ABSTRACT
DNA alkylating agents form the first line of cancer chemotherapy. They not only kill cells but also behave as potential carcinogens. MNU, a DNA methylating agent, is well known to induce mammary tumours in rodents. However, the mechanism of tumorigenesis is not well understood. Our study reports a novel role played by DNA-dependent protein kinase (DNA-PK) in methylation damage-induced transformation using three-dimensional breast acinar cultures. Here, we report that exposure of breast epithelial cells to MNU inhibited polarisation at the basolateral domain, increased dispersal of the Golgi at the apical domain and induced an epithelial-to-mesenchymal transition (EMT)-like phenotype as well as invasion. This altered Golgi phenotype correlated with impaired intracellular trafficking. Inhibition of DNA-PK resulted in almost complete reversal of the altered Golgi phenotype and partial rescue of the polarity defect and EMT-like phenotype. The results confirm that methylation damage-induced activation of DNA-PK is a major mechanism in mediating cellular transformation.
This article has an associated First Person interview with the first author of the paper.
INTRODUCTION
Many harmful chemicals in the environment challenge human DNA. Aberrant regulation of the DNA damage response leads to genome instability and various diseases including cancer. DNA damage, which is frequently used to treat cancer, can also cause cancer and therefore appears as a side effect of cancer treatment (Kastan, 2008). On similar lines, Fung and colleagues demonstrated that the reversal of apoptosis following DNA damage in cells can eventually result in oncogenic transformation (Tang et al., 2012). Additionally, Paules and colleagues reported the induction of ‘cancer-predisposition programs’ in lymphocytes following DNA damage (Innes et al., 2013).
DNA damage and its accumulation occurs as a result of exposure to many common chemicals, such as nitrosamines from cigarette smoke and nitrodialkylamines found in dietary sources, which act as potential alkylating agents capable of alkylating DNA either directly or following metabolic activation (Silber et al., 1996). Furthermore, a few endogenous processes are also known to alkylate DNA (Barrows and Magee, 1982; Vaca et al., 1988). DNA damaging agents are widely used in cancer chemotherapy, alkylating agents being some of the oldest and most commonly used chemicals (Chaney and Sancar, 1996; Cheung-Ong et al., 2013). Thus, it is essential to study the process by which DNA damage can induce cancer. The constant remodelling occurring in breast tissue in women, from menarche until menopause, makes it more prone to DNA damage caused by both endogenous and exogenous sources (Davis and Lin, 2011). Alkylating agents have also been used to induce tumours in rodent models. DMBA (7,12-dimethylbenzanthracene) and MNU (N-methyl N-nitrosourea) are two commonly used carcinogens that induce mammary tumours in rodents (Gullino et al., 1975; Medina et al., 1980). MNU, a prototypic SN1 type of DNA methylating agent, reacts with O positions (harder nucleophiles) to a greater extent than other SN1 methylating agents (Kondo et al., 2010; Beland and Poirier, 1989). Generally, methylating agents alkylate DNA at O and N positions. However, among the different adducts formed, O6-methylguanine adducts are the most significant and are known to be cytotoxic as well as mutagenic (Kondo et al., 2010). The physiological relevance of this lesion was supported in a population-based study where the DNAs of individuals were found to carry the modified base O6-methylguanine (Kyrtopoulos, 1998). Furthermore, such adducts were found in larger proportions in cancer patients treated with chemotherapy than in untreated patients (Povey, 2000).
MNU has been widely used to induce mammary tumours in rodent models. These induced tumours have been demonstrated to have close similarities to human breast tumours (Tsubura et al., 2011). MNU-induced tumours in rats are known to carry a mutation (glycine to alanine transition) in the 12th codon of the H-RAS gene (Sukumar et al., 1983; Zarbl et al., 1985). By contrast, Maffini et al., demonstrated that there is no correlation of H-RAS mutation with initiation of the tumour, but neoplastic transformation was induced by the stroma (Maffini et al., 2004). Nevertheless, there are no direct reports pertaining to the mechanism of MNU-induced tumour formation, which is one of the main objectives of this work.
In this study, we report that MNU-induced DNA damage caused dispersal of the Golgi and affected intracellular trafficking through activation of DNA-dependent protein kinase (DNA-PK). This, in turn, led to transformation of non-tumorigenic breast epithelial cells grown as spheroids (as observed through disruption of apico-basal polarity), induction of an epithelial-to-mesenchymal transition (EMT)-like phenotype, as well as invasion, and acquisition of anchorage-independent growth. DMNB, a small molecule inhibitor of DNA-PK, partially reversed the EMT-like phenotype and invasiveness.
RESULTS
MNU treatment causes DNA damage and leads to increase in nuclear volume of MCF10A cells grown in 3D culture
To select a suitable sub-lethal dose of MNU, the MTT-based cytotoxicity assay was performed on MCF10A cells exposed to different doses of MNU. Doses of 1 mM and 1.5 mM were found to be suitable (more than 75% viability) (Fig. 1A). Cells were exposed to 1 mM MNU during the initial days of morphogenesis (day 0 and day 2) to ensure that only those cells that encountered DNA damage and subsequently survived were used to characterise the phenotypes in 3D cell culture. To ensure that the selected dose of MNU induced DNA damage, comet assays were performed. Alkaline and neutral comet assays (Fig. 1B-E and Fig. S1) were performed on MCF10A cells grown as 3D cultures, with or without 1 mM MNU treatment, to visualise whether breaks were formed following MNU exposure. The percentage tail DNA (Fig. 1B and D) and tail length (Fig. 1C and E) were used as measures of the extent of DNA damage. MNU (1 mM ) was found to induce formation of both single-strand (SSBs) and double-strand breaks (DSBs). These breaks were formed within 2 h of treatment and persisted up to 24 h. Thus, 1 mM MNU was found to be an appropriate sub-lethal dose for MCF10A cells and was capable of inducing both SSBs and DSBs.
MNU induced DNA damage and altered nuclear morphology. (A) MCF10A cells were treated with doses of MNU ranging from 0.2 to 3 mM. Cell viability is represented as a percentage of the viability of vehicle-treated cells. (B–E) MCF10A acini treated with 1 mM MNU for 2 h and 24 h were subjected to alkaline and neutral comet assay and imaged at 20× magnification. The % tail DNA for alkaline comet (B) and neutral comet (D), and tail length for alkaline comet (C) for neutral comet (E) were measured using ImageJ software and represented as box plots. Student's t-test was used to analyse the statistical significance of difference in tail length and % tail DNA; n=150, number of cells imaged from three independent experiments. (F,G) Immunofluorescence images of 3D cultures were captured using the Zeiss LSM 710 laser scanning confocal microscope. The images were analysed to measure the volume of acini nuclei at day 16 (F) and day 30 (G) using Huygens Professional software (SVI, Hilversum, Netherlands). Data are pooled from n>5 independent experiments. Statistical analysis was performed using the Mann–Whitney test for (F) and (G). *P<0.05, **P<0.01 and ***P<0.0001.
MNU induced DNA damage and altered nuclear morphology. (A) MCF10A cells were treated with doses of MNU ranging from 0.2 to 3 mM. Cell viability is represented as a percentage of the viability of vehicle-treated cells. (B–E) MCF10A acini treated with 1 mM MNU for 2 h and 24 h were subjected to alkaline and neutral comet assay and imaged at 20× magnification. The % tail DNA for alkaline comet (B) and neutral comet (D), and tail length for alkaline comet (C) for neutral comet (E) were measured using ImageJ software and represented as box plots. Student's t-test was used to analyse the statistical significance of difference in tail length and % tail DNA; n=150, number of cells imaged from three independent experiments. (F,G) Immunofluorescence images of 3D cultures were captured using the Zeiss LSM 710 laser scanning confocal microscope. The images were analysed to measure the volume of acini nuclei at day 16 (F) and day 30 (G) using Huygens Professional software (SVI, Hilversum, Netherlands). Data are pooled from n>5 independent experiments. Statistical analysis was performed using the Mann–Whitney test for (F) and (G). *P<0.05, **P<0.01 and ***P<0.0001.
Nuclear morphology is altered in many cancers (Zink et al., 2004); hence, we examined the nuclear morphology of MCF10A cells (using Huygens Professional software; SVI, Hilversum, Netherlands). Morphometric analysis of the nuclei revealed that treated cells had nuclei of significantly larger volume than those of control cells. The difference persisted until day 30, implying that treatment of MCF10A cells with 1 mM MNU during the early days of morphogenesis leads to an appreciable increase in nuclear volume (Fig. 1F and G).
MNU treatment of MCF10A cells results in loss of basolateral polarity
MCF10A spheroids contain a monolayer of polarised epithelial cells surrounding a hollow lumen. Loss of polarity in these cells is one of the hallmarks of transformed cells (Debnath and Brugge, 2005). We found that MNU-treated MCF10A cells showed a disruption of basolateral polarity at day 16. We quantified the number of acini showing disruption of basolateral markers. α6 integrin, which marks the basal region of the acini (Debnath et al., 2003), was found to be either mis-localised to the apical region (33% acini) or discontinuous in the basal region (43% acini), indicating disruption of basal polarity (Fig. 2A,B). About 52% of the treated acini showed loss of laminin V, a basement membrane marker (Fig. 2C,D), thus confirming loss of basal polarity in the MNU-treated breast acini.
MNU damage disrupted basolateral polarity. MCF10A cells grown as 3D cultures for 16 days untreated and treated were immunostained with basolateral polarity markers. (A) Representative images of day 16 acini immunostained with α6 integrin (green) antibody, a marker for basal polarity. (B) Graph representing quantification of α6 integrin phenotypes observed. (C) Representative images of day 16 acini immunostained for laminin V (red), a marker for basal polarity. (D) Graph representing quantification of laminin V phenotypes observed. (E) Representative images of day 16 acini immunostained for E-cadherin (green). (F) Graph representing quantification of E-cadherin phenotypes observed. (G) Representative images of day 16 acini immunostained for β-catenin (red). (H) Graph representing quantification of β-catenin phenotypes observed. (I) Representative images of day 16 acini immunostained for hDLG (green). (J) Graph representing quantification of hDLG phenotypes observed. Data are pooled from n>5 independent experiments.
MNU damage disrupted basolateral polarity. MCF10A cells grown as 3D cultures for 16 days untreated and treated were immunostained with basolateral polarity markers. (A) Representative images of day 16 acini immunostained with α6 integrin (green) antibody, a marker for basal polarity. (B) Graph representing quantification of α6 integrin phenotypes observed. (C) Representative images of day 16 acini immunostained for laminin V (red), a marker for basal polarity. (D) Graph representing quantification of laminin V phenotypes observed. (E) Representative images of day 16 acini immunostained for E-cadherin (green). (F) Graph representing quantification of E-cadherin phenotypes observed. (G) Representative images of day 16 acini immunostained for β-catenin (red). (H) Graph representing quantification of β-catenin phenotypes observed. (I) Representative images of day 16 acini immunostained for hDLG (green). (J) Graph representing quantification of hDLG phenotypes observed. Data are pooled from n>5 independent experiments.
In addition, E-cadherin, an adherens junction marker, was decreased as well as diffused from the cell–cell junctions in 56% of the acini (Fig. 2E and F). To confirm the loss of E-cadherin from cell–cell junctions, cells dissociated from day 16 acini were grown as a monolayer and analysed. As in the phenotype observed in 3D cultures, E-cadherin was lost from cell–cell junctions in the monolayer cultures (Fig. S2A). β-catenin, another marker for cell–cell junctions, also showed cytoplasmic localisation as well as a diffused staining pattern in 59% of acini (Fig. 2G and H), indicating a transformed phenotype. hDlg, a component of the Scribble complex, also showed a diffused pattern at the cell–cell junctions (Fig. 2I,J). Loss of basal and lateral components of epithelial cells indicates loss of polarity. To determine whether this loss of polarity in response to MNU treatment was a transient effect or not, the acini were grown until day 30; a similar loss phenotype of the α6 integrin basal polarity marker was observed (Fig. S2B,C).
MNU-treated MCF10A cells show pERM mis-localisation and altered Golgi morphology with impaired cellular trafficking
The Golgi plays a vital role in polarised sorting of proteins to the basolateral and apical regions of epithelial cells, thus being a central organelle in establishment and maintenance of polarity. In 3D spheroid cultures, the Golgi is known to localise to the apical region above the nuclei (Aranda et al., 2006). An interesting observation in this study was the change in Golgi morphology. After MNU treatment, the Golgi exhibited an altered phenotype and quantification revealed a significant increase in the Golgi area (Fig. 3A and Fig. S2D). Furthermore, the altered phenotype was also observed in day 30 acini, implying that the phenotype was not transient following MNU damage (Fig. S2E). In order to ascertain the integrity of apical polarity, pERM (phosphorylated ezrin, radixin and moesin), a plasma membrane marker that localises to actin-rich structures, was used (Morales et al., 2004). In MCF10A acinar structures, pERM marks apical region and the plasma membrane (Debnath et al., 2003). MNU-treated acini showed a strong basal localisation of pERM, indicating polarity disruption (Fig. 3B).
MNU induced aberrant Golgi morphology with impaired intracellular trafficking from Golgi to plasma membrane. MCF10A 3D ‘on-top’ cultures were treated with two doses of MNU (day 0 and day 2), grown for 16 days and immunostained for GM130. (A,B) Representative images of day 16 acini immunostained with apical polarity marker GM130 antibody (green) (A) and apical membrane marker pERM (red) (B). (C,D) MCF10A cells grown as monolayers were treated with 1 mM MNU for 24 h. Cells with or without treatment were transfected with ts-045 VSVG-GFP construct and shifted to non-permissive temperature of 40°C. Samples were collected at different time-points following a shift to the permissive temperature of 32°C. Surface VSVG molecules were stained with conformation specific antibody under non-permeabilising conditions. (C) Representative images of cells incubated at 32°C for 80 min and 120 min. (D) Graph represents mean±s.e.m. of the relative ratio of surface VSVG molecules (red fluorescence) to total VSVG molecules (green fluorescence). Statistical significance of the effect of MNU and time on trafficking of VSVG molecules was determined using two-way ANOVA. P<0.05 is considered statistically significant. Data are pooled from three independent experiments.
MNU induced aberrant Golgi morphology with impaired intracellular trafficking from Golgi to plasma membrane. MCF10A 3D ‘on-top’ cultures were treated with two doses of MNU (day 0 and day 2), grown for 16 days and immunostained for GM130. (A,B) Representative images of day 16 acini immunostained with apical polarity marker GM130 antibody (green) (A) and apical membrane marker pERM (red) (B). (C,D) MCF10A cells grown as monolayers were treated with 1 mM MNU for 24 h. Cells with or without treatment were transfected with ts-045 VSVG-GFP construct and shifted to non-permissive temperature of 40°C. Samples were collected at different time-points following a shift to the permissive temperature of 32°C. Surface VSVG molecules were stained with conformation specific antibody under non-permeabilising conditions. (C) Representative images of cells incubated at 32°C for 80 min and 120 min. (D) Graph represents mean±s.e.m. of the relative ratio of surface VSVG molecules (red fluorescence) to total VSVG molecules (green fluorescence). Statistical significance of the effect of MNU and time on trafficking of VSVG molecules was determined using two-way ANOVA. P<0.05 is considered statistically significant. Data are pooled from three independent experiments.
To investigate the effect of MNU-induced aberrant Golgi morphology on Golgi to plasma membrane trafficking, a GFP-fused temperature-sensitive mutant of VSVG glycoprotein was used. At the non-permissive temperature of 40°C, VSVG molecules accumulate in the endoplasmic reticulum (ER). Upon inhibition of protein synthesis and shifting to the permissive temperature of 32°C, VSVG molecules are synchronously released from ER to the Golgi (by 40 min) and then to the plasma membrane (by 120 min) (Presley et al., 1997). VSVG molecules that reached the surface were marked using the conformation-specific antibody (clone 8G5F11). The ratio of surface VSVG to total VSVG was quantified to estimate the extent of delivery to the plasma membrane. It was found that Golgi trafficking was impaired following exposure of MCF10A cells to 1 mM MNU for 24 h (Fig. 3C and D).
To test whether MNU treatment affected ER to Golgi or Golgi to plasma membrane trafficking, RUSH (retention using selective hooks) assay was performed using the Str-KDEL-ManII-SBP-EGFP construct, where Str-KDEL is the ER hook and ManII-SBP-EGFP is the reporter. Ten minutes after biotin addition, most of the MNU-treated cells did not show ManII-SBP-EGFP reporter localisation to the Golgi, in contrast to untreated cells (Fig. 4A and B). However, the percentage of cells showing Golgi localisation of the reporter increased by 20 min, indicating that ER to Golgi trafficking was impaired but not completely abrogated (Fig. 4A,C).
MNU treatment impaired intracellular trafficking. (A) Representative images of MCF10A cells with and without MNU treatment transfected with Str-KDEL-ManII-SBP-EGFP construct. Reporter (green) release was observed at 0, 10 and 20 min after biotin addition. (B,C) Graphs represent the localisation of the reporter 10 min (B) and 20 min (C) following biotin addition. C denotes control and T denotes 1 mM MNU-treated cells. (D) Representative images of Con A-labelled (green) MCF10A cells with or without 1 mM MNU treatment. (E) MCF10A cells grown on Matrigel™-coated coverslips, immunostained for α3 integrin (green). Nuclei were counterstained with Hoechst 33342 (blue).
MNU treatment impaired intracellular trafficking. (A) Representative images of MCF10A cells with and without MNU treatment transfected with Str-KDEL-ManII-SBP-EGFP construct. Reporter (green) release was observed at 0, 10 and 20 min after biotin addition. (B,C) Graphs represent the localisation of the reporter 10 min (B) and 20 min (C) following biotin addition. C denotes control and T denotes 1 mM MNU-treated cells. (D) Representative images of Con A-labelled (green) MCF10A cells with or without 1 mM MNU treatment. (E) MCF10A cells grown on Matrigel™-coated coverslips, immunostained for α3 integrin (green). Nuclei were counterstained with Hoechst 33342 (blue).
To investigate the consequences of trafficking defects, the change in plasma membrane composition was investigated. Labelling cell membranes with fluorescently conjugated lectins is a common strategy for studying defects in trafficking between ER and plasma membrane (Chia et al., 2012). We used a fluorescent conjugate of Concanavalin A (Con A) to label mannopyranosyl and glucopyranosyl residues on the plasma membrane (Byrne et al., 2012). Untreated cells exhibited uniform labelling of the plasma membrane, whereas MNU-treated cells showed non-uniform labelling of Con A at the plasma membrane (Fig. 4D, arrows). Thus, MNU damage perturbed cellular trafficking, which ultimately affected the composition of the cell membrane. To confirm this, we investigated integrin localisation in cells (Riggs et al., 2012; Caswell and Norman, 2006; Bretscher, 1989) with and without MNU treatment. α3 integrin appeared to be present in vesicles in the cytoplasm following MNU treatment, thus corroborating the conclusion that MNU treatment impaired cellular trafficking (Fig. 4E). The presence of comparable α3 integrin staining at the cell membrane in the treated cells further strengthens the claim that, although trafficking is impaired or delayed, it is not completely abrogated.
Thus, MNU-induced DNA damage leads to disruption of both Golgi morphology and intracellular trafficking. These data, in combination with available literature (Keller et al., 2001; Griffiths and Simons, 1986) on the role of the Golgi complex in mediating distribution of polarity proteins, suggest that Golgi dispersal can affect the disruption of epithelial cell polarity.
MNU-treated MCF10A cells exhibit an EMT-like phenotype as well as invasiveness
Loss of polarity coupled with loss of cell–cell junctions and attainment of fibroblast-like elongated morphology are the pre-requisites of EMT (Thiery, 2003). MNU-treated cells formed acini with disrupted polarity and cell–cell junctions. Hence, to study whether an EMT-like phenotype was induced in the MNU-treated breast spheroids, the acini grown on Matrigel® were immunostained for vimentin, an EMT marker. We observed that 69% of MNU-treated acini showed upregulation of vimentin at day 16 (Fig. 5A,B). To confirm this, whole cell lysates of day 16 3D cultures were immunoblotted and probed for different EMT markers. A significant upregulation of EMT markers such as vimentin, fibronectin, N-cadherin and snail were observed (Fig. 5C). Cytokeratin 19 and 14 as well as E-cadherin (all epithelial cell markers) showed significant downregulation following damage (Fig. 5C). The upregulation of fibronectin and N-cadherin were observed as early as day 6 and increased expression was maintained until day 30 (Fig. S2F). The transcript levels of vimentin, twist and snail showed significant upregulation in treated acini compared with controls (Fig. 5D), which corroborated our finding that MNU induced an EMT-like phenotype in breast acini. Taken together, these results imply that exposure of non-transformed epithelial cells to a DNA damaging agent leads to induction of an EMT-like phenotype.
MNU induced EMT-like phenotype and invasion in MCF10A spheroids. MCF10A cells were grown as 3D cultures, exposed to two doses of 1 mM MNU and cultured for 16 days. (A) Representative images of day 16 acini immunostained for vimentin (red), an EMT marker. (B) Quantification of phenotypes observed. (C,D) Day 16 acini were lysed and analysed for various markers of EMT by western blotting (C) and mRNA expression (D). (E) Day 16 structures on collagen:Matrigel (1:1), stained with laminin V (green) and phalloidin (red). The cells subjected to damage showed cytoplasmic localisation of laminin V as well an elongated morphology (an indication of attainment of mesenchymal phenotype). (F) Day 16 cultures of MCF10A cells with and without MNU treatment were dissociated, passaged more than three times and then seeded on coverslips coated with rat tail collagen type 1 containing DQ™ collagen type 1. Fluorescence intensity was captured using an Olympus IX81 system equipped with a Hamamatsu ORCA-R2 CCD camera. (G) Quantification of intensity per cell in the presence and absence of MNU treatment. (H) Commassie gel showing gelatinase activity of active MMP-9 in the presence and absence of MNU. Data were pooled from n>2 independent experiments. (I) Representative images of soft agar colony formation assay. (J) Graph representing the average total number of colonies observed in ten randomly selected fields per treatment across two independent biological experiments in triplicate. C denotes control and T denotes 1 mM MNU-treated cells.
MNU induced EMT-like phenotype and invasion in MCF10A spheroids. MCF10A cells were grown as 3D cultures, exposed to two doses of 1 mM MNU and cultured for 16 days. (A) Representative images of day 16 acini immunostained for vimentin (red), an EMT marker. (B) Quantification of phenotypes observed. (C,D) Day 16 acini were lysed and analysed for various markers of EMT by western blotting (C) and mRNA expression (D). (E) Day 16 structures on collagen:Matrigel (1:1), stained with laminin V (green) and phalloidin (red). The cells subjected to damage showed cytoplasmic localisation of laminin V as well an elongated morphology (an indication of attainment of mesenchymal phenotype). (F) Day 16 cultures of MCF10A cells with and without MNU treatment were dissociated, passaged more than three times and then seeded on coverslips coated with rat tail collagen type 1 containing DQ™ collagen type 1. Fluorescence intensity was captured using an Olympus IX81 system equipped with a Hamamatsu ORCA-R2 CCD camera. (G) Quantification of intensity per cell in the presence and absence of MNU treatment. (H) Commassie gel showing gelatinase activity of active MMP-9 in the presence and absence of MNU. Data were pooled from n>2 independent experiments. (I) Representative images of soft agar colony formation assay. (J) Graph representing the average total number of colonies observed in ten randomly selected fields per treatment across two independent biological experiments in triplicate. C denotes control and T denotes 1 mM MNU-treated cells.
MCF10A cells grown on collagen:Matrigel® (1:1) mixture formed acini (alveolar structures), although a few cells also formed multilayered structures resembling the ductal-like morphology described by Dhimolea et al. (Dhimolea et al., 2010). Altering the stiffness of Matrigel with collagen in a 1:1 ratio has been shown to promote the formation of invasive structures (Xiang and Muthuswamy, 2006). Interestingly, cells exposed to MNU failed to form alveolar structures as well as ductal structures; instead, they formed a cluster of cells growing as a monolayer. These structures differed from the ductal structures formed by untreated cells with respect to the localisation of laminin V. MNU-treated cells growing as a monolayer showed cytoplasmic localisation of laminin V (Fig. 5E). Such a pattern of expression is frequently observed in epithelial cancers, especially in cells at the invading front (Giannelli and Antonaci, 2000). This assay also revealed the elongated morphology of cells exposed to MNU, which supports the previous observation of acquiring an EMT-like phenotype (Fig. 5E). These results further confirmed the induction of an EMT-like phenotype upon exposure to an alkylating agent.
To investigate whether MNU-treated cells acquired invasive capabilities, the DQ™ Collagen cell invasion assay was performed. Cells were dissociated from day 16 cultures using Dispase™ and passaged. These cells were seeded on coverslips pre-coated with rat tail collagen type 1 containing DQ collagen type 1. It was observed that MNU-treated cells invaded the collagen bed, which was characterised by enhanced fluorescence (Fig. 5F and G). The conditioned media from the invasion assay was collected and subjected to gelatin zymography analysis, which showed an increase in MMP-9 activity (Fig. 5H). These results demonstrate that treatment of MCF10A breast epithelial cells with 1 mM MNU induced invasive capabilities.
To confirm that MCF10A cells exposed to MNU were transformed, the ability of these cells to form colonies on soft agar, a stringent parameter of transformation, was evaluated. Soft agar transformation assay revealed that MNU-treated cells acquired the ability to form colonies, thus confirming conclusively that the cells had undergone transformation (Fig. 5I,J).
MNU-induced aberrant Golgi morphology in MCF10A cells is through activation of DNA-PKcs
DNA damage has recently been shown to result in dispersal of the Golgi complex through phosphorylation of DNA-PK (Farber-Katz et al., 2014). To investigate whether MNU-induced DNA damage resulted in the activation of DNA-PK under the conditions of our study, MCF10A monolayer cultures were exposed to 1 mM MNU and immunostained for pDNA-PKcs (T2069) (catalytic subunit of DNA-PK complex). As expected, DNA-PKcs foci were seen within 10 min of MNU damage (Fig. 6A). To investigate whether an aberrant Golgi phenotype resulted from activation of DNA-PK, cells were exposed to MNU at different time points and immunostained for the Golgi marker GM130. An aberrant Golgi phenotype was observed 4 h after methylation damage (Fig. 6B; Fig. S3A). Thus, DNA-PK activation precedes the observation of an aberrant Golgi phenotype following MNU damage.
MNU induced Golgi dispersal in MCF10A cells through activation of DNA-PK. MCF10A cells grown as monolayers were exposed to 1 mM MNU. (A) Representative images showing pDNA-PKcs (T2069) (green) following damage at 10 min and 20 min. (B) Representative images for GM130 (green) cis-Golgi marker at 30 min and 4 h after MNU damage. (C–F) MCF10A cells grown as monolayer cultures were treated with MNU followed by 25 µM DMNB for 24 h. (C) Representative images of MCF10A cells immunostained for pDNA-PKcs (T2069) showing a reduction in the number of foci per cell after DMNB treatment. (D) Graph depicting number of pDNA-PKcs foci with and without DMNB following 24 h treatment with 1 mM MNU. (E) Representative images of GM130 (green) with and without 25 µM DMNB showing reversal of aberrant Golgi morphology following inhibition of DNA-PK. (F) Graph representing relative Golgi area measured using ImageJ software. (G) Representative images of GM130-stained MCF10A cells dissociated from day 16 acini, showing aberrant Golgi morphology as well as its reversal following DMNB treatment. (H) Graph representing relative Golgi area measured using ImageJ software. Statistical analysis for (D) was performed using the Mann–Whitney test; one way ANOVA was performed for (F) and (H); ***P<0.0001.
MNU induced Golgi dispersal in MCF10A cells through activation of DNA-PK. MCF10A cells grown as monolayers were exposed to 1 mM MNU. (A) Representative images showing pDNA-PKcs (T2069) (green) following damage at 10 min and 20 min. (B) Representative images for GM130 (green) cis-Golgi marker at 30 min and 4 h after MNU damage. (C–F) MCF10A cells grown as monolayer cultures were treated with MNU followed by 25 µM DMNB for 24 h. (C) Representative images of MCF10A cells immunostained for pDNA-PKcs (T2069) showing a reduction in the number of foci per cell after DMNB treatment. (D) Graph depicting number of pDNA-PKcs foci with and without DMNB following 24 h treatment with 1 mM MNU. (E) Representative images of GM130 (green) with and without 25 µM DMNB showing reversal of aberrant Golgi morphology following inhibition of DNA-PK. (F) Graph representing relative Golgi area measured using ImageJ software. (G) Representative images of GM130-stained MCF10A cells dissociated from day 16 acini, showing aberrant Golgi morphology as well as its reversal following DMNB treatment. (H) Graph representing relative Golgi area measured using ImageJ software. Statistical analysis for (D) was performed using the Mann–Whitney test; one way ANOVA was performed for (F) and (H); ***P<0.0001.
To confirm whether the aberrant Golgi phenotype was mediated via activation of DNA-PK, cells were treated with 25 µM DMNB (a small molecular inhibitor of DNA-PK) following MNU exposure. The aberrant Golgi phenotype was analysed by determining the Golgi area. DMNB, a vanillin derivative, is known to inhibit DNA-PK activity directly (Durant and Karran, 2003). It was observed that inhibiting DNA-PK (Fig. 6C and D) led to reversal of the altered Golgi phenotype, suggesting that DNA-PK was responsible for the MNU-induced aberrant Golgi phenotype (Fig. 6E and F). When cells dissociated from the 3D acinar cultures were treated with 25 µM DMNB, a similar reversal of aberrant Golgi phenotype was observed (Fig. 6G,H).
The presence of cells with altered Golgi morphology has been observed in various cancers (Kellokumpu et al., 2002). The MCF10A cell line series provides a unique platform for study of transformation; the series consists of non-tumorigenic MCF10A cells transformed to represent the various stages in breast cancer progression (Kadota et al., 2010; Imbalzano et al., 2009). To ascertain the pathological relevance of altered Golgi morphology, MCF10A, MCF10AT1 (pre-malignant) and MCF10CA1a (invasive) cells were immunostained with GM130 antibody to mark the Golgi (Fig. S3B). Interestingly, Golgi morphology appeared to be increasingly altered with increasing malignancy state. DNA-PK inhibition led to a reduction in the aberrant Golgi morphology (Fig. S3C,D). Taken together, these results suggest a global role of DNA-PK activation in disrupting Golgi morphology, and possibly in malignancy.
MNU damage-induced transformation via DNA-PK activation is independent of cellular trafficking defect
To test our hypothesis of the involvement of DNA-PK in transformation of MNU-treated MCF10A cells, we used DMNB and examined the effect of DNA-PK inhibition on the changes induced by MNU treatment. In cultures grown as monolayers, we observed a partial rescue of loss of E-cadherin from cell–cell junctions following DNA-PK inhibition (Fig. 7A). To further ascertain whether the MNU-induced EMT-like phenotype was dependent on activation of DNA-PK, the 3D cultures were treated with two consecutive doses of 25 µM DMNB at day 16. As expected, there was rescue of the EMT-like phenotype, as demonstrated by downregulation of mesenchymal markers such as vimentin, fibronectin and N-cadherin (Fig. 7B). We also observed a decrease in invasiveness in the DQ™ Collagen assay upon DNA-PK inhibition. These results suggest that DNA-PK plays a central role in MNU-induced transformation (Fig. 7C and D). Interestingly, we also observed a reduction in the levels of MMP-9 following DNA-PK inhibition (Fig. 7E). In addition, DMNB treatment caused a reduction in the number of colonies formed in the soft agar transformation assay, thus confirming that DNA-PK plays a vital role in the transformation of breast epithelial cells caused by MNU damage (Fig. 7F,G).
MNU-induced transformation is dependent on the activation of DNA-PK. (A) Representative images of E-cadherin showing partial rescue after two consecutive doses of 25 µM DMNB. (B) MNU-treated acini treated with two consecutive doses of 25 µM DMNB at day 16 were lysed and analysed for fibronectin, E-cadherin and vimentin. (C) Representative fluorescence images showing DQ™ collagen invasion assay with and without 25 µM DMNB. (D) Graph representing the relative fluorescence intensity of degraded DQ™ collagen per cell in the presence and absence of DMNB in control and 1 mM MNU-treated cells dissociated from day 16 acinar cultures. (E) Coomassie gel showing gelatinase activity of active MMP-9 in the conditioned media from control and 1 mM MNU-treated cells dissociated from day 16 acinar cultures with and without DMNB treatment. (F) Representative images of soft agar colony formation assay in cells with and without MNU and DMNB treatment. (G) Graph representing the average total number of colonies observed in 20 randomly selected fields per treatment across two independent biological experiments in triplicate. ***P<0.0001.
MNU-induced transformation is dependent on the activation of DNA-PK. (A) Representative images of E-cadherin showing partial rescue after two consecutive doses of 25 µM DMNB. (B) MNU-treated acini treated with two consecutive doses of 25 µM DMNB at day 16 were lysed and analysed for fibronectin, E-cadherin and vimentin. (C) Representative fluorescence images showing DQ™ collagen invasion assay with and without 25 µM DMNB. (D) Graph representing the relative fluorescence intensity of degraded DQ™ collagen per cell in the presence and absence of DMNB in control and 1 mM MNU-treated cells dissociated from day 16 acinar cultures. (E) Coomassie gel showing gelatinase activity of active MMP-9 in the conditioned media from control and 1 mM MNU-treated cells dissociated from day 16 acinar cultures with and without DMNB treatment. (F) Representative images of soft agar colony formation assay in cells with and without MNU and DMNB treatment. (G) Graph representing the average total number of colonies observed in 20 randomly selected fields per treatment across two independent biological experiments in triplicate. ***P<0.0001.
To investigate whether the transformation induced via DNA-PK activation following MNU damage is dependent on trafficking impairment, we performed ts045-vsvg-trafficking assays (Fig. 8A; Fig. S4A) and RUSH assays (Fig. 8B,C; Fig. S4B) in the presence and absence of DMNB. Our results indicate that inhibition of DNA-PK, although it reversed the morphological defect in the Golgi, was unable to reverse the trafficking impairment. Thus, our data indicate that the transformation induced by MNU damage is independent of intracellular trafficking defects.
DNA-PK-mediated transformation is independent of the trafficking defect induced by MNU. MCF10A cells with and without MNU and DMNB treatment were subjected to the ts045-vsvg trafficking assay and RUSH assay and analysed for their trafficking defect. (A) Graph representing the quantification of ts045-vsvg trafficking assay. Values indicate mean±s.e.m. of the relative ratio of surface VSVG molecules (red fluorescence) to total VSVG molecules (green fluorescence). Two-way ANOVA was used to determine the statistical significance of the effect of treatment and time of incubation in permissive temperatures on trafficking of VSVG molecules. P<0.05 is considered statistically significant. (B,C) Graphs representing the quantification of RUSH assay; percentage of cells showing localisation of Str-KDEL-ManII-SBP-EGFP (RUSH construct), in either the ER or Golgi at 10 min (B) and 20 min (C) following biotin addition. Data are pooled from three biologically independent experiments considering n>80 transfected cells. (D) Schematic showing the mechanism of MNU-induced transformation. MNU damage activates DNA-PKcs, which leads to two independent outcomes. One branch leads to activation of DNA-PK, which results in disruption of polarity and induced EMT-like phenotype and invasion in breast epithelial cells grown as 3D cultures. In the other branch, aberrant Golgi morphology follows MNU damage and leads to impaired intracellular trafficking.
DNA-PK-mediated transformation is independent of the trafficking defect induced by MNU. MCF10A cells with and without MNU and DMNB treatment were subjected to the ts045-vsvg trafficking assay and RUSH assay and analysed for their trafficking defect. (A) Graph representing the quantification of ts045-vsvg trafficking assay. Values indicate mean±s.e.m. of the relative ratio of surface VSVG molecules (red fluorescence) to total VSVG molecules (green fluorescence). Two-way ANOVA was used to determine the statistical significance of the effect of treatment and time of incubation in permissive temperatures on trafficking of VSVG molecules. P<0.05 is considered statistically significant. (B,C) Graphs representing the quantification of RUSH assay; percentage of cells showing localisation of Str-KDEL-ManII-SBP-EGFP (RUSH construct), in either the ER or Golgi at 10 min (B) and 20 min (C) following biotin addition. Data are pooled from three biologically independent experiments considering n>80 transfected cells. (D) Schematic showing the mechanism of MNU-induced transformation. MNU damage activates DNA-PKcs, which leads to two independent outcomes. One branch leads to activation of DNA-PK, which results in disruption of polarity and induced EMT-like phenotype and invasion in breast epithelial cells grown as 3D cultures. In the other branch, aberrant Golgi morphology follows MNU damage and leads to impaired intracellular trafficking.
Taken together, these results demonstrate that methylation damage leads to transformation of breast epithelial cells and that DNA-PK plays a central role in this transformation process.
DISCUSSION
DNA damaging agents often work as a double-edged sword in cancer treatment; for successful therapy, it is essential to strike a balance between efficacy and toxicity. In this report, we demonstrate the potential of a methylating agent (MNU) to induce transformation in MCF10A cells (non-tumorigenic mammary epithelial cells) through activation of DNA-PK, which in turn disrupts Golgi structure and function.
A plethora of reports in the literature have reported the gene expression profiles of human cancer samples, where various proteins in the DNA damage response pathway are deregulated (Albino et al., 2011; Leongamornlert et al., 2014; Goode et al., 2002; Lahtz and Pfeifer, 2011). A few studies have also attempted to understand the role of proteins of the DNA damage response pathway in the process of transformation (Lim et al., 2009; Sun et al., 2012; Russell et al., 2015; Pu et al., 2014; Zhang et al., 2014). However, only a very small number of studies have tried to understand the role of these proteins in the initiation of cancer. In the present study, we have attempted to delineate the process of DNA damage-induced transformation in breast epithelial cells.
Treatment of MCF10A cells with a non-lethal dose of MNU resulted in the formation of unpolarised spheroids characterised by mis-localisation of α6-integrin and pERM, and disruption of cell–cell junctions. Interestingly, the morphology of the Golgi appeared to be disrupted. The aberrant Golgi morphology observed in MNU-treated cells shares similarities with that reported by Farber-Katz et al. (2014). Furthermore, we also found that aberrant Golgi morphology occurred via phosphorylation of DNA-PKcs and that the vesicles formed showed impaired trafficking between Golgi and the plasma membrane. Because the Golgi is one of the central organelles regulating polarised sorting of proteins, it appears that the phenomenon of aberrant Golgi morphology could result in polarity disruption following DNA damage. In addition to this, we observed impaired ER to Golgi trafficking following DNA damage. The disruption of intracellular trafficking could explain the cytoplasmic localisation of laminin V observed in the collagen:Matrigel assay, the reduction of E-cadherin localisation at the cell–cell junctions and abnormal β-catenin localisation.
The α6 and β4 integrins are known to mediate cell–ECM interactions and dictate polarity (Vidi et al., 2012). Furthermore, Akhtar and Streuli have shown that loss of cell–ECM interactions (knockdown of β1 integrin) affected Golgi positioning and caused disruption of Golgi morphology, which is comparable to the loss of α6 integrin (Akhtar and Streuli, 2013) and disruption of Golgi structure observed in our present study. In addition, previously published reports indicate the controversial role of α6 integrin in cancer metastasis. The α6β4 subunits of integrin have been found to play a vital role in the formation of hemidesmosomes in epithelial cells (Sonnenberg et al., 1991). These subunits have been shown to promote tumour metastasis in prostate and pancreatic cancers (Rabinovitz et al., 1995; Cruz-Monserrate and O'Connor, 2008). The downregulation of these integrins has been observed in prostate cancer (Davis et al., 2001) and some instances of breast cancer, where loss of α6β4 integrin was seen in ductal carcinoma of the breast (grade III) (Pignatelli et al., 1992).
The loss phenotype is seen in cells that metastasise to the parenchyma and pleural cavity (Natali et al., 1992). In addition, cells are known to break down their basement membrane to migrate and invade nearby tissue (Vidi et al., 2013). We observed a loss phenotype for the basement membrane marker laminin V, thus explaining the possibility that these cells gained invasive potential, which was corroborated through the DQ™ Collagen invasion assay and the presence of active MMP-9 in the conditioned media. One of the important characteristics of transformation is the ability of cells to form colonies when grown in anchorage-independent conditions. MNU-treated cells formed colonies in soft agar, indicative of transformation.
In a quest to elucidate the mechanism by which MNU damage induces transformation, a notable observation in our study was the activation of DNA-PKcs. Similarly to Farber-Katz et al. (2014), we observed the presence of aberrant Golgi morphology even up to day 30, which was reversed upon treatment with a DNA-PK inhibitor. This implies that there is continued activation of DNA-PKcs following methylation damage. This also suggests that activation of DNA-PKcs following methylation damage orchestrated an aberrant Golgi morphology and impaired trafficking, resulting in loss of basolateral polarity. This loss of polarity, especially basal polarity, could reduce the efficiency of DSB repair (Vidi et al., 2012). Unrepaired DNA can result in continued activation of DNA-PKcs; alternatively, when such cells re-enter the cell cycle they accumulate chromosomal alterations that can result in carcinogenesis (Asaithamby et al., 2011). The increase in nuclear size observed could thus imply changes in the nuclear content resulting from mis-segregation of chromosomes during cell division. This is also supported by published reports showing that MNU induced sister chromatid exchange (Neft and Conner, 1989; Bouffler et al., 2000; Neft et al., 1989). However, this warrants further investigation. Detailed metaphase spread analysis or spectral karyotyping (SKY) might help in understanding this intriguing question.
The above argument raises the possibility that DNA-PKcs plays a vital role in the process of transformation. Partial reversal of polarity as well as reduction in the levels of mesenchymal markers upon DNA-PK inhibition confirm the role played by DNA-PK in polarity disruption and, hence, transformation. Furthermore, DMNB treatment also rescued the loss of E-cadherin phenotype. These observations are supported by the reduction in invasiveness of MNU-treated cells following treatment with DNA-PK and by reports describing upregulation of DNA-PKcs in many patient samples (Goodwin et al., 2015). The same group also identified DNA-PKcs as a contributor to tumour metastasis (Goodwin et al., 2015). DNA-PKcs is also known to interact with Snail1 (one of the inducers of EMT), stabilise it and induce its functional activity (Pyun et al., 2013). Activation of DNA-PKcs, coupled with upregulation of Snail1 observed in the present study, corroborates available reports. This interaction between DNA-PKcs and Snail1 can be hypothesised to play a vital role in the induction of EMT observed in our study, characterised by attainment of elongated morphology (in collagen:Matrigel matrix), upregulation of mesenchymal markers, loss of polarity and disruption of cell–cell junctions. However, the inability of DNA-PK inhibition to restore Golgi trafficking and the ability to reverse transformation phenotypes implies that DNA-PK-mediated transformation is independent of the trafficking defect induced by MNU.
We have established a model for DNA damage-induced transformation and demonstrated that the transformation occurred via DNA-PK activation, but was independent of impaired cellular trafficking (Fig. 8D). Although this mechanism is responsible for polarity disruption and thus transformation, the involvement of other parallel pathways cannot be overruled. It will be of great value to explore other pathways to help complete our understanding of the process of transformation induced by alkylation damage. In addition, this model can be exploited to understand how early transformation occurs and identify the initial set of gene(s) that are deregulated in the process of transformation. Understanding this phenomena and recognition of key candidate genes as biomarkers could help in the design of novel therapeutic strategies.
MATERIALS AND METHODS
Cell lines and culture conditions
The MCF10A cell line was a generous gift from Prof. Raymond C. Stevens (The Scripps Research Institute, La Jolla, CA). The cell line was authenticated by DNA typing carried out by DDC medical (Fairfield, OH). Cells were tested for mycoplasma contamination periodically using Hoechst 33342 and MycoAlert™ Mycoplasma Detection Kit (Lonza). These cells were grown in High Glucose DMEM without sodium pyruvate (Lonza) containing 5% horse serum (Invitrogen), 20 ng/ml EGF (Sigma-Aldrich), 0.5 µg/ml hydrocortisone (Sigma-Aldrich), 100 ng/ml cholera toxin (Sigma-Aldrich), 10 µg/ml insulin (Sigma-Aldrich) and 100 units/ml penicillin-streptomycin (Invitrogen). Cells were resuspended during sub-culturing in High Glucose DMEM without sodium pyruvate, containing 20% horse serum and 100 units/ml penicillin-streptomycin (Invitrogen). The overlay medium in which the cell suspension was made for seeding (assay medium) contained DMEM without sodium pyruvate, horse serum, hydrocortisone, cholera toxin, insulin, EGF and penicillin-streptomycin. Opti-MEM® used for transfection was obtained from Invitrogen. Cells were maintained in 100 mm tissue-culture treated dishes (Corning, Sigma-Aldrich) at 37°C in a humidified 5% CO2 incubator (Thermo Scientific). The 3D ‘on-top’ cultures were seeded in an 8-well chamber cover glass obtained from Nunc Lab-tek, (Thermo Scientific) or in 6-well and 12-well dishes (Corning, Sigma-Aldrich) and grown on Matrigel® Basement Membrane Matrix obtained from Corning, Sigma-Aldrich.
Chemicals and antibodies
N-nitroso-N-methylurea (MNU), dimethyl sulfoxide (DMSO), thiazolyl blue tetrazolium bromide (MTT), gelatin, biotin and cyclohexamide were purchased from Sigma-Aldrich. The DNA PK inhibitor 4,5-dimethoxy-2-nitrobenzaldehyde (DMNB), was purchased from Tocris Bioscience. Paraformaldehyde (16% w/v aqueous solution) used for immunostaining was bought from Alfa Aeser. DQ™ Collagen type1 was purchased from Thermo Fisher Scientific. Rat tail collagen type 1 came from BD Biosciences. Monoclonal antibodies used for immunofluorescence assay for laminin-V (MAB19562) and α6-integrin (MAB1378) were bought from Millipore; α-tubulin (T6199) and GAPDH (G9545) antibodies were bought from Sigma-Aldrich.
Monoclonal antibodies for immunofluorescence for vimentin (ab92547), β-catenin (ab32572), E-cadherin (ab1416), anti-integrin α3 antibody (ab11767) and DNA-PKcs (phospho T2609, ab18356) were obtained from Abcam; that for hDlg (sc-9961) was obtained from Santa Cruz Biotechnology. For western blotting, monoclonal antibodies for vimentin (ab8069), N-cadherin (ab12221), Cytokeratin 14 (ab7800) and Cytokeratin 19 (ab52625) were purchased from Abcam, along with polyclonal antibody for Snail (ab180714). GM130 monoclonal antibody (610822) for immunofluorescence studies and fibronectin monoclonal antibody (610077) for Western studies were obtained from BD Biosciences. Polyclonal antibody for pAKT (S437) (9271S) was purchased from Cell Signaling, whereas that for c-Myc (SC 40) was purchased from Santa Cruz Biotechnology. Peroxidase-conjugated AffiniPure goat anti-mouse and anti-rabbit, as well as AffiniPure F(ab′)2fragment goat anti-mouse IgG, F(ab′)2 fragment specific, were obtained from Jackson ImmunoResearch. Hoechst 33342, Concanavalin A, Alexa Fluor 488 conjugate, Alexa Fluor 568 phalloidin, Alexa Fluor 488 and Alexa Fluor 568 were bought from Invitrogen. Anti-VSVG (8G5F11) antibody (EB0010) purchased from Kerafast was used to select the extracellular domain of VSVG glycoprotein in the trafficking assay.
3D ‘on-top’ culture
The 3D ‘on-top’ culture was set up in an 8-well chamber cover glass (Nunc Lab-tek, Thermo Scientific), 6-well dishes and 12-well dishes (Corning, Sigma-Aldrich) using standard protocols (Debnath et al., 2003; Anandi et al., 2016; Bodakuntla et al., 2014). Cultures were allowed to grow for 16 days at 37°C in a humidified 5% CO2 incubator (Thermo Scientific). Medium was supplemented every 4 days. Drug treatment of MNU was given on day 0 and day 2; the day of seeding was considered as day 0.
For dissociating cells from spheroids, the cultures were incubated with 0.05% trypsin supplemented with 100 mM EDTA for 30 min. To aid in dislodging, the Matrigel layer was scraped partially with the tip. Trypsin was quenched using medium and the cells were plated in tissue culture dishes and grown at 37°C in a humidified 5% CO2 incubator.
Immunofluorescence
The MCF10A 3D cultures grown on the Matrigel-coated chamber cover glass were fixed with 200 µl of 4% paraformaldehyde for 20 min at room temperature. For immunostaining of apical markers and cell–cell junction markers, the 3D cultures were incubated with PBS-EDTA for 15 min at 4°C and then fixed as described above. Cells were permeabilised with pre-chilled 0.5% Triton X-100 in PBS for 10 min at 4°C and then rinsed with PBS-glycine three times, for 10 min each time. This was followed by a primary block using 10% goat serum in immunofluorescence (IF) buffer for 1 h. This was followed by a secondary block (10% goat serum in IF buffer plus 1% secondary blocking antibody; F(ab′)2 fragment goat anti-mouse IgG, F(ab′)2 fragment specific antibody) for 1 h. The cells were further incubated with primary antibody (1:100 dilution) at 4°C overnight. Following primary antibody incubation, cells were washed with IF buffer three times, 20 min each time, and then incubated with secondary antibody (1:200 dilution) for 1 h in the dark. The slides were rinsed with IF buffer once for 20 min, followed by two or three washes with PBS. Incubation with Hoechst 33342 was done for 5 min followed by two PBS washes of 10 min each. Mounting was done with Slow fade Gold Anti-fade reagent (Invitrogen) and images captured in a LSM710 laser scanning confocal microscope (Carl Zeiss) using a 63× oil objective. At least 50 acini were analysed per treatment group to determine the phenotype induced by treatment.
Comet assay
Comet assay was performed using standard protocols mentioned earlier (Bodakuntla et al., 2014), with a few modifications. Briefly, MCF10A cells were seeded at a density of 1.5×105 cells per well on Matrigel®-coated 12-well dishes. MNU was added to the media 2 h following seeding. At 2 h and 24 h following damage, 3D cultures were harvested by trypsinising at 37°C for 45 min. The cell suspension was collected in microcentrifuge tubes (1.5 ml) and centrifuged at 900 rpm for 5 min at 4°C. The supernatant was discarded (leaving behind about 100 µl of solution), mixed well and used for subsequent steps. Alkaline and Neutral Comet assays were performed according to standard protocols (Olive and Banáth, 2006).
To calculate the extent of DNA damage, the percentage tail length and relative DNA content were calculated using the Comet assay plugin in ImageJ software.
ts045-VSVG-GFP trafficking assay
MCF10A cells seeded in an 8-well chamber cover glass at a density of 5×104 cells per well were incubated for 16 h. The cells were treated with either DMSO or 1 mM MNU for 24 h. For rescue experiments, 25 µM DMNB was added twice for 12 h each. This was followed by transfection with pCDM8.1/VSVGts045-GFP construct (a kind gift from Dr Jennifer Lippincott-Schwartz, NIH, Bethesda, USA) using Lipofectamine 2000 in OptiMEM I Reduced Serum Medium (Invitrogen). The transfected cells were then incubated at 37°C for 4 h. Following this, the cells were re-fed with growth medium and incubated at 40°C for 16 h. Cyclohexamide (100 µg/ml) was added for 30 min. Cells were then shifted to a permissive temperature of 32°C and fixed after 0, 40, 80 and 120 min. Immunostaining of the surface VSVG molecules was done under non-permeabilisation conditions. The cells were fixed with 2% paraformaldehyde (PFA) for 15 min at 4°C, followed by washing with wash buffer and incubation with anti-VSVG (8G5F11) antibody (1:1000 dilution) for 1 h at 4°C. After extensive washing, cells were incubated with Alexa Fluor 568 conjugated goat anti-mouse antibody (1:100 dilution) at 4°C for 1 h. After washing, mounting was done using Slowfade Gold antifade reagent (Invitrogen). Images were captured with a laser scanning confocal microscope LSM710 (Carl Zeiss) using a 63× objective.
To measure the intensity of cells expressing VSVG molecules, ImageJ software was used. Cell boundaries were marked manually and fluorescence measured. Mean cellular fluorescence of a non-transfected cell, in the same field, was measured to account for background fluorescence. To calculate the corrected total cell fluorescence (CTCF), the formula used was CTCF=integrated density − (area of selected cell×mean fluorescence of background readings).
Retention using selective hook assay
MCF10A cells were seeded at a density of 4×105 cells per well. Following 1 mM MNU treatment for 24 h, the cells were transfected with ManII-SBP-EGFP construct (generous gift from Dr Franck Perez, Institut Curie, Paris, France) using transfectene reagent (Qiagen). For rescue experiments, 25 µM DMNB was added following MNU treatment, for 12 h each. After 18 h of transfection, 40 µM biotin was added and cells fixed at 0, 10 and 20 min following addition and visualised using the 63× objective of a laser scanning confocal microscope LSM710 (Carl Zeiss). The percentage of cells showing fluorescence reporter diffused in either the cytoplasm or Golgi, or both, were calculated and represented as bar graphs.
Concanavalin A labelling assay
MCF10A cells were seeded on coverslips at a cell density of 5×104 per coverslip. Following incubation at 37°C for 16 h, the cells were treated with MNU and incubated for 24 h. This was followed by washing the coverslips with PBS and incubating the cells with 1 µg/µl of Concanavalin A-Alexa Fluor 488 conjugate for 15 min at 4°C in the dark. The coverslips were then rinsed with PBS and fixed using 3.5% paraformaldehyde for 15 min at room temperature. The nuclei were counter-stained with Hoechst 33342 and mounted using Slowfade Gold antifade reagent (Invitrogen). Images were captured in a LSM710 laser scanning confocal microscope (Carl Zeiss) using a 63× oil objective.
DQ collagen invasion assay
Cells were dissociated from day 16 cultures of MCF10A cells treated with or without MNU and DMNB. Cells (2×104) were seeded in an 8-well chamber cover glass coated with 1.6 mg/ml rat tail collagen type 1 containing 25 mg/ml of DQ collagen type 1 (Invitrogen). At 48 h following incubation, the cells were imaged at 40× magnification using an Olympus IX81 system equipped with a Hamamatsu ORCA-R2 CCD camera. Fluorescence intensity per cell was measured using ImageJ. The corrected total cellular fluorescence was calculated as described in the section ‘ts045-VSVG-GFP trafficking assay’.
Soft agar assay
Cells (1×105) were mixed in 0.3% agar in DMEM and plated over 0.6% agar as the base layer. Growth medium (750 µl) was overlaid on the agar layers. The cells were re-fed with media every 4 days and maintained for about 3 weeks. The cells were stained by adding 200 µl of 1 mg/ml MTT followed by incubation for 90 min. Images were acquired using the 4× objective of a Nikon phase contrast microscope. Ten randomly selected fields were imaged and colonies counted manually.
Statistical analysis
The Student's t-test was used to analyse the statistical significance of tail length and the one-tailed Mann–Whitney U-test was used for percentage tail DNA. The Mann–Whitney U-test was also used to analyse the significance of difference between parameters in morphometric analysis of 3D cultures. Data from VSVG trafficking assays were analysed using two-way ANOVA to determine the significance of the effect of treatment and time on trafficking. The Mann–Whitney U-test was used to analyse the statistical significance of changes in relative Golgi area and relative fluorescence intensity in the DQ™ Collagen invasion assay. P<0.05 was considered statistically significant. Graph Pad Prism software (Graph Pad Software, La Jolla, CA, USA) was used to analyse data.
Acknowledgements
We thank Dr Annapoorni Rangarajan (Indian Institute of Science, Bangalore, India), Drs Richa Rikhy, Aurnab Ghose and Thomas Pucadyil (IISER Pune, India), and Dr Carsten Janke (Institut Curie, Orsay, France) for their useful suggestions. We also thank Dr Franck Perez (Institut Curie, Paris, France) for the RUSH plasmids. We appreciate Dr Nagaraj Balasubramanium (IISER Pune, India) for the Concanavalin A conjugate. We also thank Lahiri laboratory members for helpful comments and discussions. The authors greatly acknowledge help from Vijay Vittal at the IISER, Pune Microscopy facility.
Footnotes
Author contributions
Conceptualization: L.A., M.L.; Methodology: L.A., M.L.; Validation: L.A., V.C.; Formal analysis: L.A., V.A., K.A.A.; Investigation: L.A., V.C., K.A.A.; Writing - original draft: L.A., V.C., M.L.; Writing - review & editing: L.A., V.C., K.A.A., M.L.; Visualization: L.A., M.L.; Supervision: L.A., M.L.; Project administration: M.L.; Funding acquisition: M.L.
Funding
This study is supported by a grant from Department of Biotechnology, Ministry of Science and Technology (DBT) (BT/PR8699/MED/30/1018/2013) and partly by Indian Institute of Science Education and Research Pune Core funding. L.A. is funded through Department of Science and Technology, Ministry of Science and Technology (DST)-INSPIRE fellowship, V.C. was funded by DBT (BT/PR8699/MED/30/1018/2013). K.A.A. was funded by DST-INSPIRE scholarship and partly by DBT (BT/PR8699/MED/30/1018/2013). S.B. was funded partly by IISER, Pune Core funds and partly through DBT (BT/PR3156/BRB/10/961/2011).
References
Competing interests
The authors declare no competing or financial interests.