Cyanocobalamin [c-lactam] Inhibits Vitamin B₁₂ and Causes Cytotoxicity in HL60 Cells: Methionine Protects Cells Completely

THERE HAS HITHERTO been no convenient in vitro model in which to study the actions of vitamin B₁₂ at the cellular level. The clinical effects of vitamin B₁₂ deficiency include megaloblastic anemia, demyelination, neuropsychiatric abnormalities, and raised blood homocysteine concentrations, which may cause arterial and venous thrombosis.1-6 The relationship between the known biochemistry of vitamin B₁₂ and these clinical manifestations, particularly the megaloblastic anemia and the demyelination, remains poorly understood.

Vitamin B₁₂ is required for two biochemical reactions in humans: the methylation of homocysteine to methionine and the isomerization of methylmalonyl-CoA to succinyl-CoA, neither of which is directly concerned with DNA synthesis.7 Because the megaloblastic anemia that occurs in patients with a deficiency of vitamin B₁₂ is characterized by abnormal DNA synthesis, it has been difficult to find a biochemical explanation. A number of hypotheses have been proposed, including methylfolate trapping, failure of folate polyglutamate synthesis, and misincorporation of uracil into DNA, but none is entirely satisfactory.8-12 Similarly, the biochemical relationship between vitamin B₁₂ deficiency and demyelination remains unclear, although a failure of cellular methylation secondary to impaired methionine synthesis may be responsible.13 

A tissue culture model of vitamin B₁₂ deficiency would help to elucidate these biochemical problems. Serum contains protein-bound vitamin B₁₂, so that under normal tissue culture conditions, in which the media are supplemented with serum, deficiency cannot be achieved by simple means. Recently, a number of cobalamin analogs were subjected to in vivo testing in rats. Some of them caused marked elevations in the serum levels of homocysteine and methylmalonic acid, suggesting vitamin B₁₂ antagonism.14 Among the most active of these compounds was hydroxocobalamin [c-lactam], a B-ring derivative formed when cobalamin is heated in alkaline solution. The closely related compound, cyanocobalamin [c-lactam], was therefore examined in tissue culture, for the purpose of developing a model of vitamin B₁₂-deficient hematopoiesis.

Cell culture and growth media.HL60 cells were obtained from the American Typed Culture Collection (Rockville, MD). This cell line derives from a human myeloid leukemia and was chosen because, like normal cells and in contrast with some tumor cell lines, it was able to grow in the absence of methionine, when provided with homocysteine, vitamin B₁₂, and a source of folate.15-17 

RPMI 1640 was obtained free from vitamin B₁₂, folic acid, and methionine. “Met” medium was supplemented with 200 nmol/L 5-methyltetrahydrofolate, 3.7 nmol/L vitamin B₁₂ (standard vitamin B₁₂ ) 100 μmol/L L-methionine, 5 mg/L gentamicin, 5% dialyzed human serum, and 5% dialyzed fetal bovine serum. “Hcy” medium contained 200 μmol/L L-homocysteine thiolactone instead of methionine. Cyanocobalamin [c-lactam] and additional vitamin B₁₂ were used at a concentration of 7.4 μmol/L. Where indicated, pteroylglutamic acid was used instead of 5-methyltetrahydrofolate. Higher concentrations of both folates were used in some experiments. Where indicated, ascorbic acid at 284 μmol/L was also present in the culture medium. Cultures were performed in the dark, and medium changes were made every 48 hours.

L-homocysteine thiolactone was chosen in preference to homocysteine, because homocysteine in its reduced form is toxic to cells in culture, whereas the thiolactone is not, and there are data indicating that it is readily metabolized to homocysteine, which can then participate in the methionine synthase reaction.18-20 It has been shown repeatedly to support cell growth in the absence of methionine when the cultures contain vitamin B₁₂ and 5-methyltetrahydrofolate.17,18 21 A total of 200 μmol/L L-cystine or L-cysteine were unable to support cell growth in the presence of vitamin B₁₂ and 5-methyltetrahydrofolate with or without ascorbic acid at 284 μmol/L.

Dialysis of serum.Human and bovine sera were dialyzed in Spectra/Por cellulose ester wet membrane tubing with a molecular weight cut-off of 1,000 (Spectrum Inc, Houston, TX) for 24 hours at 4°C against two changes of normal saline and sterilized by filtration.

Preparation of cyanocobalamin [c-lactam].A solution of cyanocobalamin at 5 mg/mL in 0.1 mol/L sodium hydroxide was heated to 100°C for 10 minutes, cooled, and neutralized with potassium dihydrogen phosphate. The cobalamins were adsorbed onto a C₁₈ column (Waters Sep-Pak; Waters Assoc, Milford, MA), washed with water, eluted with methanol, evaporated to dryness, and dissolved in water. The carboxylic acids of vitamin B₁₂ were removed on an anion exchange column (AG1-X; Bio-Rad, Missauga, Ontario, Canada), and the resulting solution of cobalamins was concentrated on a Sep-Pak. Purification was performed by paper chromatography (butan-2-ol 880 mL/glacial acetic acid 8.2 mL/hydrogen cyanide 6.2 mmol/water to saturation [approximately 425 mL]). The cyanocobalamin [c-lactam] band was cut from the strip, eluted in water, and concentrated using a Sep-Pak. The concentrations of the cobalamin solutions were determined by spectrophotometry in 0.1 mol/L KCN.

Purity of the cyanocobalamin [c-lactam] preparation.Paper chromatography in butan-2-ol (1 L containing 2 mL aqueous ammonia [d = 0.88] and 1 mL 4% aqueous hydrogen cyanide) was performed. Under these conditions, cyanocobalamin [c-lactam] has R_F 1.15 relative to vitamin B₁₂.22 The preparation gave a single band in the expected position. High performance liquid chromatography (HPLC) was performed in H₂O:acetic acid:isopropanol (90:1:9) on a Waters μBondapak C₁₈ reverse phase column (10 μm particle size, 300 × 3.9 mm internal diameter) using a flow rate of 0.3 mL/min. The cobalamin solutions were injected in a volume of 10 μL containing approximately 1 nmol and detected by their absorbance at 365 nm.23 The preparation gave a single peak (retention time, 19.05 minutes) in a clearly different position from that of cyanocobalamin (retention time, 22.06 minutes).

Preparation of purified 5-methyltetrahydrofolic acid.5-methyltetrahydrofolate was obtained as the disodium salt (Sigma Chemical Co, St Louis, MO) and dissolved in 0.2 mol/L ammonium acetate containing 10 mmol/L 2-mercaptoethanol. Two milligrams was placed on a 40 × 1-cm column of DEAE-Sephadex A-25 equilibrated in 0.2 mol/L ammonium acetate containing 10 mmol/L 2-mercaptoethanol. The material was eluted in the dark at 4°C using a linear gradient of ammonium acetate containing 10 mmol/L 2-mercaptoethanol. A total of 200 mL of 0.2 mol/L ammonium acetate was placed in the mixing chamber, and 200 mL of 2 mol/L ammonium acetate was placed in the reservoir. The flow rate was 10 mL/h, and fractions of 2.5 mL were collected. Absorbance was recorded in the 235 to 340 nm range. The concentration of 5-methyltetrahydrofolate was determined by spectrophotometry, and the solution lyophilized. Finally, the purified 5-methyltetrahydrofolate was dissolved to 0.1 mg/mL in 6 mg/mL ascorbic acid and kept in liquid nitrogen.24 

The limiting concentration of 5-methyltetrahydrofolate and pteroylglutamic acid.Cells were cultured in vitamin B₁₂-, folate-, and methionine-free RPMI 1640 supplemented with 3.7 nmol/L vitamin B₁₂, 100 μmol/L L-methionine, 10% nondialyzed fetal bovine serum, and various concentrations of 5-methyltetrahydrofolate or pteroylglutamic acid. Deoxyuridine suppression tests were performed in microtiter plates as previously described, and the minimum concentration of 5-methyltetrahydrofolate or pteroylglutamic acid needed to maintain normal deoxyuridine suppression was established as 200 nmol/L for both folates.25 This concentration was used in all subsequent experiments, except where otherwise stated.

Incorporation of 5-[¹⁴CH₃]-tetrahydrofolate.The incorporation of 5-[¹⁴CH₃]-tetrahydrofolate by intact cells into trichloroacetic acid (TCA)-precipitable material was taken to reflect the activity of methionine synthase. 5-[¹⁴CH₃]-tetrahydrofolic acid (barium salt, 59 mCi/mmol) was obtained from Amersham International (Bucks, UK) and dissolved in Hank's balanced salt solution (HBSS) containing 284 μmol/L ascorbic acid to yield a 50 μCi/mL solution. For the assay, 80 μL of this label was added to 4 mL of either test or control medium. Test medium was vitamin B₁₂-, folate-, and methionine-free RPMI 1640 containing L-homocysteine thiolactone 200 μmol/L and ascorbic acid 284 μmol/L. Control medium contained 100 μmol/L methionine instead of homocysteine.20 

The cells were grown for several weeks in their stated culture medium until the time of assay. Between 10⁶ and 10⁷ cells were washed three times in HBSS and suspended to 0.5 mL in both test and control media. A total of 50 μL dialyzed fetal calf serum was added to each culture. After 24 hours of incubation at 37°C in 5% CO₂, 1.5 mL ice-cold phosphate-buffered saline (PBS) was added to each well and 200 μL was removed for cell counting. The remaining cells were washed twice in ice-cold PBS, and then 2 mL ice-cold 5% TCA was added for 15 minutes. The resulting precipitate was washed three times in ice-cold TCA and then dissolved in 2 mL 1 N NaOH. A total of 1.5 mL of this was neutralized with 90 μL glacial acetic acid and counted in a liquid scintillation counter using the external standard ratio to convert from counts per minute to disintegrations per minute. The results were expressed as pmol 5-[¹⁴CH₃]-tetrahydrofolic acid incorporated per 10⁶ cells per hour. The uptake in control wells (the non–methionine-suppressible uptake) was also recorded.

Incorporation of [1-¹⁴C] propionic acid.The incorporation of [1-¹⁴C] propionic acid by intact cells into TCA-precipitable material was taken to reflect the activity of methylmalonyl-CoA mutase. [1-¹⁴C] propionic acid (sodium salt, 51 mCi/mmol) was obtained from NEN Research Products (Wilmington, DE) as an ethanol solution 100 μCi/mL, and diluted to 25 μCi/mL in HBSS. The original method was modified because the cells survived poorly in HBSS. The cells were grown for several weeks in their stated culture medium until the time of assay. A total of 10⁶ to 10⁷ cells were then incubated in 0.5 mL of their specific culture medium with 20 μL label, giving a final concentration of 1 μCi/mL. After 24 hours at 37°C in 5% CO₂, the reaction was stopped with ice-cold PBS and the cells prepared for counting as for the incorporation of 5-[¹⁴CH₃]-tetrahydrofolate. The results were expressed as picomols of [1-¹⁴C] propionic acid incorporated per 10⁶ cells per hour.26 

In “Hcy” medium and dialyzed serum, cyanocobalamin [c-lactam] caused cell death within 96 hours in cells cultured with standard concentrations of vitamin B₁₂. Additional vitamin B₁₂ prevented this cell death and restored the growth rate (Fig 1). The cytotoxic effect of cyanocobalamin [c-lactam] was not seen with nondialyzed serum except for some slowing of growth in “Hcy” medium over the first few days of culture. High concentrations of 5-methyltetrahydrofolate (up to 20 μmol/L) did not reverse the cytotoxic effects of cyanocobalamin [c-lactam].

The growth rate of cells in “Met” medium containing 5-methyltetrahydrofolate and dialyzed serum was unaffected by cyanocobalamin [c-lactam] during 4 months of continuous culture (Fig 2). To ensure that the prolonged growth of cells in “Met” medium containing cyanocobalamin [c-lactam] was not facilitated by nonenzymatic oxidation of the 5-methyltetrahydrofolate and subsequent uptake of the oxidized folate via a nonmethionine synthase pathway, the cell growth experiments were repeated in the presence of ascorbic acid at 284 μmol/L. The results were identical (Fig 3).

Pteroylglutamic acid did not prevent the cytotoxic effects of cyanocobalamin [c-lactam] and, in fact, rendered cells in “Hcy” medium more susceptible to the inhibitor. Whereas 5-methyltetrahydrofolate was able to support long-term cell growth in “Hcy” medium and standard concentrations of vitamin B₁₂, pteroylglutamic acid could not do this. The cells grew normally for approximately 10 days, after which growth ceased and the cells died. Additional vitamin B₁₂ supplements maintained viability and allowed continued growth (Fig 4). High concentrations of pteroylglutamic acid (up to 20 μmol/L) did not reverse the cytotoxicity of the inhibitor.

To determine whether the cytotoxic effect of the inhibitor was caused by methionine deficiency and to gauge whether the protective effect of nondialyzed serum was caused by its methionine content, we observed the effect of adding back small quantities of methionine to HL60 cells grown in “Hcy” medium, dialyzed serum, and cyanocobalamin [c-lactam]. The methionine concentration in culture media containing nondialyzed serum was 2.8 μmol/L. Adding back as little as 2 μmol/L methionine to medium containing dialyzed serum and the inhibitor allowed cells to proliferate, but the full protective effect of methionine did not occur until the concentration exceeded 20 μmol/L (Fig 5). Above this concentration of methionine, cyanocobalamin [c-lactam] did not inhibit cell growth.

Cyanocobalamin [c-lactam] caused significant impairment of 5-[¹⁴CH₃]-tetrahydrofolate incorporation when the cells were grown in either “Met” or “Hcy” culture medium (Fig 6). This was reversed by the additional concentrations of vitamin B₁₂ in the culture medium. In the presence of standard concentrations of vitamin B₁₂, the incorporation of 5-[¹⁴CH₃]-tetrahydrofolate was higher when HL60 cells were grown in “Hcy” than in “Met” medium. When the cells were cultured with the additional concentrations of vitamin B₁₂, 5-[¹⁴CH₃]-tetrahydrofolate incorporation was stimulated, and this difference was lost.

In an attempt to determine whether some of the incorporation of 5-[¹⁴CH₃]-tetrahydrofolate might be taking place via a pathway other than methionine synthase, the uptake of label was also determined in the presence of excess methionine in the assay system. Although methionine synthase remains active in vivo in the presence of excess methionine, the dilution of the labeled end product (methionine) by the excess of exogenous unlabeled methionine should greatly reduce the incorporation of label into trichloroacetic acid-precipitable macromolecules, unless the newly synthesized methionine is channeled directly into cellular methylation reactions or protein synthesis, rather than through a general metabolic pool. Substantial uptake of 5-[¹⁴CH₃]-tetrahydrofolate did indeed continue in the presence of excess methionine, as shown by the solid bars in Fig 6. Only a small proportion of this is nonspecific uptake, because the incorporation in “Hcy” medium with cyanocobalamin [c-lactam] is very low.

The incorporation of [1-¹⁴C] propionic acid by HL60 cells was higher when they were cultured in “Met” than “Hcy” medium. Cyanocobalamin [c-lactam] had little effect on cells in “Met” medium, but suppressed [1-¹⁴C] propionic acid incorporation by cells grown in “Hcy” medium. The additional vitamin B₁₂ had no effect on cells in “Met” medium, but stimulated incorporation by those in “Hcy” medium (Fig 7).

Megaloblastic morphology did not develop at any time during the cell culture period.

These results show that cyanocobalamin [c-lactam] causes growth inhibition and cell death in vitro by antagonizing the functions of vitamin B₁₂. Nonspecific cytotoxicity is excluded by the lack of growth inhibition in the presence of methionine and because the effects are reversed by vitamin B₁₂. Consistent with this hypothesis, cyanocobalamin [c-lactam] inhibited the incorporation of both 5-[¹⁴CH₃]-tetrahydrofolate and [1-¹⁴C] propionic acid, consistent with antagonism of both methionine synthase and methylmalonyl-CoA mutase, reversible by vitamin B₁₂. The data are also consistent with previous observations on the effects of intravenous hydroxocobalamin [c-lactam] in rats, in which it caused elevated serum concentrations of both methylmalonic acid and homocysteine, suggesting antagonism of vitamin B₁₂.14 Cyanocobalamin [c-lactam] is unlikely to be a direct inhibitor of methionine synthase, because it is an efficient cofactor for the apoenzyme in a cell-free system.27 It more probably affects some other aspect of the intracellular metabolism of vitamin B₁₂, such as uptake, transport, or reduction.

Small concentrations of methionine prevented cytotoxicity, and allowed the cells to proliferate at a reduced rate. Methionine at concentrations of 20 μmol/L or more protected cells completely from the lethal effects of cyanocobalamin [c-lactam]. The cells divided at a normal rate for at least 4 months and showed no morphologic evidence of megaloblastosis. This indicates that methionine deficiency is responsible for the early cytotoxic effects of cyanocobalamin [c-lactam] and that, when the methionine concentration in the culture medium exceeds approximately 20 μmol/L, the remethylation of homocysteine is unimportant for cell growth.

Because HL60 cells grew normally for an indefinite period in the presence of cyanocobalamin [c-lactam] when provided with sufficient methionine, it seems reasonable to conclude that they did not develop intracellular folate deficiency secondary to the inhibition of vitamin B₁₂. Previous data in experimental animals with a deficiency of vitamin B₁₂ show that methionine restores the diminished levels of hepatic folates and reduces urinary excretion of formiminoglutamic acid.28-36 The same effect of methionine is seen on formiminoglutamic acid excretion in humans with a deficiency of vitamin B₁₂, and there is also improvement in the impaired incorporation of deoxyuridine into the DNA of bone marrow cells.37,38 However, methionine does not produce hematological responses in patients with vitamin B₁₂ deficiency.39 In the nitrous oxide–treated rat, in which methionine synthase is approximately 85% inhibited, the methyl group of 5-[¹⁴CH₃]-tetrahydrofolate can be slowly metabolized to CO₂ and the folate incorporated into the cell, an effect that is enhanced by methionine.40-42 

There has been controversy as to whether methionine acts by relieving a “methylfolate trap” (Fig 8) or whether it promotes metabolism of 5-methyltetrahydrofolate through a pathway alternative to methionine synthase.43 Oxidation of 5-methyltetrahydrofolate to 5,10-methylenetetrahydrofolate by quinonoid dihydrobiopterin has been reported, which is a potential candidate for such a pathway.44 The present results are against the methylfolate trap hypothesis and favor the existence of an alternative pathway, for the following reasons:

(1) The methylfolate trap hypothesis predicts that, when methionine synthase is inhibited, the growth rate and viability of cells should be enhanced when the folate supplement in the medium does not carry a methyl group at the 5 position. Our results were the opposite. Pteroylglutamic acid, a nonmethylated folate, did not rescue cells from the effects of cyanocobalamin [c-lactam], even in very high concentrations, and was unable to support prolonged cell growth in “Hcy” medium.

(2) Methionine also appeared to reverse the effects of cyanocobalamin [c-lactam] on the incorporation of [1-¹⁴C] propionic acid, which cannot be explained in terms of a folate effect. This phenomenon has not previously been observed and, if confirmed, is worthy of further study.

(3) There was substantial incorporation of label from 5-[¹⁴CH₃]-tetrahydrofolate when the assay system contained excess methionine. One explanation for this finding might be that the uptake occurred via a pathway other than methionine synthase, although channeling of the newly synthesized methionine directly into cellular methylation reactions or protein synthesis cannot be excluded.

Although these findings suggest the existence of a pathway for 5-methyltetrahydrofolate incorporation other than through methionine synthase, they do not necessarily indicate that this pathway was responsible for the protective effect of methionine against cyanocobalamin [c-lactam]. Although cells exposed to the lactam appeared to incorporate 5-[¹⁴CH₃]-tetrahydrofolate through an apparent nonmethionine-suppressible pathway more in “Met” medium than in “Hcy,” the difference was not statistically significant. However, the data in nitrous-oxide breathing rats do suggest that methionine might enhance a nonmethionine synthase pathway of 5-methyltetrahydrofolate use.42 

High concentrations of vitamin B₁₂ stimulated the incorporation of 5-[¹⁴CH₃]-tetrahydrofolate in both “Met” and “Hcy” media. Stimulation by vitamin B₁₂ is characteristic of methionine synthase in cell-free assay systems, presumably because unsaturated apoenzyme is present.27 However, studies of methionine synthase in vivo show that the majority (90% to 100%) of the enzyme is bound to vitamin B₁₂.45 This could mean that the standard concentration of vitamin B₁₂ used in our culture media was suboptimal, so that there was insufficient intracellular vitamin B₁₂ to saturate the apoenzyme, but there may also be alternative explanations, because some of the published data are inconsistent with the hypothesis that the effect of vitamin B₁₂ is simply to bind unsaturated apoenzyme.14 Cells grown in “Hcy” medium incorporated more 5-[¹⁴CH₃]-tetrahydrofolate than those in “Met” medium. This observation is consistent with previous data showing increased methionine synthase activity in cells cultured in the absence of methionine, resulting probably from the synthesis of new enzyme.21 

Spontaneous photogenic oxidation of the 5-methyltetrahydrofolate to a form able to bypass methionine synthase and enter the intracellular folate pools directly seems an improbable explanation for the data presented. All cultures were performed in the dark, and ascorbic acid, which is known to retard the oxidation of 5-methyltetrahydrofolate, had no effect on the results.46,47 The oxidation product of 5-methyltetrahydrofolate is 5-methyldihydrofolate, which is stable in solution. This compound can participate in the methionine synthase reaction, but is not a substrate for methylenetetrahydrofolate reductase, and the methyl group, therefore, cannot be oxidized to a level that could directly enter the cellular folate pools.47 Contamination of the 5-methyltetrahydrofolate by other folates able to circumvent the methionine synthase pathway is excluded because it was purified before use.

We conclude that vitamin B₁₂ is effectively antagonized in the present tissue culture model, and that the effects observed can be explained in terms of methionine deficiency. The results suggest that, in malignant HL60 cells, methylfolate is not trapped when methionine synthase is inhibited, and there may be another pathway by which methylfolate is metabolized. Although these metabolic features may be unique to the malignant state and do not disprove the folate trap hypothesis as applied to the normal blood cells on whose cultures it was developed, this model should be useful for investigating the megaloblastic state. It should be possible to manipulate the culture conditions to induce the biochemical and morphological changes of megaloblastosis, which would lead to significant insights into the understanding of megaloblastic change. It should also improve the understanding of one of the pathways of homocysteine metabolism, abnormalities of which are being increasingly recognized as important in the pathogenesis of vascular disease.