Introduction

Top Abstract Introduction Materials & Methods Results & Discussion References

The therapeutically important circulating metabolite of the anticancer drug cyclophosphamide is 4-hydroxycyclophosphamide/aldophosphamide (reviewed in ref. 1). NAD(P)-linked aldehyde dehydrogenases catalyze the oxidation of aldophosphamide to carboxyphosphamide, thereby irreversibly detoxifying it (reviewed in ref. 1). Most efficient at doing so appears to be ALDH-1.¹ Highest levels of this enzyme are ordinarily found in the liver. Thus, systemic detoxification of cyclophosphamide is thought to be catalyzed largely by hepatic ALDH-1, and, indeed, it catalyzes the majority (> 80% when the aldophosphamide concentration is pharmacological) of human hepatic aldophosphamide oxidation (2). However, lower levels of ALDH-1 are present in numerous tissues including erythrocytes (reviewed in ref. 3) where it reportedly accounts for the vast majority of aldehyde dehydrogenase-catalyzed acetaldehyde oxidation (reviewed in ref. 4). Since erythrocytes are continuously exposed to plasma 4-hydroxycyclophosphamide/aldophosphamide, the expectation is that some of it, known to be un-ionized and to readily pass through membranes (reviewed in ref. 1), will diffuse into erythrocytes and be detoxified when oxidized as a consequence of ALDH-1-mediated catalysis. Indeed, entry of 4-hydroxycyclophosphamide/aldophosphamide into erythrocytes and the accumulation therein of carboxyifosfamide, generated when aldoifosfamide, an analogue of aldophosphamide, is oxidized, have already been reported (5, 6). Salient features of the distribution and metabolism of cyclophosphamide transport forms in human blood are presented in fig. 1. Predictably, then, an analogue of "activated cyclophosphamide," viz. 4-hydroperoxycyclophosphamide, was found to be less effective in the presence of erythrocytes when it was used ex vivo to purge autologous bone marrow of neoplastic cells so that the purged marrow could be safely used in vivo to quickly repopulate the marrow and other organs of recipients who had been depleted of vital hematopoietic progenitor cells and their progeny as a consequence of having been treated with high-dose chemotherapy and/or radiotherapy in the interim (11).

View larger version (25K): [in this window] [in a new window]   Fig. 1.   Salient features of the distribution and metabolism of cyclophosphamide transport forms in human blood. The concentration of 4-hydroxycyclophosphamide/4-thiocyclophosphamide/aldophosphamide/aldophosphamide thiohemiacetal (HTAH) in erythrocytes exceeds that in plasma by about 1.8-fold (5). The concentrations of carboxyifosfamide and ifosfamide mustard (metabolites of another oxazaphosphorine, ifosfamide, that correspond to carboxyphosphamide and phosphoramide mustard, respectively) in erythrocytes each exceed those in plasma by about 2.6-fold (6). Given that formation of glutathione conjugates of 4-hydroxycyclophoshamide/aldophosphamide is favored in erythrocytes because the erythrocyte glutathione concentration, 2 - 3 mM, is vastly greater, about three orders of magnitude, than is the plasma glutathione concentration, 3 - 4 µM (7, 8) and that these conjugates are highly ionized at physiological pH and thus unable to cross membranes, the relatively greater concentration of HTAH in erythrocytes is as expected. Given the higher concentrations of HTAH in erythrocytes, the presence of ALDH-1 in erythrocytes but not in plasma, and the fact that both are highly ionized at physiological pH, the relatively greater concentration of carboxyifosfamide (carboxyphosphamide) and ifosfamide mustard (phosphoramide mustard) is also as expected. Not reflected in this scheme is the distribution of 4-hydroxycyclophosphamide/aldophosphamide and phosphoramide mustard between plasma proteins (reversibly bound) and plasma water. Whether 4-hydroxycyclophosphamide/aldophosphamide binds to specific sites on plasma proteins is uncertain; approximately 39% of plasma phosphoramide mustard appears to be reversibly bound to plasma proteins (reviewed in ref. 1). Uncertain, too, is whether the multidrug resistance protein (MRP), a membrane glycoprotein (about 190 kDa) belonging to the ATP-binding cassette (ABC) protein family and known to be present in erythrocytes and to promote the export of certain glutathione conjugates (9), promotes the export of 4-thiocyclophosphamide and/or aldophosphamide hemiacetal from erythrocytes. However, existing evidence suggests that MRP-mediated export of these conjugates from erythrocytes is likely to be minimal (10).

Unknown is to what extent erythrocyte ALDH-1 contributes to the overall systemic oxidation of circulating aldophosphamide, and, if the contribution is ordinarily significant, as was expected given the foregoing and the large number of erythrocytes that are present in the human body, to what extent it varies. Thus, aldehyde dehydrogenase-catalyzed oxidation (detoxification) of aldophosphamide in human erythrocytes was quantified in the present investigation to address these questions. Also quantified was aldo-keto reductase-catalyzed reduction of aldophosphamide to alcophosphamide, since l) alcophosphamide per se is without cytotoxic activity (reviewed in ref. 1), 2) aldo-keto reductases, e.g. aldehyde and aldose reductases, are known to be present in erythrocytes (12), and 3) both identified, e.g. aldehyde and aldose reductases, and unidentified aldo-keto reductases are known to catalyze the reduction of aldophosphamide to alcophosphamide (reviewed in ref. 1).

	Materials and Methods

Top Abstract Introduction Materials & Methods Results & Discussion References

Materials. 4-Hydroperoxycyclophosphamide was provided by Dr. Jörge Pohl, Asta Medica AG, Frankfurt, Germany. CM-Sephadex C-50, an isoelectric point marker kit (pI 4.5 - 6.5), and the Sigma Diagnostics Total Hemoglobin Determination Kit were purchased from Sigma Chemical Co. (St. Louis, MO). Isotonic buffered saline (Hematall) was purchased from Fisher Scientific (Fairlawn, NJ). Ampholine PAGplates (pH 4.5 - 6.5) were purchased from Pharmacia Biotechnology (Piscataway, NJ). Centriprep-100 concentrators were purchased from Amicon Division, W. R. Grace & Co. (Danvers, MA). All other materials were obtained from standard commercial sources or were prepared as described previously (2, 13).

Protein and Hemoglobin Determinations. Protein content was quantified by the Coomassie brilliant blue dye binding assay (17) with the aid of the commercially available Bio-Rad Protein Assay Reagent; bovine serum albumin was used as the standard. The hemoglobin content of washed erythrocytes was determined in duplicate for all samples by the cyanmethemoglobin protocol (18) using the commercially available Sigma Diagnostics Total Hemoglobin Determination Kit with methemoglobin as the standard.

Enzyme assays. Aldehyde dehydrogenase activity was quantified spectrophotometrically, essentially as described previously (2, 14). The reaction mixture (1 ml, pH 8.2) contained 4 mM NAD, 32 mM tetrasodium pyrophosphate, 0.1 mM pyrazole, 5 mM glutathione, 1 mM EDTA, the substrate of interest, and a hemoglobin-depleted erythrocyte lysate concentrate. The reaction was initiated by addition of aldehyde and was followed at 37°C by monitoring the appearance of NADH at 340 nm with a Beckman DU-70 recording spectrophotometer.

Isoelectric Focusing, Enzyme Activity Staining, and Immunoblot Analysis. Isoelectric focusing and staining for protein and aldehyde dehydrogenase activity was carried out as previously described (2) except that narrow range Ampholine PAGplates (pH 4.5 - 6.5) were used, the cathode and anode buffers were 0.1 M -alanine and 0.1 M glutamic acid/0.5 M phosphoric acid, respectively, and a narrow range standard protein pI marker kit was used to assign the pI values. Acetaldehyde (4 mM) used as substrate to visualize aldehyde dehydrogenase activity on the focused gels. Immunoblot analysis was carried out essentially as described previously (13), except that the antibody and the blocking agent (5% instant nonfat milk) were in 0.01 M potassium phosphate buffer, pH 7.2, containing 0.1 M NaCl (Buffer B), and all washes were with Buffer B.

Statistics. The Student's t-test was used to determine whether there were significant differences between groups.

	Results and Discussion

Top Abstract Introduction Materials & Methods Results & Discussion References

Aldehyde dehydrogenase-catalyzed oxidation of acetaldehyde (160 µM) occurred at a mean rate of 220 nmol/min/g Hb (6.8 atmol/min/rbc), fig. 2. This value is in good agreement with those reported previously (20-22). Also in agreement with the reports of others (reviewed in refs. 3 and 4), catalytic activity distributed over an approximately 3-fold range (fig. 2) and was independent of the donors sex (p > 0.05, data not presented).

View larger version (23K): [in this window] [in a new window]   Fig. 2.   NAD-linked aldehyde dehydrogenase-catalyzed oxidation of aldophosphamide and acetaldehyde in hemoglobin-depleted erythrocyte lysate concentrates prepared from human blood samples. NAD (4 mM) -linked enzyme-catalyzed oxidation of aldophosphamide (160 µM) and acetaldehyde (4 mM) was quantified by spectrophotometric assay as described in Materials and Methods. Points are means of duplicate determinations made on single blood samples taken from each of 35 subjects, 20 of whom were males and 15 of whom were females. Catalytic rates are expressed per gram Hb (top) and rbc (bottom). Values in the tables are the means ± SD and ranges (parentheses) of these determinations.

Aldehyde dehydrogenase-catalyzed oxidation of aldophosphamide (160 µM) occurred at mean rates of 162 nmol/min/g Hb and 5.0 atmol/min/rbc (fig. 2). Again, catalytic activity distributed over an approximately 3-fold range. The rate of aldehyde dehydrogenase-catalyzed oxidation of aldophosphamide obtained for individual erythrocyte samples was consistently about 70% that of aldehyde dehydrogenase-catalyzed oxidation of acetaldehyde, a value that is significantly different (p < 0.05) from that, viz. 106%, obtained with purified hepatic ALDH-1 under identical assay conditions.²

Subjecting the 35 hemoglobin-depleted erythrocyte lysate concentrates to isoelectric focusing invariably revealed the presence of one major and two minor bands semiquantifying enzyme-catalyzed oxidation of acetaldehyde; representative findings are shown in fig. 3. The pI value of the major band was 5.2 (fig. 3, lane b), a value that is identical to that previously reported for human liver ALDH-1 (2, 23). The pI values of the minor bands were 5.15 and 5.08. These values do not correspond to those of any known aldehyde dehydrogenases. Immunoblot analysis showed that all three bands were recognized by an antibody preparation raised to purified human liver ALDH-1 (fig. 3, lane c). It is likely that the two minor bands are slightly degraded products of ALDH-1 that differentially retain their ability to catalyze acetaldehyde oxidation as compared with their ability to catalyze aldophosphamide oxidation, thereby providing an explanation for the discrepancy noted in the previous paragraph.

View larger version (65K): [in this window] [in a new window]   Fig. 3.   Isoelectric focusing and immunoblot analysis of an erythrocyte lysate concentrate prepared from a human blood sample. A hemoglobin-depleted erythrocyte lysate concentrate was subjected to isoelectric focusing and immunoblot analysis. The amount of concentrate loaded onto the gel was sufficient to generate 5-10 nmol NADH/min as determined by a spectrophotometric aldehyde dehydrogenase assay when acetaldehyde (4 mM) was used as the substrate. Lane a, pI standards visualized, after isoelectric focusing, with the aid of Coomassie Brilliant Blue R-250. Lane b, aldehyde dehydrogenase(s) in the erythrocyte lysate concentrate visualized, after isoelectric focusing, with the aid of a nitroblue tetrazolium-based enzyme activity stain. Lane c, ALDH-1 in the erythrocyte lysate concentrate visualized, after isoelectric focusing and electrotransfer to Immobilon-PVDF transfer membranes, with the aid of polyclonal antibodies raised to purified liver ALDH-1. Additional details are presented in Materials and Methods.

These observations are consistent with the notion that erythrocyte ALDH-1-catalyzed oxidation of aldophosphamide contributes significantly to the overall systemic detoxification of circulating aldophosphamide. Pharmacologically relevant plasma concentrations of 4-hydroxycyclophosphamide/aldophosphamide are in the range of 0.1-10 µM (reviewed in ref. 1). The K_m value for ALDH-1-catalyzed oxidation of aldophosphamide is 52 µM (2). Thus, at a saturating NAD concentration and an aldophosphamide concentration of 1 µM, erythrocyte ALDH-1 would be expected to catalyze the oxidation of aldophosphamide at a rate that is 2.5% of that at which it would catalyze the reaction if the aldophosphamide concentration were 160 µM, i.e. at a rate of 0.125 atmol/min/rbc. There are approximately 3.3 × 10¹³ erythrocytes in an average 70-kg male. Collectively, they would be expected to oxidize 4.1 µmol aldophosphamide/min at saturating concentrations of NAD. The NAD level in erythrocytes is about 50 µM (24). The K_m value for aldehyde-linked ALDH-1 catalyzed reduction of NAD is also about 50 µM (25). Thus, erythrocyte ALDH-1-catalyzed oxidation of circulating aldophosphamide would be at about 2 (range: 1.1-3.5) µmol/min when the aldophosphamide concentration is 1 µM. Not taken into account in these calculations is that 4-hydroxycyclophosphamide/4-thiocyclophosphamide/aldophosphamide/aldophosphamide thiohemiacetal is concentrated in erythrocytes (fig. 1). Systemic detoxification of aldophosphamide is generally viewed as being effected largely by hepatic enzymes, principally ALDH-1. Assuming a typical weight of 1500 g in an average 70-kg man, the liver would be expected to catalyze the oxidation of aldophosphamide at a rate of about 72 µmol/min when the aldophosphamide concentration is 1 µM, given that the K_m defining this reaction is 52 µM, that catalysis occurs at a rate of 2.1 µmol/min/g liver when the aldophosphamide concentration is 160 µM and the NAD concentration is saturating (2), and that the hepatic NAD concentration is about 500 µM (26). However, hepatic oxidation of circulating aldophosphamide is probably much less rapid than this estimate because the rate at which it occurs depends, in large part, on the flow-limited rate of aldophosphamide delivery to the liver by the blood. Blood flow to the liver is typically about 1500 ml/min. Thus, at a concentration of 1 µM, delivery of aldophosphamide to the liver by the blood occurs at a rate of 1.5 µmol/min, a maximum clearance rate that is less than the aggregate erythrocyte mean clearance rate of 2.0 µmol/min, vide supra.

Catalysis of cyclophosphamide hydroxylation to yield 4-hydroxycyclophosphamide/aldophosphamide is largely, if not exclusively, by hepatic microsomal mixed-function oxidases (reviewed in ref. 1). Thus, it is likely that some of the aldophosphamide is oxidized to carboxyphosphamide before it ever leaves the liver and enters the general circulation. At first glance, this likelihood is seemingly inappropriately omitted from the foregoing deliberations. However, only the aldophosphamide that leaves the liver and enters the general circulation is ordinarily of any major importance since only it has any therapeutic or toxic potential towards any organ/tissue/cell other than the liver. In that context, then, the omission is justified.

Erythrocyte ALDH-1-effected oxidation of other aldehydes to their corresponding acids may also be of pharmacological and/or physiological significance since a wide variety of aldehydes are known to be substrates for ALDH-1 (reviewed in refs. 3, 27).

Also evident from these experiments is that care must be taken to remove all erythrocytes from tissues/organs that are to be evaluated with regard to the levels of ALDH-1 therein.

As judged by altered electrophoretic and/or kinetic properties, variants of ALDH-1 have been identified in human tissues, although the incidence is apparently very rare (reviewed in ref. 3). Unknown is whether the variant forms of ALDH-1 catalyze the oxidation of aldophosphamide. As judged by isoelectric focusing and catalytic activity, there were no variants in our sample population of 35.

Erythrocyte ALDH-1 activity is apparently reduced or inhibited in a number of populations, e.g. in chronic alcoholics (reviewed in ref. 4), cigarette smokers (28, 29), snuff users (28), and users of birth control pills (30). Further, a number of frequently used drugs such as nitrate-ester antianginals, sulfonylurea hypoglycemics, certain cephalosporins, and disulfiram are known to inhibit ALDH-1, while diazepam and chlordiazepoxide seem to activate it (reviewed in refs. 4 and 31). Thus, there is the potential for clinically favorable/adverse drug interactions. Interindividual differences in the balance of urinary aldophosphamide metabolites have been reported; in some cases, there was a complete absence of carboxyphosphamide, suggesting a deficiency in relevant aldehyde dehydrogenase activity (reviewed in ref. 1).

Given: 1) the estimated minimum mean contribution of erythrocyte ALDH-1 to the systemic detoxification of circulating 4-hydroxycyclophosphamide/aldophosphamide, 2) the 3-fold variation in constitutive levels of erythrocyte ALDH-1, 3) the potential for activating and inhibiting this enzyme, and 4) the inverse relationship between ALDH-1 levels and AUC values for 4-hydroxycyclophosphamide/aldophosphamide, and, therefore, the therapeutic efficacy of cyclophosphamide (1), it follows that some of the variation in the therapeutic efficacy of cyclophosphamide experienced clinically may be the consequence of variation in the functional ALDH-1 content of erythrocytes. In that case, dosage adjustments based on the functional ALDH-1 content of erythrocytes may be warranted. These notions are, however, yet to be directly validated experimentally.

Scatter plots of individual values obtained for NADPH-linked aldo-keto reductase-catalyzed reduction of aldophosphamide (160 µM) are shown in fig. 4. Mean rates were 8.6 nmol/min/g Hb and 0.3 atmol/min/rbc, and were independent of the donors sex (p > 0.1; data not presented).

View larger version (18K): [in this window] [in a new window]   Fig. 4.   NADPH-linked aldo-keto reductase-catalyzed reduction of aldophosphamide in hemoglobin-depleted erythrocyte lysate concentrates prepared from human blood samples. NADPH (160 µM) -linked enzyme-catalyzed reduction of aldophosphamide (160 µM) was quantified by a spectrophotometric assay as described in Materials and Methods. Points are means, rounded off for clarity of presentation to 1 nmol/min/g Hb or the nearest multiple thereof (panel A) or to 0.05 atmol/min/rbc or the nearest multiple thereof (panel B), of duplicate determinations made on single blood samples taken from each of 35 subjects. Mean ± SD values and ranges (parentheses) were 8.6 ± 6.9 (1.4 - 34.1) nmol/min/g Hb and 0.3 ± 0.2 (0.1 - 1.1) atmol/min/rbc.

It is unlikely that NADPH-linked aldo-keto reductase-catalyzed reduction of aldophosphamide in erythrocytes contributes significantly to the systemic detoxification of aldophosphamide for several reasons: 1) in erythrocytes, reduction occurs at a rate approximately one order of magnitude slower than that of NAD-linked ALDH-1-catalyzed oxidation, e.g. the rate of reduction at 160 µM is about 5% of the rate of oxidation, as judged by in vitro analysis; 2) the K_m values of 0.15 and 1.6 mM (19) for aldose and aldehyde reductase-catalyzed reduction of aldophosphamide, respectively, are higher than the K_m value of 52 µM (2) for ALDH-1-catalyzed oxidation of aldophosphamide; thus oxidation would also predominate at pharmacological aldophosphamide concentrations; 3) reduction generates a metabolite, viz. alcophosphamide, which retains cytotoxic potential, i.e. it can be reoxidized to aldophosphamide or directly converted to phosphoramide mustard (reviewed in ref. 1); oxidation is irreversible and the resulting metabolite, viz. carboxyphosphamide, lacks cytotoxic potential (reviewed in refs. 1 and 32).

Intact erythrocytes also catalyze the oxidation and reduction of retinaldehyde, viz. 0.20 and 0.02 atmol/min/rbc, respectively.³ It is likely that erythrocyte ALDH-1 and aldose reductase contribute to the oxidation and reduction of retinaldehyde, respectively, as retinaldehyde is a relatively good substrate for both of these enzymes. K_m values are 0.3 µM (2) and < 1 µM (33), respectively. However, the physiological significance of these reactions is unclear, as the erythrocytes are not a target tissue for the retinoids, and retinol, not retinaldehyde, is the predominant circulating retinoid (34).