22: Hematologic Principles

Blood is rightfully considered the vital fluid, because every organ system depends on the normal function of blood. Blood delivers oxygen and other essential substances throughout the body, removes waste products of metabolism, transports hormones from their origin to site of action, signals and defends against threatened infection, promotes healing via the inflammatory response, and maintains the vascular integrity of the circulatory ­system. It also serves as the central compartment in classical pharmacokinetics and thereby comes into direct contact with virtually every systemic xenobiotic that acts on the organism.⁸⁹ The ease and frequency with which blood is assayed, its central role in functions vital to the organism, and the ability to analyze its characteristics, at first by light microscopy and more recently with molecular techniques, have enabled a detailed understanding of blood that has advanced the frontier of molecular medicine.

In addition to transporting xenobiotics throughout the body, blood and the blood-forming organs can at times be directly affected by these same xenobiotics. For example, decreased blood cell production, increased blood cell destruction, alteration of hemoglobin, and impairment of coagulation can all result from exposure to many xenobiotics. The response in many cases depends on the nature and quantity of the xenobiotic as well as the capacity of the system to respond to the insult. In other cases, no clear and predictable dose–response relationship can be determined, especially when the interaction involves the immune system. These latter reactions are often termed idiosyncratic, reflecting an incomplete understanding of their causative mechanism. In general, such reactions can often be reclassified when advancing knowledge identifies the characteristics that render the individual vulnerable.

Hematopoiesis is the development of the cellular elements of blood. The majority of the cells of the blood system may be classified as either lymphoid (B, T, and natural killer lymphocytes) or myeloid (erythrocytes, megakaryocytes, granulocytes, and monocytes). All of these cells originate from a small common pool of totipotent cells called hematopoietic stem cells. Indeed, the study of this process and its regulation has provided fundamental insight into embryogenesis, stem cell pluripotency, and complex cell-to-cell signaling and interaction.

Bone Marrow

Marrow spaces within bone begin to form in humans at about the fifth fetal month and become the sole site of granulocyte and megakaryocyte proliferation. Erythropoiesis moves from the liver to the marrow by the end of the last trimester. At birth, all marrow contributes to blood cell formation and is red, containing very few fat cells. By adulthood, the same volume of hematopoietic marrow is normally restricted to the sternum, ribs, pelvis, upper sacrum, proximal femora and humeri, skull, vertebrae, clavicles and scapulae.­ Extramedullary hematopoiesis in the liver and spleen may reemerge as a compensatory mechanism under severe stimulation.

Progenitor cells must interact with a supportive microenvironment to sustain hematopoiesis. The hematopoietic stroma consists of macrophages, fibroblasts, adipocytes, and endothelial cells. The extracellular matrix is produced by the stromal cells and is composed of various fibrous proteins, glycoproteins, and proteoglycans, such as collagen, fibronectin, laminin, hemonectin, and thrombospondin. Hematopoietic progenitor cells have receptors to particular matrix molecules. The extracellular matrix provides a structural network to which the progenitors are anchored. As the cells approach maturity, they lose their surface receptors, allowing them to leave the hematopoietic space and enter the venous sinuses. Blood cell release depends upon the development of a pressure gradient that drives mature cells through channels in endothelial cell cytoplasm. This pressure is increased by erythropoietin and by granulocyte colony-stimulating factor.

Stem Cells

A stem cell is capable of self-renewal, as well as differentiation into a specific cell type. The pluripotent hematopoietic stem cells can therefore continuously replicate while awaiting the appropriate signal to differentiate into either a myeloid or a lymphoid stem cell (Fig. 22–1). Stem cells represent only 1 in 100,000 of the nucleated cells of the bone marrow, and most of these stem cells are usually quiescent. Nevertheless, these relatively few cells are directly responsible for the estimated 3 billion red cells, 2.5 billion platelets, and 1.5 billion leukocytes per kilogram of body weight produced each day. In response to hemolysis or infection, substantially larger numbers of blood cells can be produced.⁷⁰ Multiple steps are involved in the commitment of less-differentiated cells to more mature cell lines. The final steps in the maturation of erythrocytes alone, for example, involve extensive remodeling, the restructuring of cellular membranes, the accumulation of hemoglobin, and the loss of nuclei and organelles.

Multipotent mesenchymal stem cells capable of differentiating into other tissue lines, including hepatic, renal, muscle and perhaps neuronal, are also found in the bone marrow.⁴⁰ Furthermore, tissue-specific stem cells capable of self-renewal are believed to reside in numerous other organs and to play a fundamental role in repair and regeneration.²¹ A variety of strategies have likely evolved to protect the stem cell from injury due to xenobiotics and radiation, and a better understanding of these effects promises to improve our understanding of toxicity and treatment. The hematopoietic stem cell has provided fundamental insights into regenerative biology, and it remains a focus of intense research given the profound implications for organ homeostasis, tissue repair, and gene therapy.^58,59

Cytokines

Cytokines are soluble mediators secreted by cells for cell-to-cell communication. Initially termed growth factors, it is now recognized that not all cytokines are growth factors. Cytokines promote or inhibit the differentiation, proliferation, and trafficking of blood cells and their precursors. Importantly, they can also inhibit apoptosis, and their absence therefore results in the self-destruction of unwanted cells. They include growth factors or colony-stimulating factors (CSFs), interleukins, monokines, interferons, and chemokines. At baseline, these act in concert to maintain normal blood counts. In response to antigens or other stimuli, cytokines are released to combat perceived infection. Recombinant cytokines are being developed for therapeutic use in immunocompromised patients, transplant recipients, and those with sepsis and cancer. They have also been used in clinical ­toxicology for the treatment of colchicine and podophyllum toxicity.^29,43

The growth factors are glycoproteins necessary for the differentiation and maturation of individual or multiple cell lines. They fall into two families based on their target receptors. The ligands of the cytokine receptor family include growth hormone, interleukin-2 (IL-2), macrophage colony-stimulating factor-1 (CSF-1), granulocyte-macrophage colony-stimulating factor (GM-CSF), γ-interferon, and granulocyte colony-stimulating ­factor (G-CSF), to name a few. The second group, the tyrosine kinase family, includes Kit ligand and insulinlike growth factor-1 receptor (IGF-1R), a member of the insulin family.

Cell Surface Antigens

Using monoclonal antibody technology, cell surface antigens are used to characterize cell types. The cluster designation (CD) nomenclature was introduced by immunologists to ensure a common language when confronted with multiple antibodies to the same leukocyte cell-surface molecule, and hundreds of such unique molecules have been classified.¹⁰⁷ For example, the CD34 antigen is a 115-kilodalton (kDa) highly glycosylated transmembrane protein that is selectively expressed by primitive multipotent hematopoietic stem cells shortly after activation, but is absent from mature T and B lymphocytes.⁶¹ Advances in genomics and proteomics now complement the immunologic designation, but the ability to subtype blood cells phenotypically has transformed the approach to the leukemias, autoimmune disease, transplantation medicine, and thromboembolic disease.

The human leukocyte antigen (HLA) system denotes the major histocompatibility complex in humans, a group of genes involved in antigen presentation to T-lymphocytes, the complement system, cancer surveillance, and autoimmune disease. The rich variation in HLA genes, with more than 100 known alleles at six loci, is believed to be the result of selective pressure to deter individuals with similar HLA from mating. Such sexual selection enhancing diversity would protect a population from coevolving pathogens that attempted to mimic a specific human epitope and avoid immune attack. Indeed, humans appear to be able to distinguish and preferentially choose partners with HLA alleles different from their own using smell, promoting heterozygous diversity.⁶⁵ Exposure to blood transfusions, pregnancy, and organ grafts can generate antibodies against nonself HLAs.

Aplastic Anemia

Aplastic anemia is characterized by pancytopenia on peripheral smear, a hypocellular marrow, and delayed plasma iron clearance. Severe aplastic anemia denotes a granulocyte count of less than 500 cells/mm³, a platelet count of less than 20,000/mm³, and a reticulocyte count of less than 1% after correction for anemia. Following acute insult and depletion of extracirculatory reserves, cell line counts fall at a rate inversely proportional to their life span: granulocytes (half-life 6–12 hours in the circulation) disappear within days, platelets (life span of 7–10 days) decline by half in about one week, while erythrocytes (normal life span 120 days) decline over weeks in the absence of bleeding or hemolysis. Approximately 1000 new cases of aplastic anemia are diagnosed yearly in the United States. The incidence is two to three times higher in Asia, perhaps because of a combination of genetic, environmental, and infectious factors.¹⁰⁵ Aplastic anemia may be inborn (as in Fanconi anemia) or acquired. A variety of xenobiotics such as benzene, pesticides, and chloramphenicol are associated with acquired aplastic anemia (Table 22–1),^20,106 but causality is uncertain given uncertain case ascertainment and other biases. More recent epidemiologic studies demonstrate that specific causes are identified in relatively few cases.⁵¹

Generally speaking, the majority of cases of so-called idiosyncratic aplastic anemia are caused by autoimmune attack on CD34+ hematopoietic stem cells. Following an exposure to an inciting antigen, cytotoxic T-cells secrete interferon-γ and tumor necrosis factor α, which destroy progenitor cells, eventually depleting mature leukocytes, erythrocytes, and platelets in circulation.¹⁰⁶ As with other autoimmune diseases, certain HLA patterns are associated with a genetic predisposition for the condition, namely the HLA DR2. Clozapine-induced agranulocytosis is associated with the HLA B38, DR4, and DQ3 haplotypes.⁷¹ Immunosuppressive therapy or allogenic stem cell transplantation allows recovery of hematopoiesis, and survival is now expected.¹⁰⁵

It is important to distinguish immunologic xenobiotic-induced aplastic anemia from the direct myelotoxic effects of radiation and chemotherapy. Following exposure to ionizing radiation, a pancytopenia ensues due to injury to stem and progenitor cells. While atom bomb survivors rarely developed aplastic anemia,⁴⁸ fractionated whole body radiation of at least 10 Gy is used to intentionally destroy bone marrow stem cells prior to hematopoietic stem cell transplantation. Cancer chemotherapy is often dosed to the end point of reversible hematopoietic toxicity. Vacuolated pronormoblasts can be found in the bone marrow when aplasia is due to a myelotoxic drug, and treatment consists primarily of stopping the offending xenobiotic while supporting cell lines and preventing infection and bleeding. Other nonimmunologic mechanisms of aplasia are identified in specific cases. For example, the severe pancytopenia that occasionally occurs following 5-fluorouracil therapy is caused by a deficiency of dihydropyrimidine dehydrogenase present in 3% to 5% of whites.¹⁰¹

The erythron can be considered to be a single yet dispersed tissue, defined as the entire mass of erythroid cells from the first committed progenitor cell to the mature circulating erythrocyte. This functional definition emphasizes the integrated regulation of the erythron, both in health and disease. The primary function of the erythron is to transport molecular oxygen throughout the organism. To accomplish this, adequate numbers of circulating erythrocytes (nearly half of the blood by volume) must be maintained. These erythrocytes must be able to preserve their structure and flexibility to circuit repeatedly through the microcirculation and to resist oxidant stress accumulated during their life span.⁸¹ The erythrocyte also plays a key role in modulating vascular tone. Interactions between oxyhemoglobin and nitric oxide help match vasomotor tone to local tissue oxygen demands.^{30, 38, 42, 64, 88}

Homeostasis of erythron proliferation is primarily maintained by the equilibrium between stimulation via the hormone erythropoietin and apoptosis controlled by two receptors, Fas and FasL, expressed on the membranes of erythroid precursors. At the other extreme, erythrocytes are culled from the circulation at the end of their life span primarily by the action of the spleen. With age, erythrocytes become less able to negotiate the narrow red pulp passages in the spleen and are phagocytosed by macrophages. By filtering out these senescent cells, the spleen minimizes entrapment in the microvasculature of other organs and prevents spillage of intracellular contents including hemoglobin into the intravascular circulation.

Erythropoietin

Erythropoietin (EPO) is a glycoprotein hormone of molecular weight 34,000 Da that is produced in the epithelial cells lining the peritubular capillaries of the kidney. Anemia and hypoxemia stimulate its synthesis. EPO receptors are found in human erythroid cells, megakaryocytes, and fetal liver. EPO promotes erythroid differentiation, the mobilization of marrow progenitor cells, and the premature release of marrow reticulocytes. The cell most ­sensitive to EPO is a cell between the erythroid colony-forming unit (CFU-E) and the proerythroblast. The absence of EPO results in DNA cleavage and erythroid cell death.

The Mature Erythrocyte

The mature erythrocyte (red blood cell) is a highly specialized cell, designed primarily for oxygen transport. Accordingly, it is densely packed with hemoglobin, which constitutes approximately 90% of the dry weight of the erythrocyte. During maturation, the erythrocyte loses its nucleus, mitochondria, and other organelles, rendering it incapable of synthesizing new protein, replicating, or using the oxygen being transported for oxidative phosphorylation. Its metabolic repertoire is severely limited and largely restricted to a few pathways (described below under Metabolism). In ­general, the enzymatic pathways are those required for optimizing oxygen and carbon dioxide exchange, transiting the microcirculation while maintaining cellular integrity and flexibility, and resisting oxidant stress on the iron and protein of the cell. The characteristic biconcave discocyte shape is dynamically maintained, increasing membrane surface-to-cell volume. This shape decreases intracellular diffusion distances to the cell membrane and allows plastic deformation when squeezing through the microcirculation. The shape is the net sum of elastic and electrostatic forces within the membrane, surface tension, and osmotic and hydrostatic pressures. The cell membrane contains globular proteins floating within the phospholipid bilayer. The major blood group antigens are carried on membrane ceramide glycolipids and proteins, particularly glycophorin A and the Rh proteins. Membrane proteins generally serve to maintain the structure of the cell, to transport ions and other substances across the membrane, or to catalyze a limited number of specific chemical reactions for the cell.

Structural Proteins.

The cell membrane is coupled to, and interacts dynamically with, the cytoskeleton, allowing changes in cell shape such as tank treading or rotation of the membrane relative to the cytoplasm. This cytoskeleton consists of a hexagonal lattice of proteins, especially spectrin, actin, and protein 4.1, which interact with ankyrin and band 3 in the membrane to provide a strong but flexible structure to the membrane. Other essential structural proteins include tropomyosin, tropomodulin, and adducin. Absence or abnormalities of these proteins can result in abnormal erythrocyte shapes such as spherocytes and elliptocytes.

Transport Proteins.

Many specialized transport proteins are embedded in the erythrocyte membrane. These include anion and cation transporters, glucose and urea transporters, and water channels. The erythrocyte membrane is relatively impermeable to ion flux. Band 3 anion-exchange protein plays an important role in the chloride-bicarbonate exchanges that occur as the erythrocyte moves between the lung and tissues. ­Glucose, the sole source of energy of the erythrocyte, crosses the membrane by facilitated diffusion mediated by a transmembrane glucose transporter. Sodium-potassium adenosine triphosphatase (Na⁺-K⁺-ATPase) maintains the primary cation gradient by pumping sodium out of the erythrocyte in exchange for potassium.

Membrane-Associated Enzymes.

At least 50 membrane-bound or membrane-associated enzymes are known to exist in the human erythrocyte. Acetylcholinesterase is an externally oriented enzyme whose role in the function of the erythrocyte remains obscure. Its function is inhibited by certain xenobiotics, most notably the organic phosphorus insecticides, and it can be conveniently assayed as a marker for such exposures (Chap. 113).

Metabolism.

Lacking mitochondria and the ability to generate adenosine triphosphate (ATP) using molecular oxygen, the mature erythrocyte has a severely limited repertoire of intermediary metabolism compared to most mammalian cells. Having lost its nucleus, ribosomes, and translational apparatus, new enzymes cannot be synthesized and existing enzymes decline in function over the lifetime of the cell. Fortunately, the metabolic demands of the erythrocyte are usually modest, but under conditions of stress the capacity can be overwhelmed, especially among senescent cells. The greatest expenditure of energy under physiologic conditions is for the maintenance of transmembrane gradients and for the contraction of cytoskeletal elements. However, oxidant stress can put severe strain on the metabolic reserves of the erythrocyte, and lead ultimately to the premature destruction of the cell, a process termed hemolysis.

Figure 22–2 illustrates the main metabolic pathways and their purpose. Embden-Meyerhof glycolysis is the only source of ATP for the erythrocyte and consumes approximately 90% of the glucose imported by the cell. The reduced nicotinamide adenine dinucleotide (NADH) generated during glycolysis, which would ordinarily be used for oxidative phosphorylation in cells containing mitochondria, is directed toward the reduction of either methemoglobin to hemoglobin by cytochrome b5 reductase, or pyruvate to lactate. Both pyruvate and lactate are exported from the cell. During glycolysis, metabolism can be diverted into the Rapoport-Luebering shunt, generating 2,3-bisphosphoglycerate (2,3-BPG, formerly known as 2,3-diphosphoglycerate or 2,3-DPG) in lieu of ATP. 2,3-BPG binds to deoxyhemoglobin to modulate oxygen affinity and allow unloading of ­oxygen at the capillaries. In response to anemia, altitude, or changes in ­cellular pH, the activity of the shunt increases, thereby favoring synthesis of 2,3-BPG and increasing oxygen delivery considerably. Reduced [levels] of 2,3-BPG in stored blood result in impaired oxygen delivery following massive transfusion.⁹⁸

As an alternative to glycolysis, glucose can be directed toward the hexose monophosphate shunt during times of oxidant stress. This pathway results in the generation of reduced nicotinamide adenine dinucleotide phosphate (NADPH), which the erythrocyte uses to maintain reduced glutathione which, in turn, inactivates oxidants and protects the sulfhydryl groups of hemoglobin and other proteins. The initial, rate-limiting step of this pathway is controlled by glucose-6-phosphate dehydrogenase (G6PD). Accordingly, cells deficient in this enzyme are less able to maintain glutathione in a reduced state, and are vulnerable to irreversible damage under oxidant stress. The consequences of this deficiency are discussed in greater detail below under Glucose-6-Phosphate Dehydrogenase Deficiencies.

The erythrocyte also contains enzymes to synthesize glutathione (γ-glutamyl-cysteine synthetase and glutathione synthetase), to convert CO₂ to bicarbonate ion (carbonic anhydrase I), to remove pyrimidines resulting from the degradation of RNA (pyrimidine 5′-nucleotidase), to protect against free radicals (catalase, superoxide dismutase, glutathione peroxidase), and to conjugate glutathione to electrophiles (ρ-glutathione-S-transferase).

Hemoglobin

Hemoglobin, the major constituent of the cytoplasm of the erythrocyte, is a conjugated protein with a molecular weight of 64,500 Da. To put things in perspective, the typical adult man has approximately 75 mL/kg of blood containing 15 g/100 mL of hemoglobin, or nearly 1 kg of hemoglobin. His 0.3-kg heart must pump this entire mass of hemoglobin every minute at rest, which is a substantial work expenditure. One molecule is composed of four protein (globin) chains, each attached to a prosthetic group called heme. Heme contains an iron molecule complexed at the center of a porphyrin ring. The globin chains are held together by noncovalent electrostatic attraction into a tetrahedral array. Hemoglobin is so efficient at binding and carrying oxygen that it enables blood to transport 100 times as much oxygen as could be carried by plasma alone (Chap. 29). In addition, the capacity of hemoglobin to modulate oxygen binding under different conditions allows adaptation to a wide variety of environments and demands. Three complex and integrated pathways are required for the formation of hemoglobin: globin synthesis, protoporphyrin synthesis, and iron metabolism.

Globin Synthesis.

The protein chains of hemoglobin are produced with information from two different genetic loci. The α-globin gene cluster spans 30 kb on the short arm of chromosome 16, and codes for two identical adult α-chain genes, as well as the ζ-chain, an embryonic globulin. The β cluster is 50 kb on chromosome 11, and codes for the two adult globins β and α, as well as two nearly identical γ chains expressed in the fetus and an embryonic globulin ε. The expression of genes in each family changes during embryonic, fetal, neonatal, and adult development. Until 8 weeks of intrauterine life, ε, ζ, γ, and α chains are produced and assembled in various combinations in yolk sac–derived erythrocytes. With the shift in erythropoiesis from yolk sac to fetal liver and spleen, embryonic hemoglobin disappears, and the α and γ globin chains are paired into fetal hemoglobin (HbF = α₂γ₂). Erythrocytes containing HbF have a higher O₂ affinity than those containing adult hemoglobin, which is important for oxygen transfer across the placenta into the relatively hypoxic uterine environment. Beginning shortly before birth, expression shifts to the α and β globins, which constitute the predominant adult hemoglobin termed hemoglobin A (α₂β₂). Approximately 2.5% of normal adult hemoglobin is in the form of hemoglobin A₂(α₂δ₂). The thalassemias, a group of inherited disorders, result from defective synthesis of one or more of the globin chains. Clinically this results in a hypochromic, microcytic anemia.

Heme Synthesis.

Heme is the iron complex of protoporphyrin IX. Protoporphyrin IX is a tetramer composed of four porphyrin rings joined in a closed, flat-ring structure. The IX designation refers to the order in which it was first synthesized in Hans Fischer’s laboratory. Of the 15 possible isomers, only protoporphyrin IX occurs in living organisms. Technically, only iron complexes with the iron in the Fe²⁺ state can be called heme, but the term is commonly used to refer to the prosthetic group of metalloproteins such as peroxidase (ferric) and cytochrome c (both ferric and ferrous), whether the iron is in the Fe²⁺ or Fe³⁺ state. The terms “hemiglobin” and “ferrihemoglobin” are synonymous with methemoglobin but rarely used. All animal cells can synthesize heme, with the notable exception of mature erythrocytes. Hemoproteins are involved in a multitude of biologic functions, including oxygen binding (hemoglobin, myoglobin), oxygen metabolism (oxidases, peroxidases, catalases, and hydroxylases), and electron transport (cytochromes), as well as metabolism of xenobiotics (cytochrome P450 family). Erythroid cells synthesize 85% of total body heme, with the liver synthesizing most of the balance. Hemoglobin is the most abundant hemoprotein, containing 70% of total body iron.

The first step in the synthesis of heme takes place in the mito­chondrion and is the condensation of glycine and succinyl-coenzyme A (CoA) to form 5-aminolevulinic acid (ALA) (Fig. 22–3). The formation of 5-ALA is catalyzed by aminolevulinic acid synthase (ALAS), the rate-­limiting step of the pathway. The rate of heme synthesis is closely controlled, given that free intracellular heme is toxic and that this first step is essentially irreversible.

Of the two isoforms of ALAS known to exist in mammals, erythroid cells express the ALAS2 isoform, which resides on the X chromosome. Comparatively more is known about ALAS1 (chromosome 3), a housekeeping gene with a short half-life expressed ubiquitously, allowing the synthesis of cellular and mitochondrial hemoproteins. ALAS1 activity is induced by many factors, which can increase its expression by two orders of magnitude. Moreover, it is strongly inhibited by heme in a classical negative feedback fashion. ALAS2 is constitutively expressed at very high concentrations in erythroid precursors, allowing sustained synthesis of heme during erythropoiesis. Pyridoxal phosphate (active vitamin B₆) serves as a cofactor to both isoforms of ALAS. The clinical consequences of pyridoxine deficiency include a hypochromic, microcytic anemia, iron overload, and neurologic impairment.

The next step in the synthesis of hemoglobin is the formation of the monopyrrole porphobilinogen via the condensation of two molecules of ALA. The subsequent steps in heme synthesis involve the condensation of four molecules of porphobilinogen into a flat ring, which is transported back into the mitochondrion by an unknown mechanism. The final step is the insertion of iron into protoporphyrin IX, a reaction that is catalyzed by ferrochelatase (also known as heme synthase) to form heme.

Understanding this carefully regulated synthetic pathway is relevant to understanding the laboratory evidence of lead toxicity (Chap. 96), and predicting the response of porphyric patients to a range of xenobiotics. Most steps in the heme biosynthetic pathway are inhibited by lead. ALA dehydratase is the most sensitive, followed by ferrochelatase, coproporphyrinogen oxidase, and porphobilinogen deaminase. As a consequence, ALA is increased in plasma and especially urine. With increasing exposure, ferrochelatase inhibition coupled with iron deficiency causes zinc protoporphyrin to accumulate in erythrocytes, which can easily be detected by fluorescence. Coproporphyrinogen III also appears in the urine. Historically, these effects have served as the basis for a number of tests of lead exposure.

The porphyrias are a group of disorders resulting from an inherited deficiency of any given enzyme that follows ALAS on the heme biosynthetic pathway. As such, when ALAS activity outpaces the activity of the deficient enzyme, the rate-limiting step shifts downstream, and intermediate metabolites accumulate. These metabolites can cause characteristic neuropsychiatric symptoms and palsies (caused by ALA and porphobilinogen), and cutaneous reactions, including photosensitivity (due to the fluorescence of the porphyrins). For example, porphobilinogen is excreted in large quantities by patients with acute intermittent porphyria (porphobilinogen deaminase deficiency), and the urine darkens with exposure to air and light due to oxidation to porphobilin and to nonenzymatic assembly into porphyrin rings. A variety of xenobiotics can precipitate a crisis in susceptible individuals, by inducing ALAS1 and overloading the deficient enzyme.⁹⁷ The molecular mechanisms that allow xenobiotics to induce the ALAS1 gene closely resemble those accounting for induction of the cytochrome P450 (CYP) genes, which also require heme synthesis. The xenobiotic typically interacts with either the constitutive androstane receptor (CAR) or the pregnane X receptor (PXR), the two main so-called orphan nuclear receptors.¹⁰⁴ These transcriptional factors are DNA-binding proteins that induce the expression of a range of drug metabolizing and transporting genes in response to the presence of a xenobiotic and are termed “xenosensors.” When activated, they associate with the 9-cis retinoic acid xenobiotic receptor (RXR), and then attach to enhancer sequences near the apoCYP or ALAS1 genes to enhance transcription.⁷⁸ The multifunctional inducers capable of activating a wide range of hepatic enzymes are therefore extremely porphyrogenic and include phenobarbital, phenytoin, carbamazepine, and primidone. Furthermore, because CYP 3A4 and 2C9 represent nearly half of the hepatic CYP pool, inducers of these isoforms also stimulate heme synthesis and can induce a porphyric crisis. Examples include the anticonvulsants; nifedipine; sulfamethoxazole; rifampicin; ketoconazole; and the reproductive steroids progesterone, medroxyprogesterone, and testosterone. Glucocorticoids, on the other hand, despite binding to PXR, suppress ALAS1 induction and translation, and they are not porphyrogenic.⁹⁷

Iron Metabolism.

At equilibrium, approximately 1 to 2 mg of iron is absorbed from the diet, and a similar amount is shed from the intestinal epithelium each day. Unless appropriately chelated, free iron not bound by transport or storage proteins can generate harmful oxygen free radicals that damage cellular structures and metabolism (Chaps. 12 and 46). For this reason, serum iron circulates bound to a transfer protein, transferrin, and is stored in the tissues using ferritin (Fig. 22–4). Whereas each molecule of transferrin can bind two iron atoms, ferritin has a large internal cavity, approximately 80 Å in diameter, that can hold up to 4500 iron atoms per molecule. The amount of iron transported through plasma depends on total-body iron stores and the rate of erythropoiesis. Only about one-third of the iron-binding sites of circulating transferrin are normally saturated, as demonstrated by the usual serum iron content of 60 to 170 µg/dL (10–30 µmol/L) as compared to the total iron-binding capacity of 280 to 390 µg/dL (50–70 µmol/L).

Only transferrin can directly supply iron for hemoglobin synthesis. The iron–transferrin complex binds to transferrin receptors on the surface of developing erythroid cells in bone marrow. Iron in the erythroid cell is used for hemoglobin synthesis or is stored in ferritin.

The absorption of nonheme, free ferrous iron from the diet occurs via the divalent metal transporter one of the duodenal enterocyte, which then passes it into the circulation via ferroportin. Iron then circulates in the ferric form bound to transferrin.⁴⁵ Dietary iron complexed with heme can be absorbed via the recently discovered heme carrier protein 1, which transports either iron or zinc protoporphyrin into the enterocyte.⁹¹ The iron may then be freed by microsomal heme oxygenase and follow the transport of atomic iron, or perhaps heme itself can be transferred to the circulation via specific export proteins and circulate bound to its carrier ­protein, hemopexin.⁵

When the erythrocyte is removed from the circulation by splenic macrophages, heme is degraded by heme oxygenase to carbon monoxide and biliverdin, and the iron extracted. Some iron may remain in macrophages in the form of ferritin or hemosiderin. Most is delivered again by ferroportin back to the plasma and bound to transferrin.

Iron homeostasis is largely regulated at the level of absorption, with little physiologic control over its rate of loss. Excess absorption relative to body stores is the hallmark of hereditary hemochromatosis. The iron regulatory hormone hepcidin produced by the liver is now believed to play a ­central role in iron control, including the anemia of chronic disease caused by inflammatory signals.⁴⁵ Hepcidin is a 25-amino acid peptide that senses iron stores and controls the ferroportin-mediated release of iron from enterocytes, macrophages, and hepatocytes. The liver is an important reservoir of iron because it can store considerable amounts of iron taken from portal blood and release it when needed.

Oxygen-Carbon Dioxide Exchange

The evolutionary transition of organisms from anaerobic to aerobic life allowed the liberation of 18 times more energy from glucose. Vertebrates have developed two important systems to overcome the relatively small quantities of oxygen dissolved in aqueous solutions under atmospheric conditions: the circulatory system and hemoglobin. The circulatory system allows delivery of oxygen and removal of carbon dioxide throughout the organism. Hemoglobin plays an essential role in the transport and exchange of both gases. Moreover, the interactions between these gases and hemoglobin are directly linked in a remarkable story of molecular evolution. Understanding this interplay has allowed fundamental insight into protein conformation and the importance of allosteric interactions between molecules.

The binding of one oxygen molecule to each of the four iron atoms in heme results in conformational changes that affect binding of oxygen at the remaining sites. This phenomenon is known as cooperativity, and it is necessary to both the transport of relatively large quantities of oxygen and the unloading of most of this oxygen at tissue sites. Cooperativity results from the intramolecular interactions of the tetrameric hemoglobin, and is expressed in the sigmoidal shape of the oxyhemoglobin dissociation curve (Fig. 29–2). Conversely, the monomeric myoglobin has a hyperbolic oxygen binding curve. The partial pressure of oxygen at which 50% of the oxygen bindings sites of hemoglobin are occupied is about 26 mm Hg, in contrast with about 1 mm Hg for myoglobin. Moreover, hemoglobin is nearly 100% saturated at partial oxygen pressures of about 100 mm Hg in the pulmonary capillaries, transporting 1.34 mL of oxygen per gram of hemoglobin A. About one-third of this oxygen can be unloaded under normal conditions at tissue capillaries with partial oxygen pressures around 35 mm Hg. The proportion unloaded rises during exercise and ­sepsis, as well as with xenobiotics that uncouple oxidative phosphorylation. Elite athletes can extract up to 80% of the available oxygen under conditions of maximal aerobic effort.

The oxygen reserve, however, is only one of the benefits of the large quantity of hemoglobin in circulation. The ability of hemoglobin to buffer the acid equivalent of CO₂ in solution is equally vital to respiratory physiology, because it allows the removal of large quantities of CO₂ from metabolically active tissues with minimal changes in blood pH. Hemoglobin is by far the largest buffer in circulation, accounting for seven times the buffer capacity of the serum proteins combined (28 vs. 4 mEq H⁺/L of whole blood). For every 1 mole of oxygen unloaded in the tissue, about 0.5 mole of H⁺ is loaded onto hemoglobin.

The linked interaction between oxygen and carbon dioxide transport can be first considered from the perspective of oxygen binding to hemoglobin. The affinity of oxygen for hemoglobin is directly affected by pH, which is a function of the CO₂ content of the blood. The oxyhemoglobin dissociation curve shifts to the left in lungs, where the level of carbon dioxide, and thus carbonic acid, are kept relatively low as a result of ventilation, an effect that promotes oxygen binding. The curve shifts to the right in tissues where cellular respiration increases CO₂ concentrations. This phenomenon, known as the Bohr effect, promotes the uptake of oxygen in the lungs and the release of oxygen at tissue sites.

From the perspective of carbon dioxide transport, hemoglobin also plays an essential albeit indirect role. Carbon dioxide dissolves into plasma, and is slowly hydrated to carbonic acid which dissociates to H⁺ and HCO₃^– (pK_a6.35). The speed of the hydration reaction is accelerated from about 40 seconds to 10 msec by the abundant enzyme carbonic anhydrase located within the erythrocyte. Most carbon dioxide collected at the tissues diffuses into erythrocytes, where it becomes H⁺ and HCO₃^–. This HCO₃^– is then rapidly transported back to the serum in exchange for chloride ion via the band 3 anion exchange transporter located in the erythrocyte membrane, thereby shifting serum Cl^– into the erythrocyte (the chloride shift). The hydrogen ion is accepted by hemoglobin, largely at the imidazole ring of histidine residues, which have a pK_a of about 7.0. A small amount of CO₂ reacts directly with the amino terminal of the globin chains to form carbamino residues (HbNHCOO^–). Thus, most of the transported carbon dioxide is transformed by the erythrocyte into bicarbonate ion that is returned to the serum and hydrogen ion that is buffered by hemoglobin. Each liter of venous blood typically carries 0.8 mEq dissolved CO₂+ 16 mEq HCO₃^– in the plasma, 0.4 mEq dissolved CO₂+ 4.6 mEq HCO₃^–, + 1.2 mEq HbNHCOO^– in the erythrocyte (a total of 23 mEq CO₂, equivalent to 510 mL CO₂/L blood). Although two-thirds of the total CO₂ content is ultimately carried in the plasma, nearly all of the bicarbonate is generated within erythrocytes. In the capillaries of the lungs, the reverse reactions occur to eliminate CO₂. Because deoxyhemoglobin is better able to buffer hydrogen ions, the release of oxygen from hemoglobin at the tissues facilitates the uptake of carbon dioxide into venous blood. This effect is known as the Haldane effect. In fact, 1 L of venous blood at 70% oxygen saturation can transport an additional 20 mL of CO₂ compared to arterial blood, which is nearly 100% saturated. Both the Bohr and Haldane effects can have important consequences at the extremes of acid–base perturbations, as can occur in a number of poisonings that interfere with oxygen metabolism.

Finally, in addition to inactivating nitric oxide, hemoglobin can also reversibly bind it as S-nitrosohemoglobin, thereby playing an important role in the regulation of microvascular circulation and oxygen delivery. The ability of hemoglobin to vasodilate the surrounding microvasculature in response to oxygen desaturation using nitric oxide provides new insight into oxygen delivery and may be pivotal in such disorders as septic shock, pulmonary hypertension, and senescence of stored red blood cells.⁹²

Abnormal Hemoglobins

Several alterations of the hemoglobin molecule are encountered in clinical toxicology. A detailed understanding of their molecular basis, clinical manifestations, and effects on gas exchange is essential. Unfortunately, the nomenclature can be ambiguous and overlaps with distinct clinical entities such as oxidant injury and hemolysis. Therefore, although a detailed ­discussion of these abnormal hemoglobins appears elsewhere (Chaps. 125 and 127), an overview of the subject is presented here. It is helpful to recall that the iron atom has six binding positions. Four of these positions are attached in a single plane to the protoporphyrin ring to form heme. The remaining two binding positions lie on opposite sides of this plane. One site is ordinarily bonded to the F8 proximal histidine residue of the globin chain. The remaining site is available for binding molecular oxygen, but it can also bind carbon monoxide, nitric oxide, cyanide, hydroxide ion, or water. The E7 distal histidine residue facilitates the binding of oxygen while stearically hindering carbon monoxide binding.

Methemoglobin.

Methemoglobin (ferrihemoglobin or hemiglobin) is the oxidized form of deoxyhemoglobin in which at least one heme iron is in the oxidized (Fe³⁺) valence state. A number of valency hybrids can occur, depending on the number of ferric versus ferrous heme units within the ­tetramer. Methemoglobin therefore represents oxidation (loss of electrons) of the hemoglobin molecule at the iron atom. It occurs spontaneously as a consequence of interactions between the iron and oxygen. Normally, in deoxygenated hemoglobin, the heme iron is in the ferrous (Fe²⁺) valence state. In this state, there are six electrons in the outer shell, four of which are unpaired. When oxygen is bound, one of these electrons is partially transferred to it and the iron is reversibly oxidized. When O₂ is released, the electron is usually transferred back to heme iron, yielding the normal reduced state. Sometimes, the electron remains with the O₂ yielding a superoxide anion radical O₂^– rather than molecular oxygen. In this case, heme iron is left in the Fe³⁺, or oxidized, state and is unable to release another electron to bind oxygen. This oxidation is primarily reversed via the action of cytochrome b5 reductase, also known as NADH methemoglobin reductase, which uses the electron carrier NADH generated by glycolysis (Metabolism discussion above and Chap. 13). Minor pathways are also involved in methemoglobin reduction, including NADPH methemoglobin reductase, which normally reduces only approximately 5% of the methemoglobin, and vitamin C, a nonenzymatic reducing agent. The activity of NADPH methemoglobin reductase may be significantly accelerated by the presence of the electron donor methylene blue (Antidotes in Depth: A42 and Chap. 127) or riboflavin. Equilibrium is maintained with methemoglobin concentrations of 1% of total hemoglobin. Many xenobiotics increase the rate of methemoglobin formation by as much as 1000-fold. Nitrites, nitrates, chlorates, and quinones are capable of directly oxidizing hemoglobin.¹⁹ Certain individuals may be especially vulnerable resulting from deficient methemoglobin reduction.²⁸ The fetus and neonate are more susceptible to methemoglobinemia than the adult, because HbF is more susceptible to oxidation of the heme iron than adult hemoglobin. The newborn also has a limited capacity to reduce methemoglobin, because adult concentrations of cytochrome b5 reductase are only achieved at about 6 months of age.

Carboxyhemoglobin.

Carbon monoxide (CO) can reversibly bind to heme iron in lieu of molecular oxygen. The affinity of CO for hemoglobin is 200–300 times that of oxygen, despite the stearic hindrance of the E7 distal histidine. The presence of CO thereby precludes the binding of oxygen. In addition, CO binding within any one heme subunit degrades the cooperative binding of oxygen at the remaining heme groups of the same hemoglobin molecule. The oxyhemoglobin dissociation curve is therefore shifted to the left, reflecting the fact that oxygen is more tightly bound by hemoglobin and less able to be unloaded to the tissues. In addition, CO binds to the heme group of myoglobin and the cytochromes, interfering with cellular respiration and exacerbating the hypoxia (Chap. 125).⁴¹

Sulfhemoglobin.

Sulfhemoglobin is a bright green molecule in which the hydrosulfide anion HS^– irreversibly binds to ferrous hemoglobin. The sulfur atom is probably attached to a β carbon in the porphyrin ring and not at the normal oxygen-binding site.⁶⁷ It has a spectrophotometric absorption band at approximately 618 nm,¹⁶ is ineffective in oxygen transport, and clinically produces a condition that resembles cyanosis. The oxygen affinity of sulfhemoglobin is approximately 100 times less than that of oxyhemoglobin, shifting the oxyhemoglobin dissociation curve to the right, in favor of O₂ unbinding. Thus, the symptoms of hypoxia are not as severe with sulfhemoglobinemia as with carboxy- or methemoglobinemia.⁷⁴

Oxidation of the Globin Chain.

Oxidation can also occur at the amino acid side chains of the globin protein. In particular, sulfhydryl groups can oxidize to form disulfide links between cysteine residues, which leads to the unfolding of the protein chain, exposure of other side chains, and further oxidation. When these disulfide links join adjacent hemoglobin molecules, they cause the precipitation of the concentrated hemoglobin molecules out of solution. Covalent links can also form between hemoglobin and other cytoskeletal and membrane proteins.²⁵ Eventually, aggregates of denatured and insoluble protein are visible on light microscopy as Heinz bodies. The distortion of the cellular architecture and the loss of fluidity in particular signal reticuloendothelial macrophages to excise sections of erythrocyte membrane (“bite cells”) or to remove the entire erythrocyte from the circulation (see below). To guard against these oxidation reactions, the erythrocyte maintains a pool of reduced glutathione via the actions of the NADPH generated in the hexose monophosphate shunt (assuming adequate G6PD activity to initiate this pathway). This glutathione transfers electrons to break open disulfide links and to preserve sulfhydryl groups in their reduced state.

Hemolysis

Hemolysis is merely the acceleration of the normal process by which senescent or compromised erythrocytes are removed from the circulation.⁹³ The normal life span of a circulating erythrocyte is approximately 120 days, and any reduction in this life span represents some degree of hemolysis. If sufficiently rapid, hemolysis can overwhelm the regenerative capacity of the erythron, resulting in anemia. Intravascular hemolysis occurs when the rate of hemolysis exceeds the capacity of the reticuloendothelial macrophages to remove damaged erythrocytes, and free hemoglobin and other intracellular contents of the erythrocyte appear in the circulation.

Reticulocytosis, polychromasia, unconjugated hyperbilirubinemia, increased serum lactate dehydrogenase, and decreased serum haptoglobin are characteristic of hemolysis. A normal or elevated RBC distribution width and thrombocytosis are usually present. The presence of spherocytes on peripheral blood smear suggests an autoimmune or hereditary process and can be pursued with a Coombs’ test. Schistocytes suggest thrombotic thrombocytopenic purpura (TTP) or hemolytic uremic syndrome, disseminated intravascular coagulation, or valvular hemolysis. TTP and hemolytic uremic syndrome are characterized by a microangiopathic anemia, thrombocytopenia, and normal coagulation parameters (unlike disseminated intravascular coagulation). TTP is discussed under platelet disorders below. Hemoglobinemia, hemoglobinuria, and hemosiderinuria can occur with intravascular hemolysis. Specialized tests to measure hemolysis detect shortened erythrocyte survival, increased endogenous carbon monoxide generation from heme oxygenase, and increased fecal urobilinogen.

Table 22–2 presents a brief classification of acquired causes of hemolysis relevant to toxicology. Oxidant injury following xenobiotic exposure is one of the triggers of hemolysis, as it may cause irreversible changes in the erythrocyte. Erythrocytes deficient in G6PD by virtue of cell age or enzyme mutations are particularly vulnerable to hemolysis following oxidant stress due to limited capacity to generate NADPH and reduced glutathione.⁹³ The immune-mediated hemolytic anemias occur when ingested xenobiotics trigger an antigen antibody reaction. In general, these molecules are too small to be sensitizing agents, but antigenicity can be acquired after binding to carrier proteins in blood. The particulars of the xenobiotic-carrier immune activation sequence form the basis for the classification of this group of hemolytic anemias.⁷

Nonimmune-Mediated Causes of Xenobiotic-Induced Hemolysis

A number of xenobiotics or their reactive metabolites can cause hemolysis via oxidant injury. A Heinz-body hemolytic anemia can result, which typically resolves within a few weeks of drug discontinuation. Some xenobiotics cause hemolysis in the absence of overt oxidant injury (Table 22–2). Copper sulfate hemolysis is described in Chap. 95, while the delayed hemolysis following exposure to arsine or stibine is described here.

Arsine (AsH₃) is a colorless, odorless, nonirritating gas that is 2.5 times denser than air (Chap. 89). Clinical signs and symptoms appear after a latent period of up to 24 hours after exposure to concentrations above 30 ppm and may include headache, malaise, dyspnea, abdominal pain with nausea and vomiting, hepatomegaly, hemolysis with hemoglobinuric acute kidney injury, and death.^{23, 34, 53, 57, 77} The mechanism of hemolysis is believed to involve the fixation of arsine by sulfhydryl groups of hemoglobin and other essential proteins.^39,103 Interestingly, hemolysis is prevented in vitro by conversion to carboxy- or methemoglobin.⁴⁴ Impairment of membrane proteins, including Na⁺-K⁺-ATPase, is another potential mechanism for arsine-induced hemolysis.⁸⁰ Chronic exposure to low levels of arsine can produce clinically significant disease.²³ Stibine (SbH₃), an antimony derivative, likely causes hemolysis via similar mechanisms (Chap. 88).

Glucose-6-Phosphate Dehydrogenase Deficiencies.

G6PD is the enzyme that catalyzes the first step of the hexose monophosphate shunt: the conversion of glucose-6-phosphate to phosphogluconolactone (Fig. 13–7). In the process, NADP⁺ is reduced to NADPH, which the erythrocyte uses to maintain a supply of reduced glutathione and to defend against oxidation. It follows that erythrocytes deficient in G6PD activity are less able to resist oxidant attack and, in particular, to maintain sulfhydryl groups of hemoglobin in their reduced state, resulting in hemolysis. It is important to recognize that the term G6PD deficiency encompasses a wide range of differences in enzyme activities among individuals. These differences may result from decreased enzyme synthesis, altered catalytic activity, or reduced stability of the enzyme. Approximately 7.5% of the world population is affected to some degree, with more than 400 variants having been identified. Most cases involve relatively mild deficiency and minimal morbidity. Ethnic populations from tropical and subtropical countries (the so-called malaria belt) have a much higher prevalence of G6PD deficiency, possibly because that phenotype protects against malaria.⁶⁹

The gene that encodes for G6PD resides near the end of the long arm of the X-chromosome. Most mutations consist of a single amino acid substitution, as complete absence of this enzyme is lethal. Although men hemizygous for a deficient gene are more severely affected, women randomly inactivate one X-chromosome during cellular differentiation according to the Lyon hypothesis. Thus, women carriers heterozygous for a deficient G6PD gene have a mosaic of erythrocytes, some proportion of which expresses the deficient gene during maturation. Accordingly, approximately 10% of carrier women may be nearly as severely affected as males hemizygous for the same deficient gene. Because of the high gene frequency in certain ethnic groups, another approximately 10% of women may be homozygous for the deficient gene.

Normal G6PD has a half-life of about 60 days. Because the erythrocyte cannot synthesize new protein, the activity of G6PD normally declines by approximately 75% over its 120-day life span. Consequently, even in unaffected individuals, susceptibility to oxidant stress varies based on the age mix of circulating erythrocytes. In all cases, older erythrocytes are less likely to recover following exposure to an oxidant and will hemolyze first. Moreover, after an episode of hemolysis following acute exposure to an oxidant stress, the higher G6PD activity of surviving erythrocytes will confer some resistance against further hemolysis in most individuals with relatively mild deficiency, even if the offending xenobiotic is continued. For this reason, phenotypic testing for G6PD deficiency is best done 2 to 3 months after a hemolytic crisis, which is when the reticulocyte count has usually normalized.

The World Health Organization classification of G6PD is based on the degree of enzyme deficiency and severity of hemolysis.¹⁸ Both class I and class II patients are severely deficient, with less than 10% of normal G6PD activity. Class I individuals are prone to chronic hemolytic anemia, whereas class II patients experience intermittent hemolytic crises. Class III patients have only moderate (10%–60%) enzyme deficiency, and experience self-limiting hemolysis in response to certain xenobiotics and infections. Approximately 11% of African Americans have a class III deficiency, traditionally termed type A^-, and experience a decline of no more than 30% of the red blood cell mass during any single hemolytic episode. Another 20% of African Americans have type A⁺ G6PD enzyme, which is functionally normal, and therefore of no consequence despite a one-base substitution compared to wild-type B. The Mediterranean type found in Sardinia, Corsica, Greece, the Middle East, and India is a class II deficiency, and hemolysis can occur spontaneously or in response to ingestion of oxidants, such as the β-glycosides found in fava (Vicia fava) beans.

The most common clinical presentation of previously unrecognized G6PD deficiency is the acute hemolytic crisis. Typically, hemolysis begins 1 to 4 days following the exposure to an offending xenobiotic or infection (Table 22–3). Jaundice, pallor, and dark urine may occur with abdominal and back pain. A decrease in the concentration of hemoglobin occurs. The peripheral smear demonstrates cell fragments and cells that have had Heinz bodies excised. Bone marrow stimulation results in a reticulocytosis by day 5 and an increased erythrocyte mass. In general, a normal bone marrow can compensate for ongoing hemolysis and can return the hemoglobin concentration to normal. Most crises are self-limiting because of the higher G6PD activity of younger erythrocytes. Historically, the anemia observed when primaquine was administered to type A– military recruits for malaria prophylaxis resolved within 3 to 6 weeks in most cases.¹⁷ Some xenobiotics, including APAP, vitamin C, and sulfisoxazole, are safe at therapeutic doses but can cause hemolysis in G6PD-deficient patients following overdose.

Other presentations of more severe variants of G6PD include neonatal jaundice and kernicterus, chronic hemolysis with splenomegaly and black pigment gallstones, megaloblastic crisis caused by folate deficiency, and aplastic crisis after parvovirus B19 infection.

Megaloblastic Anemia

Vitamin B₁₂ and folate are essential for one-carbon metabolism in ­mammals. One-carbon fragments are necessary for the biosynthesis of thymidine, purines, serotonin, and methionine; the methylation of DNA, histones, and other proteins; and the complete catabolism of branched chain fatty acids and histidine. Unable to synthesize vitamin B₁₂ or folate, mammals are dependent on dietary sources and microorganisms for these cofactors. The hematologic manifestation of vitamin B₁₂ or folate deficiency is a characteristic panmyelosis termed megaloblastic anemia. The hallmark nuclear-cytoplasmic asynchrony is due to disrupted DNA synthesis, halted interphase and ineffective erythropoiesis.⁷⁹ The hyperplastic bone marrow contains precursor cells with abnormal nuclei filled with incompletely condensed chromatin. Among circulating blood cells, macrocytic anemia (macro-ovalocytes) without reticulocytosis is followed by the appearance of granulocytes with an abnormally large, distorted nucleus (hypersegmented neutrophils with six or more lobes). Lymphocytes and platelets may appear normal but are also functionally impaired.

In addition to dietary deficiencies, which have become less common, macrocytic anemia with or without megaloblastosis in adults is usually caused by chronic ethanol abuse, chemotherapeutics, or antiretrovirals, especially when mean corpuscular volumes are only moderately elevated (100–120 fL).⁸⁷ The folate antagonists aminopterin, methotrexate, hydantoins, pyrimethamine, proguanil, sulfasalazine, trimethoprim sulfamethoxazole, and valproate can interfere with DNA synthesis. Ethanol affects folate metabolism and transport. Vitamin B₁₂ deficiency can be induced by chronic exposure to nitrous oxide, biguanides, colchicine, neomycin, and the proton pump inhibitors. Purine analogs (eg, azathioprine, 6-mercaptopurine, 6-thioguanine, acyclovir) and pyrimidine analogs (eg, 5-fluorouracil, floxuridine, 5-azacitidine, and zidovudine) also disrupt nucleic acid synthesis. Hydroxyurea and cytarabine, which inhibit ribonucleotide reductase, also delay nuclear maturation and function and frequently cause megaloblastosis.

Pure Red Cell Aplasia

Pure red cell aplasia is an uncommon condition in which erythrocyte precursors are absent from an otherwise normal bone marrow. It results in a normocytic anemia with inappropriately low reticulocyte count. The other blood cell lines are unaffected, unlike aplastic anemia. Pharmaceuticals cause fewer than 5% of cases of this uncommon condition, having been implicated in fewer than 100 human reports.⁹⁶ Only phenytoin, azathioprine, and isoniazid meet criteria for definite causality; chlorpropamide and valproic acid can be considered only as possible causes.⁹⁶ Most other xenobiotics are cited only in single case reports, and drug rechallenge was not used, making the association uncertain.

In part because of its rarity, a cluster of 13 cases in France of pure red cell aplasia in hemodialysis patients receiving subcutaneous recombinant EPO ultimately led to an international effort by researchers, regulatory authorities, and industry to identify the etiology.^14,26 To reduce theoretical concerns regarding transmission of variant Creutzfeldt-Jakob disease, human serum albumin was replaced with polysorbate 80 as the stabilizer in a formulation used in Europe and Canada. It is suspected that this change allowed rubber to leach from the uncoated stopper of prefilled syringes, triggering an immune response against both recombinant and endogenous EPO in some patients.⁶³ This episode not only serves as a recent example of successful pharmacovigilance for rare adverse drug effects but has also influenced safety assessments for an emerging class of biological therapies that include simple peptides, monoclonal antibodies, and recombinant DNA proteins.^13,63

Erythrocytosis

Erythrocytosis denotes an increase in the red cell mass, either in absolute terms or relative to a reduced plasma volume. An increasingly recognized cause of drug-induced absolute erythrocytosis is the abuse of recombinant human EPO by athletes to enhance aerobic capacity (Chap. 40). Autologous blood transfusions (doping) are also used in this population, and both can cause dangerous increases in blood viscosity. Cobalt was once considered for the treatment of chronic anemia³² due to its ability to cause a secondary erythrocytosis (Chap. 94). The mechanism may involve impaired degradation of the transcription factor hypoxia-inducible factor 1α, thereby prolonging EPO transcription. This effect is more pronounced in high altitude dwellers, in whom elevated serum EPO concentrations persist despite hematocrits in excess of 75% and chronic mountain sickness is associated with increased serum cobalt concentrations.⁵²

The leukon represents all leukocytes (white blood cells), including precursor cells, cells in the circulation, and the large number of extravascular cells. It includes the granulocytes (neutrophils, eosinophils, and basophils), lymphocytes, and monocytes. Neutrophils (polymorphonuclear leukocytes) are highly specialized mediators of the inflammatory response, and are a primary focus of concern regarding hematologic toxicity of xenobiotics. B and T lymphocytes are involved in antibody production and cell-mediated immunity. Monocytes migrate out of the vascular compartment to become tissue macrophages and to regulate immune system function.

Immunity is generally divided into the innate and the adaptive responses. Innate immunity is an immediate but less specific defense that is highly conserved in evolutionary terms. It is centered on the neutrophil response, and it involves monocytes and macrophages as well as complement, cytokines, and acute phase proteins. The innate system responds primarily to extracellular pathogens, especially bacteria, by recognizing structures commonly found on pathogens, namely lipopolysaccharide (Gram-negative cell walls), lipoteichoic acid (Gram positive), and mannans (yeast). Adaptive immunity is demonstrated only in higher animals and is an antigen-specific response via T- and B-lymphocytes after antigen presentation and recognition. Although this reaction is more specific, it requires ­several days to develop, unless that antigen has previously triggered a response (so-called immune memory). This response can also at times be directed against self-antigens, resulting in autoimmune disorders.

The recruitment and activation of neutrophils provide the primary defense against the invasion of bacterial and fungal pathogens. They emerge from the bone marrow with the biochemical and metabolic machinery needed for the efficient killing of microorganisms. Activated macrophages release G-CSF and GM-CSF, which stimulate myeloid differentiation and can be recognized on blood testing as the classic neutrophil leukocytosis. Neutrophils are activated when circulating cells detect chemokines released from sites of inflammation. On activation, they undergo conformational and biochemical changes that transform them into powerful host defenders. These changes allow rolling along the endothelial lining of postcapillary venules, migration toward the site of inflammation, adherence to the endothelium, migration through the endothelium to tissue sites, ingestion, killing, and digestion of invading organisms.

Neutrophils migrate to sites of infection along gradients of chemoattractant mediators. An acute inflammatory stimulus leads to the accumulation of neutrophils along the endothelium of postcapillary venules. The major molecules involved in this process are adhesion molecules, chemoattractants, and chemokines. Loose adhesions between neutrophils and endothelium are made and broken, resulting in the slow movement of leukocytes along endothelium and a more intense exposure of neutrophils to activating factors. Chemotaxis requires responses involving actin polymerization–depolymerization adhesion events mediated by integrins and involving microfilament–membrane interactions. All leukocytes including lymphocytes localize infection using these same mediators. Colchicine depolymerizes microfilaments, causing the dissolution of the fibrillar microtubules in granulocytes and other motile cells, impairing this response.

Opsonized particles, immune complexes, and chemotactic factors activate neutrophils in tissues by binding to cell surface receptors. The neutrophil makes tight contact with its target, and the plasma membrane surrounds the organism completely enclosing it. Two mechanisms are then responsible for the destruction of the organism: the oxygen-dependent respiratory burst, and the oxygen-independent response involving cationic enzymes found in cytoplasmic granules. The respiratory burst is caused by an NADPH oxidase complex that assembles at the phagosomal membrane. Electrons are transferred from cytoplasmic NADPH to oxygen on the phagosomal side of the membrane, generating superoxide, hydrogen peroxide, hydroxyl radical, singlet oxygen, hypochlorous acid, chloramines, nitric oxide, and peroxynitrite. Cytoplasmic granules within the neutrophil fuse with the phagosome and empty their contents into it. The components of these granules include myeloperoxidase (MPO), elastase, lipases, metalloproteinases, and a pool of CD11b/CD18 proteins required for adhesion and migration. Finally, the phagocytized organism is digested and eliminated by the neutrophil. Overstimulation of this complex and highly regulated but somewhat nonspecific system can at times become deleterious, as is postulated to occur with reperfusion injury or carbon monoxide poisoning, to cite two examples.⁹⁵ Vasculitis and the systemic inflammatory response syndrome are further examples of excessive activation of the innate response.

Neutropenia and Agranulocytosis

Neutropenia is a reduction in circulating neutrophils at least 2 standard deviations below the norm, but the threshold of 1500/mm³ (1.5 × 10⁹/L) is often used instead.⁴⁷ Severe neutropenia is termed agranulocytosis and is generally defined to be an absolute neutrophil count of less than 500/mm³ (0.5 × 10⁹/L). Neutropenia can result from decreased production, increased destruction, or retention of neutrophils in the various storage pools. Their high rate of turnover renders neutrophils vulnerable to any xenobiotic that inhibits cellular reproduction. As such, the various antineoplastics including antimetabolites, alkylating agents, and antimitotics will predictably cause neutropenia. This predictable, dose-dependent reaction represents an important dose-limiting adverse effect of therapy. On the other hand, a number of xenobiotics are implicated in idiosyncratic neutropenia.^3,4 The parent drug or a metabolite usually acts as a hapten to trigger antineutrophil antibodies. Table 22–4 provides an abbreviated list.^76,94 Cocaine users are at very high risk for agranulocytosis because of the common contamination of this drug with levamisole (Chap. 78).²²

Eosinophilia

Eosinophils are primarily responsible for protecting against parasitic infection. Allergic reactions and malignancies such as lymphoma are also common causes of eosinophilia, especially where nematode infection is rare.⁸⁵ Eosinophils bind to antigen-specific IgE and discharge their large granules, which contain major basic protein, peroxidase, and eosinophil-derived neurotoxin, onto the surface of the antibody-coated organism. Two unusual toxicologic outbreaks were characterized by eosinophilia, acute cough, fever, and pulmonary infiltrates, followed by severe myalgia, neuropathy, and eosinophilia. The first outbreak, called the toxic oil syndrome, took place in central Spain in 1981, when industrial-use rapeseed oil denatured with 2% aniline was fraudulently sold as olive oil by door-to-door salesmen.³⁶ The precise causative agent remains uncertain but may include fatty acid esters of 3-(N-phenylamino)-1,2-propanediol.³⁶ The second outbreak, called the eosinophilia-myalgia syndrome, occurred during 1988 and 1989 in users of l-tryptophan supplements traced back to a single wholesaler in Japan.⁶ The causative contaminant has not been identified but is believed to have been present in only trace quantities in the l-tryptophan purified from microbial culture. Both syndromes appear to be mediated by immunologic mechanisms.

Leukemia

The leukemias represent the malignant, unregulated proliferation of hematopoietic cells. Although monoclonal in origin, they affect all cell lines derived from the progenitor cell. Acute myeloid leukemia (AML) and the myelodysplastic syndromes are the most common leukemias associated with xenobiotics. The long-recognized association between AML and occupational benzene exposure, radiation, or treatment with alkylating antineoplastics has helped to advance understanding of the molecular mechanisms underlying leukemogenesis.¹⁵ The necessary events are believed to involve several sequential genetic and epigenetic alterations, as evidenced by a distinct pattern of chromosomal deletions preceding the development of AML.^49,50 Other recognized xenobiotics that can cause leukemia include topoisomerase II inhibitors, 1,3-butanediol, styrol, ethylene oxide, and vinyl chloride.^54,75 In many cases, the latency period between exposure and illness is prolonged. For example, leukemia linked to benzene is preceded by several months of anemia, neutropenia, and thrombocytopenia.

In the absence of pathology, blood remains in a fluid form with cells in suspension. Injury triggers coagulation and thrombosis. The resulting clot formation, retraction, and dissolution involve an interaction between the vessel endothelium, soluble constituents of the coagulation system, and proteins located on and within platelets. Platelets respond to signals within their immediate environment and from injured components of the distant microcirculation. A dynamic balance must be maintained between coagulation and fibrinolysis to maintain the integrity of the circulatory system (Fig. 22–5).

Coagulation

Two basic pathways termed intrinsic and extrinsic are involved in the initiation of coagulation. The more important extrinsic, or tissue factor, pathway is activated when blood is exposed to tissue factor (also known as thromboplastin) which is normally expressed on subendothelial cells and fibroblasts not in direct contact with blood. Tissue factor activates factor VII and together form the extrinsic tenase complex activating both factors X and IX. The intrinsic, or contact activation, pathway also activates factor IX, which forms the intrinsic tenase complex with factor VIIIa. Both extrinsic and intrinsic tenase (ie, “ten” -ase) with calcium activate factor X, which binds to factor Va on the surface of activated platelets, forming the prothrombinase complex. The prothrombinase complex activates prothrombin, which results in the generation of thrombin activity. Thrombin activates platelets, promotes its own generation by activation of factors V, VIII, and XI, and converts fibrinogen to fibrin (Chap. 60).⁴⁶

Laboratory Tests of Coagulation

An understanding of several widely available coagulation tests is important to the correct interpretation of the effects of xenobiotics on hemostasis, whether for therapeutic drug monitoring or following overdose. Four studies—prothrombin time (PT, or international normalized ratio {INR}), partial thromboplastin time (PTT), thrombin time, and fibrinogen ­concentration—are particularly important in the classification of coagulation disorders.

PT assesses the extrinsic coagulation pathway and is calculated by adding standardized thromboplastin reagent (phospholipid and tissue factor) to a sample of the patient’s citrated plasma (the citrate removes calcium to prevent clotting). Calcium is then introduced, and the time to clotting measured. With the exception of factor X, the PT is unaffected by the presence or absence of factors VIII to XIII, platelets, prekallikrein, and high-molecular-weight kininogen (HMWK). An individual’s PT was formerly expressed as a ratio (PT observed to PT control). Because this ratio is directly affected by both laboratory methodology and the source of the thromboplastin reagent used, the generated results suffered from significant variability between laboratories and countries. Thus, the INR was developed in an attempt to limit this variability. The INR is calculated by raising the PT ratio to an exponent known as the International Sensitivity Index (ISI):(PT ratio)^ISI. The ISI is calibrated against an international standard to measure responsiveness of the particular thromboplastin to the effects of warfarin.

Although the use of the INR does not completely eliminate variability,^94,95 it improves warfarin dosing and comparison with international norms. It should be noted that the ISI, and thus the INR, is designed for greater consistency in measuring coagulation changes due to warfarin, and it may in fact be more variable than the PT ratio for patients with other coagulation defects. Specifically, replacing the PT with the INR for prognosis in the setting of fulminant hepatic failure, as may occur following APAP overdose, is controversial.^24,84 In these patients, the INR may be more variable between laboratories, and the calculation of a substantially different ISI may be warranted.¹¹

The PTT tests both the intrinsic and common coagulation pathway. It is measured by adding calcium and an activator (kaolin, silica, or celite) to citrated plasma and the time to clotting observed. Some tests also use phospholipids in the reagent to activate the remaining coagulation factors, thereby giving rise to the term activated partial thromboplastin time (aPTT). Because the PTT and aPTT are essentially interchangeable, the term PTT is used hereafter to represent the concept. The PTT is not affected by alterations in factors VII, XIII, or platelets.

The thrombin time (or thrombin clotting time), determined by adding exogenous thrombin to citrated plasma, evaluates the ability to convert fibrinogen to fibrin, is thus unaffected by abnormalities of factors II, V, VII to XIII, platelets, prekallikrein, or HMWK. Finally, either a fibrinogen concentration or a determination of fibrin degradation products will help distinguish between problems with clot formation and consumptive coagulopathy (disseminated intravascular coagulation). An evaluation of the combination of normal and abnormal results of these tests usually determines a patient’s clotting abnormality (Table 22–5).

Inhibitors can be diagnosed by “mixing studies,” because only a small percentage of the coagulation factors present in normal plasma are necessary to have normal clotting studies. Factor activity in blood can be as little as 25% of normal without affecting the PT or PTT. Thus, the presence of an abnormal PT or PTT that will not correct despite the addition of an equal volume of normal plasma demonstrates the presence of an inhibitor of coagulation as opposed to a factor deficiency. Heparin-induced anticoagulation results in an elevated PTT that does not correct when mixing studies are performed. More sophisticated studies can be used to identify specific coagulation factor deficiencies.

Xenobiotic-Induced Defects in Coagulation

A rapidly increasing number of xenobiotics are in therapeutic use to antagonize coagulation (Chap. 60). The recognition of a hemorrhagic disease in cattle in the 1920s and the isolation of the causative agent dicoumarol from spoiled sweet clover in the 1940s resulted in the development of the warfarin-type anticoagulants. This group of anticoagulants indirectly inhibits hepatic synthesis of coagulation factors II, VII, IX, X, and proteins C, S, and Z by inhibiting the reductase that is responsible for the regeneration of vitamin K quinone from vitamin K epoxide. Because warfarin is widely prescribed and has a narrow therapeutic index, efforts to improve our understanding of the pharmacology of warfarin are advancing the fields of pharmacogenomics and drug interactions.¹⁰²

Heparin is a highly sulfated glycosaminoglycan that is normally present in tissues. Commercial unfractionated heparin is either bovine or porcine in origin and consists of a mixture of polysaccharides with molecular weights ranging from 4000 to 30,000 Da. The anticoagulant activity of heparin is through its catalytic activation of antithrombin III, a suicide inhibitor of the serine proteases thrombin and factors IXa, Xa, XIa, and XIIa.⁶⁸

The low-molecular-weight heparins (LMWHs) have a mean molecular weight of 4000 to 6000 Da.⁴⁶ Their subcutaneous bioavailability, more predictable pharmacokinetics, lower protein binding, and longer half-life make them more convenient than unfractionated heparin.⁶⁸

Heparinoids are glycosaminoglycans not derived from heparin with anticoagulant effects. This class includes dermatan sulfate and danaparoid sodium. Fondaparinux is a synthetic pentasaccharide that also inhibits factor Xa by binding to antithrombin III and therefore acts upstream of thrombin. Recently, other oral factor Xa inhibitors have been approved for human use, including rivaroxaban and apixaban. Several direct thrombin inhibitors are also in clinical use, including parenterally administered lepirudin, argat­roban, bivalirudin, desirudin, and more recently oral dabigatran. These agents can prolong the INR, aPTT, thrombin clotting time, and whole-blood activated clotting time.

Fibrinolysis

The coagulation system is opposed by three major inhibitory systems. Components of the fibrinolytic system circulate as zymogens, activators, inhibitors, and cofactors.³³ Plasminogen can be activated to plasmin by an intrinsic pathway involving factor XII, prekallikrein, and HMWK. This produces the degradation products and fibrin monomers that are found in disseminated intravascular coagulation. The extrinsic pathway involves the release of tissue plasminogen activator (t-PA) from tissues and urokinase plasminogen activator (u-PA) from secretions.³³ Once activated, plasmin can degrade fibrinogen, fibrin, and coagulation factors V and VIII. The degradation of cross-linked fibrin strands results in the formation of D-dimers.

Several inhibitors oppose the fibrinolytic system, including α₂-antiplasmin, α₂-macroglobulin, both of which oppose plasmin activity, and plasminogen activator inhibitor (PAI) types 1 and 2, which oppose t-PA. PAI-1 and PAI-2 are opposed by activated protein C and protein S. Activated protein C is activated by thrombin. Another vitamin K–dependent glycoprotein, protein Z, accelerates the degradation of factor Xa. Congenital deficiencies of proteins C, S, and Z may result in pathologic venous thrombosis. Decreased fibrinolytic activity may result from decreased synthesis, release of t-PA, or from an elevation of the PAI-1 concentration. Both ­conditions have been observed postoperatively, with the use of oral contraceptives, in the third trimester of pregnancy, and in obesity. The activity of α₂-antiplasmin and α₂-macroglobulin are increased in pulmonary fibrosis, malignancy, infection, and myocardial infarction, and in thromboembolic disease.³³

Xenobiotic-Induced Defects in Fibrinolysis

Table 22–6 lists xenobiotics associated with an acquired defect of fibrinolysis. The chemotherapeutics may result in a reduction in serine ­protease inhibitors such as antithrombin. l-Asparaginase is associated with a reduction in circulating t-PA concentrations. Methotrexate can damage vascular endothelium, which may trigger thrombosis (Chap. 51).³³ Hemostatic drugs used therapeutically include the synthetic lysine derivatives aminocaproic acid and tranexamic acid, which bind reversibly to plasminogen; the bovine protease inhibitor aprotinin, which inhibits kallikrein; the vasopressin analog desmopressin, which increases plasma concentrations of factor VIII and vWF; and conjugated estrogens, which normalize bleeding times in uremia.⁶²

Platelets

In the resting state, platelets maintain a discoid shape. The platelet membrane is a typical trilaminal membrane with glycoproteins, glycolipids, and cholesterol embedded in a phospholipid bilayer. The plasma membrane is in direct continuity with a series of channels, the surface-connected canalicular system (SCCS), which is sometimes referred to as the open canalicular system. The SCCS provides a route of entry and exit for various molecules, a storage pool for platelet glycoproteins, and an internal reservoir of membrane that may be recruited to increase platelet surface area. This facilitates platelet spreading and pseudopod formation during the process of cell adhesion.

The glycocalyx, or outer coat, is heavily invested with glycoproteins that serve as receptors for a wide variety of stimuli. The β₁-integrin family includes receptors that mediate interactions between cells and mediators in the extracellular matrix, including collagen, laminin, and fibronectin. The β₂-integrin receptors are present in inflammatory cells and platelets and are important in immune activation. The β₃-integrin receptors (also known as cytoadhesins) include the glycoprotein (GP) IIb-IIIa fibrinogen receptor, as well as vitronectin.⁴⁶ Vitronectin has binding sites for other integrins, collagen, heparin, and components of complement. All of the integrins are active in the process of platelet adhesion to surfaces. Platelet aggregation is mediated by the GP IIb-IIIa receptors.⁴⁶

The submembrane region contains actin filaments that stabilize the platelets’ discoid shape and are involved in the formation and stabilization of pseudopods. They also generate the force needed for the movement of receptor-ligand complexes from the outer plasma membrane to the SCCS. These mobile receptors are important in the spreading of platelets on surfaces, and for binding fibrin strands and other platelets. Platelet cytoplasm contains three types of membrane bound secretory granules.⁴⁶ The α granules contain β-thromboglobulin, which mediates inflammation, binds and inactivates heparin, and blocks the endothelial release of prostacyclin. In addition, platelet factor-4, which inactivates heparin, and fibrinogen are contained within the α granules. Dense granules store adenine nucleotides, serotonin, and calcium, which are secreted during the release reaction. Platelet lysosomes contain hydrolytic enzymes. Stimulation by platelet agonists causes the granules to fuse with the channels of the SCCS, driving the contents out of the platelets and into the surrounding media.

Platelet Adhesion.

On the vessel wall, collagen, von Willebrand factor (vWF), and fibronectin are the adhesive proteins that play the most prominent role in the adhesion of platelets to vascular subendothelium. On the exposure of collagen (eg, following a laceration or the rupture of an atherosclerotic plaque), platelet adhesion is triggered. Under conditions of high shear (flowing blood), platelet adhesion is mediated by the binding of GP Ib-V-IX receptors on platelet membranes to vWF in the vascular subendothelium.⁴⁶ Following adherence of platelets to subendothelial vWF, a conformational change in GP IIb-IIIa on platelet membrane occurs, activating this receptor complex to ligate vWF and fibrinogen. The result is the amplification of platelet adhesion and aggregation. An important interaction occurs between thrombosis and inflammation. Platelet-activating factor is synthesized and coexpressed with P-selectin on the surface of the endothelium in response to mediators such as histamine or thrombin. Platelet-activating factor interacts with a receptor on the surface of neutrophils that activates the CD11/CD18 adhesion complex, and results in adhesion of neutrophils to endothelium and to platelets. This results in the synthesis of leukotrienes and other mediators of inflammation.

Platelet Activation.

Thrombin, collagen, and epinephrine can activate platelets. In response to thrombin, granules fuse with each other and with elements of the SCCS to form secretory vesicles. These vesicles are believed to fuse with the surface membrane, releasing their contents into the surrounding medium. The membranes of the secretory granules become incorporated into the platelet surface membrane.

Platelet Aggregation.

Following activation, GP IIb-IIIa is expressed in active form on platelet surface, serving as the final common pathway for platelet aggregation regardless of inciting stimulus. This receptor binds exogenous calcium and fibrinogen. GP IIb-IIIa ligates fibrinogen along with fibronectin, vitronectin, and vWF, resulting in the binding of platelets to other platelets, and ultimately the formation of the platelet plug. Collagen-induced platelet aggregation is mediated by adenosine diphosphate (ADP) and thromboxane A₂. ADP binds to the metabotropic purine receptors P2Y1 and P2Y12, leading to transient and sustained aggregation, respectively.³⁵ Thromboxane A₂ is formed from arachidonic acid by the action of cyclooxygenase-1. It is a potent vasoconstrictor and inducer of platelet aggregation and release reactions.⁴⁶ Platelets participate in triggering the coagulation cascade by binding coagulation factors II, VII, IX, and X to membrane phospholipid, a calcium-dependent process.

Antiplatelet Xenobiotics

Aspirin.

Aspirin inhibits prostaglandin H synthase (cyclooxygenase {COX}) by irreversibly acetylating a serine residue at the active site of the enzyme. Aspirin inhibition of the COX-1 isoform of this enzyme is 100–150 times more potent than its inhibition of the COX-2 isoform. The inhibition of COX-1 results in the irreversible inhibition of thromboxane A₂ formation. Because platelets can be activated by other mechanisms including thrombin, thrombosis can develop despite aspirin therapy (Chap. 39).⁹⁰

Selective COX-2 Inhibitors.

Platelets express primarily COX-1 and use it to produce mostly thromboxane A₂, which leads to platelet aggregation and vasoconstriction. Endothelial cells express COX-2 and use it to produce prostaglandin I₂, an inhibitor of platelet aggregation and a vasodilator. Whereas aspirin and traditional (nonselective) nonsteroidal antiinflammatory medications inhibit the production of thromboxane A₂ and prostaglandin I₂ at both sites, the selective COX-2 inhibitors do not affect platelet-derived thromboxane A₂, perhaps accounting for the increase in cardiovascular events associated with long-term use of some of these xenobiotics.^27,99

Glycoprotein IIb-IIIa Antagonists.

The GP IIb-IIIa antagonist abciximab is a chimeric human-murine monoclonal antibody that binds the GP IIb-IIIa receptor of platelets and megakaryocytes. Two synthetic ligand-mimetic antagonists have also been developed: eptifibatide and tirofiban. These parenteral antagonists are used primarily in patients undergoing percutaneous coronary interventions.⁴⁶ By blocking the fibrogen binding site of IIb-IIIa, platelet aggregation is blocked and even reversed regardless of the inciting activation. Reversible thrombocytopenia can occur within hours of initiation of these xenobiotics.⁸⁶

ADP Receptor Inhibitors.

The thienopyridines clopidogrel, ticlopidine, and prasugrel antagonize platelet aggregation by irreversible inhibition of ADP binding to the P2Y₁₂ receptor. All three prodrugs are associated with TTP, as well as neutropenia and aplastic anemia.^{12, 72, 73} The purine analogs ticagrelor and cangrelor are direct-acting, reversible allosteric antagonists that bind to a different site on the P2Y₁₂ receptor.

Dipyridamole.

The pyrimidopyrimidine derivative dipyridamole inhibits cyclic nucleotide phosphodiesterase in platelets, resulting in the accumulation of cyclic adenosine monophosphate and perhaps cyclic guanine monophosphate. It also inhibits cellular reuptake of adenosine, blocks thromboxane synthase, and inhibits the thromboxane receptor.

Xenobiotic-Induced Thrombocytopenia

Multiple xenobiotics are reported to cause thrombocytopenia, either via the formation of drug-dependent antiplatelet antibodies or as TTP. Drug-induced platelet antibodies are estimated to occur in 1 in 100,000 drug exposures. Reversible drug binding to platelet epitopes, such as GP Ib-V-IX, GP IIb-IIIa, and platelet–endothelial cell adhesion molecule-1, leads to a structural change that can form or expose a neoepitope target for antibody formation. The presence of the drug is required for antibody binding and increased platelet destruction, but there is no covalent bond (as occurs in the hapten model of penicillin binding to the erythrocyte membrane).

Thrombocytopenia can also occur as a result of spurious clumping, pregnancy, hypersplenism with cirrhosis, idiopathic thrombocytopenic purpura, heparin-induced thrombocytopenia (Chap. 60), bone marrow toxicity, and TTP. After excluding these conditions and nontherapeutic exposures, a systematic literature search updated annually lists more than 1000 cases reported in English involving more than 150 xenobiotics. Table 22–7 lists the xenobiotics appearing in multiple cases, satisfying criteria for probable to definite causality including drug rechallenge. Nevertheless, a common clinical problem is to distinguish drug-induced thrombocytopenia from idiopathic thrombocytopenic purpura in a patient on multiple medications who develops thrombocytopenia. In the absence of validated laboratory assays for drug-dependent platelet antibodies other than heparin, diagnosis still depends on the clinical course following drug discontinuation and perhaps rechallenge. Large databases provide some guidance regarding past reported experience.^{37, 60, 82} Severe (<20,000 platelets/mm³), or acute transient thrombocytopenia are more likely to be drug-induced.¹⁰

In patients administered the sensitizing agent de novo, 5 to 7 days are typically required for the development of the immune response. During rechallenge, thrombocytopenia can develop within 12 hours.¹⁰ Interestingly, the unique ability of GP IIb/IIIa inhibitors such as abciximab to cause thrombocytopenia within hours of first use suggests the presence of preformed antibodies directed against platelet epitopes, perhaps accounting for ex vivo clumping of platelets observed in approximately one out of 500 normal patients.⁸³ Clinically, fever, chills, pruritus, and lethargy may occur. The onset of life-threatening bleeding may be abrupt. Hemorrhagic vesicles may be seen in the oral mucosa. Laboratory investigations will demonstrate an absence of platelets on peripheral smear, prolongation of the bleeding time, deficient clot retraction, and an abnormal prothrombin consumption test. Bone marrow aspiration will demonstrate normal or increased numbers of megakaryocytes and immature forms. Treatment includes the transfusion of blood products, glucocorticoids, and the withdrawal of the offending xenobiotic.¹⁰

Thrombotic Thrombocytopenic Purpura.

TTP is characterized by the triad of microangiopathic hemolytic anemia, severe thrombocytopenia, and fluctuating neurologic abnormalities.⁶⁶ Fever and acute kidney injury are also common, although overt kidney failure is rare. The hallmark is the presence of platelet aggregates throughout the microvasculature, without fibrin clot, unlike the fibrin-rich thrombi seen in disseminated intravascular coagulation or the hemolytic uremic syndrome. In the acquired form, drug-induced autoantibodies inactivate a circulating zinc metalloprotease ADAMTS13, thereby blocking its ability to depolymerize large multimers of vWF and leading to platelet clumping.^{8, 9, 12, 100} Plasma exchange with fresh frozen plasma replenishes ADAMTS13 and removes the inhibitory antibodies. Prior to understanding of the molecular mechanism, other causes of microvascular hemolysis with thrombocytopenia were often confused with TTP. These causes included hemolytic uremic syndrome due to shiga toxin, the HELLP syndrome of pregnancy, Rocky Mountain spotted fever, and paroxysmal nocturnal hemoglobinuria. Also included were cases secondary to xenobiotics such as cyclosporine, cocaine, gemcitabine, mitomycin C, and cisplatin, in which ADAMTS13 antibodies are not present.^31,100

Heparin-Induced Thrombocytopenia.

An immune response to both unfractionated and low-molecular weight heparin, manifested clinically by the development of thrombocytopenia and, at times, venous thrombosis, is now recognized to result from IgG antibodies against platelet factor 4, a heparin-binding chemokine.¹ The diagnosis is confirmed by testing for these antibodies. The heparin–platelet factor 4 antibody complex results in the release of procoagulant microparticles which release tissue factor, provide an anionic phospholipid membrane surface on which clotting cascade complexes assemble, and can lead to paradoxical thrombosis, which can be limb- or life-threatening^55,56 (Chap. 60).

• The mechanisms of toxic injury to the blood are extremely varied, reflecting the complexity of this vital fluid.

• Virtually all xenobiotics come into contact with blood, and have the potential to disrupt its essential functions of oxygen transport, gas exchange, signaling, host defense, and hemostasis.

• Xenobiotics may directly injure mature cells, or prohibit their development by injuring the stem cell pool.

• Oxygen transport can be disrupted in several distinct ways, including oxidation of the heme iron, protein globin chain disruption, and shift of the oxygen-hemoglobin dissociation curve and hemolysis.

• A common theme in hematologic toxicity is the perturbation of homeostatic equilibria that exist between cell proliferation and apoptosis, between immune activation and suppression, or between thrombophilia and thrombolysis.

Acknowledgment

Diane Sauter, MD, contributed to this chapter in earlier editions.

References