Follicularr Variant of Papillary Thyroid Carcinoma

17 downloads 3151 Views 2MB Size Report
Aberrations in PPARG were uniquely detected in FVPTC and FTA/FTC. ...... Supportedd by: Faculty Research Fellowship from the American College of Surgeons ...
Chapterr 17

Follicularr Variant of Papillary Thyroid Carcinoma:: Genome-wide Appraisal of a Controversiall Entity

Volkertt B. Wreesmann,1 Ronald A. Ghossein,1 Michael Hezel, Debendranathh Banerjee, Ashok R. Shaha, R. Michael Tuttle, Jatin P. Shah,, Pulivarthi H. Rao, Bhuvanesh Singh +

Equall contributors

Genes,, Chromosomes & Cancer 2004; 40(4):355-364

285 5

286 6

Follicularr Variant of Papillary Thyroid Cancer Abstract t Thee majority of thyroid tumors are classified as papillary (papillary thyroid carcinomas; PTCs)) or follicular neoplasms (follicular thyroid adenomas and carcinomas; FT A/FTC) based onn nuclear features and the cellular growth pattern. However, classification of the follicular variantt of papillary thyroid carcinoma (FVPTC) remains an issue of debate. These tumors contain a predominantlyy follicular growth pattern but display nuclear features and overall clinical behaviorr consistent with PTC. In this study, we used comparative genomic hybridization (CGH) to comparee the global chromosomal aberrations in FVPTC to the PTC of classical variant (classicall PTC) and FT A/FTC. In addition, we assessed the presence of peroxisome proliferator-activatedd receptor-gamma (PPARG) alteration, a genetic event specific to FTA/FTC, using Southernn blot and immunohistochemistry analyses. In sharp contrast to the findings in classical PTCC (4% of cases), CGH analysis demonstrated that both FVPTC (59% of cases) and FTA/FTC (36%% of cases) were commonly characterized by aneuploidy (p=0.0002). Moreover, the pattern of chromosomall aberrations (gains at chromosome arms 2q, 4q, 5q, 6q, 8q, and 13q and deletions att lp, 9q, 16q, 17q, 19q, and 22q) in the follicular variant of PTC closely resembled that of FTA/FTC.. Aberrations in PPARG were uniquely detected in FVPTC and FTA/FTC. Our findingss suggest a stronger relationship between the FVPTC and FTA/FTC than previously appreciatedd and support further consideration of the current classification of thyroid neoplasms.

T

hee vast majority of thyroid neoplasms are well-differentiated tumors of follicularr cell origin. The differential diagnosis of these lesions includes papillaryy carcinomas (PTCs) and follicular neoplasms consisting of follicular carcinomass (FTCs) and benign follicular adenomas (FTAs). Clinically, PTC is characterizedd by a tendency for multicentric involvement, a predilection for lymphogenous spread,, and a typically indolent clinical course, with overall survival exceeding 90% afterr 20 years. In contrast, FTCs are characterized by unifocal disease, presence of tumor encapsulation,, a propensity for hematogenous dissemination, and lower overall survival relativee to PTC (80% at 20 years). Given their divergent biology, the pathologic distinctionn among PTC, FT A, and FTC has important clinical implications [28, 42]. Thee presence of characteristic nuclear features (optically clear nuclei and nuclear grooves/pseudo-inclusions)) in a neoplastic thyroid lesion is a condition sine qua non for thee diagnosis of PTC. [42] Thyroid neoplasms with classical nuclear features and a papillaryy growth pattern (i.e., the classical variant of PTC, or classical PTC) leave little doubtt about their diagnosis. However, the presence of these nuclear features in a thyroidd neoplasm with an entirely follicular growth pattern may generate significant diagnosticc debate. [2] Given their common growth pattern, these neoplasms traditionally weree considered to be either FTC or FT A, depending on the respective presence or absencee of capsular/vascular invasion. [34, 2] In 1980, their reclassification as a variant off PTC (follicular variant of PTC; FVPTC) was justified by their overall clinical behaviorr in short-term follow-up studies. [14, 6, 33,8, 32, S3] Nonetheless, several reports continuedd to suggest that individual cases of FVPTC may mimic the pathologic features

287 7

Chapterr 17 andd clinical behavior of FT A/FTC. [5/, 48, 13, 1, 23, 22, 54, 55] For example, FVPTCs weree shown to have a higher rate of tumor encapsulation and hematogenous disseminationn and a lower rate of lymphatic metastasis compared to classical PTC. [48, 1, 23, 54, 55]55] At present, a genetic basis for these observations remains obscure. Here,, we report on a genome-wide comparison of FVPTC, classical PTC, and FTT A/FTC, in which we used comparative genomic hybridization (CGH) [24] analysis to definee pathogenetic relationships. Our data demonstrate that the presence and profile of chromosomall aberrations in FVPTC are strikingly different from those in classical PTC butt similar to FT A/FTC. In addition, we show that alterations in peroxisome proliferator-activatedd receptor-gamma (PPARG) [27, 31, 35, 7] are unique to FVPTC and FTT A/FTC. These data suggest that the pathogenesis of (at least) some cases of FVPTC iss more closely related to that of FT A/FTC than to that of PTC. Alll cases of thyroid neoplasia operated on at the Memorial Sloan-Kettering Cancer Centerr with available fresh frozen tissue obtained according to the guidelines establishedd by the institutional review board were identified by a search of the Department off Pathology database. A detailed histopathologic review of cases was performed by two pathologistss with extensive experience in oncologic pathology. The categorization of our casess was based on criteria established by the most recent Armed Forces Institute of Pathologyy fascicle on thyroid neoplasia. [42]. In brief, classical PTCs were defined by theirr typical nuclear features and the presence of a papillary growth pattern. The diagnosiss of FVPTC was contingent upon the presence of tumor nuclei featuring all five of thee following features: (1) ground-glass appearance (empty, "Orphan Annie-eyed" nuclei),, (2) overlapping, (3) enlargement, (4) irregular shape, (5) nuclear grooves, and (6) pseudo-- inclusions and a follicular growth pattern comprising more than 99% of the tumor.. Most of the FVPTCs also had dense eosinophilic colloid with scalloped edges, a characteristicc feature of FVPTC. In contrast, follicular adenomas/carcinomas were solitaryy encapsulated tumors with an entirely follicular growth pattern and lacked the char-

Tablee 1. Comparison of ClinicopathologicalCharacteristics soff Study Population Clinicall factor Mediann age (range) Mediann tumor size Femalee sex Cervicall lymph node metastasis Extrathyroidall extension Distantt metastasis Recurrence e Deathh of disease

CPTC C (nn = 25) 41(20-77) ) 2.00 (0.6-4.5) 122 (48%) 122 (48%) 100 (40%) 11 (4%) 22 (8%) 00 (0%)

FVPTC C (nn = 17) 400 (25-75) 2.55 (0.8-6.5) 44 (24%) 44 (24%) 11 (6%) 00 (0%) 11 (6%) 00 (0%)

FTA A (nn = l l ) 466 (26-67) 3.2(1.5-5.0) ) 10(91%) ) 00 (0%) 00 (0%) 00 (0%) 00 (0%) 00 (0%)

FTC C (nn = 3) 466 (35-46) 3.22 (3.0-3.5) 11 (33%) 00 (0%) 11 (30%) 00 (0%) 00 (0%) 00 (0%)

PP value* NS S NS S NS S NS S 0.016 6 NS S NS S NS S

CPTC-classicall papillary thyroid carcinoma; FVPTC-follicular variant of papillary thyroid carcinoma; FTA-follicular thyroidd adenoma; FTC-follicular thyroid carcinoma; NS-not signi.cant. *PP value comparing cPTC to FVPTC.

288 8

Follicularr Variant of Papillary Thyroid Cancer acteristicc nuclear features of papillary carcinoma. The distinction of follicular carcinomass from adenomas was based on the respective presence or absence of microscopically detectablee invasion (vascular/capsular). Hurthle cell (oncocytic) tumors, defined by the presencee of >75% of oncocytic cells, were excluded from the study. In addition, cases showingg evidence of poor differentiation or anaplasia, such as signi.cant mitotic activity,, necrosis, or nuclear pleomorphism, were excluded. Based on these criteria, cases of well-differentiatedd PTC of the classical (n=25) or follicular (n = 17) variant, follicular adenomaa (n-11), and well-differentiated follicular carcinoma (n-3) with available frozenn tissue from the primary tumor were included in the study. The CGH profiles of 133 of the PTC (7 FVPTC) cases included in this study were reported previously. [Singh ett ah, 2000] A detailed comparison of clinicopathological features of the study cases is givenn in Table 1. All cases of FVPTC were assessed for the presence/absence of tumor encapsulation.. Cases with encapsulation were assessed for the presence of vascular and/orr capsular invasion. This analysis revealed encapsulation in 14 of 17 cases of FVPTC,, 12 of which were devoid of capsular/vascular invasion. Four of 17 FVPTC cases developedd clinically apparent cervical lymph node metastases, all but one of which were unencapsulatedd tumors. The single encapsulated case of FVPTC with lymph node metastasess contained multifocal capsular invasion. Genomicc DNA extraction was performed as described previously from fresh-frozen tissuee sections confirmed to contain >70% tumor tissues. [45, 46] Tumor DNA was labeledd by nick translation (Life Technologies, Rockville, MD) with fluorescein-12dUTP.. Reference DNA was extracted from normal placenta and labeled with Texas red5-dUTPP (New England Nuclear-Dupont, Boston, MA). CGH was performed according too previously published methods. [24, 45] Forr analysis, 7-10 separate metaphase cells were captured and processed with a quantitativee image processing system (Quips Pathvysion System; Applied Imaging, Santa Clara,, CA). Red, green, and blue fluorescence intensities were analyzed for all metaphase spreads,, normalized to a standard length, and statistically combined to show the red: greenn signal ratio and 95% confidence intervals for the entire chromosome. Copy numberr changes were detected based on the variance of the red: green ratio profile from the standardd of 1. Based on validation experiments performed in our laboratory [20, 21, 52], ratioss of 1.2 and 2.0 were defined as thresholds for gains and amplifications, respectively,, and losses were defined as a ratio of 0.8 or less. Several control experiments were performed,, including color-switch experiments, normal-to-normal experiments, tumor-totumorr experiments, and validation of CGH by fluorescence in situ hybridization. [20, 21, 52] 52] AA tissue microarray was constructed as described in detail previously. [18] Briefly, a singlee pathologist conducted a critical histologic slide review of all cases of PTC, identifyingg representative areas of tumor and normal thyroid. Six-millimeter tissue cores were extractedd from the defined areas and arrayed in triplicate on a recipient paraffin block

289 9

Chapterr 17 withh a precision instrument (Beecher Instruments, Silver Spring, MD). Five-micrometerr sections were taken from the constructed tissue array blocks and placed on charged poly-L-lysine-coatedd slides. These sections were used for immunoperoxidase staining. Briefly,, slides were incubated with the mouse monoclonal antibody against human PPARGG (E8) (sc-7273; Santa Cruz Biotechnology, Santa Cruz, CA), raised against the carboxyy terminus portion of human PPARG at 1:50 dilution. Microwave antigen retrievall was used for 40 min at 95°C-99°C in 10 mM citrate buffer at pH 6.0. The slides weree then incubated overnight at 4°C with the antibody.. Next, the samples were incubated with biotinylatedd anti-mouse immunoglobulins at 1:5000 dilution (Vector Laboratories, Inc., Burlingame,, CA) for 60 min, followed by peroxidase-- conjugated streptavidin at 1:500 dilution for 45-600 min at room temperature. Diaminobenzidinee was used as a chromogen and hematoxylinn as a nuclear counterstain. Full-tissue sectionss of a myxoid liposarcoma were used as a positivee control for PPARG expression, and arrayedd normal tissues served as baseline controls. .. Nuclearr staining of the tumor cells was recorded 66 i ass being faint, moderate, or strong in intensity. Thee extent of staining was defined as focal (j50% of tumor cells), 76 6

59 9

56 6

35 5

44

cPTCC

FVPTC FTA;FTC FTA

FTC

10 0

BB

4.5 5

"J J

nn 44

a n dd p o s i t i v i t y f o r P P A R G w a s d e f i n e d as d i f f u s e Figure 1. Comparison of aneuploidy in classical n u c l e a rr S t a i n i n g o f t h e t u m o r c e l l s b e i n g o f m o d - papillary carcinomas (cPTC), follicular variant T^II „ papillary thyroid carcinomas (FVPTC), follicular

eratee or strong intensity. These catena were ^ r o i d y a d ( 4 m a s

b a s e dd On t h e C o m m o n p r e s e n c e o f focal a n d f a i n t

mas

and follicular thyr0ld carcin0 .

from Memorial Sloan-Kettering Cancer

nuclearr staining in matched, nonlesional thyroid Center (FTA/FTC) andfollicularthyroid adeno,, . , ,. r mas (FTA) and follicular thyroid carcinomas tlSSUe.. M o r e o v e r , t h e y W e r e b a s e d On S t u d i e s Ol ( F T C ) reported in the literature. Follicular variant M a r q u e ss a n d N i k i f o r o v a a n d C o l l e a g u e s [31, 35], papillary thyroid carcinomas and follicular thyW h oo f o u n d t h a t d i f f u s e P P A R G i m m u n O S t a i n i n g r f o i d neoplasms displayed a significantly higher .. " frequency of aneuploidy relative to classical papil-

off moderate to severe intensity correlated with t h ee p r e s e n c e o f PAX8-PPARG . _ „ _ , _ _ ,, , , r

j ,

lary thyroi d

carcinomas. This was evident both

t r a n s l o c a t i o n b y from (A) the percentage of tumors with chromosor ._. . . . mal aberrations and (B) the median number of

RTPCR,, whereas focal and/or faint-intensity chromosomal aberrat > ns in cases with chromosostainingg did not show any translocation by RT- mai aberrations. PCR.. In the final analysis, the rate of tissue loss wass 10%, an acceptable rate that was within the range of previous experiences. [19] In additionn to the immunohistochemical analysis of PTC, PPARG staining was assessed on full-tissuee sections of the 14 follicular thyroid tumors used in this study. Forr Southern blot analysis, genomic DNA was subjected to restriction enzyme 290 0

Follicularr Variant of Papillary Thyroid Cancer

Figuree 2. (A) Ideograms showing DNA copy number changes identified by CGH in classical papillary carcinomas (a), papillaryy carcinomas of follicular variant (b), and follicular thyroid neoplasms (adenomas and carcinomas) (c). vertical liness on left side of ideogram indicate losses of the chromosomal region and those on right side indicate gains. The genomicc profile of follicular variant papillary carcinomas is strikingly different from that of classical PTCs but resembless the genomic profile of follicular thyroid tumors (adenomas and carcinomas). (B) Ideogram showing DNA copy numberr changes identified by CGH in all follicular thyroid adenomas (a) and follicular thyroid carcinomas (b) reported in thee literature [(Hemmer et al., 1998, 1999; Frisk et al., 1999; Roque et al., 2003) Hurthle cell neoplasms excluded]. Verticall lines on left side of ideogram indicate losses of the chromosomal region and those on right indicate gains.

digestionn (BamHl) and gel electrophoresis and blotted on a positively charged nylon membranee as described previously. [46] Genomic DNA extracted from genetically normall placenta was used as a negative control. Southern filters were sequentially hybridizedd with full-length cDNA clones representing PPARG (IMAGE Consortium Clonee 3447380, Research Genetics, Huntsville, AL) and PAX8 (IMAGE Consortium Clone:: 2963877, Research Genetics) after labeling with 32P by random priming (PrimeIt,, Stratagene Cloning Systems, La Jolla, CA). Southern blot experiments were repeated usingg an alternative restriction enzyme (Hindlll) for confirmation. Statisticall analyses were performed by use of the JMP4 statistical software package (SASS Institute Inc., Cary, NC). Statistical significance was defined as a two-tailedp value lesss than or equal to 0.05. Qualitative and quantitative differences were assessed with the chi-squaree test and the Kruskal-Wallis test, respectively.

291 1

Chapterr 17 CGHH analysis showed that the presence andd pattern of chromosomal aberrations in FVPTCC were significantly different from thosee in classical PTC, but comparable to follicularr thyroid tumors (Fig. 1). Chromosomal aberrationss were detected in 10 of 17 (59%) FVPTCC cases and 5 of 14 (36%) FT A/FTC casess (4 F T As and 1 minimally invasive FTC; pp = NS; Table 1). In sharp contrast, 1 of 25 classicall PTC cases contained a chromosomal aberrationn (4%; p = 0.0002). In addition, the mediann number of chromosomal aberrations inn FVPTCs (4.5) was more comparable to follicularr thyroid tumors (13) than to classical PTCss (Fig. 1). [1] Comparison of our findings too all previously reported cases of FTA and FTCC analyzed by CGH, excluding Hurthle celll tumors [IS, 11, 39] confirmed that the FVPTCss resembled these tumors in extent of aneuploidyy (Fig. 1). Twenty-three of 41 FTA Casess ( 5 6 % ) a n d 3 3 o f 4 4 F T C cases ( 7 5 % )

i..

11. jj

Wlftflf f H. .

II. .

VV V

in. . Figure 3. Panel A: Representative images of Southern

reportedd in the literature were characterized blot analysis O(PPARG(1) and PAX8(U) in papillary byy CGH-detectable genetic aberrations thyroid carcinoma - Rearrangement of PPARG was identifiedd in only one case of follicular variant papillary (p(p — Nb). A l s o , p r e v i o u s l y p u b l i s h e d M A S carcinoma [additional bands in case F2 as compared to ( 1 0 )) a n d F T C s ( 4 ) f e a t u r e d a m e d i a n n u m b e r normal (N) and cases Fl and F3 identified by arrows] Off a b n o r m a l i t i e s C o m p a r a b l e t o t h o s e i n that also showed overexpression by immunohistochemrr istry. Rearrangement ol PAXa was not identi.ed in case F2, suggesting the presence of an alternative fusion ff V I I C (p — L\j). [15, 11, 41\ I nn S h a r p c o n t r a s t tO c l a s s i c a l P T C ( c o n - p a r t n " ; T f ° r . ^ * G Panel B:lLow-power (I) and highpowerr (II) images ot immunohistochemical analysis of t a m i n gg a s i n g l e d e l e t i o n o f c h r o m o s o m e 2 2 PPARG protein dysregulation in case F2. The tumor d e t e c t e dd i n a s i n g l e c a s e ) , F V P T C d i s p l a y e d a s h o w s strong and diffuse nuclear positivity. In addiii • i i /-. r i tion, a representative H&E section of a case of FVPTC

h i g h l yy consistent profile of genomic aberra-

P ( m ) ' i s s h 0 W n.

t i o n ss w i t h m u l t i p l e i m b a l a n c e s , i n c l u d i n g r e c u r r e n tt g a i n s m o s t c o m m o n l y i n v o l v i n g .. . . , . /) 1 2 9 7 ~) A 1 1 "7/1 0 71 7 2 Z q Z t - 3 Z ,, 4 q t 3 - Z / , 3 q l 4 - Z Z , o q l l - Z 4 , ö q Z l - Z 3 ,

by cells with ground glass (empty), enlarged, irregular, overlapping nuclei with grooves. The colloid inside the follicles is dense, with scalloped edges. Other areas of this tumor showed nuclear pseudo-inclusions (x600).

The tumor is composed offollicleslined

andd 13ql4-31 and deletions involving lp3436,, 9q34, 16q22-24, 17p, 17q22-25, 19p, and 22ql3. The profile of genomic aberration in FVPTCC was similar to that in FT A/FTC (Fig. 2A), overlapping at recurrent gains of 4, 5,, 7, 12, and 13 and losses of lp, 8p, lOp, 19p, and 22. In addition, a comparison of the genomicc aberrations in FVPTC to all previously reported cases of FTA and FTC analyzedd by CGH [15,11,41] confirmed our observations (Fig. 2B). Notably, virtually every

292 2

Follicularr Variant of Papillary Thyroid Cancer abnormalityy detected in our FVPTCs matched recurrent abnormalities detected in FTAss and FTCs as reported in the literature. This finding is further supported by cytogeneticc karyotyping and allelotyping studies showing imbalance of chromosome 17 and chromosomee arms 19q and 22q to be the most common genetic abnormalities characterizingg FT A/FTC. [12, 50, 4, 25, 9, 17] Unfortunately, comparison of the CGH profile off our PTCs to previous CGH studies on PTC (3, 16, 26) is not possible because the authorss did not stratify their findings according to the histological variety of PTC. Indeed,, inclusion of morphological heterogeneous study populations (including classicall PTC, FVPTC, tall cell-variant PTC, poorly differentiated PTC, and unusual variants)) in prior studies was suggested by the rather large variation in detection rate and typess of CGH-detected abnormalities among these studies, as discussed previously. [52] Nonetheless,, it is of note that deletions of chromosome 22 were previously detected in FVPTCC using LOH and cytogenetic karyotyping. [37, 47] The exclusive detection of significantt aneuploidy (intermediate in intensity compared to that in other solid tumors; seee http://www.helsinki../cmg/cgh_data.html) in both FVPTC and FT A/FTC suggests thatt their pathogenesis is driven by chromosomal instability. [38] Because chromosomal instabilityy is a nonspecific process, the finding of multiple recurrent chromosomal alterationss overlapping in FVPTC and FT A/ FTC suggests that these aberrations were selectedd for in the course of tumor development, supporting the contention that these entitiess progress along similar pathways that are different from those involved in cPTC development.. [5, 43] Amongg the multitude of chromosomal changes present in FT A/FTC, alteration of thee chromosomal region 3p25 has been the most common genetic feature of these tumorss reported to date. [12, 4,39, 29]. Recently, the structure of one of these alterations, t(2;3)(ql3;p25),, was deciphered, showing a fusion of PAX8 (2pl3) to PPARG (3p25). [27,[27, 31, 35, 7]. Studies have implicated PPARG as the main target of this rearrangement byy suggesting that this gene may be fused to a range of alternative fusion partners, like thee rearrangements of RET in PTC. [27, 31, 7, 30, 54] The finding that overexpression off the chimeric PPARG-PAX8 protein inhibits the tumor-suppressive function of wildtypee PPARG in a dominant-negative fashion further substantiates this concept. [27, 31, 35,35, 7] We assessed our cases for aberrant PPARG protein levels by immunohistochemistryy as described previously. Overexpression of PPARG was detected in 10 of 14 follicularr thyroid tumors (8 adenomas and 2 carcinomas) and three cases of FVPTC, but none off the classical PTCs (p