I Clin Pathol 1996;49:749-754 749 Immunoglobulin light chain mRNA detected by in situ hybridisation in diagnostic fine needle J Clin Pathol: first published as 10.1136/jcp.49.9.749 on 1 September 1996. Downloaded from aspiration cytology specimens

C J R Stewart, M A Farquharson, T Kerr, J McCorriston

Abstract erative lesions is generally more Aims-To demonstrate expression of im- problematic.' Whereas malignant lymphoma munoglobulin light chain mRNA in diag- may be diagnosed on the basis of routine cyto- nostic fine needle aspiration (FNA) cytol- logical preparations,2 3 in some cases it is not ogy specimens using an in situ possible to distinguish reactive and malignant hybridisation (ISH) technique; and to lymphoid lesions on morphology alone.' 4 As in evaluate ISH in a series of reactive histopathology, special techniques often offer a lymphoid proliferations and malignant useful aid to the assessment of lymphoid lymphomas. proliferations and together with the cytomor- Methods-Forty diagnostic FNA speci- phology permit an accurate diagnosis in most mens showing a lymphoid cell population specimens.57 were examined for immunoglobulin light Immunocytochemistry has been used in chain mRNA expression using ISH. Aspi- cytology specimens mainly to demonstrate T rates were obtained from (n = and phenotypes, aberrant expression of 34), salivary gland (n = 3), subcutaneous antigens in lymphomas and immu- tissue, thyroid and breast (n = 1 each). noglobulin light chain restriction in B cell The cases included 20 B cell lymphomas, lymphomas.8"- More recently, material ob- five cases of Hodgkin's disease and 15 tained by FNA has been used for genotypic reactive lymphoid proliferations. Com- analysis including Southern blotting and parison with light chain immunoreactivity polymerase chain reaction (PCR) for demon- was made in 36 cases and histological cor- stration of T cell antigen receptor and relation from biopsy material was avail- immunoglobulin gene rearrangements.'3- The able in 24. latter are valuable techniques but are limited Results-Immunoglobulin light chain re- by loss of correlative cell morphology and the

striction was demonstrated in 14 of 20 B requirement for relatively high cell yields. http://jcp.bmj.com/ cell lymphomas using ISH and in six of 17 Monoclonal B cell proliferation may also be B cell lymphomas using immunocyto- demonstrated using in situ hybridisation (ISH) chemistry. A polytypic pattern of light for detection of restricted immunoglobulin chain expression was observed in four of light chain mRNA.'6 This technique offers the five cases of Hodgkin's disease with both specificity of genotypic analysis but with techniques, and in 12 of 15 and 11 of 14 preservation of cell morphology. ISH detection reactive lymphoid proliferations using is also unaffected by serum immunoglobulin on September 28, 2021 by guest. Protected copyright. ISH and immunocytochemistry, respec- which produces background staining and diffi- tively. culties in interpretation with some immunocy- Conclusions-The assessment of immu- tochemistry preparations. In this study we have noglobulin light chain expression is a use- examined light chain mRNA expression by ful adjunct to morphology in the diagnosis ISH and immunocytochemistry in a series of of reactive and malignant lymphoid pro- reactive and malignant lymphoid proliferations liferations in FNA specimens. Light chain obtained by FNA. restriction can be shown using either immunocytochemistry or ISH, but the latter is more sensitive in the diagnosis of Methods B cell lymphoma. DIAGNOSTIC CASES (7 Clin Pathol 1996;49:749-754) FNA specimens were obtained from 40 patients attending Glasgow Royal Infirmary Keywords: B cell lymphoma, in situ hybridisation, with clinically palpable lesions. Two or three immunocytochemistry, fine needle aspiration. aspirates were performed by Cytology staff Departments of using 23 or 25 gauge needles and 10 or 20 ml Cytology and syringes. The aspirates were obtained from Pathology, Royal Fine needle aspiration (FNA) cytology offers lymph node (34 cases), salivary gland (three Infirmary, Glasgow major advantages in the initial assessment of cases), subcutaneous tissue, thyroid and breast Correspondence to: patients with . The technique (one case each). Direct smears were prepared Dr C J R Stewart, is simple, inexpensive, safe, and is applicable to for routine staining by the Diff-Quik, Department of Pathology, May- Glasgow Royal Infirmary, both inpatient and outpatient settings. A diag- Grunwald-Giemsa or Papanicolaou methods Glasgow G4 OSF. nosis of metastatic carcinoma or melanoma is and needles thereafter rinsed in normal saline. Accepted for publication usually straightforward in lymph node FNA Cytospin preparations from the needle rinses 27 June 1996 specimens but the assessment of lymphoprolif- were used for immunocytochemistry and ISH. 750 Stewart, Farquharson, Kerr, McCorriston

The optimum dilution for cytospin prepara- Table 1 Light chain expression using in situ hybridisation was x (ISH) and immunocytochemistry (ICC) tions 150-200 10' cells/ml. J Clin Pathol: first published as 10.1136/jcp.49.9.749 on 1 September 1996. Downloaded from The cytology diagnoses, which were based Monotypic Polytypic both on morphology and the results of special techniques, fell into three main groups: Cytology diagnosis ISH ICC ISH ICC B cell lymphoma 14/17 6/14 0/17 0/14 A: B (n = 20) Suspicious of BCL 0/3 0/3 0/3 0/3 Group cell lymphoma Hodgkin's disease 0/5 0/5 4/5 4/5 Seventeen cases were considered diagnostic of Reactive lymphoid 0/15 0/14 12/15 11/14 B cell lymphoma on the FNA specimen. In proliferations three aspirates the appearances were consid- BCL = B cell lymphoma. ered highly suggestive but not conclusive of B cell lymphoma. Further subclassification of the In addition to light chain analysis, most cases lymphomas was not generally attempted on the were also examined with a panel of T and B basis of the FNA findings. cell antisera for routine diagnostic purposes. The diagnosis of B cell lymphoma was con- These included CD3 (SAPU), CD43, CD20, firmed on subsequent excision biopsy in 14 of CD79, and CD45 (all from Dako). Selected the 20 cases (1 1 of those considered diagnostic cases were also examined for immunoreactivity and all three considered suspicious on FNA). to CD 15 (SAPU) and CD30 (Dako). Four cases considered diagnostic of B cell lymphoma on FNA represented recurrent IN SITU HYBRIDISATION tumour after treatment and did not undergo The ISH technique was similar to that used for re-biopsy subsequently. The remaining two paraffin wax embedded tissue sections. Initial patients were unfit for biopsy for medical studies using aspirates from excised human reasons; however, the diagnosis of lymphoma tonsil had shown that optimum fixation was was consistent with the clinical and radiologi- achieved using neutral buffered formalin, and cal findings. that light chain mRNA could be detected up to The biopsy findings in those 18 patients with seven days after sampling in both fixed and histological correlation were of follicle centre unfixed material. cell lymphoma (seven cases), large B cell Cytospin preparations were prepared on lymphoma (five cases), low grade mucosa asso- silane treated slides and fixed for 24 hours in ciated lymphoid tissue (MALT) lymphoma neutral buffered formalin. The slides were (three cases), and lymphocytic lymphoma, my- washed twice in water and then digested with eloma and monocytoid B cell lymphoma (one 20 gg/ml proteinase K (Sigma, Poole, Dorset, case each). UK) for 30 minutes at 370C. Proteinase activ- Light chain expression was also examined by ity was stopped by washing in 0.2% glycine. immunocytochemistry in 17 cases. Cells were post-fixed with 0.4% paraformalde- hyde in phosphate buffered saline for 20

minutes at 4°C. Prior to applying the probes, http://jcp.bmj.com/ Group B: Hodgkin 's disease (n = 5) slides were washed twice in water followed by Biopsy confirmation of Hodgkin's disease was one wash in industrial alcohol, then air dried. obtained subsequently in four patients, show- All solutions used before hybridisation were ing nodular sclerosis subtype in three and treated with diethylpyrocarbonate. nodular lymphocyte predominance (LP) sub- The probes used were fluorescein isothiocy- type in one. The remaining patient who anate (FITC) labelled deoxyribonucleotide presented with a subarachnoid haemorrhage cocktails to K and X mRNA (Dako). Probes on September 28, 2021 by guest. Protected copyright. was considered medically unfit for biopsy. were diluted 1 in 1 with 30% hybridisation Light chain expression was also examined by buffer (30% formamide BDH, 10% dextran immunocytochemistry in all five cases. sulphate, 0.2% polyvinylpyrrolidone, 0.2% Ficoll, 5 mM EDTA, 50 mM Tris HCl, pH Group C: reactive lymphoid proliferation (n = 7.5); 20 p1 of diluted probe was applied to the 15) slides, which were then covered with a Biopsy in two patients with persistent and coverslip and incubated overnight at 37°C. clinically suspicious lymphadenopathy con- Coverslips were removed in 2 x SSC (0.3 M firmed the presence of a reactive lymph node. NaCl, 0.03 M sodium citrate) and sequentially The remaining 13 patients had no evidence of washed in 0.1% Triton X/TBS (0.05 M Tris lymphoma on clinical follow up of two to 24 HCI, pH 7.6, 0.15 M NaCl) for 10 minutes (average 14) months. and TBS for five minutes. Thereafter, slides Light chain expression was also examined by were incubated with alkaline phosphatase con- immunocytochemistry in 14 cases. jugated rabbit anti-FITC (Dako) for 30 minutes at room temperature. Detection was IMMUNOCYTOCHEMISTRY achieved using NBT/BCIP (Sigma) as sub- Cytospin preparations for light chain analysis strate and cells were counterstained with 1% were air dried and then fixed in acetone for 10 light green. minutes. The primary antibodies were polyclo- Reactive tonsillar tissue was used as a nal anti-K and anti-X (Dako, High Wycombe, positive control. In negative controls the UK) which were incubated for one hour at a ribonucleotide probes were omitted. The ISH dilution of 1 in 2000. Detection was with a methodology was also checked periodically by biotin/alkaline phosphatase labelled streptavi- substitution of FITC conjugated random din method using New Fuchsin (Biogenex) as oligonucleotides (Dako), and by RNase pre- substrate. treatment. Detection ofimmunoglobulin light chain mRNA in fine needle aspiration cytology specimens 751

Results proportion of cells expressing light chain varied Table 1 summarises the pattern of immu- from about 20% to 80% of those present in the noglobulin light chain expression. cytospin preparations. Morphologically, the J Clin Pathol: first published as 10.1136/jcp.49.9.749 on 1 September 1996. Downloaded from cells expressing light chain mRNA included GROUP A: B CELL LYMPHOMA (N = 20) small lymphocytes, follicle centre cells, Of the 17 cases considered diagnostic of B cell centrocyte-like cells in MALT lymphomas, lymphoma on FNA, 14 showed light chain large nucleolated immunoblasts and plasmacy- restriction on ISH (K in 10 cases and X in four). toid cells in varying proportion in the different In all cases, light chain expression seemed to be cases. completely monotypic with no reaction for the No staining for either K or X was seen in the alternative light chain mRNA. In most tu- three remaining cases considered diagnostic of mours strong cytoplasmic staining was ob- B cell lymphoma, or in the three cases consid- served (figs 1 and 2) but in three cases staining ered suspicious of lymphoma on FNA. was more subtle, appearing as a narrow Light chain restriction was demonstrated in paranuclear rim in neoplastic cells (fig 3). The six of 14 cases diagnostic of, and in none of interpretative difficulty caused by background three cases suspicious of B cell lymphoma staining with immunocytochemistry was not using immunocytochemistry. In those prepara- observed in heavily blood stained aspirates tions considered monotypic, light chain using the ISH technique (fig 1). The mRNA was expressed in about 50% to 80% of

:K... A * *

0.I .i . !L= http://jcp.bmj.com/ on September 28, 2021 by guest. Protected copyright. C > e1

D

-.01. i...A 4

Figure 1 Large B cell lymphoma. (A) Heavily blood stained FNA specimen (May-Grunwald-Giemsa). (B) Cytological detail ofatypical lymphoid cells. (C) ISHfor K showing strong staining in most cells. (D) ISHfor k is negative. 752 Stewart, Farquharson, Kerr, McCorriston

cells. There was usually at least mild back- prised mature plasma cells. No staining was ground immunoreactivity with the alternate observed in cells or their vari-

Reed-Sternberg J Clin Pathol: first published as 10.1136/jcp.49.9.749 on 1 September 1996. Downloaded from light chain antiserum. In only one specimen (a ants. ) was monoclonal light Light chain expression was not observed in chain expression demonstrated by immunocy- the remaining specimen with either ISH or tochemistry but not ISH; a similar pattern was immunocytochemistry. observed in the biopsy material from this case. The same pattern oflight chain restriction (w REACTIVE LYMPHOID PROLIFERATION (N = 15) or X) was shown by ISH and immunocyto- A polytypic staining pattern was observed in 12 chemistry, and was also demonstrated in of 15 cases using ISH and in 11 of 14 cases histology material in six cases. with immunocytochemistry. As with the speci- mens from Hodgkin's disease, relatively few HODGKIN S DISEASE (N = 5) cells (less than 20%) expressed light chain A polytypic pattern of light chain expression mRNA and these mainly comprised mature was observed in four cases both with ISH and plasma cells with occasional larger cells of immunocytochemistry. There were few reac- probable follicle centre cell origin (fig 4). tive cells, accounting for up to 10% of those in No staining was observed in the remaining the cytospin preparations and mainly com- cases using either technique.

A B a-4fi

t

.... .a .... http://jcp.bmj.com/ I ..... 2.1

Figure 2 Follicular lymphoma. (A) ISHfor K showing strong staining in most cells. (B) ISHfor X is negative. on September 28, 2021 by guest. Protected copyright. A B

.:*::..1ON

Figure 3 MALT lymphoma. (A) Staining ofparanuclear cytoplasmic margin on ISHfor X in centrocyte-like cells. (B) ISHfor K is negative. Detection ofimmunoglobulin light chain mRNA in fine needle aspiration cytology specimens 753

C.:! ity may result from diffusion or active uptake of PV*+ ,,,* and as such is

%immunoglobulin non-specific.2" J Clin Pathol: first published as 10.1136/jcp.49.9.749 on 1 September 1996. Downloaded from % * * In this series light chain restriction was shown in only six (35%) of 17 cases of B cell v lymphoma by immunocytochemistry. Simi- - ^ * ^ larly, Pilotti et al' found immunochemistry of ^ 1* e js. limited value in differentiating lymphoid hy- perplasia and lymphoma. However, other cyto- logical studies have described light chain restriction in over 90% of B cell lymphomas using immunocytochemical methods.9 10 Re- cently, Robins et al demonstrated monoclonal- ity in 57 and 58 of 63 B cell lymphomas using ¢ immunocytochemistry in cytospin prepara- i4* j tions and flow cytometry, respectively.22 *,. ¢¢ . The ISH technique, by relying on detection of cellular mRNA, offers increased specificity compared with immunocytochemistry. In ad- dition, the complete absence of background immunoglobulin staining permits detection of a monotypic B cell population even when the tumour clone comprises a minority of cells in ISH has been Figure 4 Reactive lymph node. A plasma cell strongly present cytospin preparations. expresses K light chain mRNA. Other cells are unstained.A usedpsenttoicyospdemonstraterationslight chainShasmRNAbeN similar pattern was observed on ISHfor X. expression in histological sections of bone marrow16, salivary gland23 and lymphoid tis- Discussion sue,'9 21 but to our knowledge has not been The differentiation of reactive lymphadenopa- described previously in cytology material. In thy from lymphoma presents the most com- this series ISH demonstrated light chain mon diagnostic dilemma in the assessment of restriction in 14 of 20 B cell lymphomas diag- lymphoid FNA specimens. In cytology prepa- nosed on FNA cytology. The sensitivity of the rations the architectural pattern of lymph node technique was therefore approximately twice or other tissue is lost and suspicion of that of immunochemistry, and greater than lymphoma is usually based either on the that described with ISH in tissue sections of B relatively monomorphic nature of the lym- cell lymphoma by Weiss et al. 9 In addition, the phoid cell population or on the presence of presence of both K and X light chain in aspirates obvious cytomorphological abnormalities.'"7 was of value in supporting a diagnosis of reac-

However, some B cell lymphomas, particularly tive lymphoid proliferation in 12 of 15 cases. http://jcp.bmj.com/ follicle centre cell type, initially appear poly- However, a polytypic pattern was also seen in morphous and most aspirates also include an four of five cases of Hodgkin's disease and it is admixed population of reactive cells such as T important to appreciate that co-expression of lymphocytes. In addition, overtly abnormal immunoglobulin light chain in FNA material lymphoid cells are not seen in most low grade does not exclude a diagnosis of lymph- lymphomas. The difficulty is illustrated by a oma.. 10 24 Interestingly, the tumour cells in the

recent large series of lymph node FNA in single case of nodular LP Hodgkin's disease in on September 28, 2021 by guest. Protected copyright. which low grade non-Hodgkin's lymphomas this study did not express immunoglobulin were the commonest cause of false negative light chain; this Hodgkin's subtype is now gen- diagnostic error.'8 Thus, ancillary techniques erally regarded as a lymphoma of B cell origin, which provide phenotypic and genotypic and light chain restriction was demonstrated analysis are of great value in the assess- by ISH in approximately half the biopsy cases ment of lymphoid proliferations in cytology reported by Hell et al.21 material.5 7 In summary, light chain expression may be The presence of immunoglobulin light chain demonstrated on lymphoid cells using an ISH restriction has been used to indicate a mono- technique on cytospin preparations. The use of clonal B cell proliferation as this generally ISH contributed to a 100% specificity for the implies a neoplastic rather than reactive diagnoses of B cell lymphoma, Hodgkin's lymphoid process. Immunocytochemistry is disease and reactive lymphoid proliferations in the established method of demonstrating light a series of 40 FNA specimens. ISH was more chain expression but may be of limited value in sensitive than immunocytochemistry (70% v some cases as a result of partial nodal involve- 35%) in detecting light chain restriction in the ment by lymphoma, predominance of reactive cases of B cell lymphoma. T cells in some tumours (T cell rich B cell lymphoma), non-secretory types of B cell lymphoma or background polytypic immu- 1 Orell SR, Sterrett GF, Walters MN-I, Whitaker D (eds). Manual and atlas noglobulin staining.9 '9 The latter creates the burgh: Churchill Livingstone,of fine needle1992:63-94.aspiration cytology. Edin- most common interpretative problem in our 2 Das DK, Gupta SK, Datta BN, Sharma SC. FNA cytodiag- nosis of non-Hodgkin's lymphoma and its subtyping experience, particularly in bloodstained FNA under Working Formulation of 175 cases. Diagn Cyto- specimens, and similar difficulty may be pathol 1991;7:487-98. encountered in tissue sections of mate- 3 Carter TR, Feldman PS, Innes DJ, Frierson HF, Frigy AF. biopsy The role of fine needle aspiration cytology in the diagnosis rial.20 In addition, cytoplasmic immunoreactiv- of lymphoma. Acta Cytol 1988;32:848-61. 754 Stewart, Farquharson, Kerr, McCorriston

4 Pilotti S, Di Palma S, Alasio L, Bartoli C, Rilke F. Diagnos- Southern blot analysis for antigen receptor, bcl-2, and tic assessment of enlarged superficial lymph nodes by fine c-myc gene rearrangements. Am J7 Clin Pathol 1990;

needle aspiration. Acta Cytol 1993;37:853-66. 93:754-9. J Clin Pathol: first published as 10.1136/jcp.49.9.749 on 1 September 1996. Downloaded from 5 Sneige N. Diagnosis of lymphoma and reactive lymphoid 15 Hu E, Horning S, Flynn S, Brown S, Warnke R, Sklar J. hyperplasia by immunocytochemical analysis of fine- Diagnosis of B cell lymphoma by analysis of immu- needle aspiration biopsy. Diagn Cytopathol 1990;6:39-43. noglobulin gene rearrangements in biopsy specimens 6 Skoog L, Tani E. The role offine-needle aspiration cytology obtained by fine needle aspiration. J Clin Oncol 1986; in the diagnosis ofnon-Hodgkin's lymphoma. Diagn Oncol 4:278-83. 1991;1:12-18. 16 Akhtar N, Ruprai A, Pringle JH, Lauder I, Durrant STS. In 7 Katz RL, Caraway NP. FNA lymphoproliferative diseases: situ hybridization detection of light chain mRNA in myths and legends. Diagn Cytopathol 1995;12:99-100. routine bone marrow trephines from patients with suspected myeloma. BrJ7 Haematol 1989;73:296-301. 8 Cafferty LL, Katz RL, Ordonez NG, Carrasco CH, 17 Pontifex AH, Haley L. Fine-needle aspiration cytology in Cabanillas FR. Fine needle aspiration diagnosis of lymphomas and related disorders. Diagn Cytopathol 1989; intraabdominal and retroperitoneal lymphomas by a mor- 5:432-5. phologic and immunocytochemical approach. Cancer 18 Steel BL, Schwartz MR, Ramzy I. Fine needle aspiration 1990;65:72-7. biopsy in the diagnosis of lymphadenopathy in 1,103 9 Sneige N, Dekmezian RH, Katz RL, Fanning TV, Lukeman patients. Role, limitations and analysis of diagnostic JL, Ordonez NG, et al. Morphologic and immunocyto- pitfalls. Acta Cytol 1994;38:76-81. chemical evaluation of 220 fine needle aspirates of malig- 19 Weiss LM, Movahed LA, Chen YY, Shin SS, Stroup RM, nant lymphoma and lymphoid hyperplasia. Acta Cytol Bui N, et al. Detection of immunoglobulin light-chain 1990;34:311-22. mRNA in lymphoid tissues using a practical in situ 10 Oertel J, Oertel B, Kastner M, Lobeck H, Huhn D. The hybridisation method. Am Jf Pathol 1990;137:979-88. value of immunocytochemical staining of lymph node 20 Warnke RA, Rouse RV. Limitations encountered in the aspirates in diagnostic cytology. Br J Haematol 1988; application of tissue section immunodiagnosis to the study 70:307-16. of lymphomas and related disorders. Hum Pathol 1985; 11 Tani EM, Christensson B, Porwit A, Skoog L. Immunocyto- 16:326-31. chemical analysis and cytomorphologic diagnosis on fine 21 Hell K, Pringle JH, Hansmann M-L, Lorenzen J, Colloby P, needle aspirates of lymphoproliferative disease. Acta Cytol Lauder I, et al. Demonstration of light chain mRNA in 1988;32:209-15. Hodgkin's disease. J Pathol 1993;171:137-43. 12 Daskalopoulou D, Harhalakis N, Maouni N, Markidou SG. 22 Robins DB, Katz RL, Swan F Jr, Atkinson EN, Ordonez Fine needle aspiration cytology ofnon-Hodgkin's lympho- NG, Huh YO. Immunotyping oflymphoma by fine-needle mas. A morphologic and immunophenotypic study. Acta aspiration. A comparative study of cytospin preparations Cytol 1995;39:180-6. and flow cytometry. Am GClin Pathol 1994;101:569-76. 13 Cartagena N, Katz RL, Hirsch-Ginsberg C, Childs CC, 23 Speight PM, Jordan R, Colloby P, Nandha H, Pringle JH. Ordonez N, Cabanillas F. Accuracy of diagnosis of malig- Early detection of lymphomas in Sjogren's syndrome by in nant lymphoma by combining fine-needle aspiration cyto- situ hybridisation for and light chain mRNA in labial sali- morphology with immunocytochemistry and in seclected vary glands. EurJ Cancer Oral Oncol 1994;30B:244-7. cases Southern blotting of aspirated cells. Diagn Cytopathol 24 Sneige N, Dekmezian R, El-Naggar A, Manning J. 1992;8:456-64. Cytomorphologic, immunocytochemical and nucleic acid 14 Williams ME, Frierson HF, Tabbarah S, Ennis PS. flow cytometric study of 50 lymph nodes by fine-needle Fine-needle aspiration of non-Hodgkin's lymphoma: aspiration. Cancer 1991;67:1003-7. http://jcp.bmj.com/ on September 28, 2021 by guest. Protected copyright.