Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2021 Mar 20;13(6):1429.
doi: 10.3390/cancers13061429.

Identifying Molecular Signatures of Distinct Modes of Collective Migration in Response to the Microenvironment Using Three-Dimensional Breast Cancer Models

Diana Catalina Ardila  1 , Vaishali Aggarwal  1 , Manjulata Singh  1 , Ansuman Chattopadhyay  2 , Srilakshmi Chaparala  2 , Shilpa Sant  1   3   4   5
Affiliations

Identifying Molecular Signatures of Distinct Modes of Collective Migration in Response to the Microenvironment Using Three-Dimensional Breast Cancer Models

Diana Catalina Ardila et al. Cancers (Basel). .

Abstract

Collective cell migration is a key feature of transition of ductal carcinoma in situ (DCIS) to invasive ductal carcinoma (IDC) among many other cancers, yet the microenvironmental factors and underlying mechanisms that trigger collective migration remain poorly understood. Here, we investigated two microenvironmental factors, tumor-intrinsic hypoxia and tumor-secreted factors (secretome), as triggers of collective migration using three-dimensional (3D) discrete-sized microtumor models that recapitulate hallmarks of DCIS-IDC transition. Interestingly, the two factors induced two distinct modes of collective migration: directional and radial migration in the 3D microtumors generated from the same breast cancer cell line model, T47D. Without external stimulus, large (600 µm) T47D microtumors exhibited tumor-intrinsic hypoxia and directional migration, while small (150 µm), non-hypoxic microtumors exhibited radial migration only when exposed to the secretome of large microtumors. To investigate the mechanisms underlying hypoxia- and secretome-induced directional vs. radial migration modes, we performed differential gene expression analysis of hypoxia- and secretome-induced migratory microtumors compared with non-hypoxic, non-migratory small microtumors as controls. We propose unique gene signature sets related to tumor-intrinsic hypoxia, hypoxia-induced epithelial-mesenchymal transition (EMT), as well as hypoxia-induced directional migration and secretome-induced radial migration. Gene Set Enrichment Analysis (GSEA) and protein-protein interaction (PPI) network analysis revealed enrichment and potential interaction between hypoxia, EMT, and migration gene signatures for the hypoxia-induced directional migration. In contrast, hypoxia and EMT were not enriched in the secretome-induced radial migration, suggesting that complete EMT may not be required for radial migration. Survival analysis identified unique genes associated with low survival rate and poor prognosis in TCGA-breast invasive carcinoma dataset from our tumor-intrinsic hypoxia gene signature (CXCR4, FOXO3, LDH, NDRG1), hypoxia-induced EMT gene signature (EFEMP2, MGP), and directional migration gene signature (MAP3K3, PI3K3R3). NOS3 was common between hypoxia and migration gene signature. Survival analysis from secretome-induced radial migration identified ATM, KCNMA1 (hypoxia gene signature), and KLF4, IFITM1, EFNA1, TGFBR1 (migration gene signature) to be associated with poor survival rate. In conclusion, our unique 3D cultures with controlled microenvironments respond to different microenvironmental factors, tumor-intrinsic hypoxia, and secretome by adopting distinct collective migration modes and their gene expression analysis highlights the phenotypic heterogeneity and plasticity of epithelial cancer cells.

Keywords: bioinformatic analysis; collective migration; epithelial-mesenchymal transition (EMT); microarray; microtumors; three-dimensional cultures; tumor microenvironment; tumor-intrinsic hypoxia.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Figures

Scheme 1
Scheme 1
Method workflow.
Figure 1
Figure 1
3D models of breast cancer recapitulate two distinct modes of collective migration due to different microenvironmental factors: tumor intrinsic hypoxia and secretome. (A) Schematic representation of our 3D hydrogel microwell system in which we can generate small (150 µm) and large (600 µm) microtumors. (B) Small microtumors are non-hypoxic and non-migratory, whereas large microtumors develop size-related intrinsic hypoxia and migrate directionally. (C) We can induce radial migration by treating the small non-hypoxic non-migratory tumors with the secretome of large hypoxic directional migrating tumors. (D) We performed bioinformatic analysis of microarray data to study the changes in gene expression of large hypoxic directional migrating microtumors and small non-hypoxic radially migrating microtumors compared to non-hypoxic and non-migratory microtumors.
Figure 2
Figure 2
Global changes in gene expression caused by tumor-intrinsic hypoxia and secretome. (A) Volcano plots showing the significantly upregulated (red) and downregulated (green) genes in tumor intrinsic hypoxia-induced directional migration (left) and secretome-induced radial migration (right) using a False Discovery Rate (FDR) ≤ 0.05, p-value ≤ 0.05, and fold change ± 2. (B) GSEA analysis showing upregulated (red) and downregulated (green) hallmarks for B(i) hypoxia-induced directional migration and B(ii) secretome-induced radial migration using minimum size cut off = 5. (C) Gene ontology (GO) analysis showing oppositely regulated biological processes: C(i) upregulated (red) in tumor-intrinsic hypoxia-induced directional migration, and downregulated (green) in secretome-induced radial migration and C(ii) downregulated (green) in tumor-intrinsic hypoxia-induced directional migration and upregulated (red) in secretome-induced radial migration. (D) Ingenuity Pathway Analysis (IPA) enriched and differentially regulated due to tumor-intrinsic hypoxia and secretome. Upregulated pathways are shown in red and downregulated in green. The differentially regulated canonical pathways were identified with a threshold –log10 (p-value) of >1.3.
Figure 3
Figure 3
Hypoxia is highly enriched in large migratory tumors as compared to small migratory tumors. (A) Venn diagram of correlation engine analysis showing the 33 common genes between tumor intrinsic hypoxia-induced directional migration and ‘response to hypoxia’ gene ontology (GO) term, as well as the 8 common genes between secretome-induced radial migration and ‘response to hypoxia’ GO term. (B) Significance of overlap (p-value and % of overlap) of differentially expressed genes (DEGs) in tumor intrinsic hypoxia-induced directional migration and secretome-induced radial migration with ‘response to hypoxia’ GO term. (C) GSEA enrichment plot of DEGs showing 55 genes enriched in tumor intrinsic hypoxia-induced directional migration out of 87 in ‘Hallmark hypoxia’ and a Normalized Enrichment Score (NES) of 4.19. (D) Correlation Engine meta-analysis, comparing the significance of overlap with the GO term ‘response to hypoxia’ of hypoxia-induced directional migration, secretome-induced radial migration, and seven publicly available genomic studies (GSE3893, GSE19123, GSE29406, GSE70805, GSE47533, GSE9649, GSE30019) that examined the gene expression changes in breast cancer tissue or cell lines in response to hypoxic conditions (details in Supplementary Table S1). (E) Heat map of tumor-intrinsic hypoxia 79-gene signature set that compiles genes in ‘response to hypoxia’ GO term, and ’hallmark hypoxia’ common to the DEGs in the directional migration. (F) Heat map of radial migration hypoxia 8-gene signature set that compiles genes in ’response to hypoxia’ GO term common to the DEGs in the radial migration. (G) Venn diagram displaying the 3 common genes between directional migration tumor intrinsic hypoxia 79-gene signature set and radial migration hypoxia 8-gene signature set. (H) Hypoxic stain with Ru-dpp of H(i) hypoxic directionally migrating tumors, and H(ii) secretome-treated non-hypoxic radially migratory tumors. Yellow dashed circular outline delineates the microwell. White arrows indicate the direction of migration.
Figure 4
Figure 4
Epithelial to mesenchymal transition is enriched in hypoxia-induced directional migration compared to secretome-induced radial migration. (A) GSEA enrichment plot of DEGs showing 26 genes enriched in tumor-intrinsic hypoxia-induced directional migration out of 200 in ’Hallmark Epithelial Mesenchymal Transition’ and a Normalized Enrichment Score (NES) of 2.32. (B) Correlation engine meta-analysis, comparing the significance of overlap with the GO term ’Epithelial Mesenchymal Transition’ of hypoxia-induced directional migration, and seven publicly available genomic studies (GSE3893, GSE19123, GSE29406, GSE70805, GSE47533, GSE9649, GSE30019). (C) Heat map of tumor-intrinsic hypoxia 26-gene signature set that compiles genes in ‘Hallmark Epithelial Mesenchymal Transition’ common to the DEGs in the directional migration.
Figure 5
Figure 5
Migration is equally enriched in both large and small migratory tumors displaying different patterns of migration. (A) Venn diagram of correlation engine analysis showing the 40 common genes between tumor intrinsic hypoxia-induced directional migration and ‘regulation of cell migration’ gene ontology (GO) term, as well as the 14 common genes between secretome-induced radial migration and ‘regulation of cell migration’ GO term. (B) Significance of overlap (p-value and % of overlap) of differentially expressed genes (DEGs) in hypoxia-induced directional migration and secretome-induced radial migration with ‘regulation of cell migration’ GO term. (C) GSEA enrichment plot of DEGs in GO ‘tissue migration’ showing C (i) 42 genes enriched in hypoxia-induced directional migration out of 374 in ‘GO tissue migration’ and a Normalized Enrichment Score (NES) of 1.724, and C (ii) 10 genes enriched in secretome-induced radial migration out of 374 in GO ‘tissue migration’ and a Normalized Enrichment Score (NES) of 1.721 (D) Correlation engine meta-analysis, comparing the significance of overlap with the GO term ‘regulation of cell migration’ of hypoxia-induced directional migration, secretome-induced redial migration, and seven publicly available genomic studies(GSE3893, GSE19123, GSE29406, GSE70805, GSE47533, GSE9649, GSE30019). (E) Heat map of 69-gene directional migration signature set that compiles genes in ‘regulation of cell migration’ GO term, and ‘GO tissue migration’ that are common to the DEGs in hypoxia-induced directional migration. (F) Heat map of 21-gene radial migration signature set that compiles genes in ‘regulation of cell migration’ GO term, and ‘GO tissue migration’ that are common to the DEGs in in the secretome-induced radial migration. (G) Venn diagram displaying the 6 common genes between the 69-gene directional migration signature set and the 21-gene radial migration signature set.
Figure 6
Figure 6
The drivers of directional vs. radial migration are different, have different PPI, and yet participate in the same migration-related signaling pathways: (A) Venn diagram showing the intersection of tumor-intrinsic hypoxia, EMT and directional migration signature sets. There are 8 genes common only between the hypoxia and migration signature gene sets, 4 genes common only between hypoxia and EMT, 4 genes common only between migration and EMT, and 4 common genes shared by hypoxia, migration and EMT. The heat map displays the expression of the 4 common genes in the intersection between hypoxia, migration, and EMT in the directional migration compared with non-hypoxic, non-migratory tumors. (B) Zero-order protein-protein interaction (PPI) network combining the hypoxia, migration, and EMT signature gene sets from hypoxia-induced directional migration. The zero-order network consists of seed genes that have direct interaction without intermediate nodes. (C) KEGG pathway analysis on the minimum interaction PPI network for hypoxia induced-directional migration. Y-axis shows the enrichment p-value. The numbers above the bar plot represent the number of hits for each pathway. (D) Venn diagram showing the intersection for secretome-induced hypoxia and radial migration signature sets. There are two common genes between the hypoxia and migration signature gene sets. The heat map displays the expression of these two common genes in radial migration compared with non-hypoxic, non-migratory tumors. (E) Minimum protein-protein interaction (PPI) network combining the hypoxia and migration signature gene sets from secretome-induced radial migration. The signature gene sets in radial migration do not have a zero-order network. Instead, the minimum network has intermediate nodes necessary to connect the seed genes. (F) KEGG pathway analysis on the minimum interaction PPI network for secretome induced-radial migration. Y-axis shows the enrichment p-value. The numbers above the bar plot represent the number of hits for each pathway.
Figure 7
Figure 7
Modulators of migration in large and small migratory tumors are associated with low survival rate and prognosis: Kaplan survival plots and histological images from Human Protein Atlas of ductal carcinoma and normal breast tissue of significant genes in (A) the hypoxia, EMT, and migration signature sets from directional migration, and (B) hypoxia and migration signature sets from radial migration. Kaplan plots show high expression in red and low expression in green.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Ernster V.L., Ballard-Barbash R., Barlow W.E., Zheng Y., Weaver D.L., Cutter G., Yankaskas B.C., Rosenberg R., Carney P.A., Kerlikowske K. Detection of ductal carcinoma in situ in women undergoing screening mammography. J. Natl. Cancer Inst. 2002;20:1546–1554. doi: 10.1093/jnci/94.20.1546. - DOI - PubMed
    1. Schnitt S.J. The transition from ductal carcinoma in situ to invasive breast cancer: The other side of the coin. Breast Cancer Res. 2009;11:101. doi: 10.1186/bcr2228. - DOI - PMC - PubMed
    1. Hernandez L., Wilkerson P.M., Lambros M.B., Campion-Flora A., Rodrigues D.N., Gauthier A., Cabral C., Pawar V., Mackay A., A’Hern R. Genomic and mutational profiling of ductal carcinomas in situ and matched adjacent invasive breast cancers reveals intra-tumour genetic heterogeneity and clonal selection. J. Pathol. 2012;227:42–52. doi: 10.1002/path.3990. - DOI - PMC - PubMed
    1. Yeong J., Thike A.A., Tan P.H., Iqbal J. Identifying progression predictors of breast ductal carcinoma in situ. J. Clin. Pathol. 2017;70:102–108. doi: 10.1136/jclinpath-2016-204154. - DOI - PubMed
    1. Semenza G.L. The hypoxic tumor microenvironment: A driving force for breast cancer progression. Cell Res. 2016;1863:382–391. doi: 10.1016/j.bbamcr.2015.05.036. - DOI - PMC - PubMed

LinkOut - more resources

-