Brk/PTK6 and Involucrin Expression May Predict Breast Cancer Cell Responses to Vitamin D3

The process of human embryonic mammary development gives rise to the structures in which mammary cells share a developmental lineage with skin epithelial cells such as keratinocytes. As some breast carcinomas have previously been shown to express high levels of involucrin, a marker of keratinocyte differentiation, we hypothesised that some breast tumours may de-differentiate to a keratinocyte-derived ‘evolutionary history’. To confirm our hypothesis, we investigated the frequency of involucrin expression along with that of Brk, a tyrosine kinase expressed in up to 86% of breast carcinomas whose normal expression patterns are restricted to differentiating epithelial cells, most notably those in the skin (keratinocytes) and the gastrointestinal tract. We found that involucrin, a keratinocyte differentiation marker, was expressed in a high proportion (78%) of breast carcinoma samples and cell lines. Interestingly, tumour samples found to express high levels of involucrin were also shown to express Brk. 1,25-dihydroxyvitamin D3, a known differentiation agent and potential anti-cancer agent, decreased proliferation in the breast cancer cell lines that expressed both involucrin and Brk, whereas the Brk/involucrin negative cell lines tested were less susceptible. In addition, responses to 1,25-dihydroxyvitamin D3 were not correlated with vitamin D receptor expression. These data contribute to the growing body of evidence suggesting that cellular responses to 1,25-dihydroxyvitamin D3 are potentially independent of vitamin D receptor status and provide an insight into potential markers, such as Brk and/or involucrin that could predict therapeutic responses to 1,25-dihydroxyvitamin D3.


Introduction
The Brk tyrosine kinase has been described as a potential therapeutic target for breast cancer treatment since it influences major signalling pathways implicated in tumour development [1]. Previous studies have shown that Brk expression increases anchorageindependent growth and survival [2][3][4][5] and that its suppression by RNA interference reduces the rate of breast cancer cell proliferation in vitro [6]. Brk is expressed in a majority of breast carcinomas [7][8][9]; however, its expression in normal tissues is relatively restricted. Normal expression is limited primarily to cells in differentiating epithelia such as those in the gastrointestinal (GI) tract, where it is present within the non-dividing villus epithelium of the small intestine, as well as in the crypt cells post-irradiation [10][11][12]. Brk expression has also been detected in the nuclei of normal luminal prostate epithelial cells and oral epithelia [13,14].
Brk expression has previously been reported to be low/undetectable in normal human mammary tissue samples, and studies identifying the mouse orthologue, Sik, show that it is not detected at any stage of murine embryonic mammary development [10], although more recent data suggest that Brk is expressed in normal human mammary cells [15].
In human skin, Brk expression is largely restricted to suprabasal keratinocytes in the differentiating layers [16]. Several in vitro studies suggest that Brk/Sik is involved in calcium-induced keratinocyte differentiation, which is characterised by increased expression of epidermal differentiation markers such as Keratin10 or Filaggrin [16][17][18].
The first signs of human mammary gland formation begin early in gestation. They are characterised by the thickening of the ectoderm to form milk streaks, or lines, which then fragment. The thoracic region thickens and widens, while the anterior region begins to involute. A solid bud of cells gives rise to the discrete structures that will form the mammary tissue. The primordium begins to grow down into the underlying mesenchyme and the skin ceases to protrude. Mammary capillaries then start to form and the edge of the anlage forms outgrowths. These outgrowths continue into the mesenchyme and branch to form the epithelial cords that, in turn, give rise to the mammary ducts [19,20].
Ultrastructural studies carried out on week 13 human embryos have shown that, although there was no sign of myoepithelial differentiation in the mammary basal cell layer, cells present at the epithelial-stromal junction sustained the hemidesmosomes and keratin aggregates of skin basal cells. Thus, the embryonic ectoderm gives rise to the epithelia of both the skin and the mammary gland, meaning there are likely to be shared aspects of development.
A high proportion of human breast carcinomas express Brk; however, not all the underlying regulatory mechanisms are known. The expression of Brk in normal skin, coupled with the embryonically-related origins of skin and mammary epithelial cells, has led us to hypothesise that some mammary carcinomas might retain an evolutionary history characteristic of skin epithelial cells. This suggests that the development of some breast tumours might reflect an aberrant differentiation process.
Involucrin is a cytoplasmic protein that becomes cross-linked in terminally differentiating keratinocytes [21,22] enabling its expression to be used as a marker of terminal keratinocyte differentiation. We have previously shown a link between Brk and keratinocyte differentiation markers [17]. This, combined with other studies reporting that around a quarter of breast tumours express involucrin [23,24], implies a possible link between differentiation processes and breast cancer.
We investigated whether Brk and involucrin were co-expressed in breast tumours. In this way, agents that are known to induce differentiation in keratinocytes (such as vitamin D3) may be shown to be more effective in Brk-positive breast cancer cells compared with Brk-negative cells, potentially identifying alternative therapy options.

Results
To test the hypothesis that Brk-positive breast tumour cells have a partial keratinocytelike phenotype, whole cell lysates generated from eight breast cancer cell lines were analysed by Western blotting for the presence of Brk and involucrin ( Figure 1). The cell lines were selected for their known Brk status; four were positive and four were negative, based on previous findings [4,7]. The Brk-positive cell lines, T-47D, MDA-MB-361, GI101 and SkBr3 had detectable levels of involucrin protein. The four Brk-negative cell lines, MDA-MB157, MDA-MB468, Cal51 and PMC42, showed no detectable involucrin expression. These results indicate that Brk and involucrin may be co-expressed in breast cancer cell lines. To establish whether this finding was an artefact of cultured cell lines and, therefore, not representative of tumours in vivo, quantitative RT-PCR was carried out using mRNA extracted from 44 patient surgical tumour samples of known grade as well as 7 normal breast tissue samples ( Figure 2). Surprisingly, we were able to detect involucrin expression in our samples. The normal mammary samples were found to express the lowest levels of involucrin mRNA. All 46 tumour samples assessed were initially positive for involucrin expression; however, when we subtracted the median expression level of the normal samples, 36 out of the 46 (78%) tumour samples analysed expressed elevated levels of involucrin compared to normal tissues. Median relative expressions levels were 0.296 (for normal tissues) and 1.658, 1.335 and 1.351 in samples from Grades 1, 2 and 3 carcinomas (respectively), indicating that a small number of outliers with high expression (Figure 2) are skewing the overall mathematical mean expression levels, although the differences between grades did not reach statistical significance. The 17 tumour samples that expressed the highest levels of involucrin (ten-fold or more than the median level expressed by the normal samples), were all found to be positive for Brk mRNA expression. Five out of the 6 samples (83%) that expressed between five-and ten-fold more involucrin than the normal samples were also Brk positive (Table  1). Brk expression in these tumours has been analysed previously and shown to correlate with tumour grade [4]. While there is no direct correlation between absolute levels of Brk and involucrin mRNA expression, it would appear that Brk is expressed in higher-grade involucrin-positive tumours. These data lend support to the hypothesis that some Brkpositive breast cancers may share features of differentiating keratinocytes and, while not co-dependent, Brk and involucrin expression may be co-regulated in some cases. The 17 tumour samples that expressed the highest levels of involucrin (ten-fold or more than the median level expressed by the normal samples), were all found to be positive for Brk mRNA expression. Five out of the 6 samples (83%) that expressed between fiveand ten-fold more involucrin than the normal samples were also Brk positive (Table 1). Brk expression in these tumours has been analysed previously and shown to correlate with tumour grade [4]. While there is no direct correlation between absolute levels of Brk and involucrin mRNA expression, it would appear that Brk is expressed in higher-grade involucrin-positive tumours. These data lend support to the hypothesis that some Brkpositive breast cancers may share features of differentiating keratinocytes and, while not co-dependent, Brk and involucrin expression may be co-regulated in some cases. Keratinocytes are known to differentiate when cultured in suspension [25,26], and both Brk and involucrin are co-regulated during keratinocyte differentiation [17]. Therefore, the effect of culturing breast cancer cells in suspension on Brk and involucrin expression was examined. T-47D and MDA-MB-361 were selected based on their medium Brk and involucrin expression levels (compared to the other cell lines shown in Figure 1). This meant that any changes in protein levels (up or down) could be visualised. Western blotting of whole cell lysates showed that involucrin levels were increased in T-47D and decreased in MDA-MB-361 after 7 days of growth in suspension in polyHEMA-coated wells compared with adherent cells grown as a monolayer ( Figure 3A). The levels of Brk expression also differed between the adherent and suspension cultures and mirrored involucrin levels, again suggesting that Brk and involucrin may be co-regulated. Interestingly, the Brk/involucrin positive cells also exhibited characteristics of the primary keratinocytes by forming dense spheres in suspension culture that could not be disaggregated upon pipetting. Conversely, the Brk-negative MDA-MB-468 cells remained as single cells in suspension ( Figure 3B). If Brk expression is related to differentiation capacity, especially to a keratino like phenotype, we hypothesised that Brk-positive breast cancer cells may be suscep to growth inhibition by keratinocyte-differentiating agents such as vitamin D3. Ther 1,25-dihydroxyvitamin D3 (calcitriol), the biologically active metabolite of vitamin was tested on cultures of breast cancer cell lines. Cells were treated with 1,25droxyvitamin D3 and counted after 72 h. The most consistent responses were observ the cell lines that were positive for both Brk and involucrin expression ( Figure 4A). If Brk expression is related to differentiation capacity, especially to a keratinocyte-like phenotype, we hypothesised that Brk-positive breast cancer cells may be susceptible to growth inhibition by keratinocyte-differentiating agents such as vitamin D3. Therefore 1,25-dihydroxyvitamin D3 (calcitriol), the biologically active metabolite of vitamin D3, was tested on cultures of breast cancer cell lines. Cells were treated with 1,25-dihydroxyvitamin D3 and counted after 72 h. The most consistent responses were observed in the cell lines that were positive for both Brk and involucrin expression ( Figure 4A).
The fold increase in cell number over the 72 h period was also determined (and expressed as a percentage of the vehicle control) to compare the effects between the different cell lines more clearly. The four Brk/involucrin-positive cell lines showed smaller increases in total cell number compared with the four Brk/involucrin-negative cell lines ( Figure 4B). Interestingly, the cell line with the highest involucrin expression (Sk-Br-3) (Figure 1) had the most marked response to 1,25-dihydroxyvitamin D3 treatment, suggesting that involucrin might be a greater influencing factor in Vitamin D responses than Brk. The decreases in viable cell numbers in the Brk/involucrin positive cells compared with the untreated con-trols were not accompanied by any consistent alterations in the cell cycle profile ( Figure 5). There was not a consistent increase in cell death across all 4 cell lines tested in response to 1,25-dihydroxyvitamin D3 (determined by flow cytometry and based on the absence of an increased sub-G0/G1 peak).    Finally, we examined whether responses to 1,25-dihydroxyvitamin D3 correlated with vitamin D eeceptor (VDR) expression. There was no direct correlation between cell responsiveness to 1,25-dihydroxyvitamin D3 and VDR levels ( Figure 6). Figure 6. VDR expression does not predict response to 1,25-dihydroxyvitamin D3. Cell lysates were prepared and analysed for VDR expression by Western blotting. The lysates were loaded onto the gel in order of 1,25-dihydroxyvitamin D3 responsiveness (based on the data presented in Figure 4).

Discussion
Mammary tissue is formed from invaginations of the ectoderm and, as differentiating skin cells express Brk, this study attempted to address whether Brk-positive breast tumours had retained/regained some of this previous developmental keratinocyte-like differentiation capacity.
In the human breast cancer cell lines we studied, all of the Brk-positive cultures also expressed detectable levels of involucrin. A total of 77% of the clinical cancer samples were involucrin positive, a much higher proportion than previously reported [23,24], Finally, we examined whether responses to 1,25-dihydroxyvitamin D3 correlated with vitamin D eeceptor (VDR) expression. There was no direct correlation between cell responsiveness to 1,25-dihydroxyvitamin D3 and VDR levels ( Figure 6). Finally, we examined whether responses to 1,25-dihydroxyvitamin D3 correlated with vitamin D eeceptor (VDR) expression. There was no direct correlation between cell responsiveness to 1,25-dihydroxyvitamin D3 and VDR levels ( Figure 6). Figure 6. VDR expression does not predict response to 1,25-dihydroxyvitamin D3. Cell lysates were prepared and analysed for VDR expression by Western blotting. The lysates were loaded onto the gel in order of 1,25-dihydroxyvitamin D3 responsiveness (based on the data presented in Figure 4).

Discussion
Mammary tissue is formed from invaginations of the ectoderm and, as differentiating skin cells express Brk, this study attempted to address whether Brk-positive breast tumours had retained/regained some of this previous developmental keratinocyte-like differentiation capacity.
In the human breast cancer cell lines we studied, all of the Brk-positive cultures also expressed detectable levels of involucrin. A total of 77% of the clinical cancer samples were involucrin positive, a much higher proportion than previously reported [23,24], Figure 6. VDR expression does not predict response to 1,25-dihydroxyvitamin D3. Cell lysates were prepared and analysed for VDR expression by Western blotting. The lysates were loaded onto the gel in order of 1,25-dihydroxyvitamin D3 responsiveness (based on the data presented in Figure 4).

Discussion
Mammary tissue is formed from invaginations of the ectoderm and, as differentiating skin cells express Brk, this study attempted to address whether Brk-positive breast tumours had retained/regained some of this previous developmental keratinocyte-like differentiation capacity.
In the human breast cancer cell lines we studied, all of the Brk-positive cultures also expressed detectable levels of involucrin. A total of 77% of the clinical cancer samples were involucrin positive, a much higher proportion than previously reported [23,24], possibly due to the increased sensitivity of RT-PCR in comparison with immunohistochemistry, although it remains a possibility that a subset of tumours expresses involucrin mRNA that is not translated. Tsuda and colleagues showed that involucrin expression correlated with the presence of squamoid features and noted that involucrin expression was associated with higher-grade atypia [24].
In this study, we have shown that involucrin expression can be detected in all tumour grades and that expression levels increased with tumour grade (although the increases did not reach statistical significance Student's t-test p > 0.05). In addition, all the tumours that highly overexpressed involucrin also expressed Brk, suggesting that some Brk-positive breast cancers may have a more differentiated keratinocyte-like (or squamoid) phenotype at a cellular level.
Brk expression increases with tumour grade and we have previously suggested that Brk may function in the development of metastatic disease by increasing cell survival in suspension. This would allow tumour cells to survive during dissemination [4]. When keratinocytes begin to differentiate, they migrate away from the basal layer and basement membrane, thereby losing contact with some extracellular matrix proteins. It is possible that acquiring Brk expression during the differentiation process protects cells from cell death on the loss of anchorage and facilitates cell survival until terminal differentiation and cornification commence.
Treatment of keratinocytes with vitamin D or its analogues results in suppressed proliferation and induced differentiation [27,28]. Evidence is accumulating to suggest that vitamin D3 derivatives may have potential benefits as tumour therapies [29], and some agents have been shown to be tolerated and have activity in clinical trials [30,31].
1,25-dihydroxyvitamin D3, the biologically active and far more potent metabolite of vitamin D3 [32], has previously been shown to decrease proliferation [33][34][35] and cause cell cycle arrest at the G1/S-phase transition [36,37]. Along with its analogues, it induces a more differentiated phenotype as determined by lipid droplet formation in MCF7 and MDA-MD-231 breast cancer cell lines at concentrations between 10 −10 and 10 −6 M [38]. In our study, the presence of Brk and involucrin appeared to correlate with whether 1,25-dihydroxyvitamin D3 was able to decrease cell proliferation. The Brk-positive cell lines showed reduced cell numbers relative to controls, whereas the Brk-negative breast cancer cell lines were less affected by the doses used (range of 10 nM to 10 µM). Three of the Brk-negative cell lines studied (MDA-MB-468, MDA-MB-157 and CAL-51) were unaffected by these concentrations of 1,25-dihydroxyvitamin D3. Whilst the presence of Brk and involucrin may influence responses to 1,25-dihydroxyvitamin D3, the most marked response was in the Sk-Br-3 cell line which had the most involucrin, suggesting that involucrin levels may play a greater role than those of Brk. Dissecting this will be a focus of ongoing work.
Previous studies have shown that vitamin D3 analogues have more potent antiproliferative effects in transformed cells compared with their non-transformed counterparts and that hormone sensitivity may play a role. Proliferation and cell cycle progression of untransformed MCF12F cells were unaffected by vitamin D3 analogues; however, the transformed sub-lines MCF12F-DMBA and -NMU were sensitive to both 1,25-dihydroxyvitamin D3 and its analogue 1α-hydroxyvitamin D5 [34]. This study also reported that the analogue had no effect on the oestrogen receptor (ER) negative cell line MDA-MB-231, but did in ER-positive cell lines BT474 and MCF7 [34], despite the fact that MDA-MB-231 is positive for the vitamin D receptor (VDR) [35]. Our data also support the notion that there are cell responses to vitamins that are independent of VDR expression (Table 2).
More recent evidence suggests that VDR expression is linked to that of the ER [42] and that vitamin D responses may also be mediated independently of the VDR in ER-positive cell lines [33]. Given that both MDA-MB-468 (non-responsive to 1,25-dihydroxyvitamin D3, Figure 4) and Sk-Br-3 (the strongest responding cell line, Figure 4) have similar levels of the receptor for 1,25-dihydroxyvitamin D3 (VDR) [43], our data add support to the idea that responses may be mediated separately from receptor expression. As these cell lines are also ER-negative [44], it would appear that the suggestion posed by Lopes and colleagues (2010) is not limited to ER-positive cell lines [42]. It is worth noting, however, that, of the eight breast cancer cell lines we examined, the ER-negative Sk-Br-3 had the highest level of involucrin expression which could be an indicator of responsiveness irrespective of ER status.
Unlike some previous studies in breast cells, we found that although cell proliferation was reduced, there was little change in the cell cycle profile. A similar phenomenon was also observed in U937 leukaemia cells where 100 nM 1,25-dihydroxyvitamin D3 caused a decrease in colony formation but had minimal effects on the cell cycle stage [39]. This suggests that decreased rates of proliferation could occur due to a more uniform decrease in cell cycle progression rather than an accumulation of cells in a particular phase.
We have previously reported that Brk is a potential target for anti-proliferative therapy [6,45]. An alternative approach to preventing proliferation could be to induce keratinocyte-like differentiation with agents such as 1,25-dihydroxyvitamin D3, thereby preventing disseminating cells from dividing. Clinical data would support this approach as up to 70% of women diagnosed with breast cancer are vitamin D3 deficient [46]. Clinical trials with vitamin D supplementation as a therapeutic agent are currently ongoing in breast cancer. Results from prostate cancer patients show potential with patients treated with a combination of high-dose oral 1,25-dihydroxyvitamin D3 and docetaxel having a significantly improved overall survival compared to those treated with docetaxel alone [47].
The data presented in this study suggest that cells that are positive for Brk and involucrin might respond more readily to differentiation-inducing agents, such as vitamin D, than those that do not express Brk or involucrin. They contribute to growing knowledge enabling more precise predictions of cellular responses to 1,25-dihydroxyvitamin D3 by identifying biomarkers that are potentially independent of vitamin D receptor status. They were all cultured in RPMI-1640 medium supplemented with 10% foetal bovine serum, 2 mM L-glutamine, 100 U/mL penicillin and 100 µg/mL streptomycin (all Invitrogen) at 37 • C in a humidified atmosphere of 5% CO 2 in the air. GI101 (Basal) were acquired through our collaboration with the Eccles laboratory and cultured in the same medium with the additional supplement of 5 µg/mL insulin. Cells were sub-cultured twice weekly.

Vitamin D Treatment
Cells were seeded at 50,000 cells per well into 24-well plates and allowed to adhere for 24 h. Cell numbers were determined (T = 0) by haemocytometer counting of replicate wells and the medium was removed from remaining wells and substituted with fresh medium containing either 1,25-dihydroxyvitamin D3 (SelleckChem, Planegg, Germany) at the indicated concentrations or 100% ethanol diluted to 0.1% in culture medium as vehicle control. After a further 72 h, cell numbers were counted. The fold increase in cell number from T = 0 was then calculated.

Breast Carcinoma Samples
Human breast tumour samples: Forty-eight women aged 40-88 years (median age 58 years) provided samples of 50 primary breast neoplasms after informed consent. Two patients who had bilateral disease gave two samples, one from each side. Forty-eight of the neoplasms were invasive carcinomas, and two were high-grade ductal carcinoma in situ, one of which showed foci of micro-invasion. All samples were taken at surgery. The specimens were sent immediately after removal to the laboratory, where, after inking the margins and slicing the specimen, the histopathologist removed a small part of the tumour. Specimens were prepared and snap-frozen in the Norfolk and Norwich University Hospital's tumour tissue bank as described previously [48]. The rest of the surgical specimen was fixed in 10% formol saline and processed for routine histopathological diagnosis and establishment of prognostic factors.
Human normal mammary tissue: Samples of normal mammary tissue were obtained from 7 patients with informed consent at reduction mammoplasty. All samples were snap-frozen and stored as described for the cancers [48]. Samples were collected with the approval of the Norwich District Ethics Committee, the Norwich Research and Development Committee and the Partners in Cancer Research Tissue Bank Committee.

RNA Extraction
RNA was extracted from 44 of the frozen human breast tumour surgical samples for which histological grade had been determined, and from the seven normal human mammary tissue samples. Total RNA was isolated using a modification of the SV Total RNA Isolation System (Promega, Madison, WI, USA) [49]: 50-100 mg of tumour tissue were homogenized in 1 mL of RNazol B reagent (Biogenesis Ltd., Poole, UK) using an UltraTurrax T8 homogenizer (IKA). The homogenate was stored at −80 • C pending the following step of the protocol: the samples were allowed to thaw completely at room temperature, centrifuged at 14,000 rpm for 10 min and the clear supernatants collected into 200 µL of chloroform. They were then shaken vigorously for 15 s, incubated for 3 min at room temperature and centrifuged for 15 min at 14,000 rpm. The upper phase was collected into 200 µL of 95% ethanol, mixed and transferred into a spin basket assembly. Then, the Promega protocol was followed from step 7 to the end according to the manufacturer's instructions.

Reverse Transcription
One µg of total RNA was primed with 0.5 µg of random hexameric primers (Promega) and reverse transcribed into cDNA in a 20 µL reaction volume containing 200 units of SuperScript II reverse transcriptase, 5 × first-strand buffer [final concentrations: 50 mM Tris-HCl (pH8.3), 75 mM KCl, and 3 mM MgCl 2 ], 10 mM DTT, 0.5 mM of each deoxynucleotide triphosphate (dATP, dCTP, dGTP, and dTTP; Life Technologies, Inc., Carlsbad, CA, USA), and 40 units of RNasin RNase inhibitor (Promega, Southampton, UK). Each priming reaction was carried out at 70 • C for 10 min and stopped by placing samples on ice; the reverse transcription was carried out at 42 • C for 1 h followed by incubation at 70 • C for 10 min.

TaqMan Real-Time PCR
Forward and reverse primers and fluorescence-labelled oligonucleotide probes (using 6-carboxy-fluorescein (FAM) as the reporter dye and 6-carboxytetramethylrhodamine (TAMRA) as the quencher dye) were designed for the human Brk/PTK6 gene using Primer Express 1.0 software (PE Applied Biosystems, Foster City, CA, USA) and synthesized by PE Applied Biosystems (Warrington, UK). BLASTN searches were undertaken on our designed primers to confirm gene specificity. To ensure against genomic DNA amplification, the forward and reverse primers were placed in separate adjacent exons. A standard curve with concentrations ranging from 25 ng to 1 ng was produced using human placental cDNA dissected from foetal villi as the template. An XY scatter plot for each gene was produced using Microsoft Excel Chart Wizard software (Microsoft Office 2001), and values for the equation y = mx + b (where m = the slope of the standard curve and b = the y intercept of that line) and R2 were obtained. The 18S rRNA gene was used as an endogenous control to normalize for differences in the amount of total RNA in each sample. PCR reactions for all samples were performed in duplicate in 96-well optical plates with 5 ng of cDNA (1 ng of cDNA for the 18S gene), 100 nM probe, 200 nM each primer and 12.5 µL of TaqMan Universal 2 × PCR Master Mix (PE Applied Biosystems, Warrington, UK) in a 25 µL reaction volume. The amplification reaction was carried out over 40 cycles with the following parameters: an initial holding stage of 2 min at 50 • C and then 10 min at 95 • C, followed by a two-step cycling program of 15 s at 95 • C and 1 min at 60 • C.

Flow Cytometry
Cells treated with either 100 nM 1,25-dihydroxyvitamin D3 or 0.1% (v/v) ethanol (controls) for 72 h were detached, counted and resuspended at 10 6 /mL in 70% (v/v) ethanol. After overnight incubation at 4 • C, cells were centrifuged at 550× g for 5 min. Cells were then resuspended in PBS containing 4 µg/mL propidium iodide and 10 µg/mL ribonuclease A and incubated at 37 • C for 45 min, then 4 • C overnight. Samples were analysed on a Beckman Coulter FC500 flow cytometer.

Suspension Culture
Tissue culture plates were coated with a volume of poly(2-hydroxyethylmethacrylate) (polyHEMA, Sigma, Poole, UK) in 98% ethanol (20 mg/mL) to give a final polyHEMA concentration of 1.5 mg/cm 2 . Plates were air-dried in a tissue culture cabinet [4]. Cells were seeded into polyHEMA-coated wells at 0.75 × 104 cells/cm 2 . After 7 days in suspension culture, cells were harvested and lysed in 2 × SDS-PAGE lysis buffer.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Data Availability Statement: No datasets were generated during this study.