Follicular Dendritic Cell Immunohistochemical Markers in Angioimmunoblastic T-Cell Lymphoma

Transcription

1 RESEARCH ARTICLE Follicular Dendritic Cell Immunohistochemical Markers in Angioimmunoblastic T-Cell Lymphoma Megan L. Troxell, MD, PhD,* Erich J. Schwartz, MD, PhD,* Matt van de Rijn, MD, PhD,* Douglas T. Ross, MD, PhD, Roger A. Warnke, MD,* John P. Higgins, MD,* and Yasodha Natkunam, MD, PhD* Abstract: Angioimmunoblastic T-cell lymphoma is characterized by a paracortical proliferation of medium to large neoplastic T cells, often with clear cytoplasm, in a background of arborizing high endothelial venules, many surrounded by follicular dendritic cells (FDCs). IHC staining may be applied to highlight these extrafollicular FDCs, traditionally using CD21, or CD23. Several alternative FDC markers have been described, including CNA.42, cystatin A/acid cysteine proteinase inhibitor (ACPI, involved in antigen presentation), and fascin (an actin binding protein). The authors stained a collection of 45 angioimmunoblastic T-cell lymphomas with CD21, CD23, CNA.42, cystatin A, and fascin for direct comparison of FDC staining characteristics in this setting. CD21 highlighted the expected dendritic network of cell processes, within residual follicles and outside of follicles, often adjacent to proliferating vessels. CD23 exhibited similar staining quality but was less sensitive than CD21. CNA.42 showed only diffuse weak labeling of FDCs. Cystatin A stained the cytoplasm of follicular dendritic cells within and outside of follicles; however, staining was often not sharply localized to dendritic cell processes, and scoring was further complicated by reactivity with other cell types in over half of the cases. Likewise, fascin stained a variety of cell types, including strong staining of interdigitating dendriticlike cells, moderate staining of endothelial cells, and only weak staining of follicular dendritic cells within and outside of follicles. Thus, CD21 remains the most reliable marker of follicular dendritic cells in angioimmunoblastic T-cell lymphoma. Key Words: angioimmunoblastic T-cell lymphoma, follicular dendritic cell, interdigitating dendritic cell, cystatin A, fascin, CD21, immunohistochemistry (Appl Immunohistochem Mol Morphol 2005;13: ) The subclassification of T-cell lymphomas is continuing to evolve as morphologic, immunophenotypic, and molecularcytogenetic features are becoming better characterized. Angioimmunoblastic T-cell lymphoma is redefined in the REAL Received for publication April 15, 2005; accepted May 12, From the *Department of Pathology, Stanford University School of Medicine, Stanford, CA; and Applied Genomics Inc, Sunnyvale, CA. Correspondence: Yasodha Natkunam, Department of Pathology, Stanford University School of Medicine, L Pasteur Drive, L235 Stanford, CA ( yaso@stanford.edu). Copyright Ó 2005 by Lippincott Williams & Wilkins and WHO classification systems to include cases previously considered AILD-like T-cell lymphoma (angioimmunoblastic lymphadenopathy with dysproteinemia), and accounts for approximately 2% of non-hodgkin lymphomas. 1 4 The diagnosis of angioimmunoblastic T-cell lymphoma is problematic, and other entities, including atypical T-zone hyperplasia, Castleman s disease, peripheral T-cell lymphoma NOS, T-cell and histiocyte-rich large B-cell lymphoma, and Hodgkin lymphoma, enter the differential. 2,3 Morphologic features of angioimmunoblastic T-cell lymphoma include pleomorphic small to medium-sized tumor cells with clear cytoplasm and minimal cytologic atypia. 2 4 The polymorphous background includes an admixture of eosinophils, plasma cells, mast cells, and small lymphocytes, with occasional large B cells, some of which may mimic Reed-Sternberg cells; they have been shown to be EBV positive in many cases. Involved lymph nodes may contain variable numbers of residual B-cell follicles. 2 5 The neoplastic T cells are generally CD3 positive, with a predominance of CD4-positive cells, though reactive CD8-positive cells remain admixed; pan-t antigen loss is relatively rare. 3 Recent studies have shown CD10 co-expression in tumor cells, along with Bcl-6. 3,4,6 8 Architectural features are diagnostically important, including prominent arborizing high endothelial venules, and the identification of follicular dendritic cells (FDCs) present outside of lymphoid follicles, often seen in association with the vascular proliferation. 2 4,9 11 IHC staining has become essential for demonstration of follicular dendritic cell processes in these cases. Antibodies to CD21 (C3d complement receptor) or CD23 (low-affinity IgE receptor) have been well characterized as FDC markers in normal lymph nodes, hyperplastic lymph nodes, and various B-cell and T-cell lymphomas However, staining with these two markers is not always concordant. Novel antibodies to a number of other FDC proteins have become available. The monoclonal antibody CNA.42, originally described in 1997 by Raymond et al, is reactive with a 120-kd glycosylated antigen mainly expressed by FDCs and is functional in formalin-fixed paraffin-embedded tissue sections. 16 Cystatin A (acid-cysteineproteinase inhibitor, ACPI) received attention over 20 years ago as a 12-kd protease inhibitor involved in antigen presentation and localized to dendritic cells within skin, reactive tonsil, lymph node, and to a lesser extent follicular lymphoma Fascin is an actin bundling protein and is thought to be important in maintaining the structure of dendritic process in follicular and interdigitating dendritic cells. Fascin also labels Reed-Sternberg cells of classic Hodgkin lymphoma Appl Immunohistochem Mol Morphol Volume 13, Number 4, December

2 Troxell et al Appl Immunohistochem Mol Morphol Volume 13, Number 4, December 2005 We sought to investigate the FDC staining of these markers in the specific context of angioimmunoblastic T-cell lymphomas, in comparison with CD21 and CD23. METHODS Case Selection Cases of angioimmunoblastic T-cell lymphoma were collected from the surgical and consultation files of the Department of Pathology at Stanford University Medical Center from 1995 to Over 75 cases were identified, but cases were excluded if no additional unstained sections or block were available. This yielded a total of 45 cases, although not all cases had material available for all five stains (cystatin A/S147, fascin, CNA.42, CD21, CD23). Additionally, S100 staining was performed on six cases to highlight interdigitating dendritic cells. H&E and selected immunostains performed at the time of diagnosis were reviewed by two pathologists to confirm the diagnosis of angioimmunoblastic T-cell lymphoma, using criteria established by the WHO classification. 2 H&Estained slides were no longer available in eight cases; however, the IHC and morphology on counterstained sections allowed the diagnoses to be confirmed. Stanford University School of Medicine Institutional Review Board approval was obtained for this study. Production of Antiserum A rabbit affinity-purified polyclonal antibody was raised by injecting three peptides derived from the cystatin A gene sequence (Applied Genomics Inc, Hunstville, AL). The peptides were synthesized by standard FMOC chemistry: peptide 1, MIPGGLSEAKPATPEIQEIV; peptide 2, NETYGKLEAV- QYKTQ; peptide 3, DLVLTGYQVDKNKDDELTGF. The peptides were conjugated to KLH and injected into two outbred rabbits. The serum (S147) was harvested after the rabbits showed a significant anti-peptide titer. Affinity-purified antibodies were obtained by passing the antiserum over an affinity column conjugated with the three peptides; bound antibodies were eluted with a ph gradient. IHC Staining and Scoring Serial sections of 4 mm were cut and deparaffinized in xylene and hydrated in a series of graded alcohol. Primary antibodies, dilutions, and pretreatment are listed in Table 1. Endogenous peroxidase was blocked by preincubation with 1% hydrogen peroxide in phosphate-buffered saline. Detection was performed using a modified biotin-streptavidin method. 23 Detection for CD21 was performed on an automated staining machine (Ventana Medical Systems Inc, Tucson, Arizona). Positive controls for antibody staining included normal tonsil, and negative controls included omission of primary antibody. The extent and distribution of staining were evaluated for multiple cell types for each antibody, using a semiquantitative scale. Staining was evaluated in FDCs located within residual lymphoid follicles, as well as FDCs located outside of follicles, which were often seen in association with vessels. Initial observations indicated that fascin also showed reactivity with endothelial cells and interdigitating dendritic-like cells, and staining of these components was also tabulated. Cystatin A also stained non-follicular dendritic cells in a number of cases; these were of varied morphology and were simply tabulated scored as other. The intensity of staining was scored on a four-point scale (0, 1+, 2+, 3+). Cases with no detectable labeling at low power (403) were considered negative. The extent of staining in each compartment was estimated as focal within the node, patchy, or extensive in positive cases. For comparison between markers, cases were scored as concordant if the staining intensity was within 61 and the extent of staining was of the same category as, or more extensive than, CD21. If CD21-stained sections were not available, comparisons were made with CD23-stained sections. RESULTS Immunostaining for CD21 Thirty-nine of 45 cases of angioimmunoblastic T-cell lymphoma had CD21 staining available for analysis (Table 2). In our hands, CD21 reproducibly resulted in an intense brown stain, which localized preferentially to the dendritic processes of FDC cells. No staining of cells without dendritic morphology was observed. Within follicles, the network of FDCs appeared as a meshwork, with CD21 highlighting concentric interwoven cell processes (Fig. 1). Thirty-three of 39 cases (85%) showed CD21 staining of FDCs in residual follicles, most of which were atrophic; 6 cases lacked staining, with essentially no remaining follicular centers. As previously observed, CD21 also stained dendritic processes of FDC cells located outside of follicles in cases of angioimmunoblastic TABLE 1. Antibody Reagents Used in IHC Antibody Dilution Pretreatment Source CD21 (1F8) 1:10 Protease digestion, 4 min DAKO, Carpenteria, CA CD23 (MHM6) 1:40 DAKO Target Retrieval System DAKO, Carpenteria, CA CNA.42 1:20 DAKO Target Retrieval System DAKO, Carpenteria, CA Cystatin A (S147) 1:1,000 Microwave in citric acid buffer (10 mm, ph 6.0) for 10 min Applied Genomics Inc, Sunnyvale, CA Fascin (IM20) 1:200 Microwave in citric acid Novocastra, Newcastle, UK buffer (10 mm, ph 6.0) for 10 min S100 (polyclonal) 1:1,000 Trypsin digestion, 20 min DAKO, Carpenteria, CA 298 q 2005 Lippincott Williams & Wilkins

3 Appl Immunohistochem Mol Morphol Volume 13, Number 4, December 2005 FDC Markers in AITL TABLE 2. IHC Staining Results Stain and Pattern No. Positive/Total % Positive % Concordant CD21/23 CD21 Follicular 33/39 85 NA Extrafollicular 39/ NA CD23 Follicular 22/ Extrafollicular 25/ CNA.42 Follicular 22/ Extrafollicular 27/ Cystatin A Follicular 28/ Extrafollicular 40/ Other 23/43 53 NA Fascin Follicular 14/ Extrafollicular 35/ IDC-like 44/ NA Endothelial 30/44 68 NA T-cell lymphoma. CD21 staining generally showed intermeshed cell processes vaguely parallel to plump venules in involved nodes. The stain product remained crisply localized to these narrow processes. All stained cases in the study group showed extrafollicular CD21-positive FDC processes (100%), as CD21 staining was one of the inclusion criteria. Immunostaining for CD23 CD23 staining was available for 28 cases in our study group of angioimmunoblastic T-cell lymphoma. The immunomorphologic features of CD23 staining were similar to that of CD21, with the stain highlighting delicate cellular process within residual follicles and along proliferating vessels. CD23 showed FDCs in residual follicles in 22 of 28 cases (79%) and FDCs outside of follicles in 25 of 28 cases (89%). Twenty-two of 28 cases also had CD21 staining for direct comparison; staining was scored as concordant with CD21 within follicles in 15 cases and outside of follicles in only 13 cases (68%, and 59%, respectively). In 9 of these 22 cases (41%), CD23 staining was weaker and/or failed to highlight collections of extrafollicular FDCs marked by CD21. In only three cases was CD23 staining more robust than CD21. In rare individual cases, CD23 showed scattered small lymphocytes with cytoplasmic labeling, which did not interfere with interpretation of FDC staining. Immunostaining for CNA.42 Thirty-five cases stained with CNA.42 cases were available for analysis. Twenty-two of 35 (63%) cases showed labeling of FDCs within residual follicles; however, in 12 of these cases staining was only focal with weak (1+) intensity. Staining was diffuse and lacked crisp localization to dendritic processes. Staining was scored as concordant with CD21/CD23 in 13 of 35 cases (37%). In the remaining cases, CNA.42 staining failed to label follicles demonstrated by CD21 or CD23. CNA.42 showed FDCs cells outside of follicles in 27 of 35 stained cases. Once again, a significant fraction of these cases (9) showed only focal, weak staining of extrafollicular dendritic cell networks present on CD21/23-stained slides, for a concordance of 34% with CD21/CD23. Immunostaining for Cystatin A A novel antiserum, S147, reacting with the cystatin A protein was recently characterized by tissue microarray analysis. Forty-three cases were available for analysis with S147. Qualitatively, cystatin A stained follicular dendritic cells within and outside of residual follicles; staining was generally less intense and more diffuse than seen with CD21 or CD23. Light-brown staining was seen in broad areas in the vicinity of FDCs and was not well localized to delicate cell processes. Cystatin A highlighted FDCs within follicles in 28 of 43 cases (65%) and extrafollicular FDCs in 40 of 43 cases (93%). Staining was scored as concordant with CD21/CD23 in follicles in 51% of cases and concordant outside of follicles in 56% of cases. However, this also reflects that cystatin A was relatively less specific. Staining of individual cells, not part of a network, was also seen between follicles in areas negative for CD21 or CD23 in 53% of cases. In many cases, these positive cells also showed dendritic-like processes and were suggestive of interdigitating dendritic cells or fibroblastic reticulum cells (also see fascin, below). In 13 cases (30%) this cross reactivity prohibited interpretation of FDC staining. However, in five cases, cystatin A appeared to stain yet a different population of cells, with the morphology of medium-sized lymphocytes or monocytes, without evidence of cell processes. Immunostaining for Fascin Fascin staining was evaluated in 44 cases. Fascin was observed to stain FDCs, interdigitating dendritic cells, and endothelial cells (Fig. 2). Fascin stained FDCs both within and outside of follicles, in a distribution that could be correlated with CD21/CD23 staining on serial sections. However, fascin staining of FDCs was generally faint and poorly localized to cell processes. Fascin highlighted residual follicles in 14 of 44 cases (32%) and showed extrafollicular FDCs in 35 of 44 cases (80%). Extrafollicular FDC staining was concordant with CD21/CD23 in 18 of 44 (41%) cases; in the rest, it was markedly weaker than CD21/CD23 staining, or FDC networks could not be definitively identified on fascin-stained sections. This was related to difficulty in interpreting FDC staining associated with strong fascin labeling of other cell populations. Fascin labeled a population of individual cells with dendritic process outside of follicles. This staining was intensely dark brown and seen in all 44 cases, in most cases throughout the node. This population was not associated with residual follicles or proliferating vessels and likely represented interdigitating dendritic cells based on comparison with S100 staining in six cases. This strongly staining population obscured assessment of FDCs in nearly all cases. We also observed weak to moderate fascin staining of endothelial cells in 30 of 44 cases (68%). q 2005 Lippincott Williams & Wilkins 299

4 Troxell et al Appl Immunohistochem Mol Morphol Volume 13, Number 4, December 2005 FIGURE 1. Comparison of staining characteristics. A E, Low-power view showing residual follicle (arrow) together with extrafollicular FDC staining (arrowhead). There are numerous darkly staining interdigitating dendritic-like cells demonstrated by cystatin A, fascin, and S100 (original magnification 1003). F J, Perivascular FDC proliferation (original magnification 2003). K O, Representative case showing typical weak CNA.42, S147, and fascin staining of extrafollicular FDCs (original magnification 1503). 300 q 2005 Lippincott Williams & Wilkins

5 Appl Immunohistochem Mol Morphol Volume 13, Number 4, December 2005 FDC Markers in AITL FIGURE 2. Staining characteristics of other cell types. A, Fascin highlights FDCs in small follicle (arrow) as well as endothelial cells (arrowheads) and surrounding interdigitating dendritic-like cells. B, Fascin labels endothelial cells (arrowheads show subset of small vessels). C, S100 staining of field corresponding to A. D, Example of rare case in which cystatin A labeled non-dendritic, mononuclear cells (small arrows; large arrow shows follicle). (Original magnification 2003 A C, 4003 D). DISCUSSION The finding of FDCs associated with a proliferation of postcapillary venules outside of follicles in angioimmunoblastic T-cell lymphomas was incorporated into the histologic definition of T-cell lymphoma of AILD by Lennert et al in and confirmed by Leung et al in This characteristic distribution of FDCs has become accepted as one of the diagnostic features of angioimmunoblastic T-cell lymphoma and is part of the current WHO classification. 2 4 This study addresses the staining of FDCs by markers other than CD21 or CD23 in angioimmunoblastic T-cell lymphomas. Staining with cystatin A, fascin, and CNA.42 confirmed the distribution of FDCs in association with vessels, outside of follicles. None of these three markers had advantages over CD21 or CD23 for the diagnosis of angioimmunoblastic T-cell lymphoma. In our hands, CD23 localized precisely to dendritic processes, with similar architectural features to CD21, although CD23 staining tended to be weaker and less sensitive than CD21. However, in three individual cases, CD23 detected FDCs not labeled by CD21. Similarly, CNA.42 was also less sensitive than CD21. In addition, CNA.42 staining was often weak and diffuse, without crisp localization to dendritic processes, which made interpretation more difficult. Although cystatin A and fascin stained FDCs more robustly than CNA.42, interpretation was hampered by strong staining of other cell types. In the case of fascin, this population is most likely interdigitating dendritic cells, based on comparison with S100 staining. However, we cannot rule out the possibility that some of this population may overlap with fibroblastic reticulum cells, as described by Jones et al. 11 Fascin also labeled endothelial cells, which complicated the evaluation of dendritic cell staining in the vicinity of proliferating venules of angioimmunoblastic T-cell lymphomas. For cystatin A, some of the individual labeled cells also had dendritic-like processes. However, distribution did not exactly mirror that of IDC-like cells identified on serial sections stained with fascin; again, these may represent fibroblastic reticulum cells. Without extensive double-labeling experiments, we cannot be certain of their identity. Further, in five cases, a population of small to medium-sized round cells without processes was also labeled by cystatin A; although we did not attempt to further characterize these cells, morphologically they are consistent with monocytes/histiocytes. In our prior experience and that of others, CD23 is aberrantly co-expressed on cells of CLL/SLL in a significant proportion of cases, and CD23 labels dendritic processes within benign and neoplastic follicles. 2 4,14,15 Other than cytoplasmic staining of few small lymphocytes in individual cases, CD23 appeared to be limited to FDCs in angioimmunoblastic T-cell lymphomas. However, our data show that CD23 is less sensitive than CD21 as an FDC marker. This finding is likely to be clinically significant in cases of angioimmunoblastic T-cell lymphoma where FDC proliferation in association with vessels is subtle. Bagdi et al found reactivity of CD21, CD23, and CD35 with clusters of dendritic cells in association with postcapillary venules in all nine cases of angioimmunoblastic T cells in a study of FDC markers in formalin-fixed, paraffinembedded reactive and neoplastic lymph nodes 15 Our larger q 2005 Lippincott Williams & Wilkins 301

6 Troxell et al Appl Immunohistochem Mol Morphol Volume 13, Number 4, December 2005 study confirms these observations regarding CD21 and CD23; however, we did not have additional sections available for CD35 staining. Although cystatin A (ACPI) was previously characterized in dendritic cells within various lymphoid tissues and has also been reported to stain Reed-Sternberg cells in a proportion of cases of Hodgkin disease, 17 19,24 it has largely been overlooked in the current lymphoma literature. To our knowledge, this study is the first to report analysis of cystatin A in angioimmunoblastic T-cell lymphoma, although its clinical utility is likely to be limited for the reasons described above. The finding of fascin staining numerous IDC-like cells is reminiscent of the series of HIV-related lymphoid hyperplasias published by Said et al, including nine cases of AILD-like abnormal lymphoid proliferations in which a marked proliferation of interdigitating dendritic cells was observed. 25 Further, we were able to identify fascin staining of residual FDCs in follicles, as well as extrafollicular proliferations. Our study shows that the abnormally localized FDCs in angioimmunoblastic T-cell lymphoma maintain some degree of fascin expression, whereas FDCs in association with follicular B-cell lymphoma lose expression of fascin. 20 A single report has identified increased fascin expression in B cells infected by EBV in vitro; this observation may be relevant to the current series, as increased EBV-positive cells may be seen in the setting of angioimmunoblastic T-cell lymphoma. 2,26 29 However, this report has not been replicated by staining of tissue sections, and our cases were not systematically tested for EBV. CNA.42 was previously applied to seven cases of angioimmunoblastic T-cell lymphoma as part of its initial characterization by Raymond et al. 16 Labeling of expanded FDC networks was described in all of their tested cases; further, CNA.42 staining of neoplastic T cells was observed in other types of T-cell lymphoma, but not in angioimmunoblastic T-cell lymphoma. 16 Our larger study confirms these findings. Further, Maeda et al published a technical study including 22 different antibody reagents against proteins expressed on FDCs and a variety of antigen retrieval conditions in sections of normal tonsil and hyperplastic lymph node. 30 They reported that CNA.42 produced an intense reticular labeling within lymphoid follicles; however, in our hands this stain resulted in a weak, poorly localized stain product. This difference may result from differences in fixation or retrieval conditions, as well as case selection (lymphoma vs. normal/hyperplasia). However, our results are similar to their observations for CD21 (clone 1F8), CD23, and fascin, though different monoclonal antibodies were tested in the case of CD23 and fascin. 30 In our laboratory, CD21 is currently the standard marker applied to highlight FDCs in clinical cases. Despite the development of novel FDC markers, this study provides data supporting CD21 as the primary FDC marker, with CD23 as a secondary marker. Although we have specifically examined these reagents in angioimmunoblastic T-cell lymphomas, because we found similar results in FDCs localized to follicles as well as those outside of follicles, our results should be applicable to the staining of FDCs in the diagnosis of other hematolymphoid conditions. ACKNOWLEDGMENTS The authors thank Elizabeth Domanay, Lourdes Villanueva, and Edward Gilbert for expert technical assistance with immunohistochemical staining. REFERENCES 1. Harris NL, Jaffe ES, Stein H, et al. A revised European-American classification of lymphoid neoplasms: a proposal from the International Lymphoma Study Group. Blood. 1994;84: Jaffe ES, Ralfkiaer E. Angioimmunoblastic T-cell lymphoma. In: Jaffe ES, Harris NL, Stein H, Vardiman JW, eds. WHO Classification of Tumors. Pathology and Genetics: Tumours of Haematopoietic and Lymphoid Tissues Lyon: IARC Press, 2001: Ferry JA. Angioimmunoblastic T-cell lymphoma. Adv Anat Pathol. 2002; 9: Dogan A, Attygalle AD, Kyriakou C. Angioimmunoblastic T-cell lymphoma. Br J Haematol. 2003;121: Ree HJ, Kadin ME, Kikuchi M, et al. Angioimmunoblastic lymphoma (AILD-type T-cell lymphoma) with hyperplastic germinal centers. Am J Pathol. 1998;22: Ree HJ, Kadin ME, Kikuchi M, et al. Bcl-6 expression in reactive follicular hyperplasia, follicular lymphoma, and angioimmunoblastic T-cell lymphoma with hyperplastic germinal centers: heterogeneity of intrafollicular T-cells and their altered distribution in the pathogenesis of angioimmunoblastic T-cell lymphoma. Hum Pathol. 1999;30: Attygalle A, Al-Jehani R, Diss TC, et al. Neoplastic T cells in angioimmunoblastic T-cell lymphoma express CD10. Blood. 2002;99: Yuan C, Vergilio J, Harris NL, et al. Angioimmunoblastic T-cell lymphoma: a neoplasm of intrafollicular CD10+, Bcl-6+, CD4+ memory T-cells? Mod Pathol. 2002;15:271A. 9. Lennert K, Feller AC. T-cell lymphoma of AILD (LgX) type. In: Lennert K, Feller AC, eds. Histopathology of Non-Hodgkin s Lymphomas, 2d ed. Berlin: Springer-Verlag, 1992: Leung CY, Ho FCS, Srivastava G, et al. Usefulness of follicular dendritic cell pattern in classification of peripheral T-cell lymphomas. Histopathology. 1993;23: Jones D, Jorgensen JL, Shahsafaei A, et al. Characteristic proliferation of reticular and dendritic cells in angioimmunoblastic lymphoma. Am J Surg Pathol. 1998;22: Gerdes G, Stein H, Mason DY, et al. Human dendritic reticulum cells of lymphoid follicles: their antigenic profile and their identification as multinucleated giant cells. Virchows Arch Cell Pathol. 1983;42: Sellheyer K, Schwarting R, Stein H. Isolation and antigenic profile of follicular dendritic cells. Clin Exp Immunol. 1989;78: Knowles DM. Immunophenotypic markers useful in the diagnosis and classification of hematopoietic neoplasms. In: Knowles DM, ed. Neoplastic Hematopathology. Philadelphia: Lippincott Williams & Wilkins, 2001: Bagdi E, Krenacs L, Krenacs T, et al. Follicular dendritic cells in reactive and neoplastic lymphoid tissues: a reevaluation of staining patterns of CD21, CD23, and CD35 antibodies in paraffin sections after wet heatinduced epitope retrieval. Appl Immunohistochem Mol Morphol. 2001;9: Raymond I, Al Saati T, Tkaczuk J, et al. CNA.42, a new monoclonal antibody directed against a fixative-resistant antigen of follicular dendritic reticulum cells. Am J Pathol. 1997;151: Alavaikko M, Rinne A, Jarvinen M, et al. Acid-cysteine-proteinase inhibitor, a new characteristic of reticulum cells in lymphoid secondary follicles. Acta Histochem. 1985;77: Alavaikko M, Aine R, Jarvinen M, et al. Behavior of dendritic reticulum cells possessing immunoreactive acid cysteine-proteinase inhibitor in human lymphoid secondary follicles and in follicular-centre cell lymphomas. Int J Cancer. 1985;35: Rinne A, Alavaikko M, Jarvinen M, et al. Demonstration of immunoreactive acid cysteine-proteinase inhibitor in reticulum cells of lymph node germinal centres. Virchows Arch B Cell Pathol. 1983;43: Said JW, Pinkus JL, Shintaku IP, et al. Alterations in fascin expressing germinal center dendritic cells in neoplastic follicles of B-cell lymphomas. Mod Pathol. 1998;11: q 2005 Lippincott Williams & Wilkins

7 Appl Immunohistochem Mol Morphol Volume 13, Number 4, December 2005 FDC Markers in AITL 21. Jaffe R, DeVaughn D, Langhoff E. Fascin and the differential diagnosis of childhood histiocytic lesions. Pediatr Dev Pathol. 1998;1: Pinkus GS, Pinkus JL, Langhoff E, et al. Fascin, a sensitive new marker for Reed-Sternberg cells of Hodgkin s disease. Am J Pathol. 1997;150: Bindl J, Warnke RA. Advantages of detecting monoclonal antibody binding to tissue sections with biotin and avidin reagents in Coplin jars. Am J Clin Pathol. 1986;85: Soderstrom KO, Rinne R, Hopsu-Havu VK, et al. Identification of acid cysteine proteinase inhibitor (cystatin A) in the human thymus. Anat Rec. 1994;240: Said JW, Pinkus JL, Yamashita J, et al. The role of follicular and interdigitating dendritic cells in HIV-related lymphoid hyperplasia: localization of fascin. Mod Pathol. 1997;10: Mosialos G, Yamashiro S, Baughman RW, et al. Epstein-Barr virus infection induces expression in B lymphocytes of a novel gene encoding an evolutionarily conserved 55-kilodalton actin-bundling protein. J Virol. 1994;68: Weiss LM, Jaffe ES, Liu XF, et al. Detection and localization of Epstein- Barr viral genomes in angioimmunoblastic lymphadenopathy and angioimmunoblastic lymphadenopathy-like lymphoma. Blood. 1992;7: Anagnostopoulos I, Hummel M, Finn T, et al. Heterogeneous Epstein- Barr virus infection patterns in peripheral T-cell lymphoma of angioimmunoblastic lymphadenopathy type. Blood. 1992;80: Ohshima K, Takeo H, Kikuchi M, et al. Heterogeneity of Epstein-Barr virus infection in angioimmunoblastic lymphadenopathy type T-cell lymphoma. Histopathology. 1994;25: Maeda K, Matsuda M, Suzuki H, et al. Immunohistochemical recognition of human follicular dendritic cells (FDCs) in routinely processed paraffin sections. J Histochem Cytochem. 2002;50: q 2005 Lippincott Williams & Wilkins 303