Opinion on the Optimal Histologic Evaluation of the Bone Marrow in Nonclinical Toxicity Studies

Abstract

Identification of bone marrow toxicity is an important issue in drug development and toxicologic pathologists play a critical role in that identification. Knowledge of the general components of bone marrow, relevant anatomical and species differences, and the standard approach (routine systematic histological evaluation of the bone marrow in conjunction with analysis of the peripheral complete blood count data) will be reviewed. Specific morphologic features that anatomic pathologists should look for in the various components of bone marrow as well as suggested terminology for bone marrow findings will be discussed. Finally, an opinion on the limitations of the standard approach to bone marrow evaluation will be provided including general recommendations on when additional methods (image analysis of hematoxylin and eosin stained slides, flow cytometry or Sysmex XT 2000iV analysis, cytological evaluation of bone marrow smears, in vitro models, and transmission electron microscopy) might be useful in the detection or further characterization of bone marrow toxicity.

This is an opinion article submitted to the Toxicologic Pathology Forum. It represents the views of the author. It does not constitute an official position of the Society of Toxicologic Pathology, British Society of Toxicological Pathology, or European Society of Toxicologic Pathology, and the views expressed might not reflect the best practices recommended by these Societies. This article should not be construed to represent the policies, positions, or opinions of their respective organizations, employers, or regulatory agencies.

Introduction

Due to the rapidly dividing nature of bone marrow precursor cells, hematotoxicity is an important and often limiting complication of some test articles. ^1,2 As such, it is important to evaluate the bone marrow in nonclinical toxicity studies. ^3,4 The standard is to evaluate the peripheral complete blood count (CBC) data in conjunction with bone marrow histology. ^3,4 Hematologic data should be interpreted with respect to all study data (study procedures, in life findings, toxicokinetic data, temporal relationship to dosing, other histologic findings, etc), and dose relationships should be assessed through comparison with age and sex-matched concurrent controls. ^{2 –5} Even when this standard approach is followed, there are limitations to what can be determined by histological evaluation of the bone marrow and important anatomical and species differences of which to be aware. ⁴

Bone marrow is found in both axial and long bones and is protected between the outer cortical bone and inner trabeculae of cancellous bone. ^6,7 It is primarily composed of foci of hematopoietic cells, adipose tissue, and interspersed highly vascular connective tissue stroma and stromal cells. ^6,7 The stromal components, including endothelial cells, reticular cells, fibroblasts, adipocytes, macrophages, mast cells, nerve fibers, and reticulin, provide structural support and secrete soluble mediators essential for the maintenance, differentiation, and growth of hematopoietic stem cells. ^7,8 For example, endothelial cells regulate hematopoiesis by producing extracellular matrix, stem cell factor, and various other cytokines. ⁷ Recent research suggests that bone marrow adipocytes are metabolically active and also likely play a role in hematopoiesis through direct contact with cells as well as secretion of adipocyte-derived factors such as adiponectin, leptin, prostaglandins, and interleukin-6 ⁹ .

Hematopoiesis begins with a pluripotent cell which gives rise to either a lymphoid stem cell or a multipotent myeloid stem cell, respectively capable of developing into lymphocytes or erythrocytes, granulocytes, platelets, monocytes, or mast cells. ^6,7 In the bone marrow, erythropoiesis takes place in distinct islands supported by “nurse” macrophages, while granulopoiesis occurs in less distinct regions, and megakaryopoiesis occurs next to sinus endothelium. ^6,7 Mature platelets are released directly into the blood from megakaryocyte cytoplasmic processes that penetrate through the wall of the sinuses. ⁶ On the other hand, erythroid and myeloid cells traverse the wall of the venous sinuses to enter the blood stream. ⁶ Under normal homeostatic conditions, hematopoiesis is a continuous process such that the loss of mature circulating blood cells is replaced by the production of new ones. ^6,7 However, all stages of hematopoiesis are sensitive to external influences and can be suppressed in response to many factors including decreased nutritional state, chronic inflammation, proliferative or neoplastic disorders, or direct test article-related toxicity. ^6,10

Importance of a Consistent Integrative Approach to Bone Marrow Evaluation

The need to holistically interpret all study findings is best exemplified by the fact that feed restriction, decreased body weight, and/or decreased food intake, particularly in rodents, can have a major (secondary) impact on the hematopoietic system and should not be mistaken for a direct test article-related effect. ⁴ In rats, a study looking at the effect of food restriction of 25% or greater for 2 weeks resulted in decreased circulating white blood cells, primarily lymphocytes, as well as decreased circulating platelet counts. ¹¹ In the bone marrow, depending on the degree (25% to 75%) of feed restriction, there was a slight to moderate decrease in overall hematopoietic cellularity with mild feed restriction and bone marrow necrosis with more severe feed restriction. ¹¹ In mice, decreased protein intake has been associated with bone marrow cellular depletion and a halt in the hematopoietic precursor cell cycle. ¹² A similar effect, although typically of lower magnitude, has been observed in various dog breeds, including beagles, where after 4 weeks of food restriction they had decreased reticulocyte ratios and white blood cells as well as decreased cellularity of the bone marrow. ¹³ In humans with anorexia nervosa, changes in the CBC (decreases in red blood cell [RBC] mass parameters, white blood cells, and platelets) and morphologic alterations in the bone marrow (bone marrow atrophy and gelatinous transformation) are frequently reported. ^14,15 The peripheral changes are not predictive of the severity of bone marrow atrophy and all of these changes resolve with nutritional rehabilitation. ^14,15 In nonclinical toxicity studies, these hematologic changes are consistent with commonly observed stress-related decreases in circulating white blood cells and decreases in overall cellularity of the bone marrow. ¹⁶

Similarly, due to possible inherent variability in bone marrow morphology resulting from various factors (age, sex, strain, study conditions, etc), the evaluation of bone marrow histology requires comparison of treatment groups to concurrent age and sex-matched controls of the same anatomic site at the same timepoint in the study. ⁴ The importance of appropriate age-matched controls in nonclinical toxicity studies is supported by a publication documenting differences in bone marrow sample cellularity in Fisher 344 rats of various ages. ¹⁷ It is known that bone marrow cellularity generally decreases with age. ¹⁷ Active marrow has a predominance of mature erythrocytes, resulting in a macroscopically observed deep red color while less active marrow has more adipocytes and a yellow color. ¹⁸ In addition, in Fisher 344 rats, it was reported that the cellularity of bone marrow samples varied between different bones. ¹⁷ In the author’s experience, similar variations have been observed in the cellularity of sternal versus femoral bone marrow sections in Wistar Han rats used in nonclinical toxicity studies at my institution (All procedures performed on animals at the authors institution were in accordance with regulations and established guidelines and were reviewed and approved by an Institutional Animal Care and Use Committee). Lastly, long bones, such as the proximal femur and humerus, should be avoided in dogs, as they typically contain primarily adipose tissue in the diaphyseal marrow cavities; thus, the sternum, rib, or vertebrae should be sampled for histologic evaluation of canine bone marrow. ¹⁹

Awareness of Species Differences in Bone Marrow Assessment

As nonclinical evaluation of test articles often includes an evaluation of both a rodent (mouse or rat) and nonrodent species (dog, nonhuman primate, or less often mini pig), it is also important to have a general knowledge of what is normal in the species being evaluated when assessing for test article-related microscopic effects on the bone marrow. There are a few differences in rodents as compared to the typical nonrodent species used in nonclinical toxicity studies. First, rodents can have substantial extramedullary hematopoiesis in the spleen and, to a lesser degree, the liver. ^3,4 This means that in the overall assessment of the hematopoietic system of rodents, histology of the spleen, and possibly the liver, in addition to histology of the bone marrow, is warranted. ^4,7 In general, rodents have a greater proportion of hematopoietic tissue as compared to adipose tissue in their bone marrow than other species. ¹⁰ Rats typically have relatively high numbers of mast cells (up to 1%-3% of nucleated bone marrow cells) in their bone marrow. ^10,20 Lastly, rodents typically have a higher number of lymphocytes in their bone marrow as compared to large animal species. ⁴ Flow cytometry to differentiate bone marrow components of nonclinical species into myeloid, erythroid, and lymphoid components demonstrated generally higher lymphoid percentages in mice and rats (13.3-16.0 and 21.7-24.5, respectively) than dogs and monkeys (4.7-5.3 and 10.5-14.8, respectively). ²¹ Despite this, lymphoid cell aggregates or lymphoid follicles are not usually observed in the bone marrow of normal rodents or even after immunization.^4,6 In hematoxylin and eosin (H&E)-stained sections of bone marrow, when lymphocytes are aggregated or forming follicles, they can be readily differentiated, but scattered individual mature lymphocytes cannot be unequivocally differentiated from other mononuclear bone marrow cells due to their indistinct morphology and/or low cell numbers. ^3,4 The inability to differentiate mature lymphocytes from other bone marrow cells is particularly relevant in rodents since they have a higher number of lymphocytes in their bone marrow compared to nonrodent species. Immunohistochemical (IHC) staining methods to identify lymphoid cells in formalin-fixed, paraffin-embedded sections of various rodent tissues including bone marrow have been described ²² ; however, standard processing of bone marrow samples for routine H&E histologic examination includes decalcification, which may not be ideal for IHC. An alternative option would be to collect a cast (or scoop) of bone marrow from the diaphysis of a long bone at necropsy and fix and process without decalcification if bone marrow IHC is anticipated to be needed. ¹⁰

General Approach to Bone Marrow Histology

For optimal results, histology of the bone marrow should begin with an adequate quality H&E-stained slide which is typically generated from 10% neutral-buffered formalin-fixed samples that are decalcified, paraffin embedded, sectioned, mounted on a glass slide then stained with H&E. ⁴ Histological evaluation of the bone marrow allows for a subjective evaluation of overall hematopoietic cellularity, megakaryocyte number, and limited morphology, with a general estimation of the proportion and maturation of granulocytic and erythroid cells, as well as iron content stored as hemosiderin in macrophages. ⁴ Other abnormalities such as necrosis, inflammation, vascular or stromal proliferations, and proliferative or neoplastic disorders can also be identified. ^3,4 Since the structure of bone marrow tissue is preserved, histology is more sensitive than cytology for identifying subtle focal lesions and other changes to the structural environment. ⁴ Histology is also better for estimating megakaryocyte numbers than cytology, because in cytologic preparations, megakaryocytes are associated with bone spicules and the numbers of spicules, and therefore megakaryocytes, can vary widely. ⁵ Specific cell types that can be identified by histological evaluation of the bone marrow include more mature granulocytic, erythroid, and megakaryocytic cells as well as adipocytes, macrophages (with or without hemosiderin), and mast cells. ^4,7,18 In general, erythroid cells are smaller and darker with round, dense, deeply basophilic nuclei until these are extruded, after which the deeply basophilic cytoplasm gets more eosinophilic with maturation. ⁷ Erythroid precursors, in particular, can be difficult to distinguish from individual lymphocytes with histological evaluation. ⁴ Granulocytes are often larger and lighter, with larger bean or ring-shaped nuclei that are less basophilic and more vesicular than other hematopoietic cell types, along with more eosinophilic cytoplasm. ⁷ Individual megakaryocytes are readily recognized because of their large size and multilobulated nuclei though care should be taken not to confuse them with multinucleated osteoclasts lining Howship’s lacunae. ⁷ Megakaryocytes go through a sequence of nuclear duplication without cell division (endomitosis) and this results in a large cell with abundant cytoplasm and a polyploid nucleus. ^18,23 Adipocytes, macrophages, and mast cells look the same as in other tissues. Hemosiderin in macrophages appears as a golden brown pigment in H&E-stained bone marrow sections. ⁷ Cells that cannot reliably be identified by histological evaluation of the bone marrow include hematopoietic stem cells, immature granulocytic, erythroid cells, and megakaryocytic cells, as well as monocytes, lymphocytes, and stromal cells due to their low cell numbers and/or indistinctive morphology. ^4,18 However, there are some exceptions to this including the identification of stromal cells when there are proliferative stromal lesions and lymphocytes in cases of lymphoma due to the larger cell numbers being present. ^3,18

The approach to bone marrow histology should start with examination at low magnification to first evaluate section and staining adequacy as well as tissue architecture. ^4,18 Then, overall hematopoietic cellularity can be estimated as normal, decreased, or increased by estimating the proportion of hematopoietic cells to adipocytes, as their content varies inversely, and comparing these percentages to concurrent control animals as well. ^3,4 The proportions vary between species, but a general rule of thumb is normal dog bone marrow has 50% hematopoietic tissue and 50% adipocytes. ⁶ In rodents, bone marrow has been shown to have approximately 70% to 80% hematopoietic tissue and approximately 20% to 30% adipocytes. ⁶ A qualitative myeloid: erythroid (M:E) ratio can be determined by estimating the proportion of lighter staining myeloid to darker staining erythroid cells. ⁴ Although the M:E ratio varies by species, in most mammals it is typically >1 and is independent of total bone marrow cellularity. ⁸ In rodents, because they have a larger proportion of lymphocytes in their bone marrow that cannot easily be distinguished from erythrocytic cells, the cells that are perceived to be erythrocytic likely also include lymphocytes. ⁴ Standard histologic examination of bone marrow should also include an evaluation of megakaryocyte number and distribution, including an assessment of megakaryocyte morphology at higher magnifications. ⁴ The cellular morphology of myeloid and erythroid elements should also be evaluated at higher magnifications along with an estimate of the relative proportions of immature versus mature precursors, synchrony of maturation, and evidence of dysplasia or neoplasia. ⁴ Additionally, the stroma including vasculature, as well as hemosiderin stores, and any findings observed at lower magnification should be evaluated at higher magnifications. ⁴

Specific Bone Marrow Histologic Findings

Test article-related microscopic findings in the bone marrow should be recorded and graded in the individual animal data, ideally using descriptive or semiquantitative terms. ⁴ A summary of the more common histologic findings in the bone marrow follows although additional detailed information and/or images are also available, including relevant International Harmonization of Nomenclature and Diagnostic Criteria (INHAND) documents ^10,19,24 and the National Toxicology Program nonneoplastic lesion atlas. ²⁰ Generally, these findings include variations in cellularity, dyshematopoiesis, bone marrow structural changes, degenerative changes, inflammation, and neoplasia. ⁷

Variation in cellularity of hematopoietic tissues can involve one or more of the erythroid, myeloid (granulocytic or monocytic), or megakaryocytic cell lines. ³ If only one cell line is affected and it is possible to determine the lineage, this should be documented; if these changes involve primarily the erythroid or myeloid cell lines, the M:E ratio may also be altered. ³ It should be noted that a finding of increased or decreased M:E ratio is very difficult to interpret in isolation, so ideally the use of this term should be accompanied by a description of what is altering the M:E ratio (eg, increased myeloid cellularity, decreased erythroid cellularity, etc). If the M:E ratio is increased, there is either increased myeloid cellularity or decreased erythroid cellularity or a combination of both; the CBC results can be used to help determine which is present. If circulating neutrophil count is high and RBC mass parameters are normal, then these CBC parameters likely correlate with increased myeloid cellularity. Conversely, if there was a normal circulating neutrophil count and decreased RBC mass parameters then this could be interpreted as decreased erythroid cellularity. Of note, high circulating neutrophil count and decreased RBC mass parameters could be associated with both increased myeloid cellularity and decreased erythroid cellularity.

Increased hematopoietic cellularity is generally characterized by an increase in the proportion of hematopoietic cells relative to adipocytes as compared to concurrent controls. ³ When increased in number, hematopoietic cells typically retain normal morphology and orderly maturation. ³ Increased hematopoietic cellularity can be a direct response to compound administration (eg, cytokines), but more commonly is a secondary response to increased cell demand. ^3,20 Increased erythroid cellularity is expected with increased peripheral demand for RBCs and this may be associated with decreased hemosiderin content in bone marrow macrophages and/or increased extramedullary hematopoiesis in the spleen and or liver, particularly in rodents. ^3,5,7 Increased granulocyte cellularity often reflects inflammation and/or necrosis somewhere in the body or can be a direct response to compound administration (eg, granulocyte-macrophage colony-stimulating factor). ^3,5,7 When there is peripheral demand for granulocytes, the bone marrow reserve will accommodate resulting in a shift to more immature myeloid forms in the bone marrow. ^3,5 Increased megakaryocyte cellularity is typically associated with decreased circulating platelets, often due to destruction or increased consumption. ^3,5 There can also be alterations in bone marrow mast cell numbers, which are typically characterized by increased numbers of loosely scattered mature mast cells. ⁷ However, it is important to compare mast cell numbers with concurrent controls, because as previously stated mast cells are normally observed in the bone marrow, particularly in older rodents. ^7,10,20

Decreased hematopoietic cellularity is characterized by a decrease in hematopoietic cells relative to adipocytes as compared to concurrent controls. ³ Decreases in hematopoietic cellularity can be caused by multiple factors such as toxicity, decreased nutritional status, inflammation, neoplasia, and the normal aging process. ^3,20 More specifically, decreased erythroid cellularity can be due to certain endocrine diseases or chronic renal or hepatic disease, in which case the CBC may reflect a non or poorly regenerative anemia. ^3,7 Decreases in granulocytic cells or megakaryocytes most commonly occur with concurrent changes in erythroid cells as a pancytopenia, but each cell line can be individually affected. ⁷ Decreases in granulocytic cells in the bone marrow are usually accompanied by neutropenia. ⁷ Similarly, decreases in bone marrow megakaryocyte numbers are usually accompanied by thrombocytopenia. ⁷ In this case, it is important to be aware that, by histological evaluation, it is difficult to tell the difference between megakaryocyte precursors that cannot yet produce platelets from immature megakaryocytes that have lower ploidy but can produce smaller numbers of larger sized and more active platelets than more mature megakaryocytes. ²³ This is an example where cytological examination of bone marrow is favored. Due to the short life span of neutrophils in circulation, when myelotoxicity occurs, neutropenia is often the earliest change detected in the peripheral blood; it can occur as early as 1 to 3 days postexposure to high doses of cytotoxic agents. ¹⁸ Decreased reticulocyte numbers may be detected as early as 2 to 3 days postexposure, and thrombocytopenia may be detected after 9 to 14 days postexposure, although effects on RBC mass parameters may not be detectable given the relatively long life span of erythrocytes in circulation. ¹⁸

As mentioned above, changes in hematopoietic cellularity are typically associated with opposite changes in adipocyte cellularity. A common change with bone marrow adipocytes is a decrease due to displacement by increased hematopoietic cells or due to decreased nutritional status which in severe cases can progress to serous atrophy of fat. ^3,10 Similarly, there can be an increase in bone marrow adipocytes, but care should be taken to limit this diagnosis to instances where it is considered a primary change and not simply a response to a decrease in hematopoietic cells. ^10,20

Dyshematopoiesis refers to disorders with abnormal hematopoietic cell morphology or maturation with ineffective hematopoiesis and associated cytopenia. ⁷ Observable changes can include hematopoietic cells with altered cell and/or nuclear size, nuclear/cytoplasmic ratio, morphology, maturation, and/or asynchronous nuclear and cytoplasmic development. ¹⁰ There can be accompanying changes in M:E ratio, cellularity, and/or the presence of blasts and peripheral blood smears may contain abnormal cells. ¹⁰ This is typically identified by cytological evaluation of bone marrow, as this allows better assessment of cellular morphology and maturation. ^3,10

Two additional microscopic findings observed in hematopoietic cells that are described in relevant INHAND documents are hypersegmentation of granulocytes and emperipolesis. Hypersegmentation of granulocytes, in rodents, is characterized by increased numbers of mature granulocytes with hypersegmented (6 or more nuclear lobes) nuclei in the peripheral blood and in the bone marrow. ¹⁰ One of the most common causes of hypersegmented granulocytes is endogenous or exogenous corticosteroids. ¹⁰ Emperipolesis is characterized by the active engulfment of one cell by another cell, with both cells remaining intact and it is most often observed in megakaryocytes that contain neutrophils or other hematopoietic cells. ¹⁹ Emperipolesis of primarily neutrophils by megakaryocytes has been reported to occur in severely feed-restricted rats and debilitated animals in general as well as rodents given a growth factor. ^3,11 As this finding has been observed in healthy rats, cynomolgus monkeys, and humans, its significance is unknown. ²⁵

Microscopic findings involving other bone marrow stroma or vasculature include fibrosis, angiectasis, and increased macrophages. Fibrosis in the bone marrow medullary cavity is characterized by increased extracellular matrix (collagen and/or reticular fibers) with or without an associated proliferation of cellular elements (fibroblasts and/or reticular cells). ^7,10 In cases of extensive fibrosis, there may not be any remaining hematopoietic tissue. ⁷ Histology of routine H&E-stained slides is often sufficient to diagnose fibrosis but if needed it can be confirmed with a histochemical stain (trichrome). ⁵ Bone marrow fibrosis is not a common primary finding in nonclinical toxicity studies but anatomic pathologists should be aware that it has occasionally been observed as a focal background finding in rats. ^7,10,20 When observed, it is typically secondary to another process such as inflammation, necrosis, or neoplasia and cytokines from bone marrow macrophages, megakaryocytes, and platelets seem to play a role. ^10,20 The term myelofibrosis is associated with a human disorder, and as a similar disorder has not been documented in rodents, this term should not be used in nonclinical toxicity studies. ^7,20 Angiectasis is characterized by abnormally dilated vascular spaces lined by normal-appearing endothelial cells. ¹⁰ The continuous endothelial lining of the vascular spaces differentiates this finding from acute hemorrhage. ⁷ Angiectasis can be found in association with other bone marrow findings, such as decreased hematopoietic cells, inflammation, or neoplasia. ²⁰ Increased numbers of macrophages can be seen in the bone marrow, often in the form of disorganized aggregates. ⁷ Macrophages may increase in number and/or size due to increased demand for phagocytosis or to support erythropoiesis. ¹⁰ The cytoplasm of macrophages may contain phagocytized material and/or vacuoles. ¹⁰

Bone marrow necrosis may occur with severe injury and is usually the coagulative type. ¹⁸ It is best recognized by histology and is characterized by remnants of hematopoietic and stromal cells with indistinct borders, cytoplasmic vacuolation or eosinophilia, and nuclear pyknosis, karyolysis, or karyorrhexis. ^5,18,20 This can range from small numbers of scattered individual necrotic hematopoietic cells to larger areas of confluent necrosis; the latter likely being indicative of vascular compromise (infarction) such as that evident with thrombosis. ^3,20 When severe, there may be replacement of normal stroma by debris, mineralization, and/or hemorrhage. ¹⁸ Necrosis needs to be distinguished from autolysis and processing artifact. ³ It has been reported that megakaryocytes are the first cells to show signs of autolysis, in particular pyknosis of the nucleus. ³ Causes of bone marrow necrosis include direct toxicity from test article administration, infectious disease and, as previously mentioned, obstruction of the blood supply or severe feed restriction in rats. ^3,20 Serous atrophy of fat (or gelatinous transformation) is an uncommonly diagnosed degenerative change of adipocytes in nonclinical toxicity studies; this finding is characterized by atrophy of adipocytes, decreased numbers of hematopoietic cells, and replacement by eosinophilic extracellular ground substance. ^3,20 As this finding is associated with cachexia or advanced malnutrition, it is rarely seen in nonclinical toxicity studies because animals are typically electively euthanized before reaching the severe degree of weight loss associated with this change. ^10,20

In addition to increased cellular infiltrates without tissue disruption, there can also be cellular accumulations that are part of an active inflammatory process which is often associated with other changes in the tissue. ²⁰ In the bone marrow, granulomatous and suppurative/neutrophilic inflammation are the 2 types of inflammation that can sometimes be observed (although rarely); other categories of inflammation are also possible but are often difficult to appreciate in the bone marrow by histological examination because the cells are part of the resident population or are difficult to distinguish from hematopoietic cells. ^3,20 Granulomatous inflammation is characterized by vacuolated or epithelioid macrophages that occur in aggregates or granulomas, which can be associated with lymphocytes, plasma cells, neutrophils, or fibroblasts. ²⁰ In canine bone marrow, granulomatous inflammation with small lipid vacuoles present both within macrophages and extracellularly has been described. ²⁰ Suppurative/neutrophilic inflammation is characterized by aggregates of intact and necrotic neutrophils and there can also be fewer numbers of scattered lymphocytes, plasma cells, and macrophages as well as edema, fibrin, hemorrhage and/or vascular dilation. ²⁰ Bone marrow inflammation may be associated with test article administration or infectious agents. ^3,20

Hematopoietic neoplasms are rarely observed in short-term toxicology studies but can be spontaneous findings in older animals. ^2,26 Hematopoietic neoplasms are often associated with increased bone marrow cellularity and are generally classified as lymphoid or nonlymphoid. ^2,27 Lymphoid neoplasms include several types of lymphomas as well as plasma cell neoplasia. ²⁷ Typically the lymphocyte population in bone marrow is relatively small and difficult to definitively differentiate from other mononuclear cell types using routine histological evaluation of H&E-stained sections but, in the case of lymphoma, the neoplastic process can often be identified given that neoplastic lymphocytes can account for ≥30% of the nucleated hematopoietic cell population. ^18,27 Nonetheless given the morphological similarities between hematopoietic neoplasms, it can still be difficult to determine which bone marrow mononuclear cell lineage is neoplastic without the aid of additional techniques such as cytological evaluation and IHC to confirm cell lineage and/or lymphocyte subtypes (B or T cells). ^10,22,27 Nonlymphoid hematopoietic neoplasms include erythroid leukemia, megakaryocytic leukemia, myeloid leukemia, mast cell leukemia, and leukemia not otherwise specified (NOS). ^10,27 Erythroid leukemia consists of an increase in cells of the erythroid lineage at various stages of differentiation. ¹⁰ Myeloid leukemia develops from the granulocytic or monocytic cell lineages and can consist of immature to mature forms. ¹⁰ Megakaryocytic leukemia consists of an increase in megakaryocytes at various stages of differentiation and is rare in conventional mice. ¹⁰ Mast cell neoplasia in the bone marrow of rats is characterized by sheets of cells with compression of adjacent tissue architecture. ⁷ Mast cell leukemia has also been reported as a spontaneous neoplasm in a Sinclair minipig. ²⁴ Leukemia NOS can be used when a definitive cell lineage cannot be determined. ¹⁰ Bone marrow cytology and/or peripheral blood smear evaluation can be beneficial for the diagnosis of hematopoietic neoplasia. ¹⁰ Other primary neoplasia in the bone marrow includes fibrosarcoma, histiocytic sarcoma, and hemangiosarcoma; metastatic neoplasia is also possible. ^3,27

Other Techniques for Analysis of Bone Marrow

Recently, an automated image analysis method that quantitates overall bone marrow cellularity using H&E-stained slides was developed. ²⁸ The slides were scanned with a Nanozoomer HT 2.0 slide scanner, and Definiens Developer v2.6 software was used for image analysis. ²⁸ This automated method detected all changes in bone marrow cellularity that were detected by qualitative histologic assessment and, in some cases, the automated method found changes at lower dose levels that were not detected with routine histologic assessment by an anatomic pathologist. ²⁸ This method was primarily tested in rat sternal bone marrow, although it also was applicable to other species (mice, dogs, and monkeys) and other sites, including the femur and tibia. ²⁸ Subsequent development of the algorithm by the same group expanded its application to quantification of M:E and lymphoid ratios, as well as megakaryocyte cell density. ²⁹ This method has promise as a rapid quantitative screening procedure similar to flow cytometry (discussed below), but without the need to use specialized collection techniques or additional samples beyond those already routinely available (eg, preexisting H&E-stained bone marrow slides). ^28,29

If toxic effects on the bone marrow are identified but incompletely characterized using CBC data and bone marrow histology, then additional techniques can be employed to provide additional information. These techniques include cytological evaluation of bone marrow smears, flow cytometry or Sysmex XT 2000iV analysis, and, less frequently, clonogenic assays, liquid cultures, and transmission electron microscopy (TEM). ^{4,30 –33} Ideally, bone marrow smears are made proactively at the time of necropsy. These smears, stained with various Romanowsky stains (modified Wright’s-Giemsa, Giemsa, or May-Grunwald Giemsa), are often not evaluated unless subsequent data suggest there would be value in doing so. ⁴ If indicated, these bone marrow smears can be cytologically evaluated by a clinical pathologist to quantitatively assess cell types to determine an M:E ratio and maturation index. ⁴ In addition, as bone marrow smears are the ideal sample for evaluation of individual cell morphology, these can be used to better characterize alterations in bone marrow cellularity observed in the routine evaluation of H&E-stained bone marrow, often providing additional information on the cell types involved (eg, lymphoid vs erythroid cells).

If evaluating cellular morphology is not useful or sufficient on its own, then flow cytometry or Sysmex XT 2000iV analysis can be utilized to provide some comparable information to a bone marrow smear cytological evaluation, and data can often be obtained faster and more precisely, as they are automated processes that evaluate a larger number of cells. ^{4,30 –32} Flow cytometry can classify hematopoietic cells into major categories such as neutrophilic granulocytic, erythrocytic, and lymphocytic lineages and can provide some information regarding maturation but not a complete classification of all cell types (eg, megakaryocytic and eosinophilic/basophilic granulocytic cell lineages cannot be characterized with standard flow cytometry). ⁴ Additionally, if the entire marrow of a long bone is processed from rodents, then a total nucleated cell count can be determined. ⁴ However, it should be noted that samples for flow cytometry cannot be preserved, eliminating the possibility of conducting retrospective analyses, and sometimes resulting in a need to conduct follow-up studies. ⁴ Similarly, the Sysmex XT 2000iV hematology analyzer (when coupled with magnetic bead isolation, purification and sequestration of bone marrow lymphoid cells, and applying customized gating) can produce an automated M:E ratio and a 5-part differential count (proliferating myeloid, maturing myeloid, proliferating erythroid, maturing erythroid, and lymphocytes) in rats, mice, dogs, and monkeys equivalent to those derived from cytological evaluation of bone marrow smears. ^{30 –32}

Samples can also be collected for TEM if there is a morphologic change that needs to be further classified or for clonogenic assays if an in vitro system is needed. ⁴ If there are issues with platelets, liquid culture may be needed for in vitro studies of megakaryocytes. ³³ Although these advanced options are not routinely used for the evaluation of bone marrow in nonclinical toxicity studies, they can be useful in the investigation of the mechanism of bone marrow toxicity as described in the publication by Reagan et al. ⁴

Summary of Bone Marrow Evaluation in Nonclinical Toxicity Studies

Toxicologic pathologists evaluating nonclinical toxicity studies play a critical role in the identification of bone marrow toxicity. As such, general knowledge of bone marrow morphology, relevant anatomical and species differences, and suggested terminology for bone marrow findings is essential. The systematic histologic review of the bone marrow is facilitated by a concurrent review of CBC data, which includes age and sex-matched controls. Despite adherence to this standard approach, there are important limitations to the histological evaluation of the bone marrow. Bone marrow smears are the ideal sample for cytological evaluation of individual cell morphology (eg, to distinguish lymphoid vs erythroid cells), can be used to better characterize alterations in bone marrow cellularity (eg, in cases of inflammation, increased or decreased cellularity, dyshematopoiesis, and hematopoietic neoplasia) observed in the routine evaluation of H&E-stained bone marrow, and to quantitively assess cell types to determine a M:E ratio and maturation index. Additional options to detect or further characterize bone marrow toxicity include flow cytometry, Sysmex XT 2000iV or image analysis, in vitro assays, and TEM.