Radiol Oncol 2004; 38(3): 217-25. Diagnosis and classification of spontaneously developed and radiation-induced murine haematopoietic neoplasms. The murine models for the research on the human haematopoietic neoplasms Hanna Szymańska1, Joanna Piskorowska1, Elnbieta Krysiak1, Henryk Skurzak2, Alina Czarnomska1, Peter Demant3 1Department of Genetics and Laboratory Animal Breeding, The Maria Skłodowska-Curie Memorial Cancer Center and Institute of Oncology, Warsaw, Poland; 2Department of Immunology, The Maria Skłodowska-Curie Memorial Cancer Center and Institute of Oncology, Warsaw, Poland; 3Department of Cellular and Molecular Biology, Roswell Park Cancer Institute, Buffalo, N.Y., USA The Haematopathology Subcommittee of Mouse Models of Human Cancer Consortium (MMHC) proposed a classification that can be readily compared with the human WHO classification 2001 1 and appropriate-ly delineates the diseases that occur in mice. The mouse lymphoid and nonlymphoid neoplasms develop spontaneously in certain strains and in genetically engineered mice (GEM) or follow induction with ionis-ing radiation or chemical carcinogens or viruses. In the study, the haematopoietic neoplasms that developed in the three investigated mouse strains were identified according to the above classification. They can be use-ful as mouse models of human lymphoid and nonlymphoid haematopoietic neoplasms. Key words: T-cell lymphoma, B-cell lymphoma; models, mice Received 18 March 2004 Accepted 29 July 2004 Correspondence to: Hanna Szymańska, PhD, The Maria Skłodowska-Curie Memorial Cancer Center and Institute of Oncology, Department of Genetics and Laboratory Animal Breeding, Roentgen Str 5, Warsaw, Poland; E-mail: hanszym@yahoo.com This paper was presented at the “3rd Conference on Experimental and Translational Oncology”, Kranjska gora, Slovenia, March 18-21, 2004. Introduction Uniform classification of tumours of murine haematopoietic system has been extensively studied.1,2 The precise diagnosis will make it possible to compare and contrast murine dis-eases with human lesions and enable model-ling human haematopoietic neoplasms in mice. Classifications of murine haematopoietic neoplasms has been changing over the last decades. The first classification formulated 218 Szymańska H et al. / Diagnosis and classification of murine neoplasms by Dunn (1954) was based on the morpholo-gy of neoplastic cells and showed similarities to Rappaport's (1966) nomenclature of human lymphomas.3,4 In 1981, Pattengale and Taylor proposed a histopathological and immunological scheme of murine lymphomas based on the concept stating that human and murine lymphoid cells of neoplasms represent neoplastic con-version of T or B cell lineage.5,6 That classifi-cation showed a very close relation to Lukes and Collins' (1974) and Kiel's classification (1981).7,8 Pattengale and Taylor adopted the term of "lymphoid neoplasm" for all haemato-poietic neoplasms containing transformed cells that have fully or partially differentiated into T-cells or B-cells or natural killer cells and showing monoclonal proliferation. The latest classification of murine lym-phoid neoplasms was recommended by Haematopathology Subcommittee of MMHCC (Mouse Models of Human Cancers Consor-tium) and published by Morse et al. (2002).9 It is worth stressing that the classification can be compared with the latest WHO classifica-tion of human haematopoietic neoplasms.1 The classification is also known as "Bethesda proposals for classification of lymphoid neo-plasms in mice".9 In 2002, the same Subcommittee of MMHCC formulated the latest classification for murine nonlymphoid haematopoietic neo-plasms. It was published by Scott C. Kogan et al. and it is known as "Bethesda proposals for classification of nonlymphoid neoplasms in mice".10 The term "nonlymphoid haematopoi-etic neoplasm" was adopted for haematopo-ietic neoplasms arising from other lineages than lymphoid ones. The Haematopoietic Subcommittee of MMHCC recommends recognition of the following types of murine lymphoid neoplasms: B-cell neoplasms Precursor B-cell neoplasm Precursor B-cell lymphoblastic lym-phoma/leukaemia Radiol Oncol 2004; 38(3): 217-25. Mature B-cell neoplasms Small B-cell lymphoma Splenic marginal zone B-cell lymphoma Follicular B-cell lymphoma Diffuse large B-cell lymphoma - Centroblastic - Immunoblastic - Histiocyte associated Classic Burkitt's lymphoma Burkitt's-like lymphoma Plasma cell neoplasm B natural killer cell lymphoma T-cell - neoplasms Precursor T-cell neoplasms Precursor T-cell lymphoblastic lym-phoma/leukaemia Mature T-cell neoplasm Small T-cell lymphoma T-natural killer cell lymphoma T-cell neoplasm, character undetermined Large cell anaplastic lymphoma The marked types were not recognized in Pattengale and Taylor classification as separate categories. E.g., the diffuse large B-cell lymphoma was a subtype of follicular centre cell lymphoma from large cells. It is also worth pointing out that some types of murine lymphomas occur only in mice as special models. E.g. B-natural killer lymphomas can develop only in thymec-tomized (SL/Kh x AKR/Ms) F1 mice.11 The Haematopoietic Subcommittee of MMHCC recommends also four categories of murine nonlymphoid haematopoietic neo-plasms with subtypes: 1. Nonlymphoid leukemias - Myeloid leukemias (granulocytic leukemia) - Erythroid leukemia - Megakaryocytic leukemia - Biphenotypic leukemia 2. Nonlymphoid haematopoietic sarcomas - Granulocytic sarcoma - Histiocytic sarcoma - Mast cell sarcoma 3. Myeloid dysplasias 4. Myeloid proliferations (nonreactive) Szymańska H et al. / Diagnosis and classification of murine neoplasms 219 The diseases represented in that classification are not very readily compared with human lesions. It is due to the fact that some nonlymphoid haematopoietic neoplasms have not been clearly described in mice. Additional murine models of human nonlymphoid haematopoietic neoplasms are still anticipated. The purpose of this paper is to show that the murine haematopoietic neoplasms developed in our three investigated mouse strains and identified according to the latest classification can be recognized as murine models of human haematopoietic neoplasms. Material and methods Tumours The murine haematopoietic neoplasms developed: - spontaneously in recombinant congenic strain OcB/Dem - 20 cases. The OcB/Dem mice were bred from the pairs of mice sent to Cancer Centre in Warsaw from the Netherlands Cancer Institute,12 - spontaneously in AKR/W mice carrying endogenous ecotropic provirus which induced potential lymphomas - 46 cases, 13 - in back-cross (CcS17 x CcS2) x CcS/Dem mice exposed to y radiation - 72 cases. Mice were exposed to four whole-body y irradiation with the doses of 1.7 Gy at one week intervals.14 All mice were sacrificed when they were visibly sick with poor grooming, hunched posture, weight loss and enlargement of the thymus, lymph nodes or/and spleen detected by palpation. Pathological procedures Autopsies were done on each animal. The thymus, mesenteric lymph nodes, spleen, liver as well as other organs with visible neoplastic lesions were fixed in EAFS (ethanol, acetic acid, formol, 0.9% NaCl), and embed- ded in paraffin; 4 ^im thick paraffin sections were stained with H&E or prepared for im-munohistochemistry. Immunophenotyping was performed using two techniques of immunohistochemistry ABComplex and MOM® (Mouse on Mouse) -(immunodetection kit designed to localize murine primary antibodies on mouse tissue), and flow cytometry with the appropriate monoclonal antibodies. The specific monoclonal antibodies (MAbs) for flow cytometry were conjugated with FITC (CD90.1, CD90.2, CD3e, CD8, CD5, CD19, CD45R) or PE (CD4, TCRaß, RAM KAPPA - (rat anti mouse k), and for immunohistochemistry, they were biotin or pure (anti IgM, anti IgD, anti IgK, anti Igk1\2X3, Gr-1).15 All antibodies were produced by PharMingen Germany. Additionally, two histochemical stainings were performed on the air-dried imprints of tumours, of the spleen, liver or of other organs with neoplastic lesions: naphtol ASBI method - acid phosphatase focal staining is considered to be a specific marker for T lym-phoblasts and ASD method - assessment of chloroacetate esterase activity is needed if granulocytic leukemia is to be diagnosed. Blood smears were stained with Giemsa and the number of neoplastic haematopoietic cells was estimated. Results We recognized three types of lymphoid and two types of nonlymphoid haematopoietic neoplasms among 138 classified haematopoietic neoplasms: LYMPHOID NEOPLASMS: T-cell derived lymphoma: - precursor T-cell lymphoblastic lym-phoma/leukaemia B-cell derived lymphoma: - follicular B-cell lymphoma Radiol Oncol 2004; 38(3): 217-25. 220 Szymańska H et al. / Diagnosis and classification of murine neoplasms - diffuse large B-cell lyphoma subtype centroblastic (CB) NONLYMPHOID HAEMATOPOIETIC NEOPLASMS: - granulocytic leukaemia - granulocytic sarcoma Precursor T-cell lymphoblastic lymphoma Human counterpart of mouse precursor T-cell lymphoblastic lymphoma is the lymphoma of the same nomenclature. The micro-scopic characteristic of examined lymphomas was as follow: - cells were monomorphic, medium size with scant cytoplasm; - nuclei were round with fine immature chromatin; - cells exhibited numerous mitosis; - 1-2 nucleoli were placed in the centre of the nucleus; - the spleen was filled up with sheets of neoplastic lymphoid cells; - in the liver - sheets of neoplastic cells were placed in sinusoids and/or around the vessels. The detailed immunophenotypes of the ex-amined precursor T-cell lymphoblastic lymphomas are demonstrated in Table 1. The lymphomas exhibited one out of three im- munophenotypes (CD4/CD8)+; CD4+/CD8–; CD4–/CD8+. The double positive phenotype was the most frequent conversely phenotype CD4–/CD8–. The expression of CD4 or/and CD8 and CD90 demonstrated by FACS scan analyses is shown in Figures 1A, B, C. Those lymphomas were positive for acid phos-phatase staining. Representative images of T-cell lymphoblastic lymphoma are shown in Figures 2A, B, C. Follicular B-cell lymphoma Human counterpart is also follicular B-cell lymphoma. Microscopic changes were as follow: - diffuse pattern of lymphoma; - neoplastic cells were small – centrocytic or large – centroblastic; - cytoplasm was scant; - nuclei were cleaved, usually with charac-teristic "heart" shape or noncleaved, round; - 2-3 nucleoli were prominent and adher-ent to the nuclear membrane; - mitoses were visible only among centrob-lastic cells. Follicular B-cell lymphomas showed im-munophenotype of B-cells – sIg+, B220+, CD19+. Clonality was confirmed by the ex- Table 1. Immunophenotypes of tested mouse haematopoietic neoplasms Marker T-cell derived lymphomas B-cell derived lymphomas Granulocytic leukemia / Granulocytic sarcoma CD90/Thy.1. CD3e CD4 CD8 CD5 TCRaß IgM IgD IgK / RAM KAPPA IgX1X2X3 CD19 CD45R B220 Gr-1 + - positive +/– - more often positive than negative –/+ - more often negative than positive – - negative Radiol Oncol 2004; 38(3): 217-25. + –– + –– + or – – – + or – – – + –– + –– – +– – +/– – – +– – –/+ – – +/– – – +– – –+ Szymańska H et al. / Diagnosis and classification of murine neoplasms 221 Figure 1. FACS analysis: Tumor of thymus, T-cell lymphoma (A), (B), (C). Positive reaction with CD 90.1 (Thy 1.1) FITC (left) co-expression (CD4/CD8)+ or expression CD4+ CD8– or CD4– CD8+ (right). Tumor of mesenteric lymph nodes, B-cell lymphoma (D). Negative with CD90 FITC (left) and positive reaction with RAM KAPPA PE (right). Radiol Oncol 2004; 38(3): 217-25. 222 Szymańska H et al. / Diagnosis and classification of murine neoplasms Figure 2. Precursor T-cell lymphoblastic lymphoma/leukaemia (A), (B), (C). (A) Tumor of thymus - uniform population of medium sized cells with central prominent nucleoli (red arrow) and numerous mitosis (green arrow). H&E x400. (B) Tumor of mesenteric lymph nodes - focal positive acid phosphatase reaction on air-dried imprint ASBI method x1000. (C) Tumor of thymus - immunohistochemical staining on air-dried imprints positive for CD90. ABC method x1000. Follicular B-cell lymphoma (D). Tumor of mesenteric lymph nodes. Population of large centroblasts (green arrow). and cleaved "heart" shaped centrocytes (red arrow) H&E x1000. Granulocytic leukaemia (E), (F), (G) (E) Tumor of mesenteric lymph nodes – infiltration of ring forms of granulocytes (red arrow) H&E x1000. (F) Tumor of mesenteric lymph nodes – ASD positive staining for chloroacetate esterase activity on air-dried imprint. ASD method x1000. (G) Spleen – numerous ring shaped forms of granulocytes; positive reaction with Gr-1. ABC method x1000. Radiol Oncol 2004; 38(3): 217-25. Szymańska H et al. / Diagnosis and classification of murine neoplasms 223 pression of one or two heavy chains of immunoglobulin \i, ö (IgM or IgD) and one of light chains k, X (IgK or Igk). The detailed immunophenotype of follicular B-cell lymphoma is shown in Table 1. Those lymphomas exhibited the expression of RAM KAPPA and no expression of CD90 (Thy 1.) (Figure 1D). Representative image of follicular lymphoma is shown in Figure 2D. Diffuse large B-cell lymphoma The human counterpart is also diffuse large Bcell lymphoma - variant centroblastic. We classified B-cell derived lymphomas with more than 50% of neoplastic lymphoid cells as diffuse large B-cell lymphoma - centroblastic. Those lymphomas showed immunophenotype of mature B-cell lymphoma sIg+, B220+, CD19+, IgK/RAM KAPPA+. Granulocytic leukemia The most severely affected organs were the spleen and mesenteric lymph nodes. The thymus was not involved. The immature forms of granulocytes were found in the peripheral blood and in the periportal, sinusoidal liver and around the liver vessels. Infiltration into the kidney and lung was not observed in our material. Leukaemic cells were bean or, more often ring-shaped and were in one stage of maturation. Representative images of granulocytic leukemia are shown in Figures 2E, F, G. Granulocytic sarcoma The lesion was primarily a solid tumour developed in a mesenteric lymph node mass without "spillover" to the peripheral blood. Granulocytic leukaemia and granulocytic sar-coma exhibited positive staining for chloroac-etate esterase and showed the expression of Ly-6G Gr-1 in ABC method (Table 1). Assessment of murine tumours of the investigated mouse strains as a model of human haematopoiet-ic neoplasms The vast majority of examined haemato-poietic neoplasms were derived from lym-phoid lineages -127 cases, while nonlym-phoid haematopoietic neoplasms developed only in 11 cases. The occurrence of haemato-poietic neoplasms in examined strains is shown in Table 2. In all examined strains, a T-cell derived lymphoma – a precursor T-cell lymphoblastic lymphoma - was the prevalent type of lym-phoid neoplasms - 106 cases. As expected, in the AKR/W mice that type occurred in significantly higher proportion than in the other strains, due to the fact that the AKR/W mice are the most appropriate mouse models of the human precursor T-cell lymphoblastic lymphoma. Two types of B-cell derived lymphomas, follicular B-cell lymphoma and diffuse large B-cell lymphoma occurred in 21 cases. Those lymphomas developed in the OcB and irradi-ated Bc(CcS17x CcS2)xCcS2/Dem strains more frequently than in the AKR/W mice. Both mouse strains could be recommended as mouse models for studying human counter-part of those B-cell derived lymphomas. Granulocytic leukaemia and granulocytic sarcoma were diagnosed mainly in the Bc(CcS17xCcS2)xCcS2/Dem strain. Those Table 2. The occurrence of neoplasms with given immunophenotype in examined mouse strains Mouse strains No of tumours T-cell derived lymphoma B-cell derived lymphoma Nonlymphoid haematopoietic neoplasms OcB/Dem AKR/W Bc (CcS17 x CcS2) x CcS2/Dem 20 46 72 12 45 49 7 1 13 1 0 10 TOTAL 138 106 21 11 Radiol Oncol 2004; 38(3): 217-25. 224 Szymańska H et al. / Diagnosis and classification of murine neoplasms animals were prepared especially to test the genetic control of susceptibility to radiation induced B-cell derived lymphomas and gran-ulocytic leukaemia. Therefore, the nonlym-phoid haematopoietic neoplasms developed in that strain could be the potential mouse model of human neoplasms of the same nomenclature. Discussion The mouse precursor T-cell lymphoblastic lymphoma is a very appropriate and well-known counterpart of human lymphoma of the same nomenclature, observed especially in children and young adults.16 However, it should be stressed that the nuclear convolu-tion, often observed in the human lymphomas, is not seen in the analogous murine lymphomas. Those lymphomas have been extensively studied in the AKR mice. Follicular B-cell lymphomas, which devel-oped spontaneously in the OcB/Dem mice and, as a result of irradiation, in the Bc (CcS17 x CcS2) x CcS2/Dem mice, showed a diffuse pattern. While that pattern is com-mon in mice, the follicular structure is usual-ly recognised in human, and the diffuse variant is seen very rarely. Despite those differ-ences, the cytology of these murine and human lymphomas is similar. The other described type of B-cell derived lymphoma – diffuse large B-cell lymphoma in mice always requires a differential diagnosis with the progression of follicular B-cell lymphoma or progression of splenic marginal zone lymphoma.17 It is worth stressing that there is an obsta-cle in confirmation of clonality of B-cell de-rived lymphomas in mice, due to the fact that most (approximately 95%) murine light chains are the ? type. Therefore, more informative is the restriction to IgM or IgD than the restriction to light chain ?. Radiol Oncol 2004; 38(3): 217-25. There is also the difficulty in the diagnosis of granulocytic leukemia and granulocytic sarcoma because, in mice, extramedullary haematopoiesis continues in the spleen throughout life. The infiltration in the liver parenchyma by immature forms of granulo-cytes is more informative than the sheets of immature granulocytes in the spleen. Unfortunately, in our material, in some cases, the liver was not involved and there were no immature granulocytes in the peripheral blood. In those cases, the restriction to one stage of maturation of granulocytes in the spleen was the only criterion for diagnosis. Due to the fact that the investigated strains developed haematopoietic neoplasms that can be identified according to the classifica-tion proposed by the Haematopathology Subcommittee of Mouse Models of Human Cancer Consortium, they can be used as murine models of the human diseases. The observations made on these haematopoietic neoplasms can be translated to their human counterparts. References 1. Jaffe ES, Harris NL, Stein H, Vardiman J, eds. Pathology and genetics of tumours of haematopoi-etic and lymphoid tissues: WHO Classification of tumours. Lyon, France 2001: IARC Press. 2. Jaffe ES, Banks PM, Nathwani B, Said J, Swerdlow SH. Recommendations for the reporting of lym-phoid neoplasms: A report from the Association of Directors of Anatomic and Surgical Pathology. Mod Pathol 2004; 17: 131-5. 3. Dunn TB. Normal and pathologic anatomy of the reticular tissue in laboratory mice, with classifica-tion and discussion of neoplasms. J Natl Cancer Inst 1954; 14: 1281-433. 4. Rappaport H. Tumours of the haematopoietic system. [In:] Atlas of tumour pathology, Sec. 3. fasc.8. Washington 1966. 5. Pattengale PK, Taylor CR, Immunomorphologic classification of murine lymphomas and related leukemias. Proceedings of the Rodent Lymphoma Workshop. 1981: 22-3. Szymańska H et al. / Diagnosis and classification of murine neoplasms 225 6. Pattengale PK, Taylor CR, Experimental models of lymphoproliferative disease. The mouse as a model for human Non-Hodgkin's lymphomas and re-lated leukemias. Am J Pathol 1983; 113: 237-65. 7. Lukes R, Collins RD, Immunological characteriza-tion of human malignant lymphomas. Cancer 1974; 34: 1488-503. 8. Lennert K. Histopathology of Non-Hodgkin's Lymphomas: Based on the Kiel Classification. New York 1981. 9. Morse HC III, Anver MR, Fredrickson TN, et al. Bethesda proposals for the classification of lym-phoid neoplasms in mice. Blood 2002; 100: 246-58. 10. Kogan SC, Ward JM, Anver MR, Berman JJ, Brayton C, Cardiff RD, et al. 3rd. Bethesda pro-posals for classification of nonlymphoid neo-plasms in mice. Blood 2002; 100: 238-45. 11. Lu L-M, Hiai H. Mixed phenotype lymphomas in thymectomized (SL/Kh x AKR/Ms) F1 mice. Jpn J Cancer Res 1999; 90: 1218-23. 12. Demant P, Hart AA. Recombinant congenic strains - a new tool for analyzing genetic traits de-termined by more than one gene. Immunogenetics 1986; 24: 416-22. 13. Klein O, Staroselsky A, Huszar M, Hiss J, Kay S, Donin N, et al. Biological behaviour and cell prop-erties of new AKR/W lymphoma malignancy vari-ants. Tissue Cell 1998; 30: 95-103. 14. Szymanska H, Sitarz M, Krysiak E, Piskorowska J, Czarnomska A, Skurzak H, et al. Genetics of sus-ceptibility to radiation-induced lymphomas, leukemias and lung tumours studied in recombi-nant congenic strains. Int J Cancer 1999; 83: 674-8. 15. Lai L, Alaverdi N, Maltais L, Morse HC. Mouse cell surface antigens: nomenclature and im-munophenotyping. J Immunol 1998; 160: 3861-8. 16. Panke TW, Langlinais PC, Vriend J, McCue MJ. An animal model for childhood convoluted T-cell lymphoma. Am J Pathol. 1978; 92: 595-610. 17. Fredrickson TN, Lennert K, Chattopadhyay SK, Morse HC, Hartley JW. Splenic marginal zone lymphomas of mice. Am J Pathol 1999; 154: 805-12. Radiol Oncol 2004; 38(3): 217-25.