this is the author manuscript accepted for publication and

Herrington Edited

Title: An Epithelial to Mesenchymal Transition programme does not usually

drive the phenotype of Invasive Lobular Carcinomas.

Short Title: EMT does not usually underpin the classic ILC phenotype

Amy E McCart Reed1*, Jamie R Kutasovic1*, Ana Cristina Vargas1, Janani

Jayanthan1, Amel Al-Murrani1, Lynne E Reid1, Rachael Chambers1,2, Leonard

Da Silva1, Lewis Melville1,3, Elizabeth Evans4, Alan Porter4, David Papadimos2,

Erik W Thompson5,6, Sunil R Lakhani1,3,7, Peter T Simpson1,7.

1The University of Queensland, UQ Centre for Clinical Research, Herston,

4029, QLD, Australia; 2Sullivan Nicolaides Pathology, Taringa, 4068, QLD, Australia; 3Pathology Queensland, The Royal Brisbane & Women’s Hospital, Herston,

4029, QLD, Australia;

4Wesley Breast Clinic, Wesley Hospital, Auchenflower, 4066, QLD, Australia; 5Institute of Health and Biomedical Innovation and School of Biomedical

Sciences, Queensland University of Technology, Kelvin Grove, 4059, QLD,

Australia 6University of Melbourne, Department of Surgery, St Vincent’s Hospital,

Melbourne, 3065, Vic Australia 7The University of Queensland, School of Medicine, Herston, 4006, QLD,

Australia. *These authors made an equal contribution

1

This article is protected by copyright. All rights reserved.

This is the author manuscript accepted for publication and has undergone full peer review but hasnot been through the copyediting, typesetting, pagination and proofreading process, which maylead to differences between this version and the Version of Record. Please cite this article as doi:10.1002/PATH.4668

http://dx.doi.org/10.1002/PATH.4668

http://dx.doi.org/10.1002/PATH.4668

Conflict of Interest

The authors declare they have no conflicts of interest

Abstract

Epithelial to mesenchymal transition (EMT) is a cellular phenotype switching

phenomenon, which occurs during normal development and is proposed to

promote tumour cell invasive capabilities during tumour progression. Invasive

lobular carcinoma (ILC) is a histological special type of breast cancer with a

peculiar aetiology – the tumour cells display an invasive growth pattern, with

detached, single cells or single-files of cells, and a canonical feature is the

loss of E-cadherin expression. These characteristics are indicative of an EMT,

or at the very least that they represent some plasticity between phenotypes.

While some gene expression profiling data support this view, the tumour cells

remain epithelial and limited immunohistochemistry data suggest that EMT

markers may not feature prominently in ILC. We assessed the expression of a

panel of EMT markers (Fibronectin, Vimentin, N-cadherin, Smooth Muscle

Actin, Osteonectin, Snail, Twist) in 148 ILC and performed a meta-analysis of

publically available molecular data from 155 ILC. Three out of 148 (2%) ILC

demonstrated an early and coordinated alteration of multiple EMT markers

(downregulation of E-cadherin, nuclear TWIST and upregulation of vimentin,

osteonectin and smooth muscle actin). However the data overall do not

2


support a role for EMT in defining the phenotypic peculiarities of the majority

of ILC.

Keywords (3-10)

Breast cancer; Invasive Lobular Carcinoma; Epithelial to Mesenchymal

Transition (EMT); E-cadherin

3


Introduction

Epithelial to Mesenchymal Transition (EMT) is a rapid and reversible cellular

phenotype modulation whereby the epithelial properties of a cell, such as

adhesion, apico-basal polarity and expression of E-cadherin, are reduced.

Concomitantly, increased expression of the mesenchymal-type proteins,

including N-cadherin, vimentin and smooth muscle actin, results in phenotypic

characteristics such as anchorage-independent growth and motility. Three

broad instances of EMT have been described: Type 1 is well-defined and

associated with normal embryonic growth and organ differentiation; Type 2 is

a mechanism of wound healing and repair; and Type 3 is related to cancer

progression and enhances tumour cell invasive and metastatic properties by

reducing cell-cell adhesion and improving migratory capacity [1].

The strongest evidence of EMT playing a role in the behaviour of breast

cancer cells arises from studies of invasive carcinomas with basal-like

features. These tumours are most frequently high grade, poorly differentiated

tumours and include metaplastic carcinomas, a special histological type of

breast cancer [2] [3] [4]. These tumours show evidence of cadherin-switching

(reduced E-cadherin and activated N-cadherin expression) [5] and the

activated expression of typical mesenchymal markers (e.g. Smooth Muscle

Actin (SMA) and Vimentin) in tumour epithelial cells [3]. This is supported by

in vitro data, whereby a core EMT gene expression signature was defined

following the induction of EMT [4]. A comparative analysis with a panel of

4


breast cancer cell lines and invasive tumours showed that the basal-B

molecular cell line subtype and metaplastic and claudin-low breast tumours

were most closely associated with this core EMT signature [4].

Invasive Lobular Carcinomas (ILC) are at the opposite end of the breast

cancer spectrum relative to those with basal-like characteristics, being low

grade, well differentiated, hormone receptor positive tumours and with a

generally good short-term prognosis. The characteristic infiltrative growth

pattern of ILC, combined with discohesion and loss of the archetypal epithelial

marker, E-cadherin, in the vast majority of cases has led some to postulate

that ILCs undergo EMT. E-cadherin inactivation is fundamental to this

hypothesis, since it contributes to dysfunctional cell-cell adhesion [6],

acquisition of anoikis resistance [7] and the typical single file invasive growth

pattern [8-10]. E-cadherin can be dysregulated in multiple ways including by

gene mutation, deletion and/or methylation and by transcriptional or post-

transcriptional mechanisms. A combination of genomic and epigenetic

inactivation is most common in ILC, but this does not account for all tumours

that show deregulated E-cadherin ([11] and see below).

It has been suggested both that EMT is important in the aetiology of the

lobular phenotype [9], and that EMT is permanent and not a dynamic

phenomenon, in ILC [12]. TWIST is an oncogenic transcription factor that

upregulates N-cadherin and down-regulates E-cadherin expression; a master

5


switch of EMT [5]. Yang et al reported that high TWIST mRNA levels were

associated with ILC implicating EMT as a driver of the lobular phenotype [10]

and TWIST expression has been shown to increase through the progression

of lobular neoplasia to ILC [13]. Countering these arguments is the evidence

that, despite the loss of E-cadherin, the neoplastic cells of ILC retain

morphological features associated with epithelial cells (i.e. large intracellular

vacuoles) [14], express epithelial cytokeratins (CK8/18, CK19) and data from

small cohorts of ILC suggest that they infrequently express EMT markers

[3,15]. However, there is some suggestion that ILC may undergo partial or

incomplete EMT [16,17] to facilitate tumour cell invasion.

Evidence against a role for EMT in ILC is also beginning to come to light from

in vitro and in vivo data. The isolated loss of E-cadherin expression in non-

tumorigenic MCF10a mammary epithelial cells to create a CDH1-/- null line

was not sufficient to drive an EMT programme [18]. Furthermore, in a murine

ILC model established by conditional mutation of CDH1 and knock out of

TP53 [19], an epithelial tumour cell phenotype was retained and E-cadherin

loss was not sufficient for EMT induction, in this context. Hollestelle et al [15]

showed that loss of E-cadherin is neither necessary, nor sufficient to induce

EMT in cultured breast cancer cells, and this was corroborated in a subset of

clinical samples [3] [15].

6


The connection between EMT and the lobular phenotype therefore remains

tenuous. To resolve this controversy we have assessed a large cohort of 148

archival ILCs for the expression of a range of EMT markers and performed a

meta-analysis of publically available molecular data from The Cancer

Genome Atlas (TCGA).

Materials and Methods

Tumour blocks for 148 ILC were retrieved from Sullivan Nicolaides Pathology

with approval from relevant human research ethics committees. Tissue

microarrays were constructed with duplicate 0.6 mm diameter tumour cores

and immunohistochemistry for a range of EMT markers was performed and

scored as detailed (Table 1). Clinical and molecular data were accessed from

TCGA [20,21] and analysed using GraphPad Prism 6.

Results and Discussion

The cohort of ILC reflected typical characteristics for this tumour type: 78%

were grade 2, 96.5% were Oestrogen Receptor (ER) and/or Progesterone

Receptor (PR) positive, 1.4% were HER2 positive, 1.4% were triple negative

(ER, PR, HER2) and 36.6% were Ki67 positive. CK19 (98.6%) and CK8/18

(100%) positivity confirmed the epithelial nature of the tumours. E-cadherin

was negative in 75.8% and aberrantly distributed in 22.8% of cases, and

p120-catenin was aberrantly located to the cytoplasm in 90% of samples.

7


We observed infrequent and focal expression of mesenchymal markers in

tumour epithelial cells (Table 2, Fig1) suggesting dynamic but not coordinated

regulation of expression. There was no significant difference in the

proportions of EMT marker expression between exclusively E-cadherin

negative cases, and those that were E-cadherin positive/aberrant (Fisher’s

exact test; data not shown). In 3 tumours (2%) there was an apparent

coordinated expression of multiple EMT markers, including nuclear

localisation of Snail and/or TWIST, as well as activated expression of

Vimentin, SMA and Osteonectin in all tumour epithelial cells, while Fibronectin

was negative (Fig1, Supplementary Figs). Cadherin switching is regarded as

an important feature of complete EMT and although E-cadherin was lost in

these cases, there was no reciprocal expression of N-cadherin. These three

tumours were HER2-negative, and CK8/18 and CK19 positive.

We consider, therefore, that these tumours exhibited a partially activated EMT

program, based on the fact that there was down-regulation of E-cadherin,

activation and nuclear localisation of TWIST and the activation of multiple

mesenchymal markers. We consider this only a partial activation based on the

fact that cells remain epithelial in their appearance, continue to express

epithelial markers, and fail to express N-cadherin. Based on these criteria, all

other tumours did not demonstrate characteristics of EMT.

8


Interestingly, one of these ‘EMT-like’ cases was triple negative and another

was basal-like, with focal CK14 positivity. The third case (Fig1) was grade 2,

ER and PR positive, but also Ki67 and p53 positive. Therefore, these three

‘EMT-like’ positive cases exhibit phenotypic characteristics consistent with an

unusual lobular phenotype and perhaps an aggressive tumour biology. One

case had co-occurring lobular carcinoma in situ (LCIS), which was also E-

cadherin negative, nuclear TWIST positive and Vimentin positive, suggesting

the activation of this ‘EMT-like program’ occurred early in the natural history of

the tumour (Fig 1).

We wondered whether this coordinated expression of EMT markers

represented an alternative mechanism of down-regulating E-cadherin to that

seen in most ILC (i.e. typically driven by (epi)genomic alterations).

Unfortunately there was either insufficient or only poor quality DNA/RNA for

these three cases to enable CDH1 gene mutation, copy number or

methylation analysis or for a more global investigation of the molecular

mechanisms underlying EMT in these tumours. We therefore performed an in

silico investigation of the TCGA breast cancer dataset using the cBioportal

[20,21] to see if there was any evidence of dysregulated EMT marker

expression in a similarly large ILC cohort. Most tumours display classic

genomic features of ILC, with haploinsufficiency at the CDH1 locus (128/154,

83.1%), and CDH1 mutation in 78/154 (50.6%). Fifteen (9.7%) ILC were

9


CDH1 wildtype (i.e. diploid and no mutation) with low CDH1 methylation

scores (Fig2), leading us to question whether a role for EMT, particularly in

this context could bypass the archetypal lobular E-cadherin

genomic/epigenetic hit(s). Six cases of the 15 showed evidence for

upregulation of at least 1 EMT marker, however, the diagnostic slides of these

cases (viewed through the Cancer Digital Slide archive) showed that 3/5 were

pleomorphic rather than classic ILC (2/5), and one was in fact an invasive

carcinoma no special type (IC NST; which we eliminated from further

analysis). The overall expression snapshot of nine EMT markers in the

wildtype/diploid ILC cohort does not differ from that of the CDH1

mutant/haploid or the CDH1 wildtype/haploid ILC cohorts (Fig2). ER positive

and negative IC-NST cohorts, as well as metaplastic carcinomas, trend

towards higher levels of mesenchymal marker expression than ILC (Fig2).

Yang et al [10] previously reported that 70% of ILC had high levels of Twist1

mRNA. This was not evident in the TCGA ILC cohort (Figure 2), where only

8/155 (5%) showed an alteration in TWIST1; one case each of gene

amplification and point mutation, and only six cases of upregulated mRNA

(3.8%). We observed TWIST protein expression in stromal fibroblasts, which

could be the source of high TWIST1 mRNA levels reported in earlier gene

expression studies. Considering the combined data from 154 TCGA samples

and our 148 archival samples, we counter the Yang [10] study’s assertion that

TWIST-driven EMT is a significant factor in ILC.

10


The apparent dynamic nature of EMT phenotype switching means it is

challenging to investigate in clinical samples. This has therefore raised

contention as to whether the phenomenon, which is readily observed in vitro,

translates to the clinical scenario of tumour invasion and metastasis, or even

whether subtle cell level changes can simply be detected.

Immunohistochemical and gene expression profiling approaches have

demonstrated that some breast cancers show broad evidence of EMT

[3,22,23], and here using the same selection of immunohistochemical markers

we identify three cases that exhibit tumour-level evidence of a partial EMT

phenotype. There are, indubitably, other possible markers of EMT not studied

here (e.g. cell polarity markers and other transcription factors) suggesting the

frequency of EMT may be higher than reported; however the coordinated

expression of multiple ‘end-stage’ mesenchymal markers remains rare.

Ideally, molecular profiling and IHC would have been performed on the same

cohort, however this was not feasible due to technical limitations.

In summary, the morphological traits of classic ILC are aligned with the

characteristic features necessary for the phenotypic switch of EMT: loss of

cell-cell adhesion, loss of apico-basal polarity and enhanced acquisition of

invasive capability [1]. Tumour cells of ILC, however, remain intrinsically

epithelial in nature, and while three ILC showed a coordinated and stable

activation of a partial EMT program, there is limited evidence to suggest that

11


the characteristic invasive growth pattern aligned with E-cadherin down-

regulation/dysfunction is an indicator of EMT, even in cases with no

detectable genomic/epigenetic alteration of the CDH1 gene.

Acknowledgements

Thank you to Admire Matsika for insightful discussions. Peter Simpson was

the recipient of a fellowship from the National Breast Cancer Foundation,

Australia; Ana Cristina Vargas was the recipient of a clinical fellowship from

the Ludwig Institute for Cancer Research. This work was funded in part by

grants from the Wesley Research Institute and Cancer Council Queensland.

Statement of author contributions

A McCart Reed scored the IHC, analysed data, designed and performed

meta-analyses and wrote the paper; J Kutasovic performed the IHC and

analysed data; AC Vargas made the TMAs and performed pathology review;

L Melville, R Chambers and L Da Silva performed pathology review; J

Jayanthan, A Al-Murrani and L Reid performed the IHC; E Evans and A Porter

provided funding, tissue resources and clinical data; D Papadimos provided

tissue resources and clinical data; E Thompson interpreted data; S Lakhani

provided funding, tissue resources and supervised the pathology review; P

Simpson scored the IHC, analysed data, wrote the paper, provided resources

and conceived the study.

12


13


References

1. Thiery JP, Acloque H, Huang RY, et al. Epithelial-mesenchymal transitions in development and disease. Cell 2009; 139: 871-890.

2. Weigelt B, Ng CK, Shen R, et al. Metastatic breast carcinomas display genomic and transcriptomic heterogeneity. Mod Pathol 2015; 28: 340-351.

3. Sarrio D, Rodriguez-Pinilla SM, Hardisson D, et al. Epithelial-mesenchymal transition in breast cancer relates to the basal-like phenotype. Cancer research 2008; 68: 989-997.

4. Taube JH, Herschkowitz JI, Komurov K, et al. Core epithelial-to-mesenchymal transition interactome gene-expression signature is associated with claudin-low and metaplastic breast cancer subtypes. Proc Natl Acad Sci USA 2010; 107: 15449-15454.

5. Hazan RB, Qiao R, Keren R, et al. Cadherin switch in tumor progression. Annals of the New York Academy of Sciences 2004; 1014: 155-163.

6. Rakha EA, Ellis IO. Lobular breast carcinoma and its variants. Seminars in diagnostic pathology 2010; 27: 49-61.

7. Schackmann RC, van Amersfoort M, Haarhuis JH, et al. Cytosolic p120-catenin regulates growth of metastatic lobular carcinoma through Rock1-mediated anoikis resistance. The Journal of clinical investigation 2011; 121: 3176-3188.

8. Micalizzi DS, Farabaugh SM, Ford HL. Epithelial-mesenchymal transition in cancer: parallels between normal development and tumor progression. Journal of mammary gland biology and neoplasia 2010; 15: 117-134.

9. Turashvili G, Bouchal J, Ehrmann J, et al. Novel immunohistochemical markers for the differentiation of lobular and ductal invasive breast carcinomas. Biomedical papers of the Medical Faculty of the University Palacky, Olomouc, Czechoslovakia 2007; 151: 59-64.

10. Yang J, Mani SA, Donaher JL, et al. Twist, a master regulator of morphogenesis, plays an essential role in tumor metastasis. Cell 2004; 117: 927-939.

11. Sarrio D, Moreno-Bueno G, Hardisson D, et al. Epigenetic and genetic alterations of APC and CDH1 genes in lobular breast cancer: relationships with abnormal E-cadherin and catenin expression and microsatellite instability. International journal of cancer Journal international du cancer 2003; 106: 208-215.

12. Tsai JH, Yang J. Epithelial-mesenchymal plasticity in carcinoma metastasis. Genes & development 2013; 27: 2192-2206.

13. Morrogh M, Andrade VP, Giri D, et al. Cadherin-catenin complex dissociation in lobular neoplasia of the breast. Breast cancer research and treatment 2012; 132: 641-652.

14


14. Tarin D, Thompson EW, Newgreen DF. The fallacy of epithelial mesenchymal transition in neoplasia. Cancer research 2005; 65: 5996-6000; discussion 6000-5991.

15. Hollestelle A, Peeters JK, Smid M, et al. Loss of E-cadherin is not a necessity for epithelial to mesenchymal transition in human breast cancer. Breast cancer research and treatment 2013; 138: 47-57.

16. Savagner P. The epithelial-mesenchymal transition (EMT) phenomenon. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO 2010; 21 Suppl 7: vii89-92.

17. Zou D, Yoon HS, Anjomshoaa A, et al. Increased levels of active c-Src distinguish invasive from in situ lobular lesions. Breast cancer research : BCR 2009; 11: R45.

18. Chen A, Beetham H, Black MA, et al. E-cadherin loss alters cytoskeletal organization and adhesion in non-malignant breast cells but is insufficient to induce an epithelial-mesenchymal transition. BMC cancer 2014; 14: 552.

19. Derksen PW, Liu X, Saridin F, et al. Somatic inactivation of E-cadherin and p53 in mice leads to metastatic lobular mammary carcinoma through induction of anoikis resistance and angiogenesis. Cancer cell 2006; 10: 437-449.

20. Cerami E, Gao J, Dogrusoz U, et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer discovery 2012; 2: 401-404.

21. Gao J, Aksoy BA, Dogrusoz U, et al. Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Science signaling 2013; 6: pl1.

22 Lehmann BD, Bauer JA, Chen X, et al. Identification of human triple-negative breast cancer subtypes and preclinical models for selection of targeted therapies. The Journal of clinical investigation 2011; 121: 2750-2767.

23 Prat A, Parker JS, Karginova O, et al. Phenotypic and molecular characterization of the claudin-low intrinsic subtype of breast cancer. Breast cancer research : BCR 2010; 12: R68.

15


Supporting Information

An additional PDF file containing whole section H&Es, and additional stained TMA

cores is provided online.

16


Figure legends:

Figure 1: ILCs rarely express markers associated with EMT; presented are two of the

3/148 mesenchymal marker positive cases. The case shown in A-H is a classic/solid

type, as shown by the Haematoxylin and Eosin core in (A) and is ER and PR

positive, Her2 negative and Ki67 positive. N-cadherin is not expressed (B), in spite of

the loss of E-cadherin (C), so there is not a clear ‘cadherin switch’. There is

detectable expression of Osteonectin (D), nuclear Twist (E), nuclear Snail (F),

Vimentin (G) and SMA (H). The case shown in I-K depicts the only mesenchymal

marker positive case in which there was co-incident LCIS; (I) negative for E-

cadherin; and positive staining for Twist (J) and Vimentin (K), in both the in situ cells

and surrounding targetoid invasive cells. TMA core is approximately 600 µm; scale

bar for whole sections is 500 µm.

Figure 2: CDH1 and EMT marker expression in The Cancer Genome Atlas breast

cancer expression dataset. An RNASeq z score (defined as a value indicating the

number of standard deviations away from the mean of expression in the reference

population (reference population being either all tumors that are diploid for the gene

in question, or, when available, normal adjacent tissue)) for selected markers was

plotted in GraphPad Prism 6. Graphs A-C represent subsets of the TCGA ILC cohort

and depict the relationship (as assessed by linear regression) between the CDH1

gene methylation status (HM450 array data; x axes presented as a proportion of

assayed CpGs that are methylated) and the CDH1 mRNA levels (RNASeq z score);

(A) cohort of tumours WT and diploid for CDH1 (R2=0.3184); (B) the CDH1 WT,

haploid cohort (R2=0.2393); (C) all CDH1 mutant ILCs, with the 3 diploid cases as

open squares and the haploid cases as solid circles (R2=0.04872). Graphs D-I

describe mesenchymal marker expression across a series of breast cohorts; (D) ILC

tumours without a CDH1 mutation or chromosomal loss at that locus; (E) ILCs with a

CDH1 mutation and a chromosomal loss at the CDH1 locus; (F) ILC tumours without

17


a CDH1 mutation but with a chromosomal loss at that locus; (G) ER positive IC-NST

tumours; (H) ER negative IC-NST tumours; and, (I) subset of breast cancers

specifically recorded as the metaplastic histological type (which are reportedly more

likely to have an EMT profile). Note that CDH1 and CDH2 are the genes encoding E-

cadherin and N-cadherin, respectively. WT = wildtype.

18


Table 1: Immunohistochemistry reagents and scoring details.

Marker Clone Supplier Dilution Scoring notes ER 6F11 Novocastra 1/70 1% cut off PR 1A6 Novocastra 1/50 1% cut off HER2 Dako HercepTestTM guidelines CK19 RCK108 Dako 1/40 any positivity CK8/18 5D3 Novocastra 1/100 any positivity CK14 LL02 Novocastra 1/40 10% cut off CK5/6 D5/16B4 Chemicon 1/400 10% cut off EGFR 31G7 Zymed 1/50 any positivity E-Cadherin 36B5 Novocastra 1/50 aberrant or membrane β-catenin 17C2 Novocastra 1/40 aberrant or membrane p120 catenin G10133 Bioscience 1/100 aberrant or membrane TP53 DO7 Dako 1/200 10% cut off Ki-67 MIB-1 Dako 1/100 20% cut off SMA 1A4 Dako 1/700 any positivity Fibronectin 568 Novocastra 1/100 3+ positivity only N-cadherin 3B9 Invitrogen 1/150 membrane positivity Osteonectin 15G12 Novocastra 1/80 any positivity Snail SN9H2 Cell Signalling 1/40 nuclear positivity TWIST 2C1A Abcam 1/50 nuclear positivity Vimentin M0725 Dako 1/400 any positivity

19


Table 2: Pathology overview and EMT phenotype of Invasive Lobular Carcinoma cohort (n=148).

Histopathology Grade Multicentric Multifocal LVI

G1 4.5% 8.1% 21.6% 11.5% G2 78.0% DCIS LNa G3 17.5% 39.2% 87.2%

Clinical diagnostic biomarkers Basal/luminal markers ER PR CK5/6 CK14

95.2% 76.8% 5.6% 8.0% HER2+ Triple negative EGFR Basal-likea 1.4% 1.4% 0% 13.0% P53 Ki67 CK19 CK8/18

15.5% 36.6% 98.6% 100% E-cadherin complex

Membrane Aberrant Negative E-cadherin 1.4% 22.8% 75.8% β-catenin 9.2% 4.9% 85.9%

P120 9.9% 90.0% 0% Mesenchymal markers

Vimentin N-Cadherin SMA Twist 14.2% 3.5% 20.0% 3.8%

Fibronectin Osteonectin Snail 1.1% 28.7% 2.2%

a LN (Lobular neoplasia) is the presence of Atypical Lobular Hyperplasia (ALH) and/or LCIS b Basal-like is defined as any positivity for CK5/6, CK14 or EGFR LVI, Lymphovascular invasion; SMA, smooth muscle actin

20


Figure 1_final.tif


Figure 2.tif


Minerva Access is the Institutional Repository of The University of Melbourne

Author/s:

Reed, AEM; Kutasovic, JR; Vargas, AC; Jayanthan, J; Al-Murrani, A; Reid, LE; Chambers, R;

Da Silva, L; Melville, L; Evans, E; Porter, A; Papadimos, D; Thompson, EW; Lakhani, SR;

Simpson, PT

Title:

An epithelial to mesenchymal transition programme does not usually drive the phenotype of

invasive lobular carcinomas

Date:

2016-03-01

Citation:

Reed, A. E. M., Kutasovic, J. R., Vargas, A. C., Jayanthan, J., Al-Murrani, A., Reid, L. E.,

Chambers, R., Da Silva, L., Melville, L., Evans, E., Porter, A., Papadimos, D., Thompson, E.

W., Lakhani, S. R. & Simpson, P. T. (2016). An epithelial to mesenchymal transition

programme does not usually drive the phenotype of invasive lobular carcinomas. JOURNAL

OF PATHOLOGY, 238 (4), pp.489-494. https://doi.org/10.1002/path.4668.

Persistent Link:

http://hdl.handle.net/11343/290881

File Description:

Accepted version

this is the author manuscript accepted for publication and

Documents