than in native HL-60 cells. In both types of cells, influx of cADPR was competitively inhibited by 16 mM guanosine or by 50 mM inosine (not shown), confirming the involvement of nucleoside transporters in the cADPR uptake as previously observed in murine 3T3 fibroblasts (19).
To establish the roles of equilibrative versus concentrative nucleoside transporters in the cADPR uptake by native and Me 2 SO-differentiated cells, we measured the effects of various reagents and conditions known to affect the specific processes of nucleoside transport. In particular, we investigated the effects of: (i) replacement of Na ϩ with Li ϩ in the medium (a condition that abrogates the nucleoside/Na ϩ concentrative symport processes), (ii) the addition of 50 M dipyridamole or of 100 M NBMPR (both inhibiting the equilibrative NT and also the concentrative NT cs and csg, Refs. [25][26][27], and (iii) the addition of 10 nM NBMPR (this concentration is able to inhibit only cs and csg, Ref. 26). The results shown in Fig. 4 demonstrate that in the undifferentiated HL-60 cells, the uptake of cADPR is not affected by replacement of Na ϩ with Li ϩ in the buffers, but is severely reduced by 50 M dipyridamole (100 M NBMPR, not indicated, afforded a comparable inhibition) and is not impaired by 10 nM NBMPR. These results, coupled with , demonstrate that influx of cADPR into the undifferentiated HL-60 cells takes place via an equilibrative mechanism mediated by ENT2.
In the Me 2 SO-differentiated cells, the cADPR influx was comparatively much more extensive (see also Fig. 3). Moreover, it was decreased by 85% upon replacing Na ϩ with Li ϩ in the buffers, thus indicating a substantial concentrative cADPR/ Na ϩ symport mechanism (Fig. 4). Dipyridamole (50 M) decreased the extent of cADPR influx by ϳ30%. NBMPR, at 10 nM, inhibited cADPR influx by 20%, supporting the view of partial cADPR translocation across the cs-csg system under these experimental conditions. These patterns allow us to conclude that in the Me 2 SO-differentiated HL-60 cells, the role of concentrative mechanisms in the total cADPR uptake largely prevails over that of the equilibrative system (identified with ENT2) accounting for ϳ15% of total transport activity (as measured by the residual transport occurring in Li ϩ buffers). Specifically, the concentrative cADPR/Na ϩ symport processes responsible for ϳ85% of the total cADPR uptake, measured at 100 M extracellular cADPR, seem to be mediated mostly by CNT3 and, to a comparatively lower extent, by the cs-csg system (Fig. 4).
Therefore, analysis of NT mRNAs (Fig. 1) and of the properties of cADPR uptake by HL-60 cells (Figs. 3 and 4) indicate that Me 2 SO-differentiated cells, expressing 6-fold more CNT3 mRNA than the native cells, have a readily measurable concentrative cADPR transport. Conversely, the native cells, expressing comparatively lower levels of CNT3 mRNA, do not show any detectable mechanism of concentrative cADPR translocation. Accordingly, the limited amount of CNT3 transcripts might fail to be translated into functionally detectable CNT3 protein in the undifferentiated HL-60 cells. A similar situation had been observed in HeLa cells, where no concentrative cADPR influx could be measured despite occurrence of low levels of CNT3 mRNA (19).  5B) from which estimates of the binding parameters were consistent with two discrete classes of sites: (i) a smaller fraction bound NBMPR with high affinity (B max ϭ 7.80 Ϯ 3.01 fmol/10 6 cells; K d ϭ 0.23 Ϯ 0.05 nM); (ii) a larger fraction showed low binding affinity, with a K d of 0.92 Ϯ 0.13, identical to that of the NBMPR binding sites in the undifferentiated cells, and a B max of 27.21 Ϯ 2.05 fmol/10 6 cells. This latter value is ϳ25% less than that estimated for the NBMPR binding sites in the undifferentiated cells. Binding parameters were determined from the mean values of K d and of B max calculated for each of five different experiments.

Binding of [H 3 ]NBMPR to Native and Me 2 SO-differentiated HL-60 Cells-
Together, these results suggest the occurrence, in native HL-60 cells, of a single class of NBMPR binding sites identifiable with the equilibrative, non cADPR-translocating nucleoside transporter ENT1. The calculated K d value (0.91 nM) of this class of sites is close to the range of 0.3-1 nM, which was previously estimated for NBMPR binding to erythrocyte ENT1 (39) and to recombinant hENT1 expressed in Xenopus laevis oocytes (40). Conversely, the Me 2 SO-differentiated cells still express the ENT1 transporter (K d ϭ 0.92 nM), although at a reduced density (25% less) of NBMPR binding sites, and in addition a new population of binding sites with high affinity for NBMPR (K d ϭ 0.23 nM), which seems to be coincident with the cs-csg transporter system (25)(26)(27).
Roles of Specific Nucleoside Transporters in Mediating cADPR Uptake by Native and Me 2 SO-differentiated HL-60 Cells-The sharply different patterns of expression of cADPR-

FIG. 4. Effect of inhibitors of nucleoside transporters on cADPR influx in HL-60 cells.
Both native (white columns) and Me 2 SO-differentiated cells (black columns) were resuspended in Na ϩ buffer (total transport) or in Li ϩ buffer (equilibrative transport) at 10 7 cells/ml. Cells resuspended in Na ϩ buffer were preincubated for 15 min at 22°C in the presence of either 50 M dipyridamole (to measure concentrative, dipyridamole-insensitive transport) or 10 nM NBMPR (to measure equilibrative and concentrative, NBMPR-insensitive transport). Exposure to 100 M extracellular cADPR was performed for 5 min at 22°C. cADPR content was measured by the enzymatic cyclic assay translocating NT in the native and in the Me 2 SO-differentiated-induced HL-60 cells, respectively, prompted us to undergo kinetic experiments aiming at dissecting the specific roles of dipyridamole/NBMPR-sensitive cADPR transporters (ENT2 and cs-csg) from that of dipyridamole/NBMPR-insensitive CNT3. The results of these experiments are displayed in Fig. 6 and in Table I.
The concentration-dependent patterns of cADPR uptake by native, undifferentiated, and by Me 2 SO-differentiated cells are comparatively shown in Fig. 6A. In the undifferentiated cells, the extent of cADPR association recorded at low extracellular concentrations (50 nM to 5 M) was hardly compatible with true influx of the cyclic nucleotide. This became apparent, instead, from cADPR concentrations above 10 -20 M, at which a saturable cADPR uptake could be observed and measured.
On the contrary, the Me 2 SO-differentiated HL-60 cells showed significant cADPR uptake, occurring in a saturable fashion, even in the lowest range of concentrations. This fraction of cADPR uptake, as illustrated in Fig. 6B (which indicates influx at the lowest concentrations of the extracellular cyclic nucleotide), was almost completely abolished by preincubation of the differentiated cells with 50 M dipyridamole. Therefore, under these conditions, the content of cADPR in the differentiated cells dropped down to almost undetectable levels observed with native cells (Fig. 6B). Since ENT2 cannot mediate any cADPR uptake in the low concentration range, and CNT3 is not inhibited by dipyridamole, the fraction of saturable cADPR transport, which is characterized by high affinity (with an apparent K m value of 200 nM, see Table I) and by susceptibility to dipyridamole inhibition, proved to be identifiable with the cs-csg NT. Fig. 6C illustrates the patterns of cADPR influx into the native and the Me 2 SO-differentiated HL-60 cells, respectively, measured at the highest concentrations of extracellular cADPR (up to 4 mM). In both cases, saturable kinetics was observed. The kinetic parameters of cADPR transport in the native and in the differentiated HL-60 cells are summarized in Table I: the undifferentiated cells, expressing ENT2 only, had an apparent K m of 300 Ϯ 19 M and a V max of 18 Ϯ 2.1 pmol/mg/30 s. Conversely, the Me 2 SO-differentiated cells showed two sharply different patterns of cADPR influx: (i) the first pattern, observed at low extracellular cADPR (see Fig. 6B), characterized by an apparent K m (K m1 ) of 200 Ϯ 11 nM and a V max (V max1 ) of 10.32 Ϯ 1.63 pmol/mg/30 s; (ii) the second pattern, recorded at high cADPR (Fig. 6C) and characterized by a K m (K m2 ) of 250 Ϯ 22 M and by a V max (V max2 ) of 44 Ϯ 3 pmol/mg/30 s.
The affinity of ENT2 for cADPR consistently measured in the present experiments on the undifferentiated HL-60 cells (K m ϭ 300 M) was quite different from that indicated by previous experiments on HeLa cells (19) (K m 2-3 mM). Preliminary data seem to suggest that this difference could be attributed to a different extent of ENT2 glycosylation: indeed, surface deglycosylation of the native undifferentiated HL-60 cells, as obtained by incubation with the peptide N-glycosidase F (41), proved to decrease the affinity of cADPR transport.  The kinetic profiles recorded at high cADPR concentrations (Fig. 6C) result from the sum of individual cADPR-translocating systems represented by ENT2, CNT3, and cs-csg. Therefore, we tried to identify the kinetic properties of the fraction of cADPR influx contributed only by CNT3. To this purpose, either 50 M dipyridamole or 100 M NBMPR was used in an effort to inhibit both ENT2 and cs-csg. However, even with the native HL-60 cells (expressing ENT2 only), was the inhibition partial, probably as a result of competition of millimolar cADPR at the dipyridamole/NBMPR binding site that identifies ENT2. Therefore, the contribution of CNT3 to the total cADPR translocation occurring in the Me 2 SO-differentiated HL-60 cells could not be measured, and the kinetic parameters (K m2 and V max2 ) shown in Table I are indicative of a complex translocation process mediated by all NT expressed in the differentiated HL-60 cells. In any case, an important role of CNT3 in this global process is clearly indicated by the results shown in Fig. 4, at least when extracellular concentrations as high as 100 M were used.
Transient Expression of CNT3 in COS-7 cells-In an attempt to unequivocally define the involvement of CNT3 in the concentrative influx of cADPR, we performed transient transfection of COS-7 cells with hCNT3 cDNA. This was recovered from RNA of Me 2 SO-differentiated cells (see Fig. 1B). COS-7 cells are known to express both ENT1 and ENT2 and, in addition, low levels of CNT2 (19). Preliminary experiments confirmed these data, because [ 14 C]inosine influx was 80% inhibited by 10 M NBMPR and 20% inhibited by replacing Na ϩ with Li ϩ in the buffers, respectively (not shown).
RT-PCR analyses on pcDNA3.1/hCNT3-transfected cells showed a high level of CNT3 mRNA, whereas the control cells transfected with empty pcDNA3.1 plasmid were completely negative (Fig. 7A). The uptake of 50 M cADPR was then investigated in COS-7 cells transiently transfected with CNT3 cDNA in the presence of 100 M NBMBR to abrogate the equilibrative transport caused by ENT2 and to investigate only the concentrative processes involved therein. Results shown in Fig. 7B demonstrate that the extent of cADPR uptake by pcDNA3.1/CNT3-transfected cells increased about 5-fold over that measured in the control, pcDNA3.1-transfected cells (16 Ϯ 3 pmol of cADPR/mg of protein). Replacement of Na ϩ with Li ϩ in the buffer in the CNT3-transfected samples resulted in 87% inhibition of cADPR uptake, demonstrating the involvement of a concentrative Na ϩ -dependent transporter, identified as CNT3. In control cells transfected with empty pcDNA3.1, the same replacement of Na ϩ with Li ϩ induced only a very limited inhibition (10%) of cADPR influx, likely to reflect the small contribution of the concentrative transporter protein CNT2, which is weakly expressed in native COS-7 cells (19).
Me 2 SO-differentiated Cells, but Not Native HL-60 Cells, Accumulate Functionally Active cADPR When Grown in Co-culture over CD38 ϩ 3T3 Fibroblasts-Transwell co-culture of either CD38 Ϫ murine fibroblasts or of CD38 low human hemopoietic progenitors on CD38-transfected, cADPR-generating 3T3 fibroblasts, was previously shown to result in intracellular cADPR accumulation, in consequent [Ca 2ϩ ] i increases and in calcium-mediated enhanced cell proliferation (22,23). Therefore, in order to investigate the role of individual cADPRtranslocating NT in both native and Me 2 SO-differentiated HL-60 cells, we performed co-culture experiments using either CD38 Ϫ or CD38 ϩ 3T3 fibroblasts as feeders. The cADPR uptake, the [Ca 2ϩ ] i levels and the extent of differentiation were measured following a 48-h co-culture, as described under "Experimental Procedures." As illustrated in Fig. 8A, the native HL-60 cells did not undergo any significant increase of intracellular cADPR when grown on either the CD38 ϩ or the CD38 Ϫ feeders, nor was any modification afforded by preincubation with 1 M NBMPR. Failure to record any significant change in the cADPR influx (Fig. 8A) and in the [Ca 2ϩ ] i levels as well (Fig. 8B) indicated that ENT2, the only cADPR transporter expressed in the native cells, is not permeable to cADPR in the presence of the nanomolar concentrations of this cyclic nucleotide occurring in the co-culture media (19). This finding confirms the data shown in Fig. 6B.
The Me 2 SO-differentiated HL-60 cells exhibited a completely different behavior in terms of both cADPR uptake and of [Ca 2ϩ ] i . Thus, when grown on the CD38 ϩ 3T3 feeders, they were characterized by an extensive influx of cADPR, accounting for 4 -5-fold compared with the same cells grown on the CD38 Ϫ feeders (Fig. 8A). Comparable co-cultures of Me 2 SOdifferentiated HL-60 over CD38 ϩ 3T3 fibroblasts cells in the presence of 1 M NBMPR showed a significant decrease (60%) of intracellular cADPR. These results unequivocally demonstrate the extraordinary potency of the concentrative mechanism of cADPR influx that occurs in the Me 2 SO-differentiated HL-60 cells under the co-culture conditions. In addition, they indicate a substantial role of the cs-csg NT in the concentrative process occurring in this experimental setting of cell-to-cell interaction. Analysis of [Ca 2ϩ ] i levels (Fig. 8B) lanes 1 and 3) and pcDNA3.1/CNT3 (lanes 2 and 4)-transfected COS-7 cells was reversetranscribed, and cDNA was subjected to PCR using specific primers for hCNT3 (lanes 1 and 2) and for ␤-actin (lanes 3 and 4). B, uptake of 50 M extracellular cADPR by pcDNA3.1 (control) or pcDNA3.1/CNT3-transfected cells (1 ϫ 10 7 cells/ml) was measured following 5 min of incubation at 22°C in Na ϩ buffer (control, white columns) or Li ϩ buffer (black columns), in the presence of 100 M NBMPR in order to inhibit the equilibrative transport activity. Uptake of cADPR in the control was 16 Ϯ 3 pmol/mg. cADPR was measured in the perchloric acid-deproteinized cell extracts by the enzymatic cyclic assay (31). Values are means Ϯ S.D. of four different experiments. entiated HL-60 cells co-cultured on the CD38 Ϫ 3T3 feeder cells (p Ͻ 0.05). As observed for intracellular levels of cADPR, this increase of [Ca 2ϩ ] i proved to be prevented by presence of 1 M NBMPR in the co-culture medium (p Ͻ 0.05) (Fig. 8B). It is interesting to note, from comparative inspection of Figs. 4 and 8, that the specific contribution of the dipyridamole/NBMPRinhibitable fraction of concentrative cADPR influx to the total extent of uptake is inversely correlated to the concentrations of extracellular cADPR, i.e. higher at nanomolar concentrations in the medium and lower at 100 M cADPR.
As shown in Table II, the differentiation levels of native HL-60 cells co-cultured on CD38 ϩ 3T3 feeder cells, measured as NBT absorbance, did not show any increase compared with the same cells grown on CD38 Ϫ feeder cells. More importantly, the differentiation level reached by HL-60 cells stimulated for 5 days with Me 2 SO did not undergo any further modification following 48 h of co-culture with CD38 ϩ feeder cells, nor was it inhibited by the presence 1 M NBMPR in the co-culture medium.
In an attempt to elucidate whether the enhanced expression of the concentrative NT plays a causal role in mediating differentiation, native HL-60 cells were exposed for 5 days to 10 M extracellular cADPR under the same conditions used to promote Me 2 SO-induced differentiation (see "Experimental Procedures"). No increase of NBT absorbance was recorded in these conditions, nor was extracellular cADPR able to enhance the Me 2 SO-induced differentiation of native HL-60 cells (not shown). Accordingly, the increased uptake of extracellular cADPR that accompanies the Me 2 SO-induced HL-60 cells differentiation is a consequence, rather than a determinant, of the differentiation process. DISCUSSION The results of this and of a previous study conducted on 3T3 murine fibroblasts (19) indicate that cell types lacking CD38 or any other ADP-ribosyl cyclase can still feature cADPR-mediated calcium responses, even in physiological conditions, provided that they express concentrative nucleoside transporters competent for cADPR influx and RyR-controlled calcium stores. This view is supported by the occurrence of nanomolar or even lower concentrations of cADPR in blood plasma and extracel-lular fluids (19). Comparably low concentrations are present in co-culture media conditioned by CD38 ϩ 3T3 cell feeders, and yet they are active in inducing cADPR-mediated [Ca 2ϩ ] i increases and calcium-stimulated proliferation in constitutive CD38 Ϫ cells (15).
Our present findings point to the requirement of a threshold of expression of cADPR-translocating CNT in order for a cell to be able to take up this cyclic nucleotide against a concentration gradient. (The equilibrative ENT2, although competent for cADPR transport, cannot do so at minute extracellular concentrations of cADPR, Ref. 19.) More specifically, very low levels of CNT3, as those expressed by HeLa cells, were apparently inadequate to mediate cADPR influx (19). CNT3 proved in fact to be responsible for measurable cADPR transport, which clearly correlated with the extent of expression of CNT3 itself. Indeed, a significantly increased transport activity was observed both in the Me 2 SO-differentiated versus the native HL-60 cells (Fig.  4) and in the transiently transfected versus the control COS-7 cells as well (Fig. 7).
In addition to the levels of cADPR-translocating concentrative NT, a critical parameter that might be physiologically significant in determining increases of [Ca 2ϩ ] i in cADPR-responsive cells is certainly the K m for cADPR. Although the relative affinities of CNT3 and cs-csg NT toward cADPR escaped determination (see "Results"), the specific contribution of the latter CNT system in taking up the minute cADPR concentrations that are found in the co-culture media is clearly indicated by the remarkable inhibition by NBMPR that was observed in this experimental setting (Fig. 8). On the other hand, a major role of cs-csg in the uptake of extracellular cADPR was also observed in mixed co-cultures, in which NBMPR completely prevented the expansion of human hemopoietic progenitors when these were grown over CD38-transfected 3T3 fibroblasts. 2 Therefore, the high potential of the transwell coculture system in providing information on the paracrine mechanisms of NAD ϩ /cADPR signaling is re-established. In particular, this experimental setting may mimic the situation occurring at sites of inflammation. Here, several types of CD38 ϩ cells are recruited (42) and can generate high amounts of extracellular cADPR, with consequently increased influx into CNT-expressing neutrophils. Altogether, the present data indicate that the cs-csg system can potentially surrogate the absence of ADP-ribosyl cyclases in triggering cADPR-dependent [Ca 2ϩ ] i increases and up-modulate specific calcium-dependent functions, at least in some types of cells.
The human promyelocytic leukemia cells line HL-60 is a well known target of the cADPR/calcium system (24,29,36,43). Failure of native HL-60 cells to express either CD38 or concentrative cADPR transporters contrasts with the properties of terminally differentiated cell types induced by a number of different agents. Among these, granulocytic differentiation of HL-60 cells induced by retinoic acid is accompanied by a massive expression of CD38 and several lines of evidence support the view that the time-dependent accumulation of intracellular cADPR is causally responsible for this cellular differentiation (24). In contrast to RA, Me 2 SO does not elicit any CD38 expression. Instead, it triggers overexpression of concentrative cADPR-translocating NT, CNT3, and cs-csg, which may take up extracellular cADPR and eventually accumulate it in the Me 2 SO-differentiated HL-60 cells.
The presence of functionally active cADPR concentrations in granulocytes is a substantial prerequisite for important functions of these cells to take place. In fact, severely decreased levels of intracellular cADPR in CD38 knockout mice have been shown to be compatible with normality of neutrophils, unless they are required to develop chemotactic responses to bacterialderived peptides (44). Specifically, the cADPR-negative neutrophils of CD38 knockout mice failed to be recruited at sites of bacterial infection with Streptococcus pneumoniae. In addition, cADPR was demonstrated to be required in normal granulocytes to trigger calcium release from RyR-gated stores and to induce subsequent capacitative calcium influx from an external medium; both processes being causally involved in the mechanisms of chemotaxis (44). Unfortunately, Me 2 SO-differentiated HL-60 cells are not suitable to the development of an in vivo model, e.g. immunodeficient NOD/SCID mice, where engraftment of human cells takes place. The reason for this inadequacy is the limited survival of the Me 2 SO-differentiated HL-60 cells in culture, i.e. 5-6 days (not shown). Therefore, HL-60 cells induced to differentiate into granulocytes, irrespective of the inducing agent, cannot provide clues for the role of cADPR in the differentiation and in the physiology of neutrophils in vivo (24), where several redundant mechanisms may be involved: (i) CD38 itself, (ii) other ADP-ribosyl cyclases, e.g. some soluble forms thereof, whose occurrence is suggested by detectable activities in some organs and tissues from CD38 knockout mice (44), as well as by direct evidence obtained in T lymphoid cells (45) and in peripheral blood mononuclear cells (46); (iii) processes of concentrative cADPR influx as those reported in this study. Although with these limitations, the Me 2 SO-differentiated HL-60 cells seem to represent a good model system for in vitro studies aimed at elucidating cADPR-related physiological functions of neutrophils and especially mechanisms underlying their role in innate immunity.