3855991698

305

PARK2/FRA6E and Afadin in breast cancer

A Letessier et al

FRA6EJPARK2. The multivariate analysis showed that Afadin could be a good prognosis marker. In associa-tion with gra de and Ki67 status, Afadin status may help in the detection of patients with poor prognosis in the lymph node-negative population. Moreover, in the FHIT-negative patient population, Afadin is a marker of poor prognosis. The combined analysis of FHIT and Afadin could be useful to discriminate patients with adverse outeome in the whole population. The coordi-nated loss of FHIT and WWOX expression has bccn found in breast cancers (Guler et al.. 2005). We have similarly shown here that loss of break of PARK2 and loss of FHIT cxpression arc concomitantly found in some samples. These cases are associatcd with a poor prognosis. We have not tested WWOX expression on our series.

Finally, we showed that loss of Afadin expression affects adherence of cells in culture. This is in perfect agreement with in vivo data; celi celi adherens and tight junctions are improperly organized in the ectoderm of Af-6 (-/—) mice and embryoid bodies (Ikeda et a!., 1999; Zhadanov et a/., 1999). A thorough study has recently described the role of Afadin in the reeruitment of E-cadherin and tight junction compo-nents at celi celi junctions (Sato et al., 2006). Like E-cadherin, Afadin exprcssion may be lost in a subgroup of breast cancers.

In conclusion. our data suggest that: (i) frequent breaks in tumors at a CFS should not automatically point to an intrasite gene as involved in cancer; breaks may inducc loss of nearby TSG; (ii) Afadin may be a new marker of advcrsc evolution in patients with apparent good prognosis at diagnosis; (iii) breast tumors with concomitant loss of FRA6E break, Afadin loss and FHIT loss may constitute a subclass with inereased genomie fragility; (iv) Afadin may have a role in mammary oncogenesis. Ii acts as a tumor suppressor whose loss of cxpression disrupts epithelial integrity and may favor metastasis.

Patients aml methods

Patients and histological samples

A consecutive series of 547 unilaterał localized invasive breast carcinomas from women treated at the Institut Paoli-Calmettes between October 1987 and December 1999 was studied. According to the WHO classification, this series comprised 386 ductal, 72 lobular, 37 tubular, 8 medullary carcinomas and 44 other histological types. The average age at diagnosis was 59 years (rangę 25 94 years). A total of 254 tumors were associatcd with lymph node invasion and 403 were positive for estrogen receptor. Of the 547 cases, 190 cases were available for FISH analysis, and 473 and 352 were available for Parkin and Afadin immunostaining, respectively (Supplementary Table 1).

The various histoclinical factors collected for this series included: patient age. invasive histological type, pathological tumor si/.e, ScarfF-Bloom-Richardson (SBR) grade (I-III), peritumoral vascular invasion, axillary lymph node status, estrogen receptor expression (ER), progesterone receptor expression (PR), P53, Parkin, Afadin, as evaluated by IHC with a positivity cutoff value of 1%. ERBB2 status, evaluated by IHC with the 0 3+ score as illustrated by the HercepTest kit scoring guidelines (DakoCytomation, Coppenhagen. Denmark), and Ki67 status as evaluated by IHC with a positive cutoff value at 20%.

T\1 A construction

TM A were prepared as described previously (Ginestier et a/.. 2002). For each tumor. three representative tumor areas were carefully selected from a hematoxylin-eosin-safran stained section of a donor błock. Core cylinders with a diameter of 0.6 mm each were punched from each of these areas and deposited into three separate reeipient paraffin blocks using a specific arraying device (Beecher Instruments, Silver Spring. MD. USA). In addition to tumor tissues, the reeipient błock also received normal breast tissues and celi lines pellets. Sections 5-/im of the resulting microarrays błock were madę and used for FISH and IHC analysis after transfer onto glass slides.

Fluorescence in situ hybridization analysis FISH on TM A w as carried out according to a published protocol (Chin et a!., 2003; Huang et al., 2004). Based on the split-signal FISH approach (van der Burg et al.. 2004), we used a combination of two di [Teren tly labeled pools of BAC clones overlapping the PARK2 loeus as probes (Figurę la): from telomere to centromere, biotinylated RP11-157B17 (chr6: 163,481.350-163,680.324). RP11-117116 (AC058815: chr6: 163,177,577- 163,342,875), RP11-15318 (chr6: 162,898,135-163.041.573) (revealed in green, FITC) and digoxigenin-labeled RP11-431E19 (chr6: 161,659,375- 161.837,993), RP11-479C23 (chr6:    161,442,754-    161.620.713).    RP11-158E9 (chr6:

161.336.365-161,484,919) (revealed in red, TRITC). PARK2 is located on chromosome arm 6q, in the 161,740,081-163,119,211 genomie interval. RP11-15318 and RP11-431E19 were the two overlapping BAC clones of PARK2 used in this combination. They overlap PARK2 on 221 kb and 98kb, respectively (Figurę la). Genomie information was taken from the UCSC Genome Browser on Humań (http://genome.ucsc.edu May 2004 Assembly), which is based on NCBI Build 35 (National Center for Biotechnology Information, National Library of Medicine. Bethesda. USA).

DNA from BAC clones were purified, labeled and indivi-dually Yerified for their specificity for chromosome 6. Ali BAC clones were obtained from the BACPAC resource (ChildreiTs Hospital Oakland - BACPAC Resources, Oakland, CA, USA). After counterstaining with Vectashield containing 4,6-diamidino-2-phenylindoIe (DAPI) (Vector. Burlingame. C’A. USA), images were analysed with a microscope (DMRXA, Leica Microsystems, Marseille. France), captured with a CC'D camera, filtered and processed with ISIS software (In Situ Imaging Systems, Metasystems Hard- und Software GmbH, Altlussheim, Germany) (described in www.metasystems.de). Fluorescence was scored on a minimum of 50 nuclei per tumor. The 50 nuclei of cancer cells were representative of the overall celi heterogeneity of the tumor. Two observers (AL and CG) read the TMA independently.

Inummohistochenucal analysis

The characteristics of the antibodies used are listed in Table 1. IHC was carried out on five-/un sections of tissue fixed in alcohol formalin for 24 h and included in paraffin. Sections were deparaffinized in Histolemon (Carlo Erba Reagenti. Rodano. Italy) and rehydrated in graded alcohol solution. Antigen enhancement was carried out by ineubating the sections in target retrieval solution (DakoCytomation, Coppenhagen. Denmark) as recommended. The reactions were

Oncogene