The acid-base transport proteins NHE1 and NBCn1 regulate cell cycle progression in human breast cancer cells

ABSTRACT

Precise acid-base homeostasis is essential for maintaining normal cell proliferation and growth. Conversely, dysregulated acid-base homeostasis, with increased acid extrusion and marked extracellular acidification, is an enabling feature of solid tumors, yet the mechanisms through which intra- and extracellular pH (pH_i, pH_e) impact proliferation and growth are incompletely understood. The aim of this study was to determine the impact of pH, and specifically of the Na⁺/H⁺ exchanger NHE1 and Na⁺, HCO₃⁻ transporter NBCn1, on cell cycle progression and its regulators in human breast cancer cells. Reduction of pH_e to 6.5, a common condition in tumors, significantly delayed cell cycle progression in MCF-7 human breast cancer cells. The NHE1 protein level peaked in S phase and that of NBCn1 in G2/M. Steady state pH_i changed through the cell cycle, from 7.1 in early S phase to 6.8 in G2, recovering again in M phase. This pattern, as well as net acid extrusion capacity, was dependent on NHE1 and NBCn1. Accordingly, knockdown of either NHE1 or NBCn1 reduced proliferation, prolonged cell cycle progression in a manner involving S phase prolongation and delayed G2/M transition, and altered the expression pattern and phosphorylation of cell cycle regulatory proteins. Our work demonstrates, for the first time, that both NHE1 and NBCn1 regulate cell cycle progression in breast cancer cells, and we propose that this involves cell cycle phase-specific pH_i regulation by the two transporters.

Introduction

Breast cancer is one of the most commonly diagnosed cancers globally and by far the most frequent cancer among women, with almost 250,000 new cases in 2016 [1]. Dysregulation of cell cycle control is a hallmark of most cancers, including breast cancers [2,3]. Molecular analysis of human tumors has revealed frequent mutations and epigenetic changes in cell cycle regulators, including deregulation and overexpression of cyclins and CDKs as well as loss of CDK inhibitors [4]. Overexpression of cyclin D [5,6] and cyclin E [7,8] is associated with breast cancer, and small molecule CDK inhibitors, in particular CDK4/6 inhibitors, are approved therapies for some breast cancer subtypes [9,10].

It is well recognized that the microenvironmental context plays a pivotal role in the control of proliferation of breast epithelial cells and can be determinant for whether they remain normally differentiated or switch to a de-differentiated state with uncontrolled proliferation and eventually full-blown mammary cancer [11,12]. However, the focus has been on the role of the stromal microenvironment (ibid.), and it is essentially unknown how physico-chemical microenvironment conditions may impact breast cancer cell proliferation. A hallmark of the microenvironment in solid tumors is a profoundly dysregulated pH homeostasis resulting from elevated metabolic acid production and compensatory upregulation of acid extrusion in the cancer cells. This leads to very acidic extracellular pH (pH_e) values (generally pH 6–6.8) in solid tumors, while cytoplasmic intracellular pH (pH_i) remains normal or even slightly alkaline [13–16]. We and others have demonstrated that the Na⁺/H⁺-exchanger NHE1 (SLC9A1) and the Na⁺/HCO₃^–cotransporter NBCn1 (SLC4A7) are upregulated in many breast cancer cells and play central roles in their pH regulation [17–21] as well as in their growth as 3-dimensional spheroids and in vivo tumors [22–24].

It has long been recognized that a slightly alkaline pH_i is a prerequisite for mitogen-induced cell proliferation and growth [25–27]. Early studies in sea urchin eggs demonstrated that activation of Na⁺/H⁺-exchange occurred very early in metabolic activation after fertilization, leading to an increase in pH_i and activation of protein synthesis [28]. Later work in mammalian cells demonstrated that under HCO₃⁻ free conditions, Na⁺/H⁺-exchanger activation, leading to cytoplasmic alkalinization, was important for cell proliferation, at least in part by activating DNA and protein synthesis [29–33]. A central role for NHE1 in the timing of G2/M entry and transition in PS120 fibroblasts was subsequently demonstrated and was suggested to involve pH_i-dependent regulation of the cell cycle regulators cyclin B1 and Cdc2 [34]. However, with a few exceptions [34] these studies were carried out under HCO₃⁻ free conditions, and none have directly addressed the roles of HCO₃⁻ transporters such as NBCn1. Given the known, major role of NBCn1 in mammary cancer cell pH regulation and primary tumor development [18,19,22–24,35,36], this raises the question of the relative importance and mechanisms of cell cycle control by NHE1 and NBCn1.

The aim of this study was, therefore, to determine the impact of pH, and specifically of NHE1 and NBCn1, on cell cycle regulation in human breast cancer cells in presence of HCO₃⁻ to allow assessment of the contribution of NBCn1. We show that reduction of pH_e significantly delayed cell cycle progression in MCF-7 breast cancer cells. The NHE1 protein level peaked in S phase and that of NBCn1 in G2/M. Steady state pH_i changed through the cell cycle in a manner dependent on NHE1 and NBCn1. Accordingly, knockdown (KD) of either NHE1 or NBCn1 reduced proliferation, delayed cell cycle progression in a manner involving S phase prolongation and delayed G2/M transition, and altered the expression pattern and phosphorylation of cell cycle regulatory proteins. Our work demonstrates that both NHE1 and NBCn1 regulate cell cycle progression in breast cancer cells, and we propose that this involves phase-specific pH_i regulation by the two transporters.

Results

Cell cycle progression of MCF-7 cells is reduced by decreasing pH_e

Most studies of the pH dependence of cell proliferation have been performed in absence of HCO₃⁻, precluding contributions from HCO₃⁻ dependent transporters. To determine the importance of pH for cell cycle progression in MCF-7 human breast cancer cells under pH- and HCO₃⁻ conditions relevant to solid tumors, we adjusted growth medium pH to 6.5, 7.4, or 7.6 by changing [HCO₃⁻] under constant pCO₂ (5%). This will additionally elicit a corresponding albeit lesser, change in pH_i [37] (in MCF-7 pLKO.1 cells, approximately 6.8, 7.1, and 7.2 at pH_e 6.5, 7.4, and 7.6, respective (JS, SFP, unpublished)), while maintaining physiological HCO₃⁻ conditions. Cells were synchronized by double thymidine block, and released to monitor progression through the cell cycle at selected time points by flow cytometric measurement of DNA content (Figure 1). Notably, a decrease in pH_e to 6.5 resulted in a significant delay in cell cycle progression, with the S-phase peak (Figure 1(b)) shifting from 3 h after release at pH_e 7.4, to ~7.5 h after release at pH_e 6.5. In acidic pH_e, a significantly higher number of cells were in G0/G1 at all times (Figure 1(a)), and the fraction of cells reaching G2/M phase at 10 h after release was significantly reduced (Figure 1(c)). Further, a greater fraction of cells (~20%) were in G2/M phase immediately after release (1.5–3 h) as compared to cells grown at pH_e 7.4 or 7.6 (~10%), suggesting arrest in G2/M during the synchronization protocol. Notably, in acidic pH_e, the difference between the maximum and minimum number of cells in each cell cycle phase is ~10%. This indicates that only 10% of acid pH_e-grown cells are actively cycling, and the rest are arrested in G1 or G2/M during synchronization. Interestingly, a change to pH 7.6 also slightly delayed cell cycle progression, suggesting that pH_e 7.4 is optimal for cell cycle progression of MCF-7 cells under these conditions.

Figure 1.

Cell cycle progression of synchronous MCF-7 cells pLKO.1 cells is delayed in acidic extracellular pH.

MCF-7 pLKO.1 cells were grown under acidic (pH 6.5), neutral (pH 7.4) and alkaline (pH 7.6) conditions in presence of HCO₃⁻/CO₂ for 24 h, and were synchronized at the G1/S-phase border by a double thymidine block. Cells were subsequently released and collected, DNA was stained with propidium iodide, and cell cycle progression was analyzed by flow cytometry. (a) Quantification of cells (%) in G0/G1 phase. (b) Quantification of cells (%) in S-phase. (c) Quantification of cells (%) in G2/M phase. Data is represented as means with SEM error bars (n = 3). Paired, Two-Way ANOVA with repeated measures and Dunnet’s multiple comparison post-test was used to determine statistically significant differences between the control cell line grown under neutral pH conditions (7.4) and the cell lines grown under acidic (pH 6.5) and alkaline (pH 7.6) conditions. *, **, ***, and **** indicate p < 0.05, 0.01, 0.001, and 0.0001, respectively.

Proliferation is reduced by knockdown of either NHE1 or NBCn1

This result showed that changing pH conditions profoundly impacts cell cycle progression, but did not allow us to distinguish between effects of reduced pH_e or secondary changes in pH_i. NHE1 and NBCn1 are robustly expressed in MCF-7 cells and contribute to their growth as tumor spheroids [23]. To address the role of pH_i and gain insight into the transporters involved, we therefore next employed MCF-7 cells with stable lentiviral knockdown (KD) of either NHE1 or NBCn1, and the corresponding pLKO.1 empty vector control cells (hereafter control cells). The mean KD was approximately 50% for NHE1 and 55% for NBCn1 (Figure 2(a), Suppl. Figure 6), similar to our previous report [23]. MCF-7 cells only grow exponentially for a short period of time, presumably because they immediately form tight islands of cells, restricting growth by contact inhibition. Thus, doubling times were calculated from the first, brief exponential period of growth, indicated by the shaded region in Figure 2(b). Proliferation of asynchronous cells was significantly reduced by KD of NBCn1 or NHE1, with doubling times of approximately 22 h and 18 h in the NBCn1 and NHE1 KD cells, respectively, compared to 10 h in control cells (Figure 2(b-c)). Confirming this finding, the frequency of Ki-67-positive nuclei (reflecting actively cycling cells) in asynchronous MCF-7 cells decreased with KD of NHE1 or NBCn1 (Figure 2(d-e)).

Figure 2.

Proliferation of asynchronous MCF-7 cells is reduced by knockdown of NHE1 or NBCn1.

(a) Western blot analysis of NHE1 and NBCn1 protein expression in MCF-7 NHE1 KD, NBCn1 KD, and empty vector control (pLKO.1) cells to verify KD efficiency (see Suppl. Figure 6 for quantification). (b) Cell confluency was analyzed using the IncuCyte ZOOM live-cell imaging system. Images were acquired every 2 h for at least 72 h. Proliferation curves (mean with SEM error bars, n = 6) pLKO.1 (black), NBCn1 KD (red) and NHE1 KD (grey) cells are shown. The grey area represents the region in which growth was approximately exponential, and within which doubling times were calculated. Paired Two-Way ANOVA with repeated measures and Dunnet’s multiple comparison post-test was used to determine statistically significant differences between the control cell line (pLKO.1) and the KD cell lines. **** indicates p < 0.0001. (c) An exponential growth curve was fitted to data obtained from 0–15 h, for each cell line and best-fit values for doubling times were plotted and compared by a paired Student’s t-test. * and ** indicate p < 0.05 and 0.01, respectively. (d) Representative immunofluorescence images of Ki-67 expression in MCF-7 cell lines. Upper panel shows Ki-67 only, and lower panel shows overlays of Ki-67 and DAPI staining (e) The percent Ki-67 positive cells calculated from the experiments shown in (d).

Collectively, these results show that both NHE1 and NBCn1 support proliferation of asynchronously growing MCF-7 cells.

Cell cycle progression is delayed by KD of NHE1 or NBCn1

To determine whether NHE1 and NBCn1 are important for cell cycle progression, cells were synchronized as above and cell cycle progression was monitored by flow cytometric measurement of DNA content. Representative DNA histograms from each time point and the corresponding quantifications are shown in Suppl. Figures 2(A-B, 3(A-B) and 4(A-B) for control, NBCn1- and NHE1 KD cells, respectively. The percentage of NHE1- and NBCn1 KD cells in either phase at time zero did not differ from that in control cells (Suppl. Figure 1), justifying normalization to 1 at this time. At the time of release, approximately 50% of control cells were in either G0/G1 phases or S phase (Figure 3(a,b)), confirming successful synchronization at the G1/S phase border. After 1.5 h, S phase progression was initiated (Figure 3(b)). The subsequent decrease in S phase population (3–6 h) was mirrored by an increase in G2/M phase cells, reflecting S/G2 transition (Figure 3(b,c)). The G2/M population reached a maximum at 10 h, followed by a decrease (10–12 h), mirrored by a reciprocal increase in G0/G1 phase cells indicating M/G1 transition (Figure 3(a,c)). Parallel immunoblot analyses demonstrated that the relative expression of cyclin A and cyclin B1 was increased in G2/M, Tyr15-phosphorylation of Cdc2 was increased during G1/S, and Ser10-phosphorylation of histone (H)3 was only detectable in G2/M, consistent with the known expression pattern of these proteins [38] (Suppl. Fig. 2C, 3C, and 4C, for control, NBCn1-, and NHE1 KD cells, respectively).

Figure 3.

Cell cycle progression of synchronous MCF-7 cell lines is delayed by knockdown of NHE1 or NBCn1.

Cells were synchronized at the G1/S-phase border by a double thymidine block. Cells were subsequently released, stained with PI, and subjected to flow cytometric analysis. For each cell line, and for each cell cycle phase, the percentage of cells at a given time point was normalized to the respective initial percentage of cells at 0 h, to determine the relative time-dependent cell cycle progression of each cell line. (a) Relative cell count in G0/G1 phase. (b) Relative cell count in S-phase. (c) Relative cell count in G2/M phase. Data is represented as means with SEM error bars (n = 3). Paired Two-Way ANOVA with repeated measures and Dunnet’s multiple comparison post-test was used to evaluate statistical significance relative to the control cell line (pLKO.1). *, **, *** and **** indicate p < 0.050.01, 0.001, and 0.001, respectively.

The G0/G1 population was not detectably affected by KD of either NHE1 or NBCn1 (Figure 3(a)). Notably, however, the S phase was prolonged in both NBCn1- and NHE1 KD cells compared to that in control cells (Figure 3(b)). This correlated with delayed entry into G2/M: Whereas control cells started to progress into G2/M phase 3 h after release, peaking after 10 h, NBCn1- and NHE1 KD cells did not start G2/M progression until approximately 5 h after release (Figure 3(c)). Furthermore, the relative cell counts for the NHE1 and NBCn1 KD cells in G2/M were decreased with approximately 40% at peak values compared to that of control cells. This suggests that the fraction of cells undergoing division was reduced by KD of NHE1 or NBCn1, consistent with the increased doubling times (Figure 2(c)) and decreased frequency of Ki-67 positive nuclei in these cells (Figure 2(e)).

Thus, KD of either NBCn1 or NHE1 resulted in S phase prolongation, delayed G2/M entry, and a significant decrease of the fraction of cells undergoing cell division.

NHE1 and NBCn1 expression, steady state pH_i and net acid extrusion capacity vary through the cell cycle

An NHE1-dependent increase in pH_i at the end of S phase has previously been reported [34], yet the possible involvement of NBCn1 in cell cycle progression has to our knowledge never been addressed, nor has the expression pattern of NHE1 and NBCn1 through the cell cycle.

To address this, MCF-7 cells were synchronized and NHE1 and NBCn1 expression analyzed by immunoblotting over a 12 h period after release. For NHE1, data are shown for control and NBCn1 KD cells (Figure 4(a)) and for NBCn1, data are shown for control and NHE1 KD cells (Figure 4(b)), to evaluate the impact of possible compensatory changes in expression levels. The absolute time zero expression of NHE1 in NBCn1 KD cells and of NBCn1 in NHE1 KD cells, respectively, did not differ from that in control cells (Suppl. Figure 7), justifying normalization to 1 at this time. NHE1 expression was highest in early S phase and declined gradually through the rest of the cell cycle. In NBCn1 KD cells, NHE1 expression appeared to increase in late S and G2, declining again toward M phase (Figure 4(a)). In contrast, relative NBCn1 expression was stable until the start of G2, after which it tended to increase. Surprisingly, the relative NBCn1 expression in NHE1 KD cells was, in contrast, modestly decreased from late S onward (Figure 4(b)).

Figure 4.

Protein expression of NHE1 and NBCn1 changes during cell cycle progression.

MCF-7 pLKO.1 and KD cells were synchronized at the G1/S-phase border by a double thymidine block. Cells were subsequently released, lysed, and analyzed by WB analysis. (a) Upper panel: WB of NHE1 protein expression in MCF-7 pLKO.1 and NBCn1 KD cells over 12 h after thymidine release. Lower panel: Quantification of blots normalized to their respective loading controls and 0 h samples. p150 was used as a loading control. The data is shown as means with SEM error bars (n = 6). (b) Upper panel: WB of NBCn1 protein expression in MCF-7 pLKO.1 and NHE1 KD cells over 12 h after thymidine release. Lower panel: Quantification of blots normalized to their respective loading controls and 0 h samples. β-actin was used as a loading control. The data is shown as means with SEM error bars (n = 8). An unpaired Two-Way ANOVA with repeated measures and Sidaks multiple comparison post-test was used to evaluate statistical significance relative to the control cell line (pLKO.1). Paired two-tailed Student’s t-test was used to evaluate statistical significance difference between the graphs. *, **, and *** indicate p < 0.05, 0.01, and 0.001, respectively.

To assess whether steady state pH_i and pH_i regulatory capacity varied during cell cycle progression and evaluate its dependence on NHE1 and NBCn1, control cells, NHE1- and NBCn1 KD cells were synchronized and pH_i homeostasis evaluated at time points corresponding to early S, late S, G2 and M phase (1.5, 3, 7 and 10 h after release, respectively). All experiments were performed at 5% CO₂, pH_e 7.4, 25 mM HCO₃⁻ to allow contributions from HCO₃⁻ dependent transporters. In control cells, steady state pH_i was approximately 7.1 in early S phase, declining to approximately 6.8 in G2 and partially recovering, to about 7.0, in M phase (Figure 5(a)). Steady state pH_i tended to be decreased in early S in both NHE1 and NBCn1 KD cells compared to that in control cells, albeit with substantial variability and not reaching statistical significance (Figure 5(a)). In control cells, the net acid extrusion capacity (pH_i recovery rate) after an NH₄Cl prepulse was approximately 0.075 min⁻¹ in early S phase, increasing slightly in late S phase and remaining at about 0.085 min⁻¹ for the rest of the cell cycle (Figure 5(b)). Throughout the cell cycle, the acid extrusion capacity tended to be highest in control cells and to be reduced by KD of either NHE1 or NBCn1 (Figure 5(b)).

Figure 5.

pH_i changes during cell cycle progression.

MCF-7 pLKO.1 and KD cells were synchronized at the G1/S-phase border by a double thymidine block. Cells were subsequently released and loaded with the fluorescent pH_i indicator BCECF at the following time points: 1.5 h, 3 h, 7 h, and 10 h, corresponding to early S phase, late S phase, G2 and M phase, respectively. (a) Steady State pH_i (SSpH_i) at the different cell cycle phases (n = 8). (b) pH_i recovery rate (pH_i recovery rate (min⁻¹)) at the different cell cycle phases (n = 7). Representative original traces for each cell line are shown in suppl. Figure 5. The data is shown as box plots with min-to-max whiskers. For each bar, the line represents the median and the “+” the mean value (n = 7). An unpaired Student’s t-test was used to evaluate statistical significance relative to the control cell line (pLKO.1). * and (*)indicate p < 0.05 and 0.05 < p < 0.06 respectively.

These results show that NHE1 and NBCn1 expression varies through the cell cycle, NHE1 peaking at early S, and NBCn1 in G2/M. Steady state pH_i is highest in G1/S, dips upon entry into G2, and increases again toward M phase. Finally, NHE1 and NBCn1 are important for steady state pH_i and net acid extrusion capacity through the cell cycle.

KD of NHE1 or NBCn1 alters the temporal pattern of expression of cell cycle regulators

In order to gain insight into the specific roles of NHE1 and NBCn1, we next asked how the expression patterns of regulators and markers of cell cycle progression were affected by transporter KD during cell cycle progression. The absolute time zero expression of all cell cycle regulators in NHE1 and NBCn1 KD cells did not differ from that in control cells (Suppl. Figure 6), justifying normalization to 1 at this time. Consistent with the flow cytometry results, Ser10-H3 phosphorylation, corresponding to entry into mitosis, was delayed in both NHE1- and NBCn1 KD cells compared to that in control cells (Figure 6(a)). As expected, the relative expression of the transcription factor E2F1 was maximal in G1/S (when its suppression by pRb is reduced [39,40], followed by a marked, gradual decrease (Figure 6(b)). Interestingly, KD of NBCn1 was associated with a significant decrease, and KD of NHE1 with an increase, in E2F1 expression in G1/S, compared to that in control cells (Figure 6(b)). The phosphorylation of ribosomal protein S6 (pS6), indicating capacity for protein synthesis, was rather stable throughout the cell cycle in pLKO.1 and NHE1 KD cells, yet peaked toward M phase in NBCn1 KD cells (Figure 6(c)). These data confirm the regulation of cell cycle progression by NHE1 and NBCn1, and points to an impact of acid-base regulation on the protein synthesis machinery, in congruence with earlier reports [31].

Figure 6.

Changes in phospho-H3, E2F1, and phospho-S6 protein levels during cell cycle progression.

MCF-7 pLKO.1 and KD cells were synchronized at the G1/S-phase border by a double thymidine block. Cells were subsequently released, lysed, and analyzed by WB analysis. (a-c) Upper panels: WB analysis of protein expression during cell cycle progression. Lower panels: Quantification of blots normalized to their respective loading control and the value at 0 h (0 h data are shown in Suppl. Figure 6). (a) phospho-Ser10 H3 (n = 4). (b) E2F1 (n = 6). (c) pS6 (n = 4). p150 and β-actin were used as loading controls. Data is shown as means with SEM error bars, and an unpaired Two-Way ANOVA with repeated measures and Dunnet’s multiple comparison post-test was used to evaluate statistical significance relative to the control cell line (pLKO.1). *, **, ***, and **** indicate p < 0.05, 0.01, 0.001, and 0.0001, respectively.

We next evaluated the impact of transporter KD on Cdc2-phosphorylation on Tyr15, removal of which by Cdc25 is necessary to allow mitotic entry [41], Cyclin B1, which forms a complex with Cdc2 required for M phase transition [41], Cyclin A2, which contributes to G1/S as well as G2/M progression [42], and Cyclin D, which in complex with CDK4/6 promotes pRb hyperphosphorylation, allowing cell cycle entry [40]. The relative pCdc2 level in pLKO.1 cells was stable until 6 h after release, followed by a peak in late G2 to G2/M, before the expected decrease during G2/M and M phase. KD of either NHE1 or NBCn1 resulted in an earlier increase in the level of the inhibitory pCdc2 phosphorylation (Figure 7(a)). Cyclin B1 expression showed the expected pattern of gradually peaking as the cells move toward and into M phase. This pattern was largely unaffected by transporter KD, although NBCn1 KD appeared to slightly increase relative Cyclin B1 expression during S-, and G2 phase (Figure 7(b)). Cyclin D expression was high throughout the cell cycle in control cells (Figure 7(c)). The overall expression pattern was similar in all three cell lines, with the exception of a peak in relative Cyclin D expression during S and G2 in both KD cell lines and an apparent decrease in cyclin D expression during M phase in NHE1 KD cells (Figure 7(c)). Finally, Cyclin A2 expression showed the expected pattern of peaking during G2/M in control cells. Notably however, KD of NHE1 or NBCn1 appeared to prolong the Cyclin A2 peak with several hours (Figure 7(d)). The absolute protein expression of the cell cycle regulators in asynchronous NHE1- and NBCn1 KD cells did not differ from that in control cells (Suppl. Figure 7), indicating cell cycle specific regulation of their expression by NHE1 and NBCn1.

Figure 7.

Changes in phospho-Cdc2, Cyclin B1, Cyclin D, and Cyclin A2 levels during cell cycle progression.

MCF-7 pLKO.1 and KD cells were synchronized at the G1/S-phase border by a double thymidine block. Cells were subsequently released, lysed, and analyzed by WB analysis. (a-d) Upper panel: WB analysis of protein expression during cell cycle progression. Lower panel: Quantification of blots normalized to their respective loading control and 0 h sample. (a) phospho-Tyr 15 Cdc2 (n = 5). (b) Cyclin B1 (n = 5). (c) Cyclin D (n = 8). (d) Cyclin A2 (n = 4). p150 was used as a loading control. Data is represented as mean with SEM and an unpaired Two-Way ANOVA with repeated measures and Dunnet’s multiple comparison post-test was used to evaluate statistical significance relative to the control cell line (pLKO.1). * indicates p < 0.05.

Taken together, these results show that KD of NHE1 or NBCn1 modestly alters the temporal pattern of multiple cell cycle regulators, including increased early Cdc2 phosphorylation, delayed cyclin A expression, and strongly delayed H3 phosphorylation.

Discussion

We show here that under physiological CO₂/HCO₃⁻ conditions, reduction of pH_e, or KD of either of the net acid extruding transporters NHE1 and NBCn1 delays cell cycle progression of MCF-7 human breast cancer cells. Similarly, KD of either NHE1 or NBCn1 reduces proliferation, alters the pattern of steady state pH_i during the cell cycle, and shifts the expression pattern of several cell cycle regulatory proteins. Collectively, this demonstrates that both transporters contribute to sustaining MCF-7 cell proliferation under physiological CO₂/HCO₃⁻ conditions.

Numerous studies have addressed the impact of NHE1 inhibition or ablation on cell proliferation [26,29–31,43–45]. However, with very few exceptions [34], previous such studies have been limited to effects on asynchronous proliferation or on mitogen-induced G0-G1 transition, and the great majority were performed in absence of CO₂/HCO₃⁻, limiting contributions from HCO₃⁻ dependent transporters. Furthermore, most were performed in fibroblasts with exogenous overexpression of NHE1 [26,27]. It has been widely debated whether NHE1 activation is an integral part of the mitogenic signal or a permissive condition establishing the minimal requirements for growth. Although mitogens generally increase pH_i at least in absence of HCO₃⁻, alkalinization per se is generally not sufficient for mitogen-induced proliferation of mammalian cells [26]. Existing work indicates that under acidic pH_e (≤ 7.0) conditions or in the absence of CO₂/HCO₃⁻, NHE1 activity is most often required for proliferation, whereas in alkaline pH_e or in presence of CO₂/HCO₃⁻, this is generally not the case, ostensibly because pH_i is often higher under these conditions [26,27]. We and others have demonstrated roles for NBCn1 [22] and other HCO₃⁻ transporters [36] in tumor cell proliferation, and a role for CO₂/HCO₃⁻ dependent transport in vascular smooth muscle cell proliferation was recently reported [46] NHE1 and NBCn1 play central roles in regulation of pH_i in MCF-7 cells [18]. In congruence with this, we show here that pH_i and net acid extrusion capacity during the cell cycle tended to be reduced upon KD of either NHE1 or NBCn1. The requirement for both transporters is fully in line with our recent reports showing that both transporters are important for growth of 3D spheroids of MCF-7 cells [23] and for xenograft growth of MDA-MB-231 breast cancer cells [24].

We report here that cell cycle progression is associated with a gradual decrease in pH_i from early S phase to G2, followed by recovery toward M phase. KD of either transporter altered this pattern, prolonged S phase duration, and delayed and impaired G2/M. Transporter KD prolonged the doubling time of asynchronously growing cells more profoundly than it shifted the cell cycle progression of synchronized cells. This likely reflects that in contrast to synchronized cells, the great majority (~70%) of asynchronous cells are in G1 phase. In conjunction with the extensive G0/G1 arrest when pH_e is reduced to 6.5, this could indicate the existence of a pH sensitive step gating G1/S transition, which is attenuated by the lower pH_i of the KD cells. Such a scenario has previously been suggested to involve pH sensitivity of DNA and protein synthesis [30,31,33], but it could also relate to NHE1/pH_i-dependent regulation of cell volume, which has been suggested to gate the timing of the G1/S [49] and G2/M transition 47, ensuring balance between growth and proliferation rates [48,51]. Additionally, our results show that acidic pH_e causes further arrest of cells in G2/M, suggesting that a pH sensitive checkpoint may also guard G2/M transition. This would be consistent with previous reports showing altered expression/activity of G2/M regulatory proteins in NHE1-deficient fibroblasts [34,52]. Tentatively, such a checkpoint could further involve pH sensitivity of the mitotic spindle, since microtubule polymerization is favoured at higher pH_i values [53], and could thus play a role in regulating the timing of mitosis, as recently proposed [54].

NHE1 expression was highest in early S phase, coinciding with a pH_i peak at this time, whereas NBCn1 expression was highest in late G2/M, coinciding with the second pH_i peak. This suggests that NHE1 may be most important in early S phase and NBCn1 in G2/M. Supporting this notion, fibroblasts lacking NHE1 exhibited S phase prolongation, impaired entry into G2/M, and partial loss of a pH_i peak at the end of S phase [34,45,55,56]. The NHE1 upregulation in late S/G2 in NBCn1 KD cells also fits with such a scenario, i.e., that the peak of NHE1 expression is shifted because of the loss of the NBCn1 peak at this time. The turnover time of both transporters in non-synchronized/non-cycling cells is long compared to the duration of a cell cycle phase: ~ 20–24 h for NHE1 [57,58] and ~75–95 h for NBCn1 [58,59]. This suggests the interesting possibility that the transporters may undergo cell cycle specific regulation. Since S and G2/M populations are only 15–20% of asynchronous cells, and much lower if cells are confluent and thus predominantly in G0, such a difference would contribute little to the total turnover time, but could be functionally very important during cell cycle progression. The specific molecular mechanisms regulating transporter protein levels during the cell cycle were not addressed here, and should be addressed in future work. Role(s) for previously described cell cycle regulatory signaling pathways [40] seems likely, however, also the pH changes during the cell cycle could be involved. For example, ChIP-seq analysis has shown that the NHE1 promoter interacts with, among many others, the transcription factors E2F1 and Egr-1 [60]. E2F1 is a major regulator of cell cycle dependent transcription [40], and it is interesting to note that its expression profile as shown in this study fits well with the NHE1 expression profile during the cell cycle (compare Figures 4(a) and 6(b)). Because the KD cells exhibit reduced pH_i recovery capacity in G2 compared to pLKO.1 cells it is tempting to suggest that this is what drives the compensatory NHE1 upregulation in NBCn1 KD cells. Egr-1 is a candidate such regulator, since it is upregulated by acidosis and its binding to the promoter of the closely related NHE2 is required for acid-induced NHE2 upregulation [61].

As mentioned, DNA-[32] and protein synthesis [31,33] have been shown to be inhibited at acidic pH_i, and some G2/M regulatory proteins were proposed to exhibit altered expression/activity in NHE1-deficient fibroblasts [34,52]. However, beyond this, the mechanisms through which pH_i might drive or favor cell cycle progression are not known. Our work indicates that H3 phosphorylation was delayed, inhibitory Cdc2 phosphorylation was increased from late S and into G2/M, and the G2 Cyclin A peak was delayed, in both the NHE1 and NBCn1 KD cell lines, consistent with the delayed cell cycle progression. Interestingly, E2F1 levels in S phase were increased in NHE1- and decreased in NBCn1 KD cells, and S6 phosphorylation was also differentially affected by KD of NHE1 and NBCn1. Together with the differential pattern of transporter expression and pH_i changes this points to at least partially different roles of the two transporters, either due to their temporal expression pattern, or because of transporter-specific regulation of activity or downstream effects. In this regard, future work should address the possibility that transporter-dependent regulation of [Na⁺]_i, pH_e, and/or cell volume contributes to the impact of these transporters on cell cycle progression and proliferation.

Regardless of the mechanisms involved, the strong impact of both NHE1 and NBCn1 on cell cycle progression of breast cancer cells reinforces previous work by us and others demonstrating roles of these transporters in growth of primary tumors [19,22,24,62,63] and 3D spheroids [23]; and is consistent with our recent demonstration of the upregulation of these transporters in patient tumor tissue [20,24].

In conclusion, our work shows, for the first time, that NHE1 and NBCn1 are both important for cell cycle progression and proliferation of human breast cancer cells. Reduction of pH_e, significantly delayed cell cycle progression in MCF-7 breast cancer cells, as did KD of either NHE1 or NBCn1. The NHE1 protein level peaked in S phase and that of NBCn1 in G2/M phase. Steady state pH_i changed through the cell cycle in a manner dependent on both NHE1 and NBCn1. Accordingly, knockdown of either NHE1 or NBCn1 reduced proliferation, delayed cell cycle progression in a manner involving S phase prolongation and delayed G2/M transition, and altered the expression pattern and phosphorylation of cell cycle regulatory proteins. Our work demonstrates that both NHE1 and NBCn1 regulate cell cycle progression in breast cancer cells, and we propose that this involves phase-specific pH_i regulation by the two transporters.

Materials and methods

Antibodies and reagents

Antibodies against cyclin B1 (#12231), E2F1 (#3742), GAPDH (#2118), Ki-67 (#9449), phospho-H3 (pSer10) (#9706), phospho-Cdc2 (pTyr15) (#9111), and phospho-S6 (pSer 235/236) (4856) were purchased from Cell Signaling Technology. Antibodies against cyclin A2 (#sc-53229) and Na⁺/H⁺ exchanger (NHE1) (#sc-136239) were from Santa Cruz Biotechnology, and the antibody against cyclin D (#06–137) was from Millipore. Antibody against p150^Glued (#610473) was from BD Transduction Laboratories and the β-actin antibody (#A5441) and polyclonal goat anti-mouse immunoglobulins/HRP (#P0447) and polyclonal goat anti-rabbit immunoglobulins/HRP (#P0448) -conjugated secondary antibodies were from Dako. The polyclonal NBCn1 antibody was a kind gift from Jeppe Praetorius, Aarhus University, Denmark.

Cell lines and general cell culture

MCF-7 cells with stable KD of NHE1, NBCn1 or empty vector pLKO.1 [23] were grown in DMEM 1885-medium (incl.NaHCO₃⁻) (Substrat og SterilCentralen, Panum 22–2-24, #015) supplemented with 5% heat-inactivated fetal bovine serum (FBS; Gibco, #10 106–177)), 1% penicillin/streptomycin (Pen/Strep; Invitrogen, #5140–148), 1% MEM Non-Essential Amino Acids (Gibco/Invitrogen, #11140–035), and 1 μg/ml puromycin (Gibco, #A11138-02). Cells were grown at 37°C, 95% humidity, 5% CO₂ and passaged when a confluency of 70–80% was reached. Cell cultures were discarded when they reached passage 22.

Synchronization of cell cultures by double thymidine block

MCF-7 pLKO.1, NHE1 KD and NBCn1 KD cells were grown in either pH-manipulated media (pH 6.5, 7.4 or 7.6, containing puromycin) (Figure 1) or in plain KD growth media (pH 7.4, containing puromycin) (Figure 3). The pH-manipulated media were prepared from RPMI-1640 medium (Sigma-Aldrich, cat # R1383), for each pH replacing NaCl with the appropriate amount of NaHCO₃ to result in the desired pH at 5% CO₂ (pH 6.5: 3 mM HCO₃⁻, pH 7.4: 24 mM HCO₃⁻, pH 7.6: 38 mM HCO₃⁻). Cells were kept at 5% CO₂ at all times. Cells were seeded at 20% confluency prior to synchronization. The cells were blocked by addition of 2 mM thymidine (Sigma-Aldrich, #T1895) to the respective media for 17 h, arresting the cells at the G1/S phase or late S phase border. After 17 h the cells were released from the thymidine block by aspirating the media and washed in 37°C phosphate-buffered saline (PBS (137 mM NaCl, 2.7 mM KCl, 8.1 mM Na₂HPO₄-2H₂O and 1.5 mM KH₂PO₄)), followed by re-addition of fresh medium without thymidine for 10 h. The release allows the cells to progress to G2/M or G1, depending on their point of arrest. After the 10 h, 2 mM thymidine was added to the media for 17 h, allowing progression of the two cell populations through G1, ensuring a collective arrest at the G1/S phase border and a synchronized population. After the second thymidine block cells were released using trypsin and diluted with growth media (either pH manipulated or plain growth medium), to allow re-seeding. The samples were collected at 15 different time points over a course of 12 h, corresponding to different phases of the cell cycle. The samples were collected by trypsinizing and washing cells in ice-cold PBS with 1% FBS (Gibco, #10 106–177), followed by fixation and permeabilization with ice-cold 96% ethanol. The samples were kept on ice for 30 min or at −20°C until staining with Propidium Iodide (PI).

Flow cytometric analysis

Fixed cells were spun down and washed in ice-cold PBS with 1% FBS (Gibco, #10 106–177), followed by resuspension in ice-cold PBS containing 10 μg/mL PI (Invitrogen, #P3566) and 10 μg/mL DNase-free RNase (Sigma-Aldrich, #11119915001). The samples were transferred to polystyrene tubes (BD Falcon) and incubated at 37°C for 30 min. A negative sample only treated with RNAse was included to control for autofluorescence. Cell cycle distribution of synchronous samples was determined by flow cytometry analysis of nuclear DNA content using a Calibur flow cytometer and CellQuest software (BD Biosciences).

SDS-PAGE and immunoblotting

Cells were grown to 70–90% confluency, washed in ice-cold PBS and lysed in lysis buffer (1% SDS, 10 mM Tris-HCl, 1 mM NaVO₃, pH 7.5, heated to 95°C). The cell lysates were homogenized by sonication and centrifuged for 5 min at 20,000 g at 4°C to remove cell debris. Lysate protein content was determined (DC Protein Assay kit (Bio-Rad Laboratories, #500–0113, #500–0114, #500–0115)), equalized with ddH₂O, and NuPAGE LDS 4x Sample Buffer (Invitrogen, #NP0007) and Dithiothreitol (DTT) were added. Proteins were separated by SDS-PAGE under denaturing and reducing conditions. The gel electrophoresis was carried out using a Bio-Rad system, precast 10% Bis-Tris gels with either 18 or 26 wells (Bio-Rad, #567–1034 and #567–1035, respectively) and Tris/Glycine/SDS (10X, 25 mM Tris, 192 mM glycine, 0.1% SDS, pH 8.3, BioRad, #161–0732) Running Buffer, and Benchmark protein ladder (Invitrogen, #10747–012). Separated proteins were transferred to Trans-Blot® Turbo™, Midi format, 0.2 μm Nitrocellulose membranes (Bio-Rad, #170–4159) using Trans Blot® Turbo™ Transfer system (Bio-Rad, #10022518). Membranes were stained with Ponceau S (Sigma-Aldrich, #P7170-1 L), blocked in blocking buffer (5% nonfat dry milk in TBST (0.01 M Tris/HCl, 0.15 M NaCl, 0.1% Tween 20)) for 1 h at 37°C, incubated overnight at 4°C with primary antibodies diluted in blocking buffer, incubated with horseradish peroxidase conjugated secondary antibodies diluted in blocking buffer, washed in TBST and developed by chemiluminescence using Pierce® ECL Western blotting substrate (Thermo Scientific, #32106). Blots were developed on a Fusion FX developer (Vilber Lourmat), and the intensity of the bands was quantified using ImageJ software.

Immunocytochemical analysis

Cells were seeded on coverslips and grown to 80% confluency. The cells were washed in cold PBS, fixed for 30 min on ice in 4% paraformaldehyde (PFA, Sigma-Aldrich, #47608), washed again in PBS, permeabilized in 0.5% Triton-X-100 (Plusone, #17–1315-01) for 10 min and blocked in 5% Bovine Serum Albumin (BSA, Sigma-Aldrich, #A7906) for 30 min. The cells were incubated with primary antibodies over night at 4°C, washed in PBS and incubated with secondary antibodies for 1 h at room temperature. Cells were treated with DAPI (1:1000) (Invitrogen, #C10595) for 5 min to stain nuclei, washed in 1% BSA and mounted in 2% N-propyl gallate (Sigma-Aldrich, #P-3130) and sealed with nail polish. Fluorescence labeled proteins were visualized using Olympus BX63 epifluorescence microscope (40X objective). Overlays of the images and adjustments of the intensities were performed using ImageJ software.

Intracellular pH measurements

Synchronous MCF-7 KD cells were released from the double thymidine block, and the intracellular pH was measured at timepoints corresponding to early S (1.5 h after release), late S (3 h after release), G2 (7 h after release) or M phase (10 h after release), as determined from the initial flow cytometry analysis. In detail, cells were seeded in WillCo glass-bottom dishes (WillCo Wells, #3522) after release, and were loaded with 2ʹ,7ʹ-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein acetoxymethyl ester (BCECF-AM, 1.6 μM) (Invitrogen, #B1150) in growth medium for 30 min at 37°C, prior to sample collection at the indicated time points. Cells were washed once in HCO₃⁻ containing Ringer solution (118 mM NaCl, 25 mM NaHCO₃, 5 mM KCl, 1 mM MgSO₄, 1 mM Na₂HPO₄, 1 mM CaCl₂, 3.3 mM 3-(-N-morpholino)propanesulfonic acid (MOPS), 3.3 mM 2-[Tris(hydroxymethyl)-methylamino]-ethanesulfonic acid (TES), 5 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), adjusted with NaOH to pH 7.4 at 37°C), placed in a temperature controlled compartment (37°C) equipped with gas and solute perfusion. pH_i measurements were carried out using a Nikon Eclipse Ti microscope equipped with a Xenon lamp for fluorescence excitation, a 40x oil 1.4 NA objective and EasyRatioPro imaging software (PTI). Emission was measured at 520 nm after excitation at 440 nm and 485 nm. For acid loading, cells were exposed for 10 min to NaHCO₃ containing saline supplemented with 20 mM NH₄Cl. A high-K⁺ calibration solution (156 mM KCl, 1 mM MgSO₄, 1 mM K₂HPO₄, 3.3 mM 3-(-N-morpholino)propanesulfonic acid (MOPS), 3.3 mM 2-[Tris(hydroxymethyl)-methylamino]-ethanesulfonic acid (TES), 5 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), of pH 6.6, 7.0, 7.4 and 7.8, plus 10 µM nigericin Sigma-Aldrich, #N-7143) and a 4-point linear calibration curve was used to calibrate pH_i values. Fluorescence measured from the two excitation channels (440 nm and 485 nm) were corrected for their respective background fluorescence, which was assessed by measuring in a cell-free area during the experiment. The ratio 485 nm/440 nm was then calculated, and the calibration data was fitted to a linear function in the applied pH range, in which the experimental data was inserted and thereby converted to corrected pH values. Average pH_i measurements for each measured time-point and their standard deviations were calculated. The recovery rate was found by a linear fit to the initial phase of the pH_i recovery.

Manipulation of pH

MCF-7 KD cells were grown in control media (2.1 g RPMI-1640 (Sigma-Aldrich, R1383), 10 mM D-(+)-Glucose solution (Sigma-Aldrich, #G8644), 5% FBS (Gibco, #10 106–177), 1% Pen/Strep (Invitrogen, #15140–148), 1% MEM Non-Essential Amino Acids (Gibco/Invitrogen, #11140–035), and 1 μg/ml puromycin (Gibco, #A11138-02) with pH 6.5 (26 mM NaCl, 3 mM sodium bicarbonate solution 7.5% (NaHCO₃; Sigma-Aldrich, #S8761), 7.4 (17 mM NaCl, 12 mM sodium bicarbonate solution 7.5% (NaHCO₃; Sigma-Aldrich, #S8761) and 7.6 (38 mM sodium bicarbonate solution 7.5% (NaHCO₃; Sigma-Aldrich, #S8761) for 24 h, followed by synchronization under the same media conditions, and analyzation by flow cytometry.

Statistical analysis

All data are shown as representative images or as mean measurements with standard error of means (SEM) error bars, and represent at least 3 independent experiments unless stated otherwise. A two-tailed paired (when applicable, otherwise unpaired) Students t-test was applied to test for statistically significant differences between two groups. When compairing more than two groups a one-way analysis of variance (ANOVA) with Dunnett’s multiple comparison post-test was used. For comparisons between multiple groups with more than one factor (e.g. cell line and timepoint) a two-way ANOVA with repeated measures by both factors were used, with Dunnett’s multiple comparison post-test. For comparison of samples containing more than one factor, but only two groups, a two-way ANOVA with Sidak’s multiple comparison post-test was applied. *, **, ***, and **** denotes p < 0.05, p < 0.01, p < 0.001, and p < 0.0001, respectively.

Funding Statement

This work was supported by the Novo Nordisk Foundation (grant NNF16OC0023194) and the Independent Research Fund Denmark (grants 12-127290 and 12-126942), all to SFP.

Acknowledgments

We are grateful to Katrine F. Mark for excellent technical assistance, to Anna Fossum, BRIC, UCPH, for training and use of the flow cytometry facility, and to Jeppe Praetorius, Aarhus University, for his kind gift of NBCn1 antibody.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary materials

Supplementary materials for this article can be accessed here