BRIEF REPORT CASE REPORT

Transcription

1 BRIEF REPORT Alpha-Interferon Secreting Blastic Plasmacytoid Dendritic Cells Neoplasm. A Case Report With Histological, Molecular Genetics and Long-Term Tumor Cells Culture Studies Tony Petrella, MD,* Geneviève Hervé, MD,* Bernard Bonnotte, MD, PhD, François Girodon, MD, PhD, John Andrew Carlson, MD, Jean-Baptiste Bour, PhD,k Pierre Lebon, MD,** Francine Mugneret, MD, and Patrick Callier, PhD Abstract: We report a new case of blastic plasmacytoid dendritic cell neoplasm (BPDCN) with extensive immunophenoptyping, genotyping (karyotype, array-comparative genomic hybridization, and fluorescent in situ hybridization), and long-term tumor cells culture. BPDCN is a very rare and aggressive disease clinically characterized by a skin revealing localization more or less rapidly disseminating to the bone marrow and other organs with or without and leukemia. The disease was initially phenotypically characterized by the expression of both CD4 and CD56 antigens, whereas lymphoid and myeloid lineage antigens were negative. A phenotypic link with alpha-interferon (IFN-I)-producing plasmacytoid dendritic cells was demonstrated. The data collected in this case report provide additional biological and genotypical data on tumor cells of BPDCN. This study confirms the capability of tumor cells to secrete IFN-I, demonstrated by biological IFN-I activity of cultured cells and immunohistochemical expression of Mx-1 protein. Although a common genetic profile involving chromosomes 5, 6, 9, 12, 13, and 15 has been identified, no specific genetic marker has been demonstrated that is specific to BPDCN. The demonstration of ETV6 gene deletion in this case deserves further investigations as a putative BPDCN marker. Key Words: blastic plasmacytoid dendritic cells neoplasm, alphainterferon, CD123, CD303, ETV6 gene (Am J Dermatopathol 2012;0:1 6) INTRODUCTION This disease named blastic plasmacytoid dendritic cell neoplasm (BPDCN) in the last WHO classification of Tumors of Hematopoietic and Lymphoid Tissues 1 and formerly known as CD4+CD56+ hematodermic neoplasm 2 or From the *Department of Pathology, CHU, Dijon, France; Center of Pathology, 33 rue Nicolas Bornier, Dijon, France; Department of Internal Medicine, CHU, Dijon, France; Department of Hematology Biology, CHU, Dijon, France; Department of Pathology, Albany, Medical College, Albany, NY; kdepartment of Virology, CHU, Dijon, France; **Department of Virology, Hospital St Vincent de Paul, APHP, Paris, France; and Department of Cytogenetics, CHU, Dijon, France. The authors declare no conflicts of interest. Reprints: Tony Petrella, MD, Department of Pathology, Plateau Technique de Biologie, 2 Rue Angélique Ducoudray, Dijon, France ( tony.petrella@chu-dijon.fr). Copyright 2012 by Lippincott Williams & Wilkins blastic NK-cell lymphoma 3 is a recently identified disease arising from interferon (IFN)-producing plasmacytoid dendritic cells (PDCs). It is a highly malignant disease, clinically characterized by a striking cutaneous tropism and rapid, lethal dissemination, with or without leukemic phase, within several months. It is a rare disease representing approximately 0.8% of the primary cutaneous lymphomas. 4 Initially, it was phenotypically defined by the coexpression of CD4 and CD56 antigens in the absence of any common myeloid, T-, B-, NK lymphoid lineage markers. 1 Recent phenotypical advances and availability of new antibodies, 5 working on paraffin sections and recognizing PDC-related antigens (eg, BDCA-2, BCL11a, CD123, CD2AP, and TCL1), now help pathologists to confirm the diagnosis of BPDCN. 1,6,7 Likely due to BPDCN s rarity, data are limited and variable regarding its ontogeny, its genotype, and the properties of its constituent tumor cells. In this case report, the authors document the results of extensive immunophenoptyping, genotyping [karyotype, array-comparative genomic hybridization (CGH), and fluorescent in situ hybridization (FISH)], and long-term tumor cell culture. CASE REPORT Clinical Data This 82-year-old man presented in November 2007 for the evaluation of a reddish, purpuric cutaneous nodule of the forehead, which appeared several weeks earlier and weakness. At the time of consultation, the nodule measured 3 cm in diameter (Fig. 1A). A complete blood cell count displayed anemia and thrombocytopenia and no circulating blast cells. A skin biopsy demonstrated BPDCN. During the 5-week interval between the diagnosis and hospitalization, the patient developed, disseminated purpuric lesions (Fig. 1B) and rapidly increasing circulating blast cells. Staging revealed bone marrow involvement but not lymph node enlargement or hepatosplenomegaly. Treatment by mini-chop (adriamycine, cyclophosphamide, vincristine, prednisone) was started 1 month after the diagnosis. Disease remission occurred after 3 courses of chemotherapy with both skin lesions and circulating blast cells quickly clearing after the first course of therapy. Five months after the diagnosis, while still in remission, he developed meningoencephalic tuberculosis and disseminated intravascular coagulation. Seven months after the diagnosis, the disease relapsed and he expired. Am J Dermatopathol Volume 0, Number 0, Month

2 Petrella et al Am J Dermatopathol Volume 0, Number 0, Month 2012 FIGURE 1. Nodule of the forehead (A), purpuric lesions of the upper chest (B). Histology and Immunohistochemistry A snap-frozen skin biopsy, a formalin-fixed paraffin-embedded skin biopsy, and a formalin-fixed paraffin embedded bone marrow biopsy were available for investigation. The routine skin biopsy showed that the tumor infiltrated the dermis and obliterated adnexal structures but spared the epidermis and the subepidermal region (Fig. 2A). There were no features of angiodestruction. Tumor cells were predominantly medium sized (Fig. 2B) with a few large cells and had minimal cytoplasm without granulation on Giemsa stain. Nuclei were irregular, with finely dispersed chromatin, and had several small or medium-sized nucleoli. Mitotic figures were rare. No plasma cells or eosinophils were found. Histochemical butyrate esterase staining was negative on frozen sections. Histological study of the bone marrow biopsy showed a discrete infiltration by similar appearing tumor cells. Blood smears of the blastic leukemic phase showed lymphoid-appearing undifferentiated blasts with round or slightly irregular eccentric nuclei containing small nucleoli. Nuclei were surrounded by slightly or moderately gray blue cytoplasm. Occasionally, cytoplasmic microvacuoles could be seen. Histochemical myeloperoxidase and butyrate esterase stains were both negative. The immunohistochemistry was performed on paraffin sections using a Benchmark automate (Ventana Medical System, FIGURE 2. Biopsy of the forehead nodule. A, There is a dense infiltrate of atypical cells within the dermis with a Grenz zone (H&E 120). B, High power magnification shows a proliferation of medium-sized blast cells with irregular nuclei and small nucleoli (H&E 630). Illkich, France) (Table 1). Tumor cells expressed: CD4, CD56, CD68 (dots), CD123 (Fig. 3A), CD303 (BDCA2) (Fig. 3B), TCL1 (Fig. 3C), and CD2AP. Mib-1 labeled 10% of the blastic cells. No tumor cells expressed CD2, CD3, CD5, CD7, CD20, CD33, TIA-1, granzyme B, TdT, myeloperoxidase. Tumor cells also expressed Mx-1 protein (Fig 3D). Mx-1 is a GTPase induced by IFN-I, which plays an important antiviral role in host defense. 8 Immunocytochemistry was also performed on air-dried cytospin smears of cultured cells with CD4 CD56, CD123, and CD303. Cultured cells also expressed these 4 antigens. EBER In Situ Hybridization and Polymerase Chain Reaction Studies Epstein-Barr virus encoded small RNA in situ hybridization was negative. By polymerase chain reaction (PCR), no clonal T or B-cell genes rearrangements were detected. 2 Ó 2012 Lippincott Williams & Wilkins

3 Am J Dermatopathol Volume 0, Number 0, Month 2012 Alpha-Interferon Secreting BPDCN TABLE 1. Immunohistochemistry Study Antibodies Source Results CD2 Menarini 2 CD2AP Pr David Mason (Oxford) + CD3 Dako 2 CD4 Menarini + CD5 Menarini 2 CD7 Menarini 2 CD8 Menarini 2 CD20 Dako 2 CD33 Dako 2 CD56 Menarini + CD68 Menarini + (dots) CD123 Clinisciences + CD303 Dendritics + Granzyme B Menarini 2 Mib-1 Dako 10% Myeloperoxidase Dako 2 Mx-1 protein Atlas + TCL1 Pr MikeTeitell (UCLA) + TdT Menarini TIA1 Beckman Coulter 2 Beckman Coulter, Brea, United States; Clinisciences, Montrouge, France; Dako, Glostrup, Denmark; Dendritics, Lyon, France; Menarini Diagnostics, Rungis, France. Karyotype, Array-CGH, and Fluorescent In Situ hybridization Hybridization Karyotyping Cytogenetic analyses of involved bone marrow samples using R and G banding techniques showed normal karyotype. Comparative Genomic Hybridization To carry out array-cgh on DNA extracted from snap-frozen skin biopsy and bone marrow samples, we used the IntegraChip genomewide bacterial chromosome (BAC) array of 4898 BAC clones (600-kb median spacing) of IntegraGen (Evry, France). The genomic DNA-labeling Kit (Enzo, Farmingdale, NY) was used to label 1 mg of the reference DNA with Cy3. The labeled products were purified with QiaQuick PCR purification kits (Qiagen GmbH, Hilden, Germany). After the addition of trna (48 mg) plus a quantity of Human COT DNA (Invitrogen, Paisley, United Kingdom) equal to 50 times the mass of the labeled DNAs, the products were concentrated by vacuum centrifugation, resuspended in 43 ml of hybridization solution (Ambion, Austin, TX) and deposited on the array under a coverslip. Hybridization in a sealed chamber (Corming, Schiphol-Rijk, The Netherlands) lasted for 66 hours at 55 C and posthybridization treatment was that of a published protocol withminor modifications. The slides were scanned with a confocal laser scanner (Agilent Technologies, Santa Clara, CA) and analyzed with GenoCensus (IntegraGen). Data were extracted with GenePix 6 (Axon Instruments, Westburg BV, Leusden, The Netherlands), and copy number analysis was performed with GenoCensus (Integra- Gen), which includes block Loess normalization. For the skin biopsy, the array-cgh study demonstrated losses on chromosomes 6, 12, and 13: del (6) (q11.1q16.3) (37 Mb), del(12) (p12.1p13.31) (13 Mb), del (13) (q13.1q13.3) (6 Mb), del (13) (q14.11q14.12) (5 Mb), and del (13) (q21.1q21.31) (9 Mb). For the bone marrow sample, array-cgh study detected the same losses as the skin sample and additional losses on chromosomes 2 and 5: del (2) (p11.2) (2.7 Mb), del (2) (p16.1) (4.3 Mb), del (2) (p21p22.1) (5.3 Mb), del (2) (p24.1) (2.8 Mb), del (5) (q23.2) (3 Mb), del (5) (q31.3q33.1) (7.5 Mb), and del (5) (q33.2q33.3) (7 Mb). Fluorescent In Situ Hybridization FISH analysis was performed by using bacterial artificial chromosome (BACs) probes of the region for chromosomes 2p, 5q, 13q using and commercial probes for 12p (ETV6, DAKO) and 17p (TP53, Vysis) according to the manufacturer s instructions. For bone marrow, metaphase spreads were hybridized overnight at 37 C with labeled. After posthybridization washes, the chromosomes were counterstained with DAPI (Vectashield, Vector, Burlingame, CA). Cell images were captured using the Isis Digital Imaging System (Metasystem, Altlusssheim, Germany). FISH analysis on bone marrow samples confirmed all the deletions with specific probes for chromosomal regions 2p, 5q, 12p (ETV6), and 13q. However, the 17p (TP53) deletion was not detected. Furthermore, ETV6 (red and green) and CEP 12 (aqua) FISH probes showed loss of 12p region in 80% of interphase nuclei, either monoallelic (50%) or bi-allelic (30%) (Fig. 4). FISH was also performed on snap-frozen skin sample sections, exhibiting the same deletions and matched the distribution of monoallelic and biallelic deletion of ETV6. Long-Term Tumor Cells Culture Tumor cells were isolated from leukocyte-rich decanted heparinized plasma. They were cultured on feeder-layer (MRC5) at 37 C (5% CO 2 ) in RPMI completed with 10% MRC5-supernatant(minimal essential medium containing 7.5% fetal calf serum). IL3 at a concentration of 10 ng/ml was added in the medium of culture during the 4 first months of culture. Biological alpha-ifn (IFN-I) activity was assayed in the supernatant of culture at 2 weeks and 24 weeks. IFN-I levels were measured by assessing the reduction of the cytopathic effect of vesicular stomatitis virus infected Madin-Darby bovine kidney cells as previously described. This method of reference, used in the laboratories of virology, is the most sensible for detection of IFN-I. 9 Results are given in international units per milliliter with a threshold of 2 IU/mL. Dosage was repeated twice at each time. Free RPMI with 10% MRC5-supernatant was used as negative control. Clusters of blastoid cells were obtained immediately. Four clusters (oligoclones) were picked up and cultured separately in the same way, adding 20% fetal calf serum. Tumor cells were cultured during 6 months through 17 passages. At the fifth month, the growth began to decrease progressively. When stressed by flask shaking or heating, the tumor cells tended to acquire dendritic morphology and stick to the MRC5 layer. After 6 months, the growth was over. Results of the IFN-I activity showed significant levels. At 2 weeks, the activity was of 50 IU/mL and decreased progressively to 6 IU/mL at 24 weeks (Fig. 5). Negative controls showed 0 IU/mL. DISCUSSION Since the first description of the entity in 1999, 10 significant improvements have been made in the knowledge of the disease leading to the recognition of a specific tumor entity likely originating from PDC (or its precursors/stem cell). 1 Clinically, BPDCN is a well-defined hematological tumor characterized by strong skin tropism, whose presence is heralded in most patients by a skin tumor that rapidly spreads to involve lymph nodes and/or bone marrow and/or peripheral blood. BPDCN s distinct immunophenotype, CD4+, CD56+, PDC markers+, and Ó 2012 Lippincott Williams & Wilkins 3

4 Petrella et al Am J Dermatopathol Volume 0, Number 0, Month 2012 FIGURE 3. Immunohistochemistry. The lesional cells strongly express CD123 (A), CD303 (B), TCL1 (C, nuclear and cytoplasmic staining), Mx-1 Protein (D). negative pan B-, pan T-cell, and myelo-monocytic-cell markers, sets it apart from other hematologic malignancies. 6 Based on clinical and phenotypical findings, the case presented here is a typical case of BPDCN. The disease began with a skin nodule that rapidly disseminated to other cutaneous surfaces, bone marrow, and blood. The pattern of skin involvement, the tumor cell cytology, and its immunophenotype are all characteristic of BPDCN. Moreover, tumor cells expressed the PDC markers CD303 (BDCA2), CD123, CD2AP, and TCL1, which is a reference panel for the diagnosis of BPDCN. 6 Genetic data are limited for this malignancy. Karyotype often shows complex aberrations, but specific abnormalities have not been identified. However, 6 major recurrent chromosomal targets are recognized, namely, 5q (72%), 12p (64%), 13q (64%), 6q (50%), 15q (43%), and 9 (28%). 11 A first array-cgh and gene expression profiling study conducted on 5 patients from Dijkman et al 12 demonstrated recurrent deletion of regions on 4q34, 9p13-p11, 9q12- q34, and 13q12-q31, resulting in the loss of tumor suppressor genes such as RB1 and LATS2. Jardin et al 13 performed highresolution array-cgh and quantitative multiplex polymerase chain reaction of short fluorescent fragments (QMPSF) on 9 BPDCN cases, showing losses of multiple genes involved in the G1/S transition. Especially common were deletions of chromosomes 9 (harboring the CDKN2A and CDKN2B genes), 13q (containing the RB1 locus), 17p (containing TP53), or 12p (including the tumor suppressor genes CDKN1B and ETV6). In our case, the karyotype did not show any chromosomal rearrangement, but array-cgh profile demonstrated deletions, in accordance to what should be expected from the previous studies cited above, particularly for chromosomes 5, 6, 12, and 13 (without RB1 deletion). However, contrarily to Jardin et al, 13 Wiesner et al, 14 and Lucioni et al, 15 we did not detect any abnormalities on chromosome 9 and chromosome 17 (P53 gene) by array-cgh and FISH analysis. In their recent study, Lucioni et al showed that the 9p21.3 deletion (p16-ink4a gene) had a worse prognosis. Although our patient did not demonstrate this deletion, he still died very quickly, 7 months after the diagnosis. 4 Ó 2012 Lippincott Williams & Wilkins

5 Am J Dermatopathol Volume 0, Number 0, Month 2012 Alpha-Interferon Secreting BPDCN FIGURE 4. FISH analysis performed on tissue sections. Interphase nuclei hybridized with alpha satellite centromeric probes for chromosome 12 (CEP 12, Spectrum Aqua, Vysis) (aqua spots) and LSI ETV6 probe (Dako). The FISH DNA probes are a mixture of a Texas Red-labeled DNA probe (ETV6-Upstream) (red spot) covering 264-kb telomeric to the ETV6 breakpoint cluster region and a fluorescein-labeled DNA probe (ETV6- Downstream) (yellow spot) covering 483-kb centromeric to the ETV6 breakpoint cluster region. Cells without deletion show 2 blue spots (CEP 12) and in each blue spot a yellow and red spot (ETV6). Cells with a biallelic deletion show 2 blue spots (CEP 12) without a yellow and red spot (ETV6). Cells with a monoallelic deletion show 2 blue spots (CEP12) with only 1yellowandredspot(ETV6). In our patient, the CGH study showed a discrepancy between skin and bone marrow with additional genomic abnormalities found in bone marrow involving chromosomes 2 and 5. The acquisition of additional genetic defects in the disseminated disease may reflect tumor progression and a high degree of genomic (chromosomal) instability. This finding may support the hypothesis, previously discussed, 4 that BPDCN could be a primary skin disease, which secondary involves other organs and blood. The CGH analysis showed, on skin and bone marrow, 12p12.1p13.31 region deletion, which contains CDKN1B and ETV6 tumor suppressor genes. In hematological malignancies, alterations of ETV6 gene are frequently observed. 14 One allele is involved in a translocation resulting in the formation of a new fusion gene, whereas the second allele is lost. 16 However, in BPDCN, translocations involving ETV6 have not yet been reported by karyotyping. 14 These concomitant translocations and deletions led to the suggestion that ETV6 may act as a tumor suppressor. However, the oncogenetic relevance seems to be related to the formation of fusion genes, not deletions. 16 ETV6 gene acts as a transcriptional repressor, and the relevance of the ETV6 loss alone is at present unclear We cultured in vitro tumor cells during 6 months. The main goal of this culture was to look for IFN-I production by tumor cells. We could demonstrate by biological dosage that FIGURE 5. Results of IFN-I biological activity in the supernatant of the culture (tumor cells and control). The chart shows a very high IFN-1 activity at 2 weeks of culture (50 IU/mL and control at 0 IU/mL) and a lower but still abnormal IFN-1 activity at 24 weeks of culture (6 IU/mL and control at 0 IU/mL). the tumor cells in the supernatant of culture secreted a high amount of IFN-I. The concentration of INF-I decreased in parallel to the impoverishment of the culture cell. Maeda et al 17 published in 2004 a cell line named CAL-1. They showed that CAL-1 cells were able to secrete large amount of alpha tumor necrosis factor, but little IFN-I after stimulation by TLR9 ligand of CPG ODNs. However, they utilized a commercially available enzyme-linked immunosorbent assay kit, instead of the assay used in this report. Interestingly, they also showed that CAL-1 cells could transform into mature dendritic cells, which we also observed in our cultured cells. Heat shock and mechanical stress from culture flask shaking triggered growth of long dendrites, which had a strong tendency to stick to the bottom of the flask like dendritic cells. This capability of tumor to mature into dendritic cells was also noticed in a short culture by Chaperot et al. 18 Inducing maturation of blast cells in vivo might be a key for the treatment of the disease as it could be for other hematologic malignancies such as acute promyelocytic leukemia (AML3). 19 CONCLUSIONS The data collected in this case report provides additional biological and genotypic data on tumor cells of BPDCN. This study confirms the capability of tumor cells to secrete IFN-I, demonstrated by the biological IFN-I activity of cultured cells and immunohistochemical expression of Mx-1 protein. Although a common genetic profile involving chromosomes 5, 6, 9, 12, 13, and 15 has been identified, no specific genetic marker has been demonstrated that is specific to BPDCN. Nonetheless, our finding of ETV6 gene deletion deserves further investigation as a putative BPDCN marker. REFERENCES 1. Facchetti F, Jones D, Petrella T. Blastic plasmacytoid dendritic cell neoplasm. In: Tumors WHOCo, ed. WHO Classification of Tumors of Haematopoietic and Lymphoid Tissues. 4th ed. Lyon, France: IARC Press; 2008: Ó 2012 Lippincott Williams & Wilkins 5

6 Petrella et al Am J Dermatopathol Volume 0, Number 0, Month Willemze R, Jaffe ES, Burg G, et al. WHO-EORTC classification for cutaneous lymphomas. Blood. 2005;105: Jaffe ES, Ralfkiaer E. Mature T-cell and NK-cell neoplasms. In: Tumors WHOCo, ed. Pathology and Genetics of Tumors of Haematopoietic and Lymphoid Tissues. 3rd ed. Lyon, France: IARC press; 2001: Petrella T, Bagot M, Willemze R, et al. Blastic NK-cell lymphomas (agranular CD4+CD56+ hematodermic neoplasms): a review. Am J Clin Pathol. 2005;123: Marafioti T, Paterson JC, Ballabio E, et al. Novel markers of normal and neoplastic human plasmacytoid dendritic cells. Blood. 2008;111: Petrella T, Facchetti F. Tumoral aspects of plasmacytoid dendritic cells: What do we know in 2009? Autoimmunity. 2010;43: Jegalian AG, Facchetti F, Jaffe ES. Plasmacytoid dendritic cells: physiologic roles and pathologic states. Adv Anat Pathol. 2009;16: Haller O, Staeheli P, Kochs G. Interferon-induced Mx proteins in antiviral host defense. Biochimie. 2007;89: Diop OM, Ploquin MJ, Mortara L, et al. Plasmacytoid dendritic cell dynamics and alpha interferon production during Simian immunodeficiency virus infection with a nonpathogenic outcome. J Virol. 2008;82: Petrella T, Dalac S, Maynadie M, et al. CD4+ CD56+ cutaneous neoplasms: a distinct hematological entity? Groupe Francais d Etude des Lymphomes Cutanes (GFELC). Am J Surg Pathol. 1999;23: Leroux D, Mugneret F, Callanan M, et al. CD4(+), CD56(+) DC2 acute leukemia is characterized by recurrent clonal chromosomal changes affecting 6 major targets: a study of 21 cases by the Groupe Francais de Cytogenetique Hematologique. Blood. 2002;99: Dijkman R, van Doorn R, Szuhai K, et al. Gene-expression profiling and array-based CGH classify CD4+CD56+ hematodermic neoplasm and cutaneous myelomonocytic leukemia as distinct disease entities. Blood. 2007;109: Jardin F, Callanan M, Penther D, et al. Recurrent genomic aberrations combined with deletions of various tumour suppressor genes may deregulate the G1/S transition in CD4+CD56+ haematodermic neoplasms and contribute to the aggressiveness of the disease. Leukemia. 2009;23: Wiesner T, Obenauf AC, Cota C, et al. Alterations of the cell-cycle inhibitors p27(kip1) and p16(ink4a) are frequent in blastic plasmacytoid dendritic cell neoplasms. J Invest Dermatol. 2010;130: Lucioni M, Novara F, Fiandrino G, et al. Twenty-one cases of blastic plasmacytoid dendritic cell neoplasm: focus on biallelic locus 9p21.3 deletion. Blood. 2011;118: Bohlander SK. ETV6: a versatile player in leukemogenesis. Semin Cancer Biol. 2005;15: Maeda T, Murata K, Fukushima T, et al. A novel plasmacytoid dendritic cell line, CAL-1, established from a patient with blastic natural killer cell lymphoma. Int J Hematol. 2005;81: Chaperot L, Bendriss N, Manches O, et al. Identification of a leukemic counterpart of the plasmacytoid dendritic cells. Blood. 2001;97: Fenaux P, Jouet JP, Zandecki M, et al. Chronic and subacute myelomonocytic leukaemia in the adult: a report of 60 cases with special reference to prognostic factors. Br J Haematol. 1987;65: Ó 2012 Lippincott Williams & Wilkins