Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
The mature T-cell and natural killer (NK)-cell leukemias form a heterogeneous group of diseases with diverse etiologies and markedly varied clinical behavior. When applied to T cells, the term mature or peripheral is often used to describe nonblastic T-cell neoplasms marked by their resemblance to post-thymic T lymphocytes and characterized by the absence of terminal deoxynucleotidyl transferase (TdT) expression and usually by the presence of surface CD3 expression (except in those cases in which CD3 is aberrantly deleted). CD3 is a structure of several polypeptides that is intimately associated with the surface T-cell receptor complex. It is present in the cytoplasm of early T-cell precursors but is present on the cell surface only in the post-thymic stage of T-cell maturation. The diagnosis and appropriate classification of T-cell and NK-cell leukemias rely heavily on the correlation of morphologic, immunophenotypic, genetic, and clinical data.
Thorough morphologic assessment of a carefully prepared peripheral blood smear can yield valuable clues in the diagnosis of mature T-cell leukemias. As with other lymphoproliferative disorders, the peripheral blood morphology of various T-cell leukemias is often not specific as to subtype; however, specific morphologic findings can be of use in the differential diagnosis and subclassification of these disorders. For example, the distinct morphology of large granular lymphocyte (LGL) leukemias helps to set them apart from other T-cell leukemias or lymphomas (e.g., hepatosplenic T-cell lymphoma) that can manifest with similar clinical patterns of anatomic distribution. Likewise, the bizarre, flower-like appearance of the neoplastic cells in human T-lymphotropic virus type 1 (HTLV-1)–associated adult T-cell leukemia–lymphoma (ATLL) can be contrasted with the more subtle, delicate nuclear folding of cerebriform lymphocytes in Sézary syndrome (SS)–mycosis fungoides (MF).
Although peripheral blood smear examination is an indispensable part of the diagnostic evaluation in T-cell leukemias, it is important not to get dogmatic regarding morphologic correlates in these diseases. The abnormal lymphocytes in a variety of T-cell lymphomas can mimic the cells of more specific entities such as SS or ATLL. Likewise, the LGLs synonymous with LGL leukemia can be present in a variety of reactive lymphocytoses. As with other diagnostic modalities, examination of the blood smear should be considered necessary but not sufficient for the specific diagnosis and classification of mature T-cell leukemias.
The value of flow cytometric immunophenotyping in the diagnosis and monitoring of T-cell leukemias and lymphomas is generally underappreciated, in part due to long-heralded truisms. The common wisdom states that because T cells do not have an easily assessable phenotypic measure of clonality (analogous to immunoglobulin light chain expression on B cells), flow cytometry is of limited value in the diagnosis and management of T-cell leukemias and lymphomas. This concept would be true if clonality were the only important determinant in diagnosis. In fact, flow cytometry is highly useful in the workup of T-cell neoplasms for two main reasons. First, a majority (>90% in one study) of T-cell neoplasms show at least one demonstrable immunophenotypic aberrance on flow cytometric analysis (e.g., deletion of normally expressed antigens, differences in intensity of antigen expression compared to normal T cells), providing immunophenotypic evidence of malignancy. Second, the specific pattern of aberrance is often characteristic of a particular type of T-cell lymphoma or leukemia and therefore can be of help in subclassification. Most recently, the use of large T-cell receptor Vβ family–specific antibody panels that cover most of the T-cell receptor family repertoire has made immunophenotypic analysis for restricted T-cell receptor expression (as a surrogate for T-cell monoclonality) possible. However, antibodies to some families are not yet commercially available, and interpretation of a clonal population requires marked skewing of the normal T-cell repertoire distribution. Nonetheless, a high percentage of T-cell neoplasms show some immunophenotypic aberrancy. With these concepts and the fact that peripheral blood and bone marrow are excellent substrates for flow cytometry, there is a strong case for the routine flow cytometric analysis of either blood or bone marrow in the diagnostic evaluation of T-cell leukemias.
NK cells and some cytotoxic T-cell subsets also express a class of receptors known as killer-associated immunoglobulin-like receptors (KIRs). Recently, KIR isoforms have been used in the assessment of clonality of NK or T-LGL proliferations, analogous to the use of κ and λ light chain for clonality detection in B cells or the use of Vβ isoforms for clonality in T cells.
The effective application of flow cytometry to the diagnosis of T-cell leukemias also requires knowledge of the array of normal T- and NK-lymphocyte subsets that may be found in peripheral blood and bone marrow. A lack of knowledge of physiologic immunophenotypes of these subsets may lead to overdiagnosis of abnormal T-cell populations. For example, the normal (usually minor) subset of T cells that expresses the T-γδ receptor complex manifests as a population of T cells with modally brighter CD3 than other T cells and the absence of CD4, CD5, and CD8. Physiologic expansions of this normal subset may be interpreted erroneously as immunophenotypically aberrant T-cell populations. Similarly, dermal-derived T lymphocytes may show downregulation of CD7 that could be erroneously interpreted as aberrant.
In the case of B-cell leukemias and lymphomas, clonality can be inferred by the restriction of B-cell populations to the expression of a single class of immunoglobulin light chain (either κ or λ). Unfortunately, the molecular components of the T-cell receptor complex come not in two types but in dozens, making their routine assessment challenging for diagnostic laboratories. As a result, the DNA-based assessment for clonal patterns of T-cell antigen receptor gene rearrangement has become the standard for assessment of T-cell clonality.
As noted in the previous discussion of flow cytometry, assessment of clonality per se is not always necessary in the diagnosis of either B- or T-cell neoplasms. However, because T-cell leukemias and lymphomas often present diagnostic challenges, and non-neoplastic conditions may enter the differential diagnosis, gene rearrangement studies are a relatively common part of their workup. With advances in polymerase chain reaction–based T-cell gene rearrangement studies, this once cumbersome analysis has evolved to become relatively simple, while preserving high levels of sensitivity and specificity.
The diagnostic contribution of karyotypic analysis to mature T-cell leukemias varies depending on the specific disease entity, as noted in the individual descriptions that follow. In general, the cytogenetic analysis of mature, nonblastic leukemias is challenging because the cells in these disorders often do not have high proliferative rates, and standard cytogenetic analysis relies on the isolation of dividing cells. As a result, cells in such disorders can be analyzed using either cultures stimulated by mitogens or interphase methods (i.e., fluorescence in situ hybridization) that do not require the presence of dividing cells. In some instances, such as T-cell prolymphocytic leukemia (T-PLL), cells have a high proliferative rate and yield characteristic karyotypic abnormalities, making cytogenetic analysis an important aspect of diagnosis and classification.
In the mid-1970s, there were reports of a distinct type of lymphoproliferative disorder marked by the leukemic proliferation of T cells containing prominent azurophilic cytoplasmic granules and expressing Fc receptors for the γ heavy chain of the immunoglobulin G molecule. This disorder has since been known by many names, including T-gamma lymphocytosis and large granular lymphocytosis, and it is now classified as T-cell LGL leukemia (T-LGL leukemia; Fig. 13.1 ). Although generally a leukemia of cytotoxic T lymphocytes, a subset of LGL leukemias may show a true NK-cell immunophenotype, as discussed in more detail later in this section.
Splenomegaly (common)
Rheumatoid factor positive in most patients, with frank rheumatoid arthritis in approximately one fourth of patients
Infectious complications related to neutropenia
Variable absolute lymphocyte count (but usually >4 × 10 9 /L)
Predominance of LGLs
Neutropenia (common)
May be associated with pure red cell aplasia
Lymphoid infiltrates may be scant or subtle
Despite peripheral neutropenia, marrow generally shows normal granulopoiesis
Usually the neoplastic cells are cytotoxic T cells that are positive for CD2, CD3, CD5, CD7, CD8, CD16, and CD57 and negative for CD4 and CD56
Sometimes it is CD56 + ; this has been associated with more aggressive behavior in some studies
Occasionally patients with this specific clinical syndrome will show a true NK-cell phenotype (CD2 + , CD3 − , CD5 − , CD7 −/+ , CD8 +/− , CD16 + , CD56 + , and CD57 −/+ )
Clonal cytogenetic abnormalities may be present, but no single abnormality is characteristic
Molecular studies show clonal T-cell antigen receptor gene rearrangement pattern in T-LGL leukemia. Cases with a true NK immunophenotype will not show clonal T-cell gene rearrangement pattern
Activating signal transducer and activator of transcription (STAT3) mutations in up to 75% of cases of T-LGL and 48% of NK-LGL lymphocytosis
LGL leukemia in its typical form is a clinically indolent disorder and usually manifests with a distinct array of clinical and laboratory findings. Median age at diagnosis is 60 years, with a relatively even prevalence in men and women. Although many reports stress that absolute peripheral blood lymphocyte counts may be normal at diagnosis, the median lymphocyte count in one large series was 8 × 10 9 /L (approximately twice the upper limit of normal for adults), and most patients exhibit peripheral blood absolute lymphocytosis. Splenomegaly in the absence of lymphadenopathy is common at clinical presentation. Although approximately 60% of patients with LGL leukemia have a positive rheumatoid factor, frank rheumatoid arthritis is present in approximately one fourth of cases. Neutropenia is also common at presentation in LGL leukemia and is a major cause of morbidity in this disorder. Of interest, bone marrow biopsy specimens from patients with LGL leukemia generally show adequate granulocyte precursors. The mechanism of neutropenia in this disorder has not been fully characterized, but recent data suggest increased peripheral neutrophil apoptosis, possibly linked to increased levels of circulating Fas ligand (part of the Fas apoptotic pathway) in these patients. The complications of chronic neutropenia are commonly the reason for eventual therapeutic intervention in this otherwise generally nonaggressive neoplasm. LGL leukemia also has been linked to pure red cell aplasia in some patients. In fact, some studies indicate that large granular lymphocytic leukemia may be the most common underlying disease in patients with noncongenital pure red cell aplasia.
Following the initial descriptions of this disorder, LGL leukemia was often referred to as large granular lymphocytosis, because the disease typically follows an indolent clinical course (median survival >13 years in one series), and clonality was, at the time, difficult to determine in T-cell or NK-cell populations. Since then, however, the detection of clonal T-cell receptor gene rearrangements and clonal cytogenetic abnormalities in patients with the classic clinical and laboratory features of large granular lymphocytosis has established that this syndrome generally represents a clonal neoplastic disorder. Recent reports have indicated a potential association between LGL leukemia and subclinical detectable populations of clonal B cells (so-called monoclonal B-cell lymphocytosis) or monoclonal gammopathies.
The leukemic cells of T-LGL are large lymphocytes with mature chromatin and moderate amounts of pale cytoplasm. Some cases may show mild nuclear indentation, but deep clefts or lobulations are not seen. The cytoplasm contains variably prominent azurophilic granules. Ultrastructurally, the azurophilic granules consist of parallel tubular arrays containing cytotoxic enzymes such as T-cell intracellular antigen (TIA-1), perforin, and granzyme. Bone marrow biopsy and immunostaining for CD3 or cytotoxic molecules characteristically shows a sinusoidal infiltrate.
A majority of LGL leukemias consist of cytotoxic T lymphocytes that express pan T-cell–associated antigens (e.g., CD2, CD3, CD5) and generally express the cytotoxic T-cell marker CD8 but are negative for the T-helper antigen CD4. Expression of CD5 is often decreased compared with normal T-cell populations, and in some cases CD5 may be absent. T-LGL leukemia cells also typically express the NK-cell–associated antigens CD16 (the Fc receptor-γ noted in original descriptions of LGL leukemia) and CD57 but are usually negative for the NK-cell–associated antigen CD56 (the neural cell adhesion molecule). Expression patterns of other markers not typically used in diagnostic immunophenotyping—specifically the absence of CD27, CD28, CD45RO, and CD62L (L-selectin) and the expression of CD45RA—are consistent with a constitutively activated terminal-effector memory T-cell phenotype, which may explain at least in part some of the autoimmune phenomena observed in many cases of LGL leukemia. The immunophenotype in LGL leukemia is characteristic but may not show frank antigenic aberrance compared with physiologic subsets of cytotoxic T lymphocytes. For that reason, T-cell antigen receptor gene rearrangement analysis should be used to confirm the diagnosis.
A minority of LGL leukemia cases consist of true NK lymphocytes. Indolent forms of NK-LGL leukemia may appear as a clinicopathologic syndrome indistinguishable from T-LGL leukemia. However, there are also forms of indolent NK LGL lymphocytosis that lack the full spectrum of LGL leukemia features such as organomegaly. True NK-LGL leukemias do not express surface CD3 but do express the NK-cell–associated antigens CD16 and CD56. In contrast to aggressive NK-cell leukemia, indolent NK-cell LGL proliferations are Epstein-Barr virus negative. Clonality has been demonstrated in a few cases by human androgen receptor assay, and more recently by demonstration of monotypic restriction of KIR subtype expression ( Fig. 13.2 ). In addition, aberrant expression patterns of NK-associated antigens such as CD94 or CD161 have been cited as evidence of neoplasia in NK-cell populations in lieu of markers of clonality. Some investigators believe that at least some cases of chronic NK-cell lymphocytosis represent a chronic reactive or perhaps a clonal nonmalignant condition.
T-LGL leukemia is a monoclonal disorder, and, therefore, T-cell receptor gene rearrangement studies show clonal rearrangement patterns. Both the TCR-β and TCR-γ loci can usually be shown to be rearranged. The neoplastic nature of T-LGL leukemia was first revealed by the discovery of clonal cytogenetic abnormalities in this disorder. The initial report of clonal cytogenetic abnormalities in T-LGL leukemia described two cases: one with clonal trisomy 8 and one with clonal trisomy 14. Since then, numerous cytogenetic abnormalities have been described in T-LGL leukemia, but none is considered a unique or defining abnormality. These abnormalities include, among others, structural abnormalities of chromosomes 7 and 14 that are common to other T-cell neoplasms.
NK-LGL leukemia does not show clonal T-cell receptor gene rearrangements, but clonality or aberrancy can be demonstrated in many cases by immunophenotypic assessment of KIR expression or abnormal deletion of NK-associated antigens.
Activating STAT3 mutations predominantly in exons 21 and 22, encoding the Src homology 2 domain, occur in 28% to 75% of T-LGL and 30% to 48% of NK-LGL cases and support the neoplastic nature of these indolent proliferations. The clinical significance of these mutations is uncertain.
Some types of transient or non-neoplastic lymphocyte proliferations may be marked by a relative or absolute increase in circulating LGLs. The characteristic reactive T lymphocytes of infectious mononucleosis may occasionally harbor cytoplasmic granules; however, infectious mononucleosis typically is marked by striking morphologic heterogeneity among circulating lymphocytes, in contrast to the monotonous LGL proliferations typical of T-LGL leukemia.
A common but underrecognized form of reactive peripheral blood lymphocytosis, known as stress lymphocytosis , may occur following marked tissue injury (e.g., myocardial infarct, cerebrovascular accident, pulmonary embolus, and physical trauma). Stress lymphocytosis generally consists of a mixture of lymphocyte types, but circulating LGLs may be increased. However, stress lymphocytosis is typically transient, subsiding within hours or days and giving way to absolute peripheral neutrophilia. Some patients with HIV infection may demonstrate a transient CD8 + T-cell lymphocytosis. A subset of these patients may develop persistent T-cell clones that phenotypically mimic T-LGL leukemia but lack typical clinical features of T-LGL leukemia. These appear to represent immune response to viral infection, and one should avoid overinterpretation of T-cell clones in the blood of such patients. If CD56 is expressed in a potential case of T-LGL leukemia, the leukemic phase of hepatosplenic T-cell lymphoma should be excluded. Features supporting the latter include hepatosplenomegaly, aggressive clinical course, hemophagocytosis, and presence of isochromosome 7q.
Persistent T-cell LGL proliferations have been described in recipients of allogeneic hematopoietic stem-cell transplants. These proliferations may be polyclonal, oligoclonal, or monoclonal. Some data indicate that a relatively common immunophenotype in such cases includes expression of CD57 in the absence of either CD16 or CD56, in contrast to the CD16 + ,CD57 + phenotype typical of de novo T-LGL leukemias. Rare reports suggest that some of these posttransplantation LGL proliferations represent true neoplasms; however, the bulk of evidence seems to support that, despite the clonal nature of some of these cases, they likely represent non-neoplastic graft-versus-host phenomena in bone marrow or stem-cell transplant recipients. Recent data suggest lower relapse rates of primary diseases in patients with more robust LGL responses after hematopoietic stem-cell transplant, further supporting the notion that this represents a graft-versus-host phenomenon. Longer term follow-up, even in cases with substantial lymphocytosis and documented clonality, supports classification of this phenomenon as reactive rather than neoplastic. Similar LGL proliferations have also been described in some solid organ transplant recipients.
Mycosis fungoides (MF) is a clinicopathologically distinct form of primary cutaneous T-cell lymphoma. A minority of patients with MF develop a syndrome of diffuse erythroderma, diffuse lymphadenopathy, and leukemic involvement by the lymphoma. This constellation of findings is termed Sézary syndrome (SS), and the circulating lymphoma cells are termed Sézary cells ( Fig. 13.3 ). The International Society for Cutaneous Lymphomas has set forth a proposed hematologic definition of Sézary syndrome that includes erythroderma plus one or more of the following:
Minimum of 1000/mm 3 Sézary cells
CD4-to-CD8 ratio greater than 10, owing to an increase in CD4 + or CD3 + cells by flow cytometry
Aberrant expression of pan–T-cell antigens by flow cytometry
Increased lymphocyte count with T-cell monoclonality by molecular techniques
Abnormal clonal karyotype
MF has a variable clinical course, with numerous emerging therapeutic options. True SS is clinically aggressive. A recent study evaluating the International Society of Cutaneous Lymphomas/European Organization for Research and Treatment of Cancer (ISCL/EORTC) staging criteria observed a median survival of true SS of 3 years, with 26% survival at 5 years and 12% at 10 years.
Leukemic phase of MF (cutaneous T-cell lymphoma)
Marked by diffuse erythroderma, lymphadenopathy, and circulating lymphoma cells (Sézary cells)
Circulating Sézary cells: morphologically distinct, convoluted (cerebriform) cells
Typical immunophenotype is T-helper cell with deletion of expression of CD7 (CD2 + , CD3 + , CD4 + , CD5 + , CD7 − , CD8 − ), but other T-cell antigen deletions may be seen
Expression of CD25 (interleukin 2 [IL-2] receptor) is variable or absent despite this being a therapeutic target for denileukin diftitox
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here