Resistance to 2-Chloro-2′-deoxyadenosine of the Human B-cell Leukemia Cell Line EHEB¹

CdA,³ an adenosine deaminase-resistant analogue of the natural nucleoside 2′-deoxyadenosine, is highly toxic to both proliferating and resting human lymphocytes (1, 2, 3). For this reason, it is widely used for the treatment of chronic lymphoid malignancies, particularly hairy cell leukemia and CLL (reviewed in Ref. 4). To exert its antileukemic effect, CdA has to be phosphorylated by dCK to CdAMP and then converted into CdADP and CdATP. CdATP inhibits various enzymes involved in DNA replication and repair, including ribonucleotide reductase and DNA polymerases α and β; moreover, it can be incorporated into newly synthesized DNA, causing chain termination (reviewed in Ref. 5). Together, these actions result in the progressive accumulation of DNA strand breaks, leading to initiation of apoptosis by mechanisms that are not entirely clear (2). Recently, CdATP has been shown to contribute to caspase-3 activation in cell-free extracts, which may play a role in CdA-induced apoptosis in resting cells (6, 7).

With the aim of improving our understanding of the mechanisms by which CdA induces apoptosis in B-cell CLL lymphocytes, we studied the continuous B-cell CLL cell line EHEB (8). Unexpectedly, we observed that these cells were much less sensitive to CdA than freshly isolated B-cell CLL lymphocytes and several lymphoblastic cell lines and that this lower sensitivity could not be related to a reduced accumulation of CdATP, which is reportedly a determining factor in the effect of CdA (9). Further investigations showed that inhibition in situ of ribonucleotide reductase required higher concentrations of CdA in EHEB cells than in other proliferating cell lines (10). Moreover, CdA, at concentrations that are usually inhibitory (1, 10, 11, 12), increased both dThd incorporation into DNA and the proportion of cells in S phase of the cell cycle. This paradoxical, hitherto unreported effect of CdA on DNA synthesis was further investigated in relation with the resistance to CdA of the EHEB cell line.

Unlabeled CdA (>99.9% purity) was synthesized and supplied by Prof. L. Ghosez (Laboratory of Organic Chemistry, Louvain-la-Neuve, Belgium). Stock solutions were prepared in ethanol/150 mm NaCl (1:1 by volume). [8-³H]CdA (4.7 mCi/mmol) and [5-³H]Cyd (18.1 Ci/mmol) were purchased from Moravek Biochemicals (La Brea, CA). [methyl-³H]dThd (84 Ci/mmol) and [5-³H]dCyd (24 Ci/mmol) were from Amersham International (Buckinghamshire, United Kingdom). MTT, hydroxyurea, BrdUrd, PI, and FITC-conjugated anti-BrdUrd antibodies were purchased from Sigma Chemical Co.-Aldrich (Bornem, Belgium). The mouse monoclonal BrdUrd-specific BR3 antibody (MD 5300) was from Caltag Laboratories, Inc. (San Fransico, CA). FCS was purchased from BioWhittaker Europe (Verviers, Belgium), RPMI 1640 was from Life Technologies Inc. (Merelbeke, Belgium), and all tissue culture reagents were from Gibco (Eggenstein, Germany). All other chemicals, materials, and reagents were of the highest quality available.

The continuous cell lines EHEB and CCRF-CEM were cultured in RPMI 1640 with glutamate and supplemented with 10% heat-inactivated FCS at 37°C in an atmosphere of 5% CO₂ in air and routinely tested for Mycoplasma contamination. Freshly obtained peripheral blood from B-cell CLL patients was fractionated by Ficoll-Paque sedimentation. Mononuclear cells were washed and resuspended in RPMI 1640 supplemented with 10% FCS and 1% penicillin-streptomycin. Cells were counted, diluted to indicated concentration in RPMI 1640, and incubated at 37°C in 5% CO₂ in air. All patients had a confirmed diagnosis of B-cell CLL by cytological and immunological studies and were free of any cytotoxic therapy for at least 6 months. Only patients with lymphocyte counts over 30,000/μl were selected.

EHEB and CCRF-CEM cells (resuspended at a concentration of 0.2 × 10⁶ cells/ml) and B-cell CLL cells (resuspended at a concentration of 1.5 × 10⁶ cells) were incubated in the presence of various concentrations of CdA in 96-well plates. After 96 h, cell viability was measured using the MTT assay as described by Mosmann (13), with the modifications of Morabito et al. (14). Controls and CdA concentrations were set up in triplicate. The absorbance of each well was measured at 540 nm with a Multiwell Scanning Spectrophotometer (Molecular Devices, Sunnyvale, CA). Cell viability, expressed as a percentage, was calculated by the equation: (mean absorbance of treated well/mean absorbance of control wells) × 100. The IC₅₀ was determined graphically. Cell viability was also estimated in a hemocytometer by trypan blue dye exclusion.

High molecular weight DNA fragmentation was analyzed by pulsed field gel electrophoresis as described by Huang et al. (15). Apopain activity was determined with the FluoroAce Apopain Assay from Bio-Rad.

To measure nucleotides synthesized from CdA, EHEB cells were incubated at a concentration of 1 × 10⁶ cells/ml in the presence of [8-H³]CdA (2.5 μCi/ml) and various concentrations of unlabeled CdA. Preparation of perchloric cell extracts and separation of labeled CdA nucleotides by high-performance liquid chromatography were performed as described previously (16). The amounts of nucleotides synthesized were calculated from the specific radioactivity of CdA. To measure the amount of CdA incorporated into nucleic acids, radioactivity was determined in the acid-insoluble fraction of the cellular extract after two washings with 1 m perchloric acid and digestion with soluene.

As described by Griffig et al. (10), ribonucleotide reductase can be investigated in intact cells by measuring the conversion of Cyd into deoxynucleotides and its incorporation into DNA. Cells (5 × 10⁶ cells/5 ml) were incubated for 24 h in the presence of various concentrations of CdA before the addition of [5-³H]Cyd (1 μCi/ml). One h after the addition of the latter precursor, incubation was stopped, and cells were treated as described previously (16). Acid-soluble ³H-deoxynucleotides were quantified by high-performance liquid chromatography after destruction of ribonucleotides with NaIO₄ (17). Radioactivity incorporated into DNA was determined after removal of RNA from the acid-insoluble pellet by incubation with 0.5 n NaOH for 2 h at 37°C (18).

Perchloric cell extracts were prepared as described for the measurement of labeled CdA nucleotides (16). Concentrations of dNTPs were determined using the DNA polymerase assay of Sherman and Fyfe (19), with the modifications of van Moorsel et al. (20). In each separate experiment, measurements were performed in triplicate. The intracellular concentrations of dNTPs were estimated from calibration curves obtained with pure standards.

DNA synthesis was investigated by measuring the incorporation of [³H]dThd into the acid-insoluble cellular fraction. Cell suspension (800 μl) containing 2 × 10⁵ cells/ml was incubated for 4 h with [³H]dThd (1 μCi/ml). Subsequently, 200 μl were taken in triplicate, harvested on a multiscreen assay system 96-well filtration plate (Millipore; catalogue number MAGV N22), and washed successively with 100 μl of PBS, 200 μl of TCA 8%, 4 × 100 μl of H₂O, and finally with 1 × 200 and 3 × 100 μl of 70% ethanol. Remaining radioactivity associated with filters was counted. Incorporation of [³H]dCyd into DNA was determined by the same procedure.

Frozen cell pellets corresponding to 10 × 10⁶ cells were suspended in 0.5 ml of 50 mm Tris-HCl (pH 7.2) and disrupted by three freeze/thawings. Assays were performed on the supernatant obtained after centrifugation. CdA phosphorylation was assayed with 50 μm [8-³H]CdA (∼200 cpm/pmol) and 0.03–0.05 mg of cellular protein in a final volume of 100 μl (21). TK and dCK activities were measured under the same assay conditions with 10 μm [³H]dThd or 10 μm [5-³H]dCyd (∼1000 cpm/pmol) as substrates. The enzyme assays were carried out at 37°C, and 10-μl aliquots were taken after several appropriate time intervals, within 30 min, to yield a linear reaction rate. The protein content of cell extracts was measured by the method of Bradford (22), using BSA as the standard.

After incubation for the times indicated, with or without 10 μm CdA, samples containing 10 × 10⁶ cells were pulse-labeled for 1 h with 10 μm BrdUrd, which is incorporated into DNA in place of dThd during DNA synthesis. The proportion of cells in S phase was determined by bivariate flow cytometric analysis using FITC-conjugated anti-BrdUrd antibody to reveal nascent DNA and PI for total DNA content (23). Briefly, nuclei from BrdUrd pulse-labeled cells were isolated by pepsin treatment (0.04% in 0.1 m HCl). DNA was denaturated in 2 n HCl for 20 min at 37°C, neutralized with 0.1 m Na₂B₄O₇, and washed with PBTB. Pellets were resuspended in 150 μl of the same medium containing 500-fold diluted (v/v) mouse monoclonal BrdUrd antibody and incubated for 45 min at room temperature. Nuclei were washed in PBTB, centrifuged, and incubated for an additional 45 min with 150 μl of FITC-labeled goat antimouse IgG (diluted 50-fold in PBTB). After washing, nuclei were resuspended in 2 ml of PI solution (10 μg/ml PI in PBTB). RNase was added just before flow cytometry analysis to a final concentration of 5 μg/ml.

All measurements were performed with a FACScan fluorescence-activated cell analyzer (Becton Dickinson). Data were analyzed by CellQuest I software (Becton Dickinson). For each sample, 15,000 events were collected. Debris and doublets were excluded from the analysis. Dot plots generated by FACScan analysis (see Fig. 8) show DNA content on the X axis (linear) and DNA synthesis on the Y axis (log). Both cell cycle distribution and DNA synthesis capacity can be estimated by gating BrdUrd-labeled (S phase) from unlabeled and single from double DNA content nuclei.

All results of repeated experiments are given as the means ± SE. Significance was estimated by the paired two-tailed Student’s t test.

Sensitivity to CdA of the B-cell CLL cell line EHEB was evaluated by the MTT reduction assay after 4 days of incubation and compared with that of B-cell CLL lymphocytes freshly isolated from patients (Fig. 1). With an IC₅₀ of approximately 5 μm, EHEB cells appear to be 25-fold less sensitive to CdA than B-cell CLL lymphocytes, for which an IC₅₀ of 0.2 μm was calculated. An IC₅₀ of less than 0.5 μm was determined by the same procedure in CCRF-CEM cells (results not shown). Identical results were obtained when cell sensitivity to CdA was investigated by cell count in the presence of 0.1% trypan blue. Because of their quite long doubling time (48–72 h), the sensitivity of EHEB cells to CdA was also measured after 7 days of incubation with the drug, which reduced the IC₅₀ slightly from 5 to 2 μm. This low sensitivity to CdA of EHEB cells contrasts with that of other human lymphoblastic cell lines, in which inhibition of cell growth or cytotoxicity was recorded with nanomolar concentrations of CdA (1, 3, 11, 24, 25, 26, 27, 28).

High molecular weight DNA fragments and activation of apopain/caspase-3 were detected in EHEB cells at concentrations of CdA that induced a loss of viability (data not shown). This indicates that cytotoxicity of CdA in EHEB cells, as in other cell lines and in B-cell CLL lymphocytes, results at least in part from induction of apoptosis.

Phosphorylation of CdA by dCK, leading to intracellular accumulation of CdATP, is a prerequisite for its cytotoxicity. Cell lines deficient in dCK are indeed resistant to the toxic effect of CdA (24), and a correlation has been evidenced between the level of phosphorylation of CdA and its cytotoxic effect (9). The rate of CdA phosphorylation (mean ± SE of three separate experiments), measured at a substrate concentration of 50 μm, was 0.12 ± 0.02 nmol/min/mg protein in EHEB cell extracts as compared with 0.22 ± 0.03 nmol/min/mg protein in B-cell CLL cell extracts. These rates are in the same range as those determined by others in various lymphoblastoid cell lines (1, 29) and in B-cell CLL lymphocytes (21). Kinetic properties of the CdA phosphorylating activity were also investigated in EHEB cell extracts: a Michaelis-Menten curve with a K_m of 20 μm was obtained, which accords with values found for recombinant dCK by Johansson and Karlsson (30).

Because an active dCK could be counteracted by a high cytoplasmic 5′-deoxynucleotidase activity, leading to impaired accumulation of CdATP, we investigated the metabolism of CdA in intact EHEB cells. EHEB cells rapidly formed CdAMP and CdATP, whereas CdADP was nearly undetectable. Stable levels of CdA nucleotides were reached after 2 h and maintained for at least 24 h. Values measured after a 4-h incubation with CdA are presented in Fig. 2, which shows that maximal levels of CdAMP and CdATP were obtained at 20 μm CdA. The calculated concentrations of CdATP were 1.73 ± 0.27 μm at 0.5 μm CdA and 9.9 ± 1.1 μm at 20 μm CdA (means ± SE of three separate experiments). These values are in the same range as those measured in other CdA-sensitive leukemia cell lines (3, 10). As also illustrated in Fig. 2, EHEB cells actively incorporated CdA into their DNA at a rate of 2.7 ± 0.7 pmol/10⁸ cells/h, which is comparable with that determined in CCRF-CEM lymphoblasts (1, 10). Taken together, these results dismiss impaired metabolism of CdA as the cause of the resistance of EHEB cells to this drug.

CdATP is one of the most potent inhibitors of ribonucleotide reductase identified to date, with an IC₅₀ of 0.1–0.3 μm in cellular extracts (10). Inhibition of ribonucleotide reductase is hence assumed to be a determining factor in inhibition of DNA synthesis and cytotoxicity induced by CdA, at least in dividing cells. To evaluate ribonucleotide reductase activity in situ, we analyzed the effect of CdA on the conversion of labeled Cyd into deoxynucleotides and its incorporation into DNA (10). As depicted in Fig. 3, sizeable inhibition of the conversion of [³H]Cyd into dCTP and of its incorporation into DNA was not observed at concentrations of CdA below 10 μm. It was verified that incorporation of [³H]Cyd into DNA could be inhibited in EHEB cells by 93 ± 2% (n = 3) by 1 mm hydroxyurea, an established ribonucleotide reductase inhibitor. As in Ref. 10, we observed an inhibition of ribonucleotide reductase in situ by CdA with an IC₅₀ of less than 0.1 μm in CCRF-CEM cells. Thus, inhibition of ribonucleotide reduction required more than 100-fold higher concentrations of CdA in EHEB cells than in CCRF-CEM cells.

In CdA-sensitive cell lines, CdA has been shown to lower the concentration of all or some of the dNTPs as a consequence of inhibition of ribonucleotide reductase by CdATP (10, 31, 32, 33). The fact that ribonucleotide reductase was only weakly inhibited by CdA in EHEB cells led us to measure the dNTP concentrations in these cells. Not only did a 24-h incubation in the presence of increasing concentrations of CdA fail to decrease the concentrations of the dNTPs, but it tended to elevate them (Fig. 4), with a maximal effect at around 1 μm CdA. Only dATP decreased significantly below the control values on incubation with 10 and 50 μm CdA.

An inhibitory effect of CdA on DNA synthesis, as assessed by precursor incorporation, has been demonstrated in various sensitive cell lines (1, 10, 11, 24, 31) and in human resting and tonsillar lymphocytes (12, 34). The effect of CdA on the incorporation of [³H]dThd into DNA of EHEB cells is shown in Fig. 5. After a 4-h incubation, incorporation of [³H]dThd into DNA was inhibited by concentrations of CdA above 5 μm. In marked contrast, after 24 h, it was significantly enhanced by the nucleoside at concentrations ranging from 2–10 μm. At 10 μm CdA, this incorporation was increased approximately 2-fold. This effect was still observed 48 and 72 h after the addition of CdA (results not shown). As illustrated in Fig. 6, incorporation of [5-³H]dCyd into DNA was also increased up to 4-fold by a 24-h incubation with 10 μm CdA. This effect was already detectable after a 4-h incubation. Both the increased dThd and dCyd incorporations were completely prevented by the addition of 100 μm dCyd, which competitively inhibits CdA phosphorylation (data not shown).

The fact that CdA elicited an increase in precursor incorporation and in the concentration of dNTPs in EHEB cells led us to investigate its effects on TK and dCK activities. As illustrated in Fig. 7, CdA induced a dose-dependent increase of the activities of both TK and dCK, as measured 24 h after its addition. A maximal effect was observed at ∼5 μm CdA for TK and at ∼1 μm CdA for dCK. For the latter enzyme (but not for TK), an increase of activity was already observable after 4 h [an increase from 15.5 ± 3.5 pmol/min/mg protein in the absence of CdA to 29.1 ± 4.2 pmol/min/mg protein in the presence of 10 μm CdA (means ± SE of four separate experiments); P < 0.05]. The increase in TK activity was due to TK1, the cytoplasmic isoform, rather than to TK2, the mitochondrial isoform, because TK1 mRNA but not TK2 mRNA was detectable in the EHEB cells (results not shown). The stimulatory effect of CdA on dCK and TK activities was suppressed by coincubation with an excess of dCyd (data not shown), indicating that CdA has to be phosphorylated to exert this effect. The increase of these enzyme activities was not suppressed by dialysis of the cell lysates.

Double labeling with PI and BrdUrd allows measurements of DNA content simultaneously with analysis of DNA synthesis activity because BrdUrd is an S-phase marker. Fig. 8 shows the scattergrams obtained by FACScan analysis of cells incubated for 24 h without (Fig. 8,A) or with 10 μm CdA (Fig. 8,B). The dashed lines separate the BrdUrd-incorporating cells (cells actively synthesizing DNA) from those that did not incorporate BrdUrd. Two cell subpopulations were distinguished among the BrdUrd-incorporating cells: (a) the early S-phase cells, with a DNA content close to 2 C and a variable degree of BrdUrd incorporation, which had entered S-phase during the 1 h-pulse of BrdUrd; and (b) the mid-late S-phase cells with a variable DNA content. The two subpopulations of cells that did not incorporate BrdUrd and have a DNA content of 2 or 4 C are G₁ or G₂-M cells, respectively, that did not synthesize DNA during the pulse. Separation of the regions around the G₁, early S, mid-late S, and G₂-M subpopulations in the two scattergrams allowed determination of the percentages of populations within the cell cycle. Results obtained from four separate experiments, similar to those illustrated in Fig. 8, showed (Table 1) that the proportion of cells in G₁ phase was significantly decreased after a 24-h incubation with 10 μm CdA, whereas the proportion of cells synthesizing DNA (in either early or mid-late S phase) was significantly increased. The fact that the height of the crescent was the same without (Fig. 8,A) or with CdA (Fig. 8,B) indicates that CdA did not increase the rate of DNA synthesis but only increased the number of cells synthesizing DNA. As also shown in Table 1, the proportion of cells in G₂-M was only slightly decreased. Additional studies (data not shown) showed that a 6-h exposure of EHEB cells to CdA did not significantly modify the distribution of cells among the phases of the cell cycle, whereas a pattern similar to that given in Table 1 was observed 48 h after CdA addition. In contrast with EHEB cells, exposure of CCRF-CEM cells to CdA for 6 or 24 h suppressed the subpopulation of cells incorporating BrdUrd (results not shown).

This study shows that EHEB cells are approximately 25-fold less sensitive to CdA than lymphocytes freshly isolated from patients with B-cell CLL (Fig. 1) and 10–1000-fold less sensitive to CdA than other lymphoblastic cells (1, 3, 11, 24, 25, 26, 27, 28). The importance of the level of CdA phosphorylation in the cytotoxic effect of CdA has been emphasized by several authors (1, 9, 26, 29), and although there is not a simple relationship between the intracellular concentration of CdATP and the cytotoxicity of CdA, less sensitive cell lines tend to accumulate less CdATP than more sensitive ones (3). However, as mentioned in “Results,” the relative resistance of EHEB cells to CdA could not be explained by a defect in the accumulation of CdATP. Indeed, CdATP accumulated to the same levels in EHEB cells as in other sensitive cell lines (1, 3, 10, 35). In addition, EHEB cells incorporated CdA into DNA at a rate that was in the same range as that seen in the CCRF-CEM cell line, which is very sensitive to CdA (1, 10). Taken together, these data preclude a defect in the metabolism of CdA as the mechanism of resistance of EHEB cells to this drug.

CdATP is a potent inhibitor of ribonucleotide reductase; thus, its intracellular accumulation is expected to provoke a strong inhibition of the conversion of Cyd into dCTP and its incorporation into DNA (10). However, these processes were only marginally inhibited in EHEB cells at CdA concentrations below 10 μm (Fig. 3), whereas they were nearly totally inhibited at 0.1 μm in CCRF-CEM cells. These findings are in accord with recent reports that suggest that differences in purine analogue cytotoxicity could also be related to differences in inhibition or regulation of ribonucleotide reductase (27, 36). In accordance with the weak inhibition of ribonucleotide reduction, dNTP concentrations, which decrease in CdA-sensitive cell lines as a consequence of the inhibition of ribonucleotide reductase, were not lowered in EHEB cells and even tended to increase under CdA (Fig. 4).

In CdA-sensitive cell lines, DNA synthesis was found to be nearly completely inhibited at concentrations of CdA around 1 μm (1, 10, 11, 12, 24, 31). At these concentrations, DNA synthesis, as measured 4 h after the addition of CdA by the incorporation of dThd, was not inhibited in EHEB cells (Fig. 5). On the contrary, it was dose-dependently increased after a 24-h incubation in the presence of up to 10 μm nucleoside (Fig. 5). This stimulatory effect was observed not only on the incorporation of dThd into DNA but also on that of dCyd into DNA (Fig. 6), which, for unexplained reasons, was markedly more stimulated than that of dThd.

Cell cycle analysis (Fig. 8 and Table 1) showed that 10 μm CdA significantly enlarged the proportion of cells in S phase and that this increase was correlated with an increase in the number of cells incorporating BrdUrd in either early S phase or mid-late S phase. These results indicate that the increase of dThd incorporation observed after CdA in EHEB cells is not due to an increase in the rate of DNA synthesis but rather to an increase in the number of cells synthesizing DNA. These effects of CdA on the cell cycle and dThd incorporation have, to our knowledge, never been reported, although DNA synthesis has not been systematically investigated in CdA-resistant cell lines. They contrast with those reported in CCRF-CEM cells, in which CdA inhibits DNA synthesis and provokes an accumulation of most cells in either early S phase or at the G₁-S border, accompanied by a progressive depletion of cells in S and G₂-M phase (11). We have also observed in CCRF-CEM cells that the latter events are accompanied by a suppression of BrdUrd incorporation. Our observation that EHEB cells did not accumulate in G₂ phase indicates that the accumulation of cells in S phase is not due to a block at the G₂-M boundary. It remains to be determined whether accumulation of EHEB cells in S phase is due to a stimulation of entry into this phase or to a delay in progression therein. An increase in DNA repair by CdA can also be suspected because of the accumulation of EHEB cells in early S phase. Accumulation of cells in S phase most likely accounts for the elevation of activity of TK1, an S-phase enzyme involved in the nucleoside salvage pathway (37, 38, 39), after the addition of CdA (Fig. 7). The accumulation of cells in S phase, the increase in TK and dCK activities, and the weak inhibition of ribonucleotide reductase, which is also an S-phase enzyme, could account for the increase in dNTP concentrations observed after incubation with CdA.

Activation of dCK by short-term treatment with CdA has already been reported in human tonsillar lymphocytes (12, 40). In the latter cells, however, DNA synthesis was concomitantly strongly inhibited by CdA, and TK activity was not stimulated. A posttranslational modification of dCK has been proposed to explain its CdA-induced activation in human tonsillar lymphocytes, based on the observation that the amount of dCK protein was not modified by CdA treatment (12). Moreover, it has been shown that dCK can be inactivated by protein phosphatase treatment (41). In human tonsillar lymphocytes, activation of dCK by CdA can be prevented by an excess of dCyd, which competes with CdA for phosphorylation (12). Similarly, we observed that an excess of dCyd prevented the activation of TK as well as dCK, and suppressed the increase of dThd and dCyd incorporation into DNA by CdA in EHEB cells. These results indicate that the latter process and the increases of TK and dCK activities are linked and that they occur secondary to CdA phosphorylation.

We conclude that the resistance of the human B-leukemia cell line EHEB to CdA does not result from an impairment of its phosphorylation into CdATP but from a complex series of events encompassing poor inhibition of ribonucleotide reductase and an increase in the proportion of DNA-synthesizing cells in S phase. Questions such as why ribonucleotide reductase in situ is less sensitive to CdA and why cells accumulate in S phase remain to be unraveled. Elucidation of these processes may shed light on the resistance to CdA in clinical situations and lead to its rational association with other drugs to improve its efficacy.

We thank Dr. P. van der Bruggen for the use of the multiwell scanning spectrophotometer, Dr. K. van Moorsel for help with dNTP determination by the DNA polymerase assay, and Dr. V. Gregoire for expert advice.