Practical Murine Hematopathology: A Comparative Review and Implications for Research

Abstract

Hematologic parameters are important markers of disease in human and veterinary medicine. Biomedical research has benefited from mouse models that recapitulate such disease, thus expanding knowledge of pathogenetic mechanisms and investigative therapies that translate across species. Mice in health have many notable hematologic differences from humans and other veterinary species, including smaller erythrocytes, higher percentage of circulating reticulocytes or polychromasia, lower peripheral blood neutrophil and higher peripheral blood and bone marrow lymphocyte percentages, variable leukocyte morphologies, physiologic splenic hematopoiesis and iron storage, and more numerous and shorter-lived erythrocytes and platelets. For accurate and complete hematologic analyses of disease and response to investigative therapeutic interventions, these differences and the unique features of murine hematopathology must be understood. Here we review murine hematology and hematopathology for practical application to translational investigation.

Hematology is an important adjunct to both clinical medicine and biomedical research, with more than 1700 currently funded NIH projects¹⁰⁹ and more than 3400 research articles published over the past 5 years using mouse models.¹²⁰ There are now more than 6000 genetically engineered mouse (GEM) models of disease, with 500 new GEM created each year at the Jackson Laboratory alone, and several large projects are underway to thoroughly phenotype each new mutant mouse strain (https://www.komp.org/).^13,176 A mouse tumor database (http://tumor.informatics.jax.org/mtbwi/index.do) is available to provide information regarding mouse models of human cancer, and the Mouse Phenome Database at the Jackson Laboratory provides links to phenotypic data for many GEM models (http://phenome.jax.org/).⁸ The defined components to complete the phenotyping of GEM models have been recently reviewed.^13,157,176 In addition, 21 inbred strains of mice are commonly used for investigations into such topics as response to infectious and genetically induced disease and dietary and pharmacologic therapies. These commonly used laboratory mouse strains have, for example, inherent differences in immunology or iron trafficking, which can affect research outcomes.^16,47,137 These interstrain differences are important to recognize and understand as a component of effective study design and prior to strain selection for laboratory investigations, especially when hematologic responses to disease need to be considered.^13,16,137

For any appropriately designed experiment, concurrent age-, sex-, and strain-matched control mice must be included to accurately compare the effects of a disease, genetic manipulation or therapeutic intervention;^13,155 alternatively, individual mice can be used as their own controls in some studies. Several important guidelines exist to ensure that appropriate numbers of experimental and control mice are incorporated into a study design to maximize statistical power yet minimize waste.^{13,40-42,71,72,176} During and between studies, consistent blood collection methods are essential for accurate comparative analyses. Species-appropriate hematologic instrumentation and timely analysis of fresh blood are necessary to minimize preanalytic hematologic errors.^3,37,71 Especially important for mice and their restricted available blood volume are the use of practical, accurate, species-specific, and up-to-date hematologic methods.

Here we comprehensively review murine hematology and hematopathologic responses to disease in the context of biomedical research, discovery, and phenotyping studies. To maximize the opportunity for detecting phenotypes, disease, and responses to therapeutic interventions in mice, we focus on providing a practical summary of methods and analysis for accurate hematologic studies and on describing the morphologic assessment of mouse hematopathology in peripheral blood and bone marrow in ways that will be useful to those—veterinarians and researchers alike—who work with murine species.

Murine Hematology: General Considerations

Minimizing preanalytic variables.

Several factors impact the ability to obtain meaningful hematologic results from blood samples in mice with limited available blood volume. Frequency of blood collection, size and age of the mice, and available hematologic instrumentation all play a role in blood volume limitations and ultimately in the quality of the data generated. In particular, variability in fasting or anesthetic protocols or blood collection site can affect results.^142,155 For example, fasting protocols not standardized between studies may introduce preanalytic variation because mice consume less water while fasting than otherwise, potentially resulting in hemoconcentration³⁶ and causing an artifactual increase in Hct. Preanalytic factors to consider include sex, strain, age, altitude, and vendor¹¹⁶ and environmental variables such as diet, housing, and SPF status of the mouse colony.¹³ A factor specific to hematologic analyses is the propensity for mouse platelets to clump.^105,176 Therefore, the use of a consistent collection site and method, handling that minimizes stress and allows for the collection of a sufficient volume of blood, and appropriate anticoagulation are imperative to minimizing preanalytic variables and thus ensuring accurate hematologic analysis and interpretation.

Features of peripheral blood.

Mice have small erythrocytes (RBC), compared with other mammalian species, and the lifespans of erythrocytes and platelets in mice are generally shorter than those in humans and other veterinary species. Consequently mice maintain a somewhat regenerative state, normally having between 1% to 6% circulating reticulocytes, resulting in polychromasia and slight anisocytosis on Wright–Giemsa-stained blood films.^37,130 ‘Ringform’ nuclear morphology, characterized by a circular nucleus, is a normal feature of mouse neutrophils, eosinophils, and monocytes.¹⁰ Mice have very high platelet counts (9 to 16 × 10⁵/μL) compared with those of other mammals,^{37,100,105,119} a contributing factor to the aforementioned potential for platelet clumping.

Features of bone marrow.

A complete assessment of hematologic status in mice should include evaluation of bone marrow and spleen in addition to peripheral blood. Preparation of bone marrow samples for cytology in mice is best accomplished by using the ‘paintbrush technique’ (see Methods), which preserves cellular morphology and evenly distributes cells on glass slides.³⁷ Bone marrow cellularity is higher for mice than other species and does not decrease with age, as it does in humans, but the proliferative capacity of murine hematopoietic stem cells is decreased.^37,100,130 Both granulocytic and monocytic precursors can have ring-shaped nuclei.

Features of the spleen.

In mice, the spleen is the primary site for body iron storage^114,153 and remains an active site for hematopoiesis throughout life. The potential for exuberant extramedullary hematopoiesis in response to anemia must be recognized in this species.^16,17,156 Therefore, the role of the spleen should be considered in any evaluation of hematopoiesis, and this organ should be evaluated in addition to the bone marrow to fully characterize a hematologic phenotype or response to therapy.^47,127

Methods for Collecting High-Quality Blood Samples and for Accurate Hematologic Analyses

Study Design.

The sample size for any experimental design ultimately depends on the variability of the outcomes; for hematology, variability can be established by using appropriate control groups. Pilot studies can be helpful for predicting variability. A statistician should participate in the determination of sample size by power analysis to ensure that hematologic data meet test assumptions and to account for experimental error¹⁷⁶(http://www.uml.edu/Research/OIC/animal-use/helpful-links.aspx). Experimental design for animal studies has been reviewed elsewhere.^{41,42,71,72,131,155} The goal should always be minimizing use of animals while allowing for sufficient power to determine the effects of an intervention.

Blood collection.

Volume.

The average reported blood volume in mice is 7.8 mL per 100 g of body weight.^37,90 Therefore the total blood volume in 9- to 10-wk-old mice is approximately 2 mL, making the maximal volume that can be collected safely at a single survival time point approximately 200 μL. Previous studies in rodents indicate that although serial blood sampling is possible, it requires a postphlebotomy recovery period that depends on the withdrawn volume;^20,37 the maximal collection of 15% blood volume with a 4-wk recovery period has been recommended for mice.^37,90 More frequent or greater blood volume collection than this recommendation has been reported, with recovery defined as the return of mean Hgb values to within 2 SD of mean baseline values provided that no weight loss, behavioral changes, or clinically significant anemia has occurred.¹²³ However, adhering to these alternative criteria may lead to important changes in the hemogram that could alter study outcomes.

Collection site.

The method and site of blood collection in rodents can influence results.^87,100 The submandibular venipuncture blood collection method is recognized to obviate the need for anesthesia, thus removing that potentially confounding variable while minimizing animal distress.^39,48 When performed by trained personnel and with immediate and appropriate mixing of anticoagulated blood samples, we have found that blood collection by submandibular venipuncture can significantly reduce platelet clumping. Although excessive bleeding after submandibular collection in mice with abnormal coagulation is reported,⁶¹ this complication can occur at any collection site in these models, which may therefore require prolonged compression and monitoring after blood collection. Scientific studies frequently report platelet clumping as an issue when collecting blood from mice; therefore some authors subsequently either do not report platelet counts or report artifactually low counts when clumping is suspected.^94,105 Because platelet clumps lead to decreased automated platelet counts, such data always warrant review of a blood film^84,90 to verify thrombocytopenia. Another site historically included for survival blood collection is the retroorbital sinus, but this site is no longer recommended in light of tissue trauma, contaminated and clumped samples, postcollection morbidity, and the need for anesthesia.^37,59,87,155 For terminal samples, cardiocentesis can be performed in sufficiently anesthetized mice and, when performed quickly and efficiently, enables the collection of large sample volumes without platelet clumping. Care must be taken during cardiocentesis to avoid puncturing other viscera and potential contamination of the sample with nonblood cells. Other perimortem collection sites include the aorta and caudal vena cava.³⁷

Sample handling.

Blood samples can be collected directly into an anticoagulant, such as EDTA. Because EDTA (K₂ or K₃) causes less postcollection platelet clumping and provides better staining characteristics, it is preferred over heparin as an anticoagulant for rodent blood.^37,51,82,155 In addition, EDTA is the preferred anticoagulant for most automated analyzers. For greatest accuracy, blood films should be prepared (Figure 1 A) and samples analyzed by automated methods within 4 h and not longer than 24 h after collection. One study using an automated analyzer found increased MCV, RBC distribution width (RDW), and MPV and decreased MCHC and monocyte counts in CD1 mice after storage of blood for 24 h at 4 °C.³ Blood collection tubes should be filled to recommended volume to ensure the correct blood:anticoagulant ratio, and blood should be mixed by inversion gently and immediately after collection to ensure adequate distribution of the anticoagulant. To avoid artifactual hemolysis of samples (Figure 1 B), the blood tube should not be shaken, and the needle should be removed from the syringe prior to dispensing blood into the collection tube to avoid shearing of cells. Although mouse blood samples can be diluted for automated analyzers that require large sample volumes, the accuracy and precision of diluted samples are highly variable,³⁷ and modern automated veterinary analyzers require relatively small sample volumes for CBC analysis (for example, 20 μL [Heska HemaTrue], 50 μL [Idexx ProCyte]), making survival and sequential sampling designs feasible.

Figure 1.

Features of mouse peripheral blood: RBC and platelet morphologies. (A) Example of a blood film stained with a Romanowsky type stain showing the classic ‘half-moon’ appearance and feathered edge (arrow); bracket highlights the monolayer counting area. (B) Examples of hemolyzed plasma, which is due to poor collection method or disease and can interfere with automated hematologic analyses. (C) Normal erythrocytes and platelets with a normal percentage (approximately 1% in this case) of polychromatophilic erythrocytes (arrows). (D) New methylene blue staining showing remnant RNA in reticulocytes (arrowhead). (E) Prussian blue staining showing atypical iron inclusions in an erythroid precursor (top arrow) and erythrocyte (bottom arrow). (F) A megaplatelet (arrow) is markedly larger than surrounding erythrocytes. (G) Giant platelets (arrow) are approximately the same size as erythrocytes, compared with the much smaller, normal platelets (arrowhead). (H) Circulating atypical nucleated erythrocyte (nRBC) undergoing nuclear division. (I) Sickled RBC (arrowheads) induced by a genetic mutation. (J) Iron deficiency induced by a genetic mutation (tmprss6^−/−) has led to anisocytosis, fragmented (thin arrows) and microcytic RBC (arrowhead). (K) Echinocytes (spiculated RBC, arrowhead) in a mouse model of polycystic kidney disease (CD1^pcy/pcy) with anemia and renal failure. Wright–Giemsa staining except where noted; original magnification, ×100.

Slide preparation.

Fresh blood films should be prepared at room temperature and stained within 4 h of collection by using anticoagulated blood that is well-mixed. Mixing is important and can be performed manually by gently inverting the tube 5 to 10 times or by placing on a tilting or rotating rack designed for mixing blood. A clean microhematocrit tube or pipette then is used to dispense a drop of blood onto one end of a clean microscope slide. A second slide is placed at an angle of 30° to 45° in front of the drop of blood and then is backed into the drop of blood. Once the blood droplet has spread along the edge of the angled slide, it is then pushed forward in a single rapid motion to create the classic ‘half-moon’ profile that provides a monolayer for cell counting (Figure 1 A).¹⁶⁸ To avoid cellular morphology artifacts, the blood film must be allowed to air dry fully (typically at least 30 to 45 min, depending on humidity and temperature) before being stained with a Romanowsky-type stain for review (Figure 1 C).^62,168 Unfixed slides can be saved for future use (for example, additional stains, confirmation of results).

Manual WBC counts and correction for nucleated RBC (nRBC).

A manual leukocyte differential count is important for hematologic analyses in all species, and mice in particular, given that available veterinary analyzers have not yet been validated fully for use in mice;^94,158 in this analysis, 100 (or 200 to 500, for improved accuracy) WBC are counted and categorized.⁶² Cell-type percentages then are multiplied by the total WBC count (generally obtained from an automated analyzer) to determine the absolute count for each cell type. To correct the WBC count when nRBC comprise greater than 5% of the total nucleated cell count (TNCC), the number of nRBC per 100 nucleated cells is determined through blood-film review, and the TNCC (or total WBC count), typically from an automated analyzer, is multiplied by 100 / (no. of nRBC + 100).¹⁴⁸

Ancillary tests.

At the time of slide preparation, it is also useful to perform a PCV or spun Hct measurement. A small aliquot of blood is drawn into a capillary tube and sealed with tube sealant on one end. The tube then is spun in a microhematocrit centrifuge for more than 3 min, separating the blood components into 3 layers—plasma, buffy coat (WBC), and packed erythrocytes. Icterus, lipemia, and hemolysis can be detected by holding the capillary tube against a white background. The total protein concentration can be estimated by breaking the tube and loading the plasma into a refractometer. Reticulocyte counts can be obtained from some automated analyzers (ProCyte, Advia), if validated, or can be performed manually. For manual reticulocyte enumeration, whole anticoagulated blood is mixed with new methylene blue (NMB) and allowed to incubate for 10 to 20 min, after which a blood film is made (Figure 1 D).⁶² An absolute reticulocyte count is achieved by multiplying the percentage of reticulocytes (usually determined by counting the number of reticulocytes per 500 to 1000 RBC) by the absolute RBC count.²⁸

Reference Intervals

Historical reference intervals are available in textbooks,^66,100,119 journal articles^94,105,116 and on the Internet (Jackson Laboratory's Mouse Phenome Database, Knockout Mouse Project, https://www.komp.org/). However, because of variability due to laboratory instrumentation, methods, collection sites, strain, age, sex, and environmental factors, reference intervals are best generated inhouse for any specific experimental population.^105,155 Individual mice in study groups can be compared with their own baselines, or group comparisons can be made.¹²⁷ In addition, individual reference intervals are starting to gain popularity¹⁶² and can be an option for mice now that serial, survival CBC counts are practical given the availability of microvolume, automated analyzers. Reference intervals should be used as a tool, not as the sole guide to determine normalcy.¹¹⁹ In our experience, the single best way to detect an important hematologic phenotype, effect of disease, or response to therapy is to generate concurrent age-, strain-, and sex-matched controls for experiments, by using consistent sample collection and analysis methodologies including collection time, site, fasting status, and automated analyzer.

Peripheral Blood: General Considerations and Responses to Disease

Erythrocytes in health and disease.

Murine RBC are spherical, anucleate biconcave discs with central pallor (Figure 1 C) that are approximately 4 to 7 μm in diameter. In general, total RBC counts range from 7800 to 10,600 per microliter.^87,119 However, because counts vary with mouse strain, automated analyzer, and other factors, experiment-specific age-, strain-, and sex-matched controls should always be included to detect important experimental differences. Hct, which is a measure of RBC volume, is lower in very young and very old mice because of their lower RBC absolute count and expanded plasma volume, respectively.^37,130 In healthy control mice, Hct ranges from 35% to 52%, and in general, should be 3 times the Hgb concentration.^{37,100,130,166,168} Importantly, when automated analyzers use spectral analysis, hemolysis can alter MCV and therefore the calculated Hct value. A spun PCV analysis, which is essentially a manual Hct determination, can be performed to confirm an automated Hct value and should be used when there is a mismatch of the 3:1 Hct:Hgb ratio. Mice have small RBC compared with that of most other mammalian species, with an average MCV of 45 to 55 fL in health. The normal mouse MCHC usually is 30 to 38 g/dL; the RDW, which reflects variation in RBC size, can fluctuate greatly depending on instrumentation¹¹⁹ and the number of circulating larger immature (polychromatophilic) RBC, even in healthy mice.

Anisocytosis (that is, variation in RBC size) is a common feature seen on Wright–Giemsa-stained blood films because of the normal presence of polychromatophilic cells, which are larger than mature RBC, and is reflected in the RDW value. Mouse RBC have a halflife of 38 to 52 d, which is shorter than that in humans and other veterinary species.^37,100,130 Polychromasia is identified on Wright–Giemsa-stained blood films as increased cytoplasmic basophilia due to increased RNA content. Polychromatophilic erythrocytes are anucleate and generally larger relative to the orange-red mature RBC (Figure 1 C), and when increased in number, indicate an erythropoietic response to anemia. Reticulocytes are polychromatophilic erythrocytes that are identified and manually counted by using NMB-stained preparations, in which RNA networks are visible as aggregates of dark-blue stain (Figure 1 D). Alternatively reticulocytes can be counted by using automated analyzers with murine-validated reticulocyte enumeration capability. Reticulocyte counts are reported as a percentage of the total RBC count or as an absolute number, with absolute numbers generally preferred for the interpretation of a regenerative response.^28,119 In addition, low numbers of nucleated RBC precursors or metarubricytes (<5%) can occur in peripheral blood films of healthy mice and increase during both physiologic and pathologic erythroid regenerative responses.^100,174 Correction of the WBC count is recommended when nucleated RBC (nRBC) are greater than 5% of the TNCC (see Methods section). In addition, Howell–Jolly bodies (micronuclei) can occur in mouse RBC (Figure 2 H).^37,100 The number of nRBC or Howell–Jolly bodies (or both) may increase in response to diseases causing anemia, inadequate splenic function, or myelodysplastic syndrome and myeloproliferative disorders.^{37,64,101,121,125,133,174} Heinz bodies, which are cellular inclusions of denatured hemoglobin, result from oxidative injury, such as occurs after phenylhydrazine administration.²³

Figure 2.

Features of mouse peripheral blood leukocytes: WBC morphologies. (A) Normal mature neutrophil (segmented neutrophil), characterized by faint, finely granular cytoplasm and dense chromatin. (B) ‘Figure-8’ segmented neutrophil, a common finding in peripheral mouse blood. (C) Ringform monocyte, characterized by abundant pale blue-gray cytoplasm, open chromatin pattern, and cytoplasmic vacuoles. (D) Ringform eosinophil, characterized by multilobed nucleus with typical dense chromatin pattern and abundant eosinophilic cytoplasmic granules. (E) Pseudo-Pelger–Huet-type anomaly in a mouse with myelodysplasia (Xist^−/−), showing hyposegmented nucleus with mature dense chromatin in a neutrophil. (F) Atypical trilobed leukocyte in a mouse with myelodysplasia (Xist^−/−). (G) Circulating lymphoblasts representative of lymphoid leukemia in an aged mouse. (H) Reactive lymphocyte (arrowhead) and Howell–Jolly bodies (arrows) in erythrocytes; the presence of low numbers of Howell–Jolly bodies is considered a normal finding. I. A reactive lymphocyte (arrowhead) in comparison to a normal small lymphocyte. (J) Vacuolated lymphocyte (arrow indicates a vacuole) in a murine model of juvenile neuronal ceroid lipofuscinosis. (K) Vacuolated lymphocyte (arrow) and eosinophil (arrowhead) in a mouse model of juvenile neuronal ceroid lipofuscinosis. Wright–Giemsa staining; original magnification, ×100.

Abnormalities in total RBC numbers or cellular morphology can be detected on review of Wright–Giemsa-stained blood films. Decreased RBC counts (anemia) occur secondary to conditions that include blood loss, immune-mediated hemolysis, inflammatory disease, renal disease, iron deficiency, myelodysplastic disease, genetic disorders, and neoplasia.^{16,17,26,74,89,147,152,174} CD47 is important to the recognition of RBC by phagocytes, because the absence of this protein leads to severe immune-mediated hemolytic anemia.¹¹⁰ Mice with exuberant regenerative responses to anemia can have increased anisocytosis (increased RDW) or macrocytosis (increased MCV) and low numbers of circulating early erythroid precursors, such as rubriblasts, prorubricytes, and metarubricytes, which are all nucleated forms of RBC. Importantly, GEM models of disease and disease states such as myelodysplasia and neoplasia can present with erythroid dysplastic changes including increased numbers of circulating megaloblastic rubricytes, sideroblasts, and siderocytes (identified by using Prussian blue staining; Figure 1 E), normochromic macrocytes and other nuclear:cytoplasmic asynchrony, and nuclear changes including multiple (Figure 1 H), fragmented, or lobulated nuclei; atypical mitoses; abnormal chromatin patterns; and prominent nucleoli.^{11,125,147,159,174} An increased RBC count (polycythemia) may be relative or absolute. Relative polycythemia is caused by hemoconcentration secondary to dehydration, high altitude, myelodysplasia, or neoplasia.³⁷ Absolute polycythemia may be caused by increased erythropoietin production in response to anemia or hypoxia or by genetic mutations, such as overexpression of the Kit protein, that cause increased erythropoiesis.^15,32,130

Poikilocytosis (variation in RBC shape) can take many forms, including acanthocytes, dacryocytes, eccentrocytes, echinocytes, keratocytes, schistocytes, spherocytes, stomatocytes, and target cells (Figure 1 I, K). RBC morphology can provide clues to pathogenesis in many diseases, and characteristics identified by a simple blood film review can aid tremendously in the interpretation of disease model phenotypes. Some examples of disease and their characteristic RBC morphologies include iron deficiency (schistocytes, microcytes, keratocytes; Figure 1J), liver disease (target cells, acanthocytes), disseminated intravascular coagulation (schistocytes), immune-mediated hemolytic anemia (spherocytes), renal disease (echinocytes; Figure 1K), and myelofibrosis (dacryocytes).

Leukocytes in health and disease.

The typical WBC count in mice is 2000 to 10,000 per microliter.^100,119 In general, an automated total WBC count is more precise and accurate than is a manual count, because many more cells are counted during automated analysis. However, reviewing the blood film is an important quality-control check, especially when results are beyond the normal range, to confirm TNCC and automated WBC differential counts, define severe leukopenia or leukocytosis, determine left shifts (increased number of immature granulocytes in peripheral blood), and identify cells with atypical morphology (Figure 2 E, G). A quick scan of the feathered edge (Figure 1 A) is an important component of any slide review and serves to identify large atypical cells, platelet clumps, mast cells, nonhematopoietic cells, and parasites.¹⁶⁸ Many disease phenotypes can be better characterized and diagnosed after careful review of a stained blood film. Automated analyzers tend to undercount mouse monocytes, for example, and often will not differentiate atypical lymphocytes, granulocyte types (Figure 2 A, B), bands and early myeloid precursors, nRBC, or neoplastic cells.^{4,7,17,53,,113,145} Differentiating types of leukocytes can pose a challenge in mice due to their unique morphologic characteristics including, for example, the ring forms normally seen in both granulocytic and monocytic lineages (Figure 2 C, D).¹⁰ As nucleated cells, nRBC will be automatically included by automated veterinary analyzers in the TNCC and not differentiated as nRBC and thus will increase total WBC counts artifactually. Therefore, nRBC should be enumerated as a percentage during blood film review. Correction of the total WBC count is recommended when nRBC account for more than 5% of the TNCC (see Methods section). Newer technology in automated analyzers shows promise for identifying the presence of nRBC and bands.¹¹⁸

Lymphocytes.

Lymphocytes are the predominant leukocyte in most strains of healthy wild-type mice, making up 70% to 80% of the WBC differential count.^37,87,119 They are typically 10 to 15 μm in diameter with scant blue cytoplasm, a smooth chromatin pattern, and a round, oval, or slightly indented central nucleus (Figure 2 I).^37,56,87 However, lymphocyte morphology can vary even in health, and variants include larger forms with more dispersed chromatin patterns and increased cytoplasm, which ranges from pale to dark blue and can include azurophilic granules (large granular lymphocytes).⁸⁷

In mice, lymphocyte counts can decrease with handling or other stressors¹³⁵ and with age as neutrophil counts increase,^66,119 again demonstrating the need for strain- and age-matched controls in all studies. Vacuolated lymphocytes can be seen in murine lysosomal storage disease models including juvenile neuronal ceroid lipofuscinosis (Figure 2 J, K), sialic acid storage disease, mannosidosis, and Pompe disease^{63,95,124,147,178} as well as in chronic active inflammation with lipidosis.¹⁷ The numbers of activated lymphocytes (Figure 2 H) and large granular lymphocytes in circulation can increase due to lymphoma or leukemia, immune stimulation, and some viral infections,^108,177 and circulating lymphoblasts can be seen with lymphoid leukemia in aged mice (Figure 2 G).

Neutrophils.

Neutrophils generally comprise 20% to 30%¹¹⁹ of the WBC count in mice and are the most common granulocyte. Specific morphologic characteristics include pale, finely granular cytoplasm and a segmented nucleus with areas of both pale and condensed chromatin (Figure 2 A). Unique features of mouse neutrophils include ringform and ‘figure 8’ nuclei (Figure 2 B)^10,114 and high numbers (5 or 6) of nuclear indentations, which can be interpreted mistakenly as hypersegmentation.¹¹⁹ Neutrophils have a small storage pool and a 7- to 10-h circulating halflife.⁶ Once released from the bone marrow, neutrophils are allocated into the marginal pool and the circulating pool; the circulating pool is included in the leukocyte count. The proportion of neutrophils in the marginal compared with the circulating pool varies by mouse strain.^79,144 The circulating neutrophil count depends on the rate of their release from the bone marrow, the distribution between the marginal and circulating pools, and the rate of migration into tissue.¹⁴⁴

Increased neutrophil counts (neutrophilia) are associated with responses to stress or excitement¹³⁵ and infectious diseases, and typically increase in cases of bacterial infection and acute inflammation^16,17,38,88 given their primary phagocytic and microbiocidal roles. Neutrophil counts can also increase during myeloproliferative disease and myelo- and myelomonocytic leukemia.¹⁷⁴ Due to the small storage pool of neutrophils, both immature and toxic neutrophils may circulate during inflammatory diseases. Immature neutrophils have band or horse-shoe-shaped nuclei (that is, nuclei are not yet lobulated); band cells typically are larger than are mature neutrophils. Dysplastic changes include cytoplasmic hypogranulation, bizarre granulation, nuclear hyposegmentation and bilobation (pseudo-Pelger–Huet anomaly, Figure 2 E), and atypical chromatin condensation.^{2,27,125,140,174} In addition, a spontaneous mutant mouse model (mouse ichthyosis) of Pelger–Huet anomaly with a laminin B receptor gene mutation, is similar to that detected in the human disease.¹⁴³ The immature neutrophils seen during inflammation can include both bands and young forms such as metamyelocytes and myelocytes, with less mature forms being consecutively less abundant.¹⁴⁴ Bone marrow is the predominant site for increased granulopoiesis during acute inflammation, whereas the spleen is the predominant site for increased erythropoiesis during acute erythropoietic responses.^17,37 The main growth factors for neutrophils include GM-CSF and G-CSF, and their primary chemoattractants include the IL8 homologs MIP2, LIX, and KC.^{37,60,81,86,117}

Eosinophils.

Eosinophils are granulocytes that are characterized by bright orange to red, round, cytoplasmic granules that are uniformly sized with indistinct borders.¹¹⁹ These cells generally comprise 0% to 7% of the murine WBC differential count.¹¹⁹ The nuclei of mature eosinophils typically are multilobulated with condensed chromatin and can be ringform (Figure 2 D). Immature eosinophils are band in form. Eosinophils are involved in parasitic and allergic reactions, and their counts are increased in some GEM models of neoplasia, including chronic eosinophilic leukemia.¹⁷³ IL5 is thought to function as their main growth factor and eotaxin as their primary chemoattractant.^37,173 The cytoplasm of eosinophils contains discrete vacuoles in mice with juvenile neuronal ceroid lipofuscinosis (Figure 2 K).

Basophils.

Basophils are rare in mice; these cells have few but large, round, deeply basophilic cytoplasmic granules and segmented nuclei. Basophils, like eosinophils, increase in number during parasitic and allergic responses.^44,166 Basophils must be differentiated from mast cells, which occasionally are present in the circulation of mice with inflammatory disease, necrosis, tissue injury, or severe regenerative anemia.⁶⁵ Mast cells are larger than are basophils and have round, nonsegmented nuclear morphology and more dense metachromatic cytoplasmic granules (Figure 3 C).^66,166 Basophil counts may be increased when blood is collected from the tail, as the cells are squeezed from the tissue into the blood during collection.¹⁰⁰ IL3 is the primary growth factor³⁷ for basophils, and they produce IL4.^98,136

Figure 3.

Other cells that can be seen in mouse blood films. (A) Hepatocytes on feathered edge after aspiration of liver during cardiocentesis. (B) Plasmodium sp (arrowhead) infecting approximately 50% of erythrocytes and markedly increased polychromasia in an anemic mouse; mouse model of malaria. There are also lymphocytes (L) and a neutrophil (N) in this field. (C) Mast cell (atypical finding in mouse peripheral blood) at feathered edge. (D) Large platelet clump at the feathered edge; automated platelet counts will be artifactually low due to clumping. Wright–Giemsa staining; original magnification, ×100.

Monocytes.

Monocytes are the largest leukocyte and typically make up less than 2% of the total WBC count in mice.¹⁰⁰ However, automated analyzers may undercount this population of cells, and values should be verified by manual review of blood films.^{4,6,17,53,,113,145} Monocytes are characterized by their abundant pale gray-blue cytoplasm which often contains vacuoles and occasional faintly eosinophilic granules; nuclei have loose chromatin and are generally bi- or trilobed, reniform, or horseshoe-shaped.^37,87 Monocytes can, also display ringform nuclear morphology (Figure 2C).¹⁰ Mouse monocytes have a 17-h halflife in circulation; 40% of the population of peripheral blood monocytes is circulating, whereas 60% of monocytes are marginated.^37,85,160 The main monocyte growth factors are M-CSF, GM-CSF, and IL3,^37,85 and MCP1 is a primary chemoattractant.³² Monocytes are a major source of cytokines in the blood, including IL1β, TNFα, and IL6, and monocytosis has been associated with intracellular bacterial infections,²⁹ chronic inflammation,^{16,17,141,166} and neoplasia.¹⁷⁴ Monocytosis also occurs with experimental hemoparasitism, such as trypanosomiasis and malaria,^30,111 and with viral infections, as with mouse cytomegalovirus.¹⁴¹ Immature forms and promonocytes can be seen during neoplastic conditions.¹⁷⁴ When present on freshly prepared blood films, hemophagocytosis should be noted and may accompany monocytosis, neoplastic disease,¹⁷⁴ or hemolytic anemia. Circulating immature monocytes should be differentiated as part of the phenotypic description of disease models.¹⁷⁴

Platelets in health and disease.

Compared with other mammals, mice have very high platelet counts (900,000 to 1,600,000 per microliter).^37,100,119 Platelet activation in mice can be spontaneous and strain-dependent.³⁷ Adenosine diphosphate, collagen, arachidonate, and thrombin are potent agonists of platelet clumping in mice.²³ Platelets originate from megakaryocytes in the bone marrow and spleen of mice, and their primary growth factor is thrombopoietin, which is predominantly produced in the liver.^{21,103,104,122,149} The lifespan of platelets in mice, approximately 5 d, is shorter than that in other species.^23,37 Automated analyzers often underestimate platelets counts in mice, due to both the small size of platelets and their propensity to clump (Figure 3 D).^37,87 For example, one analyzer (the Advia 120) may falsely report platelet clumps as eosinophils, because the highly variable size and granularity of these clumps cause them to appear as a heterogenous population in the area of the dot plot where eosinophils normally appear. As for other veterinary species, the feathered edge of a blood film (Figure 1 A) should be scanned for platelet clumps to aid in judging the accuracy of automated platelet counts in mice.

On blood films, mouse platelets are 1 to 4 μm in diameter, anucleate, with discoid, spheroid, or elongated or spindloid morphology, and central basophilic, eosinophilic, or metachromatic granules scattered throughout faintly pink to gray cytoplasm.^37,87,119 Cell membranes may have a few fine threadlike surface projections. Both mature and reticulated (young) platelets can be counted by flow cytometry.^119,130 In addition, mice may have circulating giant or megaplatelets, described as such when they are equal in size to or larger than an RBC, respectively (Figure 1 F, G). Giant platelets increase in number in response to accelerated hematopoiesis; this morphology can be correlated with an increased MPV.⁶⁶

The main function of platelets is primary hemostasis, and platelet production can increase due to inflammatory disease, myeloproliferative disease and neoplasia, and iron deficiency.^57,119,138 Conversely, platelet production can decrease due to myeloproliferative disease, neoplasia, and erythropoietin administration and in various GEM models.^58,97,125 Dysplastic changes, including retained nuclei, circulating micromegakaryocytes,⁷⁶ and atypical cytoplasmic granulation,¹⁷⁴ are associated with myeloproliferative disorders, for example. Giant and megaplatelets can occur with leukemias, myelofibrosis, thrombocythemia, and polycythemia vera^83,139,174 and are released from the bone marrow in response to thrombocytopenia and when the peripheral blood halflife of normal circulating platelets is decreased (Figure 1 F, G).³⁷ Populations of giant and megaplatelets as well as microcytic and fragmented erythrocytes can overlap in automated analyzers that sort cells according to size, thus skewing both cell counts (Figure 4 E).^75,152 The MPV is a sensitive indicator of increased platelet production and increases in response to hypoxia-induced thrombocytopenia^70,97 and various genetically or physiologically induced causes of thrombocytopenia.²³

Figure 4.

Features of an automated analyzer, the Heska Hematrue. The manual micropipette adapter (MPA, 20 µL) enables analysis of small volume blood samples. Using the MPA, blood is withdrawn by capillary action into (A) a specialized microhematocrit tube held by provided forceps. The microhematocrit tube then is placed into the (B) MPA device and inserted into (C) the MPA holder. The analyzer prints cellular histograms showing 3 populations (WBC, RBC, and platelets) for evaluation and comparison with results from review of blood films. (D) Histograms from a naïve C57Bl/6 mouse sample showing normocytic RBC with an MCV of 41.8 fl. (E) Histograms from an anemic transgenic iron-deficient mouse (tmprss6 ^−/−), showing a microcytic RBC population (shifted left on the x axis) with an MCV of 33.8 fl.

Other cells.

Circulating mast cells (Figure 3 C) occur occasionally, depending on the blood collection site, and must be differentiated from basophils (described earlier). In addition, epithelial cells can be seen secondary to a poorly performed cardiac puncture when other internal organs are aspirated inadvertently. For example, when the liver is aspirated during cardiocentesis, hepatocytes might collect along the feathered edge of the blood film (Figure 3 A). Furthermore, blood parasites sometimes are present in blood films from various mouse models with infections such as malaria (Figure 3 B), babesiosis, and trypanosomiasis.^12,52,92

Automated Methods in Mouse Hematology

Several options are available for automated veterinary analyzers, including impedance, laser, and combination instruments. Given the cellular differences we already have described, human automated analyzers will not provide accurate results for mice. Although veterinary analyzers are more often evaluated and reviewed for companion animal medicine,^7,74,113 both Bayer (Tarrytown, New York) and Abbott (Abbott Park, Illinois) have software for WBC differentiation for mice,³⁷ and Heska (Loveland, Colorado) and Idexx (Westbrook, Maine) have recently added mouse-specific software and microvolume sample options (Figure 4 A–C). Instrument operators must be trained in maintenance procedures, error flags, troubleshooting, and appropriate quality control,¹⁶⁸ and instruments should go through a standardized validation procedure and be compared with a ‘gold-standard’ method, such as a manual method or a previously validated instrument.^50,91,161 The leukocyte differential is one component of the automated CBC analysis that has still not replaced manual methods as the gold standard in veterinary medicine.¹⁶¹ Mice have several physiologic variables that may decrease the efficacy of automated analyses including small RBC, relatively high numbers of circulating immature and nucleated RBC, variable leukocyte morphology, and platelet clumping. Therefore, manual leukocyte differential and blood-film review are still warranted. In addition to these physiologic factors, the mouse's utility for biomedical research includes the ease with which its physiology and genetics can be manipulated. These manipulations can lead to unpredictable changes in hematologic variables; therefore it is important to consult a blinded, well-trained observer familiar with veterinary, and particularly murine, clinical pathology to detect and appropriately interpret changes in blood cell counts and indices, distribution, and morphology by using appropriately matched controls.

Genetic mutations, iron deficiency, and inflammatory disease are examples of disorders that can cause changes in cell morphology or size, thus resulting in overlap of various cell populations by automated analyzers. Therefore, evaluating not only a blood film but also instrument-generated cell-population histograms (Figure 4 D, E) becomes necessary.^23,168 Previous studies show that markedly microcytic or fragmented erythrocytes (as well as other cell debris) can overlap with the size-based platelet population,^{17,43,152,168} thus skewing both RBC and platelet counts (Figure 4 E). Results from automated analyzers should always be verified by manual blood film review, especially when the automated analyzer reports atypically high or low values.

Methods for Collecting Quality Samples and Analysis of Bone Marrow

Collection for cytology.

The optimal method for generating samples for mouse bone-marrow cytology is the brush preparation. When properly performed, this method preserves cellular morphology without dilutional or mechanically induced cellular damage. Samples should be collected immediately postmortem from the sternum or femur.¹²⁷ Briefly, to allow maximal exposure of the marrow, the bone (usually the femur, for the most sample) is bisected lengthwise by using a clean razor or surgical blade. A small paintbrush dampened with PBS is gently brushed along the marrow surface to collect the cells. Then the collected cells are brushed gently lengthwise along a slide in long rows, as the brush is rotated for each row to deposit the cells in a monolayer^37,156(Figure 5 A). Other methods of slide preparation include push slides, squash preps, and cytocentrifugation, in which the marrow is flushed similar to preparation for flow cytometric evaluation,¹²⁷ however, these methods often result in increased numbers of broken and smeared cells, which cannot be evaluated. Bone marrow slides can be stained routinely with Wright–Giemsa, Giemsa, or Prussian blue for evaluation by light microscopy and are generally stained twice as long as peripheral blood films.^156,168 Cellularity and morphology will be excellent when gentle handling is used.

Figure 5.

Features of mouse bone marrow. (A) Representative slide resulting from the ‘brush preparation’ method for bone marrow cytology. (B) Mouse bone marrow sample with normal cellularity and myeloid:erythroid ratio (approximately 1:1 in this case). Wright–Giemsa staining; original magnification, ×50. (C) Empiropolesis of a neutrophil (arrow) through a megakaryocyte, a common finding in mouse bone marrow. Wright–Giemsa staining; original magnification, ×50. (D) Typical megakaryocyte (arrowhead), characterized by multilobulated nucleus with dense chromatin pattern and abundant medium-blue cytoplasm. Wright–Giemsa staining; original magnification, 100×. (E) Myeloid hyperplasia (increased myeloid:erythroid ratio, approximately 7:1 in this case), characterized by increased numbers of mature neutrophils; bone marrow cytology from a colony mouse with cervical lymphadenitis. Wright–Giemsa staining; original magnification, ×100. (F) Formalin-fixed, paraffin-embedded bone marrow from control mouse showing iron-containing macrophages (arrow). Prussian blue staining; original magnification, ×20. (G) Bone marrow from control mouse showing macrophages (arrowhead) containing iron as hemosiderin. Prussian blue staining; original magnification, ×50. (H) Markedly increased reticulin staining (myelofibrosis, black fibers) in formalin-fixed, paraffin-embedded bone marrow from a 9-mo-old mouse with myeloproliferative disease (Xist−/−). Original magnification, ×60. (I) Close-up of normal mouse bone marrow, showing typical ringform neutrophils and their precursors and erythroid precursors (darker cells). Wright–Giemsa staining; original magnification, ×100.

Collection for histology.

Bone marrow from the sternum, vertebrae, humerus, or femur can be collected for histology,^36,127 and consistency of site collection is recommended for best comparison within or between studies. Samples collected at necropsy should be fixed as soon as possible (within 20 min) in 10% neutral buffered formalin or Bouin solution and decalcified in a chelating agent (for example, EDTA) or a weak organic acid.^127,156 Excess decalcification should be avoided to preserve cellular morphology and when sections are to be used for special staining. Alternatively decalcification may be unnecessary before sectioning very small bones. Sections are processed routinely through graded alcohols and embedded in paraffin; 3- to 4-μm tissue sections are recommended for the best cellular detail. Slides then can be stained with hematoxylin and eosin and special stains (discussed later). Bone marrow histology is most useful for assessing architecture, overall cellularity, neoplastic infiltrates, and myelofibrosis.

Collection for flow cytometry.

Flow cytometric analysis can be a useful adjunct to morphologic assessments. The femur is the typical collection site for flow cytometry, allowing for 10 to 30 million cells from a single bone.^23,130 Once the femur is dissected, both ends can be removed by using a razor. A 12-mL syringe is filled with buffer and, with a 21- to 23-gauge needle attached, is inserted into one end of the femur and flushed by using alternating fast and slow pulses to maximize cell yield. Flow cytometric analyses must be performed immediately after cell collection.¹²⁷ These analyses are beyond the scope of this paper but have been reviewed recently.^{31,90,127,170} As with most aspects of phenotypic characterization, flow cytometric parameters can vary with mouse strain and other preanalytic factors.¹¹²

Cytochemical Staining of Blood and Bone Marrow

Routine staining for bone marrow and blood includes use of a Romanowsky-type stain (Figures 1 C, 5 B). Reticulocytes are best visualized and enumerated by NMB staining (Figure 1 D). Polychromasia can be estimated from Wright–Giemsa-stained preparations (Figure 1 C).²⁸ In mouse bone marrow (and peripheral blood with certain manipulations) it can be very challenging to differentiate immature cell types by morphology alone. For example, lymphoblasts, monoblasts, and myeloblasts can overlap in morphologic features, as can atypical or immature cells in cases of myelodysplasia or neoplasia. In some cases, cellular origin cannot be differentiated by routine microscopy, but some cells contain inclusions or storage material that can be used to confirm identification. In such cases, special stains such as periodic acid–Schiff for glycogen inclusions, Masson trichrome for stroma, Prussian blue for iron-positive inclusions (Figure 5 F, G), and reticulin for bone marrow myelofibrosis (Figure 5 H) may be necessary. In addition, unstained slides can be used for other special stains and cytochemical analyses (Figure 6).

Figure 6.

Examples of common cytochemical stains and their uses for mouse bone marrow.

Immunocytochemistry and Immunohistochemistry

Immunocytochemistry and immunohistochemistry can be performed on fresh unstained blood or bone marrow slides, destained slides,⁹³ or histology sections of bone marrow and other tissue. Briefly, slides are fixed with methanol, washed with buffer, and incubated with a protein blocker, then a primary antibody, followed by a secondary antibody, and finally a color-development enzyme solution.¹⁹ Specific antibodies useful in blood and bone marrow samples include those to CD71, Ter119, B220, CD3, CD1b, Gr1, F4/80, Mac2, and CD41.^{17,22,80,165,174} Immature forms can be identified by using various combinations of stem cell markers.²⁵ Additional antibodies useful in the hematopathology of mice using formalin-fixed, paraffin-embedded material are reviewed elsewhere.^77,128 Most antibodies tend to yield higher signals on frozen tissue (avoids the extra steps required for fixed tissues), but frozen samples are suboptimal for assessing tissue and cellular morphology. Although immunocytochemistry and immunohistochemistry provide important diagnostic information, most markers are associated with but are not specific for various cell lineages, so panels are often required for cell identification, and concurrent control samples from age-, sex-, strain-, and disease-matched mice are needed.

Bone Marrow: General Considerations and Responses to Disease

Bone marrow cytologic evaluation is needed to determine the myeloid:erythroid ratio, to differentiate hematopoietic precursors, characterize changes in bone marrow hematopoiesis relative to peripheral blood cell counts, assess individual cell morphology, differentiate lymphoid and erythroid cells, and assess features of myelodysplasia or neoplasia.¹²⁷ In addition, bone marrow is evaluated by histopathology to assess overall cellularity and architecture and to identify necrosis, inflammation, or infiltrative disease.^127,156 Collection sites should remain consistent within and between experiments (see Methods), and concurrent controls are needed. Bone marrow cytology is especially useful when circulating atypical cells have been detected and when erythrocyte indices suggest abnormal hematopoiesis.¹²⁷ Wright–Giemsa-stained preparations are best evaluated by using the 100× objective to assess cellular morphologic detail including megakaryocytes (Figure 5 C, D) and to determine the myeloid:erythroid ratio and cellular percentages (Figure 5 B, I).⁸⁷ At least 500 cells should be counted and classified by type—including myeloid (granulocytic and monocytic), erythroid, and megakaryocytic lineages as well as lymphocytes, macrophages, plasma cells, mast cells—and cellular stage of maturation.¹⁵⁶ Slides should be prepared concurrently from appropriate controls to evaluate for hematopathology. For complete hematologic evaluation in in mice, bone marrow cytologic evaluation is performed as an adjunct to peripheral blood assessment (CBC and blood film review), bone marrow histology, and histopathology of the spleen.^17,127,153 Special stains used for mouse blood and bone marrow are outlined in Figure 6. In general, bone marrow cellularity is higher for mice than for other species, with as much as 90% to 95% of medullary space in the femur and vertebral column occupied by marrow;^100,130 strain- and age-associated variability is best identified by the evaluation of concurrent controls and by using consistent collection sites. The detailed architecture of murine bone marrow has been reviewed.¹⁵⁴ Flow cytometry is a useful adjunct to detailed morphologic assessment of the bone marrow, especially when multiple cell-surface markers are used.^{31,90,127,170}

The normal myeloid:erythroid ratio in mice ranges from 0.8 to 2.8:1 (average, 1.5:1),³⁷ and age and strain must be considered during comparative analyses. The nuclei and cytoplasm should mature together, asynchronous maturation is abnormal, and cellular maturation should be complete and orderly for all lineages. Megakaryocyte emperipolesis (the passage of an intact leukocyte through another cell) is considered a normal finding and is common in bone marrow from mice (Figure 5 C).¹¹⁹ Bone marrow cellularity tends to increase with inflammatory disease (Figure 5 E) and age. However, functional capacity declines with age, and myelofibrosis (fibro-osseous lesions) can occur in mice before they are 2 y of age.^129,151 This abnormality is more common in female than male mice and may represent an estrogenic effect.^46,132,163 In addition, myelofibrosis can be an important feature of bone marrow disease in younger, especially genetically manipulated, mice.^73,83,174 As in human hematopathology practice, identification and grading of myelofibrosis in murine bone marrow by using slides stained with reticulin (Figure 5 H) or trichrome (for type I collagen) can be helpful in the assessment of disease.^78,174 Adipose tissue often increases concurrently with decreasing marrow cellularity, and indeed, adipocytes have been shown to suppress hematopoiesis in mice.^107,154 Importantly, hematopoietic neoplasia occurs with relatively high frequency in different strains of aging mice (Figure 2 G), and this fact must be considered carefully when attempting to differentiate a neoplastic phenotype from a strain-related background lesion and to characterize such neoplasms correctly.^14,77 New genetic mutations are being discovered or induced in mice at an ever-increasing rate, and many of these have important direct or indirect effects on hematopoietic development, can cause immunosuppression, and that must be considered in the context of the background strain when interpreting bone marrow findings.^1,137,171

Erythroid lineage.

Erythropoiesis begins in erythroblastic islands, which consist of a central ‘nurse’ macrophage surrounded by RBC at various stages of differentiation. Rubriblasts are large cells with large round nuclei, finely stippled and reddish chromatin, nucleoli, and a narrow rim of deeply basophilic cytoplasm. The next stage of maturation is the prorubricyte, with more coarse chromatin, and loss of nucleoli. Rubricytes are the most mature stage capable of mitosis. These cells are smaller than the earlier forms, with very coarse chromatin and light blue to gray cytoplasm (Figure 5 I). Metarubricytes are smaller than rubricytes, with a pyknotic nucleus and polychromatic cytoplasm. Removal of the nucleus from metarubricytes leads to the formation of polychromatophilic erythrocytes, which are reticulocytes that contain aggregates of RNA which stain with NMB. The final stage of erythroid maturation is the mature erythrocyte, which is anucleate and has pink to red hemoglobinized cytoplasm.¹⁵⁶ In human medicine, the term ‘normoblast’ is used for RBC precursors, with pronormoblasts roughly equivalent to rubriblasts, basophilic normoblasts to prorubricytes, polychromatophilic normoblasts to rubricytes, and orthochromatic normoblasts to metarubricytes.⁴⁹ Erythroid precursors can comprise 20% to 50% of the TNCC in bone marrow, and with normal maturation, the more mature forms are present in greater numbers than are immature forms. The primary erythroid growth factor is erythropoietin, a protein secreted by the kidneys.^37,156 Dysplastic changes of erythroid cells include binucleate precursors, siderocytes and ring sideroblasts, nuclear cytoplasmic asynchrony, and atypical mitoses.^125,159,174 Orderly, progressive maturation or maturation arrest should be assessed and blast percentages enumerated as part of an evaluation for myeloproliferative neoplasia.

Myeloid lineage.

The granulocytic myeloid lineage normally progresses from myeloblast (early stage) to promyelocyte, myelocyte, and then metamyelocyte.¹⁶⁹ Band neutrophils and segmented neutrophils (most mature stage) predominate in normal maturation (Figure 5 I). The monocytic myeloid lineage starts with the same bipotential granulocyte–macrophage colony-forming unit (CFU-GM) as for the granulocytic lineage but then is influenced by IL3, GM-CSF, and M-CSF to accomplish monoblastic differentiation.^9,85 Promonocytes develop with cytoplasmic vacuolation and irregular cell membranes. Myeloid precursors account for 30% to 50% of the TNCC in bone marrow. Mature forms should outnumber immature forms, with blasts comprising 2% of the total myeloid component.⁸⁷ Increased blasts may indicate neoplasia of either the myeloid or lymphoid lineage.¹²⁸ Ring forms can be present starting at the promyelocyte stage in neutrophils and eosinophils and in the monocyte lineage. Myeloid hyperplasia can occur during infectious disease and myeloproliferative disease.^16,174 Histiocytic sarcoma is a common tumor of aging mice and is characterized by sheets of round to elongate cells with abundant variably foamy eosinophilic cytoplasm, vesiculate nuclei, and prominent nucleoli.⁵⁵ Erythrophagocytosis and multinucleated giant cells can occur with infiltrative histiocytic sarcoma of the bone marrow and spleen^164,174 and with other lympho- or myeloproliferative neoplastic and inflammatory diseases including granulomatous diseases⁶⁷ and intracellular bacterial infections such as salmonellosis.^16,17,96

Megakaryocytes.

Megakaryopoiesis occurs adjacent to the sinus endothelium.¹⁵⁶ Megakaryocytes are platelet precursors and form in the bone marrow first as megakaryoblasts, which are large with 1 to 4 reddish nuclei and a small amount of deeply basophilic cytoplasm. These then progress to promegakaryocytes, in which nuclei multiply and may fuse into a common mass with a narrow rim of cytoplasm; finally, megakaryocytes are formed (Figure 5 D). Megakaryocytes are the largest (20 to 160 μm diameter) hematopoietic precursors in the bone marrow. They are usually round, with a single, multilobed nuclear mass, abundant pale cytoplasm, and numerous small azurophilic granules. There are about twice as many megakaryocytes in adult mouse bone marrow as compared with human.¹³⁴ Dysplastic changes include hypolobation, atypical mitoses, small forms, and clustering.^35,150,174 Mouse megakaryocytes typically have 16 nuclei, but this count varies by strain.^65,134

Other cells.

Lymphocytes are more abundant (7% to 21% of nucleated cells) in the bone marrow of mice than other species,¹¹⁹ and small lymphocytes predominate. Young mice may have increased numbers of lymphocytes; bone marrow lymphocyte density does not correlate well with peripheral blood lymphocyte counts¹²⁷ and may vary by age, sex, strain, and GEM. Although distinguishing lymphoid from erythroid cells in histologic sections can be difficult, mouse-specific lymphoid and erythroid markers are available for immunohistochemistry.¹²⁸ Erythroid cells can be identified in mice by using the Ter119 antibody.^146,174 Lymphoblasts are present in low numbers (less than 2%) in normal mice, and immunohistochemistry or flow cytometry can be helpful in differentiating cell lineages.^31,172 Other cells normally present in low numbers in the bone marrow of mice include macrophages, plasma cells, osteoclasts, mast cells, and endothelial cells. Bone marrow macrophages are often phagocytic and notably increase in various infectious diseases^16,17 and neoplastic conditions.^77,174 Plasma cells, which should account for less than 3% of the TNCC, are round to oval, with abundant deeply basophilic cytoplasm and round eccentric nuclei with a perinuclear clear zone, and the cytoplasm may be filled with Russell bodies (immunoglobulins).^87,172 Osteoclasts are large irregular cells with multiple oval nuclei and pink granular cytoplasm, which may increase with bone remodeling, changes in hormone and vitamin D levels, and sarcomas.^18,44,54 Osteoclasts can be differentiated from megakaryocytes by the separated nuclei (as compared with as a single mass) and more open chromatin pattern of osteoclasts. Mast cell percentages can increase during innate immune responses and neutrophil recruitment.⁴⁵ Both osteoclast and mast cell numbers in the bone marrow vary with mouse strain.^45,106

Splenic Extramedullary Hematopoiesis

Because the major hematopoietic tissues of mice include bone marrow, spleen, and, to a lesser extent, liver,^37,114,153 analysis of these tissues should be included with CBC and blood film review for complete hematologic evaluation in this species. The spleen is the primary site for iron storage in mice,^114,153 but the iron storage amounts vary by strain.⁴⁷ Splenic extramedullary hematopoiesis persists throughout life in normal mice and comprises about 30% of hematopoiesis^114,130. Splenic extramedullary hematopoiesis produces all 3 hematopoietic lineages: myeloid precursors, erythroid precursors, and megakaryocytes.¹⁵³ This degree of splenic extramedullary hematopoiesis is unique to rodents and is an important consideration in regard to comparative pathology when developing or assessing mouse models of human disease.¹⁷⁵ Under normal conditions, low levels of extramedullary hematopoiesis may occur in the liver as well as other sites, including lymph nodes.^68,73,167 Lymphoma, histiocytic sarcoma, mast cell tumor, hemangioma and hemangiosarcoma, and leukemia can occur in the spleens of rodents.¹⁵³ Whether from anemia, inflammatory disease, altered iron trafficking, storage disease, or other pathologies, marked splenomegaly is a unique and common response to increased hematopoiesis in mice and should not be misconstrued as a neoplastic phenomenon only.^{16,17,24,47,99,147,153,174} In general, the spleen is the primary responding tissue for increased erythropoiesis, whereas the bone marrow is the main responder for myelopoiesis.^17,37,167

Conclusions

Herein we have reviewed murine hematology and provided examples of important hematopathologic responses in mice, thus demonstrating the value of complete murine hematologic analyses during biomedical research. An important component of biomedical research is the potential translation of findings to human and veterinary medicine. Appropriately designed studies with correct and consistent sample collection and evaluation and relevant controls support the elucidation of important phenotypes, hematopathologies, and responsiveness to investigational therapeutics. Blinded observations, randomization,¹¹⁵ and studies designed with appropriate statistical power⁴² enhance the quality of research and allow for repeatability and subsequent translation of important findings. A strategic plan for hematologic analyses in mouse models should include the evaluation of GEM and other disease models for both expected and unexpected hematologic phenotypes.

High-quality hematologic analyses for laboratory mice are possible and practical and can provide insight into the phenotypes and pathogenesis of mouse models of human and veterinary diseases and the evaluation of responsiveness to novel therapeutic investigations. In this manuscript, we describe various practical methods for the collection and processing of samples for automated and manual analyses and provide examples of the utility of these laboratory tests for modern biomedical research. We hope this article provides a reference framework for improved experimental design, understanding of hematophysiology, interpretation of results and important findings for murine hematologic studies, and description of important hematopathologic responses and cellular morphologies clearly demonstrable in mouse blood and bone marrow.