Article Text

PDF

Acute lymphoblastic leukaemia: correlation between morphological/immunohistochemical and molecular biological findings in bone marrow biopsy specimens
  1. S M Kröber1,
  2. A Greschniok1,
  3. E Kaiserling1,
  4. H-P Horny1
  1. 1Institute of Pathology, University of Tübingen, D-72076 Tübingen, Germany
  1. Dr Kröber, Institut für Pathologie, Eberhard-Karls-Universität, Liebermeisterstrasse 8, D-72076 Tübingen, Germany email: stefan-martin.kroeber{at}med.uni-tuebingen.de

Abstract

Background—Although numerous antibodies suitable for use on paraffin wax embedded sections are available for the subtyping of acute leukaemia (acute myelogenous leukaemia (AML) and acute lymphoblastic leukaemia (ALL)) in bone marrow biopsy sections, unequivocal identification of the cell line involved is sometimes impossible.

Methods—Forty eight formalin fixed, paraffin wax embedded bone marrow biopsy specimens that had been decalcified in EDTA were investigated, including 42 thought to exhibit ALL on the basis of bone marrow smears. Five specimens exhibited AML and one biphenotypic leukaemia, as diagnosed immunohistochemically in bone marrow biopsies. Immunostaining was performed with antibodies against relatively specific B and T cell antigens. The blasts were investigated for rearrangements of the immunoglobulin heavy chain (IgH) and the T cell antigen receptor (TCR) genes.

Results—Amplifiable DNA was obtained from all 48 specimens. An IgH gene rearrangement was detected in 20 of 23 c-ALL specimens. Four of seven T cell ALL (T-ALL) specimens had a TCR-γ gene rearrangement, and the one B cell ALL (B-ALL) specimen exhibited a clonal IgH gene. Three of four cases of unclassifiable ALL could be assigned to the B cell lineage on the basis of gene rearrangement analysis. Seven cases originally diagnosed in smears as ALL were rediagnosed as AML (n = 5) or biphenotypic leukaemia (n = 2) because of immunohistochemical reactivity for myeloperoxidase or lysozyme. Two of these AML cases and two of three cases of biphenotypic leukaemia exhibited a monoclonal IgH gene rearrangement.

Conclusions—Acute leukaemia can be subtyped in bone marrow sections with a limited panel of antibodies suitable for use on paraffin wax embedded sections (against CD3, CD10, CD20, CD79a, myeloperoxidase, and lysozyme). In patients with ALL and a diagnostically equivocal immunophenotype, gene rearrangement analysis might indicate whether the B or T cell lineage is involved.

Statistics from Altmetric.com

The term acute lymphoblastic leukaemia (ALL) describes a heterogeneous group of acute leukaemias that usually manifest themselves primarily in the bone marrow and blood as a neoplastic proliferation of lymphoblasts. It is usually possible to subtype ALL in sections of bone marrow biopsy specimens (taken from the iliac crest) using immunohistochemical techniques.1 A particularly large panel of antibodies that can be used on paraffin wax embedded tissue is necessary when the lymphoblasts have unusual immunophenotypes.

However, in a few cases, definitive immunohistochemical identification of the lineage (B or T cell) is not possible. We have tried to solve this problem by using the polymerase chain reaction (PCR). The lineage of lymphoid cells can be identified by investigations to detect rearrangement of the immunoglobulin heavy chain (IgH) and/or T cell antigen receptor (TCR) genes.2–6 Unlike Southern blotting, the PCR method can be used on very small, formalin fixed, paraffin wax embedded specimens.

We assessed the diagnostic value of gene rearrangement analysis in ALL using a large series of patients with acute leukaemia from our department, and with particular reference to those in which the tumour cell lineage could not be identified with certainty by immunostaining.

Material and methods

A total of 42 bone marrow biopsy specimens shown by cytomorphological, enzyme cytochemical, and immunocytochemical investigations to contain infiltrates of ALL were included in our study (all cases included in a previous study).1 Five cases of acute myelogenous leukaemia (AML) and one of mixed myelogenous/lymphoblastic (biphenotypic) leukaemia, all of which had been diagnosed immunohistochemically, served as controls. All the specimens were fixed in 5% buffered formalin, mildly decalcified in EDTA,7,8 and embedded in paraffin wax. Serial sections were cut at 4 μm and stained with haematoxylin and eosin, Giemsa, the periodic acid-Schiff reaction, and the naphthol AS-D chloroacetate esterase reaction. Immunohistochemical staining was performed by the avidin–biotin–peroxidase complex method.9 Although a very broad panel of antibodies was used, particular emphasis in the evaluation was placed on those antibodies with high lineage specificity; that is, those antibodies directed against: myeloperoxidase (MPO; Dako, Hamburg, Germany), lysozyme (Lys; Dako), terminal deoxynucleotidyl transferase (TdT; Dako), CD3 (Novocastra, Newcastle upon Tyne, UK), CD10 (Novocastra), CD20 (Dako), CD79a (Dako), and βF1 (β chain of TCR; T-cell Diagnostics Inc, Cambridge, UK).1

Immunostaining gave definite confirmation of the diagnosis of ALL in 35 cases: c-ALL (n = 23), T cell ALL (T-ALL; n = 7), B cell ALL (B-ALL; n = 1), and u-ALL (not classifiable; n = 4) (table 1). Seven cases that had originally been diagnosed in smears as ALL were shown by staining with antibodies against myeloperoxidase and lysozyme to be either myeloid leukaemia (AML-M1, n = 2; AML-M4, n = 2; and AML-M5, n = 1) or biphenotypic leukaemia (n = 2). Together with the control cases, this made a total of 12 cases of AML and three cases of biphenotypic leukaemia.

Table 1

Immunophenotype, clonality, and subtype of the cases of acute leukaemia investigated

The PCR investigations for IgH and TCR-γ gene rearrangements were performed according to a standardised protocol.10 DNA extracts from a lymph node with diffuse infiltration by B cell chronic lymphoblastic leukaemia (B-CLL) and a large cell anaplastic T cell lymphoma were used as positive controls.

Results

Amplifiable DNA was extracted from all 48 bone marrow specimens. Monoclonal or polyclonal PCR products of the IgH and TCR-γ gene rearrangements could be identified in all the 35 cases of ALL that could be analysed by PCR (that is, β globin positive cases). Table 1 gives a summary of the molecular biological findings and diagnostically relevant immunohistochemical findings with highly lineage specific antibodies.

Monoclonal IgH rearrangement patterns were found in 20 of the 23 cases of c-ALL (fig 1). One specimen (case 8), with the immunophenotype TdT positive, CD10 positive, Cd79a positive, exhibited monoclonal rearrangements of both the IgH gene and the TCR-γ gene. A further c-ALL case had monoclonal rearrangement of only the TCR-γ gene; the immunophenotype of the blasts here was TdT positive, CD10 positive (case 18). Among the seven cases of T-ALL, monoclonal rearrangement of the TCR-γ gene was found in four, and of the IgH gene in one. The TCR-γ rearrangement was polyclonal in the latter case (case 30) and the blast cells expressed TdT and βF1. As expected, the one case of B-ALL (case 31) exhibited a monoclonal rearrangement of the IgH gene. A monoclonal rearrangement of the IgH gene was found in all four cases of u-ALL; one of these (case 35), which was immunoreactive only for TdT, also exhibited a monoclonal rearrangement of the TCR-γ gene.

Figure 1

Trephine biopsy specimen taken from the iliac crest of a patient with acute lymphoblastic leukaemia (c-ALL; case 6). Left: extremely hypercellular marrow with subtotal depletion of fat cells and normal blood cell precursors. Most of the cells are reactive for the B cell associated antigen CD79a. Note the pronounced membrane associated (“annular”) immunostaining, which indicates that the staining is specific. Avidin–biotin–peroxidase complex (ABC) method, anti-CD79a. Right: ethidium bromide stained agarose gel showing PCR analysis of the immunoglobulin heavy chain (IgH) gene rearrangements in complementarity determining region 3 (CDR3). Note the clear monoclonal band of lane 3 (case 6). M, molecular weight markers (100 bp ladder); lane 1, no DNA; lane 2, case 31; lane 3, case 6.

A monoclonal IgH gene rearrangement pattern was found in two of the three cases of biphenotypic leukaemia (cases 36 and 38; fig 2) and in two of the 10 cases diagnosed as AML on the basis of the immunohistochemical findings (cases 39 and 48). None of the AML cases exhibited a monoclonal TCR-γ gene.

Figure 2

Trephine biopsy specimen taken from the iliac crest of a patient with biphenotypic acute leukaemia (case 38). Left: hypercellular marrow infiltrated by blast cells, almost all of which are reactive for myeloperoxidase (MPO). In addition, most of the blast cells were also immunoreactive for CD79a and terminal deoxynucleotidyl transferase (TdT) (not shown). Right: ethidium bromide stained agarose gel showing PCR analysis of the immunoglobulin heavy chain (IgH) gene rearrangements in complementarity determining region 2 (CDR2). Note the clear monoclonal band of lane 4 (case 38). M, molecular weight markers (100 bp ladder); lane 1, case 5; lane 2, case 11; lane 3, case 17; lane 4, case 38.

In summary, the diagnosis of ALL and the subtype identified were confirmed by PCR in 25 of the 31 cases (table 2). It was possible to classify three of the four cases of u-ALL as deriving from the B cell lineage. Two of the three cases of biphenotypic leukaemia exhibited a monoclonal rearrangement of the IgH gene.

Table 2

Molecular biological findings in relation to the leukaemia subtype

Discussion

The most commonly used systems of classification of acute leukaemia are based mainly on cytomorphological, enzyme cytochemical, and immunocytochemical findings in blood and bone marrow smears. If bone marrow aspiration is unsuccessful, the investigation of bone marrow biopsy specimens (from the iliac crest) becomes of major diagnostic importance.

Various investigators have studied the diagnostic value of antibodies suitable for use on paraffin wax embedded sections in the diagnosis of acute leukaemia in sections from bone marrow biopsy specimens,11–16 and the following panel is generally considered suitable for subtyping in this situation: MPO, TdT, lysozyme, CD3, CD10, CD79a, CD20, and βF1.1 The new WHO classification system of the acute leukaemias is based on a revision of the French–American–British (FAB) system, and takes particular account of cytogenetic and molecular genetic findings, especially in the case of ALL.17–19 Therefore, we investigated the diagnostic value of molecular biological rearrangement analysis in acute leukaemia, particularly in cases in which routine investigation of bone marrow biopsy specimens had not revealed a diagnostically unequivocal immunophenotype.

The finding of a high proportion of cases with monoclonal rearrangement patterns (> 80%) was surprising, because immature neoplasms of the lymphoid lineage (and therefore ALL) are thought to exhibit false negative findings concerning monoclonality in many cases.2,20,21 However, analysis of unfixed cells from bone marrow aspirates in ALL has revealed informative IgH gene rearrangements in up to 95% of cases and TCR-γ or TCR-δ rearrangements in up to 84% of cases.22,23

Gene analysis is based on a comparative PCR. The main source of error of a false positive result is therefore unsuitably small numbers of cells from which the DNA has been extracted.24 The rearranged VDJ sequences of the template DNA of any given cells (including non-neoplastic cells) can bind preferentially to the PCR primers.25 However, in our study, the clonality of the IgH and TCR-γ gene rearrangements was determined in all cases in DNA extracts from iliac crest bone marrow biopsy specimens with diffuse, compact infiltration by blasts. It can therefore be assumed that the cell numbers here were adequate.

Only cases that had been diagnosed as ALL on the basis of blood and bone marrow smears were included in our study. The PCR findings in the one B-ALL case and in four of the seven T-ALL cases confirmed the lineage identified by the immunohistochemical investigation of bone marrow biopsy specimens. False negative findings of gene analysis can be expected more often in T cell neoplasms than in B cell neoplasms because the method is less sensitive and these neoplasms sometimes exhibit a germ line configuration—that is, non-rearranged configuration of the TCR gene—despite CD3 positivity.26,27 As expected, a monoclonal IgH gene rearrangement was found in most (19 of 23) cases of c-ALL, confirming that they are indeed B cell in origin. The molecular biological findings were of particular interest in the four cases that remained unclassifiable on the basis of the immunohistochemical findings, that is, cases of u-ALL. In three of these four cases, the finding of a monoclonal IgH rearrangement showed the blasts to be of B cell origin. This suggests that in cases where the lineage of a TdT positive leukaemia cannot be identified by immunohistochemical investigation, molecular genetic analysis can be of major diagnostic value.

Six control cases were included to test the specificity of the method: five AML cases and one case of biphenotypic leukaemia, all diagnosed immunohistochemically. Therefore, a total of 13 cases of AML and biphenotypic leukaemia was investigated. In two of the biphenotypic leukaemias, at least some of the blasts were found to be of B cell origin by the identification of a monoclonal IgH gene rearrangement. In one of these, immunohistochemical investigation revealed about 20% of the blasts to be CD79a positive, a finding that supports the diagnosis of mixed leukaemia, with one blast cell population with a myeloid phenotype, and one blast cell population with a B cell genotype.

Among the cases with a definite myeloid immunophenotype were two in which the molecular biological finding of a monoclonal IgH rearrangement was inconsistent. In one of these, a case of AML-M1, almost 100% of the blasts were found to express MPO, and CD20 positive B lymphocytes accounted for far less than 10% of the cells. In the other case, a case of AML-M5 with a monoclonal IgH rearrangement, more than 90% of the blast cells were positive for lysozyme and, again, few CD20 positive B lymphocytes were identified by immunohistochemical staining. The inconsistency between the immunohistochemical and molecular biological findings might have been caused by a false positive PCR result deriving from a sampling error, as can occur in very lymphocyte depleted or microdissected tissue.24,28,29 The finding of a monoclonal IgH gene rearrangement could also result from an aberrant recombination of the IgH gene in the tumour cells with a myeloid immunophenotype,27,30,31 but it was not possible to determine whether this was the case with the method used here. It can only be concluded from our investigations and the criteria of the WHO classification system that a diagnosis of biphenotypic leukaemia should be reserved for cases in which the cytogenetic and molecular biological findings indicate transformation of a stem cell with differentiation towards the lymphoid and myeloid cell lines.19

Inconsistencies between the immunohistochemical and molecular biological findings were also noted in a relatively small proportion of the ALL cases with lymphoid differentiation alone (four of 35). In one case of T-ALL with TdT positive and βF1 positive blasts, there was a monoclonal IgH gene rearrangement and a polyclonal TCR-γ gene rearrangement, although the blasts expressed none of the B cell markers investigated. It is not possible to say with certainty whether this represented an ALL with bilinear differentiation; that is, a T cell associated immunophenotype with a B cell associated IgH rearrangement, because such illegitimate IgH rearrangements are said to be extremely rare.32 Only occasional cases of IgH monoclonal T cell neoplasms and TCR monoclonal B cell neoplasms have been described.26,33–35 Coexistent monoclonal IgH and TCR gene rearrangements, as in bilinear lymphoblastic leukaemia, have been found relatively often in unfixed bone marrow aspirates, and monoclonal TCR rearrangements have been reported to have a relatively high incidence in both B-ALL and T-ALL.23,36,37 A further case with a genotype/phenotype discrepancy was an IgH monoclonal c-ALL that also exhibited a TCR-γ gene rearrangement. However, the lymphoblasts in this case expressed only B cell markers, such as CD10 and CD79a. No T cell antigens were detected, although it must be said that not all the relevant markers (especially CD7) would have been detected by the panel of antibodies used. These two last cases exhibited findings consistent with the diagnoses of ALL with B cell genotype and ALL with bilinear genotype. One case of c-ALL exhibited a monoclonal TCR-γ gene rearrangement and a polyclonal IgH gene rearrangement. More than 50% of the blasts in this case expressed OPD4, a T helper cell associated marker.1 These findings led us to amend the diagnosis to T-ALL.

In conclusion, from our findings and published data it appears that the most suitable approach to the diagnosis of acute leukaemia is the evaluation of a bone marrow biopsy, as well as blood and bone marrow smears, by enzyme cytochemical and immunohistochemical techniques. In cases of ALL with a diagnostically equivocal immunophenotype, it is usually possible to determine the lineage (B or T cell) by molecular biological analysis.

Acknowledgments

The authors are grateful to Ms H Ableiter and Ms A Mall for excellent technical assistance, Dr B Toth for help in assessing immunohistochemical data, and Dr M Ruck for translation of the manuscript. Thanks also to Dako Diagnostica GmbH, Hamburg, Germany for contributing towards the reproduction costs of the colour artwork.

References

View Abstract

Request permissions

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.