Aims T cell large granular lymphocytes (T-LGLs) are commonly increased in reactive conditions as well as T-LGL leukaemia. This differential diagnosis often requires a combined assessment of clonality and tumour burden. In this study we assessed the utility of flow cytometric (FC) analysis of T cell receptor β chain variable region (TCR-Vβ) expression by using 24 antibodies reactive to 70% of the TCR-Vβ repertoire.
Methods Analyses were performed on peripheral blood samples obtained from 20 patients with a confirmed diagnosis of T-LGL leukaemia and 18 patients without known T cell lymphoproliferative diseases.
Results The results were compared with TCR gene rearrangement status assessed by PCR. By FC analysis, 19/20 T-LGL leukaemia cases were CD3+CD8+ and one case was CD3+CD4+. All the cases demonstrated at least one immunophenotypic aberration, with altered CD5 expression being most frequent. Abnormal Vβ expression was detected by FC in 19 of 20 (95%) T-LGL leukaemia cases, but in none of the controls; this showed 100% concordance with TCR gene rearrangement studies. In addition to establishing clonality, FC Vβ analysis enables calculation of absolute numbers of clonal T cells; this is important in monitoring tumour burden after treatment.
Conclusions It is concluded that FC Vβ analysis is a fast, reliable and quantitative method that can simultaneously assess T-LGL leukaemia clonality and tumour burden.
Statistics from Altmetric.com
T cell large granular lymphocyte (T-LGL) leukaemia is a rare mature T cell neoplasm characterised by a monoclonal proliferation of mature large granular lymphocytes (LGLs) in the range of 2.0×109/l to 20×109/l.1–3 T-LGL leukaemia mainly affects older patients with an average age of 60 years. The disease is indolent and early stage cases can be asymptomatic but common clinical manifestations include fever, recurrent infection, weight loss and moderate splenomegaly.3–6 Approximately 60–80% patents present with neutropenia at the time of initial diagnosis.7 Lymphadenopathy is rare. Bone marrow involvement is almost always present but can be subtle and difficult to recognise morphologically.8 The differential diagnosis is broad and includes benign reactive conditions such as viral infection, rheumatoid arthritis and other types of autoimmune disease, and reactive conditions associated with stem cell or organ transplantation,7 9 and neoplastic conditions such as T prolymphocytic leukaemia, hepatosplenic T cell lymphoma, aggressive natural killer (NK) cell leukaemia, other mature T cell lymphoma/leukaemia and NK cell lymphoma in leukaemic phase. In elderly patients (>65 years) it is also important to remember that small numbers of monoclonal T cell populations can sometimes be detected in patients without clinical evidence of disease.10 The significance of this T cell equivalent to “benign monoclonal gammopathy”, and the frequency of subsequent evolution to T cell neoplasia, is uncertain.
With regard to immunophenotype, T-LGL cells typically show a mature cytotoxic T cell pattern, positive for CD3, CD5 (slightly dimmer than most other T cells), CD8, CD16, CD57, TIA-1, granzyme M, granzyme B and TCRα/β (detectable by immunohistochemistry using the βF1 antibody). Rare cases are positive for CD4.10 Diminished expression or loss of CD5 and/or CD7 are common findings, and together they are detected in up to 80% of cases.11 12
Persistent elevation of circulating T-LGLs for at least 6 months at a level equal to or greater than 2.0×109/l is generally required to establish the diagnosis of T-LGL leukaemia.1 13 In the presence of clonality, however, a diagnosis of T-LGL leukaemia can be made even when the blood LGL level is below 2.0×109/l, especially when other clinical and pathological criteria are met (clinical symptoms, splenomegaly, histological involvement of bone marrow, etc.).1 13 14 Flow cytometric (FC) identification of an expanded T-LGL population in bone marrow and/or peripheral blood can be very helpful in assisting the diagnosis of T-LGL leukaemia, especially when an aberrant immunophenotype is present. However, T cell antigenic alterations are not entirely specific for T cell lymphoma/leukaemia and have been observed in some reactive conditions with relatively high frequencies.15–17
Clonality analysis of T-LGL is routinely performed by assessment for monoclonal TCR gene rearrangement, most often using PCR or, less commonly, Southern blot analysis.6 18 19 The recent availability of FC assays for assessing expression of Vβ repertoire has been shown to be useful and advantageous for assessing T cell clonality.20 21 One study assessed Vβ expression in 98 cases with persistent expansion of LGLs, and correlated the results with molecular studies of TCR gene rearrangements. The study showed that LGL expansion/proliferation encompasses a spectrum of cases with polyclonal, oligoclonal or monoclonal patterns, as shown by Vβ FC analysis or TCR gene rearrangement study. For this reason, it is helpful to incorporate tumour burden (ie, the numbers of clonal T cells) as one of the criteria to establish the diagnosis of T-LGL leukaemia.
In our institution, TCR-Vβ analysis by FC is selectively performed on cases that show either a marked expansion of T-LGL cells or have T cells with an aberrant immunophenotype demonstrated by FC immunophenotyping. We report here our experience using FC TCR-Vβ analysis in assessing clonality and tumour burden in 20 cases of T-LGL leukaemia.
Materials and methods
All patients were seen at The University of Texas MD Anderson Cancer Center (MDACC) during a 4-year period beginning in May 2004 when the FC analysis of TCR Vβ expression was first implemented in the clinical laboratory. FC Vβ analysis was selectively performed in patients whose T cells exhibited an aberrant immunophenotype and/or an abnormal CD4:CD8 ratio by standard FC analysis in suspected T cell lymphoproliferative processes. All cases included in this study had a confirmed diagnosis of T-LGL leukaemia rendered by haematologists and haematopathologists after assessment of patients' clinical presentation, laboratory data, peripheral blood smears, bone marrow biopsies and aspirates, FC immunophenotypic data and TCR gene rearrangement studies by PCR.
We used a control group of 18 patients who had no evidence of a T cell malignancy. These patients sought medical care for a variety of illnesses including: benign dermatitis (n=5), Epstein–Barr virus (EBV) infection (n=2), hypereosinophilic syndrome (n=1), Langerhans cell histiocytosis (n=1), follicular lymphoma (n=1), plasma cell dyscrasia (n=1), myelodysplastic syndrome (n=4) and acute myeloid leukaemia (n=3).
FC immunophenotypic analysis of Vβ repertoires
Peripheral blood (PB) specimens were collected in EDTA tubes. After incubation of whole blood with monoclonal antibodies for 10 min at 4°C, erythrocytes were lysed with NH4Cl for 10 min, followed by two washing steps using phosphate-buffered saline solution. The cells then were resuspended and fixed with 1% paraformaldehyde. Four-color FC was performed using FACSCalibur cytometers (BD Biosciences, San Jose, California, USA). The panel of monoclonal antibodies used included CD2, CD3, CD4, CD5, CD8, CD7, CD16, CD25, CD26, CD56, CD57, CD94, TCR α/β and TCR γ/δ (BD Biosciences). FC immunophenotypic aberrancies were defined as altered expression levels of CD2, CD3, CD4, CD5, CD7, CD8, CD56 and/or CD57 on the gated T cell population.
FC analysis of TCR Vβ expression was assessed by using an eight-tube panel containing 24 monoclonal antibodies known to react with specific TCR-Vβ families (IOTest Beta Mark TCR-Vβ Repertoire Kit; Beckman Coulter, Miami, Florida, USA). Each tube contained a mixture of three different antibodies conjugated to phycoerythrin, fluorescein isothiocyanate, or phycoerythrin and fluorescein isothiocyanate, thus permitting simultaneous analysis of expression of three Vβ families. This panel of antibodies covers approximately 70% of the TCR-Vβ repertoires of normal T cells. In most cases CD3-APC and CD8-PerCP-Cy5.5 were included in each tube for gating. FC Vβ analysis was performed primarily on CD3+CD8+ T cells. In one case with CD4+ neoplastic cells, CD4 was used for gating instead of CD8. To improve detection sensitivity, if an aberrant T cell population showed altered levels of CD3 and/or CD8, Vβ expression was analysed specifically on the aberrant T cells. The FC results were considered to be positive for clonality if: (1) a single TCR-Vβ was expressed by ≥50% of a gated T cell subset; or (2) a TCR-Vβ was expressed at a frequency ≥10 times above its normal limit, based on the manufacturer's reported ranges for total T cells or T cell subsets; or (3) ≥70% of gated T cells failed to react to any of the TCR-Vβ antibodies, presumably due to expression of a TCR-Vβ not recognised by the antibody panel. Henceforth, the last set of cases is referred to as “non-reactive” in the text.
Quantitative T cell subset analysis
Absolute cell counts were obtained using dual platform analysis. The total white blood cell count was derived using a haematology analyser (CellDyn 3500; Abbott Diagnostics, Des Plaines, Illinois, USA). Instrument performance was checked daily by recording fluorescence intensity with calibrating beads (BD Biosciences). Reagent and antibody performance was checked by analysing control cells (CDChex; Streck Laboratories, Omaha, Nebraska, USA) and peripheral blood from blood bank donors at MDACC.
TCR gene rearrangement analyses
Genomic DNA was extracted from PB samples and the TCR γ gene was analysed by multiplex PCR utilising four sets of fluorescently labelled, multicolor primers and capillary electrophoresis.22 The TCR β gene was assessed by multiplex PCR using BIOMED primers, a two-tube multicolor system, and capillary eletrophoresis. Monoclonal peaks were scored, with or without a low-level or more prominent smear (polyclonal amplification) pattern. An oligoclonal pattern was defined as three prominent peaks, with peak height substantially higher than the baseline polyclonal amplification pattern. The analytical sensitivities of the TCR γ and β PCR assays were 5–10% and 10–15%, respectively.
Statistical comparisons were performed as follows: (1) continuous numerical data were compared using a two-sided Student t test, or one way ANOVA, and (2) proportions were compared using Fisher's exact test or the χ2 test. For all statistical tests, statistical significance was defined by a p value of 0.05 or less.
Twenty patients with T-LGL leukaemia and 18 control cases were included (tables 1 and 2). The LGL leukaemia diagnosis was made based on a combination of clinical symptoms, elevated circulating LGL assessed by cytomorphology on PB smears and FC immunophenotyping, identification of bone marrow infiltration and molecular evidence of clonal TCR gene rearrangement.
Among 20 patients with T-LGL, eight patients were asymptomatic (40%), but an abnormal complete blood count (CBC) (cytopenia, lymphocytosis or both) was found at the time of physician visit. The remaining 12 patients presented with symptoms such as neutropenia and anaemia (table 1). Four of 20 cases (20%) had a prior history of malignancy (one prostate cancer, two skin cancers, one colon cancer), one patient had thymoma, nine patients (45%) had prior surgeries, three patients (15%) had a history of autoimmune disease including one with rheumatoid arthritis, and six patients (30%) presented with splenomegaly.
Laboratory investigations revealed significant anaemia (<10 g/dl) in seven (35%) patients, significant neutropenia (<1.5×109/l) in 10 (50%) patients, and significant thrombocytopenia (<100×109/l) in four (20%) patients. Lymphocytosis (>2.0×109/l) was present in 13 (65%) patients. Among the three patients with lymphocytes <1.0×109/l (table 1: patient no. 1, 11 and 18), patient 1 was treated with steroids and patient 11 was treated with 2-chlorodeoxyadenosine (2CDA) prior to being referred to MDACC. Bone marrow involvement was present in all 20 cases, showing a diffuse or interstitial lymphocytic infiltrate as demonstrated by immunohistochemistry stains, and the bone marrow lymphocytes ranged from 8% to 62% by counting aspirate smears. Conventional karyotyping revealed a complex karyotype in one patient, t(2;6)(q31;q25) in one patient, loss of Y in two patients (most likely an age-related change) and a normal karyotype in 16 patients (table 1).
Compared with the patients without a T cell lymphoproliferative disease, patients with T-LGL leukaemia showed similar age distributions (p=0.197) and WBC counts (p=0.450), higher lymphocyte percentages but not significantly higher absolute lymphocyte counts (p=0.004 and p=0.143, respectively), and significantly lower CD4:CD8 ratios (p<0.001) and absolute CD8+ T cell counts (p=0.034) (table 2). None of the control cases showed an abnormal bone marrow lymphocytic infiltrate.
Immunophenotypic aberrancy of T-LGL cells
FC immunophenotyping was performed on the blood samples from all patients. Nineteen of 20 T-LGL leukaemia cases were CD3+CD4−CD8+ and one case was CD3+CD4+CD8−. All cases were TCR α/β positive, TCR γ/δ negative and demonstrated at least one immunophenotypic aberrancy. Altered CD5 expression was most common and was seen in 13/20 cases (loss of CD5 in seven cases and CD5 expression dimmer than normal LGL in six cases). The second most common aberrancy was decreased CD7 expression, observed in 11/20 patients (loss of CD7 in four cases and dim CD7 expression in seven cases). Seven cases showed decreased CD57 expression. One case showed dim CD2 expression. All 20 cases were CD56−.
The PB CD4:CD8 ratio showed a median of 0.14 (range 0.03–5.36) (table 2, figure 1). Two cases had a CD4:CD8 ratio >1, including one case of CD3+CD4+CD8− LGL leukaemia (patient no. 9), showing a ratio of 5.3, and patient no. 18 with a ratio of 3.59 who had received 2CDA treatment prior being referred to MDACC (table 1).
In comparison, 4 of 18 (22%) control cases showed at least one immunophenotypic aberrancy: three cases with dim CD5 and one case with dim CD7 expression. The median CD4:CD8 ratio for the control group was 1.22, ranging from 0.31 to 12.97 (table 2).
Flow cytometry Vβ analysis and correlation with gene rearrangement data
FC Vβ repertoire analyses were performed on the PB samples of all 20 patients with T-LGL leukaemia. An example histogram of a positive case is shown in figure 2. In this example, the neoplastic cells were CD8 dim+ as compared with the residual normal CD8 cells. A total of 93% of the CD8 dim+ neoplastic cells were positive for surface Vβ 8 expression, confirming T cell clonality, while the normal CD8 T cells showed a polyclonal pattern. Using this approach, clonality was established in 19 of 20 LGL cases (tables 1 and 2), of which 15 cases showed expansion of single Vβ subtype. Four cases showed an expanded CD8+ T cell population (≥70%) non-reactive to the panel of anti-Vβ antibodies, consistent with clonal expansion of T cells expressing a Vβ subtype not included in the anti-Vβ panel. One case showed increased non-reactive cells (62%), but failed to meet the cut-off of 70%. None of the control cases showed a clonal expansion of T cells by FC Vβ analysis.
Monoclonal TCR gene rearrangements were detected by PCR in all 18 cases tested. Overall, there was a 100% (18/18) concordance rate between clonality assays by FC for Vβ expression, versus PCR for TCR gene rearrangements. TCR gene rearrangements were also detected by PCR in 3 of 12 control cases tested, including one patient with acute myeloid leukaemia, one with myelodysplastic syndrome, and one patient with spongiotic dermatitis.
Correlation between clonal Vβ T cells and lymphocyte subsets
There are strong correlations between the absolute count of the expanded single Vβ clone (median 2.77, range 0.23–17.67 (×109/l)), the absolute number of lymphocytes (median 3.53, range 0.81–23.94 (×109/l)), and the T-LGL immunophenotype (CD3+ CD8+CD57+) cells (median 2.85, range 0.62–17.81 (×109/l)) (r=0.9 and 0.85, respectively). The CD4:CD8 ratio (median 0.14, range 0.03–3.56) was often reversed in T-LGL leukaemia, but its correlation with levels of clonal Vβ+ T cells was poor (r=0.33).
We assessed here the utility of FC Vβ analysis in assessing T cell clonality and circulating tumour burden in 20 cases of T-LGL leukaemia, and correlated the results with clinical and haematological features. Vβ FC analysis was positive for clonality in 19 of 20 (95%) T-LGL leukaemia cases, and negative in all control cases without a T cell lymphoproliferative process. By contrast, monoclonal TCR gene rearrangements detected by PCR were found in all 18 (100%) tested patients with T-LGL leukaemia, but monoclonal rearrangements were also found in 3/12 controls. Thus, FC Vβ clonality assessment had equal sensitivity and superior specificity to PCR analysis for TCR gene rearrangements.
The recommendation for establishing a diagnosis of T-LGL leukaemia is to combine the assessment of clinical presentation, bone marrow infiltration, clonality testing and tumour volume. In most laboratories, T cell clonality is assessed by analysing TCR gene rearrangements, usually by PCR, which is a qualitative but not quantitative assay. The tumour volume is usually assessed by total lymphocyte count. This approach is practical, but imprecise. In contrast, FC Vβ analysis can assess the quantity and clonality of lymphocytes simultaneously; also, FC Vβ analysis is easy to combine with FC immunophenotyping. Therefore, we suggest that FC Vβ analysis is a more convenient and precise approach for enumerating tumour burden in patients with T-LGL.
A limitation of FC Vβ analysis is that the available anti-TCRβ antibody panel covers only approximately 70% of all the TCRβ families expressed. Therefore, a subset of T cell neoplasms will not be detected by the current antibody panel. In these cases, however, clonality can be inferred from the non-reactive pattern, which contrasts with normal polyclonal T cell populations that express various Vβ families. Four cases showed >70% T cells non-reactive to any of the anti-Vβ antibodies, and were reported as positive, according to previously suggested thresholds.20 One case (patient no. 19) showed 62% non-reactive T cells; this is a marked increase, but fails to meet the 70% cut-off, indicating that this cut-off for non-reactive cases may be slightly too conservative in some cases.
A second potential limitation of this assay is that a clonal population that expresses a Vβ chain detected by the antibody panel may nonetheless be masked by background polytypic normal T cells, since the diagnostic thresholds are based on the proportion of gated T cells positive for a single Vβ family. In this study, we gated primarily on CD3+CD8+ cells, further defined by FSC/SSC, in order to remove polytypic normal CD4+ T cells. It is noteworthy that many PB specimens contain neoplastic CD3+CD8+ T cells and substantial numbers of normal CD3+CD8+ T lymphocytes. CD3 and CD8 expression levels are infrequently altered in T-LGL leukaemia (0/20 and 1/20 cases in our series, respectively). Thus in most cases, one cannot distinguish clonal and non-clonal CD8+ T cell populations based on CD3 and/or CD8 aberrancy for gating. With the development of multicolour FC in recent years, incorporating more T cell markers into FC Vβ analysis, such as CD5 and CD57, may improve the detection sensitivity.
We have shown that there is a strong correlation between the absolute count of clonal Vβ+ cells and the absolute count of CD3+CD8+CD57+ T-LGL cells, as well as between the percentages of these two populations. In most cases, the clonal Vβ+ cell count also correlates with a high proportion of total lymphocytes, total T cells and CD8+ T cells. We therefore conclude that the total lymphocyte count, and CD3+CD8+ T cell percentage and count, are all approximate surrogates for the quantity of clonal T cells, and can reflect the T-LGL tumour burden at the initial diagnosis. However, in patients post-treatment, in which there may be a reduction of tumour burden and reconstitution of the normal immune system, this correlation may no longer exist. In such cases, FC Vβ analysis with a single tube containing the relevant Vβ antibody can accurately measure tumour burden in the post-therapy settings (unpublished observations).
In summary, we have shown that FC Vβ analysis is sensitive and specific in assessing clonality in T-LGL leukaemia. FC Vβ analysis can assess clonality and quantify neoplastic cells simultaneously. This capacity to assess tumour volume in patients with LGL expansion gives FC Vβ analysis an advantage over PCR studies in establishing a diagnosis of LGL leukaemia. In addition, this assay is rapid, convenient and potentially useful in monitoring disease after treatment.
Flow cytometric (FC) Vβ analysis is sensitive and specific in assessing clonality in T cell large granular lymphocyte (T-LGL) leukaemia.
FC Vβ analysis can assess clonality and quantify neoplastic cells simultaneously, showing advantage over PCR studies in establishing a diagnosis of large granular lymphocyte leukaemia.
FC Vβ analysis is rapid, convenient and potentially useful in monitoring disease after treatment.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.