research paper

Immunogenetic factors determining the evolution of T-cell large granular lymphocyte leukaemia and associated cytopenias

Zachary P. Nearman,1 Marcin Wlodarski,1,3 Anna M. Jankowska,1 Evan Howe,1 Yadira Narvaez,1 Edward Ball2 and Jaroslaw P. Maciejewski1 1

Experimental Hematology and Hematopoiesis

Section, Taussig Cancer Center, 2Allogen Laboratories, Cleveland Clinic, Cleveland, OH, USA, and 3Institute of Immunology Charite Medical School, Berlin, Germany

Received 17 August 2006; accepted for publication 29 September 2006; first published online 30 November 2006 Correspondence: Jaroslaw Maciejewski, MD, PhD, Experimental Hematology and Hematopoiesis Section, Taussig Cancer Center

Summary T-cell large granular lymphocyte leukaemia (T-LGL) is a chronic clonal proliferation of cytotoxic T lymphocytes (CTL). T-LGL presents with cytopenias, often accompanied by autoimmune diseases, suggesting clonal transformation arising from an initially polyclonal immune response. Various immunogenetic predisposition factors, previously described for both immune-mediated bone marrow failure and autoimmune conditions, may promote T-LGL evolution and/or development of cytopenias. The association of T-LGL was analysed with a number of immunogenetic factors in 66 patients, including human leucocyte antigen (HLA) and killer-cell immunoglobulin-like receptor (KIR) genotype, KIR/KIR-L mismatch, CTLA4 (+49 A/G), CD16 )158V/F, CD45 polymorphisms, cytokine single nucleotide polymorphisms including: TNF-a ()308G/A), TGF-b1 (codons 10 C/T, 25 G/C), IL-10 ()1082 G/A), IL-6 ()174 C/G), and IFN-c (+874 T/ A). A statistically significant increase in A/A genotype for TNF-a )308, IL10–1082, andCTLA-4 +49 was observed in T-LGL patients compared with control, suggesting that the G allele serves a protective role in each case. No association was found between specific KIR/HLA profile and disease. KIR/ KIR-L analysis revealed significant mismatches between KIR3DL2 and KIR2DS1 and their ligands HLA-A3/11 and HLA-C group 2 (P ¼ 0Æ03 and 0Æ01 respectively); the biological relevance of this finding is questionable. The significance of additional genetic polymorphisms and their clinical correlation to evolution of T-LGL requires future analysis.

R-40, Cleveland Clinic, 9500 Euclid Avenue, Cleveland, OH, USA. E-mail: [email protected]

Keywords: cytokines, single-nucleotide polymorphisms, human leucocyte antigen, leukaemia.

The inability to successfully clear inciting antigen or an exaggerated immune response were hypothesised as possible pathogenetic mechanisms in T cell large granular lymphocyte leukaemia (T-LGL). The persistent presence of an increased clonal population of cytotoxic T-lymphocytes (CTLs) suggests the initiation of an effective antigen-driven immune response that fails to adequately terminate (Kanchan & Loughran, 2003). It is possible that T-LGL may evolve from initially polyclonal autoimmune responses, explaining the association of T-LGL with a number of autoimmune disorders, most commonly rheumatoid arthritis (RA) (Loughran et al, 1985; Rovensky et al, 1991; Loughran, 1993; Sampalo et al, 1995) and immune-mediated cytopenias, including neutropenia, anaemia and, less frequently, thrombocytopenia (23%, 41% and 53% in our cohort respectively).

Various genetically determined factors can modify the quality of the physiological and pathological immune responses and may also play an important role in the development of T-LGL. Examples of such immunogenetic factors include human leucocyte antigen (HLA) type, killer immunoglobulin-like receptor (KIR) genotype, and cytokine and cytokine receptor gene single nucleotide polymorphisms (SNP). It is well established that certain HLA alleles constitute predisposition to an exaggerated immune response, such as those observed in: RA (Weyand & Goronzy, 2000), multiple sclerosis (MS) (Haines et al, 2002), systematic lupus erythematosus (SLE) (Hartung et al, 1992), ankylosing spondylitis (AS) (Dequeker et al, 1978), and many other conditions (Ghodke et al, 2005; Thorsby & Lie, 2005). HLA molecules also

ª 2006 The Authors First published online 30 November 2006 Journal Compilation ª 2006 Blackwell Publishing Ltd, British Journal of Haematology, 136, 237–248 doi:10.1111/j.1365-2141.2006.06429.x

Z. P. Nearman et al serve as ligands for KIR. Inhibitory and stimulatory KIR variants modulate the immune response by suppressing or activating cytotoxicity, depending on their matching with KIRligands (KIR-L). For example, KIR3DL2 interacts with the A3 or A11 allotypes, both of which make up the KIR-L haplotype, A3/A11. The HLA-B haplotype consists of the Bw4+ and Bw6+ (Bw4)) epitopes; the presence of a single allele belonging to the Bw4+ epitope is necessary for effective KIR3DL1 binding (Barnstable et al, 1979). HLA-C alleles are divided into two separate groups, C1 and C2, based on the amino acid residues present at positions 77 and 80; C1: Ser77-Asn80, C2: Asn77Lys80. KIR2DL1/2DS1 recognise HLA-C2 alleles, while KIR2DL2/2DL3/2DS2 ligand recognise HLA-C1 alleles (Parham, 2005). Many immune gene and promoter polymorphisms have been discovered to be associated with susceptibility to autoimmune conditions. For example, SNPs located in the promoter region of Interleukin-18 (IL-18) are a known risk factor associated with child type 1 diabetes mellitus (Kretowski et al, 2002). Tumour necrosis factor receptor (TNFR) polymorphisms and SNP in the TNF-a gene promoters have been implicated in the development of several related autoimmune disorders, including RA (McDermott et al, 1999; Aguillon et al, 2006). Protein tyrosine phosphatase receptor type C (PTPRC or CD45) is a haematopoietic cell surface antigen whose multiple isoforms are considered to play important roles in signal transduction among lymphocytes (Penninger et al, 2001). Alternative splicing of the CD45 gene produces the following isoforms: CD45RA (exon 4 spliced) CD45RB (exon 5), CD45RC (exon 6), and CD45RO (exons 4–6) (Jacobsen et al, 2000). CD45RA is a high molecular weight isoform containing the fourth or A exon, and is expressed on naı¨ve T cells. The CD45R0 isoform, of low molecular weight, which results from T cell activation and the subsequent splicing of exon 4, is expressed primarily on memory T cells. SNP in the A exon may prevent splicing of the RA isoforms, and therefore decreased expression of the memory cell phenotype, CD45R0. Variant CD45 exon 4 (C77G) expression has been described to be associated with development of MS (Jacobsen et al, 2000; Vyshkina et al, 2004). Another polymorphism, in the sixth exon, A138G, encodes a Thr to Ala amino acid substitution, which results in increased expression of the activated T cell phenotype (R0, exons 4–6 spliced out), and has been proposed to be associated with autoimmune disorders, such as Graves’ disease and Hashimoto’s thyroiditis, and viral diseases, such as hepatitis B and C (Boxall et al, 2004). In bone marrow failure states, the role of SNP has also been investigated; ‘high secretor’ gene variants of interferon (IFNGc) and tumour growth factor-b (TGFB) were found to be over-represented among patients with aplastic anaemia (AA) and paroxysmal nocturnal haemoglobinuria (PNH) (Fermo et al, 2004). Polymorphisms of immunomodulatory receptors may lead to similar functional consequences. Cytotoxic T lymphocyte associated-4 protein (CTLA-4, CD152) gene variants (CTLA4 +49 G/A) may result in decreased CTLA-4 expression and 238

inability to terminate immune response. An association of this SNP with incidence of insulin-dependent (type1) diabetes mellitus (IDDM1) has been described (Nistico et al, 1996). SNP in the extracellular domain 2 (EC2) of CD16 (FccRIIIa) (nonsynonomous T to G SNP at amino acid position 176) interferes with normal receptor function (binding of IgG1 and IgG3 and anti-CD16), modulates antibody-dependent cytotoxicity (Wu et al, 1997), and thus may increase autoimmune disease susceptibility. We hypothesised that the evolution of T-LGL is directly or indirectly associated with the presence of certain immunogenetic factors, which can either predispose to or alter the clinical course of this disease. Furthermore, we postulated that such a predisposition is the result of a complex genetic trait in the presence of appropriate exogenous triggers that act together to produce manifestation of disease. Here, we examine the rationally selected clinical variants, immunophenotypic features and genotypes of several immunoregulatory molecules that could be involved in inflammation and apoptosis and ultimately promote or alter evolution of T-LGL.

Materials and methods Patients Informed consent was obtained for peripheral blood collection according to the protocols approved by the Institutional Review Board of the Cleveland Clinic Foundation; samples from 66 patients with T-LGL were collected (mean age 62Æ0 years, range 21–81 years). Diagnosis of T-LGL leukaemia was based on clinical and laboratory criteria, such as the presence of T-cell receptor (TCR) c chain rearrangement and detection of an expanded discrete cell population (characterised by expression of CD3, CD8, CD16 and CD57) as previously described (Berliner et al, 1986; Semenzato et al, 1997; Herling et al, 2004). Clinical features of patients studied are summarised in Table I.

Flow cytometric Vb typing The individual contribution of each of the 25 Vb-subfamilies currently identifiable by specific monoclonal antibodies (mAb) was determined as previously described (Wlodarski et al, 2005), and results were expressed as the percentage of a/b CD4+ or CD8+ cells. Fresh peripheral blood was stained for Vb flow cytometry according to manufacturer’s instructions (IOTest Beta Mark kit; Beckman-Coulter, Fullerton, CA, USA). Due to incomplete coverage of the entire Vb spectrum by the Vb mAb set, Vb polymerase chain reaction (PCR) was performed for Vb families 6, 15 and 24. Clonotypic sequences of expanded Vb and Jb families were determined, and expanded CDR3 clonotypes were detected in 35 patients, as reported previously by Wlodarski et al (2005).

ª 2006 The Authors Journal Compilation ª 2006 Blackwell Publishing Ltd, British Journal of Haematology, 136, 237–248

Immunogenetics of T-LGL

DNA extraction DNA was extracted from whole blood using the PureGene system (Gentra, Minneapolis, MN, USA). Samples were resuspended in reduced Tris-EDTA buffer and the concentration was measured using a ND-1000 spectrophotometer (Nanodrop Technologies, Wilmington, DE, USA).

any of the HLA-Cw alleles, Cw2, 4, 5, 6, 15, 17 and 18 would silence KIR2DL1 and 2DS1. For example, a patient whose HLA profile shows HLA Cw3, Cw7 would show KIR ligand HLA C1/ C1 constellation; Cw2, Cw 4 would translate into KIR ligand C2/C2 constellation, both making a theoretical KIR/KIR-L mismatch possible.

Cytokine genotyping HLA and KIR typing HLA class I and II typing was performed by PCR-sequence specific primers (PCR-SSP) (Allogen Laboratories, Cleveland, OH, USA). KIR genotyping was performed using PCR-SSP to identify the presence or absence of KIR genes (Dynal Biotech; Invitrogen, Carlsbad, CA, USA). Locus-specific primer sets were utilised to amplify 1Æ5 lg of genomic DNA for each sample. Genomic DNA is mixed with KIR PCR reaction buffer containing dNTPs and Taq DNA Polymerase, then applied to a 96-well tray containing 5 ll optimised primer solution for thermal cycling. (Detailed description of this SSP amplification based method was reported by (Gomez-Lozano & Vilches, 2002; Hsu et al, 2002; Marsh et al, 2003)). Following amplification, PCR products were loaded on a 2% agarose gel containing 8 ll ethidium bromide. Upon completion of electrophoresis, the gel was photographed and interpreted. A KIR profile for each patient and control subject was determined by detecting the presence or absence of specifically amplified KIR products in each of 21 lanes containing individual allele-specific KIR primers. The KIR genes, 2DL1, 2DL2, 2DL3, 2DL4, 2DL5, 2DS1, 2DS2, 2DS3, 2DS4, 2DS5, 3DL1, 3DL2, 3DL3, 3DS1 and pseudogenes 2DP1 and 3DP1 were studied.

KIR and KIR-L assignment Ligands for KIR3DL2, 3DL1, 2DL1, 2DS1, 2DL2, 2DL3 and 2DS2 are known (Farag et al, 2002). For the purpose of this study, KIR mismatch was defined as the presence of a certain KIR gene and the lack of its corresponding HLA ligand according to the method proposed by Parham (2005). For example, KIR3DL2 recognises the A3 and/or A11 allotypes. Consequently, lack of HLA-A3 or HLA-A11 combined with the presence of KIR3DL2 can be considered a KIR3DL2 mismatch. Similarly, KIR3DL1 interacts with the Bw4 epitope; presence of two Bw6 epitopes combined with KIR3DL1 expression would result in KIR3DL1 mismatch. KIR2DL2, 2DL3, and thus 2DS2 and 2DS3 all interact with group 1 HLA-C molecules, which have a Lys residue in position 80 (http://www.dorak.info/hla/c1c2.html). HLA-Cw 1, 3, 7, 8, 12, 13, 14 and 16 alleles make up HLA-C1. The presence of any of these three KIR genes and the absence of a group 1 HLA-C allele will result in a mismatch. Similarly, KIR2DL1 and 2DS1 interact with group 2 HLA-C molecules, those having an Asn residue in position 80. The presence of

An assay designed for the detection of SNP in five different cytokine genes was used to create a cytokine SNP profile of T-LGL patients (One Lambda Inc., Canoga Park, CA, USA). Sequence-specific oligonucleotide primers for amplification of certain alleles from TGFB1, TNFA, Interleukin-6 (IL6), Interleukin-10 (IL10), and IFNG were used. Primer pairs for the amplification of a target sequence are provided for a total of 16 PCR reactions per sample per assay. Following the PCR, amplified DNA fragments were electrophoresed on a 2Æ5% agarose gel with 8 ll ethidium bromide. Positive reactions for a specific allele were discerned by the presence of a band between the larger internal control band and the smaller primer dimer band.

CTLA4 PCR and sequencing Exon I of the CTLA-4 gene was amplified from total genomic DNA using sense primer 5¢-catcgtcattgtagctaagc-3¢ and antisense primer 5¢-tactcaagccgattagc-5¢ in a total volume of 25 ll. After initial denaturation at 94C for 5 min, 30 cycles of 94C for 30 s, 57C for 35 s and 72C for 40 s were performed. PCR products were purified and 3 ll were used in a sequencing reaction with 1 ll of BigDye (Applied Biosystems, Foster City, CA, USA) 1 ll of 3 lmol/l primer in a total volume of 10 ll; products were purified and run on ABI 7500 Sequencer as described previously (Wlodarski et al, 2005).

FccRIIIa – 158V/F genotyping Allele-specific amplification of the FccRIIIa gene (FCGR3A) was performed as previously described with minor modifications (Leppers-van de Straat et al, 2000). Briefly, PCR reaction was optimised using 50 ng of template DNA, 10 pmol of valine-specific or phenylalanine-specific primers, 2Æ5 mmol/l dNTPs, 2Æ5 mmol/l MgCl2 and 1 U Taq polymerase diluted in PCR buffer (Invitrogen) to a final volume of 25ll. For PCR amplification, an initial denaturation step of 5 min at 95C was followed by 35 cycles (94C for 30 s, 60C for 30 s, 72C for 30 s) and final extension at 72C for 8 min. Samples were run in pairs using either G allele-specific (5¢-ctgaagacacatttttactcccaac-3¢) or T allele-specific (5¢-ctgaagacacatttttactcccaaa3¢) reverse primers combined with an FccRIIIa-specific forward primer (5¢-tccaaaagccacactcaaagac-3¢). The PCR products of 73 base pairs in the T or G allele-specific reaction were separated on 2Æ5% agarose gels and visualised under ultraviolet light using BioRad ChemiDoc XRS. Subjects were

ª 2006 The Authors Journal Compilation ª 2006 Blackwell Publishing Ltd, British Journal of Haematology, 136, 237–248

239

240

Age (years)

50 64 61 74 61 65 75 56 31 66 31 55 53 63 42 77 76 70 37 58 64 78 71 56 75 67 76 67 78 71 71 46 56 51 74 76 75 72

Pt

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38

2/ NIP 6Æ7/2Æ1 2/2Æ7 14/2Æ7 1/1Æ5 2/ 13Æ2/2Æ5 NIP 18/2Æ5 2/2Æ7 3/1Æ1 1/2Æ7 3/2Æ1 20/2Æ1,2Æ7 9/2Æ5 17/2Æ3 3/1Æ2,2Æ7 23/1Æ1,2Æ7 22/2Æ7 7Æ1/2Æ7 14/1Æ2 14/1Æ2 13Æ2/ 1/ 17/ 3/1Æ5 3/2Æ6 3/2Æ7 18/2Æ7 NIP 4/ NIP NIP/1Æ6 13Æ1/ 7Æ2/ 13Æ2/ 3/2Æ4

Vb/Jb clone 64 NIP 87 19 52 26 85 89 NIP 78 80 63 95 53 25 46 81 60 20 83 76 72 47 55 98 15 49 38 33 27 NIP 11 NIP NIP 88 67 95 88

% Vb clone of total CTL

Table I. Clinical characteristics of T-LGL patients.

3602 NIP 1958 1320 2012 765 8534 5093 NIP 2161 3370 1836 4902 3636 2419 5568 3785 1954 935 6496 1100 950 1452 1856 4504 863 456 1900 1452 700 NIP 175 NIP NIP 3785 525 23 750 N/A

Abs. clone size by Vb flow 2702 N/A 500 1852 2735 1130 283 N/A 1940 720 3240 N/A 2650 980 7640 1010 1800 950 1006 5140 905 4320 2163 1080 3280 1960 100 5350 810 466 3200 850 7410 N/A N/A 848 20 575 N/A

LGL count Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y

TCR rearrangement N Y N N Y N N Y N Y N Y Y Y Y Y Y Y Y Y Y N Y N Y N N Y N N/A Y N N Y N/A Y Y N

Splenomegaly Anaemia – Anaemia; leucopenia; severe neutropenia Anaemia Anaemia; severe neutropenia – Anaemia Severe neutropenia Anaemia Anaemia; leucopenia; neutropenia Pancytopenia; severe neutropenia Severe neutropenia; thrombocytopenia Anaemia; severe neutropenia Anaemia – Anaemia; thrombocytopenia Anaemia Pancytopenia; severe neutropenia – Neutropenia Pancytopenia Neutropenia; thrombocytopenia Anaemia – Anaemia; neutropenia – Pancytopenia – Anaemia – – Anaemia; leucopenia; severe neutropenia Anaemia; neutropenia Leucopenia; neutropenia Anaemia Leucopenia; severe neutropenia Leucocytosis Pancytopenia; severe neutropenia

Haem presentation

RhAr; PNH

RhAr

MDS RhAr MDS MDS

RhAr RhAr

Neuropathy

UC MDS RhAr

Uterine cancer RhAr; Felty Thymoma; PRCA RhAr Hepatitis RhAr; MS

RhAr; Felty

Colon cancer RhAr

Co-morbidities

Z. P. Nearman et al

ª 2006 The Authors Journal Compilation ª 2006 Blackwell Publishing Ltd, British Journal of Haematology, 136, 237–248

ª 2006 The Authors Journal Compilation ª 2006 Blackwell Publishing Ltd, British Journal of Haematology, 136, 237–248

75 66 60 28 50 64 65 78 70 27 38 54 21 58 76 37 71 67 79 69 65 67 79 63 81 57 74 67

39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

N/A N/A NIP 14/2Æ1 2/2Æ7 21Æ3/2Æ1 2/2Æ3 NIP/1Æ2,2Æ3 1/1Æ5 N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A 9/1Æ1 13Æ1/ 13Æ1/2Æ7 13Æ1/ 1,9/

Vb/Jb clone N/A N/A NIP 90 90 30 38 NIP 70 N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A 64 19 65 98 21, 28

% Vb clone of total CTL N/A N/A NIP N/A N/A N/A N/A NIP N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A 916 351 N/A 706, 941

Abs. clone size by Vb flow N/A 98 470 4481 N/A N/A N/A N/A N/A 506 853 2304 386 1904 798 1453 2735 322 542 839 285 735 109 346 1396 N/A N/A 1050

LGL count Y Y Y Y Y Y Y N/A N/A Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y

TCR rearrangement N N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N N/A Y N/A Y

Splenomegaly Anaemia Anaemia; thrombocytopenia Severe neutropenia Neutropenia Anaemia Severe neutropenia Severe neutropenia Anaemia; neutropenia – – Severe neutropenia; thrombocytopenia Neutropenia Severe neutropenia Severe neutropenia Severe neutropenia – Neutropenia; thrombocytopenia – – Neutropenia Severe neutropenia; thrombocytopenia Thrombocytopenia Severe neutropenia; thrombocytopenia Anaemia; neutropenia Anaemia Pancytopenia Anaemia; neutropenia Anaemia

Haem presentation

MDS

SLE

MDS

Co-morbidities

MDS, myelodysplastic syndrome; RhAr, rheumatoid arthritis; PNH, paroxysmal nocturnal haemoglobinuria; MS, multiple sclerosis; UC, ulcerative colitis; PRCA, pure red cell aplasia; SLE, systemic lupus erythematosus; NIP, not in panel; N/A, not available. Pancytopenia was defined as a deficiency in all three blood cell lineages. Anaemia was defined as an absolute reticulocyte count