Introduction

Top Summary Introduction Procedures Results Discussion References

P2 nucleotide receptors mediate the biological responses to extracellular ATP in a wide variety of tissues and cell types (Dubyak and El-Moatassim, 1994). The two major subgroups that compose the P2 receptor family include the metabotropic P2Y receptors and the seven different ionotropic P2X receptors (P2X_1-7). The similarities and differences among the P2X receptor subtypes have been reviewed (Buell et al., 1996; Evans et al., 1998). The most recently cloned P2X receptor is the P2X₇R that acts as rapidly gated cation channel in the presence of high concentrations of extracellular ATP. In contrast with the other six P2X receptor subtypes, the P2X₇R also forms a pore permeable to large organic molecules such as ethidium (342 Da) and YoPro-1 (629 Da). The long, cytoplasmic carboxyl-terminal tail of the P2X₇R is required for this permeability to large molecules, but not small cations, and this region is absent from the P2X_1-6 receptor subtypes (Surprenant et al., 1996, Rassendren et al., 1997).

The absence of high potency antagonists traditionally has limited the identification and characterization of P2 nucleotide receptors in cells and tissues. Recently, Gargett and Wiley (1997) described a novel class of P2 receptor inhibitors: the isoquinoline derivatives KN-62 [1-[N,O-bis(5-isoquinolinesulfonyl)-N-methyl-L-tyrosyl]-4-phenylpiperazine] and KN-04 [N-[1-[N-methyl-p-(5 isoquinolinesulfonyl)benzyl]-2-(4 phenylpiperazine)ethyl]-5-isoquinolinesulfonamide]. Submicromolar levels of these drugs inhibited the ability of ATP to stimulate cation influx, pore formation, phospholipase D activation, and shedding of L-selectin in human chronic lymphocytic leukemia cells. These ATP-dependent responses in lymphocytes are mediated by the P2Z receptor, a functionally defined P2 receptor that shares many properties with the cloned P2X₇R (Wiley et al., 1990; Gargett et al., 1996; Jamieson et al., 1996). Although KN-62 is a potent inhibitor of CamKII (Tokumitsu et al., 1990), Gargett and Wiley (1997) observed that KN-04, an inactive analog of KN-62, also blocked the P2Z responses in lymphocytes. This indicated that inhibition of the P2Z receptor by KN-62 is not mediated by CamKII. Previous studies have demonstrated that P2Z receptor signaling in nonlymphoid cell types also can be antagonized by isoquinolines. Blanchard et al. (1995) reported that KN-62 potently suppressed ATP-induced cytolysis of THP-1 monocytes and monocyte-derived human macrophages. We similarly demonstrated that KN-62 inhibits the ability of P2Z receptors to activate phospholipase D in THP-1 monocytes (Humphreys and Dubyak, 1996).

These various data indicate that isoquinolines are potent inhibitors of P2Z receptor signaling in human hematopoietic cell types. However, the molecular mechanism or mechanisms by which KN-62 inhibits P2Z receptor signaling remain undefined. Although expression of recombinant P2X₇R cDNA is sufficient to confer P2Z-like phenotypes to nonhematopoietic cells, it is not known whether signaling by native P2Z receptors also might involve cell-specific modulation of P2X₇R structure or function. For example, recombinant P2X₇R (human or rat) form pores that readily admit ethidium (342 Da) and the propidium dye Yo-Pro1 (629 Da). In contrast, the P2ZR pore in human leukemic lymphocytes passes ethidium but excludes propidium (414 Da) (Wiley et al., 1993). The P2ZR pore in murine thymocytes is even more restrictive given its low permeability to ethidium (Pizzo et al., 1991). The factors that underlie these altered permeation characteristics of some P2Z receptors are not known but could indicate the presence of altered subunit interactions between P2X₇Rs and other leukocyte proteins or unique post-translational regulation of the P2X₇R when expressed in hematopoietic cells. The multiexon organization of the human P2X₇R gene (GenBank accession numbers Y12851, Y12852, Y12853, Y12854, Y12855) raises the additional possibility that some cell types might express variant P2X₇ proteins encoded by alternatively spliced transcripts.

The current study was performed to determine the effect of isoquinoline derivatives on P2X₇R function in (1) cells expressing only full-length recombinant P2X₇Rs in the absence of other P2X family members and (2) various human and murine cell types that natively express both P2X₇R mRNA and P2Z-like phenotypes. To establish cell lines expressing homomeric P2X₇R, we stably transfected the rat and human P2X₇ cDNAs into HEK 293 cells that lack expression of P2X_1-7. Our results indicate that homomeric human P2X₇Rs are potently inhibited by the isoquinoline derivatives. However, we also identified a striking and unexpected difference in drug sensitivity between human and rat P2X₇Rs, and chimeric molecules were constructed to explore further the structural requirements for inhibition of huP2X₇ by the isoquinolines. Because detailed structure-function studies have not yet been performed on P2X₇Rs, little is known about the regions involved in nucleotide binding or channel/pore activation. Our results suggest that the isoquinoline derivatives will be useful pharmacological tools both for identifying THAT₇R function in human cell types and for investigating the structural determinants of THAT₇R activation.

	Experimental Procedures

Top Summary Introduction Procedures Results Discussion References

Materials. Cell culture medium was purchased from Sigma Chemie (Deisenhofen, Germany). KN-62 was from BIOMOL Research Laboratories (Plymouth Meeting, PA), and KN-04 was the generous gift of Dr. R. Ganapathi (Cleveland Clinic Foundation, Cleveland, OH). Lipofectamine was from GIBCO BRL (Gaithersburg, MD), and G418 was from Genentech (South San Francisco, CA). Recombinant human IFN- was a generous gift from Genentech. LPS was from List Biochemicals (Campbell, CA). Dr. G. Buell (Glaxo Institute for Molecular Biology, Geneva, Switzerland) kindly provided the human and rat THAT₇ cDNAs. BzATP was obtained from Sigma or synthesized as described previously (Williams and Coleman, 1982; Erb et al., 1990).

Cell culture. THP-1 cells (American Type Culture Collection, Rockville, MD) were cultured in Iscove's modified Dulbecco's medium plus 10% iron-supplemented newborn calf serum (Hyclone Laboratories, Logan, UT). At the initiation of culture, cell density was 3 × 10⁵/ml, and medium was changed when cells reached 1 × 10⁶/ml (every ~3 days). For experiments involving IFN- or LPS, cells were plated onto 100-mm culture dishes (10 ml) at a cell density of 1 × 10⁶/ml in fresh medium. IFN- (1000 units/ml) or LPS (1 µg/ml) was added from sterile stocks for 2 or 3 days, as indicated. The BAC1.2f5 macrophage cell line, a clone of the SV40 transformed murine macrophage cell line BAC1, was maintained using protocols described previously (El Moatassim and Dubyak, 1992). Murine BW5147 cells, a thymocyte line, were maintained in Dulbecco's modified Eagle's medium (Sigma) supplemented with penicillin (100 units/ml), streptomycin (100 µg/ml) and 10% iron-supplemented newborn calf serum (Hyclone). They were split every three days. HEK 293 cells (a generous gift from Dr. Cathy Carlin) were maintained in Dulbecco's minimal essential medium supplemented with 100 units/ml penicillin, 100 µg/ml streptomycin, and 10% iron-supplemented newborn calf serum (Hyclone). Stably transfected HEK lines were obtained using Lipofectamine according to the protocol supplied by the manufacturer (GIBCO BRL) or by electroporating 1 × 10⁷ cells with 20 µg of DNA at 300 mV and 500 µF with a time constant of 6-8 msec (Gene Pulser II; BioRad, Hercules, CA). G418 (19 mg/ml) was added 2 days after transfection, and >100 G418-resistant colonies were pooled 2 weeks later. The stably transfected cell lines were maintained in G418-containing medium.

Construction of P2X₇R chimeras. A HindIII/BsrG1 fragment corresponding to the 5' end of each cDNA was excised and subcloned into the opposite plasmid. This resulted in two chimeras, hu-ratP2X₇ (human 1-335, rat 336-595) and rat-huP2X₇ (rat 1-335, human 336-595). Sequencing of these recombinant receptor DNAs confirmed the expected chimeric sequences of nucleotides. Stably transfected HEK 293 cell lines were obtained by electroporation of 20 µg of DNA followed by selection in G418.

Intracellular Ca²⁺ concentration measurements. Intracellular Ca²⁺ concentration was measured using the fluorescent indicator Fura-2 in a thermostatically controlled and magnetically stirred fluorimeter exactly as described (El-Moatassim and Dubyak, 1992).

Changes in plasma membrane permeability. ATP or BzATP-activated pore function was measured with the fluorescent dye ethidium bromide. Cells (1.5 ml) of resuspended in divalent cation-free basal salt solution were incubated in the fluorimeter cuvette (37°) at a final concentration of 0.5 × 10⁶ cells/ml. Ethidium bromide was added to a final concentration of 20 µM. Cells were allowed to equilibrate for 10 min and then stimulated with nucleotide agonists, and the fluorescence was monitored at wavelengths of 360 nm excitation and 580 nm emission. In each assay, maximum cell permeabilization/ethidium uptake was defined as the fluorescence value achieved by adding digitonin (20 µg/ml) to the cuvette. Agonist-stimulated fluorescence increases were quantified as the fraction of maximum fluorescence attained 3 min after nucleotide addition. These measurements were expressed as percentages by normalizing to the maximal changes in fluorescence triggered by 1 mM ATP or 100 µM BzATP.

Electrophysiology. HEK 293 cells stably or transiently expressing the rat, human, or chimeric P2X₇R cDNAs were used in all electrophysiological experiments (Surprenant et al., 1996; Rassendren et al., 1997). Standard whole-cell recordings were obtained using an EPC9 patch-clamp amplifier and Pulse acquisition software (HEKA, Lambrecht, Germany). Patch pipettes (4-7 M) were filled with 165 mM NaCl, 10 mM HEPES, and 11 mM EGTA. Standard extracellular solution was 154 mM NaCl, 2 mM KCl, 1 mM MgCl₂, 2 mM CaCl₂, 10 mM HEPES, and 12 mM glucose. "Low divalent" solution contained no added MgCl₂ and 0.1 mM CaCl₂. All solutions were maintained at pH 7.3 and 300-310 mOsmol/liter. Agonists were applied using a fast-flow U-tube delivery system (Fenwick et al., 1982). Antagonists were present in both superfusion and fast-flow solutions. Results are shown as mean ± standard error.

Semiquantitative RT-PCR. Total RNA was isolated from THP-1, BAC1.2f5, and BW5147 cultures by the acid-guanidinium thiocyanate extraction method (Chomczynski and Satchi, 1987). RNA (1.0 µg) was reverse transcribed to cDNA in a 20-µl reaction volume containing 0.5 µg of oligo(dT) primer (Promega, Madison, WI), 40 units of RNasin (Boehringer-Mannheim, Mannheim, Germany), 10 mM MgCl₂, and 25 units of avian myeloblastosis virus reverse transcriptase (Boehringer-Mannheim) dissolved in reverse transcriptase buffer (Promega). The reactions were incubated for 1 hr at 42°, heated to 90° for 2 min, and diluted to 100 µl with RNase-free water. In each experiment, a parallel aliquot of RNA was subjected to mock reverse transcription as indicated above except avian myeloblastosis virus reverse transcriptase was left out. Diluted aliquots from these reactions were used as templates for PCR. The P2X₇ primers were from the rat cDNA sequence (sense 5'-GGCAGTTCAGGGAGGAATCATGG-3' and antisense 5'-AAAGCGCCAGGTGGCATAGCTC-3'), generating a 939-bp product. Commercial primers to human GAPDH (Stratagene, La Jolla, CA) or murine  actin (Clontech) also were used to generate 600- and 540-bp products, respectively. P2X₇R reactions included 1.0 µM concentration of each primer, 0.8 mM concentration of dNTPs, 60 mM Tris·HCl, pH 8.5, 15 mM (NH₄)₂SO₄, 3.5 mM MgCl₂, and 1.25 units of Taq polymerase (Boehringer-Mannheim) preincubated with 275 ng of TaqStart antibody (Clontech) for 5-30 min at room temperature. huGAPDH and murine  actin reactions included 1.0 µM concentration of each primer, 0.8 mM concentration of dNTPs, 10 mM Tris·HCl, pH 8.3, 50 mM KCl, 1.5 mM MgCl₂, 0.001% gelatin, and Taq polymerase pretreated as above. The PCR cycling protocol was as follows: 1 min at 94°, 2 min at 60°, and 2 min at 72°. This protocol was carried out for 35 cycles and included an initial 5-min denaturation at 94° and a final 7-min extension at 72°. Products were electrophoresed on 1% agarose gels containing ethidium bromide and photographed. In control experiments, standard curves were generated using serial dilutions of the RT reactions as templates for PCR with each primer set. The linear range of the assay thus was determined for each primer set, and products from the original 20-µl RT were diluted into the final PCR volume to ensure nonsaturation of the PCR amplification reactions. A 1:50 dilution of the RT reaction was used for P2X₇R analysis; a 1:500 dilution was used for huGAPDH and mu-actin.

Data presentation. Values are presented as average ± standard error, and each experiment was repeated two to five times with similar results.

	Results

Top Summary Introduction Procedures Results Discussion References

Functional analysis of P2X₇R pore activity by ethidium bromide uptake. Activation of the P2X₇R is accompanied by a dramatic increase in plasma membrane permeability to organic solutes up to 900 Da in mass. We used uptake of the fluorescent dye ethidium bromide as an index of flux through this gated pore. On activation of P2X₇Rs by ATP or BzATP, extracellular ethidium rapidly enters the cytoplasm and binds nucleic acids, resulting in increased fluorescence. These experiments were performed in the absence of extracellular divalent cations that inhibit P2X₇/P2Z receptor function by mechanisms involving reduction in the concentration of free ATP⁴ (Steinberg et al., 1987), reduction in receptor affinity for ATP (Virginio et al., 1997), or direct blockade of permeation (Surprenant et al., 1996). Neither ATP nor BzATP activated any ethidium influx in untransfected HEK 293 cells (data not shown). In cells stably transfected with either the human P2X₇ cDNA (HEK-hP2X₇) or the rat P2X₇ cDNA (HEK-rP2X₇), ATP elicited a dose-dependent increase in the rate of ethidium influx (Fig. 1, A and B). Although the maximal rate of ATP-induced ethidium influx for HEK-rP2X₇ was faster than that in HEK-hP2X₇, the ATP dose responses were identical (Fig. 1C).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Functional characterization of HEK 293 cells stably transfected with human or rat P2X7 receptor. A, Effects of increasing ATP concentration on ethidium influx into HEK-huP2X7 cells. Ethidium influx studies were performed in divalent free-basal salt solution as described in Experimental Procedures. Digitonin (20 µg/ml) was added to cell suspensions after 3 min to determine maximal ethidium accumulation as indicated. B, Effects of increasing ATP concentration on ethidium influx into HEK-ratP2X7 cells. C, Concentration-response relationships describing ATP-induced ethidium accumulation in HEK-huP2X7 () and HEK-ratP2X7 (). Ethidium influx at 3 min after the indicated [ATP] was measured and normalized to the maximal ATP-induced accumulation of ethidium. Values represent mean ± standard error of triplicate samples.

Effect of KN-62 on ATP-mediated ethidium influx in HEK-huP2X₇ cells versus HEK-ratP2X₇ cells. Because HEK 293 cells do not natively express any of the known P2X receptor subtypes (Evans et al., 1998), the P2X₇R will be expressed as a homomeric receptor in this cell type. We tested the ability of KN-62, a CamKII antagonist, to inhibit homomeric P2X₇R function. Fig. 2A shows that increasing concentrations of KN-62 in the nanomolar range inhibited the ethidium influx induced by 1 mM ATP in the HEK-hP2X₇cells. Unexpectedly, even a high concentration of KN-62 (3 µM) had no detectable effect on ATP-mediated ethidium uptake by the HEK-rP2X₇ cells (Fig. 2B). Stimulating the HEK-rP2X₇ cells with submaximally active concentrations of ATP failed to unmask inhibition by KN-62 (data not shown). Because the surface protein expression of HEK-rP2X₇ may differ from that of HEK-hP2X₇, we reduced the number of functional cell surface rat P2X₇R by partial inhibition with oxidized ATP (oATP), a drug that irreversibly inactivates the P2X₇R. A 1-hr incubation of HEK-rP2X₇ cells with 100 µM oATP (followed by removal of unreacted oATP) reduced the magnitude of ATP-induced ethidium influx by about half, but these cells remained KN-62 insensitive (data not shown). A longer incubation with oATP resulted in greater inhibition of rat P2X₇R activity but no increase in relative KN-62 sensitivity. Fig. 2C shows the concentration-response relationships that characterize these isoquinoline effects; KN-62 has an IC₅₀ value of ~100 nM against the human P2X₇R but no inhibitory activity against the rat P2X₇R.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Effect of the isoquinoline derivative KN-62 on ATP-evoked ethidium influx in HEK-huP2X7 and HEK-ratP2X7. A, HEK-huP2X7 cells were challenged with 1 mM ATP in the absence or presence of the indicated concentrations of KN-62 and ethidium fluorescence measured. KN-62 was added 10 min before the addition of ATP. B, HEK-ratP2X7 cells were challenged with 1 mM ATP in the absence or presence of 3 µM KN-62. C, Ethidium influx at 3 min after ATP was measured and normalized to the maximal ATP-induced accumulation assayed in the absence of isoquinoline. The data points were obtained from HEK-huP2X7 () or HEK ratP2X7 () over a range of KN-62 concentrations (10-3000 nM). Values represent mean ± standard error of triplicate samples.

KN-62 partially antagonizes BzATP-induced ethidium uptake in HEK-hP2X₇ cells. Because BzATP is the most potent and selective agonist for P2X₇/P2Z receptors, we examined the ability of KN-62 to inhibit BzATP-mediated ethidium uptake in the HEK-hP2X₇ cells. BzATP (100 µM) maximally stimulated ethidium influx (data not shown), and KN-62 markedly inhibited this influx. However, KN-62 was less efficacious as an inhibitor of BzATP-induced ethidium uptake than of ATP-induced ethidium uptake. KN-62 (3 µM) inhibited BzATP-stimulated ethidium influx by 61.6 ± 1.7% (three experiments; Fig. 3, A and B) compared with 87.5 ± 1.8% (three experiments) inhibition for ATP-stimulated influx (Fig. 2, A and C). The inhibitory effect of KN-62 was most apparent during the first 10 sec after BzATP application; the initial rate of ethidium influx was slowed greatly but then increased. Thus, at these earlier time points, inhibition by KN-62 was more dramatic than that after several minutes of BzATP stimulation (Fig. 3A). Gargett et al. (1997) also observed that KN-62 is less efficacious as an antagonist of BzATP-activated versus ATP-activated P2Z receptors in human CLL lymphocytes.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Inhibition of BzATP-mediated ethidium influx by KN-62 in HEK-huP2X7. A, Ethidium influx in response to 100 µM BzATP recorded from HEK-huP2X7 in the absence or presence of KN-62 (10-3000 nM). Cell suspensions were challenged with BzATP after a 10-min preincubation in the presence of KN-62. B, Concentration-response relationship for inhibition of BzATP-mediated ethidium uptake by KN-62. Ethidium influx at 3 min after BzATP was measured and normalized to the maximal BzATP-induced accumulation assayed in the absence of isoquinoline. Values represent mean ± standard error of triplicate samples.

Effect of KN-62 and KN-04 on ATP and BzATPinduced ionic currents. The P2X₇/P2ZR can exist in two functionally distinct states. The first is a channel that opens rapidly in response to agonist and passes small inorganic cations such as Na⁺ and Ca²⁺ but not larger molecules. With prolonged or repetitive nucleotide application, the receptor forms the pore capable of passing the larger organic ions such as N-methyl-glucamine, ethidium, and Yo-Pro (Nuttle et al., 1993, Surprenant et al., 1996). Gargett and Wiley (1997) observed that isoquinoline derivatives inhibited the ability of ATP to activate influx of both Ca²⁺ and ethidium in CLL lymphocytes. This indicated that these compounds should inhibit both functional states of the human but not the rat P2X₇R. This was confirmed by electrophysiological analysis of whole-cell ionic currents recorded from HEK 293 cells that stably express either of these recombinant receptors. As reported previously (Surprenant et al., 1996; Rassendren et al., 1997), BzATP (1-300 µM) and ATP (100-1000 µM) evoked rapid inward currents in HEK 293 cells transfected with either the rP2X₇ or hP2X₇ cDNAs (Fig. 4, A and B). The decay of inward current after a brief (1-2 sec) application of agonist was markedly slower at the rP2X₇R than at the hP2X₇R. KN-62 (10-1000 nM) was without effect on the BzATP or the ATP-evoked current recorded from rP2X₇ transfected cells (12 experiments). The reduction in the rate of decay of the inward current during repeated agonist pulses is observed in the presence or absence of KN-62 (Fig. 4B). This phenomenon was originally described by Surprenant et al.(1996) in their initial characterization of the cloned rP2X₇R. In contrast, KN-62 produced a dose-dependent inhibition of agonist-evoked currents in hP2X₇ transfected cells (Fig. 4C). The half-maximal concentrations of KN-62 required to inhibit the currents evoked by 100 µM BzATP or 1 mM ATP was 38 ± 1.4 nM (six experiments) and 34 ± 3 nM (four experiments), respectively. These latter currents were recorded from cells incubated in saline containing low concentrations of divalent cations. KN-62 exhibited a similar potency in inhibiting currents evoked by BzATP (300 µM) when the external solution contained normal divalent cations (i.e., 2 mM CaCl₂ and 1 mM MgCl₂; five experiments). The inhibition by KN-62 was only partially reversible after 10-min washout of the compound; for example, 300 nM KN-62 inhibited the BzATP-evoked current by 95 ± 2% (five experiments) within 4 min of application, and current had returned to 62 ± 4% of control amplitude 10 min after KN-62 was removed from the bath.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   KN-62 potently inhibits currents through the human P2X7 receptor but is ineffective at the rat P2X7 receptor. Representative currents recorded from HEK 293 cells expressing the human P2X7 (A) or rat P2X7 (B) receptor in response to a brief application of ATP (A) or BzATP (B) as indicated (bar above trace). A, Records are shown after 4 min in the presence of KN-62 at the indicated concentrations. B, Records are for before and during the continued presence of 1 µM KN-62 for the indicated durations. C, Concentration-response curve for inhibition of hP2X7-mediated currents induced by ATP (1 mM, ) or BzATP (100 µM, ). All results are from experiments in which low divalent cation external solutions were used. Values represent mean ± standard error of four to eight experiments.

Chimeric receptors localize KN-62 sensitivity to the amino-terminal half of huP2X₇. The striking difference in isoquinoline sensitivity between human and rat P2X₇Rs despite 80% shared amino acid identity suggested that KN-62 sensitivity might be localized to only a few amino acid residues of the human receptor homologue. Gargett and Wiley (1997) speculated that the inhibitory actions of KN-62 might be mediated by an interaction with the unique intracellular carboxyl-terminal domain of the P2X₇R. To address this possibility, we examined the isoquinoline sensitivity of HEK cells stably transfected with chimeric P2X₇Rs in which the carboxyl-terminal tails of human and rat receptors were exchanged. Fig. 5A shows that ATP-mediated ethidium uptake in cells expressing the human-rat P2X₇R construct (h-rP2X₇; human amino terminus, rat carboxyl terminus) was inhibited by 1 µM KN-62 (88.1 ± 1.2% inhibition, four experiments). In contrast, cells stably expressing the rat-human P2X₇ construct (r-hP2X₇; rat amino terminus, human carboxyl terminus) displayed ATP-stimulated ethidium uptake that was completely insensitive to KN-62 (Fig. 5B) or KN-04 (data not shown). Similar results were obtained when ATP or BzATP-evoked currents were measured in HEK cells expressing these chimeric receptors (Fig. 5C). The IC₅₀ value (40 nM) that characterized KN-62 inhibition of BzATP-activated current through the h-rP2X₇R was equivalent to the IC₅₀ observed in HEK cells expressing the full-length human P2X₇R (Fig. 4C). In contrast, even micromolar amounts of KN-62 failed to substantially inhibit current through the rat-human P2X₇R chimera. These experiments indicate that the first 335 amino acids of the hP2X₇R mediate isoquinoline sensitivity, ruling out a KN-62 binding site in the intracellular carboxyl-terminal domain of the receptor.

View larger version (13K): [in this window] [in a new window]   Fig. 5.   Chimeric P2X7R constructs localize KN-62 sensitivity to the amino-terminal 335 amino acids of the human P2X7R. A and B, P2X7R chimeras were constructed between human and rat P2X7 receptors and stably expressed in HEK 293 cells as described in Experimental Procedures. Ethidium influx was assayed in response to 1 mM ATP after a 10-min preincubation of the cells in absence or presence of 1 µM KN-62. Ethidium influx was inhibited by 88.1 + 1.2% (four experiments) in the human-rat P2X7 chimera (A), whereas the rat/human chimera (B) was insensitive to KN-62. C, KN-62 inhibition of currents evoked by ATP (300 µM) at the chimeric P2X7 receptors. Traces were obtained as described in the legend to Fig. 4. , From the rat/human chimera. , From the human/rat chimera. Three to five experiments were conducted for each value.

Effects of KN-62 on BzATP-activated ionic fluxes in leukocyte cell lines that natively express the human or murine P2X₇R. Having established that KN-62 inhibits homomeric human P2X₇R, we next examined the KN-62 sensitivity of P2X₇R as natively expressed in three different cell backgrounds: human macrophages (activated THP-1 monocytic leukemia cells), murine macrophages (the BAC1.2f5 macrophage line), and murine thymocytes (the BW5147 thymoma line). We previously established that human THP-1 monocytes treated with IFN- and LPS express both human P2X₇R mRNA and P2Z-like functional responses (Humphreys and Dubyak, 1996). Many studies of P2Z receptor function have used murine macrophage cell lines such as the BAC1.2f5 line used in our experiments. However, the sequence of the murine P2X₇R has not been reported, and its relative similarity to the human and rat P2X₇ homologues remains to be established. Finally, as noted above, the permeability characteristics of the P2Z receptor expressed in murine thymocytes differs from those observed in cells expressing the recombinant P2X₇Rs. Fig. 6A verifies that all of these leukocyte cell lines express P2X₇R mRNA. The primers used for these RT-PCR analyses were derived from the rat P2X₇R cDNA and will amplify a mRNA segment corresponding to the protein domain starting at Val245 and ending at Phe558. This encompasses 84 residues from the extracellular loop, the entire second transmembrane domain, and 85% of the 239 residue intracellular carboxyl-terminal tail. The absence of bands in the mock RT reactions rules out genomic DNA contamination of the RNA preparations, and amplification of human GAPDH or murine actin confirms the quality of the RTs. We tested the ability of KN-62 to attenuate BzATP-activated calcium influx in each of these cell types. Analysis of P2Z-mediated changes in cytosolic Ca²⁺ in human myeloid cells is complicated by coexpression of Ca²⁺-mobilizing P2Y₂ receptors, which also can be activated by 100 µM BzATP (Humphreys BD and Dubyak GR, unpublished observations). Thus, we routinely activated and desensitized these latter receptors by challenging THP-1 cells with 100 µM UTP for 10 min before the addition of BzATP. We previously confirmed that this protocol results in complete desensitization of the human P2Y₂ receptors (Humphreys and Dubyak, 1996). This desensitization step was not required with the BAC-1.2f5 macrophages because the murine P2Y₂ receptors expressed in these cells are not activated by 100 µM BzATP (El-Moatassim and Dubyak, 1992). Likewise, the BW5147 thymocytes were not pretreated with UTP because these murine lymphoid cells lack expression of any Ca²⁺-mobilizing P2Y₂ receptors (Humphreys BD and Dubyak GR, unpublished observations).

View larger version (26K): [in this window] [in a new window]   Fig. 6.   P2X7R mRNA expression and BzATP-evoked calcium responses in human and murine leukocyte cell lines. A, RNA was isolated from BW5147 murine lymphocytes, undifferentiated THP-1 human monocytes, THP-1 human monocytes differentiated with IFN (1000 units/ml), and LPS (1 µg/ml) for 3 days and BAC1.2f5 murine macrophages. RT-PCR was performed as described in Experimental Procedures using P2X7 primers specific to the unique carboxyl-terminal tail region of the P2X7R; parallel PCRs were performed on mock reverse-transcriptions to test for genomic DNA contamination. All amplified bands were of the expected size. B-D, Cells were loaded with Fura-2, equilibrated for 10 min in the absence or presence of 3 µM KN-62, and then assayed for increases in cytosolic Ca2+ in response to 100 µM BzATP. Maximal Ca2+-dependent fluorescence subsequently was determined by permeabilizing the cells with digitonin (20 µg/ml). P2Y2 receptors present on THP-1 cells were desensitized by stimulation with 100 µM UTP for 10 min before the addition of BzATP.

Effect of KN-62 on BzATP-activated pore formation in leukocyte cell lines that natively express the human or murine P2X₇R. A similar analysis was used test the effects of KN-62 on the pore-forming function of natively expressed human and murine P2X₇Rs. Although KN-62 completely inhibited BzATP-activated ethidium influx in the IFN/LPS-treated THP-1 human monocytes (Fig. 7A), it only partially inhibited the same response in the BAC1.2f5 murine macrophages (Fig. 7B). Characterization of the concentration-response relationships for KN-62 inhibition of ethidium influx in the two cell types suggested that reduced efficacy rather than reduced potency underlies the partial inhibitory effect observed in the murine macrophages (Fig. 7C). We were unable detect any BzATP-induced ethidium influx in the BW5147 thymocytes (data not shown), which is consistent with the findings from previous studies of P2Z receptor function in murine thymocytes (Pizzo et al., 1991).

View larger version (19K): [in this window] [in a new window]   Fig. 7.   KN-62 concentration response on BzATP-induced ethidium uptake in THP-1 human monocytes and BAC1.2f5 murine macrophages. A, THP-1 cells that had been incubated with IFN (1000 units/ml) and LPS (1 µg/ml) for 3 days were challenged with 100 µM BzATP in the absence or presence of KN-62 and ethidium influx was measured. B, BAC1.2f5 macrophages were subjected to the same procedure using 100 µM BzATP and different concentrations of KN-62. C, Concentration-response relationships characterizing the effects of KN-62 on ethidium influx activated by 100 µM BzATP in THP-1 monocytes () and BAC1.2f5 (). Ethidium influx at 3 min after BzATP was measured and normalized to the maximal BzATP-induced accumulation assayed in the absence of isoquinoline. Values represent the mean ± standard error of three samples.

Isoquinolines rapidly block ethidium influx through previously activated P2X₇R pores. All of the previous experiments tested inhibitory effects of isoquinolines in cells that were preincubated for 10 min with these drugs before agonist activation of the P2X₇Rs. Using IFN/LPS-treated THP-1 monocytes, we tested the effects of KN-62 and KN-04 on ethidium fluxes through P2X₇R pores preactivated by BzATP addition before exposure to the isoquinolines. Fig. 8 shows that isoquinoline addition after nucleotide challenge rapidly attenuated ethidium influx, with maximal inhibition occurring after 20 sec. Similar results were obtained in studies assaying ATP-activated ethidium influx in HEK-hP2X₇ cells (data not shown). The rapidity of inhibition is consistent with the notion that KN-62 inhibition may be mediated via its interaction with extracellular domains of the human or murine P2X₇R.

View larger version (6K): [in this window] [in a new window]   Fig. 8.   KN-62 blocks ethidium influx through previously opened P2X7 pores. A and B, To test whether the isoquinoline derivatives could inhibit ethidium influx after P2X7R activation, THP-1 cells that had been incubated for 3 days in IFN (1000 units/ml) and LPS (1 µg/ml) were challenged with 100 µM BzATP, and ethidium uptake was measured. One minute after BzATP addition, either 1 µM KN-62 (A) or 1 µM KN-04 was added (arrow), and ethidium influx compared with cells incubated without drug.

	Discussion

Top Summary Introduction Procedures Results Discussion References

Activation of P2X₇Rs by extracellular ATP triggers two distinct fluxes. The first reflects a rapid and reversible current through channels permeable to small cations, whereas the second requires either sustained agonist application or repetitive short exposures and reflects the opening of a nonselective pore through which large organic molecules such as ethidium or YoPro-1 can pass (Nuttle et al., 1993, Surprenant et al., 1996). We have shown that the isoquinoline derivatives KN-62 and KN-04 block both of these fluxes in HEK 293 cells that express recombinant human P2X₇Rs but not recombinant rat P2X₇Rs. The extreme species-specific actions of these drugs was surprising given the high degree of amino acid identity (80%) between the human and rat P2X₇Rs. Isoquinoline sensitivity is mediated by the first 335 amino acids of the human P2X₇R because transfer of the human amino-terminal domain, which included the intracellular amino-terminal tail, the first transmembrane segment, and the entire extracellular loop, to the rat P2X₇R resulted in a receptor that exhibited KN-62-sensitive ethidium influx. In contrast, KN-62 had no effect on the reciprocal chimera that transferred the corresponding amino-terminal regions of the rat P2X₇R to a segment containing the second transmembrane domain and intracellular carboxyl-terminal tail of the human receptor. The isoquinoline-sensitive span of 335 amino acids from the human receptor contains 64 amino acid differences between human and rat sequences, with one stretch of seven consecutive amino acid differences (153-159). Because P2X₇R function in murine cell types (BW5147 and BAC1.2f5) also was isoquinoline sensitive, a sequence comparison among human, rat, and murine (when available) P2X₇Rs may help to localize further the KN-62 interaction domain. It is interesting to note that human and rat homologues of the P2X₄ receptor also exhibit markedly different sensitivities to the P2 receptor antagonists suramin and pyridoxal-phosphate-6-azophenyl-2',4'-disulfonic acid despite their 89% sequence identity (Garcia-Guzman et al., 1996). This suggests that valuable structure-function information may be gained from careful analysis of drug sensitivities across the species homologues of all P2X receptor family proteins.

As first suggested by Gargett and Wiley (1997), the ability of KN-04, an inactive analog of KN-62, to inhibit human P2X₇R function suggests that CamKII does not play a role in regulation of either conductance state of this receptor. Also arguing against a role for CamKII is the failure of either isoquinoline derivative to inhibit the rat P2X₇R. The cytoplasmic carboxyl-terminal tails of both the human and rat P2X₇Rs contain homologous segments (residues 556-562; -RFGSQD- in the human receptor and -RFVSQD- in the rat protein) that comprise the minimal CamKII consensus phosphorylation site (-RXXS/T-) (Braun and Schulman, 1995). However, our chimeric receptors localized KN-62 sensitivity to the amino-terminal half of the human receptor, ruling out CamKII-mediated phosphorylation of the carboxyl-terminal tail as a mechanism for isoquinoline inhibition of P2X₇R signaling. Gargett and Wiley (1997) speculated that KN-62 might directly bind to the unique carboxyl-terminal tail of the human P2X₇R. This is unlikely given our finding that transfer of the intracellular carboxyl-terminal tail of the human protein to the rat receptor resulted in a chimeric P2X₇R that was insensitive to the isoquinolines (Figs. 4C and 5B).

Blanchard et al. (1995) proposed that CamKII could act as a specific regulator of P2ZR-mediated cytolysis because both KN-62 and the calmodulin antagonists, trifluoroperazine and calmidazolium, protected human monocyte-derived macrophages from ATP-mediated cytolysis. These authors also suggested that this effect of KN-62 was downstream of the Ca²⁺ influx response to P2ZR activation because the drug failed to block the ATP-induced increase in cytosolic Ca²⁺. This mechanism seems unlikely because we performed similar studies with human monocyte-derived macrophages and found that KN-04 inhibits ATP-induced release of lactate dehydrogenase to the same degree as KN-62 (Humphreys BD and Dubyak GR, unpublished observations). The analysis of ATP-induced changes in Ca²⁺ in macrophages is complicated by coexpression of the ionotropic P2X₇Rs and the Ca²⁺-mobilizing P2Y₂ receptors (Torre and Trautmann, 1993; Humphreys and Dubyak, 1996). Trifluoroperazine and calmidazolium can block ATP-induced calcium influx by the human P2X₇R (Humphreys BD and Dubyak GR, unpublished observations) whereas calmidazolium additionally blocks the rapid gating of inward current by ATP in HEK cells expressing the rat P2X₇R (Virginio et al., 1997). This suggests that these drugs also are direct inhibitors of P2X₇R function. It is significant to note that KN-62, calmidazolium and trifluoroperazine have been shown to directly block L-type calcium channels (Nakazawa et al., 1993, Li et al., 1992), suggesting that P2X₇Rs and L-type calcium channels share multiple structural determinants that mediate the inhibitory effect of these three drugs.

The molecular mechanism or mechanisms that underlie the inhibition of human P2X₇R function by KN-62 remain uncertain. The observed inhibitory effects of these compounds (as described in this study and by Gargett and Wiley, 1997) cannot be ascribed to action as a simple competitive or noncompetitive antagonist. Different IC₅₀ values and efficacies were observed depending on (1) the species of nucleotide agonist used to activate the P2X₇R, (2) the type of P2X₇R function being assayed, and (3) the cellular background of human P2X₇R expression. The IC₅₀ values for KN-62 in our experiments ranged from 30 to 100 nM (Figs. 2-4 and 7). This is significantly higher than the IC₅₀ value of 13 nM that characterized the effects of KN-62 on various ATP-induced responses in human CLL leukocytes (Gargett and Wiley, 1997). However, it should be noted that these latter investigators used cells suspended in high KCl/low NaCl medium, whereas our assays were performed using medium with standard extracellular NaCl concentrations. Wiley et al. (1992) reported previously that extracellular Na⁺ attenuates P2Z receptor function in CLL lymphocytes. The affinity of KN-62 for its binding site or sites on the human P2X₇R may show similar modulation by different monovalent cations or other components of the extracellular medium (e.g., pH, divalent cations).

In contrast, no simple hypothesis can explain the observed variations in KN-62 efficacy among the different experimental systems. KN-62 was a full antagonist of human P2X₇R when ATP was the agonist but was a partial antagonist when BzATP was the agonist (Figs. 2-4 and 7). The ability of KN-62 to fully antagonize ATP-activated human P2X₇R was independent of cellular background (recombinant receptor in HEK cells, native receptor in THP-1 cells) or type of assayed P2X₇R function (rapidly gated cation conductance, ethidium influx pore, phospholipase D, cytolysis). In contrast, the partial antagonism of KN-62 at BzATP-activated human P2X₇R varied with both cell background and assayed function. KN-62 fully antagonized BzATP-activated ethidium influx in THP-1 monocytes but partially inhibited the same phenomenon in HEK-hP2X7 cells. Although KN-62 only partially antagonized the ability of BzATP to induce ethidium influx (indicative of the nonselective pore) in these latter cells, it fully inhibited the ability of BzATP to the trigger the rapidly gated cation current. These findings are consistent with previous speculations that the rapidly gated ion channel and the more slowly induced pore represent functionally and structurally distinct conformations of the P2X₇R (Nuttle and Dubyak, 1994; Surprenant et al., 1996, Rassendren et al., 1997). This hypothesis has been further supported by the observation of Virginio et al. (1997) that calmidazolium can fully antagonize the rapid gating of cation channels by BzATP in HEK cells expressing recombinant rat P2X₇R while having no inhibitory effect on BzATP-activation of Yo-Pro dye uptake. Although our results do not identify the mechanism by which the isoquinolines inhibit P2X₇R-mediated transmembrane fluxes, the ability of these drugs to block previously opened P2X₇ pores (Fig. 8) suggests that these drugs may directly block pore permeation.

Although expression of recombinant P2X₇Rs confers the ATP-induced ethidium influx phenotype to HEK 293 cells, this type of membrane permeability change was absent in the BW5147 lymphocytes. Our experiments represent the first verification that murine thymocytes do indeed express P2X₇R mRNA despite their lack of the pore-forming P2X₇R phenotype. These cells also displayed a BzATP-activated and KN-62-sensitive Ca²⁺ influx response indicative of the cation-conducting conformation of the P2X₇Rs. The failure of this lymphocyte cell line to accumulate ethidium in response to extracellular ATP is consistent with previous studies of P2Z receptor function in rodent lymphocytes; organic dyes, such as ethidium, that readily permeate the macrophage P2ZR failed to pass through the P2ZR of mouse thymocytes (Pizzo et al., 1991). It remains unclear why the lymphocyte P2X₇R have a smaller size exclusion limit than P2X₇R expressed in other cell types. Lymphocytes may express another P2X subtype capable of modulating P2X₇R function because some combinations of P2XR subtypes can form heteromeric channels (Lewis et al., 1995). Alternatively, efficient pore formation may depend on the local concentration of P2X₇R subunits at the cell surface. In this regard, lymphocytes might have a lower P2X₇R surface density than either macrophages or cells engineered to overexpress recombinant P2X₇R.

Recent studies point to important physiological roles for P2X₇Rs in immune cell biology. Expression of the murine T-lymphocyte P2Z/P2X₇ is modulated during differentiation, and the receptor is proposed to play a role in thymocyte selection (Chused et al., 1996, Ross et al., 1997). Extracellular ATP is mitogenic to human T lymphocytes and pharmacological data implicates P2Z/P2X₇ in mediating the mitogenic effects of phytohemaglutinin and CD3 stimulation (Baricordi et al., 1996). In human macrophages and microglial cells, P2Z/P2X₇ activation has clearly been shown to trigger activation of the cysteine protease caspase-1 (interleukin-converting enzyme) with subsequent cleavage and release of pro-IL-1 to bioactive IL-1 (Hogquist et al., 1991; Perregaux and Gabel, 1994; Ferrari et al., 1997). Two groups have shown that P2X₇R activation by a brief ATP pulses triggers apoptosis of Bacillus-Clamette-Guerin (BCG)-infected monocytes with killing of the intracellular bacilli (Molloy et al., 1994, Lammas et al., 1997). Moreover, other apoptotic stimuli (CD95 or FAS ligation) directed at the infected macrophages induce apoptosis but do not kill the mycobacteria (Lammas et al., 1997). The high potency antagonism of the P2X₇R by isoquinoline derivatives presents a useful tool to study these effects of P2z/P2X₇ receptor activation in vivo.