Folate and cobalamin (vitamin B₁₂) play key roles in the metabolism of all cells, particularly proliferating cells.

CHEMISTRY

The group of compounds referred to as folates consists of folic acid and its various derivatives. Folic acid (pteroylglutamic acid) is composed of a pteridine moeity, a p-aminobenzoate residue, and an L-glutamic acid residue (Fig. 41–1A). The first two together are called pteroic acid.¹ In nature, folates occur largely as conjugates in which multiple glutamic acids are linked by peptide bonds involving their γ-carboxyl groups (Fig. 41–1B). Additionally, the naturally occurring polyglutamated folates are reduced in the 5, 6, 7, and 8 positions of the pteridine ring (as described below, Fig. 41–1B). Conjugates are named according to the length of the glutamate chain (e.g., pteroylmonoglutamate, pteroyldiglutamate, pteroylhexaglutamate). Therapeutic folic acid (pteroylglutamic acid, abbreviated PteGlu, or F) has one glutamic acid and the pteridine ring is not reduced.

To form a functional compound, folate must be reduced to tetrahydrofolate (FH₄; see Fig. 41–1B). In this reduction, dihydrofolate (FH₂) is an intermediate. A single enzyme, FH₂ reductase, catalyzes both F→FH₂ and FH₂→FH₄.

The folate family consists largely of FH₄ derivatives bearing a one-carbon substituent (symbolized as FH₄-C). The varieties of FH₄-C differ with regard to the identity of the one-carbon unit and the site of its attachment to FH₄. Figure 41–2 shows one-carbon substituents of biochemical significance and their major interconversions.

These substituents are attached to FH₄ through N⁵,N¹⁰, or both (see Fig. 41–2). Specific enzymes interconvert these various FH₄ derivatives through oxidations that require nicotinamide adenine dinucleotide phosphate (NADP) and reductions that utilize the reduced form of NADP, NADPH.²

Reduced derivatives of folic acid usually are sensitive to air oxidation. An important exception is N⁵-formyl FH₄, also called citrovorum factor, leucovorin, or folinic acid, which, because of its stability, is the form preferred for clinical use.

NUTRITION

Sources

Folic acid comes from many sources. The richest vegetable sources are asparagus, broccoli, endive, spinach, lettuce, and lima beans. Each vegetable contains more than 1 mg of folate per 100 g dry weight. The best fruit sources are oranges, lemons, bananas, strawberries, and melons. Folates also are abundant in liver, kidney, yeast, mushrooms, and peanuts. Since the advent of folic acid fortification of the food supply, the median daily intake of folate from an average American diet is estimated to be 350 mcg.³ Foods are readily depleted of folate by excessive cooking, especially with large amounts of water, discarded before ingestion.

Daily Requirements

In the normal adult, the minimum daily requirement for folic acid is approximately 50 mcg. The average diet contains many times this amount, but some of the folate may be unavailable. Accordingly, the officially recommended dietary allowance of food folate for an adult is 0.4 mg.³ This figure is derived through considerations of the nutrient requirements to satisfy the needs of 97 to 98 percent of healthy individuals and the relative differences in absorption and bioavailability between dietary folate and the more bioavailable synthetic folic acid. One microgram of food folate is the dietary equivalent of 0.6 mcg folic acid added to food. The body is thought to contain approximately 5 mg of folate.⁴ When folate intake is reduced to 5 mcg/day, megaloblastic anemia develops in approximately 4 months.⁵

Folic acid requirements increase in hemolytic anemia, leukemia, and other malignant diseases, in alcoholism⁶ and during growth; in pregnancy and during lactation requirements increase threefold to sixfold.⁷ Adequate folate supplies are particularly important in pregnant and lactating women, in whom the recommended daily allowance is increased to 600 and 500 mcg/day, respectively, to meet requirements.⁸

METABOLISM

Folate-Dependent Enzymes

FH₄ is an intermediate in reactions involving the transfer of one-carbon units from a donor to an acceptor. Table 41–1 summarizes the metabolic systems of animal tissues known to require folic acid coenzymes.

One-carbon units enter the folate pool principally via the serine hydroxymethyltransferase (SHMT) reaction which requires pyridoxal phosphate as a cofactor.⁹ There are both cytoplasmic (SHMT1) and mitochondrial (SHMT2) isoforms of SHMT which impart a state of functional redundancy for this important enzyme, which is the primary source of one-carbon units for purine and pyrimidine biosynthesis. The cytoplasmic form can undergo sumoylation during S-phase, allowing nuclear localization.¹⁰

Among the several one-carbon transfers mediated by folic acid, the transfer that appears to be the most important clinically is the methylation of deoxyuridylate to thymidylate, catalyzed by the enzyme thymidylate synthase.^2,10,11 This reaction is an essential step in the synthesis of DNA (Fig. 41–3). In carrying out this reaction, N⁵,N¹⁰-methylene FH₄ simultaneously transfers and reduces a one-carbon group, itself serving as the hydrogen donor for the reduction.¹² The reaction generates FH₂, which must be reduced again to FH₄ by FH₂ reductase and NADPH before it can again be utilized as a coenzyme:

where dUMP = deoxyuridine monophosphate; dTMP = deoxythymidine monophosphate; and NADP = nicotinamide adenine dinucleotide phosphate. Limitation of thymidylate synthesis in folic acid deficiency causes incorporation of uracil instead of thymine into DNA.¹³

A key enzyme, methylenetetrahydrofolate reductase (MTHFR) regulates the distribution of reduced folates by controlling the rate of NADPH-mediated conversion of N⁵,N¹⁰-methylene FH₄ to N-methyltetrahydrofolate. Because N⁵,N¹⁰-methylene FH₄ is the obligate one-carbon donor for thymidylate synthesis, its conversion to N-methyltetrahydrofolate serves as a brake on DNA synthesis and repair, while diverting more folate to the methionine synthase reaction (see Fig. 41–2). The activity of MTHFR therefore serves as a checkpoint for intracellular folate trafficking and distribution, and thus becomes more critical in states of folate depletion.

A polymorphic form of MTHFR, MTHFR 677C→T, is of some clinical importance. The mutation results in a thermolabile form of the enzyme with a higher K_m (Michaelis-Menten dissociation constant) for its methylene-FH₄ substrate. Retardation of the folate methylation cycle makes more methylene-FH₄ available for thymidylate synthesis (Fig. 41–4), and also affects the levels of homocysteine, an amino acid whose rate of production depends on both folate and cobalamin.

Folate deficiency diminishes purine biosynthesis by slowing (1) the folate-dependent formylation of glycinamide ribotide to N-formylglycinamide ribotide, the reaction that places the C-8 in the purine ring, and (2) the folate-dependent conversion of 5-amino-4-imidazole carboxamide ribotide (AICAR) to 5-formamido-4-imidazole carboxy-amide ribotide, the reaction that places the C-2 in the purine ring.¹⁴ Additional reactions dependent on biopterin, a nonfolate pteridine derivative, that are of potential metabolic importance are hydroxylation of phenylalanine to tyrosine, oxidation of long-chain alkyl ethers of glycerol to fatty acid, hydroxylation of tryptophan to 6-hydroxytryptophan (a precursor of serotonin), 17α-hydroxylation of progesterone,¹⁵ and production of nitric oxide.¹⁶ Tetrahydrofolic acid is weakly active in some of these systems in vitro¹⁷; whether it plays any such role in vivo is unknown.

Significance of Folylpolyglutamates

Intracellular folates exist primarily as polyglutamate conjugates.¹⁸ Approximately 75 percent of the folate in human erythrocytes and leukocytes is polyglutamylated.¹⁹ Plasma folate consists largely of the monoglutamate N⁵-methyltetrahydrofolate and is transported into the cells in this form.²⁰ Inside the cells, the polyglutamate chain is sequentially built up by an ATP-dependent folylpoly-γ-glutamyl synthase (FPGS).²¹ The activity of human FPGS depends strongly on the form of the folate substrate, declining in the order FH₄ > N¹⁰-formyl FH₄ > N⁵-methyltetrahydrofolate, toward which the enzyme is almost inert.²² In humans, conjugated folates carry on average seven to eight glutamyl residues.²³ Intracellular folylmonoglutamates leak out of the cells at a fairly rapid rate whereas polyglutamates do not, presumably because of the highly charged polyglutamate tail.²⁴ Therefore, attachment of the polyglutamate chain is essential for retaining folates within cells. Folylpolyglutamates are superior to monoglutamates as substrates for folate-dependent enzyme reactions.¹⁹

PHYSIOLOGY

Intestinal Absorption

The duodenum and proximal jejunum is the principal site of folate absorption. Absorption of a dose of either monoglutamate or polyglutamate forms of folate begins within minutes. Peak plasma levels are reached in 1 to 2 hours. Because only folylmonoglutamate appears in plasma, all folylpolyglutamates must first be hydrolyzed by the enzyme glutamate carboxypeptidase II (folate hydrolase)²⁵ during absorption across the intestine.²⁶ This enzyme (previously referred to as “conjugase”) is located on the brush-border membrane and plays an important role in the intestinal absorption of folate.²⁷ Folylpolyglutamate is hydrolyzed at the brush-border of the intestinal cell (Fig. 41–5). Folate hydrolase purified from human jejunum catalyzes the Zn²⁺-dependent deconjugation of folate polyglutamates ranging from PteGlu₂ to at least PteGlu₇.²⁸ It is an exopeptidase that successively removes single glutamate residues from the end of the polyglutamate chain, ultimately yielding the folylmonoglutamate. The monoglutamate forms are then taken up by one of two folate-specific transporters located on the apical brush-border, the reduced folate carrier (RFC) or the proton-coupled folate transporter (PCFT).²⁵ Although RFC has a pH optimum of 7.4, PCFT is a high-affinity folate transporter that uses a proton-coupled system to facilitate folate absorption and shows maximum transport activity at a low pH.^25,29 This property is consistent with the observation that cancer cells retain a high affinity for the new-generation antifolate pemetrexed. Defects in the PCFT are the underlying cause of hereditary folate malabsorption.³⁰

Folate hydrolases also are found outside the intestine. For example, human plasma contains sufficient hydrolase activity to convert polyglutamates containing more than three glutamyl residues to monoglutamates. Other γ-glutamyl hydrolases appear to be lysosomal carboxypeptidases³¹ that are not involved in absorption of folates from the intestine but that play a role in the release of folate from storage sites in the liver and kidney.²⁵

Folate monoglutamates are actively transported across the intestinal epithelium by PCFT-mediated transport (K_m = 1 to 2 μM) that is independent of Na⁺, K⁺, and transmembrane potential.³² The mechanism uses the pH gradient between the jejunal lumen (pH ~6) and the interior of the epithelial cell to drive folate into the cell against a concentration gradient.³³ Passive transport also may occur.³⁴ In the intestinal cell, the absorbed folate monoglutamates are reduced if necessary, and then converted to N⁵-methyltetrahydrofolate (some N¹⁰-formyl FH₄ also is made) and transported into the bloodstream without further change.³⁵

Folate undergoes an enterohepatic cycle in which it is first secreted against a concentration gradient into the bile, appearing there chiefly as N⁵-methyltetrahydrofolate monoglutamate, and then is reabsorbed from the small intestine.³⁶ Bile contains approximately two to 10 times the folate concentration of normal serum, with biliary excretion accounting for up to 0.1 mg of folate per day. This quantity is sufficiently large that interruption of the enterohepatic cycle by biliary diversion causes serum folate levels to fall by more than 50 percent in less than 1 day.³⁷ The enterohepatic cycle has been proposed to redistribute folate between hepatic stores and peripheral tissues according to the state of the exogenous folate supplies.³⁸

METABOLISM

Tritiated folylmonoglutamate (³H-F) administered intravenously is almost completely removed from the bloodstream in a few minutes.³⁹ Uptake involves two classes of folate-binding proteins⁴⁰: high-affinity folate receptors⁴¹ that concentrate folate in intracellular vesicles and a membrane folate transporter that transports folate from the vesicles into the cytosol. The high-affinity receptors, which are attached to the outer surface of the cell membrane by glycosyl-phosphatidylinositol linkages and lack an intracytoplasmic portion,⁴² bind very tightly (K_d [distribution coefficient] in the nanomolar range) to most physiologic folate monoglutamates,⁴³ particularly N⁵-methyltetrahydrofolate, the major circulating folate.⁴⁴ The folate receptors exist in various isoforms (of which α and β are the most important). Folate receptor-α, despite its missing cytoplasmic extension, is effective in mediating endocytosis.⁴⁵ Their very high affinity enables the receptors to take up N⁵-methyltetrahydrofolate from the plasma, even at its ambient concentration of approximately 10 nM. The membrane folate transporter is a probenecid-inhibitable organic anion carrier that, among other functions, carries reduced folates and methotrexate in and out of the cytoplasm.⁴⁰ Its K_m for folate is in the micromolar range. Once internalized, the folates are retained by the cells partly through polyglutamylation⁴⁶ as well as through tight association with a set of intracellular folate-binding proteins.⁴⁷ Three of these proteins are enzymes involved in methyl group metabolism: sarcosine dehydrogenase and dimethylglycine dehydrogenase (mitochondrial)⁴⁸ and glycine N-methyl transferase (cytosolic).⁴⁹ Why these enzymes bind folate so avidly or whether this binding affects overall methyl group metabolism is unknown, although glycine N-methyl transferase is speculated to regulate methyl group metabolism by controlling the tissue concentration of S-adenosylhomocysteine (SAH), one of its reaction products and a potent inhibitor of most methyltransferases.

Folates have been found in all body tissues that have been analyzed. The principal form of the vitamin in tissues and in blood appears to be the N⁵-methyl form.⁵⁰ The total folate pool turns over very slowly.⁵¹ Degradation accounts for a portion of this turnover. p-Aminobenzoylglutamate has been identified as a breakdown product. The fate of the pteridine moiety is unknown.

FOLATE-BINDING PROTEINS OF SERUM AND MILK

The soluble folate-binding proteins of serum and milk are high-affinity folate receptors that are released from cell membranes by proteolysis.⁵² These proteins can be detected in approximately 15 percent of normal individuals⁵³ and are found at increased levels in some pregnant women, women taking oral contraceptives, folate-deficient alcoholics (but not patients with cobalamin deficiency),⁵⁴ and patients with uremia, hepatic cirrhosis, and chronic myelogenous leukemia.⁵⁵ In normal subjects, the proteins are approximately two-thirds saturated and have a total folate-binding capacity of approximately 175 pg/mL of serum.⁵⁶ The proteins may not be detected in some subjects because of complete saturation of the proteins with endogenous unlabeled folate.⁵⁷ Serum folate-binding protein has an Mr of 40,000 and prefers oxidized to reduced folates.⁵⁵

Folate-binding proteins have been found in milk and in normal granulocytes.⁵⁸ Folate bound to the milk folate binder in suckling animals is absorbed chiefly in the ileum⁵⁹ rather than the jejunum, the principal site of absorption of free folate. The milk folate binder, a glycoprotein, also promotes folate transport into the liver via the asialoglycoprotein receptor.⁶⁰ The milk folate binder is speculated to protect an infant’s folate supply by preventing bacteria from sequestering the vitamin away from the intestinal absorptive surface. The folate-binding protein in granulocytes has been localized to the specific granules, from which it is released when the granulocytes are stimulated.

EXCRETION

Folates are both resorbed and secreted by the kidney. Resorption is accomplished by a membrane-bound high-affinity folate receptor (K_m for N⁵-methyltetrahydrofolate = 0.4 nM) located in the brush-borders of the proximal tubules.⁶¹ Filtered folate may thus be returned to the bloodstream. There is resorption of most, but not all, of the filtered folate.

In humans, intact folates and their cleavage products are excreted by the kidney at a rate of 2 to 5 mcg/day.⁶² A small percentage of parenterally administered labeled folate is recoverable in the feces and mainly represents unreabsorbed folate from the enterohepatic cycle.⁶³

ASSAY OF SERUM FOLATE

Folates are measured by chemiluminescent methods using various folate binders. These assays are identical in principle to the radioligand binding assays that they have replaced.

CHEMISTRY

Structure and Nomenclature

The cobalamin molecule has two major portions: a porphyrin-like near-planar macrocycle known as corrin, and a nucleotide that lies almost perpendicular to the corrin ring (Fig. 41–6). The corrin moiety contains four reduced pyrrole rings that bind a central cobalt atom whose two remaining coordination positions are occupied by a 5,6-dimethylbenzimidazolyl (5,6-DMB) group, below the ring and various ligands (in this case, —N in the form of CN) above the ring.⁶⁴

Compounds containing the corrin ring are known as corrinoids. The cobalamins are corrinoids whose nucleotide contains 5,6-DMB. Two connections exist between the corrin and the nucleotide: (1) a bond between the nucleotide phosphate and a side chain in ring D, and (2) a bond between cobalt and a nitrogen atom of benzimidazole. Figure 41–7 summarizes the numbering and ring designations of the corrin system.

The term vitamin B₁₂ is sometimes used as a generic term for the cobalamins. The term probably is best reserved, however, as an alternative name for cyanocobalamin, the usual therapeutic form of cobalamin.

Four cobalamins are important in animal cell metabolism. Two are cyanocobalamin (CnCbl; vitamin B₁₂) and hydroxocobalamin (OHCbl) or aquocobalamin (HOH Cbl). The other two cobalamins are alkyl derivatives that are synthesized from OHCbl and serve as coenzymes. In one, adenosylcobalamin (AdoCbl), a 5′-deoxyadenosyl replaces OH as the cobalt ligand above the ring (Fig. 41–8).⁶⁵ In the second, methylcobalamin (MeCbl), the upper ligand is a methyl group. MeCbl is the major form of cobalamin in human blood plasma.⁶⁶

NUTRITION

Sources

Cobalamin is synthesized only by certain microorganisms; animals ultimately depend on microbial synthesis for their cobalamin supply. Foods that contain cobalamin are of animal origin: meat, liver, seafood, and dairy products. Cobalamin has not been found in plants.

Daily Requirements

The average daily diet in Western countries contains 5 to 30 mcg of cobalamin, of which 1 to 5 mcg is absorbed.⁶⁷ Less than 250 ng appears in the urine; the unabsorbed remainder appears in the feces. Total body content is 2 to 5 mg in an adult,⁶⁸ with approximately 1 mg in the liver. The kidneys also are rich in cobalamin.⁶⁹ Relative to the daily requirement, body reserves of cobalamin are much larger than those of folate.

Cobalamin has a daily rate of obligatory loss of approximately 0.1 percent of the total-body pool, irrespective of the pool size. For this reason, a deficiency state does not develop for several years after cessation of cobalamin intake. The officially recommended dietary allowance (RDA) for adults is 2.4 mcg²; growth, hypermetabolic states, and pregnancy increase daily requirements. The RDA for children ages 1 to 13 years is 0.9 to 1.8 mcg. Because of insufficient data, no RDA has been established for infants. Instead, adequate intakes of 0.4 mcg for age 0 to 6 months and 0.5 mcg for age 7 to 12 months have been estimated.

ROLE IN METABOLISM

The only two recognized cobalamin-dependent enzymes in human cells are AdoCbl-dependent methylmalonyl CoA (coenzyme A) mutase and MeCbl-dependent methyltetrahydrofolate-homocysteine methyltransferase.

Methylmalonyl Coenzyme A Mutase

Methylmalonyl CoA mutase is a mitochondrial enzyme that participates in the disposal of the propionate formed during breakdown of valine, isoleucine, and odd-carbon fatty acids. The enzyme is a homodimer of a 78-kDa subunit that is encoded by a gene on chromosome 6.⁷⁰ In the reaction catalyzed by methylmalonyl CoA mutase, methylmalonyl CoA, which is produced during catabolism of propionate,⁷¹ is converted to succinyl CoA, a Krebs cycle intermediate. In the course of this reaction, a hydrogen on the methyl carbon of the substrate exchanges places with the —COSCoA group (Fig. 41–9).

The coenzyme serves as an intermediate hydrogen carrier, accepting the hydrogen from the substrate in the initial phase of the reaction and returning it to the product after migration of —COSCoA.

N⁵-Methyltetrahydrofolate-Homocysteine Methyltransferase

MeCbl participates in cobalamin-dependent synthesis of methionine from homocysteine by the enzyme N⁵-methyltetrahydrofolate–homocysteine methyltransferase. S-adenosylmethionine (SAMe) and a second enzyme, methionine synthase reductase are required for methyltransferase activity.⁷² The reductase converts the oxidized cobalt to the readily alkalizable Co⁺, which then accepts a methyl group from SAMe, a powerful biologic methylating agent, thereby restoring activity of the methyltransferase. In humans, this pathway also serves as a mechanism critical for converting N⁵-methyltetrahydrofolate to FH₄ required for synthesis of polyglutamates as well as other important one-carbon adducts of folate. The demethylation of N⁵-methyltetrahydrofolate is a prerequisite for attachment of the polyglutamate chain to newly acquired folate, which is largely taken up by the cell in the form of N⁵-methyltetrahydrofolate monoglutamate.²⁴ Nitrous oxide (N₂O) impairs the methyltransferase by oxidizing cob(I)alamin (a catalytic intermediate in the methyltransferase reaction) to cob(II)alamin. This reaction depletes MeCbl and produces a cobalamin deficiency-like state.

Nonenzymatic Metabolism

Because cobalamin has the capacity to bind cyanide, it may participate in detoxification of cyanide. Tobacco and certain foods (fruits, beans, tubers, and nuts) contain cyanide in the form of thiocyanate. Although the evidence is inconclusive, cobalamin is believed to play a role in neutralizing cyanide taken in via these substances.⁶⁴

FOLATE–COBALAMIN RELATIONSHIP

In both folate deficiency and cobalamin deficiency, the megaloblastic anemias are fully corrected by treatment with the appropriate vitamin. The megaloblastic anemia of cobalamin deficiency also is variably corrected by folic acid supplementation even if no cobalamin is given, although the remission may be partial and only temporary. Conversely, the anemia of folate deficiency is generally not helped at all by cobalamin although partial responses to high doses of cobalamin have been reported in some patients with folate deficiency.⁷³ These clinical observations indicate that the megaloblastic anemia in cobalamin deficiency actually results from an abnormality in folate metabolism.²⁴ The observation that urinary excretion of formiminoglutamic acid (FIGlu) and AICAR, normally regarded as a sign of folate deficiency, is seen occasionally in pure cobalamin deficiency⁷⁴ provides further evidence that folate metabolism is deranged by cobalamin deficiency. Two explanations have been proposed to account for the folate responsiveness of cobalamin-deficient megaloblastic anemia: (1) the methylfolate trap hypothesis, which is accepted by the majority of authorities, and (2) the formate starvation hypothesis (Fig. 41–10).

METHYLFOLATE TRAP HYPOTHESIS

The methylfolate trap hypothesis⁷⁵ is based on the fact that the folate-requiring enzyme N⁵-methyltetrahydrofolate–homocysteine methyltransferase is also dependent on cobalamin. The hypothesis states that in cobalamin deficiency tissue folates are gradually diverted into the N⁵-methyltetrahydrofolate pool because of slowing of the methyltransferase reaction,⁷⁶ the only route out of that pool for folate. As N⁵-methyltetrahydrofolate levels increase, the levels of other forms of folate decline, with a consequent fall in the rates of reactions in which those forms participate. In particular, because the MTHFR reaction is irreversible, methylene-FH₄ becomes depleted, the synthesis of dTMP is slowed, and megaloblastic anemia ensues.

In its simplest form, the hypothesis predicts that in cobalamin deficiency tissue levels of N⁵-methyltetrahydrofolate are abnormally high and those of other forms of folate are abnormally low. Although serum N⁵-methyltetrahydrofolate levels are frequently elevated in cobalamin deficiency,⁷⁷ tissue folate levels, predominantly polyglutamates, decline.⁷⁸ The decreased level appears to be related to the substrate specificity of the folate-conjugating enzyme. This enzyme works very poorly with N⁵-methyltetrahydrofolate; therefore, it is unable to carry out normal γ-glutamylation of newly internalized N⁵-methyltetrahydrofolate monoglutamate in cobalamin-deficient cells because the freshly acquired folate cannot be converted into a suitable substrate (i.e., free FH₄ or formyl FH₄). Thus, although sequestration of tissue folates in an expanded N⁵-methyltetrahydrofolate pool may account for some of the effects of the blockade in methyltransferase activity, the major problem seems to be a failure to convert newly acquired folate into a form that can be retained by the cell. The upshot is development of tissue folate deficiency as the unconjugated folate leaks out (see Fig. 41–10). The whole process is aggravated by a drop in tissue levels of SAMe as the methionine supply is curtailed because of the diminished activity of the methyltransferase.⁷⁹ SAMe, which is necessary for methyltransferase activity, is also a powerful inhibitor of N⁵,N¹⁰-methylene FH₄ reductase MTHFR,⁸⁰ the enzyme responsible for production of N⁵-methyltetrahydrofolate. The relief of this inhibition as SAMe levels fall accelerates the flow of folates toward N⁵-methyltetrahydrofolate, further aggravating the metabolic imbalance resulting from impairment in methyltransferase activity.

This problem could be overcome if N⁵-methyltetrahydrofolate were converted into a substrate for the conjugating enzyme by another route. In theory, this could be accomplished by reversal of the N⁵,N¹⁰-methylene FH₄ reductase reaction. For practical purposes, however, the N⁵,N¹⁰-methylene FH₄ reductase reaction is irreversible in vivo.⁸¹

FORMATE STARVATION HYPOTHESIS

This hypothesis holds that formate starvation is the basis for folate-responsive megaloblastic anemia of cobalamin deficiency.⁸² This theory is based on the diminished capacity of cobalamin-deficient lymphoblasts to incorporate formaldehyde into purine and methionine⁷⁹ and on experiments showing that N⁵-formyl FH₄ is more effective than FH₄ at correcting some of the abnormalities in folate metabolism seen in cobalamin deficiency.⁸³ The hypothesis states that with the decrease in methionine production in cobalamin-deficient conditions, the generation of formate is depressed (because normally the methyl group of excess methionine is rapidly oxidized to formate),⁸⁴ leading to a decline in the production of N⁵-formyl FH₄.

INTESTINAL ABSORPTION

Intrinsic Factor

Intrinsic factor is one of a number of binding proteins in which cobalamin is ensconced as it makes its way through the body (Table 41–2). Intrinsic factor is needed for the absorption of cobalamins taken orally at physiologic dosage levels. Human intrinsic factor is a glycoprotein (Mr approximately 44,000) encoded by a gene on chromosome 11.⁸⁵ It has binding sites for cobalamin and a specific ileal receptor, the former situated near the carboxy-terminus and the latter near the aminoterminus of the intrinsic factor molecule.⁸⁶ Binding to cobalamin is very tight, and involves the 5,6-DMB lower axial ligand of the molecule. This specificity allows for the exclusion of other noncobalamin corrinoids during the tightly regulated absorptive process.⁶⁴ The entrapment of the vitamin alters the conformation of intrinsic factor, producing a more compact form that is resistant to proteolytic digestion.

In humans, intrinsic factor is synthesized and secreted by the parietal cells of the cardiac and fundic mucosa.⁸⁷ Secretion of intrinsic factor usually parallels that of hydrochloric acid (HCl). It is enhanced by the presence of food in the stomach, vagal stimulation, histamine, and gastrin. Gastric juice also contains other cobalamin-binding glycoproteins.⁸⁸ These proteins were known as the R proteins because of their rapid electrophoretic mobility compared with intrinsic factor. Elucidation of the primary protein structure of the R proteins reveals that they belong to the same family of isoproteins as the plasma haptocorrin (HC) binder (previously known as transcobalamins I and III). These HC-like proteins are produced mainly by the salivary glands.

Absorption of Cobalamin: Cubilin

Cobalamins in foods are liberated in the stomach by peptic digestion.⁸⁹ They are then bound not to intrinsic factor but to the HC-like protein because cobalamin binds much more tightly to HC than to intrinsic factor at the acid pH of the stomach.⁹⁰ Upon entering the duodenum, cobalamin is released from the cobalamin–HC protein complex through digestion by pancreatic proteases, which in normal subjects act by selectively degrading HC and the cobalamin–HC complex while sparing intrinsic factor.⁹⁰ Only at this point can cobalamin bind to intrinsic factor to form the intrinsic factor–cobalamin complex.

The intrinsic factor–cobalamin complex, which is very resistant to digestion,⁹¹ traverses the intestine until it reaches the intrinsic factor receptor, cubilin,⁹² a 460-kDa peripheral membrane glycoprotein located in the microvillus pits of the ileal mucosa brush-border. Cubilin forms part of a multifunctional epithelial receptor complex also found in the yolk sac and renal proximal tubule cells.⁹³ In the kidney, it appears to serve a role in the overall body economy through tubular reabsorption of cobalamin,⁹⁴ but the function of the cubilin receptor complex in the kidney and other polarized epithelial surfaces extends beyond cobalamin. The ileal cubilin receptor complex consists of two proteins, cubilin (CUB) and amnionless (AMN), the product of two distinct genes, CUB and AMN. Both proteins, which together have been designated the “CUBAM complex,” colocalize in the endocytic compartment and are required for the process of assimilation of cobalamin,⁹⁵ AMN serving as a chaperon for endosomal targeting. Mutations affecting either of the two proteins disrupt the normal process of the intestinal phase of cobalamin absorption. In addition to the tightly embracing components of the CUBAM complex, a distinct large multifunctional protein, megalin, which belongs to the low-density lipoprotein family,⁹⁶ also participates in the conformational changes that accompany internalization. The concentration of the CUBAM complex rises progressively to a maximum near the terminal ileum.⁹⁷ A specific site on the intrinsic factor molecule avidly attaches to the receptor in a binding reaction that requires a pH of 5.4 or greater and Ca²⁺ (or other divalent cations) but no energy.⁹⁸

The intrinsic factor–cobalamin receptor complex is taken into the ileal mucosal cells over 30 to 60 minutes by endocytosis,⁹⁹ where the vitamin is processed and released into the portal blood over many hours. The receptors recycle to the microvillus surface to shuttle another load of intrinsic factor–cobalamin complex.⁹⁹ That this process has a limited capacity is evident from estimates of the maximum amount of cobalamin that can be absorbed from a single dose via this physiologic pathway.^64,100 Defects in the genes that regulate the complex mechanism of ileal absorption are implicated in autosomal recessive megaloblastic anemia (MGA1), caused by intestinal malabsorption of cobalamin (see “Selective Malabsorption of Cobalamin, Autosomal Recessive Megaloblastic Anemia, Imerslund-Gräsbeck Disease” below).

During its sojourn in the ileal enterocyte, the vitamin first appears in the lysosomes, but by 4 hours most of the vitamin is located in the cytosol.¹⁰¹ During absorption, the entire intrinsic factor–cobalamin complex appears to be taken into the cell, where the cobalamin is released while the intrinsic factor is degraded.¹⁰²

Cobalamin from a small oral dose (10 to 20 mcg) starts to appear in the blood after 3 to 4 hours, and the vitamin reaches a peak level in 6 to 12 hours. In the portal blood, the cobalamin is complexed with a cobalamin-transporting protein known as transcobalamin (TC) previously known as TC II.¹⁰³ There is evidence that cobalamin leaves the enterocyte through a portal that is part of the ABC drug transport system, ABCC1 (also known as the multidrug resistance protein1 [MRP1]) located on the basolateral surface of the intestinal epithelium, as well as other polar cells and nonpolar cells (macrophages).¹⁰⁴ The cobalamin–TC complex is now believed to be formed as it exits the ileal enterocyte, one of a variety of cells that synthesize TC, including the neighboring vascular endothelial cells in the submucosa.¹⁰⁵ Large oral doses (1 mg) of cobalamin are absorbed inefficiently (1 to 2 percent of an oral dose) by simple diffusion that is not mediated by intrinsic factor.¹⁰⁰ In these instances, the vitamin appears in blood within minutes, again as the cobalamin–TC complex.

Like the folates, the cobalamins undergo appreciable enterohepatic recycling.¹⁰⁶ In humans, between 0.5 and 9 mcg/day of cobalamin is secreted into the bile, where it is bound to HC.¹⁰⁷ After entering the intestine, the cobalamin–HC complexes of biliary origin are treated exactly like those delivered from the stomach. The cobalamin is released by digestion of the HC by pancreatic proteases, and then is taken up by intrinsic factor and reabsorbed. From 65 to 75 percent of biliary cobalamin is estimated to be reabsorbed by this mechanism.¹⁰⁸ Because of the size of the cobalamin storage pool and the existence of this enterohepatic circulation, a very long time—as long as 20 years—is required for a clinically significant cobalamin deficiency to develop from a diet providing insufficient cobalamin (e.g., a strictly vegetarian diet).¹⁰⁹ Patients who are unable to absorb the vitamin, however, become clinically deficient in only 3 to 6 years because the absorption of both biliary and dietary cobalamin are interdicted.¹¹⁰

COBALAMIN IN THE CELL: TRANSCOBALAMIN

Uptake of Cobalamin By Cells

TC is the plasma protein that mediates the transport of cobalamin into the tissues.¹¹¹ A β-globulin protein with a calculated molecular weight of 45,538 from the deduced amino acid sequence,^112,113 TC binds cobalamin with exceedingly high affinity (K_a = 10^–11 M).¹¹⁴ Unlike intrinsic factor, whose binding is highly specific for cobalamins, TC shows some promiscuity and also can bind certain corrins that are chemically related to the cobalamins but have no function in mammalian systems and are known as cobalamin “analogues.”¹¹⁵ TC is synthesized by many types of cells, including enterocytes, hepatocytes, endothelial cells, mononuclear phagocytes, fibroblasts, and hematopoietic precursors in the marrow.⁶⁴ Although circulating TC carries only a minor fraction of the cobalamin in the plasma, it is the protein on which cobalamin absorbed through the intestine and is transported into the portal blood as the preformed cobalamin–TC complex. It also is the protein with which cobalamin given parenterally associates almost immediately.¹¹⁶ These cobalamin–TC complexes are transported into the tissues within minutes of appearing in the bloodstream.¹¹⁷ The transport process begins with binding of the cobalamin–TC complex to a specific membrane receptor that is present on a wide variety of cells.¹¹⁸ The protein and gene encoding the TC receptor has been purified from placental membranes and characterized.¹¹⁹ Designated as CD320, the receptor belongs to the low-density lipoprotein receptor family and its internalization involves megalin. The receptor-bound complex is internalized by receptor-mediated endocytosis and delivered to a lysosome, where the TC is digested and the cobalamin is freed.^120,121

Formation of Adenosylcobalamin and Methylcobalamin

To become metabolically active, CnCbl and OHCbl must first be converted to AdoCbl and MeCbl, the coenzymatically active cobalamins. The conversion is accomplished by reduction and alkylation. CnCbl and OHCbl are first reduced to the Co⁺⁺ form [cob(II)alamin] by NADPH- and NADH (nicotinamide adenine dinucleotide phosphate)-dependent reductases that are present in mitochondria and microsomes.¹²² CN– and OH– are displaced from the metal during reduction. Some of the cob(II)alamin is reduced further in the mitochondria to the intensely nucleophilic Co⁺ form [cob(I)alamin]. This is then alkylated by ATP to form AdoCbl in a reaction in which the 5′-deoxyadenosyl moiety of ATP is transferred to the cobalamin and the three phosphates of ATP are released as inorganic triphosphate (Fig. 41–11). The rest of the cobalamin binds to cytosolic N⁵-methyltetrahydrofolate–homocysteine methyltransferase, where it is converted to MeCbl. The several steps involved in the conversion of cobalamin to its coenzymatically active forms are regulated by genes that play a critical role in the processing of the vitamin. There are a number of inherited metabolic errors that correspond to one or more of these specific steps and that result in characteristic syndromes affecting aspects of cobalamin metabolism that are discussed later in this chapter.

PLASMA HAPTOCORRIN (TRANSCOBALAMINS I AND III; “R” PROTEINS)

The HCs (previously known as R proteins) are a group of immunologically related proteins of apparent Mr approximately 60,000, consisting of a single polypeptide species variably substituted with oligosaccharides that terminate with different quantities of sialic acid.¹²³ They are found in milk, plasma, saliva, gastric juice, and numerous other body fluids. They appear to be synthesized by mucosal cells of the organs that secrete them¹²⁴ and by phagocytes.¹²⁵ Although the HCs bind cobalamin, they lack intrinsic factor activity, that is, they are unable to promote the intestinal absorption of the vitamin.

Plasma HC carries most (70 to 90 percent) of the circulating cobalamin. It contains nine potential glycosylation sites¹²⁶ and is encoded by a gene on chromosome 11, the same chromosome that carries the intrinsic factor gene.¹²⁷ In contrast to TC, HC clearance from the plasma is very slow (half-life [T_1/2]: 9 to 10 days).¹²⁸ The asialoglycoprotein receptor carries the cobalamin–HC complexes into the hepatocytes, where they are chiefly eliminated. The complexes are degraded, and their load of cobalamins is excreted in the bile.^106,129 HC binds its ligands more tightly than does either intrinsic factor or TC. Furthermore, HC is less restrictive than either intrinsic factor or TC with respect to ligand specificity; it avidly takes up corrinoids of widely varying structure.¹³⁰ The ligand-binding properties of HC and its mode of clearance by the liver suggest that HC helps clear the system of nonphysiologic cobalamin analogues that may have been acquired or may have arisen through degradation of cobalamin.^131,132 As the liver metabolizes analogue–HC complexes, it secretes the analogues into the bile. Because these analogues are bound poorly by intrinsic factor,¹³⁰ they are poorly reabsorbed from the intestine and are eliminated in the feces. The precise role of HC is unknown, although it may play a role in the body economy of cobalamin by facilitating excretion of cobalamin analogues while conserving cobalamin through enterohepatic recycling. Additionally, it has been proposed that HC may serve an antimicrobial role.⁶⁴

ASSAY OF SERUM COBALAMIN AND THE TRANSCOBALAMINS

As with folate, cobalamin is usually measured with automated competitive displacement assays using intrinsic factor as a cobalamin-binding protein. The misleading results previously provided by competitive ligand displacement assays were explained by the discovery in serum and tissue of a class of cobalamin analogues that are detected by the radioisotope assay when HC-type proteins and not intrinsic factor were used as the binding protein.¹³³ Current assays use intrinsic factor as the binder and give more reliable values for serum cobalamin. The chemical nature and biologic significance of the analogues are unknown,¹³⁴ but recent evidence suggests that they may arise in the gastrointestinal tract.^131,132

TC and HC are present in plasma in trace quantities (approximately 7 and 20 mcg/L, respectively). In fasting plasma, at least 70 percent of the circulating cobalamin is bound to HC.¹³⁵ TC binds only 10 to 25 percent of the total plasma cobalamin,¹³⁶ but provides the majority (approximately 75 percent) of the total unsaturated cobalamin-binding capacity of plasma.¹³⁵ Table 41–3 lists alterations in unsaturated cobalamin-binding capacity and in HC and TC levels in various disease states. In recent years, assays have been developed that measure the fraction of the plasma cobalamin that is bound to TC. This component, known as holotranscobalamin (holoTC), shows improved specificity compared with the standard cobalamin assay for identifying true cobalamin deficiency, although the assays appear to be generally comparable with respect to sensitivity.^{137,138,139,140,141,142,143}

DEFINITION

Megaloblastic anemias are disorders caused by impaired DNA synthesis. The presence of megaloblastic cells is the morphologic hallmark of this group of anemias. Megaloblastic red cell precursors are larger than normal and have more cytoplasm relative to the size of the nucleus. Promegaloblasts show a blue granule-free cytoplasm and a fine “salt and pepper” granular chromatin that contrasts with the ground-glass texture of its normal counterpart. As the cell differentiates, the chromatin condenses more slowly than normal into darker aggregates that coalesce, but do not fuse homogeneously, giving the nucleus a characteristic fenestrated appearance. Continuing maturation of the cytoplasm as it acquires hemoglobin contrasts with the immature-looking nucleus—a feature termed nuclear-cytoplasmic asynchrony.

Megaloblastic granulocyte precursors are also larger than normal and show nuclear-cytoplasmic asynchrony. A characteristic cell is the giant metamyelocyte, which has a large horseshoe-shaped nucleus, sometimes irregularly shaped, containing ragged open chromatin.

Megaloblastic megakaryocytes may be abnormally large and polylobated, with deficient granulation of the cytoplasm. In severe megaloblastosis, the nucleus may show detached lobes. Further details are provided in “Laboratory Features” below and in Figs. 41–12 and 41–13.

ETIOLOGY AND PATHOGENESIS

Table 41–4 lists the causes of megaloblastic anemia. By far the most common causes worldwide are folate deficiency and cobalamin deficiency. There has, however, been a marked reduction in the prevalence of folate deficiency in North America and a growing number of other countries that have implemented folic acid fortification of the food supply.

Megaloblastic cells have much more cytoplasm and RNA than do their normal counterparts, but they have a relatively normal amount of DNA,¹⁴⁴ suggesting that cytoplasmic constituents (RNA and protein) are synthesized faster than is DNA. Evidence that maturation is retarded in megaloblastic precursors supports this conclusion.¹⁴⁵ DNA synthesis is impaired,¹⁴⁶ and migration of the DNA replication fork and the joining of DNA fragments synthesized from the lagging strand (Okazaki fragments) are delayed,¹⁴⁷ and the S-phase is prolonged.¹⁴⁶

Slowing of DNA replication in the megaloblastic anemias of folate and cobalamin deficiency appears to arise from failure of the folate-dependent conversion of dUMP to dTMP. Because of this failure, deoxyuridine triphosphate (dUTP) levels become abundant and because DNA polymerase is promiscuous with respect to its substrate specificity, allows dUTP to become incorporated into the DNA of folate-deficient cells in place of deoxythymidine triphosphate (dTTP).¹⁴⁸ DNA excision–repair mechanisms to repair the DNA by replacing uridine with thymidine fail for the same reason that uridine triphosphate was incorporated into the DNA in the first place. The result is a repetitive iteration of flawed DNA repair that ultimately leads to DNA strand breaks, fragmentation, and apoptotic cell death.¹⁴⁹

Addition of deoxyuridine (dU) to marrow cells in culture normally decreases the incorporation of tritiated thymidine into DNA, because the dU is converted via dUMP→dTMP to unlabeled dTTP, which competes with the tritiated thymidine. In megaloblastic cells, this effect of added dU is greatly diminished. This finding is consistent with impairment in the dUMP→dTMP reaction in the megaloblastic cells and was the basis for the now defunct dU suppression test.¹⁵⁰ The failed excision–repair model following dUTP misincorporation into DNA also explains the chromosome breaks and other abnormalities that occur in megaloblastic cells.¹⁵¹

CLINICAL FEATURES

All megaloblastic anemias share certain general clinical features. Because the anemia develops slowly, with opportunity for cardiopulmonary and intraerythrocytic compensatory changes,¹⁵² it produces few symptoms until the hematocrit is severely depressed. Symptoms, when they appear, are those of anemia: weakness, palpitation, fatigue, light-headedness, and shortness of breath. The blending of severe pallor and slight jaundice caused by a combination of intramedullary and extravascular hemolysis produce a characteristic lemon-yellow skin. Leukocyte and platelet counts may be low, but rarely cause clinical problems. Details of the clinical manifestations are given in the sections on the specific forms of megaloblastic anemia later in this chapter.

LABORATORY FEATURES

Blood Cells

All cell lines are affected. Erythrocytes vary markedly in size and shape, often are large and oval, and in severe cases can show basophilic stippling and nuclear remnants (Cabot rings and Howell-Jolly bodies). Erythroid activity in the marrow is enhanced, although the megaloblastic cells usually die before they are released, accounting for the reduced reticulocyte count. The more severe the anemia, the more pronounced the morphologic changes in the red cells. When the hematocrit is less than 20 percent, erythroblasts with megaloblastic nuclei, including an occasional promegaloblast, may appear in the blood. The anemia is macrocytic (mean corpuscular volume [MCV] is 100 to 150 fL or more), although coexisting iron deficiency, thalassemia trait,¹⁵³ or inflammation can prevent macrocytosis.¹⁵⁴ Slight macrocytosis may be the earliest sign of megaloblastic anemia. Because of the progressive nature of gradual replacement of normocytic red cells with the macrocytic progeny of a megaloblastic marrow, the earliest observable change in red cell indices is an increase in the red cell distribution width (RDW), reflecting an increase in anisocytosis.

Neutrophil nuclei often have more than the usual three to five lobes (see Fig. 41–12).¹⁵⁵ Typically, more than 5 percent of the neutrophils have five lobes. Cells may contain six or more lobes, a morphology rarely seen in normal neutrophils but not pathognomonic of megaloblastic hematopoiesis. In nutritional megaloblastic anemias caused by folate deficiency, hypersegmented neutrophils are an early sign of megaloblastosis⁵ and persist in the blood for many days after treatment.¹⁵⁵ Neutrophil hypersegmentation was not found to be a sensitive test for mild cobalamin deficiency.¹⁵⁶ Cytogenetic studies are nonspecific and show chromosomes that are elongated and broken. Specific therapy corrects these abnormalities, usually within 2 days, although some abnormalities do not disappear for months.^151,157 Platelets are often reduced in number and slightly smaller than normal with a wider variation in size (increased platelet distribution width [PDW]).¹⁵⁸ The morphologic features of megaloblastic anemia may be grossly exaggerated in patients who have been splenectomized or lack a functional spleen as occurs in celiac disease or sickle cell anemia. Numerous circulating megaloblasts and bizarre red cell morphology may be present.¹⁵⁹

Marrow

Aspirated marrow is cellular and shows striking megaloblastic changes, especially in the erythroid series with well-hemoglobinized erythroblasts containing nuclei that possess less-mature, more-open nuclear chromatin than their normal counterparts. There is a preponderance of earlier basophilic erythroblasts over more mature forms which gives the overall impression of a maturation arrest (see Fig. 41–13). Sideroblasts are increased in number and contain increased numbers of iron granules. The ratio of myeloid to erythroid precursors falls to 1:1 or lower, and granulocyte reserves may be decreased. In severe cases, promegaloblasts containing an unusually large number of mitotic figures are plentiful. Macrophage iron content often is increased. Megaloblastic features in the granulocytic series is also usually present with giant forms and large horseshoe-shaped nuclei. Occasionally megakaryocytes with hyperlobated nuclei are present.

Coexisting Microcytic Anemia

Many features of megaloblastic anemia may be masked when megaloblastic anemia is combined with a microcytic anemia.¹⁵⁴ The anemia can be normocytic or even microcytic, whereas the blood film may show both microcytes and macroovalocytes (a “dimorphic anemia”). The marrow may contain “intermediate” megaloblasts¹⁶⁰ that are smaller and look less “megaloblastic” than usual. In this kind of mixed anemia, the microcytic component usually is iron-deficiency anemia,¹⁵⁴ but it may be thalassemia minor¹⁵³ or the anemia of chronic disease. Even megaloblastic anemia masked by a severe microcytic anemia usually shows hypersegmented neutrophils in the blood and giant metamyelocytes and bands in the marrow. Neutrophil myeloperoxidase levels are high.¹⁶¹

Less commonly, the megaloblastic component of a mixed iron-deficiency anemia can be overlooked, and the patient may be treated only with iron. In this situation, the anemia may respond only partly to therapy, and megaloblastic features become more conspicuous as iron stores fill. The masking of macrocytosis in these situations may be responsible for delay or difficulty in diagnosis of pernicious anemia, particularly in certain geographic areas and ethnic groups where there is a high incidence of thalassemia and microcytic hemoglobinopathies.^153,162,163 There are several situations that favor the coexistence of a megaloblastic state with iron deficiency. Both folate and iron deficiency occur in celiac disease,¹⁶⁴ and cobalamin and iron deficiency both complicate gastric reduction surgery for morbid obesity.¹⁶⁵ Furthermore, Helicobacter pylori infection is associated with gastric atrophy that can result first in iron deficiency and later lead to cobalamin malabsorption and perhaps even predispose to pernicious anemia.^166,167

Incomplete Megaloblastic Anemia

If a patient with a full-blown megaloblastic anemia receives cobalamin or folate before marrow aspiration, the anemia persists but the megaloblastic changes may be obscured. Attenuated megaloblastic changes also are seen in patients with early megaloblastic anemia, in patients with coexisting infection,¹⁵⁴ or in patients after transfusion.

Megaloblastic Anemia Misdiagnosed as Acute Leukemia

Occasionally, very severe megaloblastic anemia produces marrow morphology so bizarre as to be mistaken for acute leukemia. In some cases, the erythroid series does not mature, and the megaloblastic pronormoblast dominates the marrow with prominent mitotic figures and dysmorphic forms, raising the possibility of erythroid leukemia.

Megaloblastic Changes in Other Cells

In most forms of megaloblastic anemia, cytologic abnormalities resembling megaloblastosis may appear in other proliferating cells. Epithelial cells from the mouth, stomach, small intestine, and cervix uteri may look megaloblastic, appearing larger than their normal counterparts and containing atypical immature-looking nuclei. Distinguishing these “megaloblastic” changes from the changes of malignancy may be difficult.¹⁶⁸

Chemical Changes in Body Fluids