Cytohistological and immunological classification of canine malignant lymphomas: Comparison with human non-Hodgkin's lymphomas

A

J. Comp.Path. 1997Vol. 117, 35 59

Cytohistologieal and Immunological Classification of Canine Malignant Lymphomas" Comparison with Human Non-Hodgkin's Lymphomas C. F o u r n e l - F l e u r y , J. P. M a g n o l , P. B r i c a i r e , T. M a r c h a l , L. C h a b a n n e , A. D e l v e r d i e r * , P. A. Bryon~ a n d P. F e l m a n ~ Laboratoire d'Hdmatologie-Cytologie-Immunopathologie, Ecole V~tdrinairede Lyon, B.P 83, 69280 Marcy l'Etoile (Lyon), *Laboratoire d'Anatomie Pathologique du Sud-Ouest, 85 All~e de Brienne, B.P 81, 31901 Toulouse, and ~Laboratoire d'Hdmatologie, H~pital E. Herriot, 5 Place d'Arsonval, 69008 Lyon et Centre Hospitalier Lyon-Sud, 69310 Pierre B~nite, France

Summary Non-Hodgkin's lymphomas (NHLs) in man are on the increase. They are also common in dogs, which, as close companions of man, may constitute a useful experimental model. However, comparisons cannot be made without a reliable morphological and immunological classification of canine NHL. Canine NHLs (n= 134) were classified on the basis of fine-needle lymphnode aspirates according to the Kiel classification, and 92 were re-classified according to the Working Formulation and the updated Kiel classification, in a histological and immunological study. The immunophenotype was determined (1) in 92 cases by the use of the pan-T anti-CD3 polyclonal antibody and the pan-B anti-mbl monoclonal antibody on paraffin waxembedded tissue sections, and (2) in 47 cases by the use of a panel of polyclonal and monoclonal antibodies on fresh preparations and frozen tissue. Cytological analysis showed a predominance of high-grade lymphomas (73-9%) over low-grade lymphomas (26"1%); it also demonstrated forms not reported in other species (small-cell variants, lymphomas with macronucleolated medium-sized cells [MMCs], and polymorphic lymphomas with a centroblastic component). Histological examination revealed the rarity of follicular lymphomas (2"2% of cases), an appearance suggestive of T-cell neoplasia (8"7% of cases), and evidence that some MMC lymphomas originated in the marginal perifollicular zones. Some (26%) of the lymphomas were of the T-cell phenotype: the majority of these consisted of small-cell, low-grade lymphomas and mycosis fungoides, the rest being either highgrade pleomorphic lymphomas (mixed or large-cell) or, rarely, high-grade, small noncleaved-cell, plasmacytoid lymphomas. No lymphoma expressed a double (T and B) phenotype. This study revealed similarities with, but also notable differences from, human NHL. In particular, the MMC lymphomas may constitute an interesting equivalent of human marginal zone B-cell lymphomas. 9 1997 W.B. Saunders Company Limited

Introduction N o n - H o d g k i n ' s l y m p h o m a s (NHLs) are c o m m o n t u m o u r s in dogs, being o u t n u m b e r e d only by m a m m a r y a n d c u t a n e o u s t u m o u r s (Priester, 1979). T h e y 0021 9975/97/050035+ 25 $12.00/0

9 1997W.B. Saunders CompanyLimited

36

C. Fournel-Fleury et

al.

are highly aggressive (Squire et al., 1973; Greenlee et al., 1990; Teske et al., 1994) and their aetiology remains obscure. In contrast to similar tumours in cats, no viral cause has been demonstrated, though it has been suggested (Onions, 1980; Tomley et al., 1983). NHLs are also common in man and are on the increase (Coiffier, 1993). The dog, as a close companion of man, and subject to the same environmental influences, may prove to be a useful model for the study ofaetiology (MacEwen, 1990; Hayes et al., 1991) and therapy (MacEwen, 1990; MacEwen et al., 1992). However, given the heterogeneity of these tumours in the dog and in man, the value of such a model would depend on the availability of a reliable scheme for the classification of canine malignant lymphomas, which could be related to the schemes currently used for human lymphomas. The existing morphological classifications of canine NHLs have been based successively on the different human classifications, i.e., that of Rappaport (1966), the Working Formulation (National Cancer Institute, 1982) and the Kiel classification (Lennert et al., 1978). The various studies of canine NHLs all emphasize the rareness of lymphomas of low-grade malignancy, and in particular follicular lymphomas, but they sometimes differ as to the frequency of occurrences of the different morphological sub-types. Some such studies were purely histological (Appelbaum et al., 1984; Delverdier et al., 1988; Greenlee et al., 1990), and others cytohistological (Carter et al., 1986; Carter and Valli, 1988; Parodi et al., 1988; Teske et al., 1994). More recently, the immunophenotype has been taken into account, though this has been hampered by the small number of monoclonal markers available for the canine species. Studies of this kind are also difficult to compare in the absence of a consensus among the different authors on the specificity of the various markers used. Standardization in this area came about only recently, in July 1993 during the first Canine Leukocyte Antigen Workshop (CLAW) in Cambridge (Cobbold and Metcalfe, 1994). This Workshop demonstrated that, although the available T-cell markers (antibodies recognizing the equivalents of the CD3, CD5, CD4 and CD8 molecules [Ferrer et al., 1992; Moore et al., 1992; Chabanne et al., 1994; Cobbold and Metcalfe, 1994]), provided a satisfactory means of identifying this lineage, B-cell markers other than for surface immunoglobulins (sIg) and cytoplasmic immunoglobulins (cIg) were not readily available, except for an antibody against the CD21 molecule (Moore et al., 1990). Furthermore, there was no real equivalent to the pan-B antibodies used against the CD19 and CD20 molecules in man. Recently, a monoclonal antibody against the m b l molecule in man (Mason et al., 1991, 1992) has been shown to cross-react with numerous mammalian species, including the dog (Jones, 1993; Jones et al., 1993; Fisher et al., 1995) and can therefore be used as a pan-B marker. The first results concerning the immunophenotype of canine NHLs indicated a high percentage oflymphomas expressing a double (B and T) phenotype (Appelbaum et al., 1984; Ladiges et al., 1988; Greenlee-et al., 1990), but these findings were not confirmed by Teske et al. (1994). The latter authors found a high percentage of T lymphomas (37"9%), but their morphology usually gave no clue as to the immunophenotype. In this study, we present 134 cases of canine NHL, classified mainly on

Classification of Canine Malignant L y m p h o m a s

37

cytological data; the latter were s u p p l e m e n t e d b y histological d a t a in 92 cases. An i m m u n o l o g i c a l study was p e r f o r m e d on paraffin w a x - e m b e d d e d sections in these 92 cases, a n d o n fresh cell p r e p a r a t i o n s a n d frozen tissue in 47 cases. Because o f m o r p h o l o g i c a l discrepancies b e t w e e n canine a n d h u m a n N H L s , we used the Kiel classification, which is based on cytology, a d a p t i n g it to the canine species and establishing comparisons with the W o r k i n g F o r m u l a t i o n a n d the u p d a t e d K i e l classification ( L e n n e r t a n d Feller, 1991). Materials and Methods

Animals Cytological analysis was carried out, between 1990 and 1995, on 136 fine-needle aspirates from lymph nodes, obtained when malignant lymphoma was suspected. The animals, which were aged 4 to 14 years, had not been subjected to chemotherapy. In total, 134 cases of canine N H L were diagnosed; two cases of myeloid proliferation of monocytic origin were excluded on morphological and cytoenzymological grounds (positive naphtol ASD-acetate esterase [NASDA] reaction, inhibited by fluoride treatment). Histological analysis of a lymph node, removed surgically or post mortem, was carried o u t in 92 cases. Among these were seven cases of mycosis fungoides, from which cutaneous biopsies were also examined. An immunophenotypieal study was carried out on fresh preparations (imprint smears, or cytocentrifuge preparations of cell suspensions, or both) and frozen tissue in 47 cases, and on paraffin-wax sections in the 92 cases referred to above. In the remaining 42 cases the owners were opposed to the removal of a lymph node; diagnosis was therefore based on repeated cytological examinations of lymph node, haematopoietic bone marrow, blood, and, if applicable, abdominal, pleural or pericardial effusions, as well as on a clinical follow-up that continued until the death of the animal.

Cytological Studies Cytological smears obtained by fine-needle aspiration were air-dried, fixed, and stained by the May Granwald Giemsa (MGG) technique. The morphological classification criteria were based on cell size ("medium", "small" or "large", i.e., nucleus equal to, smaller than, or larger than two red blood cells) and on the shape of the nucleus, the density of the chromatin, the number, size and distribution of the nucleoli, the extension and basophilia of the cytoplasm, and the mitotic index. The mitotic rate was estimated in cytological specimens by scanning five fields at x 50 magnification and counting mitotic figures. A low mitotic index was defined as 0 to 1 mitosis/five fields, medium as 2 to 4/five fields, and high as >_5/five fields. The 134 canine NHLs were classified according to the Kiel classification, with some adaptations to account for morphological types not included in the human classification. After immunophenotyping, 92 cases were re-classified according to the updated Kiel classification.

Histological Studies Biopsy specimens for light-microscopical examination and immunolabelling were fixed in 10% neutral buffered formalin at room temperature for 4 8 h and processed for routine paraffin-wax embedding. Sections (1"5 gm) were cut and stained with haematoxylin and eosin (HE) and Giemsa for light-microscopical examination. If required for further characterization, a reticulin stain (Laidlaw's silver stain) was used.

38

C. F o u r n e l - F l e u r y et al.

The mitotic rate was estimated in histological specimens by scanning 10 fields at x 40 magnification and counting mitotic figures. A low mitotic index was defined as 0 to 2 mitoses/field, medium as 3 to 5/field, and high as _>6/field (Carter et al., 1986). Ninety-two NHLs were classified according to the Working Formulation, which specifies the follicular or diffuse character of a lymphomatous proliferation. Immunological Studies Immunophenotyping of 92 cases was performed on paraffin wax-embedded sections with antibodies raised against human antigens but cross-reacting with equivalent canine antigens. These antibodies were: polyclonal antibodies against CD3 as a panT marker (Ferrer et al., 1992) and against kappa (~) and lambda (X) immunoglobulin light chains (Sandusky et al., 1986); and monoclonal antibody directed against mbl as pan-B marker (Jones, 1993). Canine-specific polyclonal antibodies were used to detect G, M and A immunoglobulins (IgG Fc-specific, g-chain-specific and a-chainspecific). Immunolabelling was performed by a standard avidin-biotin immunoperoxidase procedure with Biomeda reagents (AutoProbe II 08-802X; Biomeda, Foster City, CA, USA). After pre-treatment (Table 1), the slides were covered for 10 min with a blocking reagent (tissue conditioner), and then incubated with primary antibodies under the dilution and incubation conditions listed in Table 1. After two rinses in phosphate-buffered saline (PBS), a biotinylated secondary antibody (reagent 1A) preincubated with normal dog serum was applied to each slide for 20 min at room temperature. After two washes, streptavidin-peroxidase (reagent 2) was applied for 20 min at room temperature. After two rinses in PBS, 3-3' diaminobenzidine was applied for 15 rain at room temperature as a chromogen (reagent 3). The sections were then washed with distilled water and counterstained with Mayer's haematoxylin. Non-immune serum was used as a negative control for each specimen. A human lymph node provided a positive control when antisera to human antigens were used. For immunophenotyping on fresh or frozen preparations (47 cases), four canine monoclonal antibodies (against CO5, CO4, CO8 [Moore et al., 1992; Cobbold and Metcalfe, 1994] and CO21 [Moore et al., 1990]) were added. For these cases,' a second specimen of the same lymph node was frozen in nitrogen-cooled isopentane for the preparation of frozen sections, and a third specimen was mechanically disrupted to obtain cytocentrifuge preparations which were either used immediately or stored at - 70~ Immunohistochemical examination was carried out on frozen tissue sections and cytocentrifuge preparations according to the standard avidin-biotinimmunoperoxidase procedure described above, but without any pre-treatment. Lymphomas were classified as T-cell type when they reacted positively with antiCO3 and one or more of the other T-cell markers (CD5, CO4, CD8) and negatively with anti-mb 1. They were defined as B-cell type when they reacted positively with anti-mb 1 or anti-CO21 (or both), anti-IgM, anti-IgG or anti-IgA, and negatively with anti-CO3.

Results

Cytological Studies T h e results obtained, with the Kiel classification a d a p t e d to the canine species, are listed in Table 2. Thirty-five cases (26" 1%) were classified as low-grade lymphomas, a n d 99 (73"9%) as high-grade lymphomas. O f the low-grade l y m p h o m a s , 16 belonged to the category of small-cell l y m p h o m a s , two of these being lymphoplasmacytic, one centrocytic, and 13 " v a r i a n t " forms without a n y real equivalent in h u m a n classifications. A m o n g these 13, four were of the small-cell type with irregular or n o t c h e d nuclei,

Classification of Canine Malignant Lymphomas ooco ~o~o~

>.>.

o

"n

~

~o~o~

~ ~

"~0 o

o~:o~

o~

o m ~onoooooo

-" o- o' n~

Co >.

~~ .,~

c o

o

o

= o

c o

c ~=

~

> o

~-=-

~

~

~ -

=

o ~

~

~

o

o

o 5~

~-c

c

~uz

o ~

o

o ~

o

u= u=

o

~m

=

o

~.o o= o =

= o

= o

= o

~ o

o

39

40

C. Fournel-Fleury e t

0 o~

al.

r

9

g

0

o

t~

P~

=~

"0

0

~Tx "~'~

G~

0

~g

!oo

o

o

o

o

9

0

o~ o

.~I

G

=

>..

~o oG

o

o

t~

~'n

~

~8~

~ o ~'~

~

~

~ ~ ~

"~

~

~ ~ ~o

0

~_~.~o~ 0

~ .~.~

~,~.~.~

E ~g

Classification of Canine Malignant Lymphomas

41

and noncleaved in the strict sense ("centrocytic-like") (Fig. 1); five had eccentric cytoplasm, extended and pale, often containing fine azurophilic granules ("clear cells") (Fig. 2); and four had nuclei with fine chromatin, distinctly nucleolated ("prolymphocytic-like") (Fig. 3). A single lymphoma was classified as centroblastic-centrocytic type. There were seven cases of infiltration of the lymph node by mycosis fungoides (MF), with variable morphology, showing cells with cerebriform nuclei and cells of a monocytoid appearance (Fig. 4A) with fine chromatin and large, pale cytoplasm, as well as cells of the "large granular lymphocyte" type (Fig. 4B), with pale cytoplasm, containing large azurophilic granules. A further type of lymphoma showed macronucleolated medium-sized cells (MMCs) (11 cases), characterized by a particularly prominent nucleolus in small- to medium-sized nuclei (not larger than two red blood cells, and with an average size of 1-5 red blood cells), with cytoplasm which was moderately basophilic and not highly extended; such tumours were classified in the low-grade NHL group, on the basis of cell size, low mitotic index and weakly basophilic cytoplasm (Fig. 5) (Magnol et al., 1995). Of the 99 lymphomas classified as high-grade (Table 2), 67 were centroblastic, 20 immunoblastic, one lymphoblastic, one large-cell anaplastic, four smail-noncleaved-cell (one Burkitt-type and three plasmacytoid), two pleomorphic mixed small- and large-cell, and four pleomorphic large-cell. Of the 67 centroblastic lymphomas, only two were of the monomorphic sub-type, as judged by the homogeneous proliferation of large cells with round nuclei, fine chromatin, discrete juxta-membranous nucleoli and a narrow cytoplasmic rim of intermediate basophilia. In the polymorphic sub-type, which represented the great majority (65/67) of cases, fine cytological examination showed a mixture of the following four cell components: (1) a minor, variable component of MMCs (often <20%); (2) a major component (up to 80%) of small blastic cells with round nuclei and decondensed chromatin, irregular juxta-membranous and perinucleolar chromatin stippling, prominent multiple nucleoli and a narrow hyperbasophilic cytoplasmic rim; (3) a component (20 to 50%) of medium-sized cells with nuclei having less prominent nucleoli and dispersed chromatin, resembling centroblasts; and, (4) a minor component (often <20%), of immunoblastic cells characterized by large size, large nucleolar volume (made up of a single voluminous central nucleolus, or several nucleoli of variable size) and extended hyperbasophilic cytoplasm, sometimes with a perinuclear chromophobic zone. In this type of NHL, though "starry sky" patterns were almost always observed, the mitotic index varied from moderate to very high. Because of this cellular diversity, the name "polymorphic lymphomas with a centroblastic component" was given to this category, within which we identified two sub-types, one with predominantly small cells (PSC) (Fig. 6A), which constituted a majority (43/65), and the other (22/65), with predominantly large cells (PLC) (Fig. 6B). The term "immunoblastic lymphomas" was given to the 20 cases which showed a clear predominance (>80%) of immunoblasts (Fig. 7). These were sometimes characterized by strongly plasmablastic differentiation. The mitotic index was uniformly high. This category included two unusual lymphomas characterized by a nucleus with an exceptionally voluminous central nucleolus (Fig. 8). Only one anaplastic

42

C. F o u r n e l - F l e u r y et al.

Figs. ! 6.

Classification

of Canine

Malignant Lymphomas

43

lymphoma was identified, characterized by large cell size, significant nuclear pleomorphism, the presence of binucleated cells of the pseudo-Sternberg type, and extended, strongly basophilic cytoplasm (Fig. 9A and 9B). A single lymphoblastic lymphoma was identified; it showed predominantly thymic infiltration, characterized by proliferation of small- to medium-sized cells, numerous closely-grouped mitoses, a nucleus that was sometimes convoluted, with fine chromatin and discrete nucleoli, and cytoplasm of moderate extension and basophilia. Four tumours, grouped together as small-noncleaved-cell highgrade lymphomas, comprised one with homogeneous proliferation, similar to a Burkitt's lymphoma (Fig. 10), and three of plasmacytoid appearance (Fig. 11A and 11B). All four had a high mitotic index. Finally, six pleomorphic lymphomas were identified: of these, two were mixed, with small and large cells, and four had large cells with irregular nuclei, fine chromatin and pale or moderately basophilic cytoplasm, suggesting a T phenotype (Fig. 12).

Histological Studies A parallel histological study was carried out in 92 cases; its relationship to the cytological study is given in Table 2. A follicular-type architecture, confirmed by the staining of the reticulum fibres, was found in only two of the 92 cases (2"2%)--a small-cell lymphoma classified according to the Kiel classification as a centrocytic lymphoma, and a centroblastic monomorphic lymphoma. Six cases showed a pseudo-follicular organization, although this was not confirmed by the staining of the reticulum fibers. An origin in the marginal perifollicular zones was obvious in three of the nine cases of MMC lymphoma, in which histological examination showed a marginal zone pattern (Fig. 13A). All the other lymphomas were diffuse. An architecture suggesnve of T neoplasia characterized by interfollicular proliferation (Fig. 14) and the presence of post-capillary hyperplastic venules (Fig. 15), sometimes associated with histiocytic cells and eosinophils, was found in eight of 92 cases (8"7%). Four of these were sniall-cell low-grade lymphomas (three with clear cells, one centrocytic-like), and four were lymphomas with pleomorphic cells, of which one was mixed and three had large cells. There were also two immunoblastic lymphomas with some indication of angiocentric development. Fig. 1. ' Fig. 2. Fig. 3. Fig. 4. Fig. 5.

Fig. 6.

(opposite) Centrocytic-like lymphoma (T-cell). Small cells with irregularly indented nuclear outlines or deep clefts (arrows). MGG. x 625. Clear-cell lymphoma (T-cell). Small cells with extended pale cytoplasm. MGG. x 625. Prolymphocytic-like lymphoma (B-cell). Small cells with clumped chromatin and prominent central nucleolus. MGG. x 625. Mycosis fungoides. A: Small and medium-sized cells of a monocytoid appearance with fine chromatin and abundant pale cytoplasm. Two mitoses are present. B: Large granular-lymphocyte type. MGG. x 625. Macronucleolated medium-sized cell lymphoma. Medium-sized cells with fine chromatin, particularly prominent nucleolus, and moderately extended ar/d weakly basophilic cytoplasm. MGG. x 625. Polymorphic lymphoma with a centroblastic component. A: Predominantly small-cell type (PSC). Mixture of macronucleolated medium-sized cells and small cells with prominent multiple nucleoli and centroblasts. B: Predominantly large-cell type (PLC). A majority of large cells. MGG. x 625.

44

C. F o u r n e l - F l e u r y et al.

Figs. 7 12.

Classification

of Canine Malignant Lymphomas

45

A moderately to highly pronounced starry-sky pattern was found in most of the polymorphic lymphomas with a centroblastic component, in all the immunoblastic lymphomas, in the small-cell Burkitt-type lymphoma and in the lymphoblastic lymphoma. This feature was never observed in the smallcell low-grade lymphomas, the MMC lymphomas or the mixed or large-cell pleomorphic lymphomas. All the small-cell low-grade lymphomas, and also the MMC lymphomas, had a low mitotic index (Fig. 13B). The mitotic index was high in all the blastic lymphomas (Fig. 16) except one, and also in the pleomorphic lymphomas, the highest value (>13) being found in some immunoblastic lymphomas and small-noncleaved high-grade lymphomas, and particularly in the lymphoma of the Burkitt type. The histological analysis of cell morphology emphasized the presence of the four different cell populations in the pleomorphic lymphomas with a centroblastic component and the often irregular appearance of the nucleus in proliferations suggestive of T neoplasia (5/8 cases). Table 2 gives the distribution of these 92 cases according to the Working Formulation, which lists the follicular lymphomas. Of the 92 cases of canine NHL, 48 (52"2%) were classified as being of intermediate-grade malignancy.

Immunological Studies Of the 92 cases studied in paraffin wax-embedded tissue sections, a T-cell phenotype (CD3 +, rob 1-) was found in 24 (26%), of which 15 were low-grade lymphomas and nine were high-grade lymphomas (Table 3). Of the same 24 cases, eight gave histological results suggestive of a T-cell proliferation, though the two angiocentric immunoblastic lymphomas were not among these. The low-grade T-cell lymphomas comprised eight small-cell lymphomas (out of 13 in all, i.e., 61-5%), all the variant forms with clear cells (4/4), half of the prolymphocytic-like variants (2/4), two centrocytic-like lymphomas reclassified "pleomorphic small-cell" in the updated Kiel classification (2/2), and the seven-cases of MF (7/7). The nine high-grade T-cell lymphomas included the single case of lymphoblastic lymphoma, the six pleomorphic lymphomas and two high-grade small-noncleaved-cell lymphomas of a plasmacytoid appearance (Fig. 11B). A B-cell phenotype was established in all the other 68 cases (74%) and in particular all the MMC lymphomas (Fig. 17), all the immunoblastic lymphomas, and all those with a centroblastic component. Fig. 7.

(opposite) Immunoblastic lymphoma. Large cells with a large nucleolar volume and extended

basophilic cytoplasm. MGG. x 625. Immunoblastic lymphoma. Unusual type, with an exceptionally voluminous central nucleolus. MGG. x 625. Fig. 9. Anaplastic lymphoma. A: Low magnification, showing the very large cell size and the nuclear pleomorphism. MGG. x 150. B: Pseudo-Sternberg cell at high magnification. MGG. x 625. Fig. 10. Small non-cleaved-cell high-grade lymphoma, Burkitt type. Homogeneous proliferation of small cells with prominent nucleoli and hyperbasophilic cytoplasmic rim. MGG. x 625. Fig. 11. Small-non-cleaved-cell high-grade lymphoma, plasmacytoid type. A: B-cell lymphoma, B: T-cell lymphoma. MGG. x 625. Fig. 12. Pleomorphic large-cell lymphoma (T-cell). The neoplastic cells all have irregular nuclei and pale cytoplasm, but vary in size and shape. MGG. x 625. Fig. 8.

46

Fig. 13.

C. Fournel-Fleury

e t al.

Macronucleolated mediuin-sized cell lymphoma. Histology. A: The lymph node cortical area shows highly developed marginal zones. Also visible, from the periphery inwards, are the mantle zone and the germinal centre. Deeper down, marginal zones are continuous with lymphomatous areas (arrow). HE. x 24. B: The homogeneity of the lymphomatous population is striking, as are the central, prominent nucleoli. Note the presence of an immunoblast in the centre of the field. Giemsa. x 320.

Classification of Canine Malignant Lymphomas

Fig. 14.

47

Malignant lymphoma of the paracortical areas (T-cell). Diffuse and interfollicular proliferation associated with numerous secondary follicles. HE. x 24.

Fig. 15.

T-cell lymphoma (small-cell type). Hyperplastic venules. HE. x 320.

48

C. F o u r n e l - F l e u r y e t al.

Fig. 16. High-grade immunoblastic lymphoma. High mitotic index. HE. • 320.

Table 3 Updated Kiel classification of 92 malignant lymphomas B-cell lymphomas

Low-grade malignancy Small cell Lymphocytic Lymphoplasmacytic Prolymphocytic Centrocytic Centroblastic/centrocytic Macronucleolated medium-sized

Number of cases

T-cell lymphomas

Low-grade malignancy Small cell 0 2 2 1 1 9

Clear cell Prolymphocytic Pleomorphic small cell

cell*

Mycosis fungoides

High-grade malignancy

High-grade malignancy

Centroblastic

Pleomorphic, mixed, small and large cell Pleomorphic large cell

Monomorphic Polymorphic Immunoblastic Small cell, unclassifiable Burkkt-type Plasmacytoid Lymphoblastic

1 37 13 1 1 0

Immunoblastic Small cell, unclassifiahle, plasmacytoid Lymphoblastic

* Canine malignant lymphoma without established human counterpart.

Number of cases

Classification of Canine Malignant L y m p h o m a s

Fig. 17.

49

Macronucleolated medium-sized cell lymphoma. Immunolabelling by anti-mb 1 antibody (HM57). The lymphomatous marginal zones are strongly labelled, by contrast with the weakly labelled normal germinal centres, x 200.

Sixty-three expressed an mbl +, clg-, CD3- phenotype, and five cases were cIg +, mb 1-, C D3-. Of the latter, two were cIgM + lymphoplasmacytic lymphomas associated with a serum monoclonal IgM, and three were immunoblastic lymphomas (one cIgG +, one cIgM+G +, and one c)~+ without an associated heavy chain). The Burkitt-type lymphoma was~of the B-cell type. A complementary analysis, with fresh or frozen preparations, was carried out to identify surface immunoglobulins (sIg) and equivalents of the CD21, CD5, CD4 and CD8 molecules in 47 cases (29 B-cell lymphomas and 18 Tcell lymphomas) (Table 4), including 14 low-grade NHLs (seven MF, four small-cell lymphomas, and three MMC lymphomas) and 33 cases of highgrade NHL (16 polymorphic lymphomas with a centroblastic component, eight immunoblastic lymphomas, one lymphoblastic lymphoma, two plasmacytoid small-noncleaved-cell lymphomas, and six pleomorphic lymphomas, mixed or large-cell). In the B-cell lymphomas, the surface immunoglobulins, expressed in 22 cases, were IgM in most cases (15/22, 68"2%), I g M + G in one case, IgG in four, and IgA in two. The CD21 molecule was expressed in 24 cases, namely, in all 16 lymphomas with a centroblastic component, in all three MMC lymphomas, and in the majority of the immunoblastic lymphomas (5/8, 62"5%). The three MMC lymphomas studied with fresh preparations and frozen tissue were all CD21 + and sIgM +. Furthermore, for all the B lymphomas, light chains (either intracytoplasmic or on the surface) were ~ in 85"2% of cases but 1< in only 14"8%. Most of the T-cell lymphomas were CD5 + (14/18, 77"8%) and either CD8 +

50

C. Fournel-Fleury e t

al.

cO 0

: 0 0 0 0 Ill 0 0 0 0

.

~ 0 0 0 0 0 0

zP~\ ( 0 0 . ~ 0 0

: 0 0 0 0 0 0 0

0 0 0 0 0 0

el

0

~.~ ~

.~

~

,.~

~0 , ~ ~.o ~ . ~

~

~

,-Q '-~ I~ , r ~ 0 0 0

Classification of Canine Malignant Lymphomas

51

13/18, 72"2%), CD4 + (4/18, 22"2%), or, in just one case, CD4- CDS-. The seven cases of MF were all CD3 +, CD5 +, CD8 +, and CD4-. The lymphoblastic lymphoma was CD3 +, CD5-, CD4-, and CD8-, and the two plasmacytoid small-noncleaved-cell lymphomas were, respectively, CD3 +, CD5 +, CD8 +, CD4- and CD3 +, CD5 +, CD4 +, CD8-. Three of the four pleomorphic largecell lymphomas were CD3 +, CD5-, CD8 + and CD4-, while the other was CD3 +, CD5 +, CD4 + and CD8-. No lymphomas expressed both B and T markers. The results of the morphological and immunophenotypic studies made it possible to re-classify the 92 cases according to the updated Kiel classification, adapted to the canine species (Table 3).

Discussion

In human and veterinary medicine, the value of cytological analysis by fineneedle aspiration and imprint smears in the diagnosis and follow-up of NHL is now well established (Spieler and Schmid, 1978; Van Heerde et al., 1984; Carter et al., 1988; Carter and Valli, 1988; Skoog and Tani, 1991; Caniatti et al., 1996). In this study, collection of fine-needle aspirates from 134 dogs with NHL made possible an exhaustive cytological analysis with statistical evaluation, which was more valuable than a histological study alone. The 92 cases examined histologically excluded a large number of high-grade lymphomas, for which it was possible to make a diagnosis by cytological examination alone. Using the Kiel classification, we confirmed the 'predominance of high-grade lymphomas (73"9%) over low-grade lymphomas (26"1%) in dogs, as reported previously by others (Appelbaum et al., 1984; Carter et al., 1986; Carter and Valli, 1988; Delverdier et al., 1988; Parodi et al., 1988; Greenlee et al., 1990; Baron, 1992; Teske et al., 1994). This differs from findings in man (Bryon, 1993). The cytological study revealed notable differences from human NHL, both in terms of simple morphological changes and of categories of lymphoma referred to only briefly, if at all, in the veterinary literature. Thus, among low-grade lymphomas, small-cell lymphomas were often "variant" forms of types well known in man. We identified proliferations containing small and irregular nuclei, quite distinct from the centrocytes referred to in human classifications, and also different from the centrocytes of reactive lymph nodes in dogs (Fournel-Fleury et al., 1994). These centrocytic-like lymphomas may be T-cell derived, as observed by Greenlee et al. (1990), and should not be considered to be the equivalent of human centrocytic lymphomas, which by definition are B-cell derived. (There was only one case of true centrocytic lymphoma in our series.) The clear-cell variant, which was invariably of the T phenotype, may correspond to the plasmacytoid variants noted by Greenlee et al. (1990), and probably, in part, to the immunocytomas described by Teske et al. (1994), of which 83% were T cell-derived. This illustrates the value of cytological analysis and of nuclear and cytoplasmic criteria for the prediction of the immunophenotype. We also identified a category of low-grade B

52

C. Fournel-Fleury e t

al.

lymphoma, provisionally named "lymphoma with macronucleolated mediumsized cells" (Magnol et al., 1995). Despite the presence of a prominent central nucleolus, it would seem doubtful whether these small to medium-sized cells, with their scant, weakly basophilic cytoplasm and low mitotic index, should be identified as small-sized immunoblasts, as they seem to have been by others (Carter et al., 1986; Carter and Valli, 1988; Greenlee et al., 1990; Teske et al., 1994). The presence of a comparable type of cell in the perifollicular zone of the reactive lymph node in the dog, with its similarities to the marginal zone of the human lymph node (Fournel-Fleury et al., 1995), suggests an origin in this compartment (Magnol et al., 1995). These lymphomas may therefore represent the equivalent of the lymphomas that develop in the marginal zones in man (Isaacson and Wright, 1983; Cousar et al., 1987; Isaacson and Spencer, 1987; Sheibani et al., 1988; Nizze et al., 1991). Of the high-grade lymphomas, 48-5% were centroblastic polymorphic lymphomas, all of the B-cell phenotype, similar to those described by others (Delverdier et al., 1988; Parodi et al., 1988; Greenlee et al., 1990). However, the heterogeneity of the population, the frequent predominance of small cells, the prominence of nucleoli and the frequent occurrence of a starry-sky pattern were reminiscent of the high-grade, small-noncleaved-cell lymphomas in the study of Carter and Valli (1988). In their report the percentage of lymphomas of this category was exceptionally high (34"6%). Some descriptions of smallnoncleaved non-Burkitt lymphomas in man (Grogan et al., 1982; Miliauskas et al., 1982; Pavlova et al., 1987; Raphael et al., 1991) have recorded similar findings. The frequent presence of a minor component of MMCs is reminiscent of small-noncleaved non-Burkitt lymphomas in man (Miliauskas et al., 1982). Furthermore, in centroblastic lymphomas the percentage of strict centroblasts appeared to be variable, but these cells were most often in a minority, justifying the term "polymorphic lymphomas with a centroblastic component". The percentage of immunoblasts was also variable, which explains the emergence of a sub-type with large cells, distinguished from the immunoblastic lymphomas by its high degree of intermixing with the other cell categories. Immunoblastic lymphomas, none of which were of the T phenotype, were defined, like their equivalents in the reactive lymph node (Felman et al., 1987; Fournel-Fleury et al., 1994), on the basis of large cell size and high total nucleolar volume. They were characterized by a single nucleolus, or several nucleoli in a paracentral position, and extended cytoplasm, most often hyperbasophilic, with or without plasmablastic differentiation. They accounted for 14"9% of the cases. This is somewhat lower than the values put forward in most other studies (Carter et al., 1986; Carter and Valli, 1988; Delverdier et al., 1988; Parodi et al., 1988; Greenlee et al., 1990; Baron, 1992), which average out at 24%, but is much higher than the 6% found by Teske et al. (1994). This may be explained on the one hand by the identification, in our study, of the MMC lymphomas, and on the other hand by Teske's classification of all the lymphomas with a mixed population of immunoblasts and centroblasts in the polymorphic centroblastic sub-type. We also identified four lymphomas with small noncleaved cells, of high grade and with a high mitotic index, and strikingly homogeneous in terms of cell size. One of these was a B-cell lymphoma similar

Classification of Canine Malignant Lymphomas

53

.to Burkitt's lymphoma (Lennert and Feller, 1991; Raphael et al., 1991), and possibly related to the two cases observed by Carter et al. (1986). The three others presented a plasmacytoid appearance, two of them being of the T phenotype. No equivalent to any of these seems to have been previously reported in dogs or human beings. Our study confirms the rareness of lymphoblastic lymphomas, which is in agreement with all previous observations except one (Carter et al., 1986). Finally, we identified six T-pleomorphic lymphomas, morphologically comparable with peripheral T-cell lymphomas in man (Lennert and Feller, 1991; Harris et al., 1994). The results of the histological study carried out in the 92 cases are in agreement with the cytological classifications, and emphasize three observations: the rarity of follicular lymphomas, the characteristic architecture of some T-cell lymphomas, and a marginal zone pattern in three cases of MMC lymphoma. The fact that follicular lymphomas were uncommon (2"2%), which was suspected from the infrequency of centroblastic-centrocytic lymphomas, contrasts with human data, but is in keeping with all observations made so far on dogs (Appelbaum et al., 1984; Carter et al., 1986; Carter and Valli, 1988; Delverdier et al., 1988; Parodi et al., 1988; Greenlee et al., 1990; Teske et al., 1994), and may be explained by the rapid, diffuse invasion of highly proliferating blastic forms of follicular origin (Greenlee et al., 1990; Teske et al., 1994). It should be emphasized that in our study, as in that of Greenlee et al. (1990), follicular lymphomas with small cleaved cells were rare; this contrasts, however, with the results of Teske et al. (1994), who identified small cleaved cells in 12"1% of follicular lymphomas. It should also be emphasized that the single small-cell follicular lymphoma in our study constituted the only notable discordance between the histological and cytological data. Cytologically, it was classified as a centrocytic (or mantle-cell) lymphoma on the basis of the Kiel classification, which would indicate a non-follicular proliferation with a nodular pattern, originating in the mantle cells. The presumption o f T neoplasia, based on the interfollicular nature of a proliferation or on the hyperplasia of post-capillary venules, was confirmed by the determination of immunophenotype in all cases in which this was carried out. These findings differ from those of Teske et al. (1994), who found no correlation between morphology and immunophenotype. On the other hand, a T phen 0type was not confirmed for the two immunoblastic lymphomas which showed signs of being angiocentric. The observation that three cases of MMC lymphoma originated in the marginal perifollicular zones is a further argument for a possible parallel with human lymphomas originating in marginal zone cells. Determination of immunophenotype is the basis of current classifications in human pathology (the updated Kiel classification and the Revised EuropeanAmerican Classification of Lymphoid Neoplasm [REAL classification]) (Harris et al., 1994). The detailed identification of the different surface or cytoplasmic markers supports the hypothesis of a parallel between normal cells and the lymphomatous cells that are supposed to result from a more or less complete blockage of the maturation process in normal lymphoid differentiation (Weisenburger et al., 1990; Bryon, 1993; Delsol et al., 1993). Unfortunately, the

54

C. Fournel-Fleury

e t al.

panel of monoclonal and polyclonal antibodies available for the identification of these markers in the dog is still incomplete, and the standardization of the recognized specificities of the different antibodies is very recent (see Cobbold and Metcalfe, 1994). This probably explains the disparities in the results of earlier studies. Among the T markers, the equivalents of the CD3, CD5, CD4 and CD8 molecules recognized by the corresponding heterospecific or homospecific antibodies seem to be well established. On the other hand, some antibodies, originally assumed to behave like pan-T, F3-20-7 (MacKenzie and Fabre, 1981; Teske et al., 1994) or CA1-4G8 (Moore et al., 1990; Teske et al., 1994), are not recognized as truly specific for the T lineage (Chabanne, 1994; Cobbold and Metcalfe, 1994). Finally, some antibodies, such as LQ1 (Greenlee et al., 1990), had not been classified in 1993 when the CLAW was convened. In our study, we found, among the 92 cases tested, 26% of T-cell NHLs and 74% of B-cell NHLs. Most of the T-cell lymphomas belonged to the lowgrade group (15 cases) and only nine to the high-grade group. These findings differ from those of Teske et al. (1994), who observed that 80% of the T lymphomas were high-grade. If cases of MF are excluded, the small-cell lymphomas in our study were mainly T cell-derived (61"5% of cases). This contrasts with the human data provided by the study of Harris et al. (1995), who listed 2"5% of low-grade small-cell T lymphomas and 47% of low-grade small-ceil B lymphomas. The seven cases of MF were of the CD8 + phenotype, which accords with most veterinary observations (Moore et al., 1994; Magnol et al., 1996); in man, however, MF cases are of the CD4 + phenotype (Lennert and Feller, 1991). The nine high-grade T-cell lymphomas comprised one lymphoblastic lymphoma of immature T phenotype and six lymphomas with pleomorphic cells, whose morphology and mature T-cell phenotype were comparable with those of the mixed or large-cell peripheral T lymphomas in human classifications (updated Kiel classification, REAL classification). T h e r e were also two small-noncleaved-cell lymphomas of the plasmacytoid variant, which do not appear to have a human equivalent. All the MMC lymphomas, all the polymorphic lymphomas with a centroblastic component and all the immunoblastic lymphomas were B-cell lymphomas; most were mb 1-positive (82" 7 %) and most expressed sIgM (65"5 %), which confirms their B origin, as in man. The lack of expression of the mb 1 molecule in lymphomas seems to be associated with an advanced plasmacytic differentiation and a strong expression of cIg, in spite of the fact that this molecule is expressed in plasma cells in normal human differentiation (Mason et al., 1992). Unlike Teske et al. (1994), we found no polymorphic lymphomas with a centroblastic component of the T phenotype. These were most often classified by Teske et al. (1994) in the centrocytoid sub-type, and may correspond to some of our pleomorphic T lymphomas. The non-appearance, in our series, of T-cell polymorphic lymphomas with a centroblastic component suggests a downward revision of the percentage of CD3 + lymphomas, given that around half of the polymorphic lymphomas with a centroblastic component were not subjected to a phenotypic study. The true percentage of T lymphomas would therefore seem to be less than the 26% observed, and much less than the 37"9% observed by Teske et al. (1994).

Classification of Canine Malignant Lymphomas

55

The phenotype of three cases of MMC lymphoma (all m b l +, CD21 +, sIgM +) supports the hypothesis put forward concerning a cell originating in the marginal zone, and accords with the sIgM + B phenotype of the cells in this functional compartment in human beings. Their limited proliferating power, which in our study was demonstrated by a low mitotic index and has also been confirmed by a low level of expression of the Ki-67 proliferation antigen (Fournel-Fleury et al., 1997), is also in agreement with this hypothesis, and argues against the designation "small-cell immunoblastic lymphoma", suggested by some authors. None of the lymphomas tested expressed a double (T and B)phenotype; this accords with human data and the results ofTeske et al. (1994), as opposed to other observations (Appelbaum et al., 1984; Ladiges et al., 1988; Greenlee et al., 1990). This discrepancy can probably be explained by the lack of specificity of the pan-T markers used by the latter authors. In this study, the combination of morphological and phenotypic data proved valuable, with cytology a useful complement to histological analysis. Given the rarity of follicular lymphomas, the value of the Working Formulation for the canine species appears limited. In this study, and that of Teske et al. (1994), the Working Formulation assigns the majority of canine NHLs to the intermediate grade of malignancy, which, in the light of cytological evidence, does not appear to be valid. This leads us to believe that the original and updated Kiel classifications, which are more descriptive in cytological terms, are, for the moment, the most suitable available. However, a modified REAL classification would seem potentially useful for the dog. In this, a provisional category such as " M M C lymphoma" might be equivalent to lymphoma of the marginal zone. Placing both sub-types of polymorphic B lymphomas (predominantly large-cell and predominantly small-cell) in the same category would be justified by the fact that the small noncleaved, centroblastic and immunoblastic cells, which may constitute successive evolutional stages, are always found together. Furthermore a large number of T lymphomas may be classifiable as peripheral T lymphomas with small or medium-sized and large cells. Acknowledgments

The authors thank P. F. Moore (Department of Veterinary Pathology, School of Veterinary Medicine, University of California, Davis, USA) who generously provided the monoclonal antibody against canine CD21 (CA 2-1 D6); Marie Andreani and Eliane Mari for technical assistance; and Catherine Barbet for preparing the manuscript. References

Appelbaum, F. R., Sale, G. E., Storb, R., Charrier, K., Deeg, H.J., Graham, T. and Wulff,J. C. (1984). Phenotyping of canine lymphoma with monoclonal antibodies directed against cell surface antigens: classification, morphology, clinical presentation, and response to chemotherapy. Hematological Oncology, 2, 151 158. Baron, T. (1992). Etablissement et caract~risation de lign6es lympho~'des dans les lymphosarcomes du chien. Th~se, Lyon. Bryon, P. A. (1993). Int~rfit pratique du diagnostic histopathologique des lymphomes malins. Revue du Praticien, Paris, 43, 1617-1623.

56

C. Fournel-Fleury e t

al.

Caniatti, M., Roccabianca, P., Scanziani, E., Paltrinieri, S. and Moore, P.F. (1996). Canine lymphoma: immunocytochemical analysis of fine-needle aspiration biopsy. Veterinary Pathology, 33, 204-212. Carter, R. F. and Valli, V. E. O. (1988). Advances in the cytologic diagnosis of canine lymphoma. Seminars in Veterinary Medicine and Surgery (Small Animal), 3, 167-175. Carter, R. F., Valli, V. E. O. and Lumsden, J. H. (1986). The cytology, histology and prevalence of cell types in canine lymphoma classified according to the National Cancer Institute Working Formulation. Canadian Journal of Veterinary Research, 50, 154-164. Carter, T.R., Feldman, P.S., Innes, D.J., Frierson, H.F. and Frigy, A.F. (1988). The role of fine-needle aspiration cytology in the diagnosis of lymphoma. Acta Cytologica, 32, 848-853. Chabanne, L. (1994). Les antig~nes de differenciation leucocytaires canins dbfinis /t l'aide d'anticorps monoclonaux. Applications gt l'exploration du systeme immunitaire du chien lors d'affections spontan~es. Th~se, Lyon. Chabanne, L., Marchal, T., Kaplanski, C., Fournel-Fleury, C., Magnol, J. P., Monier, J. C. and Rigal, D. (1994). Screening of 78 monoclonal antibodies directed against human leukocyte antigens for cross-reactivity with surface markers on canine lymphocytes. Tissue Antigens, 43, 202-205. Cobbold, S. P. and Metcalfe, S. (1994). Monoclonal antibodies that define canine homologues of human CD antigens: summmary of the First International Canine Leukocyte Antigen Workshop (CLAW). Tissue Antigens, 43, 137-154. Coiffier, B. (1993). Lymphomes malins non-Hodgkiniens. Revue du Praticien, Paris, 43, 1615-1616. Cousar, J., McGinn, D., Glick, A. and Collins, R. (1987). Report of an unusual lymphoma arising from parafollicular B lymphocytes or so-called "monocytoid lymphocytes". American Journal of Clinical Pathology, 87, 121-128. Delsol, G., Brousset, P. and Schlaifer, D. (1993). HistogenSse des lymphomes malins. Revue du Praticien, Paris, 43, 1627-1632. Delverdier, M., Buchet, M. and van Haverbeke, G. (1988). Histologie et cytologie des lymphomes malins canins. Etude comparative des classifications actuelles. Revue de Midecine Vitdrinaire, 139, 1141 1150. Felman, P., Gentilhomme, O., Berger, F. and Bryon, P. A. (1987). Critares cytologiques de classement des lymphomes malins non-Hodgkiniens. Revue Franfaise des Laboratoires, 164, 17-30. Ferrer, L., Fondevila, D., Rabanal, R. and Ramis, A. (1992). Detection of T lymphocytes in canine tissue embedded in paraffin wax by means of antibody to CD3 antigen. Journal of Comparative Pathology, 106, 311-314. Fisher, D.J., Naydan, D., Werner, L. L. and Moore, P. F. (1995). Immunophenotyping lymphomas in dogs: a comparison of results from fine needle aspirate and needle biopsy samples. Veterinary Clinical Pathology, 24, 118-123. Fournel-Fleury, C., Magnol,J. P., Chabanne, L., Ghernati, I., Marchal, T., Bonnefond, C., Bryon, P. A. and Felman, P. (1997). Growth fractions in canine non-Hodgkin's lymphomas as determined in situ by the expression of the Ki-67 antigen. Journal of Comparative Pathology, 117, 61-72. Fournel-Fleury, C., Magnol, J. P. and Guelfi, J. F. (1994). Atlas en Couleur de Cytologie du Cancer chez le Chien et le Chat~Color Atlas of Cancer Cytology of the Dog and Cat. Pratique Mddicale et Chirurgicale de l'Animal de Compagnie, Paris. Fournel-Fleury, C., Magnol, J.P., Marchal, T., Chabanne, L., Bryon, P.A. and Fetman, P. (1995). An original perifollicular zone cell in the canine reactive lymph node: a morphological, phenotypical and aetiological study. Journal of Comparative Pathology, 113, 217-231. Greenlee, P. G., Filippa, D. A., Quimby, F. W., Patnaik, A. K., Calvano, S. E., Matus, R. E., Kimmel, M. and Hurvitz, A. I. (1990). Lymphomas in dogs. A moi~hologic, immunologic and clinical study. Cancer, 66, 480-490. Grogan, M. D., Thomas, M., Roger, A., Warnke, M. D., Henry, S. and Kaplan, M. D.

Classification of Canine Malignant Lymphomas

57

(1982). A comparative study of Burkitt's and non-Burkitt's "undifferentiated" malignant lymphoma: immunologic, cytochemical, ultrastructural, cytologic, histopathologic, clinical and cell culture features. Cancer, 49, 1817-1828. Harris, N.L., Jaffe, E.S., Stein, H., Banks, P.M., Chan, J. K. C., Cleary, M.L., Delsol, G., De Woolf-Peeters, C., Falini, B., Gatter, K.C., Grogan, T.M., Isaacson, P. G., Knowles, D. M., Mason, D. Y., Mueller-Hermeling, H. K., Pileri, S. A., Piris, M. A., Ralfkiaer, E. and Warnke, R. A. (1994). A revised Europeanm American classification of lymphoid neoplasms: a proposal from the international study group. Blood, 84, 1361 1392. Harris, N.L., Jaffe, E.S., Stein, H., Banks, P.M., Chan, J. K. C., Cleary, M.L., Delsol, G. and 12 others (1995). Lymphoma classification proposal: clarification. Blood, 85, 857-860. H~iyes, H.M., Tarone, R.E., Cantor, K. P., Jessen, C.R., MacCurnin, D.M. and Richardson, R.C. (1991). Case-control study of canine malignant lymphoma: positive association with dog owner's use of 2,4-dichlorophenoxyacetic acid herbicides. Journal National Cancer Institute, 83, 1226-1231. lsaacson, P. and Spencer, J. (1987). Malignant lymphoma of mucosa-associated lymphoid tissue. Histopathology, 11,445-462. Isaacson, P. and Wright, D. (1983). Malignant lymphoma of mucosa-associated lymphoid tissue. A distinctive B cell lymphoma. Cancer, 52, 1410 1416. Jones, M. (1993). Peptide immunisation as a source of cross-species reactive antisera recognizing leukocyte differentiation antigens. Immunological Reagents for the Study of Disease in Companion Animals, The Wellcome Trust, London. Jones, M., Cordell, J.L., Beyers, A.D., Tse, A.G.D. and Mason, D.Y. (1993). Detection of T and B cells in many animal species using cross-reactive antipeptide antibodies. Journal of Immunology, 150, 5429-5435. Ladiges, W. C., Keast, M., Appelbaum, F. and Storb, R. (1988). Phenotypic characterization of canine lymphoma using monoclonal antibodies and a microlymphocytotoxicity assay. American Journal of Veterinary Research, 49, 870-872. Lennert, K. and Feller, C. A. (1991). Histologie des Lymphomes Malins Non-Hodgkiniens Selon la Classification de Kid Actualisie, Doin, Paris, pp. 1-307. Lennert, K., Mohri, N., Stein, H., Kaiserling, E. and Mt~ller-Hermelink, H. K. (1978). Malignant lymphomas other than Hodgkin's disease. In: Handbuch der Speziellen Pathologischen Anatomic und Histologie, Springer, Berlin, Heidelberg, pp. 1-833. MacEwen, E. G. (1990). Spontaneous tumours in dogs and cats: models for the study of cancer biology and treatment. Cancer Metastasis Revue, 9, 125-136. MacEwen, E. G., Rosenthal, R. C., Fox, L. E., Loar, A. S. and Kurzman, I. D. (1992). Evaluation of L-asparaginase: polyethylene glycol conjugate versus native Lasparaginase combined with chemotherapy. A randomized double blind study in canine lymphoma. Journal of Veterinary Internal Medicine, 6, 230-234. MacKenzie, J. L. and Fabre, J. W. (1981). Studies with a monoclonal antibody on the distribution of Thy-1 in the lymphoid and extracellular connective tissues of the dog. Transplantation, 31,275-281. Magnol, J.P., Fournel-Fleury, C., Marchal, T., Chabanne, L., Bryon, P:A. and Felman, P. (1995). Lymphomes malins/t moyennes cellules macronuclfiol6es dans l'esp6ce canine: implication d'une cellule originale de la zone marginale du ganglion lymphatique r~actionnel. Bulletin de I'Acad~mie Nationale de Mddecine, 1, 51-65. Magnol, J. P., Ghernati, I., Marchal, T., Chabanne, L., Delverdier, A. and FournelFleury, C. (1996). Donn~es cliniques, morphologiques et immunophdnotypiques &partir de 10 observations de lymphome T cutan6o-muqueux ~pith~Iiotrope du chien (analogue au mycosis fongoide). Intfirat d'un modale animal de pathologie spontan~e. Bulletin de l'Acad~mie Nationale de Midecine, 180, 449-462. Mason, D. Y., Cordell,J. L., Tse, A. G. D., van Dongen,J.J.M., van Noesel, C~J. M., Micklem, K., Pulford, K.A., Valensi, F., Comans-Bitter, W.M., Borst, J. and

58

C. Fournel-Fleury et

al.

Gatter, K.C. (1991). The IgM-associated protein rob-1 as a marker of normal and neoplastic B cells. Journal of Immunology, 147, 2474-2482. Mason, D.Y., van Noesel, C.J.M., Cordell, J. L., Comans-Bitter, W.M., Micklem, K., Tse, A.G.D., van Lier, R.A.W. and van Dongen, J.J.M. (1992). The B29 and mb-1 polypeptides are differentially expressed during human B cell differentiation. EuropeanJournal of Immunology, 22, 2753-2756. Miliauskas, J. R., Berard, C.W., Young, R.C., Garvin, A.J., Edwards, B.K. and DeVita, V. T. (1982). Undifferentiated non-Hodgkin's lymphomas (Burkitt's and non-Burkitt's types): the relevance of making this histologic distinction. Cancer, 50, 2115-2121. Moore, P. F., Olivry, T. and Naydan, D. (1994). Canine cutaneous epitheliotropic lymphoma (mycosis fungoides) is a proliferative disorder of CO8 + T cells. American Journal of Pathology, 144, 421-429. Moore, P. F., Rossitto, P. V. and Danilenko, D. M. (1990). Canine leukocyte integrins: characterization of a CD 18 homologue. 77ssueAntigens, 36, 211-220. Moore, P. F., Rossito, P. V., Danilenko, D. M., Wielenga,J.J., Raft, R. F. and Severns, E. (1992). Monoclonal antibodies specific for canine CO4 and CO8 define functional T-lymphocyte subsets and high-density expression of CO4 by canine neutrophils. 7~ssueAntigens, 40, 75-85. National Cancer Institute. (1982). The non-Hodgkin's lymphoma pathologic classification project: National Cancer Institute sponsored study of classifications of non-Hodgkin's lymphomas. Summary and description of a working formulation for clinical usage. Cancer, 49, 2112-2135. Nizze, H., C ogliatti, S., von Schilling, C., Feller, A. and Lennert, K. (1991). Monocytoid B-cell lymphoma: morphological variants and relationship to low grade B-cell lymphoma of the mucosa-associated lymphoid tissue. Histopathology, 18, 403-414. Onions, D. (1980). RNA-dependent DNA polymerase activity in canine lymphosarcoma. EuropeanJournal of Cancer, 16, 345-350. Parodi, A. L., Dargent, F. and Crespeau, F. (1988). Histological classification of canine malignant lymphomas. Journal of VeterinaryMedicine A, 35, 178 192. Pavlova, Z., Parker, J. W., Taylor, C.R., Levine, A.M., Feinstein, D. I. and Lukes, R.J. (l 987). Small noncleaved follicular center cell lymphoma: Burkitt's and nonBurkitt's variants in the US: II. Pathologic and immunologic features. Cancer, 59, 1892-1902. Priester, W. A. (1979). Epidemiology in Veterinary Cancer Medicine, G.H. Theilen and B.R. Madewell, Eds, Lea and Febiger, Philadelphia, pp. 14-32. Raphael, M., Gentilhomme, O., Tulliez, M., Bryon, P.A. and Diebold, J. (1991). Histopathologic features of high-grade non-Hodgkin's lymphomas in acquired immunodeficiency syndrome. Archives of Pathology and Laboratory Medicine, 115, 15-20. Rappaport, H. (1966). Tumors of the haematopoietic system. In: Atlas of Tumor Pathology, Section 3, Fascicle 8, Armed Forces Institute of Pathology, Washington DC, pp. 9 15. Sandusky, G. E., Horton, P.J. and Wightman, K. A. (1986). Use of monoclonal antibodies to human lymphocytes to identify lymphocyte subsets in lymph nodes of the rhesus monkey and the dog. Journal of Medicine Primatology, 15, 441 451. Sheibani, K., Burke, J., Swartz, W., Nademanee, A. and Winberg, C. (1988). Monocytoid B cell lymphoma. Clinicopathologic study of 21 cases of a unique type of low grade lymphoma. Cancer, 62, 1531 1538. Skoog, L. and Tani, E. (t991). The role of fine-needle aspiration cytology in the diagnosis of non-Hodgkin's lymphoma. Diagnostic Oncology, 1, 12-18. Spieler, P, and Schmid, U. (1978). How exact are the diagnosis and classification of malignant lymphomas from aspiration biopsy smears? PathologicalResearchPractice, 163, 232-250. Squire, R. A., Bush, M,, Melby, E, C., Neeley, L. M. and Yarbrough, B. (1973). Clinical

Classification of Canine Malignant Lymphomas

59

and pathologic study of canine lymphoma: clinical staging, cell classification, and therapy. Journal of the National CancerInstitute, 56, 565-574. Teske, E., Wisman, P., Moore, P. F. and van Heerde, P. (1994). Histologic classification and immuno-phenotyping of canine non-Hodgkin's lymphomas: unexpected high frequency of T cell lymphomas with B cell morphology. ExperimentalHematology, 22, 1179-1187. Tomley, F. M., Armstrong, S.J., Mahy, B. W.J. and Owen, L. N. (1983). Reverse transcriptase activity and particles ofretroviral density in cultured canine lymphosarcoma supernatants. British Journal of Cancer, 47, 277-284. Van Heerde, P., Go, D.M.D.S., Koolman-Schellekens, M.A. and Peterse, J. L. (1984). Cytodiagnosis of non-Hodgkin's lymphoma. HrchowsArchiv (Pathologica Anatomica), 403, 213-233. Weisenberger, D. D., Harrington, D. S. and Armitage, J. O. (1990). B-cell neoplasia. A conceptual understanding based on the normal humoral immune response. PathologyAnnual, 25, 99-115.

Received,Ju~ 23rd, 1996] Accepted,April 9th, 1997 J