Evidence for a Causal Role of CD38 Expression in Granulocytic Differentiation of Human HL-60 Cells*

Granulocytic differentiation of human HL-60 cells can be induced by retinoic acid and is accompanied by a massive expression of CD38, a multi-functional enzyme responsible for metabolizing cyclic ADP-ribose (cADPR), a Ca2+messenger. Immunofluorescence staining showed that CD38 was expressed not only on the surface of intact HL-60 cells but also intracellularly, which was revealed after permeabilization with Triton. Concomitant with CD38 expression was the accumulation of cADPR, and both time courses preceded the onset of differentiation, suggesting a causal role for CD38. Consistently, treatment of HL-60 cells with a permeant inhibitor of CD38, nicotinamide, inhibited both the CD38 activity and differentiation. More specific blockage of CD38 expression was achieved by using morpholino antisense oligonucleotides targeting its mRNA, which produced a corresponding inhibition of differentiation as well. Similar inhibitory effects were observed when CD38 expression was reduced by the RNA interference technique targeting two separate regions of the coding sequence of CD38. Further support came from transfecting HL-60 cells with a Tet-On expression vector containing a full-length CD38. Subsequent treatments with doxycycline induced both CD38 expression and differentiation in the absence of retinoic acid. These results provide the first evidence that CD38 expression and the consequential accumulation of cADPR play a causal role in mediating cellular differentiation.

CD38, first defined by monoclonal antibody typing as an antigen (1), has been widely used as a marker for lymphocyte differentiation. Sequence comparison reveals that it shares about 30% sequence identity with the Aplysia ADP-ribosyl cyclase, indicating that it is a mammalian homolog (2). This is later confirmed by studies showing CD38 can indeed catalyze the cyclization of NAD to produce cyclic ADP-ribose (cADPR) 1 (reviewed in Refs. 3 and 4), a cyclic nucleotide messenger active in mediating Ca 2ϩ signaling in a wide variety of cells spanning three biological kingdoms: protist, plant, and animal (Refs. 5-7; reviewed in Refs. 8 and 9). More remarkably, CD38 is, in fact, a multi-functional enzyme capable of using a different sub-strate, NADP, to catalyze a base exchange reaction (10) to produce nicotinic acid adenine dinucleotide phosphate, another general Ca 2ϩ messenger with a totally distinct structure and a separate mechanism of action (Ref. 11; reviewed in Refs. 9 and 12). It is now known that CD38 is not specific for lymphocytes but is ubiquitously expressed in many tissues and cells (reviewed in Ref. 3). The physiological functions that involve CD38 are equally widespread and include, for example, cell proliferation (13) and expansion of human hemopoietic progenitors (14,15). CD38 knockout mice exhibit defects in neutrophil chemotaxis (16), insulin secretion (reviewed in Ref. 17), and aberrant muscarinic Ca 2ϩ signaling in pancreatic acinar cells (18), indicating the importance of CD38 in regulating functions in vivo as well as in vitro.
A dramatic increase in CD38 expression accompanies granulocytic differentiation induced by retinoic acid in human HL-60 cells (19 -21). The cell line is derived from a patient with acute promyelocytic leukemia and can be induced to differentiate in vitro to a number of different cell types, such as granulocytes, monocytes, or macrophages (22). It is a widely used model system for elucidating hemopoietic differentiation (22,23). In this study we present evidence that the CD38 expression may play a causal role in mediating the differentiation process in HL-60 cells.
Culture of HL-60 Cells and Induction of Differentiation-HL-60 cells were obtained from the American Type Culture Collection. The cells were maintained in suspension in RPMI medium supplemented with 10% fetal calf bovine serum and kept at 37°C in a 5% CO 2 atmosphere. The cells were passaged by dilution in fresh medium to a density of about 0.2 ϫ 10 6 cells/ml. Prior to induction of differentiation by retinoic acid (RA), the cells were maintained at a logarithmic growth rate and seeded at a density of 0.2 ϫ 10 6 cells/ml. RA was added at a final concentration of 1 M by dilution from a 10 mM stock solution prepared in Me 2 SO. Control cells were treated with a similar dilution of Me 2 SO, which was found to have no effect on the differentiation or the rate of cell division. At the indicated times of continuous exposure to RA, the cells were pelleted by centrifugation at 700 ϫ g for 5 min. Differentiation of HL-60 cells was measured by adding 1 ml of cell suspension (0.5-2 ϫ 10 6 cells) to a solution containing 2 mg/ml of NBT and 20 ng/ml of phorbol myristate acetate in phosphate-buffered saline. The incubation was allowed to proceed for 1 h at 37°C and was stopped by the addition of 0.4 ml of cold 2 M HCl. The formazan product was obtained by centrifugation of the sample at 700 ϫ g for 10 min. The supernatant was discarded, and the formazan was dissolved in 1 ml of Me 2 SO. The absorbance of the solution was measured at 590 nm. The data are expressed as absorbance units/10 6 cells.
Fluorescence-activated Cell Sorter (FACS) Analyses of CD38 Expression-Following treatment of HL-60 cells with RA, 1.5-2 ϫ 10 6 cells were pelleted by centrifugation at 700 ϫ g for 5 min. The cells were resuspended in 1 ml of FACS buffer (phosphate-buffered saline containing 2.5% fetal bovine serum and 0.02% NaN 3 ) and washed once by a centrifugation step at 14,000 ϫ g for 20 s in a microcentrifuge and resuspended in 125 l of FACS buffer. The cells were incubated with the primary monoclonal antibody, IB4 (1:100), for 30 min on ice and washed once in 1 ml of FACS buffer and resuspended in 200 l. The secondary antibody, Alexa Fluor goat anti-mouse IgG, was added at a dilution of 1:200, and the sample was incubated on ice after mixing and kept in the dark to avoid bleaching. The cells were then washed twice and resuspended in 1 ml of FACS buffer containing 1% paraformaldehyde. The samples were sorted on a FACScalibur instrument, and data from 10,000 cells were collected and analyzed by the CELLQuest Pro software.
Measurements of Endogenous cADPR-Acid extracts were prepared for 5-20 ϫ 10 6 cells after centrifugation and the addition of 0.5-1 ml of cold 0.6 M perchloric acid, which could be stored at Ϫ80°C until processing. After thawing, the acid was removed by extraction with a solution (3:1) of 1,1,2-trichlorotrifluoroethane and tri-n-octylamine on ice as described (25). The neutral extracts were supplemented with 20 mM sodium phosphate, pH 8, and treated with an enzyme mixture containing 0.44 unit/ml nucleotide pyrophosphatase, 0.0625 unit/ml NADase, 12.5 units/ml alkaline phosphatase, and 2 mM MgCl 2 for 15-18 h at 37°C as described (25), to effectively remove interfering nucleotides, such as NAD, without degrading cADPR. Subsequently, the enzymes were removed by ultrafiltration with Centricon filters or 96-well Immobilon-P plates. The cADPR in the extracts was measured by an enzyme cycling assay as described previously (25). For comparison, cADPR standards were prepared in 20 mM sodium phosphate, pH 8, and processed in a manner identical to that of the cell extracts. This parallel processing of cADPR standards allowed for adjustment of losses of cADPR resulting from the procedure, which was typically about 30%. The enzyme cycling assay was performed in 96-or 384-well plates, using sample volumes of 100 or 40 l, respectively. The cADPR content was determined from the slopes of fluorescence increase of samples and compared with those produced by cADPR standards (25). The assay of cADPR was linear in the concentration range from 0 to 25 nM. The cADPR content was expressed in pmol/10 6 cells. NAD content of extracts was measured by the cycling assay before enzyme treatment.
Measurement of the Enzymatic Activity of CD38 -Extracts were prepared by centrifugation of 2.5-10 ϫ 10 6 cells and resuspension in 0.5 ml of a cold solution containing 10 mM Tris-Cl, pH 8, and 0.1 mM phenylmethylsulfonyl fluoride. These hypotonic cell extracts could be stored at Ϫ80°C. For the enzyme assay, 20 l of the thawed extract was added to a 200-l reaction mixture containing 20 mM Tris, pH 8, and 50 M NGD. The production of the fluorescent cGDPR was measured fluorimetrically (excitation wavelength, 300 nm; emission wavelength, 405 nm) using a fluorescence plate reader and calibrated by with known concentrations of cGDPR (21,26). The reactions were measured over several hours at 25°C, and the rates of cGDPR production were determined from the slopes of fluorescence increase and expressed as nmol of cGDPR formed per 10 6 cells.
Inhibition of CD38 Expression by Antisense Oligonucleotides-Morpholino antisense oligonucleotides have been recently designed to overcome some of the known limitations of regular antisense oligonucleotides. Morpholinos are assembled from four subunits, each of which contains one of the four bases linked to a six-membered morpholine ring. The subunits are joined in a specific order by nonionic phosphorodiamidate linkages. Applications of morpholino antisense oligonucleotides in different species indicate that it has improved specificity and stability against nucleases (reviewed in Refs. 27 and 28).
Morpholino antisense and sense oligonucleotides were synthesized by Gene Tools (Philomath, OR) to target a sequence residing in the 69 bases of the 5Ј cap of human CD38 mRNA (GenBank TM accession number M34461) (29). We selected the sequence 5Ј-GGTTGGCTGGGC-GAAGATGAGGC-3Ј, which starts 37 bases upstream of AUG and 32 bases into the 5Ј cap. The sequence has minimal secondary structure and the lowest self-complementarity, based upon the GC content. The invert of the antisense (sense) (5Ј-GCCTCATCTTCGCCCAGCCAACC-3Ј) was used as the control. A short stretch of each morpholino oligonucleotide (5 bases at the 3Ј end) was paired to complementary DNA, and an extra 10 bases of the DNA was added to form a 5Ј overhang. The DNA acts as an "adaptor" and binds electrostatically to the delivery reagent, a weakly basic ethoxylated polyethylenimine reagent (Gene Tools). Equimolar concentrations (1.4 M) of the oligonucleotides and ethoxylated polyethylenimine were first combined and incubated at room temperature for 20 min, following which, serum-free RPMI medium (Invitrogen) was added. HL60 cells were suspended in this solution at 1.0 ϫ 10 6 /ml and incubated at 37°C in a 10% CO 2 atmosphere for 3 h. The cells were then centrifuged at 3,000 rpm for 5 min, resuspended in RPMI with 10% serum at 0.2 ϫ 10 6 /ml, and returned to the incubator. Following a 24-h treatment with the oligonucleotides, RA (1.0 M, Sigma) was added, and the cells were harvested for analysis 3 days later.
Silencing of CD38 by Small Interfering RNA (siRNA)-Down-regulation of CD38 was facilitated by using the following primers targeting two separate coding regions of CD38 mRNA: region 3 sense primer, 5Ј-CTCTGTCTTGGCGTCAGTATTcctgtctc-3Ј; region 3 antisense primer, 5Ј-TACTGACGCCAAGACAGAGTTcctgtctc-3Ј; region 27 sense primer, 5Ј-AGGACTGCAGCAACAACCCTTcctgtctc-3Ј; and region 27 antisense primer, 5Ј-GGGTTGTTGCTGCAGTCCTTTcctgtctc-3Ј. The two regions start 70 and 533 bases downstream of ATG, respectively. Bases indicated by capital letters correspond to the region in the CD38 mRNA, while the 8-nucleotide stretch at the 3Ј end, in lowercase letters, is required for the T7 promoter primer sequence, 5Ј-TAATACGACTC-ACTATAGgagacagg-3Ј, to hybridize to the sense and antisense primers for transcription. The two thymidines, in italics, are needed for the stability of the siRNA (30). The HiScribe RNAi transcription kit from New England BioLabs was used to synthesize the double-stranded siRNA. Following synthesis, the siRNA was purified twice by ethanol precipitation and dissolved in sterile RNase-free water.
Transfection of HL60 cells was facilitated by the TransIT-TKO Transfection Reagent (Mirus Corporation, Madison, WI). 24 l of the TKO reagent was first incubated with 100 l of OPTIMEM-I medium (Invitrogen) at room temperature for 15 min before siRNA was added. After another 15 min at room temperature, 500 l of RPMI 1640 (with 10% calf serum) containing 0.8 ϫ 10 6 cells was added. The final siRNA concentration was 125 nM, and the incubation proceeded for 4 h at 37°C and 5% CO 2 , after which 3.5 ml of RPMI 1640 (with 10% calf serum) was added, and the incubation continued at 37°C for 24 h. RA was added to a final concentration of 1.0 M, and the cells were harvested 72 h later for analyses.
Tet-On Expression System for CD38 -The system was from Clontech and contains two components. The first is the pTet-On vector, which directs the expression of a regulatory protein, the reverse tetracyclincontrolled transactivator. The second is the pBI vector, which contains a bidirectional promoter, a tetracycline-responsive element flanked by two identical promoters in opposite orientations, allowing two genes of interest, CD38 and green fluorescent protein in our case, to be regulated by the tetracycline-responsive element. The inclusion of the green fluorescent protein was intended to facilitate monitoring of transfection and for selection of transfected clones.
Cationic liposomes were used for transfecting HL60 cells with the pTet-On plasmid. DC-Cholesterol, L-␣-dioleoyl phosphatidylethanolamine, and diphytanoyl phosphatidylethanolamine were purchased from Avanti Polar Lipids. Formulations used were 1:1 and 1:3 (molar ratio) of L-␣-dioleoyl phosphatidylethanolamine/DC-cholesterol and diphytanoyl phosphatidylethanolamine/DC-cholesterol. 5 mg of total lipid of each formulation were dried down and resuspended in 250 l of distilled H 2 O, followed by 5 min of sonication. 250 l of 2ϫ phosphatebuffered saline, pH. 7.4, was added, and the liposomes were further sonicated for 3 min. 5 l of the liposomal suspension was diluted in 50 l of OPTIMEM-I medium and incubated at room temperature for 5 min. 1.2 g of the ScaI-digested pTet-On plasmid was diluted in 50 l of OPTIMEM-I medium. The DNA and liposomal dilutions were combined in equal volumes and incubated at room temperature for 20 min. HL-60 cells (8 ϫ 10 4 ) in 500 l of OPTIMEM-I medium was added and incubated at 37°C for 4 h. Afterward, 5 ml of RPMI 1640 (with 10% fetal calf serum) was added, and the transfected cells were incubated at 37°C. After two cell divisions (48 h), the cells were resuspended in 5 ml of RPMI 1640 medium at a density of 2.0 ϫ 10 4 /ml with 380 g/ml G418 (Invitrogen), the selection antibiotic. The medium was replaced every 5 days. The cells stably expressing reverse tetracyclin-controlled transactivator were obtained after repeated selection with G418. Over the next 9 weeks the cells were passaged 12 times until there were virtually no dead cells in the cultures. Control experiments show that these stably transfected cells differentiate normally in response to RA. At that point a portion of the cells was frozen, and the remaining cells were used for the next transfection. All of the subsequent incubations were carried out in the presence of 380 g/ml G418 in the medium.
A pBI plasmid containing both green fluorescent protein and a full-length CD38. cDNA encoding for the full-length CD38 was inserted into the one of the two multiple of the pBI plasmid at the MluI and NheI restriction sites. The green fluorescent protein cDNA was spliced into the other cloning site at the SalI and PstI restriction sites. HL60 cells were co-transfected with the construct and the pTK-Hyg (Clontech) plasmid; the latter allowed positive selection using hygromycin. HL-60 cells containing the pTet-On vector were transfected with the construct using cationic liposomes as described above, and the positive clones were selected using 200 g/ml hygromycin.

RESULTS
Treatment of HL-60 cells with RA induces differentiation into granulocytes (22), which possess many of the functional characteristics of normal peripheral blood granulocytes, including phagocytosis and chemotaxis. The underlying mechanism is largely unknown. During phagocytosis, rapid generation of superoxide anion occurs, which can be conveniently monitored with NBT. It is a water-soluble dye, which is converted to insoluble intracellular blue formazan by phagocytizing neutrophils, a reaction mediated by superoxide (31,32). Differentiated cells that are phagocytizing are thus stained blue and black, whereas undifferentiated cells are not stained. The NBT reaction can also be monitored in cell suspensions by measuring the increase in absorbance at 590 nm. Differentiated cells produce greatly increased NBT reaction as compared with control cells. Either the absorbance changes or direct counting of NBT staining cells was used for quantifying granulocytic differentiation.
We and others have shown that accompanying differentiation, RA also induces expression of CD38 in HL-60 cells (19 -21), which can be conveniently measured by using FACS analyses or by measuring the ADP-ribosyl cyclase activity of CD38 in cell extracts using the NGD technique (21,26). CD38 cyclizes NGD, a nonfluorescent substrate analog of NAD, to cGDPR, a fluorescent product, which can be measured fluorimetrically.
It is generally believed that CD38 is an antigen and is mainly expressed on the cell surface. Fig. 1A shows immunofluorescence localization of CD38 in the differentiated cells. Intact cells (Ϫ Triton) showed ring-like immunostaining as revealed by confocal fluorescence microscopy, consistent with surface expression. Permeabilization with a detergent, Triton, before staining allowed internal access and resulted in even more intense staining that exhibited prominent intracellular structures (Fig. 1B). The RA-induced expression of CD38 is thus not limited to the cell surface but intracellularly throughout the cells as well, appropriate for a signaling role. This is supported by measuring the cellular accumulation of its enzymatic product, cADPR, as shown in Fig. 2. At various times after RA induction, aliquots of the culture were assayed for the enzymatic activity of CD38 ( Fig. 2A), cellular contents of cADPR (Fig. 2B), and differentiation (Fig. 2C). Parallel cultures without the RA treatment served as control. As seen, intracellular cADPR levels in the cells elevated progressively with a time course slightly lagging behind the appearance of CD38 activity (19). Both CD38 expression and cADPR accumulation, however, preceded cellular differentiation by more than 20 h, consistent with a causal role.
If this were the case, inhibition of CD38 expression should lead to inhibition of differentiation. Nicotinamide has been used as a cell-permeant inhibitor of the NAD cyclization activity of CD38. Mechanistically, nicotinamide actually forces the reverse of the cyclase reaction and produces NAD from cADPR (10,25,33). HL-60 cells were induced with RA and nicotinamide was added at 24 or 48 h afterward (indicated by arrows in Fig. 3), and endogenous cADPR levels were measured at 72 h after induction. As can be seen in Fig. 3A, nicotinamide produced partial inhibition of the CD38 activity. Because the cells were washed to remove nicotinamide before the extracts were  prepared, the observed inhibition actually reflected the reduction of CD38 expression. This was surprising, but the results were confirmed by using FACS analyses as shown in the inset in Fig. 3A. Treatment with nicotinamide (20 mM) during the RA induction resulted in most cells exhibiting less CD38 fluorescence.
The cellular cADPR levels exhibited similar changes to those of the CD38 activity as shown in Fig. 3B. The extent of reduction in cADPR levels appeared more pronounced than the CD38 activity, and the levels were reduced close to the basal levels of the control cells without the RA treatment. This is likely to be due to the combined effects of the inhibition of the cyclase activity of CD38 by nicotinamide as well as the actual reduction in CD38 expression in the cells. Parallel to the reduction in cADPR and CD38, there was corresponding inhibition of cellular differentiation as shown in Fig. 3C. The inhibitory effect of nicotinamide was specific for differentiation because the treated cells were not only viable throughout the 72 h of incubation but also proliferated equally well as compared with control cells not treated with the inhibitor. Furthermore, the NAD content of the cells actually doubled, from 0.35 Ϯ 0.05 to 0.70 Ϯ 0.01 nmol/10 6 cells, after 72 h of treatment with nicotinamide, indicative of the treated cell being in an energetically favorable state.
A common method for suppressing expression of a protein is to use antisense oligonucleotides. Morpholino oligonucleotides represent a recent improvement of the technique and offer better specificity and stability against nucleases than regular oligonucleotides (reviewed in Refs. 27 and 28). Fig. 4 shows that preincubation with the antisense oligonucleotides reduced both the expression of CD38 and cellular differentiation to levels similar to the control cells not treated with RA. Preincubation with the sense oligonucleotides or with just the carrier (ethoxylated polyethylenimine) affected neither the expression of CD38 nor cellular differentiation induced by RA. Similar to that described above for nicotinamide, HL-60 cells proliferated equally well in the presence of the antisense oligonucleotides as compared with control cells not treated, indicating that its inhibitory effect was specific for cellular differentiation.
In addition to antisense oligonucleotide, the RNAi technique has recently been used for the same purpose (reviewed in Ref. 34). Two separate regions (regions 3 and 27; see "Experimental Procedures") in the coding sequence of CD38 were targeted. The sequences of both regions are unique to CD38 as indicated by a sequence search. As shown in Fig. 5A, RNAi directed against region 27 produced about 40% inhibition of CD38 expression as compared with control cells treated with RA alone. Differentiation in these cells was also inhibited to a similar extent (Fig. 5B). RNAi directed against region 3 was less effective in inhibiting either the CD38 expression or differentiation. Mock incubation (Fig. 5B, TKO ϩ RA) had essentially no effect. Compared with the antisense oligonucleotide technique (Fig.  4), the RNAi method appeared to be less effective. The reason for this is unknown but may be related to the fact that the RNAi technique relies on the endogenous RNA degradation pathway being fully functioning, which may not be the case in HL-60 cells. Alternatively, the two targeted regions on the mRNA, which were selected because of their unique sequence, may not be optimal for degradation. The exact cause for the inefficiency of the RNAi was not investigated further. Nevertheless, there appeared to be a remarkably good correlation between the extent of inhibition of CD38 expression and differentiation (detailed below).
A more direct test for the causal role of CD38 is to enhance its expression artificially without activating the endogenous RA signaling pathway. This was achieved by transfecting HL-60 cells with the Tet-On expression construct containing full-length CD38. The expression system can then be activated by treating the transfected HL-60 cells with doxycycline. We first verified that doxycycline itself has no effect on either differentiation or CD38 expression of control HL-60 cells without the construct. Fig. 6 shows that in cells transfected with the construct, treatment with doxycycline induced expression of CD38 as shown by FACS analyses. This was confirmed by measuring CD38 enzymatic activity in cell extracts as compared with control cells not activated by RA (Fig. 6B, lower  panel). The extent of increase in CD38 activity induced by doxycycline is about 40% of the activity of either control or transfected cells treated with RA. It is clear that even though the artificial Tet-On expression system was not as effective as the natural system activated by RA, its activation by doxycycline was able to induce differentiation in the transfected cell to an extent similar to that observed for the increase in CD38 activity.
That there is a direct correlation between the extent of CD38 expression and cellular differentiation is shown more clearly in Fig. 7, where the results of the four different treatments described above are normalized to that induced by RA alone and plotted together. Thus, the antisense oligonucleotide treatment, which blocked CD38 expression most effectively, also inhibited differentiation most effectively. The regression line shown has an r 2 value of 0.976, close to perfect linearity. DISCUSSION In this study four different treatments were used to block CD38 in HL-60 cells, which include nicotinamide, a chemical inhibitor, an RNAi technique targeting two separate regions of the coding sequence of CD38, and antisense morpholino oligonucleotides targeting the 5Ј cap of the CD38 mRNA. All four treatments led to inhibition of granulocytic differentiation to varying degrees. Conversely, two different treatments that enhanced CD38 expression, namely, treatment with a natural inducer, RA, or artificially using doxycline to activate the Tet-On expression system, both led to induction of differentiation. In fact, the correlation between CD38 expression and differentiation was close to perfect as shown in Fig. 7. The results of the nicotinamide treatment were not included in the plot because the chemical has dual effects of not only inhibiting CD38 expression but also blocking its enzymatic activity. The regression line of the correlation extrapolated to an intercept of about 10% CD38 expression, suggesting that CD38 expression must surpass a threshold level before differentiation can be activated. That the nicotinamide treatment affected CD38 expression was unexpected but may suggest the existence of a positive feedback mechanism during the RA-induced differentiation, by which the increase in cADPR levels positively stimulated the expression.
The exact mechanism of how CD38 expression can induce differentiation remains to be elucidated but is likely to be related to the accumulation of cADPR, a Ca 2ϩ messenger, and an enzymatic product of CD38. Consistent with this notion is the time course measurements (Fig. 2) showing cADPR accumulation lagged slightly behind CD38 expression but preceded prominently cellular differentiation (19). Further support comes from the results of the nicotinamide treatment, which showed that the inhibition of differentiation correlated better with cADPR levels than with CD38 expression (Fig. 3). That cADPR level may be the causal factor is also consistent with the observation that a threshold of CD38 expression appeared to be required (Fig. 7), because it is likely that cADPR production must exceed degradation before it can accumulate and exert its signaling function.
A wide range of physiological functions have now been shown to be mediated by the Ca 2ϩ mobilizing activity of cADPR in cells spanning three biological kingdoms, from protist, to plant to animal (reviewed in Refs. 4, 8, and 9). Although the presence of cADPR-sensitive Ca 2ϩ stores in HL-60 cells has not yet been reported, it is likely to be present because a similar cell type, the mouse neutrophil, has been shown to possess them, the mobilization of which is involved in the chemotactic response of the cells (16). Also, HL-60 cells have been shown to possess Ca 2ϩ stores sensitive to nicotinic acid adenine dinucleotide phosphate (35), another Ca 2ϩ messenger that is also an enzymatic product of CD38 (10).
The fact that cADPR, the enzymatic product of CD38, accumulates inside HL-60 cells indicates that CD38 is expressed intracellularly. This was directly shown using immunofluorescence staining in this study (Fig. 1). The results are consistent with previous localization of CD38 to various intracellular organelles by immunoelectromicroscopy in brain (36) as well as to the nuclear envelope of cultured cells (37) and hepatocytes (38). It is thus not appropriate to think of CD38 as an antigen expressed solely on the cell surface, although it is of interest to note that even surface CD38 can exert its intracellularly signaling function by catalytically transporting cADPR into cells and also via the endocytic pathway (reviewed in Ref. 39).
Whether CD38 has a role in human neutrophil differentiation in vivo remains to be demonstrated. CD38 has already been shown to be important in mediating proliferation of HeLa cells (13), a human cell line, and expansion of human hemopoietic progenitors (14,15). On the other hand, neutrophils in CD38 knockout mice exhibit a severe defect in their chemotactic response to formyl peptides but otherwise appear to differentiate normally (16). When considering functions in vivo, the well documented presence of many compensatory mechanisms must also be considered. Indeed, ADP-ribosyl cyclase activity can still be detected in some tissues of the CD38 knockout mice. Also, cADPR contents in some tissues, especially in brain and heart, remain close to normal (16), suggesting the existence of a redundant cyclase present in the animal in addition to CD38. Another rather dramatic example is a recent report showing that the hormone, estrogen, in female mice can protect them from developing cardiac hypertrophy when their genes for the FK506 binding proteins are disrupted, whereas male knockout mice, lacking estrogen, exhibit a prominent defect (40). It is thus conceivable that some other compensatory mechanisms in the CD38 knockout mice may allow their neutrophils to develop.