www.elsevier.com/locate/humpath

Human Pathology (2012) 43, 1789–1798

Progress in pathology

Follicular T-cell lymphoma: a member of an emerging familyof follicular helper T-cell derived T-cell lymphomasShimin Hu MD, PhD, Ken H. Young MD, PhD, Sergej N. Konoplev MD, PhD,L. Jeffrey Medeiros MD⁎

Department of Hematopathology, The University of Texas MD Anderson Cancer Center, Box 72, Houston, TX 77030, USA

Received 13 March 2012; revised 30 April 2012; accepted 4 May 2012

0h

Keywords:Follicular T-celllymphoma;Follicular helper T-cells

Summary Unlike B-cell lymphomas, where knowledge of normal B-cell origin and differentiation hasgreatly contributed to their classification, the current classification of peripheral T-cell lymphomas islimited by a lack of understanding of their cellular origin. In the current World Health Organizationclassification of lymphomas, follicular T-cell lymphoma was formally recognized as a morphologicvariant of peripheral T-cell lymphoma, not otherwise specified. There is growing evidence, however, thatfollicular T-cell lymphomamay be a unique clinicopathologic entity based on its morphologic features andderivation from follicular helper T-cells. In addition, there are abundant recent data supporting the conceptthat follicular helper T-cells can give rise to other types of T-cell lymphoma, includingangioimmunoblastic T-cell lymphoma, primary cutaneous CD4+ small/medium T-cell lymphoma, anda subset of neoplasms, in addition to follicular T-cell lymphoma, currently classified as peripheral T-celllymphoma, not otherwise specified. In this review, we focus primarily on the clinicopathologic,immunophenotypic, and molecular features of follicular T-cell lymphoma and discuss its potentialrelationship with other types of T-cell lymphoma thought to be derived from follicular helper T-cells.© 2012 Elsevier Inc. All rights reserved.

1. Introduction B-cell differentiation stages and a putative cell of origin for

The current classification of B-cell lymphomas has beengreatly aided by our extensive knowledge of B-cell originand normal differentiation. In lymph nodes, normal B-cellsform unique structures, such as germinal centers, mantlezones, and marginal zones, which can be recognizedmorphologically. A subset of B-cell lymphomas can exhibitmorphologic patterns that correspond to these compartments.A follicular pattern, for example, is a common feature offollicular lymphomas. Assessment of the immunophenotypeof normal B-cells in various compartments and correlationwith B-cell lymphomas has allowed construction of normal

⁎ Corresponding author.E-mail address: ljmedeiros@mdanderson.org (L. J. Medeiros).

046-8177/$ – see front matter © 2012 Elsevier Inc. All rights reserved.ttp://dx.doi.org/10.1016/j.humpath.2012.05.002

most B-cell lymphomas. Recognition of consistent chromo-somal translocations in B-cell lymphomas also has beenhelpful in understanding cell of origin and some chromo-somal translocations are defining features of some types ofB-cell lymphoma.

In contrast, the current classification of T-cell neoplasmsis largely based on a combination of clinical and pathologicfeatures. Cases of T-cell lymphoma are grouped according totheir anatomic sites of disease, in particular, leukemic/disseminated, extranodal, cutaneous, and nodal [1]. T-celllymphomas are further classified according to their morpho-logic features and the few cytogenetic or molecularabnormalities known to occur in these tumors. This approachhas been taken, in part, because we know relatively littleabout normal T-cell differentiation. In normal lymph nodes,

1790 S. Hu et al.

there are few T-cell compartments that can be recognizedmorphologically. Most T-cell lymphomas exhibit a diffusepattern without clues to potential cellular origin or stage ofdifferentiation. Based on what is known about T-cell subsets,others have attempted to classify T-cell lymphomas on thebasis on CD4 versus CD8 expression, TH1 versus TH2function, or presence or absence of a cytotoxic immunophe-notype [1-5]. However, these attempts have failed todelineate clinically meaningful categories.

Approximately 30% of all T-cell lymphomas worldwide,in part by default based on the issues already reviewed, fallinto the category of peripheral T-cell lymphoma, nototherwise specified (PTCL-NOS) [1,6]. The tumors in thiscategory display a broad spectrum of cytologic features, fromhighly polymorphous to monomorphous, and with neoplasticcells that can be predominantly small, intermediate-sized, orlarge and pleomorphic [1]. These tumors can express T-cellmarkers corresponding to a variety of putative T-cell subsets[5]. Conventional cytogenetic analysis has shown a widevariety of cytogenetic abnormalities, often with complexkaryotypes that correlate with inferior outcome [7-9], butvery few of these abnormalities are consistent findings thatcan be used for T-cell lymphoma classification. Comparativegenomic hybridization analysis and other high-throughputtechnologies have shown that virtually all PTCL-NOS casesharbor genetic imbalances, with gains commonly outnum-bering losses [10-12]. There are no obvious genetic sub-groups of PTCL-NOS that can be appreciated, however,further supporting the heterogeneity of this group of neo-plasms. A major challenge for pathologists and researchers,therefore, is to tease out clinically and biologically mea-ningful subgroups from the category of PTCL-NOS, with thegoal perhaps being the eventual elimination of PTCL-NOSall together.

Peripheral T-cell lymphoma with a nodular/follicularpattern was first reported in 1988 [13]. The unusual mor-phologic pattern of this type of T-cell lymphoma was thefirst clue that these neoplasms may be distinctive. It wasnot until the fourth edition of the World Health Organiza-tion (WHO) classification was published in 2008, however,that this type of T-cell lymphoma was formally recognized,and considered as a morphologic variant of PTCL-NOS [1].This was an important step forward, partly attributable to aflurry of case reports and small case series regarding thistype of T-cell lymphoma during the past two decades [13-32]. In the literature, a number of designations have beenused for these tumors, including nodular T-cell lymphoma;follicular T-cell lymphoma (FTCL); peripheral T-cell lym-phoma with a nodular, follicular, or perifollicular pattern;peripheral T-cell lymphoma with follicular involvement;and follicular variant of PTCL-NOS. Henceforth, we willuse the term follicular T-cell lymphoma here to describethese tumors.

Another important step forward is the recent progress inour understanding of the ontogeny of a distinct subset ofCD4+ helper T-cells, known as follicular helper T-cells

(TFH), and their relationship with T-cell lymphomas,including FTCL [23-30,33-44]. Follicular helper T-cellsnormally have the capacity to home to germinal centers(GCs). This migration is facilitated by the interaction of oneof their surface receptors, CXCR5, with the CXCL13chemokine expressed by follicular dendritic cells in thelight zone of GCs, where TFH cells support affinity matu-ration and isotype switching, leading to the generation ofmemory B-cells and long-lived plasma cells [35-39]. TFH

cells express an array of molecules serving these functions,including cytokines and cytokine receptors (eg, interleukin-21, CXCL13, CXCR5), immune-modulating receptors/adhesion molecules (eg, programmed death-1 or PD-1,inducible co-stimulator [ICOS], CD57, CD154, CD200,signaling lymphocyte-activation molecule family), cytoplas-mic signaling molecules (e.g., signaling lymphocyte-acti-vation molecule–associating protein [SAP]), andtranscriptional factors (eg, NF-ATc1 and c-Maf) [33-40].Interestingly, the Bcl-6-MUM1/IRF4-BLIMP1 transcrip-tional axis, essential for GC B-cell differentiation, is alsocrucial for TFH cell identity [35,39,40]. Deregulation ofTFH-cell functions is implicated in a variety of autoimmunedisorders in both mouse models and humans [37].

Many of the molecules that support TFH cell functionshave been used to identify these cells in lymph nodes andother sites using a variety of methods, and identification ofTFH cells has had practical implications in the diagnosis oflymphomas [33,34,41-62]. These studies have shown thatthe TFH immunophenotype, although characteristic of FTCL,is not specific for this entity. Gene expression profilingstudies of cases of angioimmunoblastic T-cell lymphoma(AITL) have shown over-expression of many genes cha-racteristic of normal TFH cells, and the results of these studieshave been validated at the protein level, suggesting thatAITL arises from TFH cells [42-44]. These observations alsoprovide a plausible explanation for the autoimmunephenomena and broad range of histopathologic featuresassociated with AITL. In addition, a subset of nodal PTCL-NOS cases other than FTCL have a TFH-cell immunophe-notype, despite their apparently different morphologicappearances [5,23,29,30,63]. Adding to the complexity isthe recent finding that cases of primary cutaneous CD4-positive small/medium T-cell lymphoma also express manyTFH cell markers [64,65]. These studies support the conceptof an emerging family of TFH -cell derived PTCLs.

In this review, we focus on FTCL, a rare neoplasm cur-rently considered in the WHO classification to be amorphologic variant of PTCL-NOS. The data are compiledfrom a total 80 cases from 18 case reports or case seriespublished in the English literature from 1988 to through2011, as well as one previously unpublished case we haveencountered (Figs. 1 and 2) [13-30]. Cases that have beenreported more than once in different publications arecombined correspondingly [20,24,25]. An additional 20cases reported in the non-English literature without ade-quate clinical data are not included in this review [31,32].

Fig. 1 A case of FTCL showing variably sized nodules composed of medium-sized lymphoid cells with moderately abundant paleeosinophilic cytoplasm (original magnification ×20 [A], ×1000 [B]) and with proliferation of high endothelial venules in the internodular areas(C; original magnification ×500). Nodules are highlighted by CD21 immunostain (D; original magnification ×20).

1791Follicular T-cell lymphoma

We also discuss the evidence supporting the origin ofFTCL from TFH cells and the emerging family of TFH cell-derived lymphomas.

2. Clinical features and laboratory findings

Follicular T-cell lymphoma is a rare disease that repre-sents 1%-2% of all PTCL-NOS cases [6]. This frequencymay underestimate the overall incidence as this entity ischallenging to recognize and cases can be missed[14,15,18,21]. These tumors mainly affect elderly patients,in the seventh decade, with a median and mean age of 62years (Table 1) [13-29]. There is a wide age range, however,from 27 to 90 years. The male to female ratio is 1.2 to 1.Systemic (B-type) symptoms are observed in less than one-third of patients. Most patients (48/56, 86%) present withsystemic lymphadenopathy that most commonly involvesthe cervical, axillary, or inguinal regions. Extranodal in-volvement has been reported. Hepatosplenomegaly and bonemarrow involvement each occur in approximately one-fourthof patients. The tonsils, salivary glands and hard palate canbe rarely involved [18]. A subset of patients (12/51, 23%)has skin lesions, either at time of diagnosis or at relapse,including erythema or papules, and occasionally patientscan have isolated skin lesions [21,22,25,27]. Most patients(40/55, 73%) have clinical stage III or IV disease.

Laboratory tests reveal hematological, immunological,and/or biochemical abnormalities in a subset of patients withFTCL (Table 1). The most common findings include apositive Coombs’ test with or without autoimmune hemo-lytic anemia or thrombocytopenia (7/14, 50%), elevatedserum lactate dehydrogenase level (13/29, 45%), andhypergammaglobulinemia (6/32, 19%). Two patients havebeen tested for serum soluble interleukin-2 receptor (sIL-2R)and human T-cell leukemia virus (HTLV); both patients hadelevated sIL-2R levels and were negative for HTLV-1[17,28]. Other rare laboratory findings reported in FTCLpatients include hypereosinophilia and an elevated serumIgE level, reported in 2 and 1 patients, respectively [13,25].

3. Histopathology

By definition, FTCL demonstrates a nodular growth patternin lymph nodes (Fig. 1A). Depending on the morphologicpattern, lymph nodes can display minimal, partial, or totalarchitectural effacement [13-29]. The peripheral sinuses areoften spared [13,14,18,19]. Capsular fibrosis can be present[15,17]. Extranodal infiltration by tumor cells, frequently seenin AITL, is uncommon in FTCL [19,22]. In the WHOclassification, 3 morphologic patterns in FTCL are recog-nized: follicular lymphoma (FL)–like, progressive trans-formation of germinal centers (PTGC)-like (also described

Fig. 2 Immunophenotype of a case of FTCL. The neoplastic lymphoid cells positive for CD3 (A; original magnification ×400), CD4 (B;original magnification ×400), CD20 dim (original magnification ×40 [D], ×400 [E]), CD10 (F; original magnification ×400), Bcl-6 (G;original magnification ×500), and PD-1 (H; original magnification ×500) and negative for CD8 (C; original magnification ×400). Rare cellspositive for CXCL13 (I; original magnification ×500).

Table 1 Clinical features of patients with FTCL

General featuresSex ratio (male:female) 1.2:1 (36:29)Age Mean: 62

Range: 27-90B-type constitutional symptoms 29% (16/55)Advanced stages (III/IV) 73% (40/55)Organ involvementSkin lesions 23% (12/51)Multiple lymphadenopathies 86% (48/56)Hepatomegaly 23% (6/26)Splenomegaly 24% (13/54)Bone marrow involvement 26% (6/23)Laboratory testsElevated LDH 45% (13/29)Hypergammaglobulinemia 19% (6/32)AIHA/positive Coombs' test 50% (7/14)

Abbreviations: LDH, lactate dehydrogenase; AIHA, autoimmunehemolytic anemia.

1792 S. Hu et al.

in the literature as nodular lymphocyte predominant Hodgkinlymphoma-like), and marginal zone lymphoma (MZL)–like.

In cases with an FL-like pattern, representing approxi-mately 20% of all FTCL cases, the tumor cells form intra-follicular aggregates or nodules that can closely resemblefollicular lymphoma. The nodules can be surrounded bymantle zones, or mantle zones can be absent [19,20,23,26,30].The nodules can be sharply demarcated [16,21,24,25,28]. InPTGC-like cases, representing approximately 50% of allFTCL cases, the tumor cells form large irregular or vaguenodules that mimic progressively transformed germinalcenters or nodules of nodular lymphocyte predominanceHodgkin lymphoma [13,16,17,21,22,24,25,27,28]. In casesthat are MZL-like, representing approximately 30% of allFTCL cases, the neoplastic lymphoid cells show a perifolli-cular growth pattern or paracortical involvement that canmimic marginal zone lymphoma [14,15,18,20]. A mixture ofFL-like and PTGC-like patterns are frequently seen in thesame biopsy specimen.

Table 2 Immunophenotype of FTCL

CD2a 100% (56/56) CD10 65% (37/57)CD3b 99% (74/75) Bcl-6 78% (36/46)CD4+CD8- 91% (67/74) PD-1 100% (38/38)CD4+CD8+ 1%(1/73) CXCL13 88% (35/40)CD4-CD8+ 1%(1/73) ICOS 100% (13/13)CD4-CD8- 7% (5/73) SAP 100% (3/3)CD5 89% (59/66) CD57 15% (7/47)CD7 32% (17/53) NF-ATc1 100% (1/1)CD43 94% (15/16) CD25 27% (3/11)CD45 97% (32/33) CD30 12% (4/34)CD56 0% (0/22) CD31 100% (2/2)TIA1/GraB 0% (0/31) Bcl-2 33% (3/9)

Abbreviations: GraB, granzyme B; PD-1, programmed cell death-1ICOS, inducible costimulator; SAP, SLAM (signaling lymphocyte-activation molecule)-associated protein.

a Including 1 case with partial loss of CD2 as positive [19].b Including 2 cases with partial loss of CD3 and 4 cases with positive

immunohistochemistry but negative flow cytometry as positive[14,18,19].

1793Follicular T-cell lymphoma

Although these patterns are grouped together in the WHOclassification, these patterns are sufficiently different tosuggest at least some differences in pathogenesis. In par-ticular, cases of FTCL with a MZL-like pattern have somesimilarities with early phase AITL, so-called patterns I and II,as summarized by Attygalle et al [61]. The overallarchitecture is often minimally or partially effaced. Thelymphoid follicles can be hyperplastic but, more often, areregressed or atretic. In advanced lesions with a perifolliculargrowth pattern, the neoplastic cells may extend into mantlezones and T-zone regions. Internodular or intranodularepithelioid histiocytes, rarely prominent, are often encoun-tered [13,15,19,24]. Hodgkin-like cells and eosinophils aresometimes present. Proliferation of high-endothelial venules(HEV), an otherwise uncommon event in FTCL, is often seenin cases with MZL-like morphology [15]. However, ingeneral cases of FTCL differ from AITL in that FTCL casesless often have a polymorphous reactive background, aproliferation of arborizing HEV, or expansion of folliculardentritic cell networks (Fig. 1C) [13,15-17,22,24].

Agostinelli et al reported 4 cases of FTCL involvinglymph node in which the tumor nodules were not associatedwith follicular dendritic cell meshworks [29]. Although wehave not reviewed these cases, their description seemssimilar to the MZL-like pattern of FTCL. All 4 casesexpressed 3 to 6 TFH cell markers. However, most othercases of FTCL with MZL-like growth pattern have not beenassessed for TFH cell markers because these studies werepublished before TFH cells were well-characterized[14,15,18,20]. Therefore, the true frequency of the MZL-like pattern may need to be re-evaluated.

Cytologically, the neoplastic cells of FTCL involvinglymph nodes have been described as medium or mediumto large in size (58/80, 72.5%), small to medium-sized (16/80, 20%), or large (6/80, 7.5%) (Fig. 1B) [13-29]. Thetumor cells tend to have round to slightly irregular orindented nuclear contours, vesicular to coarsely granularchromatin, and usually moderate or abundant clear or paleeosinophilic cytoplasm with indistinct cell borders. Thenuclear features of FTCL cells can resemble centrocytes orcentroblasts; the cytoplasm is helpful in their recognitionas T-cells.

Follicular T-cell lymphoma also has been reported toinvolve extranodal sites in a subset of patients, includingbone marrow, liver, spleen, skin, tonsils, salivary glands,and hard palate, although as far as we are aware, descriptionsof the histologic findings at every one of these sites are notavailable. The diagnosis of FTCL at an extranodal site canbe highly challenging unless the diagnosis already has beenestablished in lymph nodes. The pattern of bone marrowinvolvement by FTCL has been reported to be paratrabecularand/or interstitial [21,26]. At extranodal mucosal sites,FTCL can be associated with lymphoepithelial lesionssimilar to those of extranodal marginal zone B-celllymphoma of mucosa-associated lymphoid tissue (MALTlymphoma) [18].

4. Immunophenotype

Follicular T-cell lymphoma is composed of lymphoidcells that express a variety of TFH cell markers (Fig. 2 andTable 2) [22-29]. PD-1/CD279 and ICOS are reported to bemost sensitive, positive in all cases tested to date, 38 and 13cases, respectively. PD-1 is a member of CD28/CTLA-4co-receptor family that delivers inhibitory signals to T-cells,and ICOS is another member of the same coreceptor familythat provides co-stimulation to T-cells. Most cases (35/40,88%) of FTCL are positive for CXCL13. Cytoplasmic SAPand transcription factor NF-ATc1 have been positive in afew cases tested [23,29]. By contrast, CD10 and Bcl-6, 2markers commonly expressed by both benign and malignantgerminal center B-cells, are less sensitive markers of FTCL.CD10 and Bcl-6 have been expressed in 37/57 (65%) and 36/46 (78%) of FTCL cases, respectively. Down-regulation ofBcl-6 and CD10, which is frequently observed whenneoplastic germinal center B-cells migrate out of the folliclesin follicular lymphoma, has been observed in small numbersof FTCL cases reported [16,19,21].

It is important to remember that individual TFH markerscan be expressed by other T-cell subsets, other types ofT-cell lymphoma (eg, mycosis fungoides) and B-cell lym-phoma (eg, chronic lymphocytic leukemia/small lympho-cytic lymphoma) as well as in reactive conditions as reportedby others [29,30,46,50,51,53,55,57,63,65-69]. Therefore, ithas been suggested that a combination of markers be used,with an arbitrary minimum of three markers being positiveto define the TFH immunophenotype [29,35,54]. Thirty-onecases of FTCL reported in the literature have been testedfor expression of 4 to 6 TFH markers, with 29 (94%)positive for at least 3 markers supporting TFH cell lineage.

The neoplastic cells of FTCL also express pan-T-cellantigens, including CD2, CD3, CD5, or CD7 (Fig. 2 and

Table 3 Karyotypes of 7 reported cases of FTCL

1. 46,XY,del(6)(q21)[3]/46,idem,add(17)(q25)[7]/46,XY[10]2. 46,XX,t(5;9)(q33;q22)[2]/46,XX,t(5;9)(q33;q22),dup(17)(q21q24)[12]/46,XX[18]

3. 46,XX,del(6)(q15q25)[8]/46,XX,der(11)t(3;11)(q13;q25),der(13)t(X;13)(q22;p11)[6]/46,XX,der(9)t(1;9)(q21;p24),der(15;17)(q10;q10),+17[4]

4. 33~41,der(X)t(X;10)(q28;q24),Y,ins(1;22)(q31;q11q13),der(2)t(1;2)(q32;q37),der(3)t(3;17)(p21;q11),−4,der(5)t(4;5)(q11;p15),der(7)(qterp22::13q?::6p216pter),der(7)t(4;7)(q11:p22),−9,−10,t(13;15)(q22;q22),der(14)t(9;14)(q32;p11),der(14)(qter-p11::10q22–10q26::2q23-2qter)[cp7]

5. 45,XY,der(7)t(1;7)(q21;q32),add(12)(q24),+2mar,inc[9]6. 47,XY,+add(X)(p22),+Y,add(1)(p34),−2,add(3)(q29),del(3)(q12q21),add(5)(q35),−6,del(6)(q15),+7,del(8)(q22),del(12)(q22),+add(12)(q24),−13,der(1;16)(q10;p10),+mar[10]/46,XY[5]

7. 45,XY,add(1)(q21),add (2)(p11.2),add (7)(p22),add(8)(p11.2),−9,add(10)(q22),−14,add (15)(q22),+mar1[18]

NOTE. References [15,20,21,28].

1794 S. Hu et al.

Table 2) [13-29]. An aberrant T-cell immunophenotype orantigen “loss” is present in many cases. CD7 has beenabsent in FTCL most often (36/53, 68%), and CD5, insome cases (7/66, 11%). Partial loss of CD2, CD3, or CD4has been reported in occasional cases. Most cases (67/74,91%) of FTCL have a CD4+CD8− helper T-cell immuno-phenotype. A small subset (7%) of cases reported wereCD4−CD8−. Aberrant expression of B-cell markers, suchas CD79a or CD20, is rarely encountered (Fig. 2D and E)[22]. T-cell receptor expression has been infrequentlyassessed in FTCLs.

Other markers seen in a small subset of cases include Bcl-2 (3/9, 33%), CD25 (3/11, 27%), CD57 (7/47, 15%), and dimCD30 (4/34, 12%). CD31 expression has been observed in 2cases tested [19,22]. They are negative for CD56 andcytotoxic antigens, such as TIA-1 and granzyme B. ScatteredEpstein-Barr virus encoded RNA–positive B cells consistentwith immunoblasts have been reported in 18 (41%) of 43cases. Reactive follicular dendritic cell meshworks revealedby CD21, CD23, and CD35 immunostains can be expandedbut within the nodular areas. The proliferation index, asassessed by Ki-67 immunoreactivity, is highly variable,ranging from 5% to 70%, with median and mean values ofapproximately 45%.

5. Molecular and genetic features

Monoclonal rearrangement of the TCR (T-cell receptor)genes has been detected in 47 of 55 cases (85%) of FTCLassessed. An oligoclonal pattern was reported in 3 cases(5.5%), and 5 cases (9.5%) were polyclonal [13-29]. Mostcases reported have been assessed by polymerase chainreaction methods, but 10 cases assessed by Southern blotanalysis all showed monoclonal TCRβ chain gene rearrange-ments [13,14,18]. Monoclonal immunoglobulin (IgH) generearrangements are frequently observed in AITL [70,71].However, no evidence of monoclonal IgH gene rearrange-ments has been detected in all 35 cases of FTCL tested; onecase was reported as oligoclonal [25].

Using conventional cytogenetic analysis, 7 (39%) of 18cases of FTCL reported had an abnormal and usually complexkaryotype (Table 3) [15,20,21,28]. One case reported carrieda t(5;9)(q33;q22) involving the fusion of 2 tyrosine kinases,ITK (IL-2-inducible T-cell kinase) and SYK (spleen tyrosinekinase), confirmed by fluorescence in situ hybridization [20].Fluorescence in situ hybridization analysis for ITK-SYK hasshown this fusion in 6 (21%) of 28 cases assessed, including 5cases with a FL-like pattern and 1 case with a MZL-likepattern [20,25]. Interestingly, t(5;9)(q33;q22) also has beenreported in 2 cases of PTCL-NOS with a diffuse growthpattern, one of which was examined for and had TFH cellimmunophenotype [5,20]. The translocation was not found inAITL or anaplastic lymphoma kinase (ALK)-negativeanaplastic large cell lymphoma [20]. Thus, t(5;9) may beunique to FTCL and rare cases of PTCL-NOS with a TFH cell

immunophenotype[5,20]. Expression of ITK-SYK in miceinduces lymphoma resembling human PTCL [72,73]. Over-expression of SYK, however, has been shown in a widevariety of T-cell lymphomas that lack t(5;9)/ITK-SYK [74].

The roles of BCL-2 and BC-L6, well known in B-celllymphomas, are less clear in FTCLs. The Bcl-6 protein isknown to be important for TFH cell differentiation [35-40]. Inone case of FTCL that we observed recently, we showedamplification of both BCL-6 and BCL-2 by fluorescence insitu hybridization analysis (unpublished data). Aberrantsomatic mutation of BCL-6 has not been identified inFTCLs [21]. To date, there are no reports of array-basedglobal examination of genetic abnormalities or geneexpression profiling in FTCL.

6. Therapy and prognosis

Comprehensive studies on the prognosis and therapeuticoptions for patients with FTCL are not available in theliterature, which is not surprising given the rarity of thesetumors. Here, we have compiled data from case reports andsmall case series of FTCL patients that have been published[13-28]. Thirty-seven of 41 patients (90%) were treated withchemotherapy; 20 of 22 patients with available informationregarding the treatment regimen received cyclophospha-mide, doxorubicin, vincristine, and prednisone (CHOP)chemotherapy, with or without other additional chemother-apeutic agents or therapeutic modalities. Six of thesepatients underwent autologous stem cell transplant[14,17,19-21,23,26,28]. After therapy, many patients hada complete or partial remission, but relapses were frequent.Clinical follow-up was available for 49 patients; 21 (43%)died after a median follow-up of 18 to 24 months (range, 1week to 124 months) [14-16,18,20-23,25]. There are 28

1795Follicular T-cell lymphoma

living patients at last follow-up, but only 16 were free ofdisease after median follow-up of 12 to 35 months (range,9-124 months) [14-17,20,21,25,27,28].

7. Other TFH cell–derived T-cell lymphomas

Other types of T-cell lymphoma that have a TFH cellimmunophenotype are known, the best known of which isAITL, the first type of T-cell lymphoma shown to have a TFH

cell immunophenotype. The morphologic findings of AITLhave been divided into 3 patterns, arbitrarily designated aspatterns I to III, which correlate with disease progression andloosely correlate with clinical progression and prognosis[61,75-77]. Cases of FTCL with an MZL-like pattern sharesome histological and clinicopathologic features with patternsI and II of AITL. It seems possible that FTCLwith aMZL-likepattern and AITL are closely related or represent a spectrum ofthe same disease. Supporting this concept, rare cases of FTCLwith coexistent AITL also have been reported [30]. Inaddition, Huang and colleagues reported that 4 patients wereinitially diagnosed with AITL but relapsed with biopsyspecimens that resembled FTCL [25]. However, relapse ofAITL as FTCL is rare as Attygalle et al followed up 30 patientswith AITL and relapse as FTCL was not reported [75].

A subset of cases currently classified as PTCL-NOSalso havebeen shown to express TFH cell markers suggesting a possiblerelationship with FTCL [5,29,30,42-44,46,50,53,55,63].These cases have been identified by gene expressionprofiling or immunohistochemical analysis using TFH-associated markers. In gene expression profiling studies,approximately 20% to 25% of PTCL-NOS cases have a TFH

gene signature [42-44]. Using immunohistochemistry,Rodriguez-Pinilla and colleagues [63] assessed for expres-sion of CD10, Bcl-6, PD-1, and CXCL13 in 81 cases ofPTCL-NOS; 16 cases (20%) expressed at least three TFH

markers supporting a TFH cell immunophenotype. Zhan et al[30] also studied the same set of TFH cell markers in 81 casesof PTCL-NOS and found 4 cases (5%) that expressed 3 ormore TFH markers. The explanation for the discrepancybetween these 2 studies is unknown, but the results ofRodriguez-Pinilla and colleagues are more in line with theresults of gene expression profiling studies [42-44].

It also appears that there are morphologic correlatesassociated with expression of TFH markers in PTCL-NOS.A subset of cases have some histologic features suggestive,although not diagnostic, of AITL [30,42,63]. It thereforeseems likely that some cases of PTCL-NOS may be earlycases of AITL, at a stage before the characteristic clinical orhistologic features of AITL are present. In addition, casesof the lymphoepithelioid variant of PTCL (so-calledLennert lymphoma), T-zone variant PTCL, and rare casesof ALK-anaplastic large cell lymphoma, and truly unclas-sifiable cases of PTCL-NOS can express multiple TFH

markers [29,30,63]. Importantly, there are definitionalissues involved in classification. Is a TFH cell immunophe-

notype adequate evidence to remove a tumor from thePTCL-NOS category and suggest another diagnosis, suchas AITL?

Primary cutaneous CD4-positive small/medium T-celllymphoma (PCSMTCL) is another type of T-cell lymphomarecently shown to have a TFH cell immunophenotype [64,65].PCSMTCL is a rare and indolent disease, currently listed as aprovisional entity in the 2008 WHO classification [1,78].These neoplasms have a predilection for the head/neck andupper trunk regions. Morphologically, PCSMTCL shows adiffuse or nodular dermal infiltrate that has a tendency toinfiltrate the subcutis and can exhibit a minor component ofepidermotropism. A reactive background is common.Aberrant T-cell immunophenotypes are uncommon andEBV infection is rare [64,78]. The TFH cell markers Bcl-6,PD-1 and CXCL13, but not CD10, are consistently expressedwhich help distinguish PCSMTCL from other types ofcutaneous T-cell lymphoma and reactive conditions [64,65].Le Tourneau et al reported 2 cases of FTCL occurring inhead/neck and upper trunk [27]. These tumors exhibited afollicular dendritic cell (FDC) meshwork and were positivefor CD10. It is not known whether these cases representcutaneous FTCL or PCSMTCL with follicular pattern.Alternatively, PCSMTCL may represent a diffuse variant ofFTCL. However, the indolent clinical behavior ofPCSMTCL distinguishes this neoplasm from FTCL whichis clinically more aggressive.

8. Conclusions

There appears to be an emerging consensus that FTCL is arare but distinctive type of T-cell lymphoma that arises fromTFH cells, a normal T-cell subset located in the germinal centersof lymphoid follicles. A variety of morphologic patterns havebeen accepted in the WHO classification, but in our opinion,this heterogeneity raises the possibility that the criteria fordiagnosis of FTCL need further refinement. A subset of FTCLis associated with a distinctive molecular abnormality,t(5;9)(q33;q22), involving SYK and TYK, further suggestingheterogeneity that will need to be addressed. Follicular T-celllymphomas are clinically aggressive and often refractory tostandard chemotherapy regimens such as CHOP.

There is also emerging evidence that there is a family of T-cell lymphomas that have a TFH cell immunophenotype,including FTCL, AITL, the lymphoepithelioid (Lennertlymphoma) and T-zone variants of PTCL-NOS, andPCSMTCL. There are also case reports and small series ofpatients who have had FTCL and AITL or PTCL-NOS with aTFH immunophenotype, simultaneously or in sequentialfashion. Whether these are examples of transformationfrom FTCL into another tumor type, or represent plasticityof TFH cell differentiation is unknown. It is essential toremember that many of these widely used TFH cell–associated markers are not entirely specific for TFH cells.As a result, the use of a single TFH cell marker for defining

1796 S. Hu et al.

TFH cell immunophenotype is inadequate. There appears tobe an emerging consensus that at least 3 TFH cell–associatedmarkers being positive are needed to support a TFH cellimmunophenotype[29,35,54].

Many questions remain to be answered. The paracorticalor perifollicular location of the neoplastic cells in cases ofFTCL with MZL-like pattern is difficult to reconcile with theGC location of TFH cells, the postulated cell of origin. Asmany cases of MZL-like FTCL were not assessed for TFH

cells at the time they were published, perhaps consideringthese cases as FTCL needs to be evaluated. Alternatively, arecent study has identified a subset of non–germinal centerT-cells with TFH cell immunophenotype [79]. These non-GCTFH cells induce the proliferation and differentiation ofnaïve, but not GC, B cells. There is also a quantitativedifference in the expression of some of the TFH cell markersbetween these 2 subsets of TH cells [35,79,80]. Anotherrecent development is the identification of TFH cellcounterparts in the peripheral blood [81,82]. It is unknownwhether these non-GC TFH cells are a completely differentsubset of TH cells, or they may represent a non-GC stage(precursor or memory stage) of GC TFH cells [35,36].

The identification of a family of T-cell lymphomas thatarise from TFH cells or show TFH cell differentiation mayhave practical implications. Others have suggested that TFH

cell–derived lymphomas have a broad spectrum and thisraises the possibility of defining these neoplasms as a groupin future lymphoma classification schemes. Although thereare similarities between FTCL and AITL, the relationshipbetween FTCL, AITL, PCSMTCL, and some cases of thelymphoepithelioid and T-zone variants of PTCL-NOS isunclear [25,29,30,62,63]. Should all of these lymphomas beclassified as subtypes under the same umbrella term of TFH

cell-derived lymphoma? In some ways, an analogy can bemade with B-cell lymphomas that have a germinal center B-cell immunophenotype, a group that includes clinicallyindolent low-grade follicular lymphomas and clinicallyaggressive diffuse large B-cell lymphoma as well as Burkittlymphoma. The umbrella term germinal center B-cellderived lymphoma has some academic value but certainlydoes not convey practical information in terms of guidingtherapeutic decisions or predicting prognosis. At least fornow, it may be premature to redefine T-cell lymphomaspurely on the basis of a TFH cell immunophenotype.

9. Note in added proof

After acceptance of this manuscript, Miyoshi et alreported 17 cases of FTCL [83]. In their study, the authorsemphasize that FTCL shares many clinical and pathologiccharacteristics with AITL, as well as a TFH cell immuno-phenotype. The authors further suggest that FTCL and AITLbelong to the same disease spectrum. The findings presentedin this study are in keeping with some of the data we havepresented in this review.

References

[1] Swerdlow SH, Campo E, Harris NL, et al, editors. WHO classificationof tumors of hematopoietic and lymphoid tissue. 4th ed. Lyon, France:IARC Press; 2008.

[2] de Leval L, Bisig B, Thielen C, Boniver J, Gaulard P. Molecularclassification of T-cell lymphoma. Crit Rev Oncol Hematol 2009;72:125-43.

[3] Went P, Agostinelli C, Gallamini A, et al. Marker expression inperipheral T-cell lymphoma: a proposed clinical-pathologic prognosticscore. J Clin Oncol 2006;24:2472-9.

[4] Rüdiger T, Geissinger E, Müller-Hermelink HK. ‘Normal counter-parts' of nodal peripheral T-cell lymphoma. Hematol Oncol 2006;24:175-80.

[5] Matsumoto Y, Horiike S, Ohshiro M, et al. Expression of masterregulators of helper T-cell differentiation in peripheral T-celllymphoma, not otherwise specified, by immunohistochemical analy-sis. Am J Clin Pathol 2010;133:281-90.

[6] Weisenburger DD, Savage KJ, Harris NL, et al. Peripheral T-celllymphoma, not otherwise specified: a report of 340 cases from theInternational Peripheral T-cell Lymphoma Project. Blood 2011;117:3402-8.

[7] Lepretre S, Buchonnet G, Stamatoullas A, et al. Chromosome abnor-malities in peripheral T-cell lymphoma. Cancer Genet Cytogenet 2000;117:71-9.

[8] Schlegelberger B, Himmler A, Godde E, Grote W, Feller AC, LennertK. Cytogenetic findings in peripheral T-cell lymphomas as a basis fordistinguishing low-grade and high-grade lymphomas. Blood 1994;83:505-11.

[9] Nelson M, Horsman DE, Weisenburger DD, et al. Cytogeneticabnormalities and clinical correlations in peripheral T-cell lymphoma.Br J Haematol 2008;141:461-9.

[10] Zettl A, Rüdiger T, Konrad MA, et al. Genomic profiling of peripheralT-cell lymphoma, unspecified, and anaplastic large T-cell lymphomadelineates novel recurrent chromosomal alterations. Am J Pathol2004;164:1837-48.

[11] Meléndez B, Díaz-Uriarte R, Cuadros M, et al. Gene expressionanalysis of chromosomal regions with gain or loss of genetic materialdetected by comparative genomic hybridization. Genes ChromosomesCancer 2004;41:353-65.

[12] Thorns C, Bastian B, Pinkel D, et al. Chromosomal aberrations inangioimmunoblastic T-cell lymphoma and peripheral T-cell lympho-ma unspecified: a matrix-based CGH approach. Genes ChromosomesCancer 2007;46:37-44.

[13] van den Oord JJ, de Wolf-Peeters C, O'Connor NT, de Vos R, TricotG, Desmet VJ. Nodular T-cell lymphoma. Report of a case studiedwith morphologic, immunohistochemical, and DNA hybridizationtechniques. Arch Pathol Lab Med 1988;112:133-8.

[14] Macon WR, Williams ME, Greer JP, Cousar JB. Paracortical nodularT-cell lymphoma. Identification of an unusual variant of peripheralT-cell lymphoma. Am J Surg Pathol 1995;19:297-303.

[15] Rudiger T, Ichinohasama R, Ott MM, et al. Peripheral T-cell lym-phoma with distinct perifollicular growth pattern: a distinct subtype ofT-cell lymphoma? Am J Surg Pathol 2000;24:117-22.

[16] de Leval L, Savilo E, Longtine J, Ferry JA, Harris NL. Peripheral T-cell lymphoma with follicular involvement and a CD4+/bcl-6+phenotype. Am J Surg Pathol 2001;25:395-400.

[17] Hatano B, Fukushima T, Honda M, et al. Peripheral T-cell lymphomawith a nodular growth pattern. Pathol Int 2002;52:400-5.

[18] Uherova P, Ross CW, Finn WG, Singleton TP, Kansal R, Schnitzer B.Peripheral T-cell lymphoma mimicking marginal zone B-cell lym-phoma. Mod Pathol 2002;15:420-5.

[19] Jiang L, Jones D, Medeiros LJ, Orduz YR, Bueso-Ramos CE.Peripheral T-cell lymphoma with a "follicular" pattern and theperifollicular sinus phenotype. Am J Clin Pathol 2005;123:448-55.

1797Follicular T-cell lymphoma

[20] Streubel B, Vinatzer U, Willheim M, Raderer M, Chott A. Novelt(5;9)(q33;q22) fuses ITK to SYK in unspecified peripheral T-celllymphoma. Leukemia 2006;20:313-8.

[21] Ikonomou IM, Tierens A, Troen G, et al. Peripheral T-cell lymphomawith involvement of the expanded mantle zone. Virchows Arch 2006;449:78-87.

[22] Starkey CR, Corn AI, Porensky RS, Viswanatha D, Wilson CS.Peripheral T-cell lymphoma with extensive dendritic cell networkmimicking follicular dendritic cell tumor: a case report withpathologic, immunophenotypic, and molecular findings. Am J ClinPathol 2006;126:230-4.

[23] Bacon CM, Paterson JC, Liu H, et al. Peripheral T-cell lymphoma witha follicular growth pattern: derivation from follicular helper T cells andrelationship to angioimmunoblastic T-cell lymphoma. Br J Haematol2008;143:439-41.

[24] Qubaja M, Audouin J, Moulin JC, et al. Nodal follicular helper T-celllymphoma may present with different patterns. A case report. HUM

PATHOL 2009;40:264-9.[25] Huang Y, Moreau A, Dupuis J, et al. Peripheral T-cell lymphomas with

a follicular growth pattern are derived from follicular helper T cells(TFH) and may show overlapping features with angioimmunoblasticT-cell lymphomas. Am J Surg Pathol 2009;33:682-90.

[26] Ortiz-Muchotrigo N, Rodriguez-Pinilla SM, Ramos R, et al. FollicularT-cell lymphoma: description of a case with characteristic findingssuggesting it is a different condition from AITL. Histopathology 2009;54:902-4.

[27] Le Tourneau A, Audouin J, Molina T, et al. Primary cutaneousfollicular variant of peripheral T-cell lymphoma NOS. A report of twocases. Histopathology 2010;56:548-51.

[28] Goto N, Tsurumi H, Sawada M, et al. Follicular variant of peripheralT-cell lymphoma mimicking follicular lymphoma: a case report with areview of the literature. Pathol Int 2011;61:326-30.

[29] Agostinelli C, Hartmann S, Klapper W, et al. Peripheral T celllymphomas with follicular T helper phenotype: a new basket or adistinct entity? Revising Karl Lennert's personal archive. Histopa-thology 2011;59:679-91.

[30] Zhan HQ, Li XQ, Zhu XZ, Lu HF, Zhou XY, Chen Y. Expression offollicular helper T cell markers in nodal peripheral T cell lym-phomas: a tissue microarray analysis of 162 cases. J Clin Pathol2011;64:319-24.

[31] Wang W, Ji FS, Chen HS, et al. Clinicopathologic features ofperipheral T-cell lymphoma, unspecified with follicular pattern.Zhonghua Bing Li Xue Za Zhi 2009;38:248-52.

[32] Zhan HQ, Zhu XZ, Li XQ, Zhou XY. Follicular variant of peripheralT-cell lymphoma: a clinicopathologic and genetic study of 2 cases.Zhonghua Bing Li Xue Za Zhi 2011;40:32-6.

[33] Kim CH, Lim HW, Kim JR, Rott L, Hillsamer P, Butcher EC. Uniquegene expression program of human germinal center T helper cells.Blood 2004;104:1952-60.

[34] Chtanova T, Tangye SG, Newton R, et al. T follicular helper cellsexpress a distinctive transcriptional profile, refecting their role as non–TH1/TH2 effector cells that provide helper for B cells. J Immunol2004;173:68-78.

[35] Yu D, Vinuesa CG. The elusive identity of T follicular helper cells.Trends Immunol 2010;31:377-83.

[36] Fazilleau N, Mark L, McHeyzer-Williams LJ, McHeyzer-WilliamsMG. Follicular helper T cells: lineage and location. Immunity 2009;30:324-35.

[37] Yu D, Batten M, Mackay CR, King C. Lineage specification andheterogeneity of T follicular helper cells. Curr Opin Immunol 2009;21:619-25.

[38] King C. New insights into the differentiation and function of Tfollicular helper cells. Nat Rev Immunol 2009;9:757-66.

[39] Nutt SL, Tarlinton DM. Germinal center B and follicular helper Tcells: siblings, cousins or just good friends? Nat Immunol 2011;12:472-7.

[40] Crotty S, Johnston RJ, Schoenberger SP. Effectors and memories: Bcl-6 and Blimp-1 in T and B lymphocyte differentiation. Nat Immunol2010;11:114-20.

[41] Gaulard P, de Leval L. Follicular helper T cells: implications inneoplastic hematopathology. Semin Diagn Pathol 2011;28:202-13.

[42] de Leval L, Rickman DS, Thielen C, et al. The gene expression profileof nodal peripheral T-cell lymphoma demonstrates a molecular linkbetween angioimmunoblastic T-cell lymphoma (AITL) and follicularhelper T (TFH) cells. Blood 2007;109:4952-63.

[43] Piccaluga PP, Agostinelli C, Califano A, et al. Gene expressionanalysis of angioimmunoblastic lymphoma indicates derivation from Tfollicular helper cells and vascular endothelial growth factorderegulation. Cancer Res 2007;67:10703-10.

[44] Iqbal J, Weisenburger DD, Greiner TC, et al. Molecular signatures toimprove diagnosis in peripheral T-cell lymphoma and prognosticationin angioimmunoblastic T-cell lymphoma. Blood 2010;115:1026-36.

[45] Grogg KL, Attygalle AD,MaconWR, Remstein ED, Kurtin PJ, DoganA. Angioimmunoblastic T-cell lymphoma: a neoplasm of germinal-center T-helper cells? Blood 2005;106:1501-12.

[46] Dupuis J, Boye K, Martin N, et al. Expression of CXCL13 byneoplastic cells in angioimmunoblastic T-cell lymphoma (AITL): anew diagnostic marker providing evidence that AITL derives fromfollicular helper T cells. Am J Surg Pathol 2006;30:490-4.

[47] Grogg KL, Attygalle AD, Macon WR, Remstein ED, Kurtin PJ,Dogan A. Expression of CXCL13, a chemokine highly upregulatedin germinal center T-helper cells, distinguishes angioimmunoblasticT-cell lymphoma from peripheral T-cell lymphoma, unspecified.Mod Pathol 2006;19:1101-7.

[48] Ortonne N, Dupuis J, Plonquet A, et al. Characterization of CXCL13+neoplastic T cells in cutaneous lesions of angioimmunoblastic T-celllymphoma (AITL). Am J Surg Pathol 2007;31:1068-76.

[49] Dorfman DM, Brown JA, Shahsafaei A, Freeman GJ. Programmeddeath-1 (PD-1) is a marker of germinal center-associated T cells andangioimmunoblastic T-cell lymphoma. Am J Surg Pathol 2006;30:802-10.

[50] Roncador G. García Verdes-Montenegro JF, Tedoldi S, et al. Expressionof two markers of germinal center T cells (SAP and PD-1) in angio-immunoblastic T-cell lymphoma. Haematologica 2007;92:1059-66.

[51] Xerri L, Chetaille B, Serriari N, et al. Programmed death 1 is a marker ofangioimmunoblastic T-cell lymphoma and B-cell small lymphocyticlymphoma/chronic lymphocytic leukemia. HUM PATHOL 2008;39:1050-8.

[52] Yu H, Shahsafaei A, Dorfman DM. Germinal-center T-helper-cellmarkers PD-1 and CXCL13 are both expressed by neoplastic cells inangioimmunoblastic T-cell lymphoma. Am J Clin Pathol 2009;131:33-41.

[53] Krishnan C, Warnke RA, Arber DA, Natkunam Y. PD-1 expression inT-cell lymphomas and reactive lymphoid entities: potential overlap instaining patterns between lymphoma and viral lymphadenitis. Am JSurg Pathol 2010;34:178-89.

[54] Laurent C, Fazilleau N, Brousset P. A novel subset of T-helper cells:follicular T-helper cells and their markers. Haematologica 2010;95:356-68.

[55] Marafioti T, Paterson JC, Ballabio E, et al. The inducible T-cell co-stimulator molecule is expressed on subsets of T cells and is a newmarker of lymphomas of T follicular helper cell-derivation. Haema-tologica 2010;95:432-9.

[56] Baseggio L, Traverse-Glehen A, Berger F, et al. CD10 and ICOSexpression by multiparametric flow cytometry in angioimmunoblasticT-cell lymphoma. Mod Pathol 2011;24:993-1003.

[57] Dorfman DM, Shahsafaei A. CD200 (OX-2 membrane glycoprotein)is expressed by follicular T helper cells and in angioimmunoblastic T-cell lymphoma. Am J Surg Pathol 2011;35:76-83.

[58] Krenacs L, Schaerli P, Kis G, Bagdi E. Phenotype of neoplastic cells inangioimmunoblastic T-cell lymphoma is consistent with activatedfollicular B helper T cells. Blood 2006;108:1110-1.

1798 S. Hu et al.

[59] Yuan CM, Vergilio JA, Zhao XF, Smith TK, Harris NL, Bagg A.CD10 and BCL6 expression in the diagnosis of angioimmunoblasticT-cell lymphoma: utility of detecting CD10+ T cells by flowcytometry. HUM PATHOL 2005;36:784-91.

[60] Murakami YI, Yatabe Y, Sakaguchi T, et al. c-Maf expression inangioimmunoblastic T-cell lymphoma. Am J Surg Pathol 2007;31:1695-702.

[61] Attygalle A, Al-Jehani R, Diss TC, et al. Neoplastic T cells inangioimmunoblastic T-cell lymphoma express CD10. Blood 2002;99:627-33.

[62] Warnke RA, Jones D, Hsi ED. Morphologic and immunophenotypicvariants of nodal T-cell lymphomas and T-cell lymphoma mimics. AmJ Clin Pathol 2007;127:511-27.

[63] Rodriguez-Pinilla SM, Atienza L, Murillo C, et al. Peripheral T-celllymphoma with follicular T-cell markers. Am J Surg Pathol 2008;32:1787-99.

[64] Rodriguez Pinilla SM, Roncador G, Rodriguez-Peralto JL, et al.Primary cutaneous CD4+ small/medium-sized pleomorphic T-celllymphoma expresses follicular T-cell markers. Am J Surg Pathol 2009;33:81-90.

[65] Cetinozman F, Jansen PM, Willemze R. Expression of programmeddeath-1 in primary cutaneous CD4-positive small/medium-sizedpleomorphic T-cell lymphoma, cutaneous pseudo–T-cell lymphoma,and other types of cutaneous T-cell lymphoma. Am J Surg Pathol2012;36:109-16.

[66] Kozako T, Yoshimitsu M, Fujiwara H, et al. PD-1/PD-L1 expressionin human T-cell leukemia virus type 1 carriers and adult T-cellleukemia/lymphoma patients. Leukemia 2009;23:375-82.

[67] Picchio MC, Scala E, Pomponi D, et al. CXCL13 is highly producedby Sézary cells and enhances their migratory ability via a synergisticmechanism involving CCL19 and CCL21 chemokines. Cancer Res2008;68:7137-46.

[68] Cook JR, Craig FE, Swerdlow SH. Benign CD10-positive T cells inreactive lymphoid proliferations and B-cell lymphomas. Mod Pathol2003;16:879-85.

[69] Chu PG, Chang KL, Weiss LM, Arber DA. Immunohistochemicaldetection of CD10 in paraffin sections of hematopoietic neoplasms: acomparison with flow cytometry detection in 56 cases. ApplImmunohistochem Mol Morphol 2000;8:257-62.

[70] Zaki MA, Wada N, Kohara M, et al. Presence of B-cell clones in T-celllymphoma. Eur J Haematol 2011;86:412-9.

[71] Tan BT, Warnke RA, Arber DA. The frequency of B- and T-cell generearrangements and Epstein-Barr virus in T-cell lymphomas: a com-

parison between angioimmunoblastic T-cell lymphoma and peripheralT-cell lymphoma, unspecified with and without associated B-cellproliferations. J Mol Diagn 2006;8:466-75.

[72] Pechloff K, Holch J, Ferch U, et al. The fusion kinase ITK-SYKmimics a T cell receptor signal and drives oncogenesis in conditionalmouse models of peripheral T cell lymphoma. J Exp Med 2010;207:1031-44.

[73] Dierks C, Adrian F, Fisch P, et al. The ITK-SYK fusion oncogeneinduces a T-cell lymphoproliferative disease in mice mimickinghuman disease. Cancer Res 2010;70:6193-204.

[74] Feldman AL, Sun DX, Law ME, et al. Overexpression of Syktyrosine kinase in peripheral T-cell lymphomas. Leukemia 2008;22:1139-43.

[75] Attygalle AD, Kyriakou C, Dupuis J, et al. Histologic evolution ofangioimmunoblastic T-cell lymphoma in consecutive biopsies: clinicalcorrelation and insights into natural history and disease progression.Am J Surg Pathol 2007;31:1077-88.

[76] de Leval L, Gisselbrecht C, Gaulard P. Advances in the understandingand management of angioimmunoblastic T-cell lymphoma. Br JHaematol 2009;148:673-89.

[77] Dogan A, Attygalle AD, Kyriakou C. Angioimmunoblastic T-celllymphoma. Br J Haematol 2003;121:681-91.

[78] Grogg KL, Jung S, Erickson LA, McClure RF, Dogan A. Primarycutaneous CD4-positive small/medium-sized pleomorphic T-celllymphoma: a clonal T-cell lymphoproliferative disorder with indolentbehavior. Mod Pathol 2008;21:708-15.

[79] Bentebibel SE, Schmitt N, Banchereau J, Ueno H. Human tonsil B-celllymphoma 6 (BCL6)-expressing CD4+ T-cell subset specialized forB-cell help outside germinal centers. Proc Natl Acad Sci U S A2011;108:E488-97.

[80] McHeyzer-Williams LJ, Pelletier N, Mark L, Fazilleau N, McHeyzer-Williams MG. Follicular helper T cells as cognate regulators of B cellimmunity. Curr Opin Immunol 2009;2:266-73.

[81] Vinuesa CG, Cook MC. Blood relatives of follicular helper T cells.Immunity 2011;34:10-2.

[82] Morita R, Schmitt N, Bentebibel SE, et al. Human blood CXCR5(+)CD4(+) T cells are counterparts of T follicular cells and containspecific subsets that differentially support antibody secretion.Immunity 2011;34:108-21.

[83] Miyoshi H, Sato K, Niino D, et al. Clinicopathologic analysis ofperipheral T-cell lymphoma, follicular variant, and comparison withangioimmunoblastic T-cell lymphoma. Am J Clin Pathol 2012;137:879-89.