• Sign In to gain access to subscriptions and/or My Tools.
Impact Factor:2.137 | Ranking:Pathology 32 out of 76 | Toxicology 51 out of 88
Source:2014 Journal Citation Reports® (Thomson Reuters, 2015)

Histopathology of Bone Marrow

  1. Gregory S. Travlos
  1. Laboratory of Experimental Pathology, National Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park, North Carolina 27709, USA
  1. Address correspondence to: Gregory S. Travlos, Laboratory of Experimental Pathology, NIEHS, NIH, 111 Alexander Drive, MD B3-06, Research Triangle Park, NC 27709, USA; e-mail: travlos{at}niehs.nih.gov


As a major hematopoietic and lymphoid organ, morphological evaluation of the bone marrow is an important component of toxicity or safety assessment studies. While definitive characterization of bone marrow lesions often requires cytological aspirates or smears, assessment of histological bone marrow sections provides information regarding tissue architecture and hematopoietic status that is relevant for the detection of direct or indirect responses to chemical exposure. A variety of lesions have been observed in bone marrow. For example, lesions involving disturbances in growth, degenerative changes, inflammatory changes and neoplasia have been described. Lesions identified in hematoxylin and eosin-stained sections typically represent changes in the hematopoietic cell lineage and/or stromal cells since definitive identification of lymphoid cells is difficult except in cases of lymphoma. This review provides a descriptive and pictorial representation of a wide range of bone marrow lesions. Since large animal-to-animal variation may exist and there can be collection site- and age-related differences, it is imperative that the pathologist reviews all potential treatment-related findings against appropriate concurrent controls.


The bone marrow is one of the largest organs in the body (Lund, 2000). It is the major hematopoietic organ and is important as both a primary and secondary lymphoid organ (Picker and Siegelman, 1999). Since the hematopoietic system is a potential target organ of chemical exposure, evaluation of the blood and bone marrow is an important component of toxicity or safety assessment studies. As part of the hematopoietic system evaluation, morphological assessment of bone marrow tissue sections provides information about bone marrow tissue architecture (e.g., cellularity, cell lineages, vascular or stromal alterations, inflammation, necrosis), estimation of iron stores and identification of other features (e.g., pigment, infectious agents, proliferative or neoplastic disorders). Used in conjunction with a complete blood count (CBC), the histological examination of bone marrow provides information regarding the hematopoietic system that might otherwise be missed by examination of peripheral blood alone.

Responses to chemical exposure may occur by direct action of the chemical or its metabolites on hematopoietic tissues or indirectly through injury to other organ systems or metabolic pathways. And, bone marrow lesions reflecting these effects may be observed on paraffin-embedded sections. The present article offers an illustrative review of a variety of lesions that have been described for the bone marrow. It should be remembered, however, that the quality of, and, hence, interpretation of, bone marrow sections is affected by numerous variables related to specimen collection and processing (Travlos, 2006). Also, there may be wide animal-to-animal, variation and sampling site- and age-related differences (Cline and Maronpot, 1985; Weiss, 1986), requiring the pathologist to review potential treatment-related findings against an appropriate concurrent control group (Figures 1 and 2).

Figure 1

Bone marrow cellularity varies between animals and between specific sites in the same animal. Representative examples of normal femoral bone marrow cellularity in two male control F344 rats in a 2-year toxicity study (Figure 1A and 1B) and 2 male control B6C3F1 mice in a 28-day toxicity study (Figure 1C and 1D). Younger animals typically have a more densely cellular marrow.

Figure 2

Animal to animal variability in normal bone marrow cellularity should be considered when assessing bone marrow changes in treated rodents. Two examples from female F344 rats at the end of a 2-year study show differences in cellularity in the distal femur (Figure 2A and 2B). Similarly, differences in cellularity in the distal femur are also present in these 2 female control Sprague–Dawley rats from a chronic toxicity study (Figure 2C and 2D).

Disturbances in Growth

Increased Cell Numbers

Increases in nucleated cells in histological sections of bone marrow usually indicates a response to increased cell demand and is most evident in the hematopoietic cell lineage since definitive identification of lymphoid cells is not possible in H&E-stained preparations. Hematopoietic cell increases (also has been referred to as hypercellularity, hematopoietic cell hyperplasia or hematopoietic cell proliferation) are diagnosed by estimating the percentage of the marrow cavity containing fat versus hematopoietic cells and comparing these percentages to concurrent control animals (Figure 3). In rodents, if the hyperplasia is marked, hematopoietic cells may fill the marrow space, even extending through the nutrient foramina (MacKenzie and Eustis, 1990). In adult dogs, marrow hyperplasia should be considered when greater than 75% of the marrow space is occupied by hematopoietic cells (Weiss, 1986). Hematopoietic cell increases may involve all cell lines (panhyperplasia) or individual cell lines. When only one line is involved, the hematopoietic cell increase should indicate the specific lineage affected (e.g., erythroid or myeloid). Cell morphology is usually unaltered and the maturation sequence is synchronous; though, scattered atypical cells (e.g., binucleate rubricytes, giant band neutrophils, Howell-Jolly bodies) may be observed related to the increased production demand (Weiss, 1986; Valli et al., 2002). Depending on the cell line affected there may also be evidence of a shift in the myeloid/erythroid (M:E) ratio. For example, an increase in erythroid cells would decrease the estimated M:E ratio. Other cellular elements of the bone marrow such as mast cells and megakaryocytes can also be increased in number (MacKenzie and Eustis, 1990). Increases in mast cell numbers (mastocytosis) have, however, also been observed in the bone marrow sections of older rats (Frith et al., 1996). Emperipolesis of neutrophils within megakaryocytes has been observed in rats and mice and has been associated with administration of a growth factor or hematopoietic cell hyperplasia related to chronic blood loss or inflammation (Lee, 1989; Mezza et al., 1991). While histologic sections may provide evidence of the linage affected, smears or cytocentrifuge preparations of bone marrow are usually necessary for accurate determination of the M:E ratio and synchrony of maturation.

Figure 3

Increased bone marrow cellularity can be associated with treatment. There is a dramatic increased in cell density affecting erythyroid and myeloid cells as well as megakaryocytes in the femoral bone marrow of this F344 male exposed to methylene blue trihydrate for 2 years (Figure 3B and 3D) compared to the concurrent vehicle control (Figure 3A and 3C).

In most species, increases in erythropoietic cells (erythroid hyperplasia) (Figure 4) usually suggest a response to anemia (e.g., early blood loss or hemolytic anemia). There are increased marrow cells (hypercellularity) and synchrony of maturation is usually unaffected. Thus, while there may be an increase in rubriblast and prorubricyte numbers, rubricytes and metarubricytes predominate. Differential cell counts of bone marrow demonstrate a decrease in the M:E ratio due to the increased numbers of erythropoietic cells. In the dog, erythroid hyperplasia occurs in primary and secondary polycythemia. Bone marrow changes are similar for both conditions and may be characterized by hypercellularity with increases in erythroid and megakaryocytic cells; occassionally, increases in granulocytes also occurs (Peterson and Randolph, 1982). The bone marrow may have decreased hemosiderin content (Weiss, 1986). In rodents, erythroid hyperplasia may or may not be evident in histological sections of the bone marrow (Valli et al., 2002). Development of a hyperplasia depends on type, duration and severity of the anemia and the age of the animal. In rodents, particularly in the mouse, an increase in erythroid cells is often seen in the spleen, and possibly the liver, in response to an anemia (Figure 5).

Figure 4

Compared to the control femoral bone marrow (Figure 4A and 4C), an increase in erythropoiesis (Figure 4B and 4D) reflects a response to anemia and is typically characterized by a normal synchrony of erythyroid cell maturation. This erythryoid hyperplasia was present in a female F344 rat treated with 2-butanone oxime for 90-days.

Figure 5

Spleen from a male B6C3F1 mouse with a moderate degree (Figure 5 A and 5B) of extramedullary hematopoiesis. The red pulp contains irregular nests and patches of nucleated hematopoietic cells. A high magnification of marked splenic extramedullary hematopoiesis from a different B6C3F1 mouse (Figure 5C) shows both erythroid and myeloid cells as well as numerous megakaryocytes. An apoptotic cell is present (arrow) among the proliferating hematopoieic cells.

Increases in granulopoietic cells (myeloid hyperplasia) (Figure 6) are often related to an inflammatory response. Histologically, there appears to be an increase in the estimated M:E ratio and maturation synchrony is often unaffected; while immature forms increase, the preponderance of myeloid cells consists of metamyelocytic, band and segmented granulocytes (Weiss, 1986; Valli et al., 2002). With an acute inflammatory response, marrow segmented neutrophil numbers may be decreased due to enhanced release from the marrow storage pool (Weiss, 1986). Yet, a myeloid hyperplasia with a normal distribution of granulocytes may be observed in the marrow associated with a leukemoid reaction (Jain, 1986a). Due to increases in granulopoietic cells, quantitative bone marrow differential cell counts will demonstrate the increased M:E ratio.

Figure 6

Femoral marrow from a strain129 wild-type mouse (Figure 6A) contains a relatively equal proportion of erythyroid and myeloid cells. In contrast, Figure 6B is from a littermate genetically engineered to overexpress TGF-alpha and shows an increased number of myeloid cells. A large island of relatively pure myeloid cells is surrounded by arrows.

Increases in bone marrow megakaryopoietic cells (megakaryocyte hyperplasia) (Figure 7) are often related to conditions where there is increased consumption or destruction of platelets. Also, as mentioned above, megakaryocyte increases may occur in some responsive anemias. It has been reported that daily IP injections of polyethyleneglycol-rHuMGDF administered to Balb/c mice for 2 weeks, induced a megakaryocytic hyperplasia, myeloid hyperplasia and erythroid hypoplasia in the bone marrow (Ulich et al., 1996). Hyperploidy of the megakaryocytes may be observed in instances of megakaryocyte hyperplasia (Valli et al., 2002). While the M:E ratio does not change, an increased percentage of megakaryocytes may be detected in bone marrow differential cell counts.

Figure 7

Low and higher magnification of a femoral marrow characterized by megakaryocytic hyperplasia. In proportion to other cells types, there is a marked increase in the number of megakaryocytes. Section is from a female dog with severe thrombocytopenia.

Increases in bone marrow lymphoid cells would be difficult to identify from H&E-stained histological sections, except in cases of lymphoma. However, relative changes in lymphoid cells can be identified in Romanowsky-stained bone marrow aspirates and cytological smears.

Decreased Cell Numbers

Decreases in nucleated cells in bone marrow sections have been identified by several terms, such as, hematopoietic cell depletion, hypocellularity, hypoplasia and atrophy. And, hematopoietic cell decreases may involve all (panhypoplasia) or individual cell lines. In H&E-stained bone marrow sections, decreased hematopoietic cells is determined by estimating the percentage of hematopoietic cells contained within the marrow cavity and comparing these findings to the concurrent controls (Figures 8 and 9). Dog bone marrow may be considered hypocellular when less than 25% of the marrow space contains hematopoietic cells (Harvey, 1984). Bone marrow panhypoplasia has been characterized by a peripheral blood pancytopenia and replacement of the hematopoietic tissue with adipose in the marrow cavity (Weiss and Armstrong, 1984). Additionally, bone marrow sections that appear normo- or hyper-cellular, may be “hypoplastic” with respect to one cell line. For instances in which one cell line is affected, the hematopoietic cell decreases should indicate the lineage involved. When either the myeloid or erythroid cell lines is affected, the M:E ratio is altered. For example, the M:E ratio is increased with erythroid hypoplasia and decreased with myeloid hypoplasia. Also, there may be an asynchrony of maturation favoring the later stages of cell development. Again, while histologic sections may provide evidence of the overall affect, smears or cytocentrifuge preparations of bone marrow are usually necessary for accurate evaluation of the M:E ratio and synchrony of maturation (i.e., the maturation index). With selective decreases in megakaryopoietic cells (megakaryocyte hypoplasia), there may be an asynchrony of maturation favoring the early stages of cell development and a megakaryocytic hypoploidy; the M:E ratio is unaffected (Valli et al., 2002).

Figure 8

Sections from female F344 rats in a 2-methoxyethanol study. Compared to the concurrent control (Figure 8A and 8C), there is a mild decrease in bone marrow hematopoietic cellular density in a treated female (Figure 8B and 8D). Photomicrographs from control and treated rats are at identical low and higher magnifications.

Figure 9

Severe bone marrow cellular depletion (Figure 9B and 9D) in a B6C3F1 male mouse treated with AZT/methadone for 90 days. The few remaining cells consist of erythroid and myeloid precursors, lymphocytes, and macrophages. A concurrent age-matched control (Figure 9A and 9C) has a highly cellular bone marrow characteristic of what is typically seen in young mice.

Decreases in bone marrow erythropoietic cells related to erythroid hypoplasia/suppression are often reflected in a non-or poorly-regenerative anemia observed from the CBC. Decreases in erythropoietic cells within the bone marrow have occurred and may suggest a direct (e.g., estrogen, chloramphenicol, antiviral agents, chemotherapeutics) or indirect (e.g., tumor necrosis factor, cupric sulfate) treatment–related effect (Manyan et al., 1972; Bloom and Lewis, 1990; NTP, 1993, 1999; Andrews, 1998; Gossett, 2000; Lund, 2000). Additionally, a variety of non-bone marrow-centered conditions, (e.g., chronic inflammation, neoplasia, inanition, hypothyroidism, hypoadrenocorticism, chronic renal or hepatic disease), may result in suppressed erythroid production (Jain, 1986b; Meierhenry, 1990; Waner and Harrus, 2000). Morphologically, bone marrow sections may appear hypocellular, the M:E ratio may be increased and there may be a shift in the maturation sequence to the later stages of cell development (Figure 10). In short-term studies, routine histologic sections of bone marrow may show minimal morphological changes (MacKenzie and Eustis, 1990).

Figure 10

Low, medium and high magnifications of bone marrow from a 14-year-old spayed female dog with an increased M:E ratio. A reasonably cellular marrow with very few recognizable erythroid cells is present. The marrow is dominated by immature and mature myeloid cells, including eosinophils and a normal complement of megakaryocytes (Figure 10A and 10B). A macrophage containing hemosiderin (arrow) is present in Figure 10C.

In anemia related to chronic inflammation, the cellular pattern and erythroid morphology of the marrow is often normal, but there are increases in stainable iron (Feldman and Kaneko, 1981; Feldman et al., 1981; Jain, 1986b; Waner and Harrus, 2000). Bone marrow macrophage iron content (hemosiderin) is increased and is observed as a golden-brown pigment with H&E stain (Figure 11) or as blue granules with a Prussian blue stain. This increase in marrow iron content can be used to differentiate the anemia of chronic inflammation from iron deficiency anemia, in which the marrow has decreased iron content (Harvey, 2000). In anemia of chronic renal disease, the marrow is may be normo- to hypo-cellular (Jain, 1986b) and, though marrow iron is present, it is not increased as in anemia of chronic inflammation (Weiss, 1986).

Figure 11

Bone marrow from a male B6C3F1 mouse treated with a high dose of phenophthalene for 2 years. Numerous hemosiderin-ladened macrophages (arrows) are present.

Drug- or chemical-induced myelopoietic cell decreases appear to occur less frequently than erythroid-related hypoplasia (Weiss, 1986). But, they have occurred with a variety of compounds including chemotherapeutics, antibiotics (e.g., chloramphenicol, cephalosporins) or antihypertensives and phenylbutazone, phenobarital and griseofulvin (Martin et al., 1985; Jain, 1986c; Weiss, 1993; Lund, 2000; Moore and Bender, 2000). Acquired thrombocytopenias related to suppressed bone marrow production have reported for a variety of drugs (e.g., chemotherapeutics, antibiotics, antiinflammatories, anticonvulsants, estrogens) (Feldman et al., 1988; Grindem et al., 1991; Ulich et al., 1995; Zimmerman, 2000). Innanition/diet restriction causes a decrease all hematopoietic cells and the cellularity of the bone marrow with an apparent increase in marrow fat; the M:E ratio appears unaffected. For example, diet restriction sufficient to stop weight gain in young rats caused a 50%, 40% and 20% decrease in erythroid, myeloid and megakaryocytic precursors, respectively (Brown, 1954). In another study, mild (75% of control) to moderate (50% of control) diet restriction for 2 weeks caused a decrease in hematopoietic tissue with an apparent increase in marrow fat located centrally in the cavity; there was no apparent change in appearance or distribution of the hematopoietic tissue remaining (Levin et al., 1993). These diet restriction studies did not provide information relative to changes in lymphoid lineage cells.

Severe decreases in hematopoietic cells involving all cell lines have been termed bone marrow aplasia, aplastic anemia or aplastic pancytopenia; drug-related decreases in all cell lines have been reported (Weiss, 2000a). This condition is characterized by decreased numbers of all hematopoietic cell lines in the bone marrow and a pancytopenia in the blood (Weiss, 2000a). Bone marrow pancytopenia should not be confused with conditions such as, pure red cell aplasia (a condition where only erythrocyte production is suppressed), myelophthistic disorders (e.g., myelofibrosis and leukemias), or myelodysplasias (where the marrow appears normo- to hyper-cellular). In H&E sections, aplastic marrow appears devoid of hematopoietic cells, consisting primarily of vascular sinuses and adipose tissue (Figure 12); lymphocytes, plasma cells, macrophages and mast cells may be observed scattered about (Weiss and Armstrong, 1984; Weiss, 1986). In less severe cases, small, scattered islands of hematopoietic cells may be found, but, in general, occupy less than 25% of the marrow cavity (Weiss, 1986, 2000a). In contrast, pure red cell aplasia demonstrates as a severe decrease in erythroid precursors, but the overall marrow cellularity may not be dramatically altered, as the production of granulocytes and platelets is unaffected (and may be increased); the M:E ratio is markedly increased and hemosiderin appears increased (Weiss, 1986, 2000b).

Figure 12

This aplastic bone marrow from a female dog is largely devoid of hematopoietic cells. Marrow spaces consist primarily of adipose tissue and vascular sinuses. Marrow cellularity is less than 5% of normal. A few lymphocytes and erythroid precursors remain in this aplastic marrow.

In the rat, the term “atrophy” has been applied to lesions in which there is an apparent decrease in marrow cellularity (MacKenzie and Eustis, 1990; Frith et al., 2000). This has been described as an uncommon, but spontaneous, lesion occurring in adult rats with an apparent predilection for females. And, it appears to be of no clinical importance (MacKenzie and Eustis, 1990). Atrophy is characterized as a focal or multi-focal lesion consisting of well-demarcated areas with reduced hematopoietic cells, increased (Firth et al., 2000) or decreased (MacKenzie and Eustis, 1990) fat cells and a prominent reticular stroma (Figure 13). Since, in some instances, there is an increase in macrophage numbers, it has been suggested that there may be some relationship between this lesion and focal granulammatous inflammation (MacKenzie and Eustis, 1990). Diffuse atrophy has been observed in old (particularly moribund) rats or in young rats that had treatment-associated, decreased weight or weight gain (Figure 14) (MacKenzie and Eustis, 1990; Meierhenry, 1990; Frith et al., 2000).

Figure 13

Two areas of focal bone marrow atrophy (Figure 13A, arrows) in the femur of a 2-year old control female F344 rat appear as regions of lower cellular density. At higher magnifications these lesions (Figure 13B and 13C) are characterized by a prominent reticular stroma.

Figure 14

Regions of decreased cellularity are scattered throughout the bone marrow (Figure 14A, arrows) in the male F344 treated with a low dose of allyl isovalerate for 2 years. At higher magnification (Figure 14B) these hypocellular areas contain increased reticular stroma and adipocytes.

Hematopoietic Cell Dysplasia

Hematopoietic cell dysplasias of the marrow have been related to chemical treatment (e.g., 3′-azido-3′-deoxythymidine, 2′, 3′-dideoxycytidine) and nutritional deficiencies (e.g., B12/folate) and may involve individual (e.g., dyserythropoiesis, dysgranulopoiesis or dysthrombopoiesis) or combinations of cell lineage (Weiss, 1986; Riley et al., 1992; NTP, 1999; Lund, 2000; Watson and Canfield, 2000; Valli et al., 2002). These are typically identified in cytological preparations. Dyserythropoiesis can be demonstrated by the presence of erythropoietic cell multinuclearity or nuclear fragmentation or megaloblastosis or ringed sideroblasts. Characteristics of dysgranulopoiesis can include giant forms, nuclear hyposegmentation or hypersegmentation, bizarre nuclear shapes, and abnormal number, size, or tinctorial properties of primary granules. Dysthrombopoiesis may consist of either small or large megakaryocytes, increased ploidy, or the presence of small, dispersed, or fragmented nuclei.

The term “myelodysplastic syndrome” describes a group of related disorders characterized by one or more cytopenias in the blood, dysplastic changes in marrow hematopoietic cells, and a propencity to progress to acute myeloid leukemia (Blue, 2000). It can be difficult to differentiate from some nutritional deficiencies (e.g., B12/folate, pyridoxine) and drug-induced dysplasias (e.g., lead, chloramphenicol). In the dog, myelodysplastic syndrome has been characterized in bone marrow sections as normocellular to hypercellular with an increased M:E ratio and dysplastic changes in all cell lines. Dysplastic features easily recognized in bone marrow aspirates (but less obvious in tissue sections) include changes such as, increased numbers of early erythroid and/or myeloid precursors, multinucleated erythropoietic cells or nuclear lobulation/fragmentation, erythroid megaloblastic change, sideroblasts, small or large hypolobulated megakaryocytes, hypo- or hyper-segmented granulocyte precursors and decreased number, size, and/or tinctorial properties of eosinophilic granules.

In chronic iron deficiency, a dyserythropoiesis in bone marrow sections is characterized by an increased cellularity (related to erythroid hyperplasia), M:E ratios of <1 and increased numbers of rubricytes and metarubricytes that have scanty cytoplasm, ragged cell borders and inconsistent cytoplasmic basophilia (Weiss, 1986; Watson and Canfield, 2000). Additionally, there is a lack of stainable iron and decreased numbers of sideroblasts in marrow sections of iron deficiency (Weiss, 1986). Some chemicals (e.g., lead) interfere with iron metabolism and/or heme/hemoglobin synthesis. Iron accumulates within the mitochondria and can be demonstrated with iron stains. Erythroid precursor cells, in which these ironcontaining mitochondria occur as a ring around the nucleus, are termed “ring sideroblasts” and the lesion is referred to as sideroblastic change (Jain, 1986d; MacKenzie and Eustis, 1990).

Alterations in Bone Marrow Stromal Cells

A variety of histologically similar and uncommon changes appear to involve primarily bone marrow stromal cells, sometimes with associated changes in hematopoietic cells and bone. These conditions are diagnostically challenging and typically involve some form of stromal cell proliferation or myelosclerosis and sometimes resemble histiocytic sarcoma. Diagnoses for these lesions include stromal cell hyperplasia, myelostromal proliferation, myelofibrosis, fibrous osteodystrophy, and fibro-osseous lesions. The myelodysplastic syndrome described here could also be considered in this category of bone marrow lesions. Since adipocytes can be considered part of the stromal cell population, focal lipomatosis can also be considered as a stromal cell alteration.

Focal Stromal Cell Hyperplasia

As a discrete entity, focal stromal cell hyperplasia has been reported as being occasionally observed in the F344 rat (MacKenzie and Eustis, 1990). However, it has been characterized as a small lesion that may be difficult to distinguish from focal atrophy. Additionally, it has been described that in some instances, the lesion is more prominent, containing cells with pale vacuolated cytoplasm and round vesicular nuclei and a single nucleolus (Figure 15) (MacKenzie and Eustis, 1990). Though the description of the nuclear shape is different, this lesion is histologically similar to myelostromal proliferation.

Figure 15

Focal stromal cell hyperplasia (Figure 15A and 15B, arrow) is a relatively uncommon lesion that is similar to focal bone marrow atrophy but is typically more prominent, consisting of cells with pale vacuolated cytoplasm and round vesicular nuclei (Figure 15C). These photomicrographs are from a vehicle control female F344 rat in a 2-year study.

Myelostromal Proliferation

Myelostromal proliferation is a term that has been used to classify a rare, poorly understood, proliferative lesion in the F344 rat (MacKenzie and Eustis, 1990). It has been described as a diffuse proliferation of cells of unknown origin (possibly stromal reticular cells or histiocytes) (Figures 16A and C). The proliferation is characterized by a generally uniform population of cells with indistinct boundaries, abundant finely vacuolated cytoplasm sometimes containing ironpositive inclusions, round to oval or slightly irregular vesicular nuclei, and single prominent nucleoli. There is a delicate network of silver-stain positive, but trichrome negative, stroma surrounding individual cells (MacKenzie and Eustis, 1990). Occasional multinucleate cells may be found but there are few mitotic figures and little or no cellular atypia observed (MacKenzie and Eustis, 1990). It appears this proliferation is limited to the marrow cavity, but may be multicentric, involving axial or long bones. Though this lesion has not been thoroughly characterized, it has been suggested that it should not be confused with histiocytic sarcoma involving the bone marrow (MacKenzie and Eustis, 1990). Interestingly, several experienced pathologists reviewed the lesion shown in Figure 15 and there were differences of opinion as to whether this represented a myelostromal proliferation or a histiocytic sarcoma. In general, histiocytic sarcoma (Figures 16B and D) has more cytoplasmic and nuclear pleomorphism, less cytoplasmic vacuolization, abundant multi-nucleated giant cells and multi-organ involvement (e.g., liver, spleen, lung) (Ogasawara et al., 1993).

Figure 16

Myelostromal proliferation in this bone marrow from a control male F344 rat is characterized as a diffuse proliferation of cells presumed to be stromal reticular cells (Figure 16A and 16C) that have displaced normal hematopoietic cells. This rare and poorly understood lesion has been distinguished from histiocytic sarcoma (Figure 16B and 16D) in which the neoplastic cells have more cytoplastic and nuclear pleomorphism, have less cytoplasic vacuolization, often have multinucleated cells (not present in this example), and typically have multi-organ involvement.

Focal Lipomatosis

Focal lipomatosis is a lesion of adipose tissue. It has been reported in the F344 rat and is characterized as a reasonably well-delineated focus of mature adipocytes within the marrow cavity (MacKenzie and Eustis, 1990). The focus of adipocytes is usually located in the marrow cavity center with hematopoietic tissue found peripherally near the cortical bone or near the epiphyses. Hematopoietic cells may also be found scattered among the adipocytes. It may be confused with the normal distribution of marrow fat in adult rats. Considered a spontaneous lesion, its significance is unknown.


Bone marrow fibrosis is characterized by an increase in stromal collagen and actively proliferating fibroblasts or fibrocytes with decreased hematopoietic cells (Figure 17) and must be differentiated from atrophy, stromal hyperplasia and fibrous osteodystrophy (Weiss, 1986; MacKenzie and Eustis, 1990; Frith et al., 2000). Focal fibrosis has been observed in young rats and may be secondarily related to instances of marrow injury, inflammation, or necrosis (Frith et al., 2000).

Figure 17

Fibrosis. In this hypocellular bone marrow from a control F344 rat, much of the marrow space is occupied by early fibrous tissue proliferation. More distinct collagen is evident along bony surfaces (Figure 18B, arrow) while elsewhere the connective tissue reaction is loose and almost myxomatous.

In laboratory species, the term myelofibrosis has been used to describe fibroproliferative lesions of the bone marrow. For example, the term myelofibrosis has been used for fibrotic lesions of the bone marrow in the dog (Reagan, 1993; Blue, 2000) and in laboratory rodents that have been exposed to whole-body radiation or that have been infected with murine leukemia virus (MacKenzie and Eustis, 1990). In humans, myelofibrosis indicates a specific myeloproliferative disorder of uncertain etiology that has hematological and immunological abnormalities and is characterized by bone marrow failure and fibrosis. In the dog, myelofibrosis (Figure 18) appears to be a fibroproliferative response secondary to variety of causes and has been associated with pyruvate kinase deficiency, gamma radiation, drugs, infectious agents and malignancies (with and without bone marrow involvement); no primary form has been described (Reagan, 1993). In some instances, there appears to be similarities between the human condition and those described for the dog (e.g., nonresponsive anemia) (Reagan, 1993). However, since similar lesions in the rat have not been associated with hematological and immunological abnormalities (MacKenzie and Eustis, 1990), it is preferable to avoid use of the diagnostic term myelofibrosis and simply diagnose this lesion as fibrosis when observed in rats.

Figure 18

In this case of myelofibrosis in an 8-year-old spayed female dog, the marrow spaces have been filled by mature collagen with a few remaining irregular collections of hematopoietic cells (Figure 18A and 18B, arrows). A dense region of mature collagen surrounding trabecular bone is shown in Figure 18D. This dog was anemic and had extramedullary hematopoiesis in the liver and spleen.

Fibro-Osseous Lesion and Fibrous Osteodystrophy

Though not primarily bone marrow lesions per se, fibro-osseous lesion (FOL) and fibrous osteodystrophy (FOD) are fibroproliferative lesions of bone that, as they progress, invade and crowd the marrow cavity. In the mouse, a specific, spontaneous bone lesion may occur which has been termed FOL; for a detailed description see Long and Leininger (1999). Briefly, it is a lesion of bone that impinges on the marrow space but, apparently, has no hematological implications. In the past, this lesion has been called a variety of names including, FOD, hyperostosis, myelofibrosis, osteoporosis, osteofibrosis, and osteosclerosis. FOL now, however, is classified as a separate entity. Since there is an accelerated osteoclastic bone resorption coupled with a proliferation of fibrovascular stroma, FOL resembles FOD histologically. Subsequent to the bone resorption there is a proliferation of osteoblasts that leads to bone formation/repair.

The proliferation of new bone leads to formation of thickened trabeculae that encroaches on the center of the marrow cavity and, eventually, may consist of a network of dense bony trabeculae separated by a fibrovascular stroma (Figure 19). As the mouse ages, the lesion can occupy a substantial portion (up to two-thirds) of the marrow cavity space. FOL can occur in all bones, but has been reported most frequently in the sternebrae, long bones, and vertebrae. The lesions have been associated with increased serum alkaline phosphatase activity but not primary or secondary hyperparathyroidism. It is a common spontaneous lesion of female B6C3F1 mice and may affect up to 100% of aged females in a study; male B6C3F1 mice are much less affected. It has been observed in female mice as early as 32 weeks of age. The etiology of FOL is unknown, but the predilection for females suggests altered sex hormone status may be involved. Though this is a spontaneous lesion of B6C3F1 mice, similar lesions have been reported in other mouse strains treated with estrogens and a prostaglandin E1 analogue.

Figure 19

Fibro-osseous lesion is a specific bone lesion of mice characterized by enhanced osteoclastic activity, formation of thickened bony trabeculae (Figure 19C and 19D), and progressive encroachment of bone marrow spaces. It is a spontaneous lesion, particularly in female B6C3F1 mice. This case was present in a female treated with sodium fluoride for 2 years.

Fibrous osteodystrophy may be a drug-induced lesion, with a fibro-osseous proliferation pattern similar of FOL and has been observed in rats administered chemicals or anticancer drugs (NTP, 1986; Courtney et al., 1991) (Figure 20). In their report, Courtney et al. (1991) characterized the lesion as an accelerated proliferation of trabecular bone and foci or bands of proliferating spindle-cells (with or without osteoid formation) that encroached on the center of the marrow cavity. When the lesion occurred, the remaining bone marrow appeared either normo- to hypo-cellular. Osteoblasts were readily apparent lining the immature woven bone.

Figure 20

Fibrous osteodystrophy in rats is similar to fibro-osseous lesion in mice. This case is from a low dose male F344 rat in a 2-year study of xylene. There is accelerated proliferation of trabecular bone and proliferating spindle cells that fill marrow spaces.

Degenerative Changes


Bone marrow degeneration (e.g., necrosis) has been associated with compound administration and must be differentiated from postmortem change (autolysis) or processing-related artifact (Figure 21) (MacKenzie, 1990). The bone marrow may be prone to autolytic change related to release of the contents from enzymes-rich granules of granulocytes and poor, or slow, penetration of the marrow by the fixative due to its encasement in bone. It has been indicated that megakaryocytes are the first cells to show postmortem changes and an early recognizable feature is the condensation (pyknosis) of the megakaryocytic nucleus (MacKenzie and Eustis, 1990).

Figure 21

Postmortem autolysis. In this mid-dose female F344 rat from a 2-year study of formamide, the bone marrow is filled by degenerating cells with pyknotic nuclei and indistinct granular cytoplasm. This rat had mononuclear cell leukemia and the majority of autolytic cells in the bone marrow spaces are neoplastic large granular lymphocytes.

Infarction of the bone marrow can occur due to interruption of the blood supply and has been associated with conditions such as thrombosis (e.g., disseminated intravascular coagulation) or vascular occlusion related to malignancy (Frith et al., 2000; Ghanayem et al., 2001; Ezov et al., 2002). The infarction results in marrow necrosis due to release of enzymes from the granulocytic granules. The lesion may be focal, multifocal or diffuse and is characterized by hematopoietic cell cytoplasmic vacuolization, nuclear pyknosis, karyorrhexis, and karyolysis (Figure 22) (MacKenzie and Eustis, 1990; Frith et al., 2000). In the affected marrow, areas of cell loss are replaced by amorphous granular eosinophilic debris and should not be confused with fibrin deposition or edema fluid; a fibrin stain may be helpful in this differentiation (Feldman et al., 1981; Weiss, 1986). With time, phagocytic macrophages increase in or near the areas of necrosis. In acute situations, the necrotic area may become filled with intact red blood cells (hemorrhage) due to injury to vessels (Levin et al., 1993; Frith et al., 2000). Hemosiderin pigment may be found in macrophages in the bone marrow associated with old hemorrhage and has been referred to as hemosiderosis; an iron stain (e.g., Prussian blue) can be used to demonstrate the presence of hemosiderin pigment. Bone marrow hemorrhage may also be observed with vascular injury/hemorrhagic diathasis unrelated to infarct/necrosis and is characterized by lakes of intact red blood cells interspersed amongst the adipose and hematopoietic tissue (Figure 23). Vascular sinus dilatation, or ectasia, has been reported in the dog (Weiss, 1986) and, if there has been a severe loss of hematopoietic tissue in the bone marrow (e.g., myelotoxicity) it has been observed in the rat (Valli et al., 2002). Bone marrow cellular loss secondary to myelotoxicity can also occur in mice (Figure 24).

Figure 22

Bone marrow necrosis secondary to a thrombus (Figure 22A and 22B, asterisk) in the central artery of a rodent is characterized by loss of hematopoietic cells, increase in reticular stroma (Figure 22C and 22D) and formation of bony trabeculae (Figure 22A, arrows). A second example of bone marrow necrosis (Figure 22E and 22F) is seen in this low dose male F344 rat from a 2-year mercaptobenzothiazole study. The cause of the necrosis was not obvious. There is decreased hematopoietic cell density and prominent bands of reticular stroma (Figure 22F).

Figure 23

In this example of bone marrow hemorrhage in an 8-year-old spayed female dog, lakes of intact mature erythrocytes are interspersed between adipocytes and hematopoietic cells. The dog presented with acute systemic hemorrhagic processes, vasculitis, and vascular necrosis.

Figure 24

Bone marrow cell loss secondary to myelotoxicity is present in this female B6C3F1 mouse treated with urethane for 90 days (Figure 24B and 24D). Dilated vascular sinuses filled with erythrocytes are interspersed among adipocytes in the marrow spaces. Comparable magnifications of an age- and sex-matched concurrent control is present in Figure 24A and 24C.

In rats, marrow necrosis, involving the hematopoietic and stromal tissues, has been associated with severe (25% of control) diet restriction for 2 weeks (Levin et al., 1993). In that study, stromal cells appeared vacuolated or necrotic, there was necrosis of hematopoietic cells, remaining hematopoietic tissue was <25% of the control group (consisting primarily of neutrophils, many appearing toxic), there were focal areas of hemorrhage, and emperipolesis of neutrophils within megakaryocytes was readily observed.

Bone marrow degeneration is a term that has been used to describe a common nonspecific change in severely debilitated animals (MacKenzie, 1990). With cachexia the marrow reveals a decreased cellularity. Also, there may be a loss of fat cells in the bone marrow with the remaining space replaced with serous fluid (serous atrophy of fat) (Figure 25). Serous atrophy of fat has not been described in the dog (Weiss, 1986). Focal degeneration, or atrophy, has been occasionally reported as a spontaneous lesion in older rats and mice (MacKenzie, 1990).

Figure 25

Severe bone marrow degeneration with serous atrophy of fat is present in this dog. The marrow is characterized by a loss of hematopoietic tissue (decreased cellularity) and presence of a homogeneous, eosinophilic, acellular material (serous atrophy) (Figure 25B and 25C, arrow). Photomicrographs courtesy of Drs. Han Harleman and Kathryn Bowenkamp.

Inflammatory Changes

In general, Inflammatory lesions of rodent and dog bone marrow are rare (Weiss, 1986; MacKenzie and Eustis, 1990; Frith et al., 2000). Focal granulomatous lesions, characterized by accumulations of macrophages (sometimes with phagocytized detritus or residual bodies in the cytoplasm), that have oval, elongated or fusiform nuclei and abundant pale eosinophilic cytoplasm and can be associated/separated by dense aggregates of mononuclear cells (Figure 26) have been reported young and aged adult rats; they appear to be more common in female than in male rats (MacKenzie and Eustis, 1990; Frith et al., 2000). Chemicals have induced granulomatous inflammatory lesions; acute inflammation of the marrow has been observed with sepsis (Frith et al., 2000). Acute inflammatory lesions of the bone marrow have been described for the dog (Weiss, 1986). They consist of a fibrinous exudate that may or may not have associated inflammatory cells. Lesions with neutrophils have been termed “acute myelitis”; lesions without neutrophils—“fibrinous myelitis.” Bone marrow granulomas have not been described in the dog (Weiss, 1986).

Figure 26

Focal granulomatous inflammation characterized by aggregates of histiocytes with phagocytized residual bodies (arrows) are present in this bone marrow from an untreated female F344 rat in a 2-year study.

A spectrum of inflammatory changes involving erythrophagocytosis as well as lymphocytes and plasma cells has been reported. Erythrophagocytosis in the bone marrow (Figure 27) has been observed with immune-mediated anemias (Jain, 1986e). Aggregates of lymphocytes may be observed in marrow sections of healthy dogs (Weiss, 1986). Increases in plasma cells, however, suggest chronic immune stimulation or myelomas (Jain, 1986f; Weiss, 1986). Immunoglobulin-engorged plasma cells are characterized by large pinkish to bluish granules (Russell bodies) and have been termed “Mott cells” (Jain, 1986f). “Flaming” plasma cells are characterized by a reddish cytoplasm located peripherally; they have been observed in dogs with IgA myeloma (Zinkl et al., 1983). It has been suggested that lymphoid cell aggregates or lymphoid follicles (with well developed germinal centers) may be observed in rat bone marrow, but are not usually found in untreated animals (Frith et al., 2000). However, Geldof et al. (1983), reported that neither the rat nor mouse bone marrow had, or developed, lymphoid-cell aggregates or structures resembling follicles, even after immunization.

Figure 27

Sinus ectasia and erythrophagocytosis are present in this femoral bone marrow from a female B6C3F1 mouse treated with methylene blue for 90 days. The phagocytized erythrocytes are most clearly evident in the highest magnification view (Figure 27C).


Neoplasia of the bone marrow may be divided into primary or secondary conditions. Primary neoplasias arise in the bone marrow and could involve hematopoietic cells (e.g., myeloid — Figure 28, lymphoid — Figure 29, mast cell —Figure 30), stromal cells (e.g., fibrosarcoma, histiocytic sarcoma — Figure 31) or endothelial cells of the vessels (e.g., hemangiosarcoma — Figure 32). Secondary neoplasias could develop from metastasis of a tumor from a distant site (e.g., pheochromocytoma — Figure 33) or locally invasive tumors (e.g., osteosarcoma — Figure 34, chordoma — Figure 35). The histological findings of malignancy in the marrow would be substantially similar to those in other organs (MacKenzie and Eustis, 1990; Frith et al., 1996). In advanced cases, the normal architecture of the marrow can be effaced and the hematopoietic/adipose tissue replaced by a monotypic population tumor cells.

Figure 28

In this example of granulocytic leukemia, the highly cellular femoral bone marrow is filled almost exclusively with leukemic cells. This case was in a nontreated p53+/ mouse in a 90-day study.

Figure 29

Lymphocytic lymphoma cells have replaced the hematopoietic cells in the sternal bone marrow and extended outside the marrow cavity in this p16+/ male mouse treated with benzene in a 90-day study. The neoplastic cells have infiltrated into adjacent muscle (Figure 29A, arrow) and periosteal tissue (Figure 29C, right side).

Figure 30

Neoplastic proliferation of mast cells occupy large areas of the femoral marrow in this male B6C3F1 mouse treated with a high dose of dibromoacetonitrile for 2 years. A toluidine blue-stained section (Figure 30C) clearly shows the granules in the neoplastic mast cells cells.

Figure 31

Examples of histiocytic sarcomas in a treated F344 rat from a 2-year 1-chloro-2-propanol study (Figure 31A and 31B), in the femur of a treated B6C3F1 mouse in a 2-year benzophenone study (Figure 31C and 31D), and in a 3-year-old spayed female dog (Figure 31E and 31F). While the neoplastic cells are relatively uniform in the rat (Figure 31A and 31B), multinucleated cells are present in the mouse (Figure 31C and 31D) and dog (Figure 31E and 31F).

Figure 32

This hemangiosarcoma was present in the femoral marrow of a female B6C3F1 mouse treated with a high dose of methalolacrylimide for 2 years. While much of this neoplasm consists of cavernous spaces filled with erythrocytes typically characteristic of hemangioma, the diagnosis of hemangiosarcoma was base on the presence of aggregates of neoplastic endothelial cells (arrows).

Figure 33

The femoral marrow spaces of this male F344 treated with a high dose of tetracycline for 2 years have been invaded by closely packed nests of neoplastic cells with an endocrine growth pattern. The primary tumor was an adrenal pheochromocytoma.

Figure 34

An invasive osteogenic sarcoma originated in the proximal tibia of this female B6C3F1 mouse treated with a high dose of formamide for 2 years. Neoplastic cells forming osteoid are present (Figure 34B).

Figure 35

Chordomas are rare neoplasms occasionally seen in 2-year rodent carcinogenicity studies. This example was present in the vertebral bone of a female F344 rat treated with a high dose of methylene blue trihydrate.

Characterization of the neoplastic cells would depend on the individual tumor type. Occasionally, bone marrow infarction and necrosis may be observed related to vascular occlusion by the neoplasm. In F344 rats, it has been reported that primary neoplasia of the hematopoietic tissue of the bone marrow is very rare and that neoplasms of the vasculature or connective tissues are occasionally found (MacKenzie and Eustis, 1990). In fact, based on the National Toxicology Program (NTP) historical database for all routes of exposure and vehicles, there are no reported incidence for primary neoplasias of the bone marrow. This report included incidence rates for diagnoses such as lymphoma or mononuclear cell leukemia as multisystemic tumors, affecting more organs than just the bone marrow (NTP, 2005a). Ogasawara et al. (1993) presented data suggesting that, while histiocytic sarcoma in the rat was considered to originate at sites other than bone marrow, the marrow appeared to be a primary site of origin in Donryu (~4.5% incidence rate) and F344 (~1.5% incidence rate) rats. In mouse bone marrow, hemangiosarcoma was reported at a rate of approximately 1% for male and 0.5% for female B6C3F1 strain (NTP, 2005b). Additionally, in male mice benign and malignant mast cell tumors occured in the bone marrow at the rate of 0.04% and 0.24%, respectively; females had an incidence rate of 0.04% for malignant mast cell tumors.


Histological examination of bone marrow provides important information regarding the hematopoietic system and effects related to drug or chemical toxicity. A variety of terms have been used to identify certain lesions. To improve consistency in evaluations, however, there has been a recent interest by regulatory and professional organizations to standardize diagnostic terms, emphasizing the use of descriptive, rather than interpretive, terminology (Kuper et al., 2000; Haley et al., 2005). Since there can be wide animal-to-animal, sampling site- and age-related variation, the pathologist must review all potential treatment-related findings against an appropriate concurrent control group and in light of potentially related lesions in other tissues. It should be remembered that the quality of, and, hence, interpretation of, bone marrow sections can only be as good as the specimen collected. Thus, attention must be paid to bone marrow collection and processing techniques.


The author thanks Robert Maronpot for invaluable support during manuscript preparation. Special thanks to Dr. Carol B. Grindem and Sandra Horton from the North Carolina State University College of Veterinary Medicine for their assistance and access to materials used in the completion of this manuscript.

Article Notes

  • This research was supported by the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences.


| Table of Contents