Anti-CD71 antibody immunohistochemistry in the diagnosis of acute myeloid leukemia, subtype acute erythroid leukemia with erythroid dominance (AML M6-Er), in a retrovirus-negative cat

Abstract

CD71 is an immunohistochemical marker used in diagnosing acute myeloid leukemia (AML) M6-Er in humans; however, to our knowledge, it has not been reportedly used for immunohistochemistry in veterinary medicine. We evaluated the pathologic features of AML M6-Er in a retrovirus-negative cat and used CD71 to support the diagnosis. A 4-y-old spayed female Scottish Fold cat was presented with lethargy, anorexia, and fever. Whole-blood PCR assay results for pro feline leukemia virus/pro feline immunodeficiency virus and feline vector-borne diseases were negative. Early erythroid precursors were observed in the peripheral blood smear. Fine-needle aspiration of the enlarged spleen and splenic lymph node showed many early erythroid precursors. Bone marrow aspirate smears revealed erythroid hyperplasia with 68.4% erythroid lineage and 3.6% rubriblasts. Dysplastic cells infiltrated other organs. The patient was diagnosed with myelodysplastic syndrome, progressing to the early phase of AML M6-Er. The patient died on day 121 despite multidrug treatments. Postmortem examination revealed neoplastic erythroblasts infiltrating the bone marrow and other organs. Neoplastic cells were immunopositive for CD71 but immunonegative for CD3, CD20, granzyme B, von Willebrand factor, CD61, myeloperoxidase, and Iba-1. Although further studies are necessary for the application of CD71, our results supported the morphologic diagnosis of AML M6-Er.

Feline acute myeloid leukemia, subtype acute erythroid leukemia with erythroid dominance (AML M6-Er), is most frequently associated with feline leukemia virus (FeLV) infection and is often preceded by myelodysplastic syndrome (MDS).⁷ Only one previous report has described AML M6-Er in a retrovirus-negative cat.⁷ The underlying mechanism of AML M6-Er induction is unclear. Leukemia is classified as AML M6 when early erythroid precursors, such as rubriblasts, prorubricytes, and basophilic rubricytes, account for >50% of marrow cells and myeloblasts and monoblasts combined are >30% of all nucleated cells (ANC). When early blast cells including rubriblasts exceed 30% of ANC, leukemia is classified as AML M6-Er in veterinary medicine.¹² According to the latest World Health Organization (WHO) classification for humans, leukemia is classified as AML–not otherwise specified (NOS), subtype acute erythroid leukemia (pure erythroid type), when immature erythroid precursors are >80% of the marrow cells with >30% proerythroblasts, and a myeloblastic component of <20%.² Distinguishing AML M6-Er from other AMLs is difficult. Some reports have described the use of the transferrin receptor CD71 (clone 10F11, 1:400; Novocastra;⁶ and clone H68.4, dilution 1:1,000; Invitrogen¹³) in paraffin-embedded bone marrow biopsy specimens in humans.^6,13 CD71 mediates the uptake of transferrin-iron complexes and is highly expressed on the surfaces of erythroid lineage cells.^6,13 In veterinary medicine, immunohistochemical diagnostic markers for AML M6-Er have not yet been identified, although a few reports have evaluated the expression of CD71 in immature erythroid cells or neoplastic cells in a cat with acute erythroid leukemia by flow cytometry.^1,15 We describe here the clinicopathologic, histopathologic, and immunohistochemical features of AML M6-Er in a retrovirus-negative cat. We used CD71 antibody to support the immunohistochemical distinction of AML M6-Er from other AMLs.

A 4-y-old spayed-female Scottish Fold cat was presented with lethargy, anorexia, and fever. Physical examination of the patient identified fever (40.8°C), pale mucous membranes, mildly labored breathing, and normal activity. Complete blood count results on day 1 (Procyte DX; Idexx) with manual differential indicated severe normocytic, normochromic, and nonregenerative anemia, mild thrombocytopenia, and mild monocytosis (Table 1). Monocytosis could have been attributed to mild chronic inflammation or tissue necrosis. Many immature erythroid precursors, including basophilic rubricytes, prorubricytes, and rubriblasts, were detected on a Wright-Giemsa–stained (Muto Pure Chemicals) peripheral blood smear (Fig. 1A–C). The percentage of erythroid precursors among nucleated cells was 47.9% in the peripheral blood smear. Moreover, apparent binucleation of polychromatophilic rubricytes (Fig. 1D), and abnormal mitotic figures (Fig. 1E), were also found sporadically.

Table 1.

Initial complete blood count for a FeLV/FIV-negative 4-y-old spayed female Scottish Fold cat with acute myeloid leukemia, subtype acute erythroid leukemia with erythroid dominance (AML M6-Er).

Analyte	Unit	Result	Reference interval
RBC	×1012/L	1.9	6.5–12.2
Hb	g/L	28	98–162
Ht	L/L	0.07	0.30–0.52
MCV	fL	38	36–53
MCHC	g/L	389	281–358
Platelets	×109/L	103	151–600
Reticulocytes	×109/L	5	3–50
TP	g/L	67	57–89
WBC	×109/L	15.2	2.9–17.0
Band neutrophils	×109/L	0	0–0.3
Segmented neutrophils	×109/L	10.0	1.5–10.3
Lymphocytes	×109/L	3.2	0.9–6.9
Monocytes	×109/L	1.8	0.0–0.7
Eosinophils	×109/L	0.2	0.2–1.6
Basophils	×109/L	0	0.0–0.3
Nucleated erythrocytes	×109/L	14.0	ND
Rubriblasts	×109/L	0.1	ND
Prorubricytes	×109/L	0.2	ND
Basophilic rubricytes	×109/L	4.5	ND
Polychromatophilic rubricytes	×109/L	5.2	ND
Metarubricytes	×109/L	3.9	ND

FeLV = feline leukemia virus; FIV = feline immunodeficiency virus; Hb = hemoglobin; Ht = hematocrit; MCHC = mean corpuscular hemoglobin concentration; MCV = mean corpuscular volume; ND = not determined; RBC = red blood cell; TP = total protein; WBC = white blood cell. All variables were determined using Idexx Procyte DX except the manual differential count. WBC count was corrected for nucleated erythroid cells. Boldface numbers indicate that the results are out of the reference intervals.

Figure 1.

Peripheral blood smear from a FeLV/FIV-negative 4-y-old spayed female Scottish Fold cat with acute myeloid leukemia, subtype acute erythroid leukemia with erythroid dominance (AML M6-Er). Many immature erythroid precursors including rubriblasts, prorubricytes, and basophilic rubricytes were present. Wright–Giemsa stain. A. A rubriblast with a single prominent large nucleolus. B. A prorubricyte. C. A binucleate basophilic rubricyte. D. An apparently binucleate polychromatophilic rubricyte. E. An abnormal mitotic figure in an erythroid precursor.

Blood coagulation analysis revealed prolongation of activated partial thromboplastin time (APTT; 87.5 s, reference interval [RI]: 20.0–42.0 s) and increased fibrinogen levels (2.64 g/L, RI: 1.20–2.40 g/L). The cause of the prolonged APTT was unknown and was not pursued. The presence of hyperfibrinogenemia indicated inflammation. A chemistry panel revealed mild hypokalemia (2.8 mmol/L, RI⁹: 3.5–5.8 mmol/L) as the only relevant abnormality. The cause of hypokalemia was speculated to be the result of decreased intake or loss, possibly as a result of renal disease. Urinalysis of samples collected by cystocentesis revealed urine specific gravity of 1.044, elevated levels of protein, bilirubin, and heme via dipstick evaluation, and cocci were found in the sediment; inflammatory cells were not found in the sediment. Proteinuria could have resulted from the dysfunction of proximal renal tubules or glomerular leakage. The presence of bacteria was potentially indicative of subclinical bacteriuria; however, the presence of cocci was not confirmed by bacterial culture. Mild bilirubinuria suggested possible cholestasis or early hepatobiliary disease. Heme was detected, and its cause was unclear.

The results of a FeLV/feline immunodeficiency virus (FIV) snap test (Idexx), a PCR assay for proFeLV/proFIV on the peripheral blood (the primer sequences are proprietary to Canine-lab Corp.), and a panel for feline vector-borne diseases including Anaplasma spp., Bartonella spp., Cytauxzoon spp., Ehrlichia spp., and Mycoplasma spp. (RealPCR panel; Idexx) were all negative.

Thoracic and abdominal radiographs on day 1 revealed no remarkable changes. However, abdominal ultrasound revealed splenomegaly and splenic lymphadenopathy. A splenic lymph node specimen obtained by fine-needle aspiration consisted mostly of early erythroid precursors, including rubriblasts, prorubricytes, and basophilic rubricytes (Fig. 2A). A few residual lymphocytes were also observed in the specimen, as well as many mitotic figures, occasional binucleation (Fig. 2B), and abnormal mitotic figures (Fig. 2C). These findings suggested a neoplastic change in the erythroid lineage and its infiltration into extramedullary tissues.

Figure 2.

Fine-needle aspiration smear of the splenic lymph node from a FeLV/FIV negative 4-y-old spayed female Scottish Fold cat with acute myeloid leukemia, subtype acute erythroid leukemia with erythroid dominance (AML M6-Er). Wright–Giemsa stain. A. The lymph node consists of mostly early erythroid precursors, including rubriblasts, prorubricytes, and basophilic rubricytes, and there are a few residual lymphocytes. B. A binucleate or bilobed prorubricyte. C. An abnormal mitotic figure.

Bone marrow aspiration biopsy was performed on both proximal humeri on day 5, which yielded high cellularity marrow particles on the preparation of cytology smears. The sample was slightly diluted with peripheral blood, but marrow particles were comprised almost entirely of hematopoietic cells; rubriblasts, prorubricytes, and basophilic rubricytes dominated the sample (Table 2, Fig. 3A). Prussian blue staining of the bone marrow revealed occasional siderotic metarubricytes and rubricytes (Fig. 3B). Dwarf megakaryocytes were occasionally detected (Fig. 3C) with 2.9 megakaryocytes per low-power field using the 10× objective. Binucleate erythroid cells were observed sporadically (Fig. 3D), along with a few polychromatophilic erythrocytes. No significant dysplastic changes were detected in the granulocytic series. The erythroid series predominated (68.4%); rubriblasts accounted for <20% (3.6%) of ANC and 5.3% of all erythroid cells. Based on these findings, we diagnosed MDS with erythroid predominance (MDS-Er). Moreover, based on the findings of the infiltration of the dysplastic cells into the extramedullary tissues, the disease was considered to be progressing to an early phase of acute myeloid leukemia (AML) M6-Er.

Table 2.

Results of a bone marrow aspiration 500-cell differential count from a FeLV/FIV-negative 4-y-old spayed female Scottish Fold cat with acute myeloid leukemia, subtype acute erythroid leukemia with erythroid dominance (AML M6-Er).

Parameter	%	Reference interval9
Erythroid series	68.4
Rubriblasts	3.6	0–0.8
Prorubricytes	19.0	0–1.6
Basophilic rubricytes	21.0	1.6–6.2
Polychromatophilic rubricytes	14.0	8.6–23.2
Metarubricytes	10.8	1.0–10.4
Myeloid series	11
Myeloblasts	0.6	0–0.4
Promyelocytes	0.2	0–3.0
Myelocytes	2.0	0.6–8.0
Metamyelocytes	1.2	4.4–13.2
Bands	4.6	12.8–16.6
Neutrophils	2.4	6.8–22.0
Eosinophils	0	0.8–3.2
Myeloid:erythroid ratio	0.16:1	1.21–2.16
Monocytes	0	0.2–1.6
Macrophages	0.8	0–0.2
Lymphocytes	19.8	11.6–21.6

FeLV = feline leukemia virus; FIV = feline immunodeficiency virus. Boldface numbers indicate that the results are out of the reference intervals.

Figure 3.

Bone marrow aspiration smear from a FeLV/FIV-negative 4-y-old spayed female Scottish Fold cat with acute myeloid leukemia, subtype acute erythroid leukemia with erythroid dominance (AML M6-Er). Wright–Giemsa stain. A. The marrow particles are highly cellular and consisted mostly of erythroid precursors. Immature erythroid cells, including rubriblasts, prorubricytes, and basophilic rubricytes, dominated. There are a few polychromatophilic erythrocytes. No significant dysplastic changes are seen in the granulocytic series. B. Occasional siderotic metarubricytes and rubricytes are present. Prussian blue stain. C. A dwarf megakaryocyte. D. A binucleate polychromatophilic rubricyte.

Stored whole-blood transfusion and bone marrow aspiration were performed on days 1 and 5, respectively (Fig. 4). Repeated blood transfusions were performed when packed cell volume (PCV) decreased. Prednisolone (5 mg q24h PO; Kyoritsu Seiyaku), dalteparin sodium (100 IU/kg q12h SC; Sawai Pharmaceutical), menatetrenone (5 mg q24h PO; Eisai), and calcitriol (0.01 µg/kg q24h PO; Hisamitsu Pharmaceutical) were administered from day 23. Menatetrenone and calcitriol were administered for differentiation induction therapy.⁸ Although there was a sustained response to blood transfusions with a slow decrease in PCV between days 30 and 63, the patient’s condition worsened, and increased numbers of immature erythroid cells were observed in the peripheral blood, reaching a high on day 91 (105 × 10⁹/L). Low-dose cytosine arabinoside therapy (20 mg/m² q24h SC; Nippon Shinyaku) was administered on day 87 for 2 wk. Despite these multidrug treatments, the patient died at home on day 121.

Figure 4.

The course of treatment and selected parameters for a FeLV/FIV-negative 4-y-old spayed female Scottish Fold cat with acute myeloid leukemia, subtype acute erythroid leukemia with erythroid dominance (AML M6-Er). Stored whole-blood transfusion and bone marrow aspiration were performed on days 1 and 5, respectively. Although there was a sustained response to blood transfusions with a slow decrease in packed cell volume (PCV) between days 30 and 63, the patient’s condition worsened, and increased numbers of immature erythroid cells were observed in the peripheral blood, reaching a maximum on day 91 (105 × 10⁹/L). The patient died at home on day 121.

Postmortem examination on day 121 revealed marked hepatomegaly, splenomegaly, pulmonary edema, cardiomegaly, jaundice, pancreaticoduodenal lymph node enlargement, and a small amount (~50 mL) of abdominal fluid. Follicles were prominent on the cut surface of the spleen. Lung lobes were swollen, and foam oozed from the cut surface. Impression smears of the bone marrow (Suppl. Fig. 1A), spleen (Suppl. Fig. 2A), and liver (not shown) contained many early erythroid precursors. In the bone marrow, neoplastic discrete round cells dominated. Among them, ~35% of the neoplastic cells had morphologic features of rubriblasts. The neoplastic cells had a high N/C (nucleus-to-cytoplasm) ratio, large round nuclei with 1 or 2 prominent large nucleoli and finely clumped chromatin, and small-to-moderate amounts of deep-blue cytoplasm. Binucleate or bilobed nuclei were seen frequently (Suppl. Figs. 1B, 1C, 2B, 2C). The myeloid-to-erythroid ratio was 0.04:1. Neoplastic cells were also observed in the direct and sedimented abdominal fluid smears (Wright–Giemsa stain).

All organs were fixed in 10% phosphate-buffered formalin solution, processed routinely, and stained using hematoxylin and eosin. Histologic examination revealed that the neoplastic cells dominated the bone marrow and invaded vessels and various organs, particularly the liver and spleen. Tumor emboli were frequently observed in the veins of the lung (Fig. 5), heart, spleen, liver, pancreas, mesentery, and mesenteric lymph node. These round-to-oval neoplastic cells had small amounts of basophilic cytoplasm and large round nuclei with large prominent single nucleoli (Fig. 6). Moderate anisokaryosis and anisocytosis were detected in these cells. The mitotic count was 28 per 2.4 mm². Neoplastic cells filled the red pulp of the spleen and infiltrated the endothelium of the splenic trabecular arteries (Fig. 7). Neoplastic cells also expanded the sinuses in the liver, compressing hepatocytes (Fig. 8). Diffuse alveolar damage and pulmonary edema were noted; endothelial injury by neoplastic cells via physiologic or biochemical mechanisms was most likely the cause of the alveolar damage. Myocardial interstitial edema and scattered myocardial necrosis were present. Multifocal mild acute tubular injury and lymphoplasmacytic interstitial nephritis were present. Mild lymphoplasmacytic infiltration was found in the small intestine. Hypoxia and multiorgan failure were likely the cause of death.

Figures 5–8.

Representative histopathologic findings of the lung, bone marrow, spleen, and liver from a FeLV/FIV-negative 4-y-old spayed female Scottish Fold cat with acute myeloid leukemia, subtype acute erythroid leukemia with erythroid dominance (AML M6-Er). Hematoxylin and eosin stain. Figure 5. Lung: a tumor embolus in a vessel. Figure 6. Bone marrow: discrete neoplastic round cells dominate the bone marrow. These neoplastic cells are round-to-oval and have small amounts of basophilic cytoplasm and large round nuclei with large prominent single nucleolus; moderate anisokaryosis and anisocytosis; 28 mitoses per 2.4 mm². Figure 7. Spleen: neoplastic cells filled the red pulp and infiltrated the endothelium of splenic trabecular arteries. Figure 8. Liver: neoplastic cells filled the sinuses, compressing hepatocytes.

Figure 10.

Immunohistochemistry for CD71 in the bone marrow of a FeLV/FIV-negative 4-y-old spayed female Scottish Fold cat with acute myeloid leukemia, subtype acute erythroid leukemia with erythroid dominance (AML M6-Er). Mayer hematoxylin counterstain. A. Neoplastic cells are immunopositive for CD71. A megakaryocyte is immunonegative for CD71. B. The membrane and cytoplasm of neoplastic cells are immunopositive for CD71. Myeloid lineage cells (arrows) are negative for CD71. C. Lymphocytes (arrowheads) are immunonegative for CD71.

Immunohistochemistry (IHC) was performed using primary antibodies for CD3, CD20, granzyme B, von Willebrand factor, CD61, myeloperoxidase, Iba-1, and CD71 (Suppl. Table 1). After the inactivation of endogenous peroxidase and nonspecific reactions following antigen retrieval, primary antibodies were applied to tissue sections. Tissue sections were treated with secondary antibodies following the manufacturer’s instructions: CD3, CD20, granzyme B, and von Willebrand factor (Super Sensitive one-step polymer-HRP IHC detection system; BioGenex); myeloperoxidase (Mouse-on-canine HRP-polymer; Biocare Medical); CD61, Iba-1, and CD71 (Histofine simple stain MAX PO; Nichirei Biosciences). The sections were developed with 3,3′-diaminobenzidine, with Mayer hematoxylin as a counterstain. Primary antibodies were replaced with phosphate-buffered saline to produce negative controls. All antibodies stained positive and negative controls as expected. Validation of the specific CD71 clone used in our study was confirmed by immunohistochemical analysis performed with normal bone marrow tissue from a healthy cat. The validation was performed twice at 2 different institutions, using the same clone and the same procedure, and the antibody stained positive and negative controls as expected. In the healthy tissue, all stages of erythroid precursors were positive for CD71 (Fig. 9A–C), but mature erythrocytes and polychromatophilic erythrocytes, myeloid cells, lymphocytes, and megakaryocytes were negative for CD71. Approximately 41% of the cells in the normal feline bone marrow were immunostained with anti-CD71. In the patient’s tumor tissue, all neoplastic cells were positive for CD71 (Fig. 10A–C) but negative for CD3, CD20, granzyme B, von Willebrand factor (Suppl. Fig. 3A), CD61 (Suppl. Fig. 3B), myeloperoxidase (Suppl. Fig. 3C), and Iba-1 (Suppl. Fig. 3D). The proportion of CD71-positive neoplastic cells was 85.2% of ANC in the patient’s bone marrow, and the morphologic features of ~35% of the neoplastic cells were similar to those of rubriblasts. Based on these findings, the diagnosis of AML M6-Er was made.¹²

Figure 9.

Immunohistochemistry (IHC) for CD71 in normal feline bone marrow. Mayer hematoxylin counterstain. A. The membrane and cytoplasm of all stages of erythroid precursors are immunopositive for CD71. Polychromatophilic erythrocytes and erythrocytes, megakaryocytes, myeloid lineage cells, and lymphocytes are immunonegative for CD71. B. The membrane and cytoplasm of erythroid precursors are immunopositive for CD71. Myeloid lineage cells (arrows) and lymphocytes (arrowhead) are immunonegative for CD71. C. Negative control of IHC for CD71 in normal feline bone marrow. There is no nonspecific reaction.

Feline AML M6-Er is most frequently associated with FeLV infection,^4,7,11,14 but can occur in the absence of FeLV as a result of spontaneous genetic mutations. Initially, our case was diagnosed as MDS-Er, and many early erythroid precursors were observed in the peripheral blood, spleen, and splenic lymph node smears. These findings suggested that MDS-Er was progressing to an early phase of AML M6-Er. The mechanisms by which AML M6-Er is induced without FeLV infection are unclear; however, AML M6-Er should be included as a differential diagnosis when immature erythroid lineages are observed in the peripheral blood smear, even in retrovirus-negative cats.

In our case, the definitive diagnosis was supported by immunohistochemical analysis. The anti-CD71 antibody is highly expressed on the surfaces of erythroid lineage cells. In humans, the usefulness of CD71 in paraffin-embedded bone marrow biopsy specimens has been examined, and it is suggested that CD71 is selectively expressed at high levels in erythroid precursors, including those at early maturation stages, and is not expressed in mature erythrocytes, such as polychromatophilic erythrocytes and erythrocytes.^6,13 In our case, neoplastic cells were positive only for CD71 antibody and were negative for other antibodies. These results suggest that CD71 is applicable as an erythroid marker for IHC in veterinary medicine. The nuclei of the neoplastic cells in the bone mallow were quite pleomorphic. This is an unusual feature for the erythroid lineage, which indicates the potential utility of CD71-positive staining to confirm erythroid lineage. In our case, the proportion of neoplastic cells exceeded 80%, and the proportion of the neoplastic cells that had similar morphologic features to those of rubriblasts was ~35%. Although AML M6-Er remains conventionally used in veterinary medicine,¹² in an attempt to apply the latest WHO criteria for hematopoietic tumors in humans, we suggest that the diagnosis of AML-NOS, subtype erythroid leukemia (pure erythroid type; AML M6-Er)² can be applied based on our findings.

In our case, blood transfusion enabled the cat’s survival for 121 d. Further research is warranted to clarify the pathogenesis of AML M6-Er, given that the PCR assay for proFeLV/proFIV was performed on EDTA whole blood, and the retroviral cause was not conclusively ruled out. From our findings, CD71 appears promising for confirmation of an erythroid lineage origin for atypical neoplastic cells; however, additional testing in cats and other species with AML is warranted to see how useful CD71 will be as a diagnostic marker for erythroid leukemia. A report suggested that Iba-1 might be a marker for feline histiocytes,¹⁰ but Iba-1 has not been validated as a monocytic marker in AML and has only been used in equine cases,^3,5 therefore additional tests are needed.