Co-localization of Her2/Neu, ER, PR, Ki67, and Cytokeratin on a Triple Positive Breast Cancer Patient Zhengyu Pang1, Judit Zubovits4 , Kashan Shaikh2, Dan Wang3, Alex Corwin2, Gina Clarke3, Sean Dinn1, Robert Filkins1, and Martin J. Yaffe3

1 Diagnostic and Biomedical Technologies, 2 Electrical Technologies & Systems, GE Global Research Center, Niskayuna, NY

3 Ontario Institute for Cancer Research and Sunnybrook Research Institute, University of Toronto, Toronto, ON

4 Department of Anatomic Pathology, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, ON

The treatment of breast cancer is based on the knowledge of

estrogen (ER), progesterone (PR) and Her2/Neu status of a tumour

(Her2). However, prediction of the effect of treatment based on

these three biomarkers alone is somewhat limited, thus motivating

the inclusion of additional molecular biomarkers. To this end, we

developed and validated a multiplexing technology to allow multiple

biomarkers to be imaged on the same tissue section and then

analyze biomarker expression at the subcellular level (nuclear,

membrane, and cytoplasmic)

Introduction

GE Global Research Center (GRC) has developed proof of concept

instrumentation and the tools for labeling, multiplexing, imaging and

analysis of onco-proteins in fixed tissues. This multiplexing platform

enabled researchers at Sunnybrook Research Institute to study

multiple biomarkers on the same tissue section. Specifically we

aimed to 1): validate multiplexing technology by comparing to

traditional histology; 2) investigate co-localization of 3 established

biomarkers in breast cancer, namely, Her2/Neu, Estrogen receptor

(ER), and progesterone receptor (PR) along with epithelial

biomarker (cytokeratin), proliferation index biomarker (Ki67) and

segmentation biomarker (NaKATPase)

Materials and Methods,

An anonymized sample of triple positive breast cancer was

retrieved from the files of Sunnybrook Health Science Centre

Department of Anatomic Pathology. Multiplexing was performed

using a proprietary microfluidic system (Fig. 2) . Slides were stained

with primary antibodies against Her2/neu and ER, and then

visualized using a secondary antibody directly conjugated with

fluorescent probes Cy3 and Cy5, respectively. After bleaching the

sample, antibodies against PR and Ki67 and directly conjugated

with Cy3 and Cy5 were subsequently used; and followed by

cytokeratin and NaKATPase direct conjugates (Fig. 1).

Results and Discussion

Immunofluorescence images confirm the positivity of the

ER/PR/Her2 as diagnosed by traditional DAB staining and

confirmed by clinical pathologist at Sunnybrook (JZ). This validated

the multiplex platform for the use in the clinical pathology lab. Next

the co-localization of Her2/Neu, ER, PR along with Ki67,

cytokeratin at specific subcellular compartment was investigated at

the single cell level. As expected, ER, PR and Ki67 expression was

localized to the nuclei, whereas cytokeratin and her2 staining was

membranous (Fig.3).

The matrix plot of these 5 biomarkers is shown in Fig. 4. The

correlation between PR and ER is very significant (R=0.77),

although there are some cells with low ER, but high PR. This

is consistent with the well-established correlation between ER

and PR expression. Correlations between Ki67 and ER/PR

are interesting. Cells cluster into two groups along the marker

axes. Cells with low Ki67 expression have high ER or PR

expression, and vice versa, suggesting that cells in the

proliferative state (high Ki67 expression) tend to lose the

ER/PR expression. There is a much less correlation between

Ki67 and Her2, as the membranous biomarker expression is

probably less dependent on the cell cycle.

R=0.158 R=-0.015 R=0.178

R=0.770 R=-0.083

R=-0.051

R=0.012

R=-0.002

R=0.079

R=0.277

Images of background fluorescence were acquired before and after

the staining. Images were registered, background fluorescence

removed, segmented and analyzed at the single cell level. Biomarker

expression was quantified for subcellular compartments such as

nucleus, cytoplasm and membrane.

Taken together, multiplexing technology should provide an

important tool in research to understand molecular

mechanisms in cancers on an individual patient basis with

the potential role of guiding personalized therapy.

Acknowledgement

Authors would like to thank Dr. Alberto Santamaria-Pang for

single cell analysis, Drs. Colin McCulloch and Yunxia Sui for

correlation study using R. This work is supported by

Molecular Imaging and Diagnostic Advanced Technology at

GE Global Research Center and a grant from Ontario

Institute for Cancer Research.

Her2 ER PR

Fig. 1. Multiplexing of Her2, ER, and PR

Fig. 2. Integrated platform for multiplexing and imaging. Right is the microfluidic insert

Fig. 3. Overlay images (left) and single cell segmentation (Right)

Fig. 4. Matrix plot of Her2, ER. PR, Ki67 and cytokeraitn

Her2 ER/PR Ki67 Memb. Cytoplasmic Nuclear

Mem.Her2

Nuc.ER

Nuc.Ki67

Cyto.Cytokeratin

Nuc.PR