Intrinsic Mitochondrial Membrane Potential and Associated Tumor Phenotype Are Independent of MUC1 Over-Expression

We have established previously that minor subpopulations of cells with stable differences in their intrinsic mitochondrial membrane potential (Δψm) exist within populations of mammary and colonic carcinoma cells and that these differences in Δψm are linked to tumorigenic phenotypes consistent with increased probability of participating in tumor progression. However, the mechanism(s) involved in generating and maintaining stable differences in intrinsic Δψm and how they are linked to phenotype are unclear. Because the mucin 1 (MUC1) oncoprotein is over-expressed in many cancers, with the cytoplasmic C-terminal fragment (MUC1 C-ter) and its integration into the outer mitochondrial membrane linked to tumorigenic phenotypes similar to those of cells with elevated intrinsic Δψm, we investigated whether endogenous differences in MUC1 levels were linked to stable differences in intrinsic Δψm and/or to the tumor phenotypes associated with the intrinsic Δψm. We report that levels of MUC1 are significantly higher in subpopulations of cells with elevated intrinsic Δψm derived from both mammary and colonic carcinoma cell lines. However, using siRNA we found that down-regulation of MUC1 failed to significantly affect either the intrinsic Δψm or the tumor phenotypes associated with increased intrinsic Δψm. Moreover, whereas pharmacologically mediated disruption of the Δψm was accompanied by attenuation of tumor phenotype, it had no impact on MUC1 levels. Therefore, while MUC1 over-expression is associated with subpopulations of cells with elevated intrinsic Δψm, it is not directly linked to the generation or maintenance of stable alterations in intrinsic Δψm, or to intrinsic Δψm associated tumor phenotypes. Since the Δψm is the focus of chemotherapeutic strategies, these data have important clinical implications in regard to effectively targeting those cells within a tumor cell population that exhibit stable elevations in intrinsic Δψm and are most likely to contribute to tumor progression.


Introduction
Heterogeneity is a fundamental property of cellular systems, including solid tumors [1], where diversity likely provides reservoirs of cells which, through evasion of preventative or therapeutic intervention and/or tolerance and rapid response to shifting micro-environmental conditions, can participate in tumor expansion and progression. Our previous work has established that within populations of colonic and mammary carcinoma cells there are minor subpopulations of cells that exhibit stable differences in intrinsic mitochondrial membrane potential (Dym) which are linked to tumorigenic phenotypes [1,2]. Although the mechanisms involved in generating and maintaining differences in Dym are unclear, they may reflect alterations in mitochondrial membranes [3][4][5][6][7][8] (Mariadason & Heerdt, unpublished).
The mucin 1 (MUC1) oncoprotein is a highly O-glycosylated heterodimeric, type-I transmembrane glycoprotein that is overexpressed in many cancers [9,10]. MUC1 is synthesized as a single polypeptide which is cleaved into an N-terminal extra cellular fragment (MUC1 N-ter) and a C-terminal fragment (MUC1 Cter), which includes transmembrane and cytoplasmic domains.
Because MUC1 is over-expressed by most human carcinomas, where it is associated with tumor progression and poor prognosis [22], we asked whether constitutive levels of MUC1 were linked to stable differences in intrinsic Dym and, as a consequence, were a component or maker of tumor cell heterogeneity in colonic and/or mammary carcinoma cell populations. Using an siRNA approach, we then investigated the impact of diminished levels of MUC1 on intrinsic Dym and on Dym linked tumorigenic phenotypes including cellular sensitivity to butyrate (NaB) mediated apoptosis and constitutive hypoxia-independent VEGF secretion. Finally, we investigated the impact of pharmacological alteration of the Dym on MUC1 and hypoxia-independent VEGF secretion levels. Our data demonstrate that while elevated constitutive MUC1 is associated with subpopulations of cells with elevated intrinsic Dym, MUC1 is not directly linked to the generation or maintenance of stable alterations in intrinsic Dym, or to intrinsic Dym associated tumor phenotypes.
Chemotherapeutic strategies that preferentially target cells with relatively higher intrinsic Dym have been investigated and some clinical responses reported [23][24][25][26][27][28][29][30][31][32][33]. Based on the data presented here, exploitation of stable elevation in intrinsic Dym may be a novel and efficacious way to target cells within tumor populations that are the most likely to contribute to tumor expansion and progression [1,2].

Cell Culture
The SW620 and SW480 human colonic carcinoma cell lines, and the MCF7 human mammary carcinoma cell line, were obtained from the American Type Culture Collection. Subcloned cell lines were derived as described [1,2].

qRT-PCR
Total RNA was purified using Qiagen RNEasy kit (Qiagen, Valencia, CA). cDNA was synthesized using Superscript III First-Strand Synthesis System for RT-PCR (Invitrogen, Carlsbad, CA). MUC1 expression was measured by RT-PCR using Gene Expression Assay (Applied Biosystems, Carlsbad, CA). Expression was normalized to GAPDH.

Transfection
Cells were seeded at 5610 5 cells/well in 6 well plates and allowed to grow overnight. The following day cells were either left untreated, mock transfected, transfected with ON-TARGET Plus non-targeting (NT) siRNA pool (Dharmacon, Lafayette, CO), or MUC1 ON-TARGET Plus Smart Pool siRNA (Dharmacon, Lafayette, CO). Transfections were performed using Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA) and Opti-MEM reduced serum medium (Invitrogen, Carlsbad, CA).

Quantitation of cell surface MUC1 N-ter
Cells were stained with FITC-conjugated anti-human MUC1 (CD227) monoclonal antibody (BD Pharmingen) and the percentage of stained cells and staining intensity were quantified by flow cytometry.

Quantitation of vascular endothelial growth factor (VEGF)
Conditioned tissue culture medium was harvested for quantitation of VEGF 165 protein by ELISA (R&D Systems, Inc., Minneapolis, MN) according to the manufacturer's protocol. The reduction of MTT to formazan [40], combined with a standard curve generated by serial dilution of SW620 cells, was used to determine the number of cells/well and standardize for variations in cell number among wells [41].

Pharmacologic modulation of Dym
We have previously shown that exposure of SW620 cells to the K + ionophore valinomycin results in a sustained disruption the Dym [35,38]. Thus, following overnight treatment with 5 mM valinomycin, cells were stained with JC-1 or FITC-conjugated anti-human MUC1 (CD227) and analyzed by flow cytometry for quantitation of Dym and cell surface MUC1 N-ter, respectively. Levels of MUC1 C-ter were determined in mitochondrial enriched fractions by immunoblotting and levels of VEGF 165 were quantified in harvested cconditioned tissue culture medium (as described above).

Statistical analyses
Data from at least 3 independent determinations were compared by Bonferroni's multiple comparison tests. Mean data were also evaluated as a function of the intrinsic Dym, determined by JC-1 staining, using linear regression analyses [1,36,41].

Results
Constitutive MUC1 levels are associated with stable differences in intrinsic Dym of subcloned cell lines derived from colonic and mammary carcinoma cell populations MUC1 over-expression is associated with tumor progression and poor prognosis [22] and is reported to diminish apoptosis signalled through the intrinsic pathway [13][14][15][18][19][20] and enhance VEGF expression in some cancer cells in vitro [17,42,43] -cellular activities that we have linked to differences in intrinsic Dym of subcloned mammary and colonic carcinoma cell lines [1,2,41]. Therefore, we investigated the potential relationship between MUC1 levels and intrinsic Dym.
Steady state mRNA levels of MUC1 were determined by qRT-PCR in the parental population of SW620 human colonic carcinoma cells and in subcloned cell lines derived from SW620 cells that have stable intrinsic Dym ranging from higher, comparable or lower than that of the population (qDym, <Dym or QDym, respectively) [1]. As shown in Figure 1A, MUC1 mRNA levels are at least 1.5 fold, and as much as 6 fold, higher in the subcloned cell lines with elevated Dym (*P,0.001 vs. the population of SW620 cells).
To investigate whether this relationship between MUC1 mRNA and intrinsic Dym extended to MUC1 C-ter protein levels, mitochondrial (mt) and accompanying Post Mitochondrial (PMF) enriched fractions were isolated from the SW620 colonic carcinoma cell population and from subcloned lines derived from SW620 cells with a range of relative intrinsic Dym [1]. Immunoblots of mt and PMF were probed with antibodies Figure 1. Constitutive MUC1 levels are associated with stable differences in intrinsic Dym of subcloned cell lines derived from SW620 colonic carcinoma cell population. A: Steady state mRNA levels of MUC1 were determined by qRT-PCR, normalized to GAPDH, in the parental population of SW620 cells and in subcloned cell lines derived from SW620 cells that have a range stable intrinsic Dym. (*; P,0.001 vs. the population of SW620 cells.) B: Immunoblots of mt and PMF from the SW620 colonic carcinoma cell population and from subcloned lines derived from SW620 cells with intrinsic Dym higher (qDym), lower (QDym) and comparable (<Dym) to that of the population were probed with anti-MUC1 Cter and anti-VDAC and chemiluminescence signals quantified and expressed as a function of the intrinsic Dym, determined by JC-1 staining, using linear regression analyses. MUC1 C-ter, but not VDAC, levels in both mt enriched and PMF are significantly associated with the intrinsic Dym (P = 0.0053 and 0.0032, respectively). doi:10.1371/journal.pone.0025207.g001 directed specifically against the C-ter fragment of MUC1 (MUC1 C-ter) or the mt outer membrane protein VDAC, and chemiluminescence signals quantified and expressed as a function of the intrinsic Dym, determined by staining cells with the Dymdependent dye JC-1, using linear regression analyses. As shown in Figure 1B, while MUC1-C-ter levels in both mt enriched and PMF are significantly associated with the intrinsic Dym (P = 0.0053 and 0.0032, respectively), VDAC levels in neither the mt nor the PMF fractions are linked to intrinsic Dym. This relationship between MUC1 C-ter and intrinsic Dym of subcloned cells lines derived from SW620 cells was confirmed by analyses of multiple immunoblots (not shown).
We next asked whether the association between intrinsic Dym and levels of endogenous mt-associated MUC1 C-ter were unique to subclones derived from SW620 colonic carcinoma cells. Thus, we investigated subcloned cell lines with stable differences in intrinsic Dym that were derived from the SW480 and MCF7 human colonic and mammary carcinoma cell lines, respectively. Similar to the subclones derived from SW620 cells, the SW480 and MCF7 derived subcloned cell lines exhibit intrinsic Dymlinked tumorigenic phenotypes consistent with enhanced probability to participate in local tumor expansion [2]. Mitochondrial enriched fractions were isolated from the parental populations of SW480 and MCF7 cells and from accompanying derived subcloned cell lines with intrinsic Dym ranging from significantly lower to significantly higher than that of the population. Mtassociated MUC1 C-ter levels were then determined by immunoblotting and expressed as a function of the relative intrinsic Dym using linear regression analyses. As shown in Figure 2, constitutive levels of mt-associated MUC1 C-ter are significantly linked to the intrinsic Dym of subclones established from SW480 colonic, and MCF7 mammary, carcinoma cells (2A and 2B;; P = 0.0118 and 0.021, respectively), thereby supporting the relationship between the Dym and mt associated MUC1 C-ter in cancer cells.

siRNA mediated down-regulation of MUC1 does not affect intrinsic Dym
Although the levels of mt-associated MUC1 C-ter are linked to the intrinsic Dym of cancer cells and MUC1 C-ter has been reported to integrate into the outer mitochondrial membrane [13,14], it is unclear whether mitochondrial MUC1 C-ter contributes to regulation or maintenance of the Dym [21]. To investigate this, the SW620 cell population and derived subcloned cell lines with a range of relative intrinsic Dyms (QDym, <Dym and qDym) were mock transfected, or transfected with nontargeting (NT) or MUC1 siRNA. Twenty-four, 48 and 72 hours later MUC1 mRNA levels were determined by qRT-PCR ( Figure 3). Compared to mock or NT-siRNA transfected cells, MUC1 expression levels were down regulated approximately 40% to 60% 24 hours after transfection of MUC-siRNA. Interestingly, the most extensive decrease in MUC1 expression was seen in the subclone with elevated Dym, in which the original levels of MUC1 were the highest. siRNA-induced down regulation of MUC1 mRNA was maintained 48 and at 72 hours after transfection and, in fact, continued to decrease in each cell line at each time point by approximately 10%, achieving approximately 87% lower levels of MUC1 mRNA in the subclone with elevated intrinsic Dym after 72 hours. Importantly, although MUC1 expression was not completely silenced, levels in the subclone with elevated intrinsic Dym were comparable to those of the mock transfected population of SW620 cells 48 and 72 hours following MUC1-siRNA transfection MUC1 expression (P.0.05).
Because MUC1 mRNA levels in the siRNA-MUC1 transfected subclone with elevated Dym approximated those of the mock transfected population of SW620 cells, we focused on these two cell lines in subsequent investigations. Thus, cells were mock, NT-siRNA or MUC1-siRNA transfected and 24, 48 and 72 hours later relative levels of cell surface MUC1 N-ter and mt-associated MUC1 C-ter were determined by flow cytometry and immunoblotting, respectively. As shown in Figure 4, siRNA-mediated Figure 2. Constitutive mt-associated MUC1 C-ter levels are associated with stable differences in intrinsic Dym of subcloned cell lines derived from SW480 colonic and MCF7 mammary carcinoma cell populations. Mt-associated MUC1 C-ter levels were determined by immunoblotting in mt enriched fractions isolated from the population of SW480 (A) and MCF7 (B) cells and from subclones derived from these cell populations that have stable intrinsic Dym ranging from significantly lower to significantly higher than that of the population. MUC1 C-ter levels are expressed as a function of the relative intrinsic Dym by linear regression analyses. doi:10.1371/journal.pone.0025207.g002 down regulation of MUC1 mRNA resulted in coincident decreases in cell surface MUC1 N-ter and mt associated MUC1 C-ter in both cell lines at each time point following transfection (4A and 4B, respectively). Moreover, although not completely eliminated, consistent with the siRNA induced reduction of MUC1 mRNA (Figure 3), the levels of MUC1 N-ter surface staining and mtassociated MUC1 C-ter in MUC1-siRNA transfected cells with elevated intrinsic Dym were comparable to those of mock transfected population of SW620 cells at each time point (P.0.05). Therefore, siRNA-induced down regulation of MUC1 mRNA is reflected in coincidently diminished levels of cell surface MUC1 N-ter and mt-associated MUC1 C-ter, especially in cells with relatively high intrinsic Dym.
We next investigated how siRNA mediated down regulation of MUC1 affected the Dym. Twenty-four, 48 and 72 hours following mock, NT-siRNA or MUC1-siRNA transfection, cells were stained with JC-1 and analyzed by flow cytometry [1,[35][36][37][38]. As shown in figure 4C, siRNA induced reduction of MUC1 had no effect on the Dym of either cell line at any time points evaluated. Therefore, despite comparable levels of MUC1 mRNA, cell surface MUC1 N-ter and mt-associated MUC1 C-ter in the subclone with elevated intrinsic Dym and mock transfected population of SW620 cells -particularly 48 and 72 hours after MUC1-siRNA transfection-the relatively higher intrinsic Dym was maintained. Similarly, siRNA mediated down regulation had no impact on the Dym of additional SW620 derived subcloned cell lines with intrinsic Dym ranging from lower to higher than that of the population (not shown). Thus, it is unlikely that MUC1 levels play a role in the regulation or maintenance of stable difference in intrinsic Dym.
To investigate the impact of siRNA mediated down regulation of MUC1 on NaB induced apoptosis, the population of SW620 cells and derived subclone with elevated intrinsic Dym were mock, NT-siRNA or MUC1-siRNA transfected. Forty-eight hours later, a time point at which the MUC1 levels in two cell lines are comparable (Figures 3 and 4A and 4B), cells were exposed to 5mM NaB, a physiologically relevant concentration [49,50], for 48 hours, a time point at which we have shown induction of significant apoptosis [1,36,38]. Cells were then stained with PI and the percentage of apoptotic cells determined by flow cytometry. As shown in Figure 5, down regulation of MUC1 did not significantly impact NaB mediated apoptosis in either cell line, and, consistent with the absence of an effect on intrinsic Dym (Figure 4C), decreased sensitivity to NaB was retained in MUC1-siRNA transfected cells with elevated intrinsic Dym.
siRNA mediated down-regulation of MUC1 does not affect intrinsic Dym-linked constitutive hypoxiaindependent VEGF secretion We have previously shown a significant association between the intrinsic Dym and levels of steady state, hypoxia-independent VEGF secretion in subcloned cell lines derived from mammary and colonic carcinoma cells [1,2,41]. Because the cytoplasmic domain of MUC1 C-ter (MUC1-CD) has been linked to transcriptional regulation of VEGF [15][16][17] and our data show that the levels of MUC1 C-ter in the PMF -which encompasses cytoplasmic components -are related to the intrinsic Dym ( Figure 1B), we investigated the effect of siRNA induced MUC1 down regulation on steady state VEGF secretion in colonic Pharmacologically mediated disruption of the Dym does not affect MUC1 levels but attenuates intrinsic Dym-linked constitutive hypoxia-independent VEGF secretion Finally, to investigate the impact of disruption of the intrinsic Dym on MUC1 and constitutive hypoxia-independent VEGF secretion levels, the population of SW620 cells and derived subclone with elevated intrinsic Dym were exposed to the K + ionophore valinomycin to dissipate the Dym [35,38]. The Dym, secreted VEGF, MUC1 N-ter and MUC1 C-ter levels were then determined. As shown in Figure 7A, whereas valinomycin treatment resulted in decreases in the Dym of both cell lines by at least 90%, accompanied by significantly decreased constitutive VEGF secretion levels (#, P,0.05), neither cell surface MUC1 Nter nor mt-associated MUC1-C-ter ( Figure 7B) levels were impacted in either cell line by disruption of the Dym. Thus, it is likely that the intrinsic Dym and associated tumor phenotype are independent of MUC1 over-expression in colonic carcinoma cells.

Discussion
Heterogeneity among cells is a fundamental property of tumors and likely allows for selection of subpopulation phenotypes that have an increased probability of participating in subsequent tumor progression [51]. Whereas the origins of, and the mechanisms by which, cellular heterogeneity contributes unequally to tumorigenesis are unclear [51], it has been know for some time that the population average Dym of cancer cells is significantly higher than that of normal cells ( [24,[52][53][54][55] and that mitochondria are important and critical elements in establishing tumorigenic phenotype [56].
Using single cell subclones, our previous work has established the existence of subpopulations of cells within mammary and colonic carcinoma cell populations that, compared to the population average Dym, have stable elevations in their intrinsic Dym [1,2]. Moreover, we have shown that these elevations in intrinsic Dym are linked to tumor phenotypes associated with increased probability of tumor progression [1,2]. Because MUC1 is over-expressed in many human carcinomas and cell lines where it is also associated with tumor progression [9,22], we investigated the relationship between intrinsic Dym and endogenous MUC1 expression. We found that constitutive MUC1 levels are significantly associated with stable differences in intrinsic Dym of subcloned cell lines derived from colonic and mammary carcinoma cell populations, with higher levels of MUC1 exhibited in cells with elevated intrinsic Dym. However, we also found that siRNA-induced down-regulation of MUC1 in colonic carcinoma cells with elevated intrinsic Dym to levels comparable to those in cells with lower Dym does not affect the intrinsic Dym or Dymlinked tumor phenotypes, including the decreased sensitivity to chemoprotective NaB induced apoptosis and the increased constitutive, hypoxia-independent VEGF secretion that are characteristic of cells with elevated intrinsic Dym [1,2]. Additionally, we found that chemically induced disruption of the Dym was associated with decreased levels of constitutive hypoxia-independent VEGF secretion, but had no impact on MUC1 levels. Therefore, these data suggest that over-expression of MUC1 is not directly involved in the generation or maintenance of stable alterations in intrinsic Dym and Dym linked phenotypes, and that  the intrinsic Dym and associated tumor phenotype are independent of MUC1 over-expression in colonic carcinoma cells.
The higher average Dym in cancer cells [24,[52][53][54][55] has been used as a rationale to develop compounds that accumulate in the mt based on the Dym (i.e., mitochondriotropic agents) to preferentially kill tumor cells [23][24][25][26][27][28][29][30][31][32][33]. Because mitochondriotropic agents accumulate in the mitochondrial matrix at a 10fold higher level with each 60 mV increase in Dym [57], they have the potential of being particularly efficacious in targeting cells with modest elevations in the intrinsic Dym-the specific cells within the tumor cell population that our data show are the most likely to participate in tumor expansion and progression [1,2].  . Pharmacologically mediated disruption of the Dym does not affect MUC1 levels but attenuates intrinsic Dym-linked constitutive hypoxia-independent VEGF secretion. Cells were exposed to 5 mM valinomycin overnight and then stained with JC-1 or FITCconjugated anti-human MUC1 (CD227) and analyzed by flow cytometry for quantitation of Dym and cell surface MUC1 N-ter, respectively. Levels of VEGF 165 were quantified in conditioned tissue culture medium harvested from untreated and valinomycin treated cells (#; P,0.05 valinomycin vs. untreated) (A). Levels of MUC1 C-ter were determined in mitochondrial enriched fractions by immunoblotting (B). doi:10.1371/journal.pone.0025207.g007