Subtype-specific characterization of breast cancer invasion using a microfluidic tumor platform

Understanding progression of breast cancers to invasive ductal carcinoma (IDC) can significantly improve breast cancer treatments. However, it is still difficult to identify genetic signatures and the role of tumor microenvironment to distinguish pathological stages of pre-invasive lesion and IDC. Presence of multiple subtypes of breast cancers makes the assessment more challenging. In this study, an in-vitro microfluidic assay was developed to quantitatively assess the subtype-specific invasion potential of breast cancers. The developed assay is a microfluidic platform in which a ductal structure of epithelial cancer cells is surrounded with a three-dimensional (3D) collagen matrix. In the developed platform, two triple negative cancer subtypes (MDA-MB-231 and SUM-159PT) invaded into the surrounding matrix but the luminal A subtype, MCF-7, did not. Among invasive subtypes, SUM-159PT cells showed significantly higher invasion and degradation of the surrounding matrix than MDA-MB-231. Interestingly, the cells cultured on the platform expressed higher levels of CD24 than in their conventional 2D cultures. This microfluidic platform may be a useful tool to characterize and predict invasive potential of breast cancer subtypes or patient-derived cells.

In this paper the authors report a 3D in vitro model of invasive ductal carcinoma using a microfluidic platform. In particular they demonstrate that invasion characteristics of three different human breast cancer cell lines is subtype-specific. The experiments are well performed and the results appear to support the proposed approach. It would be great if the authors could add some experiments with various oxygen concentrations and discussion on effects of hypoxia on invasion characteristics

Response:
We appreciate the reviewer's thoughtful suggestion. We agree that the effect of hypoxia on invasion characteristics can be investigated by using our IDC platform. However, the objective of this study is to compare the invasion characteristics of breast cancer subtypes. We add a discussion about potential use of the developed platform on hypoxia study. (Muñoz-Nájar, Neurath et al. 2006, Chen, Imanaka et al. 2010, Semenza 2012 Revision: In page 14: Throughout the invasion process, distinct biophysical and biochemical features in the tumor microenvironment (TME) such as hypoxia, the existence of fibroblast, various growth factors, and cytokines play critical roles in regulating the cell response. Specifically, the hypoxia has been reported to enhance cancer cell invasion through hypoxia-induced factor (HIF) activities, which regulate the transcription factors such as Snail, Twist, and matrix metalloproteinase. [50][51][52] Hypoxia induced Notch signaling was reported to mediate epithelial-mesenchymal transition in breast cancer through enhanced expression of Slug and Snail with e-cadherin suppression.
[50] With a benefit of the IDC-on-chip platform to recapitulate the physiological architecture of cancer invasion, the platform can be further improved by applying the complex stroma conditions.

Reviewer #2's Comments
This manuscript presents quantitative comparison of invasion process in 3D ductal carcinoma in situ (DCIS) cultures in engineered microfluidic platforms. The authors compared 2D and 3D culture of breast cancer cell lines of MDA-MB-231, SUM-159PT and MCF-7 and proved that their platform could have the invasion prediction potential of breast cancer subtypes. If the platform would realize the closer in-vivo-like microenvironments with co-culture and mechanical stress, then the platform could be a useful tool to characterize and predict invasive potential of breast cancer subtypes or patient-derived cells. The manuscript would seem of considerable interest to those working in cancer cell research in engineered microfluidic platforms. After polished based on the following critiques, this manuscript may be able to be published in PLOS ONE. I would recommend that this paper needs minor revision to be published in PLOS ONE. I recommend the current manuscript should revise to include answers for the questions below:

Comment:
There are several human breast cancer cell lines. Why did authors choose MDA-MB-231, SUM-159PT and MCF-7? In particular, MCF-7 instead of MCF 10? Were the cell lines study in 3D cultures in previous studies? Do the results show the similar to results in the previous work or to in vivo behaviors?

Response:
We selected the three cell lines based on our primary objective to study invasion potential of breast cancer subtypes from mammary duct configuration. MCF-7, MDA-MB-231, and SUM-159PT are subtypes of breast caner cell lines, specifically invasive ductal carcinoma (IDC) which is proceeded after the ductal carcinoma in situ (DCIS). DCIS model using MCF-10 has been reported in literatures. (Bischel, Beebe et al. 2015, Choi, Hyun et al. 2015) Since the developed model intends to mimic IDC, instead of MCF-10, MCF-7 was used as the Luminal A subtype IDC expressing estrogen receptor (ER) and progesterone receptor (PR). The results showed that the MCF-7 barely invaded in the IDC-on chip showing epithelial characteristics similar to MCF-10. Clarification of this point was added in the revision.

Revision:
In page 4: Since the developed model intends to mimic IDC which is a later stage of the disease after DCIS, the three subtypes of breast cancer cell lines were selected accordingly.