Survival Signals and Targets for Therapy in Breast Implant–Associated ALK⁻ Anaplastic Large Cell Lymphoma

Breast implant–associated (BIA) T-cell anaplastic large cell lymphoma (ALCL) is a recently recognized clinical entity, with 80 cases identified worldwide to date and four disease-specific fatalities (1–15). BIA-ALCL presents commonly as a late seroma and/or tumor mass attached to the scar capsule containing malignant cells an average of 5.8 years after implant placement (range, 0.4–20 years; ref. 13). While most cases are indolent and respond well to capsulectomy with local adjuvant radiation therapy, 10% of cases present with metastasis and 5% of cases are fatal (12, 13).

T-ALCL is a subset of adult peripheral T-cell lymphomas (PTCL) with strong CD30 positivity and consisting of pleomorphic epitheliod tumor cells with blast-like appearance, severe cellular and nuclear atypia, and large nuclei and nucleoli (16–18). A subset expresses the anaplastic lymphoma kinase (ALK) as a result of reciprocal (2;5) translocation between the nucleophosmin (NPM1) gene and kinase domain of the ALK (16–19). Disease is subcategorized as ALK⁺ systemic, ALK⁻ systemic, or primary cutaneous (pc-) ALCL, and each group exhibits distinct clinical behavior (16, 18). ALK⁻ systemic ALCL is aggressive, with a 5-year overall survival (OS) rate of only 49%, compared with ALK⁺ ALCL (70% 5-year OS rate) and pc-ALCL (90% 5-year OS rate; ref. 20). Seroma-associated ALCL was proposed by Roden and colleagues (5) in 2008 to address BIA-ALCL, which shares morphologic features of both primary systemic ALK⁻ ALCL and pc-ALCL but is distinct in its presentation with malignant seroma fluid and varied clinical progression (indolent to aggressive). T-ALCLs express a range of immune markers, including T-cell antigens, cytotoxic granules, and antigen presentation molecules, and, like other T-cell neoplasms, show clonal T-cell receptor (TCR) gene rearrangement (21–23).

As more cases of BIA-ALCLs are recognized, questions about tumor etiology have emerged and the identification of effective treatments becomes more important. Previously, we established the first model cell line for BIA-ALCL, designated TLBR-1, for studies of this disease (1). Since that initial report, we have established and characterized 2 new cell lines from patients with BIA-ALCL, including 1 of the 4 fatal cases, designated TLBR-2 and -3. Using these models of BIA-ALCLs, we now describe fully the phenotypic and functional features of this newly emerging clinical entity, including identification of aberrancies in cell signaling and apoptosis regulators that seem to be excellent molecular targets for therapy.

Karpas299 (Karpas), Raji, HUT102, and Jurkat cell lines were obtained from American Type Culture Collection (authentication by short tandem repeat) and maintained in RPMI-1640 with 10% fetal calf serum, 2 mmol/L l-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin in a humidified 5% CO₂, 37°C incubator. Institutional Review Board (IRB) approval from the USC Keck School of Medicine (University of Southern California, Los Angeles, CA; HS-0600579) was obtained for the collection of peripheral blood mononuclear cells from healthy donors.

Karyotype analysis was conducted by the Division of Anatomic Pathology, City of Hope (Duarte, CA) using early passages of each cell line. Patient cytogenetic information was reported by the treating physician.

Xenografts of the TLBR cell lines were attempted in 6- to 8-week-old female nude, severe combined immunodeficient (SCID; Harlan), nonobese diabetic (NOD)/SCID, and NOD/SCID-γ mice (Jackson Labs) using 10⁶ cells in a 0.2-mL subcutaneous inoculum.

Formalin-fixed, paraffin-embedded (FFPE) xenograft tumors or cultured cells were sectioned and stained using hematoxylin–eosin (H&E), Wright–Giemsa (W–G), or monoclonal antibodies (Supplementary Table S1) using immunocytochemical techniques, as described previously. Observation and image acquisition were made using a Leica DM2500 microscope (Leica Microsystems, www.leica-microsystems.com), digital SPOT RTke camera, and SPOT Advanced Software (SPOT Diagnostic Instrument Inc., www.diaginc.com). Images were resized and brightened for publication using Adobe Photoshop software (Adobe, www.adobe.com).

Single-cell suspensions were stained with fluorescence-conjugated antibodies as described previously (24). Antibodies and isotype controls were from BD Biosciences and eBiosciences (Supplementary Table S1). Samples were run in duplicate on a BD FACSCalibur flow cytometer using CellQuestPro software. Mean fluorescence intensity (MFI) and percentage of positive staining cells (difference between MFI of sample and isotype >50) were determined for 15,000 events.

PCR to assess TCRγ gene rearrangement and screen for oncogene incorporation was carried out as described previously (1, 21, 25–27). Quantitative reverse transcriptase PCR (qRT-PCR) to measure gene expression was carried out as described previously (24). Gene-specific amplification was normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and fold change calculated relative to healthy donor peripheral blood T cells. Differences in mean expression of tumor suppressor, proto-oncogenes, and apoptosis-related genes among tumor cell lines and normal donor T cells were evaluated for statistical significance by ANOVA followed by Dunnett test.

Whole-cell lysates in radioimmunoprecipitation (RIPA) buffer were fractionated on 10% Tris-glycine PAGE, electrotransferred to nitrocellulose, and probed overnight for target proteins with primary antibodies (Cell Signaling and Santa Cruz Biotech; Supplementary Table S1), as described previously (1). Protein concentration was determined by the bicinchoninic acid (BCA) assay.

Levels of cytokines in supernatants from 24-hour cultures of TLBR-1, -2, -3, Karpas299, or Jurkat cells were measured by cytometric bead array and analyzed on a BD LSRII flow cytometer using FACSDiva software for acquisition and compensation. Differences in mean levels of cytokine production were tested for statistical significance by ANOVA followed by Dunnett test with comparison to medium alone.

TLBR-1, -2, -3, and Karpas299 cells were cultured (10⁶ cells/mL) in vitro in the presence or absence of cell signaling inhibitors or chemotherapeutic drugs. For the chemotherapy studies, cells were exposed to drug or vehicle for 30 minutes, then washed twice with cold complete medium, and cultured in the absence of drug for 48 hours. Cell viability was measured by staining with Annexin V/propidium iodide (PI; Invitrogen) and analyzed on a BD LSRII flow cytometer using FACSDiva software for acquisition and compensation (≥15,000 events per sample). Reagents evaluated included WP1066, sunitinib malate, honokiol, and 4-hydroperoxycyclophosphamide (4HC; Santa Cruz); S3I-201 (EMD Chemicals); 5-aza-2′-deoxycytidine (AZA), vinblastine, doxorubicin, FBHA, DAPT, suberoyl bis-hydroxamic acid (SBHA), and valproate (Sigma). Mean percentages of positive staining cells for groups of cells treated with inhibitors or vehicle alone were evaluated by ANOVA and Dunnett post-test or pairwise comparisons by the Student t test with Bonferroni correction.

The TLBR cell lines were established from the primary tumor specimens of women with BIA-ALCL, as summarized in Table 1, under IRB-approved protocol HS-10-00254 and reported previously for TLBR-1 (1). These cases were typical of BIA-ALCLs in that the women presented with unilateral seroma fluid accumulation 3 to 15 years after elective breast augmentation with textured saline implants (Fig. 1A). The seromas uniformly recurred within months of initial drainage and were found to contain malignant cells consistent with ALK⁻ ALCLs (Fig. 1B). All patients underwent bilateral implant removal and capsulectomy and had no evidence of contralateral breast involvement, skin manifestations, or spread to regional lymph nodes at that time. Patients 1 and 3 received local radiotherapy to the affected breast and chest wall following surgery and remain disease free at the time of publication. Patient 2 developed a local recurrence 2 months after surgery with involvement of the axillary and supraclavicular lymph nodes and bilateral pleural effusions. She received radiation therapy to which she showed a dramatic response with significant tumor shrinkage. However, computed tomographic imaging of the chest 2 months later showed spread of her disease into the mediastinum with airway compression and bilateral lung infiltrates. Her status progressively worsened and she died of ALCL-related complications 9 months after initial presentation.

All 3 TLBR cell lines grow in suspension cultures and exhibit interleukin (IL)-2–dependent growth (Table 1). Wright–Giemsa staining of cytospin preparations of the TLBR cell lines showed cells with abundant cytoplasm, 1 to 4 large cytoplasmic vacuoles, enlarged nuclei, and prominent nucleoli characteristic of other ALCLs and similar to the original specimens (Fig. 1C; refs. 23, 28). Multiplex PCR analysis of TLBR-1, -2, and -3 cells showed TCRγ monoclonality, confirming a neoplastic T-cell origin of the cell lines and their derivation from the T-ALCL patient specimens (Table 1).

Conventional cytogenetic and spectral karyotyping analysis of mitotically active TLBR-2 and -3 cells showed clonally abnormal, hypertriploid, and complex karyotypes (Supplementary Fig. S1). TLBR-2 cells had a modal number of chromosomes of 76 and showed gains of chromosomes 1, 2, 5, 6, 10, 11, 14, 17, as well as clonal loss of one copy of chromosome 18, relative to a triploid genome. TLBR-3 cells showed a modal number of chromosomes of 81, gains of chromosomes X, 2, 5, 7, 8, 10, 11, 12, 14, 19, 20, 21, and 22, and clonal losses of one copy of chromosomes 9, 16, and 17, relative to a triploid genome. Cytogenetic analysis of the TLBR-1 cell line was reported previously (1). None of the TLBR cell lines show the NPM-ALK t(2;5) (consistent with the primary tumor specimens), the (7;9) translocation reported in T-cell lymphoblastic leukemia or rearrangements involving the TCR gene loci on chromosomes 7 and 14. All 3 TLBR cell lines also lacked other translocations frequently found in germinal center cell, mantle cell, diffuse large B-cell, and Burkitt lymphomas: t(14;18), t(11;14), t(3;14), t(3;22), t(8;2), t(8;14), and t(8;22) (23, 28).

Immunophenotypic characterization of TLBR-1, -2, and -3 xenograft tumors (Fig. 1D) and cells in culture [Supplementary Fig. S2 or previously shown (ref. 1)] showed similarity to the original tumor specimens. All 3 TLBR models showed strong CD30 positivity, weak expression of epithelial membrane antigen (EMA), and absent ALK-1 [t(2;5) product], keratins (squamous tissue antigen) or nuclear PAX-5 (Hodgkin lymphoma antigen; ref. 28).

To understand better the BIA-ALCL cell biology, the TLBR cell lines were characterized for expression of normal T-cell lineage markers and transcription factors. Expression of immune-related proteins by TLBR cell lines was examined by flow cytometry (Supplementary Table S2). The TLBR cell lines varied in their expression of T-cell lineage markers, CD4, CD8, and TCRαβ/γδ, suggesting arrest at different stages of maturation. Consistent with their T-cell origin and IL-2 dependence, TLBR cell lines were positive for CD25 (IL-2Rα) and CD122 (IL-2Rβ). TLBR-1, -2, and -3 cell lines showed positivity for cytotoxicity protein Granzyme B and strong expression of antigen presentation–associated markers (HLA-DR⁺CD80⁺CD86⁺) and CD71, the transferrin receptor. Expression of adhesion (CD11c⁺CD11b⁻) and myeloid (CD13⁺CD14⁻CD15⁺CD68⁻) markers was variable, and TLBR cells generally lacked B-cell (CD10⁻CD19⁻CD20⁻CD21⁻CD23⁺), dendritic cell (DC; CD1a⁻), and stem cell (c-kit⁻CD133⁻) markers. In regard to normal T-cell lineages, analysis of T-cell transcription factors [Th1 (T-bet), Th2 (GATA-3), Th17 (RORγ), and T-regulatory (FoxP3)] showed strongest positivity for T-bet and FoxP3 and weak positivity for RORγ.

Aberrant expression of cell-cycle control genes and escape from homeostatic programmed cell death pathways can facilitate lymphoma tumorigenesis and progression (29–32). In this study, expression of tumor suppressor, (proto-)oncogenes, and apoptosis regulators [antiapoptotic: survivin, BCL2L2, MCL-1 (short transcript), BCL-2; proapoptotic: BID, BAX, BBC3, BAK] was evaluated in TLBR-1, -2, and -3 and established PTCL cell lines Karpas299 and Jurkat using qRT-PCR techniques. As shown in Fig. 2A, ALCL cell lines (TLBR and Karpas299) showed significant upregulation of antiapoptotic genes survivin (P < 0.05) and BCL2L2 (P < 0.05). Strong expression of survivin by ALCL cell lines was confirmed by immunoblotting and showed similar levels of expression among the TLBR cell lines. Proapoptotic genes and tumor suppressor genes were significantly downregulated relative to normal donor T cells, most notably for BID, BAK, and BBC3, with some variance among cell lines (Fig. 2B). The TLBR cell lines were also evaluated for incorporation of oncogenic viruses human T-cell leukemia virus (HTLV)-1/2 and Epstein-Barr virus (EBV), and T-cell acute lymphocytic leukemia (T-ALL)-associated oncogenes TAL1, HOX11, LYL1 and LMO1/2, and the results of these studies were negative (data not shown).

Activation of STAT3 can upregulate survival signals and downregulate proapoptotic mediators in lymphoid cells (29–31). Immunoblotting confirmed increased translation and activation of STAT3 proteins in these models (Fig. 2C), with the greatest activity in the cell line derived from the most aggressive case of BIA-ALCL (TLBR-2), and at levels comparable with STAT3-overexpressing Karpas299 cells (33, 34). High levels of pSTAT3 were also seen in xenograft tumors of TLBR-1, -2, and -3 (Fig. 2D).

ALK expression drives STAT3 activation and survival in ALK⁺ ALCLs (34, 35), but in the absence of this translocation or activating point mutations (data not shown; ref. 36), the driver of high pSTAT3 in the BIA-ALCL cell lines was unclear. Expression of T-cell cytokines (IL-2, IFNγ, TNFα, IL-10, IL-4, IL-6, and IL-17A), immunosuppressive cytokine TGFβ, and angiogenic factor VEGF-A was measured for the TLBR cell lines in culture (Fig. 3A). The TLBR cell lines showed strong secretion of cytokines associated with multiple T-cell subsets, most notably IL-6 and IL-10, compared with other PTCL models (Jurkat, Karpas299). We hypothesized that survival signals in these cells may be driven, in part, by autocrine responses to cytokines, many of which act through JAK/STAT signaling. TLBR-1, -2, and -3 were uniformly positive for the IL-6 receptor (Fig. 3B), and TLBR-2 and -3 showed weak positivity for the IL-10 receptor (data not shown), suggesting that these cells are capable of responding to these factors. Neutralization experiments for IL-6 showed a modest but insignificant decrease in TLBR cell proliferation (data not shown), likely related to the very high levels of IL-6 produced by these models. Regulatory T cell (Treg)-like suppressive function was also suggested for TLBR cell lines by FoxP3⁺ and IL-10 and TGFβ secretion (TLBR-2 and -3). To evaluate suppressive ability, TLBR cell lines were co-cultured with naive normal donor T cells in the presence of CD3/CD28 stimulation, and T-cell proliferation was measured by carboxyfluorescein succinimidyl ester (CFSE) dilution after 3 days, as carried out routinely by our laboratory (24). Surprisingly, all 3 TLBR cell lines were found to augment T-cell proliferation (data not shown), perhaps as a result of their strong production of T-cell–activating cytokines (e.g., IFNγ, IL-2). The TLBR cell lines are strongly positive for IL-2Rα and IL-2Rβ, make detectable amounts of IL-2 in culture, exhibit IL-2–dependent growth in vitro, and show more rapid growth when cultured at higher density.

To determine the influence of JAK/STAT3 signaling on TLBR cell survival, cells were cultured in the presence of STAT3-specific inhibitors WP1066 and S3I-201 or JAK/STAT3-targeted tyrosine kinase inhibitor sunitinib, and cell viability was assessed by Annexin V/PI staining. As shown in Fig. 3C, STAT3-specific inhibition by WP1066 produced significant cell death in all 3 TLBR cell lines in a dose-dependent manner. Similar effects on cell viability were seen with S3I-201 (data not shown). Furthermore, sunitinib produced striking cell death in all TLBR cell lines across a range of doses (Fig. 3D). The ALK⁺ ALCL cell line Karpas299 was run in parallel as a positive control in these experiments.

STAT3 activation can also result from decreased levels of negative regulating phosphatase SHP-1 (33, 35). TLBR cells had significantly downregulated SHP-1 expression (P < 0.05) and decreased SHP-1/STAT3 ratios (P < 0.05) compared with normal donor T cells (Fig. 4A and B). TLBR-2 and -3 had the most dramatic loss of SHP-1 expression relative to STAT3, even relative to Karpas299, an ALCL model previously reported to have significant SHP-1 loss (37). SHP-1 activation by honokiol produced significant cell death in the TLBR cell lines, with loss of pSTAT3 confirmed in cell lysates by immunoblotting (Fig. 4C and D). In addition, the chemotherapeutic agent 5-aza-2′-deoxycytidine (AZA), which was previously shown to increase levels of SHP-1 protein in PTCL cell lines (37), produced dose-related cell death in TLBR cells (Supplementary Fig. S3).

Evaluation of TLBR and established PTCL cell lines showed strong expression of Notch1 and Notch2 receptors and unique expression of a major Notch ligand, Jagged 2, on the 3 TLBR cell lines (Supplementary Fig. S4). Aberrant expression and activation of the embryonic transcription factor Notch1 can contribute to malignant transformation in some adult PTCLs (38). Levels of cleaved, activated Notch1 protein were previously found to be elevated in TLBR-1 and Karpas299 cells (1). TLBR-2 and -3 also have significant cleaved Notch1 and Notch1 levels (Supplementary Fig. S4). The much higher levels of cleaved Notch1 in TLBR-2 cells may drive the faster division and more aggressive behavior of this model and the tumor from which it was established. However, modulation of Notch1 signaling using γ-secretase inhibitors (FBHA, DAPT) or activators (SBHA, valproate) failed to produce any significant change in TLBR-1, -2, -3, or Karpas299 cell viability (Supplementary Fig. S4).

In cases of BIA-ALCLs requiring adjuvant therapy after capsulectomy, cyclophosphamide–Adriamycin–vincristine–prednisone (CHOP) chemotherapy may be beneficial (39). To estimate BIA-ALCL sensitivity to CHOP, the TLBR model cells lines were exposed to CHOP constituent drugs [vinblastine (vincristine analogue with in vitro activity), doxorubicin, 4-hydroperoxycyclophosphamide (active metabolite of cyclophosphamide)] briefly and cell viability was then assessed. As shown in Fig. 5A, TLBR-1, -2, and -3 cells were highly sensitive to doxorubicin treatment (>80% cell death after 30-minute exposure to lower dose of 1.75 μmol/L, P < 0.001, and near-complete cell death at 17.5 μmol/L dose, P < 0.001). The TLBR cell lines showed moderate sensitivity to vinblastine (0.9 and 9 μmol/L) and to a very high dose of 4-hydroperoxycyclophosphamide (100 μmol/L; Fig. 5B and C).

As recently reported, breast implant–associated T-cell anaplastic large cell lymphomas are an emerging clinical entity (2–15). Three model cell lines, designated TLBR-1, -2, and -3, have been established from the primary tumor specimens from patients with a spectrum of indolent to aggressive BIA-ALCLs to facilitate studies of the etiology and potential therapy for these cancers. Morphologic and cytogenetic studies confirmed the ALK⁻ T-ALCL classification of the BIA-ALCL TLBR cell lines and their similarity to the original tumor biopsy specimens. The molecular features of ALK⁻ ALCLs and other ALCL subsets are largely unknown, a fact that is mirrored by the 30% to 50% of ALCL cases designated as not otherwise specified (ALCL-NOS) by histopathology (40). Functional studies of the TLBR cell lines identified high production of T-cell–associated cytokines IL-6 and IL-10, activation of JAK/STAT3 signaling pathways, and strongest expression of transcription factors associated with the T-helper cell (T_H)1 and Treg cell lineages. This molecular profile may be compared with that recently reported for ALK⁺ systemic ALCLs, namely upregulation of T_H17-related genes [e.g., IL-17A, IL-22, retinoic acid–related orphan receptor (ROR)], JAK/STAT3 signaling, and cytotoxic molecules (32).

Compared with naive, normal donor T cells, the TLBR cell lines showed dysregulation of cell-cycle and apoptotic regulators, namely survivin, and activation of JAK/STAT3 pathways. Functional characterization and in vitro studies of the TLBR cell lines yielded a working model of BIA-ALCL tumor cell biology (Fig. 5D), with an emphasis on autocrine cytokine signaling that promotes tumor cell survival. A milieu rich in immune stimulatory cytokines, like IL-6 and IL-2, which promotes rapid division of host lymphocytes may cause the initial tumorigenic changes that lead to BIA-ALCL in some patients. Chronic inflammation is well recognized as a promoter of cancer (41). Autocrine IL-6 production has been identified as a driver of tumorigenesis in some diffuse large B-cell lymphomas, as well as solid tumors including breast, lung, and ovarian carcinomas (42–44). The cytokine profile of BIA-ALCL cell lines, specifically IL-6, TGFβ, and IL-10, has also been shown to induce immune suppressor cell populations (Tregs and myeloid-derived suppressor cells) that could inhibit host antitumor immunity and facilitate cancer development (45, 46). Previous studies of women with saline and silicone breast implants found no increased risk of primary lymphoma or breast cancer compared with women without implants (15), suggesting that the present case series is directly linked to newer device features. Texturing of the implant silicone shell, an aesthetic advance introduced in the late 1980s to reduce contractures and one recurring feature in these cancer cases, results in greater silicone particle shedding in the surrounding scar capsule. Indeed, histologic analysis of mass lesions in cases of BIA-ALCLs, including patient 3 (Fig. 1A), shows nonrefractive particles consistent with shed silicone among granulomatous inflammation and tumor cells. Whether this inflammatory stimulus is increased in textured implants and may play a role in the development of BIA-ALCLs are questions that will require future investigation, but the prominent role of IL-6 found in the TLBR cell lines suggests that immune reactions are important to the progression of this disease.

It is important also to acknowledge that IL-2 signaling almost certainly contributes to BIA-ALCL cell survival, as the TLBR cell lines all die in the absence of this cytokine and have strong expression of IL-2R proteins. In experimental systems, IL-2 overexpression has been shown to produce autonomous cell growth and malignant transformation in T cells (47, 48). Because the TLBR cell lines produce only low levels of IL-2, insufficient to sustain their own survival in culture, it is likely that other immune cells in the implant microenvironment are present and actively secreting this factor. This would also be consistent with the development of BIA-ALCLs in the setting of ongoing host inflammatory responses at the implant scar capsule.

Notch1 activation in the TLBR cell lines was interesting because the highest levels were observed in TLBR-2, the cell line derived from a treatment-resistant, fatal case of BIA-ALCLs. Notch1 activation therefore might be a marker of more aggressive diseases, and studies to evaluate cleaved Notch1 levels in tumor specimens from patients with BIA-ALCL may provide useful prognostic information. γ-Secretase inhibitors failed to affect cell viability in vitro, suggesting that the cells have sufficiently strong survival signals provided by other factors or pathways to overcome inhibition of Notch1. Future studies evaluating Notch inhibition in combination with cytokine signaling interruption may identify highly effective therapies for aggressive cases of BIA-ALCLs.

Using newly established models of BIA-ALCLs, we have made significant improvements in the understanding of the biology of these tumors and identified potential targets for therapy that are readily translatable to the clinic. The identification of a successful xenotransplantation model for the TLBR cell lines should facilitate future evaluation of targeted therapies against cytokine pathways (e.g., IL-6, IL-2) and cell survival molecules (e.g., survivin), as well as confirmation of chemotherapy sensitivity, in the in vivo setting. The TLBR cell lines have been deposited with the American Tissue Culture Collection (www.atcc.org).

A.L. Epstein has a commercial research grant from Mentor Corporation and Allergan, Inc. No potential conflicts of interest were disclosed by the other authors.

Conception and design: M.G. Lechner, C.H. Church, R.B. Sevell, A.L. Epstein

Development of methodology: M.G. Lechner, C. Megiel, C.H. Church, S.M. Russell, R.B. Sevell, A.L. Epstein

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C. Megiel, C.H. Church, T.E. Angell, S.M. Russell, J.K. Jang, G.S. Brody

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.G. Lechner, C. Megiel, C.H. Church, T.E. Angell, S.M. Russell, J.K. Jang

Writing, review, and/or revision of the manuscript: M.G. Lechner, C. Megiel, C.H. Church, S.M. Russell, R.B. Sevell, J.K. Jang, A.L. Epstein, T.E. Angell

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R.B. Sevell, A.L. Epstein

Study supervision: A.L. Epstein

Execution of experiments: M.G. Lechner

The authors thank the expert work of Victoria Bedell and the City of Hope Cytogenetic Core Facility (Duarte, CA) in conducting the cytogenetic studies; and Michael F. Bohley (Aesthetic Breast Care Center, Portland, OR), Thomas W. Martin (Puget Sound Institute of Pathology, Seattle, WA), and James H. Blackburn (Plastic Surgery Bellingham, Bellingham, WA) in the clinical care of the patients and the collection of specimens and clinical information for these studies.

This work was supported by Mentor Corporation, Allergan, Inc., and Cancer Therapeutics Laboratories, Inc., of which A.L. Epstein is a co-founder.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.