Epithelial cells line the inside of blood vessels, intestines and other organs throughout the body. Any epithelial cells that become detached from their natural surroundings die by a process called anoikis (a Greek word meaning “being without a home”). This process has an important role in preventing cancer from spreading around the body because it eliminates cells that are not in their proper environment. However, some cancers that start from epithelial cells, such as breast cancer, develop resistance to anoikis. Gaining a better understanding of the cellular factors that regulate anoikis, and how resistance develops, may reveal new drug targets for the treatment of breast cancer.

Previous studies found proteins called BIM and BMF promote anoikis by inducing cell suicide. However, it is possible that other factors can also promote this process in different ways. Pedanou et al. performed a large-scale genetic screen in human breast epithelial cells and identified several new factors that promote anoikis.

Inside our cells, DNA is packaged around proteins called histones, which can influence whether a gene is switched on or off. One of the factors Pedanou et al. identified is a protein called KDM3A that can remove small chemical groups (known as methyl groups) from histones – a process that is known to switch on genes. Further experiments show that epithelial cells in their natural surroundings only produce low levels of KDM3A, but that the levels of this protein increase if these cells become detached. This promotes anoikis by activating two genes called BNIP3 and BNIP3L that induce cell suicide. However, KDM3A levels are low in human breast cancers, which suggests that these cancers become resistant to anoikis by preventing increases in KDM3A production.

Using a mouse model of breast cancer, Pedanou et al. found that switching off KDM3A in cancer cells increases their ability to move around the body. Collectively, these findings reveal a new mechanism that triggers anoikis in normal breast epithelial cells and is disabled during breast cancer development. Future challenges are to identify factors that directly regulate the production of KDM3A, and to understand how these factors are manipulated in breast cancer cells to cause anoikis resistance.

Epithelial cells that lose attachment to the extracellular matrix (ECM), or attach to an inappropriate ECM, undergo a specialized form of apoptosis called anoikis. Anoikis has an important role in preventing oncogenesis, particularly metastasis, by eliminating cells that lack proper ECM cues (Simpson et al., 2008; Zhu et al., 2001). Anoikis also functions to prevent the invasion of tumor cells into the luminal space, which is a hallmark of epithelial tumors (Debnath et al., 2002). In general, epithelial-derived cancers, such as breast cancer, develop resistance to anoikis (reviewed in Schwartz, 1997). Several signaling pathways have been shown to regulate anoikis (reviewed in Paoli et al., 2013). In particular, anoikis is suppressed by integrin signaling, which functions through focal adhesion kinase (FAK), an activator of the RAF/MEK/ERK pathway (King et al., 1997). FAK signaling is active in attached cells and is inactive following detachment (Frisch et al., 1996). Anoikis is also suppressed by integrin-mediated, ligand independent activation of the epidermal growth factor receptor (EGFR) signaling pathway (Moro et al., 1998), which, like FAK, also stimulates RAF/MEK/ERK activity.

These cell signaling pathways have been found to regulate the levels of BIM (also called BCL2L11) and BMF, two pro-apoptotic members of the BCL2 family of apoptosis regulators previously shown to contribute to anoikis (Reginato et al., 2003; Schmelzle et al., 2007). However, depletion of BIM or BMF diminishes but does not completely prevent anoikis (Reginato et al., 2003; Schmelzle et al., 2007), suggesting the existence of other factors and regulatory pathways that can promote anoikis. Moreover, the basis of anoikis resistance remains to be determined and to date has not been linked to alterations in expression or activity of BIM or BMF.

To investigate the possibility that there are additional factors and regulatory pathways that promote anoikis, we performed a large-scale RNA interference (RNAi) screen for genes whose loss of expression confer anoikis resistance. The screen was performed in MCF10A cells, an immortalized but non-transformed human breast epithelial cell line that has been frequently used to study anoikis (see, for example, Huang et al., 2010; Reginato et al., 2003; Schmelzle et al., 2007; Taube et al., 2006). A genome-wide human small hairpin RNA (shRNA) library comprising ~62,400 shRNAs directed against ~28,000 genes (Silva et al., 2003; Silva et al., 2005) was divided into 10 pools, which were packaged into retroviral particles and used to stably transduce MCF10A cells. Following selection, the cells were divided into two populations, one of which was plated on poly-2-hydroxyethylmethacrylate (HEMA)-coated plates for 10 days to inhibit cell attachment to matrix, and another that was cultured attached to matrix for 10 days as a control (Figure 1A). Surviving cells were selected and shRNAs identified by deep sequencing. Bioinformatic analysis of the two populations identified 26 shRNAs whose abundance was significantly enriched >500-fold following detachment (Figure 1—source data 1); such shRNAs presumably confer upon MCF10A cells a selective advantage by protecting them from undergoing anoikis.

Figure 1 with 4 supplements see all

Download asset Open asset

Figure 1—source data 1

List of 26 shRNAs, and the target genes, whose abundance was significantly enriched >500-fold following detachment of MCF10A cells.

https://doi.org/10.7554/eLife.16844.004

Download elife-16844-fig1-data1-v2.docx

Figure 1—source data 2

Source data for Figure 1B.

https://doi.org/10.7554/eLife.16844.005

Download elife-16844-fig1-data2-v2.xlsx

To validate candidates isolated from the primary screen, we selected the top 20 most highly enriched shRNAs and analyzed them in an independent assay for their ability to confer resistance to anoikis. Briefly, MCF10A cells were transduced with a single shRNA, detached from matrix for 96 hr, and analysed for cell death by annexin V staining. As expected, knockdown of BIM, a positive control, decreased cell death following detachment compared to the control non-silencing (NS) shRNA (Figure 1B and Figure 1—figure supplement 1). Of the 20 candidate shRNAs tested, five reduced the level of detachment-induced apoptosis compared to the NS shRNA, indicating they conferred anoikis resistance (Figure 1B and Figure 1—figure supplement 1). Similar results were obtained using a second, unrelated shRNA directed against the same target gene (Figure 1—figure supplement 2). Quantitative RT-PCR (qRT-PCR) confirmed in all cases that expression of the target gene was decreased in the knockdown cell line (Figure 1—figure supplement 3).

One of the top scoring validated candidates was KDM3A (Figure 1—source data 1), a histone demethylase that specifically demethylates mono-methylated (me1) and di-methylated (me2) histone H3 lysine 9 (H3K9) (Yamane et al., 2006). H3K9 methylation is a transcriptional repressive mark, and the identification of KDM3A raised the intriguing possibility that induction of anoikis involves transcriptional activation of specific genes through H3K9me1/2 demethylation. Therefore, our subsequent experiments focused on investigating the role of KDM3A in anoikis.

We asked whether ectopic expression of KDM3A was sufficient to promote cell death in attached cells. MCF10A cells were transduced with a retrovirus expressing wild-type KDM3A, a catalytically inactive KDM3A mutant [KDM3A(H1120G/D1122N)] (Beyer et al., 2008) or, as a control, empty vector (Figure 1—figure supplement 4), and then treated with puromycin for 10 days at which time viability was assessed by crystal violet staining. The results of Figure 1C show that ectopic expression of wild-type KDM3A but not KDM3A(H1120G/D1122N) greatly reduced MCF10A cell viability. Collectively, the results of Figure 1 demonstrate that KDM3A is necessary and sufficient for efficient induction of anoikis in breast epithelial cells.

We next examined the relationship between KDM3A expression and induction of anoikis. The immunoblot of Figure 2A shows that KDM3A protein levels were very low in attached MCF10A cells, but robustly increased in a time-dependent manner following detachment. The qRT-PCR analysis of Figure 2B shows that an increase in KDM3A expression following detachment was also detected at the mRNA level.

Figure 2 with 1 supplement see all

Download asset Open asset

Figure 2—source data 1

Source data for Figure 2B.

https://doi.org/10.7554/eLife.16844.011

Download elife-16844-fig2-data1-v2.xlsx

We next sought to understand the basis for the increase in KDM3A levels following detachment. As mentioned above, anoikis is suppressed by integrin signaling, which functions through FAK, a regulator of the RAF/MEK/ERK pathway (Frisch et al., 1996; King et al., 1997). Detachment causes a disruption in integrin–ECM contacts, resulting in a loss of FAK signaling in the detached cells (Frisch and Francis, 1994; Frisch et al., 1996), which we observed have elevated KDM3A levels (see Figures 2A and B). We therefore tested whether restoration of integrin signaling in detached cells would block the increase in KDM3A levels. The results of Figure 2C show that the addition of Matrigel basement membrane-like matrix, which restores integrin signaling, to detached cells markedly blocked the elevated levels of the BIM isoform BIM_EL, as expected, and KDM3A. Treatment of MCF10A cells with a FAK inhibitor increased the levels of KDM3A protein (Figure 2D) and mRNA (Figure 2—figure supplement 1A). Thus, the increase in KDM3A levels upon detachment of MCF10A cells is due, at least in part, to the loss of integrin/FAK signaling.

We next analyzed the relationship between the EGFR signaling pathway and KDM3A levels. In the first set of experiments, we ectopically expressed either EGFR or a constitutively active MEK mutant, MEK2(S222D/S226D) (MEK2DD) (Voisin et al., 2008), both of which have been previously shown to block anoikis in detached cells (Reginato et al., 2003). Consistent with these previous results, Figure 2E shows that in detached MCF10A cells, expression of either EGFR or MEK2DD substantially decreased the level of BIM_EL (Reginato et al., 2003). Expression of either EGFR or MEK2DD also decreased the levels of KDM3A in detached MCF10A cells. Conversely, KDM3A protein levels were increased in attached MCF10A cells treated with the EGFR inhibitor gefitinib (Barker et al., 2001; Ward et al., 1994) (Figure 2F) or the MEK inhibitor U0126 (Favata et al., 1998) (Figure 2G). Both gefitinib and U0126 treatment also resulted in increased KDM3A mRNA levels (Figure 2—figure supplement 1B,C).

The results described above suggest a model in which following detachment, the resulting increase in KDM3A demethylates H3K9me1/2 to stimulate expression of one or more pro-apoptotic genes. To test this model and identify pro-apoptotic KDM3A target genes, we took a candidate-based approach and analyzed expression of a panel of genes encoding pro-apoptotic BCL2 proteins (Boyd et al., 1994; Lomonosova and Chinnadurai, 2008; Matsushima et al., 1998) in attached MCF10A cells and detached cells expressing a NS or KDM3A shRNA. We sought to identify genes whose expression increased following detachment in control but not in KDM3A knockdown cells. We found that expression of the vast majority of genes encoding pro-apoptotic BCL2 proteins were unaffected by detachment in MCF10A cells (Figure 3A and Figure 3—figure supplement 1). Consistent with previous results (Reginato et al., 2003; Schmelzle et al., 2007), expression of BIM and BMF were increased upon detachment. However, knockdown of KDM3A did not decrease expression of either BIM or BMF. By contrast, following detachment, expression of BNIP3 and BNIP3L increased, and were the only genes whose expression was diminished more than 2-fold by KDM3A knockdown (Figure 3A and Figure 3—figure supplement 1). We therefore performed a series of experiments to determine whether BNIP3 and BNIP3L are critical KDM3A target genes that mediate anoikis.

Figure 3 with 6 supplements see all

Download asset Open asset

Figure 3—source data 1

Source data for Figure 3A, C, D and E.

https://doi.org/10.7554/eLife.16844.014

Download elife-16844-fig3-data1-v2.xlsx

In the first set of experiments we analyzed BNIP3 and BNIP3L protein levels during anoikis induction. The immunoblot of Figure 3B shows that BNIP3 and BNIP3L levels were very low in attached cells and substantially increased following detachment, with a time course similar to that of detachment-induced KDM3A expression (see Figure 2A). The chromatin immunoprecipitation (ChIP) experiment of Figure 3C shows that KDM3A was bound to the BNIP3 and BNIP3L promoters in detached but not attached cells. Moreover, the levels of H3K9me2 (Figure 3D) and H3K9me1 (Figure 3—figure supplement 2) on the BNIP3 and BNIP3L promoters were greatly diminished following detachment, which was counteracted by knockdown of KDM3A. Conversely, overexpression of KDM3A but not KDM3A(H1120G/D1122N) in attached MCF10A cells resulted in decreased levels of H3K9me1 and H3K9me2 on the BNIP3 and BNIP3L promoters and increased expression of BNIP3 and BNIP3L (Figure 3—figure supplement 3). Finally, knockdown of BNIP3 or BNIP3L (Figure 3—figure supplement 4) resulted in decreased apoptosis following detachment (Figure 3E and Figure 3—figure supplement 5). To further establish the pro-apoptotic role of BNIP3 and BNIP3L in MCF10A cells, we ectopically expressed BNIP3, BNIP3L or both in attached cells (Figure 3—figure supplement 6). Figure 3F shows that moderate cell death was observed upon ectopic expression of either BNIP3 or BNIP3L, but substantial cell death occurred in cells ectopically expressing both BNIP3 and BNIP3L. Collectively, these results establish BNIP3 and BNIP3L as critical KDM3A target genes that mediate anoikis (Figure 3G).

We considered the possibility that decreased KDM3A expression may contribute to anoikis resistance in breast cancer cells and performed a series of experiments to test this idea. We first analyzed a panel of human breast cancer cell lines (BT549, MDA-MB-231, MCF7, SUM149 and T47D) comparing, as a control, anoikis-sensitive MCF10A cells. As expected, detachment-induced apoptosis was significantly diminished in breast cancer cell lines compared to MCF10A cells, indicative of anoikis resistance (Figure 4A and Figure 4—figure supplement 1). Moreover, following detachment of the breast cancer cell lines, induction of KDM3A at both the protein (Figure 4B) and mRNA (Figure 4C) levels was much lower than that observed in MCF10A cells. However, ectopic expression of KDM3A was sufficient to induce apoptosis in each of the five breast cancer cell lines (Figure 4D). Collectively, these results indicate that anoikis-resistance of human breast cancer cells is due, at least in part, to inefficient induction of KDM3A following detachment.

Figure 4 with 9 supplements see all

Download asset Open asset

Figure 4—source data 1

Source data for Figure 4A, C, E, F, H and I.

https://doi.org/10.7554/eLife.16844.022

Download elife-16844-fig4-data1-v2.xlsx

We next analyzed KDM3A expression in human breast cancer patient samples. Interrogation of the Oncomine database (Rhodes et al., 2007) revealed decreased expression levels of KDM3A in several breast cancer datasets (Figure 4—figure supplement 2). To confirm these in silico results, we analyzed KDM3A expression by qRT-PCR in a series of human breast cancer patient samples. The results of Figure 4E show that compared to normal breast epithelium KDM3A expression was significantly decreased in a high percentage of breast cancers. Likewise, basal KDM3A expression levels were also diminished in most human breast cancer cell lines analyzed (Figure 4—figure supplement 3).

Finally, we performed a series of experiments to determine whether KDM3A affects metastatic potential. We first asked whether depletion of KDM3A would promote anoikis resistance in vivo using a mouse pulmonary survival assay. Briefly, immortalized but non-transformed mouse mammary epithelial CLS1 cells were stably transduced with an NS or Kdm3a shRNA (Figure 4—figure supplement 4) and injected into the tail vein of syngeneic mice. After 2 weeks, the lungs were harvested, dissociated into single cell suspensions, and plated in media containing puromycin to select for cells expressing the shRNA. The surviving colonies were visualized by crystal violet staining and quantified. The results of Figure 4F show that Kdm3a knockdown significantly increased the number of cells that survived in the mouse lung relative to the control NS shRNA.

In a second set of experiments, we used a well-characterized mouse breast cancer carcinoma progression series comprising isogenic cell lines with increasing metastatic potential: (1) non-invasive and non-metastatic 67NR cells, which form primary tumors, (2) invasive and non-metastatic 4T07 cells, which enter the circulation but fail to establish secondary tumors, and (3) highly metastatic 4T1 cells, which disseminate widely and colonize distant organ sites (Aslakson and Miller, 1992). qRT-PCR analysis revealed decreased Kdm3a expression in cell lines with greater metastatic potential (Figure 4—figure supplement 5). We expressed either a control NS shRNA or a Kdm3a shRNA in 67NR cells containing a luciferase reporter gene (Figure 4—figure supplement 6). Cells were injected into the tail veins of three syngeneic mice and pulmonary metastases were visualized by live animal imaging after 5 weeks. The results of Figure 4G show, as expected, that control 67NR cells failed to form pulmonary metastases in any of the three mice analyzed. By contrast, Kdm3a knockdown 67NR cells formed substantial pulmonary metastases in all three mice.

Finally, in a more stringent metastasis experiment, control and Kdm3a knockdown 4T07 cells (Figure 4—figure supplement 7), a non-metastatic mouse breast cancer cell line, were injected in the mammary fat pad of ten syngeneic mice. After 22 days the primary tumors were surgically removed and 8 weeks post-injection the animals were sacrificed and pulmonary tumors quantified. The growth of primary tumors formed by NS or Kdm3a knockdown cells was similar (Figure 4H and Figure 4—figure supplement 8). However, Kdm3a knockdown cells caused significantly increased metastatic burden in the lungs compared to control 4T07 cells (Figure 4I and Figure 4—figure supplement 9). Consistent with our results, knockdown of Bnip3 has also been shown to cause increased metastasis in similar in vivo experiments (Manka et al., 2005). Collectively, these results show that KDM3A functions to prevent metastasis.

Based on the results presented above, we propose a model of anoikis induction that is illustrated in Figure 3G and discussed below. Following detachment of non-transformed cells, integrin signaling is decreased leading to transcriptional induction of KDM3A. The increased levels of KDM3A results in its recruitment to the pro-apoptotic genes BNIP3 and BNIP3L, where it promotes demethylation of inhibitory H3K9me1/2 marks and transcriptional activation of the two genes, resulting in anoikis induction. Consistent with this model, previous studies have shown that hypoxia results in transcriptional activation of KDM3A, BNIP3 and BNIP3L (Beyer et al., 2008; Sowter et al., 2001). We have found that in anoikis-resistant human breast cancer cell lines and tumors, KDM3A expression is defective, highlighting the importance of this pathway in promoting anoikis. Collectively, our results reveal a novel transcriptional regulatory program that mediates anoikis in non-transformed cells and is disabled during cancer development.

As described above, previous studies have shown that BIM and BMF are also effectors of anoikis (Reginato et al., 2003; Schmelzle et al., 2007). However, we have found that unlike BNIP3 and BNIP3L, BIM and BMF are not regulated by KDM3A. Thus, our results reveal that anoikis is promoted by multiple non-redundant pathways, which may help prevent the development of anoikis resistance.

1. 1. Pedanou VE
   2. Green MR

   (2016) The histone H3K9 demethylase KDM3A promotes anoikis through transcriptional activation of pro-apoptotic genes BNIP3 and BNIP3L

   Publicly available at the NCBI Gene Expression Omnibus (accession no. GSE80144).

   http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE80144

1. 1. Aslakson CJ
   2. Miller FR

   (1992)

   Selective events in the metastatic process defined by analysis of the sequential dissemination of subpopulations of a mouse mammary tumor

   Cancer Research 52:1399–1405.

   • Google Scholar
2. 1. Barker AJ
   2. Gibson KH
   3. Grundy W
   4. Godfrey AA
   5. Barlow JJ
   6. Healy MP
   7. Woodburn JR
   8. Ashton SE
   9. Curry BJ
   10. Scarlett L
   11. Henthorn L
   12. Richards L

   (2001) Studies leading to the identification of ZD1839 (IRESSA): an orally active, selective epidermal growth factor receptor tyrosine kinase inhibitor targeted to the treatment of cancer

   Bioorganic & Medicinal Chemistry Letters 11:1911–1914.

   https://doi.org/10.1016/S0960-894X(01)00344-4

   • Google Scholar
3. 1. Beyer S
   2. Kristensen MM
   3. Jensen KS
   4. Johansen JV
   5. Staller P

   (2008) The histone demethylases JMJD1A and JMJD2B are transcriptional targets of hypoxia-inducible factor HIF

   Journal of Biological Chemistry 283:36542–36552.

   https://doi.org/10.1074/jbc.M804578200

   • Google Scholar
4. 1. Boyd JM
   2. Malstrom S
   3. Subramanian T
   4. Venkatesh LK
   5. Schaeper U
   6. Elangovan B
   7. D'Sa-Eipper C
   8. Chinnadurai G

   (1994) Adenovirus E1B 19 kDa and Bcl-2 proteins interact with a common set of cellular proteins

   Cell 79:341–351.

   https://doi.org/10.1016/0092-8674(94)90202-X

   • Google Scholar
5. 1. Debnath J
   2. Mills KR
   3. Collins NL
   4. Reginato MJ
   5. Muthuswamy SK
   6. Brugge JS

   (2002) The role of apoptosis in creating and maintaining luminal space within normal and oncogene-expressing mammary acini

   Cell 111:29–40.

   https://doi.org/10.1016/S0092-8674(02)01001-2

   • Google Scholar
6. 1. Favata MF
   2. Horiuchi KY
   3. Manos EJ
   4. Daulerio AJ
   5. Stradley DA
   6. Feeser WS
   7. Van Dyk DE
   8. Pitts WJ
   9. Earl RA
   10. Hobbs F
   11. Copeland RA
   12. Magolda RL
   13. Scherle PA
   14. Trzaskos JM

   (1998) Identification of a novel inhibitor of mitogen-activated protein kinase kinase

   Journal of Biological Chemistry 273:18623–18632.

   https://doi.org/10.1074/jbc.273.29.18623

   • Google Scholar
7. 1. Finak G
   2. Bertos N
   3. Pepin F
   4. Sadekova S
   5. Souleimanova M
   6. Zhao H
   7. Chen H
   8. Omeroglu G
   9. Meterissian S
   10. Omeroglu A
   11. Hallett M
   12. Park M

   (2008) Stromal gene expression predicts clinical outcome in breast cancer

   Nature Medicine 14:518–527.

   https://doi.org/10.1038/nm1764

   • Google Scholar
8. 1. Frisch SM
   2. Francis H

   (1994) Disruption of epithelial cell-matrix interactions induces apoptosis

    Journal of Cell Biology 124:619–626.

   https://doi.org/10.1083/jcb.124.4.619

   • Google Scholar
9. 1. Frisch SM
   2. Vuori K
   3. Ruoslahti E
   4. Chan-Hui PY

   (1996) Control of adhesion-dependent cell survival by focal adhesion kinase

    Journal of Cell Biology 134:793–799.

   https://doi.org/10.1083/jcb.134.3.793

   • Google Scholar
10. 1. Gazin C
    2. Wajapeyee N
    3. Gobeil S
    4. Virbasius CM
    5. Green MR

    (2007) An elaborate pathway required for Ras-mediated epigenetic silencing

    Nature 449:1073–1077.

    https://doi.org/10.1038/nature06251

    • Google Scholar
11. 1. Huang W
    2. Gonzalez ME
    3. Toy KA
    4. Banerjee M
    5. Kleer CG

    (2010) Blockade of CCN6 (WISP3) activates growth factor-independent survival and resistance to anoikis in human mammary epithelial cells

    Cancer Research 70:3340–3350.

    https://doi.org/10.1158/0008-5472.CAN-09-4225

    • Google Scholar
12. 1. Ihaka R
    2. Gentleman R

    (1996) R: A language for data analysis and graphics

    Journal of Computational and Graphical Statistics 5:299–314.

    https://doi.org/10.1080/10618600.1996.10474713

    • Google Scholar
13. 1. King WG
    2. Mattaliano MD
    3. Chan TO
    4. Tsichlis PN
    5. Brugge JS

    (1997) Phosphatidylinositol 3-kinase is required for integrin-stimulated AKT and Raf-1/mitogen-activated protein kinase pathway activation

    Molecular and Cellular Biology 17:4406–4418.

    https://doi.org/10.1128/MCB.17.8.4406

    • Google Scholar
14. 1. Lomonosova E
    2. Chinnadurai G

    (2008) BH3-only proteins in apoptosis and beyond: an overview

    Oncogene 27:S2–19.

    https://doi.org/10.1038/onc.2009.39

    • Google Scholar
15. 1. Manka D
    2. Spicer Z
    3. Millhorn DE

    (2005) Bcl-2/adenovirus E1B 19 kDa interacting protein-3 knockdown enables growth of breast cancer metastases in the lung, liver, and bone

    Cancer Research 65:11689–11693.

    https://doi.org/10.1158/0008-5472.CAN-05-3091

    • Google Scholar
16. 1. Matsushima M
    2. Fujiwara T
    3. Takahashi Ei-ichi
    4. Minaguchi T
    5. Eguchi Y
    6. Tsujimoto Y
    7. Suzumori K
    8. Nakamura Y
    9. Takahashi E

    (1998) Isolation, mapping, and functional analysis of a novel human cDNA (BNIP3L) encoding a protein homologous to human NIP3

    Genes, Chromosomes and Cancer 21:230–235.

    https://doi.org/10.1002/(SICI)1098-2264(199803)21:3<230::AID-GCC7>3.0.CO;2-0

    • Google Scholar
17. 1. Moro L
    2. Venturino M
    3. Bozzo C
    4. Silengo L
    5. Altruda F
    6. Beguinot L
    7. Tarone G
    8. Defilippi P

    (1998) Integrins induce activation of EGF receptor: role in MAP kinase induction and adhesion-dependent cell survival

    The EMBO Journal 17:6622–6632.

    https://doi.org/10.1093/emboj/17.22.6622

    • Google Scholar
18. 1. Paoli P
    2. Giannoni E
    3. Chiarugi P

    (2013) Anoikis molecular pathways and its role in cancer progression

    Biochimica Et Biophysica Acta 1833:3481–3498.

    https://doi.org/10.1016/j.bbamcr.2013.06.026

    • Google Scholar
19. 1. Reginato MJ
    2. Mills KR
    3. Paulus JK
    4. Lynch DK
    5. Sgroi DC
    6. Debnath J
    7. Muthuswamy SK
    8. Brugge JS

    (2003) Integrins and EGFR coordinately regulate the pro-apoptotic protein bim to prevent anoikis

    Nature Cell Biology 5:733–740.

    https://doi.org/10.1038/ncb1026

    • Google Scholar
20. 1. Rhodes DR
    2. Kalyana-Sundaram S
    3. Mahavisno V
    4. Varambally R
    5. Yu J
    6. Briggs BB
    7. Barrette TR
    8. Anstet MJ
    9. Kincead-Beal C
    10. Kulkarni P
    11. Varambally S
    12. Ghosh D
    13. Chinnaiyan AM

    (2007) Oncomine 3.0: genes, pathways, and networks in a collection of 18,000 cancer gene expression profiles

    Neoplasia 9:166–180.

    https://doi.org/10.1593/neo.07112

    • Google Scholar
21. 1. Santra MK
    2. Wajapeyee N
    3. Green MR

    (2009) F-box protein FBXO31 mediates cyclin D1 degradation to induce G1 arrest after DNA damage

    Nature 459:722–725.

    https://doi.org/10.1038/nature08011

    • Google Scholar
22. 1. Schmelzle T
    2. Mailleux AA
    3. Overholtzer M
    4. Carroll JS
    5. Solimini NL
    6. Lightcap ES
    7. Veiby OP
    8. Brugge JS

    (2007) Functional role and oncogene-regulated expression of the BH3-only factor Bmf in mammary epithelial anoikis and morphogenesis

    PNAS 104:3787–3792.

    https://doi.org/10.1073/pnas.0700115104

    • Google Scholar
23. 1. Schwartz MA

    (1997) Integrins, oncogenes, and anchorage independence

     Journal of Cell Biology 139:575–578.

    https://doi.org/10.1083/jcb.139.3.575

    • Google Scholar
24. 1. Silva J
    2. Mak W
    3. Zvetkova I
    4. Appanah R
    5. Nesterova TB
    6. Webster Z
    7. Peters AH
    8. Jenuwein T
    9. Otte AP
    10. Brockdorff N

    (2003) Establishment of histone h3 methylation on the inactive X chromosome requires transient recruitment of Eed-Enx1 polycomb group complexes

    Developmental Cell 4:481–495.

    https://doi.org/10.1016/S1534-5807(03)00068-6

    • Google Scholar
25. 1. Silva JM
    2. Li MZ
    3. Chang K
    4. Ge W
    5. Golding MC
    6. Rickles RJ
    7. Siolas D
    8. Hu G
    9. Paddison PJ
    10. Schlabach MR
    11. Sheth N
    12. Bradshaw J
    13. Burchard J
    14. Kulkarni A
    15. Cavet G
    16. Sachidanandam R
    17. McCombie WR
    18. Cleary MA
    19. Elledge SJ
    20. Hannon GJ

    (2005) Second-generation shRNA libraries covering the mouse and human genomes

    Nature Genetics 37:1281–1288.

    https://doi.org/10.1038/ng1650

    • Google Scholar
26. 1. Simpson CD
    2. Anyiwe K
    3. Schimmer AD

    (2008) Anoikis resistance and tumor metastasis

    Cancer Letters 272:177–185.

    https://doi.org/10.1016/j.canlet.2008.05.029

    • Google Scholar
27. 1. Sorlie T
    2. Perou CM
    3. Tibshirani R
    4. Aas T
    5. Geisler S
    6. Johnsen H
    7. Hastie T
    8. Eisen MB
    9. van de Rijn M
    10. Jeffrey SS
    11. Thorsen T
    12. Quist H
    13. Matese JC
    14. Brown PO
    15. Botstein D
    16. Lonning PE
    17. Borresen-Dale A-L

    (2001) Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications

    PNAS 98:10869–10874.

    https://doi.org/10.1073/pnas.191367098

    • Google Scholar
28. 1. Sowter HM
    2. Ratcliffe PJ
    3. Watson P
    4. Greenberg AH
    5. Harris AL

    (2001)

    HIF-1-dependent regulation of hypoxic induction of the cell death factors BNIP3 and NIX in human tumors

    Cancer Research 61:6669–6673.

    • Google Scholar
29. 1. Tabariès S
    2. Dong Z
    3. Annis MG
    4. Omeroglu A
    5. Pepin F
    6. Ouellet V
    7. Russo C
    8. Hassanain M
    9. Metrakos P
    10. Diaz Z
    11. Basik M
    12. Bertos N
    13. Park M
    14. Guettier C
    15. Adam R
    16. Hallett M
    17. Siegel PM

    (2011) Claudin-2 is selectively enriched in and promotes the formation of breast cancer liver metastases through engagement of integrin complexes

    Oncogene 30:1318–1328.

    https://doi.org/10.1038/onc.2010.518

    • Google Scholar
30. 1. Taube ME
    2. Liu XW
    3. Fridman R
    4. Kim HR

    (2006) TIMP-1 regulation of cell cycle in human breast epithelial cells via stabilization of p27(KIP1) protein

    Oncogene 25:3041–3048.

    https://doi.org/10.1038/sj.onc.1209336

    • Google Scholar
31. Website

    1. TCGA

    (2011) The cancer genome atlas - Invasive breast carcinoma gene expression data

    http://tcga-data.nci.nih.gov/tcga/
32. 1. Voisin L
    2. Julien C
    3. Duhamel S
    4. Gopalbhai K
    5. Claveau I
    6. Saba-El-Leil MK
    7. Rodrigue-Gervais IG
    8. Gaboury L
    9. Lamarre D
    10. Basik M
    11. Meloche S

    (2008) Activation of MEK1 or MEK2 isoform is sufficient to fully transform intestinal epithelial cells and induce the formation of metastatic tumors

    BMC Cancer 8:337.

    https://doi.org/10.1186/1471-2407-8-337

    • Google Scholar
33. 1. Ward WHJ
    2. Cook PN
    3. Slater AM
    4. Davies DH
    5. Holdgate GA
    6. Green LR

    (1994) Epidermal growth factor receptor tyrosine kinase

    Biochemical Pharmacology 48:659–666.

    https://doi.org/10.1016/0006-2952(94)90042-6

    • Google Scholar
34. 1. Xie L
    2. Gazin C
    3. Park SM
    4. Zhu LJ
    5. Debily MA
    6. Kittler EL
    7. Zapp ML
    8. Lapointe D
    9. Gobeil S
    10. Virbasius CM
    11. Green MR

    (2012) A synthetic interaction screen identifies factors selectively required for proliferation and TERT transcription in p53-deficient human cancer cells

    PLoS Genetics 8:e1003151.

    https://doi.org/10.1371/journal.pgen.1003151

    • Google Scholar
35. 1. Yamane K
    2. Toumazou C
    3. Tsukada Y
    4. Erdjument-Bromage H
    5. Tempst P
    6. Wong J
    7. Zhang Y

    (2006) JHDM2A, a JmjC-containing H3K9 demethylase, facilitates transcription activation by androgen receptor

    Cell 125:483–495.

    https://doi.org/10.1016/j.cell.2006.03.027

    • Google Scholar
36. 1. Zhao H
    2. Langerød A
    3. Ji Y
    4. Nowels KW
    5. Nesland JM
    6. Tibshirani R
    7. Bukholm IK
    8. Kåresen R
    9. Botstein D
    10. Børresen-Dale AL
    11. Jeffrey SS

    (2004) Different gene expression patterns in invasive lobular and ductal carcinomas of the breast

    Molecular Biology of the Cell 15:2523–2536.

    https://doi.org/10.1091/mbc.E03-11-0786

    • Google Scholar
37. 1. Zhu Z
    2. Sanchez-Sweatman O
    3. Huang X
    4. Wiltrout R
    5. Khokha R
    6. Zhao Q
    7. Gorelik E

    (2001)

    Anoikis and metastatic potential of cloudman S91 melanoma cells

    Cancer Research 61:1707–1716.

    • Google Scholar

Author details

1. Victoria E Pedanou

   1. Department of Molecular, Cell and Cancer Biology, University of Massachusetts Medical School, Worcester, United States
   2. Howard Hughes Medical Institute, University of Massachusetts Medical School, Worcester, United States

   Contribution

   VEP, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

2. Stéphane Gobeil

   1. Department of Molecular Medicine, Université Laval, Quebec City, Canada
   2. Centre de recherche du CHU de Québec, CHUL, Québec PQ, Canada

   Contribution

   SG, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

3. Sébastien Tabariès

   Department of Medicine, Goodman Cancer Research Centre, McGill University, Montreal, Canada

   Contribution

   ST, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

4. Tessa M Simone

   1. Department of Molecular, Cell and Cancer Biology, University of Massachusetts Medical School, Worcester, United States
   2. Howard Hughes Medical Institute, University of Massachusetts Medical School, Worcester, United States

   Contribution

   TMS, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

5. Lihua Julie Zhu

   1. Department of Molecular, Cell and Cancer Biology, University of Massachusetts Medical School, Worcester, United States
   2. Program in Molecular Medicine, University of Massachusetts Medical School, Worcester, United States
   3. Program in Bioinformatics and Integrative Biology, University of Massachusetts Medical School, Worcester, United States

   Contribution

   LJZ, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

6. Peter M Siegel

   Department of Medicine, Goodman Cancer Research Centre, McGill University, Montreal, Canada

   Contribution

   PMS, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

7. Michael R Green

   1. Department of Molecular, Cell and Cancer Biology, University of Massachusetts Medical School, Worcester, United States
   2. Howard Hughes Medical Institute, University of Massachusetts Medical School, Worcester, United States

   Contribution

   MRG, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   For correspondence

   michael.green@umassmed.edu

   Competing interests

   The authors declare that no competing interests exist.

   "This ORCID iD identifies the author of this article:" 0000-0003-3017-3298

• 1,823

  views

• 471

  downloads

• 22

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.