Beyond The Microscope - New Tools For The Assessment Of Hematopoietic Neoplasia

William Vernau BSc, BVMS, DVSc, PhD, Diplomate ACVP Davis CA



      Immunophenotyping refers to the application of antibodies specific for differentiation antigens of lymphocytes and accessory immune cells to identify lineages of these cells present in reactive (inflammatory) or neoplastic diseases involving the immune system. Immunophenotyping is an objective adjunct to conventional morphologic assessment. Immunophenotyping can be performed on snap frozen tissue sections, formalin-fixed paraffin sections, unfixed cytological smears, unfixed blood smears, and also on fluids (such as blood) by flow cytometry. There are numerous diseases in which the application of immunophenotyping is a crucial element in the attainment of an accurate diagnosis and therefore the provision of an accurate prognosis or institution of appropriate therapy.

      Most immunophenotyping is performed with monoclonal antibodies (Mab). Leukocyte antigen workshops have focused on clustering of Mab with similar patterns of reactivity in diverse cells and tissues. This systematic approach has led to the emergence of a common nomenclature for human antigens and their homologues in other species ("Cluster of Differentiation" or CD antigens). With few exceptions, Mab specific for leukocyte antigens label cells of diverse lineages. Therefore, to confidently identify specific cell types in lesions or in blood, it is often necessary to consider the results obtained with multiple Mab ie. a “panel” of antibodies, resulting in detection of a “constellation” of antigen expression.   


      The nomenclature and complexity pertaining to the biology of leukocyte surface molecules is intimidating.  There are currently >200 CD’s assigned in the human immune system, as well as many other defined molecules that have not yet been assigned to clusters of differentiation. Far fewer CD’s have been assigned in the canine system. Despite the large number of CD molecules identified, not all are useful in immunodiagnostics. The following glossary consists of the CD and related molecules of most importance in veterinary immunodiagnostics for which species specific monoclonal antibodies (mAb) are available. Many more specificities exist for some domestic species, but in no instance do they approach the comprehensive array of CD molecules identified in mice and humans. In most instances, these CD molecules are detectable only in unfixed cells, which can include anti-coagulated blood or bone marrow, unfixed air dried blood smears and cytological preparations and fresh tissue which has been carefully snap frozen and sectioned. In some instances, antibodies that detect epitopes resistant to formalin fixation have been developed; these are valuable reagents for the study of archived tissue in paraffin blocks and routinely processed pathological material.

CD1: Canine CD1a, CD1b and CD1c molecules have been characterized. CD1 molecules are distantly related to MHC class I; they present peptide, lipid and glycolipid antigens to T cells. CD1 molecules are expressed by cortical thymocytes, but not by mature T cells. CD1 molecules are the best markers of dendritic antigen presenting cells (APC), although subpopulations of B cells and monocytes express CD1c. The vast majority of histiocytic proliferative disorders in the dog involve dendritic APC and are best defined by expression of CD1 molecules.

T Cell Receptor / CD3: The TCR/CD3 complex is only expressed on the surface of mature T cells (and thymocytes). There are 2 types of TCR: ab and gd (heterodimeric molecule) ; each is associated with the CD3 complex (five polypeptides: CD3g, CD3d, CD3e, CD3z, CD3h), which is the signal transduction portion of the receptor complex in both TCR types (which are responsible for binding antigen). Mab specific for the CD3e component of the CD3 complex are available. CD3e expression is largely limited to mature T cells, although activated human NK cells can express CD3e in their cytoplasm. Evaluation of CD3 expression is possible in formalin fixed tissue using either a Mab (CD3-12, UC Davis) or a rabbit polyclonal antibody (A452, Dako, Carpinteria, CA) specific for a highly conserved peptide sequence from the cytoplasmic domain of CD3e from diverse species (dog, cat, horse, human etc). Demonstration of CD3 expression is one of the most useful immunophenotypic analyses currently performed in veterinary immunodiagnostics and is used for the diagnosis of T cell leukemia and T cell lymphoma. CD3 expression, with rare exceptions, confirms the presence of T cells in a lesion, although it does not distinguish ab and gd T cells. Mab specific for ab and gd T cells are available for the dog (and other species) and can be used for this purpose.

B Cell Antigen Receptor / CD79: The B cell antigen receptor complex (BCR) consists of surface Ig (sIg)(which binds antigen) complexed with 2 invariant molecules which function as signal transduction molecules (CD79a, CD79b). CD79a (MB-1) is expressed throughout all stages of B cell development and persists into the plasma cell stage (despite absent or diminished sIg on plasma cells). CD79a is a useful marker for establishing the diagnosis of B cell lymphoma and leukemia, since it is present in the BCR of all B cells regardless of the isotype of the sIg receptor. Background associated with Ig stains in tissues is also not an issue. CD79a is also useful in the diagnosis of cutaneous plasmacytoma (about 80% have focal to diffuse expression). HM57 Mab (Dako) is specific for a cytoplasmic peptide sequence (of CD79a) that is well conserved in diverse species (human and mouse) and is detectable in formalin fixed tissue sections. HM57 also has good reactivity with B cells in dog, cat and horse, although CD79a sequence is unavailable in these species.

CD4: In the dog, CD4 is expressed by MHC class II restricted T helper cells. Canine neutrophils constitutively express CD4 and in this respect differ from neutrophils of all other species for which data are available. Monocytes, macrophages and dendritic/APC can upregulate CD4 in some instances.

CD8:  CD8 is a dimeric molecule that is expressed by MHC class I restricted T cytotoxic cells. T cells usually express CD8 ab  heterodimers; although they can express CD8 aa homodimers.  A subset of natural killer (NK) cells may also express CD8 aa homodimers.

CD11 / CD18: The b2 integrins (CD11/CD18) are the major leukocyte adhesion molecule family.  The absence of b2 integrins due to mutations in CD18 results in leukocyte adhesion deficiency syndrome (LAD-I) described in humans, Irish Setter dogs, and Holstein cattle. Most leukocytes express one or more members of this family. CD18 is the b2 subunit which pairs with one of four a subunits to form a heterodimer. Therefore, positivity for CD18 indicates the presence of the b2 subunit, but does not indicate which of the 4 integrin molecules is present. The four a subunits are: CD11a (all leukocytes), CD11b (granulocytes, monocytes, some macrophages), CD11c (granulocytes, monocytes, dendritic antigen presenting cells) and CD11d (macrophages and T cells in splenic red pulp, and large granular lymphocytes). Macrophages and granulocytes express 10-fold more CD18 than do lymphocytes. Canine CD18 and CD11d are detectable with Mab in formalin fixed tissue sections. In the absence of CD3 or CD79a expression, abundant CD18 expression by large mononuclear cells (in formalin fixed, paraffin sections) is indicative of macrophage or dendritic APC differentiation. Frozen section or unfixed cytological smears stained for CD1 would be necessary for confirmation.

CD21:  CD21 is a C3dg receptor (CR2) which complexes with components of the B cell antigen receptor complex (sIg, CD79a, CD79b), CD19 and CD35 (CR1). CD21 is expressed on mature B cells and follicular dendritic cells of the germinal center. Detection of CD21 is useful in the diagnosis of B cell lymphoma and B cell leukemia.

CD45:  CD45 is the “leukocyte common antigen” family and all leukocytes express one or more CD45 isoforms. Alternate mRNA splicing generates 8 possible isoforms from 3 alternatively spliced exons. CD45RA is one of these isoforms. Canine CD45RA is detectable with Mab in formalin fixed tissue sections. CD45RA is expressed by all B cells and by 100% of B cell lymphomas involving lymph nodes. Peripheral T cell lymphomas usually occur in memory T cells in older individuals. Memory cells switch from expressing CD45RA to CD45RO. Memory cells preferentially traffick to cutaneous and mucosal sites, and T cell lymphomas (CD3+) in these sites usually do not express CD45RA. 

CD34: CD34 is a heavily glycosylated surface glycoprotein which is expressed on early lympho-hematopoietic stem and progenitor cells, small-vessel endothelial cells, embryonic fibroblasts and bone marrow stromal cells. Monoclonal antibodies specific for CD34 have been described in humans, mice and dogs. Ligands for CD34 include L-selectin (CD62-L) and E-selectin (CD62-E). Therefore, CD34 may play a role in leukocyte endothelial interactions. CD34 expression has proven useful in diagnosis of acute, immature cell leukemias of both myeloid and lymphoid types in humans. CD34 is not expressed in CLL, lymphoma or myeloma, which are malignancies of more mature cells.1 Similarly, mAb specific for canine CD34 have proven useful in the differentiation of acute lymphoid leukemia from lymphoma with leukemic involvement or chronic lymphocytic leukemia. A high proportion of acute myeloid and acute lymphoid leukemias in dogs express CD34 (see section on acute leukemia).


Background (Lessons from the human experience)

      The utility of immunophenotypic assessment of hematopoietic neoplasia has now been firmly established in people.2,3  Immunophenotypic assessment significantly impacts diagnosis, prognosis and therapy. This is because the diagnosis and precise classification of hematopoietic malignancies, both leukemias and lymphomas, by morphologic criteria alone has significant limitations.2  The basic premise of immunophenotyping is that leukemias and lymphomas are the neoplastic counterparts of subpopulations of normal lymphoid and myeloid cells.4  In general terms, they tend to maintain a constellation of antigen expression similar to their normal counterparts that reflects both their lineage and stage of maturation arrest and clonal expansion. This paradigm has not only proven very useful in the diagnosis and classification of hematopoietic neoplasia but has also assisted in the determination of normal hematopoietic ontogeny.

      For many reasons, immunophenotyping in humans has achieved a much greater level of complexity, sophistication and importance than currently exists in veterinary medicine. Nevertheless, the same principles apply and as this discipline develops in veterinary medicine, it is clear that similar useful information can be derived from immunophenotypic studies in animals. Immunophenotypic assessment has already proven very useful in the diagnosis of leukemia, lymphoma and cutaneous round cell tumors in animals.5,6

Chronic Lymphocytic Leukemia (CLL)

      Canine CLL differs markedly from human CLL. In the largest canine study to date, canine chronic lymphocytic leukemia (CLL) occurred in older dogs (mean age 9.75 years; range 1.5 - 15 years; n=73 cases).6  Blood lymphocyte counts ranged from 15,000/ul to 1,600,000/ul. Surprisingly, 73% of CLL cases involved proliferation of T lymphocytes (CD3+), and 54% of CLL cases had large granular lymphocyte (LGL) morphology.6 LGL CLL’s were almost exclusively proliferations of T cells that expressed CD8 and the leukointegrin CD11d and more frequently expressed T cell receptor (TCR)ab (69%) than TCRgd (31%). The non-LGL T cell CLL cases (19% of CLL) involved proliferation of TCRab T cells in which no consistent pattern of CD4 or CD8 expression was found. B cell CLL, based on expression of CD21 or CD79a, accounted for only 26% of canine CLL cases. These results are in marked contrast to people where greater than 95% of CLL cases involve proliferation of B lymphocytes.

Acute Leukemia.

      Similar to the experience in people, immunophenotyping also appears to be useful in the assessment of acute leukemias in dogs. In the aforementioned study of canine leukemias, a total of 38 cases of acute leukemia were also accessed and interrogated with a panel of 30 monoclonal antibodies, most of which were canine specific (the small remainder being cross reactive).6 The mean age was 7.4 years with a range of 0.5-13 years. The leukocyte counts varied from 3,300-450,000/ul (3.3-450 x 109/L, reference interval 6-17 x 109/L) with a mean of approximately 133,000/ul (133x109/L). 54% of cases had counts < 100,000/ul (100 x 109/L) and 23% of cases had counts < 50,000/ul (50 x 109/L). Probable acute myeloid leukemia (AML) accounted for 55% (21/38) of cases, B cell (CD 79a+) acute lymphocytic leukemia (ALL) accounted for 16% of cases (6/38), acute leukemia of LGL type for 18% (7/38) and acute undifferentiated leukemia (AUL) for 11% (4/31). Of the cases of LGL acute leukemia, 3 were T cell (CD3+ TCRab+CD8ab+ CD11d+ ) and 4 were considered to be most likely Natural Killer (NK) cell (CD3- CD11d+ CD8a+). When it was assessed, CD34 expression was observed in 11/12 cases (92%) of AML , 2/2 cases of B cell ALL and 0/3 cases of LGL acute leukemia. CD34 expression was not assessed in any of the cases of AUL. Similar to its chronic counterpart, LGL acute leukemia in the dog appears to be a primary splenic disease; the lack of CD34 expression in the cases of LGL acute leukemia in which it was assessed is also consistent with this conclusion.

         Many cases of acute leukemia had proliferation of cells with a primitive or immature morphology that was unhelpful in predicting possible lineage. Additionally, there appeared to be a relatively poor correlation between morphologic appearance of the neoplastic cells and immunophenotype, confirming the necessity of immunophenotyping to determine the origin or lineage of the neoplastic cells. Similar to the situation in people, routine morphologic assessment appears to have significant limitations in the diagnosis of canine acute leukemia.2,6

         Assessment of CD34 expression appears to be particularly useful when immunophenotyping canine leukemia. It can occasionally be difficult in the dog to differentiate acute lymphocytic leukemia from chronic lymphocytic leukemia, or acute lymphocytic leukemia from primary lymphoma with marked leukemic involvement, on the basis of routine historical, clinical, hematologic and morphologic information. We have found assessment of CD34 expression useful for making these important distinctions when necessary.


Lymphoproliferative disease often presents the clinician and pathologist with a diagnostic dilemma, particularly early in the course of the disease. It can be very difficult to differentiate a reactive (polyclonal) lymphoid proliferation from a neoplastic (monoclonal) one and this distinction is a fundamental prerequisite for successful therapy and patient management. This diagnostic dilemma can arise with most types of lymphoproliferative disease but especially with chronic lymphocytosis / leukemia, when it can be particularly difficult to establish a definitive diagnosis because of the indolent nature of the disease. Diagnostic uncertainty is also common with endoscopic gastro-intestinal biopsies and other “minimal” biopsies and with incipient and “mixed” cell lymphomas. While certain morphologic, architectural and immunophenotypic features may be suggestive, it is generally accepted that assessment and demonstration ofclonality by molecular genetic analysis of antigen receptor gene rearrangement provides the most objective and accurate predictor of lymphoid neoplasia.7,8,9 In human pathology, genotyping of lymphoid proliferations to assess both clonality and lineage, utilizing molecular genetic techniques, has become an extremely valuable and almost routine diagnostic adjunct. It has refined diagnostic accuracy to a level previously unimagined while simultaneously expanding the understanding of the mechanisms involved in the development and progression of lymphoproliferative disease.9

The problems associated with the use of Southern blot hybridization analysis for clonality assessment were circumvented in human pathology with the advent of the polymerase chain reaction (PCR) and its subsequent  successful adaptation for the assessment of clonality in lymphoid proliferations.9,10 The technique uses PCR to amplify the V(D)J splice junctions of the T cell receptor (TCR) or B cell receptor (BCR) gene segments in lymphocytes. The heterogeneity of N nucleotide addition between these junctions produces a unique fingerprint for any given rearrangement that provides a sensitive and specific target for PCR amplification.11 The products of a clonality PCR assay are resolved on a high resolution gel and visualized. Clonal or neoplastic proliferations produce 1 or 2 (due to rearrangement of both alleles) sharp bands on the gel while polyclonal or reactive proliferations produce a broad band or smear covering a range of product sizes (Fig. 1).11 PCR based tests are rapid, exquisitely sensitive, applicable to very small quantities of DNA (including punch biopsies and fine needle aspirates) and can be performed on formalin-fixed, paraffin-embedded tissue. These factors make it ideal as a basis for development of a reliable and convenient test for assessment of clonality. These factors also facilitate very sensitive detection of early relapse and minimal residual disease and retrospective studies on archival, formalin-fixed, paraffin-embedded material.

PCR based tests to assess clonality within lymphoid populations in the dog have been developed by several groups, although not yet routinely available, and assays to assess clonality in cats and horses are under development.6,12,13


The same major caveats apply to use of these technologies in all species. With respect to immunophenotyping, correlating the clinical outcome with specific antigens rather than the total phenotype is probably not useful. Unequivocal confirmation of neoplasia is rarely possible solely by the use of immunophenotyping. While careful assessment of clinical, morphological and immunophenotypic criteria are critical, it is generally accepted that assesssment and demonstration of clonality by molecular genetic analysis of antigen receptor genes provides the most objective and accurate predictor of lymphoid neoplasia. However, while the demonstration of clonality is strongly suggestive of neoplasia, clonality alone does not prove the neoplastic theory or necessarily imply malignancy.14,15,16 Monoclonal gammopathies have been described in asssociation with confirmed inflammatory diseases in both people and dogs.17,18,19 Similarly, benign clonal expansions of T cells have been documented in people in association with some inflammatory diseases, acute viral disease or ageing. The presence of clonality in lymphoproliferative disease must be interpreted together with the clinical and morphologic findings.16 These types of molecular genetic evaluations should be used as diagnostic adjuncts, not as replacements for more traditional methods of diagnosis. Accurate morphologic assessment must remain the cornerstone in the diagnosis of lymphoproliferative disorders.9 Therefore, optimal patient care should include the integration of immunophenotypic and clonality assessment with historical, clinical, morphologic and other (cytochemical, cytogenetic) information.

Figure 1.   Schematic depicting the principle of a polymerase chain reaction (PCR) based assay for the assessment of T cell clonality…see accompanying attachment or hard copy of Fig. 1


1.         Krause DS, Fackler MJ, Civin CI, May WS. CD34: structure, biology, and clinical utility [see comments]. Blood 87(1):1-13,1996.

2.         Davis BH, Foucar K, Szczarkowski W, Ball E, Witzig T et al. U.S.-Canadian Consensus recommendations on the immunophenotypic analysis of hematologic neoplasia by flow cytometry: medical indications. Cytometry 30(5):249-63,1997.

3.         Jennings CD, Foon KA. Recent advances in flow cytometry: application to the diagnosis of hematologic malignancy. Blood 90(8):2863-92,1997.

4.         Freedman AS. Cell surface antigens in leukemias and lymphomas. Cancer Invest 14(3):252-76,1996.

5.         Moore P et al. The use of immunological reagents in defining the pathogenesis of canine skin diseases involving proliferation in leukocytes, in Advances in Veterinary Dermatology, K. Kwotchka, T. Willemse, and C. von Tscharner, Editors. Butterworth Heinmann: Oxford. p. 77-94,1998.

6.             Vernau W, Moore PF. An immunophenotypic study of canine leukemias and preliminary assessment of clonality by polymerase chain reaction. Vet Immunol Immunopathol 69(2-4): 145-64,1999.

7.                   Griesser H, Tkachuk D, Reis MD, Mak TW. Gene rearrangements and translocations in lymphoproliferative diseases. Blood 73(6):1402-15,1989.

8.                   Knowles DM. Immunophenotypic and antigen receptor gene rearrangement analysis in T cell neoplasia. Am J Pathol 134(4):761-85,1989.

9.                   Weiss LM, Spagnolo DV. Assessment of clonality in lymphoid proliferations [comment]. Am J Pathol 142(6):1679-82,1993.

10.               Wainscoat JS, Fey MF. Assessment of clonality in human tumors: a review. Cancer Res 50(5):1355-60,1990.

11.               McCarthy KP. Molecular diagnosis of lymphomas and associated diseases. Cancer Metastasis Rev 16(1-2):109-25,1997.

12.        Burnett RC, Vernau W, Modiano JF, Swardson CJ, Moore PF, Avery AC. Diagnosis of canine lymphocytic neoplasia using clonal rearrangements of antigen receptor genes. Proceedings of the 1st International Canine Immunogenetics and Immunologic Diseases Conference. Section 1-21, Aug 1998.

13.        Dreitz MJ, Ogilvie G, Sim GK . Rearranged T lymphocyte antigen receptor genes as markers of malignant T cells. Vet Immunol Immunopathol. 69(2-4):113-119,1999.

14.        Scott CS, Richards SJ, Sivakumaran M. Disorders of large granular lymphocytes and natural killer-associated cells [letter; comment]. Blood 83(1):301-3,1994.

15.        Semenzato G, Zambello R, Starkebaum G, Oshimi K, Loughran TP Jr. The lymphoproliferative disease of granular lymphocytes: updated criteria for diagnosis. Blood 89(1):256-60,1997.

16.        Yu RC, Alaibac M. A rapid polymerase chain reaction-based technique for detecting clonal T-cell receptor gene rearrangements in cutaneous T-cell lymphomas of both the alpha beta and gamma delta varieties. Diagn Mol Pathol 5(2):121-6,1996.

17.        Breitschwerdt EB, Woody BJ, Zerbe CA, De Buysscher EV, Barta O. Monoclonal gammopathy associated with naturally occurring canine ehrlichiosis. J Vet Intern Med 1(1):2-9,1987.

18.        Diehl KJ, Lappin MR, Jones RL, Cayatte S. Monoclonal gammopathy in a dog with plasmacytic gastroenterocolitis. J Am Vet Med Assoc 201(8):1233-6,1992.

19.        Fishleder A, Tubbs R, Hesse B, Levine H. Uniform detection of immunoglobulin-gene rearrangement in benign lymphoepithelial lesions. N Engl J Med 316(18):1118-21,1987.