Bone marrow (BM) metastasis of breast cancer (BC) can recur even decades after initial diagnosis and treatment, implying the long-term survival of disseminated cancer cells in a dormant state. Here we investigated the role of microRNAs (miRNA) transmitted from BM stroma to BC cells via gap junctions and exosomes in tumor cell quiescence. MDA-MB-231 and T47D BC cells arrest in G0 phase of the cell cycle when cocultured with BM stroma. Analyses of miRNA expression profiles identified numerous miRNAs implicated in cell proliferation including miR-127, -197, -222, and -223 targeting CXCL12. Subsequently, we showed that these CXCL12-specific miRNAs are transported from BM stroma to BC cells via gap junctions, leading to reduced CXCL12 levels and decreased proliferation. Stroma-derived exosomes containing miRNAs also contributed to BC cell quiescence, although to a lesser degree than miRNAs transmitted via gap junctions. This study shows that the transfer of miRNAs from BM stroma to BC cells might play a role in the dormancy of BM metastases. Cancer Res; 71(5); 1550–60. ©2011 AACR.

Breast cancer (BC) remains a clinical dilemma despite education, aggressive intervention, and early detection (1). A significant number of BC recurrences have identified bone marrow (BM) as the source of cancer cells (2–4), supporting the existence of dormant BC cells (BCC). BC resurgence has occurred without a history of primary tumor (5). Recent mathematical models indicate dormancy by a small number of BCCs with half-life of more than 12 years (6).

Metastasis of BC to BM correlates with poor prognosis (7, 8). BC could resurge from BM after more than 10 years of remission, in the absence of hematopoietic disruption (9–11). BCCs in BM would be best targeted if there is an understanding of how BCCs coexist within the microenvironment of BM. These form gap junctional intercellular communication (GJIC) with BM stroma (11), which comprise the hematopoietic niche close to the endosteum (12–19). The location of dormant BCCs in this area of BM might explain why high-dose chemotherapy and autologous stem transplantation failed (9, 20).

Connexin 43 (Cx43) contributes to the formation of GJIC between BCCs and stroma (11). CXCL12 production is decreased in BCCs when they contact stroma, and this is linked to reduced proliferation of BCCs as well as normal hematopoietic function (11). It is presumed that the latter occurs because CXCR4-expressing hematopoietic stem cells (HSC), instead of BCCs, interact with CXCL12-expressing stroma. This study investigates the role of specific microRNAs (miRNA) in the quiescence of BCCs in contact with stroma. miRNAs have been implicated in BC metastasis and to particular secondary organs (21–24).

CXCL12 belongs to the chemokine family and interacts with CXCR4 and CXCR7 (25). It is ubiquitously expressed and is constitutively produced in BM stroma (25, 26). miRNAs are small (19–23 nucleotide) noncoding RNA (27) that mostly target the 3′-untranslated region (UTR) of mRNAs with imperfect homology (27, 28). In this study, we show exchange of miRNAs across GJIC between BCCs and BM stroma and that specific miRNAs target CXCL12 to reduce the proliferation of BCCs.

Reagents

1-Octanol and protein G Sepharose were purchased from Sigma; FCS from Hyclone Laboratories; Platinum SYBR Green SuperMix and SuperScript III from Invitrogen (www.invitrogen.com); Western Blot Stripping Buffer from Thermo Scientific; propidium iodide (PI) from BD Biosciences; RNase A from Qiagen; and CXCL12 ELISA kit from R&D Systems.

Antibodies

Secondary antibodies [fluorescein isothiocyanate (FITC) and phycoerythrin conjugated), rabbit anti-human CXCR7, and rabbit anti-cyclin C were purchased from Abcam; murine anti-cyclin D1 and CDK from Cell Signaling; mouse anti-cytokeratin and mouse anti-β-actin monoclonal antibodies (mAb) from Sigma; murine anti-human CXCL12 mAb from R&D Systems; horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG from Santa Cruz Biotechnology; and rat anti-CD63/LAMP-3 and rabbit anti-acetyl histone H3 from Upstate Cell Signaling.

Cells

MDA-MB-231 and T47D were purchased from the American Type Culture Collection (ATCC) and propagated according to their instructions. Stroma was prepared from BM aspirates of healthy donors, 18 to 25 years of age, as described (29). The protocol was approved by the Institutional Review Board of University of Medicine and Dentistry of New Jersey. At confluence, adherent cells were subcultured in α-MEM with 20% FCS and after 3 passages were negative for CD14 and more than 99% positive for prolyl-4-hydroylase.

Cocultures of BCCs and BM stroma were described previously (30). Briefly, BCCs and stroma were cocultured at equal ratios in α-MEM with 10% FCS, in the presence or absence of 300 μmol/L 1-octanol (31).

Cell separation

Cells were deadhered with accutase, resuspended at 107/mL in PBS with 2% FCS and anti-cytokeratin (1/1,000), and then incubated at 0°C for 1 hour. After this, the cells were positively selected with Dynabeads M-450 goat anti-mouse IgG (Invitrogen). The negatively selected cells were stroma. BCCs were detached with DETACHaBEAD (Invitrogen). Flow cytometry analyses indicated more than 95% cytokeratin (+) for the Dynabead fraction and more than 99% prolyl-4-hydroxylase(+) for the stromal fraction.

Western analyses

Whole cell extracts were prepared and analyzed by Western blot, as described (32). Nuclear extracts were prepared with NE-PER nuclear and cytoplasmic extraction reagent (Thermo Scientific). CXCR7 was studied with membrane extracts, which was prepared with Qproteome plasma membrane protein extraction kit (Qiagen). Blots were developed overnight with anti-CXCL12 or anti-CXCR7 (at 1/1,000) and anti-cyclin C, anti-p21, and anti-cyclin D3 (at 1/500). The primary antibody was detected with HRP–anti-mouse IgG at 1/2,000 (PerkinElmer). Anti-histone H3 and β-actin were used at 1/1,000 and 1/4,000, respectively.

Real-time reverse transcriptase PCR for miRNAs

Predicted miRNAs for the 3′-UTR of CXCL12 were determined with miRanda algorithm and then verified by real-time reverse transcriptase PCR (RT-PCR), as described (33), with primers from Ambion with 5S rRNA as the normalized values. The miRNA array was done with total RNA from T47D and stroma that were cultured alone and in cocultures. RNA was sent to Rutgers University (New Brunswick, NJ) for analyses with the ABI TaqMan miRNA v1.0.

Vectors, pre-miRs, and anti-miRs

pMIR-R/Tac1/SG has been described (33). pCMV-DsRed-Express2 was purchased from Clontech. pEZX-MT01 reporter gene with the 3′-UTR of CXCL12 (hereafter pEZX-CXCL12) was purchased from GeneCopoeia. All pre-miRs and anti-miRs were purchased from Ambion.

Antisense exchange in cocultures

BCCs were transfected with pCMV-DsRed-Express2 and then selected with 600 μg/mL G418. After this, cells with bright red fluorescence (RFP) were sorted and then established as cocultures with stroma, as described (30, 34). The stromal cells were transiently transfected with 20 μmol/L of fluorescein-labeled antisense dsRED or scrambled sequence (Sigma), using Effectene (Qiagen). The antisense 5′ acc atg gat agc act gag aac g 3′ was complementary to the region spanning the translation start site of the DsRed mRNA. The 3′ end of the antisense was tagged with fluorescein (green) to track the passage into BCCs. After 48 hours, the passage of the antisense was confirmed by flow cytometry that detects FITC (not shown). The involvement of GJIC was determined in parallel cocultures with 300 μmol/L of 1-octanol.

Reporter gene assay

Reporter gene analysis was described previously (33). Briefly, BCCs were transfected with pMIR-R/Tac1/SG with FuGENE and, after 12 hours, retransfected with miR-130a and/or miR-206, using siPORT NeoFX (Ambion). After 16 hours, cell-free lysates were analyzed for total protein (Bio-Rad) and luciferase (Promega). Luciferase was normalized with β-gal activity and presented as normalized luciferase per microgram of total protein. Luciferase activity was determined in cotransfections (16 hours) of T47D with pEZX-CXCL12 and pre-miR-127, -197, -222, and/or -223.

Proliferation/viability assessment

Proliferation and viability were assessed in cocultures of T47D and stroma with CyQuant Cell Proliferation Assay Kit (Molecular Probes) and CellTiter-Blue Cell Viability Assay (Promega), respectively.

Day 1: T47D was transfected with the anti-miR or control anti-miR, using siPORT NeoFX. Day 2: Media were removed and the transfectants were established as cocultures. Day 3: Since GJIC requires ∼24 hours (11), proliferation was determined after 3 days of GJIC (day 6). Proliferation was calculated from a standard curve of T47D and stroma. Parallel studies were done with stroma alone, and the values were subtracted from total proliferation. The percent viability was calculated from a standard curve with untreated healthy cells assigned 100% viability. Cell-free wells with reagent alone were assigned 0% viability.

Cell cycle analyses

Cell cycle analyses were studied by PI incorporation and also by double labeling with dyes Pyronin Y and Hoechst 33342. In the former analysis, cocultures were incubated with anti-cytokeratin and FITC–anti-IgG secondary antibodies to label BCCs. Controls were labeled with nonimmune isotype or FITC–anti-IgG. In the case of PI labeling, the cells were treated with RNase A (1 mg/mL) and fixed with 70% ethanol and then labeled with 20 μg/mL PI solution. G0 and G1 phases of the cell cycle were distinguished by the consecutive addition of Hoechst 33342 (5 μg/mL) at 37°C for 45 minutes and 1 μg/mL of Pyronin Y (Sigma) for 15 minutes at 37°C. Flow cytometric analyses gated the cytokeratin (+) cells, at emissions 575 nm (Pyronin Y) and 450 nm (Hoechst 33342), and the data were analyzed with BD CellQuest software.

Data analyses

Statistical analyses were done with ANOVA and the Tukey–Kramer multiple comparisons test. P < 0.05 was considered significant. The predicted functional networks of miRNA in T47D from cocultures with stroma were generated by analyzing the validated and predicted targets with Ingenuity Pathways Analysis (IPA; Ingenuity Systems; www.ingenuity.com).

Cell cycle quiescence of BCCs in cocultures with stroma

BCC proliferation is decreased upon contact with BM stroma (11). This study examined T47D and MDA-MB-231 for cell cycle proteins in time-course studies with nuclear extracts. We observed time line decreases in cyclins D3 and C but increased p21 (Fig. 1A). The first time point, 12 hours, was selected to coincide with GJIC (11). Since the Western blots required robust manipulation, we asked the same questions without subjecting the cells to selection processes. We colabeled the cocultures (n = 4) with PI and anti-cytokeratin to identify BCCs and then analyzed the cells by flow cytometry. Cocultured T47D showed significant (P < 0.05) increase in G0/G1 phase (Fig. 1C and D). These findings correlated with light bands for cyclin D1 and CDK4 (Fig. 1E). We next discriminated G1 from G0 phase with Pyronin Y and Hoechst dye and observed more than 80% T47D in G0 phase (Fig. 1F).

Figure 1.

Cell cycle analyses of BCCs and passage of DsRed in cocultures. Western blots (n = 4) were done for cell cycle proteins in T47D (A) and MDA-MB-231 (B). Cell cycle analyses of cocultured T47D (n = 4) before (C) and after 36 hours (D). Western blots (n = 3) for cyclin D1 and CDK4 with T47D nuclear extracts of coculture before and after 36 hours (E). Pyronin Y/Hoechst labeling of T47D alone or 36-hour coculture (F). Fluorescence-sorted T47D or MDA-MB-231, stably transfected with pCMV-DsRed-Express2 (G), were established as coculture with stroma. The stroma was transfected with antisense or scrambled DsRed oligonucleotides (H) in the presence or absence of 300 μmol/L 1-octanol. Images are shown at 200× with arrows indicating the loss of RFP when BCCs contact stroma.

Figure 1.

Cell cycle analyses of BCCs and passage of DsRed in cocultures. Western blots (n = 4) were done for cell cycle proteins in T47D (A) and MDA-MB-231 (B). Cell cycle analyses of cocultured T47D (n = 4) before (C) and after 36 hours (D). Western blots (n = 3) for cyclin D1 and CDK4 with T47D nuclear extracts of coculture before and after 36 hours (E). Pyronin Y/Hoechst labeling of T47D alone or 36-hour coculture (F). Fluorescence-sorted T47D or MDA-MB-231, stably transfected with pCMV-DsRed-Express2 (G), were established as coculture with stroma. The stroma was transfected with antisense or scrambled DsRed oligonucleotides (H) in the presence or absence of 300 μmol/L 1-octanol. Images are shown at 200× with arrows indicating the loss of RFP when BCCs contact stroma.

Close modal

Passage of antisense DsRed via GJIC

We previously showed GJIC between BCCs and BM stroma by dye exchange, which correlated with a decrease in the proliferation of BCCs (11). In order to study the role of GJIC in BCC quiescence, we asked if small RNA can pass through GJIC, using antisense oligos. T47D and MDA-MB-231 were stably transfected with pCMV-DsRed-Express2 (Fig. 1G) and then cocultured with stroma that was transfected with antisense or scrambled oligo DsRED. In 3 experiments (Fig. 1H), we analyzed the fluorescence of DsRED at 200× magnification with EVOS digital microscope (Advanced Microscopy Group). BCCs that contacted stroma resulted in dim to undetectable fluorescence. Similar change was not observed for scrambled oligos or, with GJIC inhibitor, 1-octanol or carbenoxolone (not shown). In summary, the results indicated that antisense DsRed oligos can pass from stroma to BCCs through GJIC.

Passage of miRNA via GJIC between BCCs and stroma

In this set of studies, we investigated if miRNAs can pass through GJIC between BCCs and stroma. We addressed this question with a system in which the miRNAs (−130a and −206) were known to suppress luciferase in pMIR-R/Tac1/SG (33). The pre-miRs were transfected in stroma and pMIR-/Tac1/SG, in BCCs. After 48 hours, luciferase levels were significantly (P < 0.05) less than negative control pre-miR (Fig. 2A and B). This effect was reversed with 1-octanol. The results of both cell lines are combined in the appropriate experimental point (Fig. 2C). The decrease in luciferase by 1-octanol was not caused by degraded pre-miRs since real-time PCR showed no change in their levels (Fig. 2D).

Figure 2.

miRNA from BCCs to stroma. pMIR-R/Tac1/SG-transfected T47D (A) or MDA-MB-231 (B) were cultured alone or as cocultures with pre-miR-130a and/or pre-miR-206-transfected stroma (C). Neg, negative. Cocultures were done in the presence or absence of 300 μmol/L 1-octanol. After 48 hours, luciferase levels were quantitated and the normalized values presented as mean ± SD, n = 4. The stability of pre-miRs in the presence of 1-octanol was studied by real-time PCR (D). *, P < 0.05 versus miRNA transfectants; **, P < 0.05 versus cocultures with pre-miR-130a or pre-miR-206; ***, P > 0.05 versus cocultures with 1-octanol; *** P > 0.05 versus open bar.

Figure 2.

miRNA from BCCs to stroma. pMIR-R/Tac1/SG-transfected T47D (A) or MDA-MB-231 (B) were cultured alone or as cocultures with pre-miR-130a and/or pre-miR-206-transfected stroma (C). Neg, negative. Cocultures were done in the presence or absence of 300 μmol/L 1-octanol. After 48 hours, luciferase levels were quantitated and the normalized values presented as mean ± SD, n = 4. The stability of pre-miRs in the presence of 1-octanol was studied by real-time PCR (D). *, P < 0.05 versus miRNA transfectants; **, P < 0.05 versus cocultures with pre-miR-130a or pre-miR-206; ***, P > 0.05 versus cocultures with 1-octanol; *** P > 0.05 versus open bar.

Close modal

miRNA levels in T47D and stroma before and after coculture

We focused on T47D to investigate miRNA passage from stroma to BCCs. The first set of studies analyzed miRNA profile in T47D and stroma, before and after coculture, using ABI TaqMan human miRNA arrays (v1.0). The raw cycle threshold (Ct) values before and after coculture are shown in Figure 3A. The circled points indicate low expression before coculture (high Ct values) and high expression after coculture (low Ct values). The increased miRs in stroma suggested that they were induced after contact with T47D.

Figure 3.

Pathway analyses of miRNA array. TaqMan miRNA array was done with T47D and stroma before and after cocultures (Cts). The changes in miRNA were analyzed with Ingenuity Pathway program (A). T47D with more than 25-fold increase after cocultures were plotted as a histogram with the corresponding values in stroma (B). T47D with more than 10-fold increase in coculture were uploaded into the Path Designer graphical application and then overlaid onto a global molecular network developed from information contained in Ingenuity's knowledge base. The image shows direct, experimentally determined relationships between miRNA and other molecules (C and D). Cell cycle analyses of T47D after culture for 48 hours with exosome-free stroma-conditioned media (E) or cultured with exosome-containing media (F). F, inset (, immunoprecipitation with anti-CD63.

Figure 3.

Pathway analyses of miRNA array. TaqMan miRNA array was done with T47D and stroma before and after cocultures (Cts). The changes in miRNA were analyzed with Ingenuity Pathway program (A). T47D with more than 25-fold increase after cocultures were plotted as a histogram with the corresponding values in stroma (B). T47D with more than 10-fold increase in coculture were uploaded into the Path Designer graphical application and then overlaid onto a global molecular network developed from information contained in Ingenuity's knowledge base. The image shows direct, experimentally determined relationships between miRNA and other molecules (C and D). Cell cycle analyses of T47D after culture for 48 hours with exosome-free stroma-conditioned media (E) or cultured with exosome-containing media (F). F, inset (, immunoprecipitation with anti-CD63.

Close modal

We next normalized the Ct values of all miRNAs to one of the internal controls (RNU48) in the ABI array (ΔCt). The ΔCt values before coculture were subtracted from the ΔCt after coculture (ΔΔCt) and those with more than 25-fold increases are presented in Figure 3B. The miRNAs levels in T47D and stroma of cocultures were reciprocal, suggesting their movement from stroma to T47D. IPAs of literature-validated targets studied the predicted functions of the miRNAs (Fig. 3B) and their network (Fig. 3C). We next selected those that showed more than 10-fold change and then submitted to IPAs. We selected the top 3 of 9 networks and edited them to focus on relationships associated with differentiation (Fig. 3D). On the basis of the data (Fig. 3C and D), we deduced that the increases in miRNAs in T47D were linked to cell cycle quiescence and functions and blunted cell differentiation.

We next determined whether exosomes, produced from stroma, contributed to cell cycle quiescence of T47D. We collected stroma-conditioned media during a 48-hour period and prepared exosome-free and exosome-containing media (method in Supplementary S1). The 2 sets of media were added to T47D, and after 48 hours we studied cell cycle analyses with PI. There was a trend toward G2/S phase with the exosome-free conditioned media (Fig. 3E and F). The presence of exosomes was confirmed by immunoprecipitation with anti-CD63 (Fig. 3F inset) and Western blot (not shown). In summary, the studies suggested that stroma-derived exosomes could have a minor role in T47D quiescence.

Changes in CXCL12-predicted miRNAs

The section focused on CXCL12 miRs in mechanistic studies to understand cell cycle quiescence. We previously showed a decrease in CXCL12 in BCCs after contact with stroma (11). miRanda algorithm (35) predicted clusters of miR-127, -197, -222, and -223 at several sites within the 3′-UTR of CXCL12 (Fig. 4A). Real-time PCR indicated high levels of miR-127, -222, and -223 in stroma but low to undetectable in T47D (Fig. 4B). After cocultures, all 4 miRs were significantly (P < 0.05) increased in T47D (open bars) as compared with in stroma (diagonal bars; Fig. 4C). These increases were reversed with 300 μmol/L 1-octanol (Fig. 4D, solid bars) but remained at high levels in stroma, except for miR-197 (Fig. 4D, diagonal bars). These results indicated that CXCL12-predicted miRNA can pass from stroma to T47D via GJIC.

Figure 4.

CXCL12 miRNAs in T47D and stroma before and after cocultures. miRanda algorithm predicted the relative positions of CXCL12 miRNAs (A). Real-time PCR quantitated the levels of the predicted miRNAs in T47D and BM stroma cultured alone (B) or from cocultures, in the absence (C) or presence of 300 μmol/L 1-octanol (D). Cells from cocultures were separated after 48 hours. Results are presented as fold change over internal control, mean ± SD, n = 4. *, P <0.05 versus Stroma; **, P < 0.05 versus T47D.

Figure 4.

CXCL12 miRNAs in T47D and stroma before and after cocultures. miRanda algorithm predicted the relative positions of CXCL12 miRNAs (A). Real-time PCR quantitated the levels of the predicted miRNAs in T47D and BM stroma cultured alone (B) or from cocultures, in the absence (C) or presence of 300 μmol/L 1-octanol (D). Cells from cocultures were separated after 48 hours. Results are presented as fold change over internal control, mean ± SD, n = 4. *, P <0.05 versus Stroma; **, P < 0.05 versus T47D.

Close modal

Validating the predicted CXCL12 miRNAs

We studied the specificity of the predicted CXCL12 miRNA, using a reporter gene vector (pEZX-CXCL12) with the 3′-UTR of CXCL12, downstream of luciferase. T47D were transfected with pEZX-CXCL12 and/or pre-miR-127, -197, -222, or -223. Each pre-miR resulted in significant (P < 0.05) decrease in luciferase as compared with negative control pre-miRs (Fig. 5A). Pre-miR-197 caused significant (P < 0.05) reduction in luciferase as compared with the other pre-miRs (Fig. 5A).

Figure 5.

Validating CXCL12 miRNAs. T47D was transfected with pEZX-CXCL12 and/or pre-miR-127, -197, -222, -223, or negative (Neg) control pre-miR. After 48 hours, luciferase activity was quantitated and presented as mean ± SD, n = 4 (A). Cocultures (n = 4) were established with pEZX-CXCL12–transfected T47D and stroma, transfected with test or Neg control pre-miR, and luciferase activity presented as above (B). Shown is a cartoon of relative overlap of predicted miRNAs within the 3′-UTR of CXCL12 (C). The studies in “A” were repeated (n = 4) with combined pre-miRs, based on “C” and the results presented as above (D). Representative Western blots for CXCL12 (n = 3) with membrane extracts from cocultures, with or without 300 μmol/L 1-octanol (E). Western blots for CXCR7 with membrane extracts from T47D and MDA-MB-231 (F). *, P < 0.05 versus transfection with miRNAs; **, P < 0.05 versus other miRNA transfectants; ***, P < 0.05 versus pre-miR.

Figure 5.

Validating CXCL12 miRNAs. T47D was transfected with pEZX-CXCL12 and/or pre-miR-127, -197, -222, -223, or negative (Neg) control pre-miR. After 48 hours, luciferase activity was quantitated and presented as mean ± SD, n = 4 (A). Cocultures (n = 4) were established with pEZX-CXCL12–transfected T47D and stroma, transfected with test or Neg control pre-miR, and luciferase activity presented as above (B). Shown is a cartoon of relative overlap of predicted miRNAs within the 3′-UTR of CXCL12 (C). The studies in “A” were repeated (n = 4) with combined pre-miRs, based on “C” and the results presented as above (D). Representative Western blots for CXCL12 (n = 3) with membrane extracts from cocultures, with or without 300 μmol/L 1-octanol (E). Western blots for CXCR7 with membrane extracts from T47D and MDA-MB-231 (F). *, P < 0.05 versus transfection with miRNAs; **, P < 0.05 versus other miRNA transfectants; ***, P < 0.05 versus pre-miR.

Close modal

We next determined if luciferase activity could be similarly decreased if the pre-miRs were transfected in stroma and pEZX-CXCL12 in T47D. Indeed, after 48 hours of coculture, luciferase levels were significantly (P < 0.05) decreased as compared with negative control miR (Fig. 5B). This indicated that the miRNAs passed from stroma to T47D. We next asked if the changes in luciferase could be caused by stroma-derived factors. We therefore placed pEZX-CXCL12-transfected stroma in fresh or stroma-conditioned media and observed similar luciferase activity: 2,426 ± 44 versus 2,421 ± 25 normalized luciferase, n = 3.

Because of predicted overlap of the miRNAs in the 3′-UTR of CXCL12 (Fig. 5C), we asked if combinations of pre-miRs can further decrease luciferase. The results showed significant (P < 0.05) reduction by more than 1 pre-miRs as compared with negative control pre-miR (Fig. 5D) and single transfectants (Fig. 5A). miR-197 was most efficient in reducing luciferase when combined with pre-miR-222 and pre-miR-223.

The effect of CXCL12-predicted miRNAs on CXCL12 protein

We further validated miR-127, -197, -222, and -223 for specificity to CXCL12 in Western blots with plasma membrane extracts from cocultured T47D and stroma. The results indicated undetectable band for T47D and a bright band when GJIC was interrupted with 300 μmol/L of 1-octanol (Fig. 5E). We next determined if the decrease in membrane-bound CXCL12 correlated with released protein. We knocked down the miRNAs in T47D with anti-miRs or further expressed them with pre-miRs and then established the T47D as cocultures with stroma. After 48 hours, the media were quantitated for CXCL12 by ELISA. The results showed significant (P < 0.05) increase when miRNAs were knocked down as compared with control anti-miR and untransfected cells (Table 1). There were similar or reduced CXCL12 levels by pre-miR as compared with untransfected T47D or control pre-miR (Table 1).

Table 1.

CXCL12 levels in anti-/pre-miR–transfected T47D in cocultures

Anti-/pre-miRsCXCL12, mean ± SD, pg/mL
Untransfected 63 ± 5 
Control anti-miR 52 ± 3 
Control pre-miR 65 ± 5 
Anti-miR-127 577 ± 21a 
Pre-miR-127 50 ± 4 
Anti-miR-197 597 ± 21a 
Pre-miR-197 28±6 
Anti-miR-222 462±28a 
Pre-miR-222 48 ± 4 
Anti-miR-223 532 ± 22a 
Pre-miR-223 40 ± 5 
Anti-/pre-miRsCXCL12, mean ± SD, pg/mL
Untransfected 63 ± 5 
Control anti-miR 52 ± 3 
Control pre-miR 65 ± 5 
Anti-miR-127 577 ± 21a 
Pre-miR-127 50 ± 4 
Anti-miR-197 597 ± 21a 
Pre-miR-197 28±6 
Anti-miR-222 462±28a 
Pre-miR-222 48 ± 4 
Anti-miR-223 532 ± 22a 
Pre-miR-223 40 ± 5 

NOTE: T47D was transfected with the anti-miR or pre-miR, listed in the table, or with control anti-miR or pre-miR. The transfectants were established as cocultures with BM stroma. After the end point of the proliferation assessment, media were collected for assessment of CXCL12 by ELISA.

aP < 0.05 versus anti-miR or untransfected T47D.

Although CXCR4, the receptor for CXCL12, is expressed on BCCs (36), CXCL12 can also interact with CXCR7 (37). This led us to investigate if CXCR7 was expressed on BCCs since this could make the cells more responsive to CXCL12, or act as a decoy by interacting with CXCL12, away from CXCR4 to prevent signaling. Western blot analyses showed CXCR7 on T47D and MDA-MB-231 (Fig. 5F).

Effect of CXCL12 miRNAs on T47D proliferation in coculture

MiR-127, -197, -222, and -223 were increased in cocultured T47D and showed specificity for CXCL12 (Figs. 4 and 5, Table 1). We therefore studied their role in the quiescence of cocultured T47D. This was determined in “gain of function” studies in which cocultured T47D was transfected with the pre-miRs. Each pre-miR showed significantly (P < 0.05) reduced proliferation as compared with negative control pre-miR (Fig. 6A). In “loss of function” studies, T47D was transfected with anti-miRs, and the results indicated significant (P < 0.05) increase in T47D proliferation as compared with negative control miRNA and untransfected T47D (Fig. 6B). There was no change in cell viability of transfectants. The proliferation was blunted by 1-octanol (not shown). In summary, miR-127, -197, -222, and -223 caused T47D proliferation in coculture with stroma.

Figure 6.

CXCL12 miRNAs in the proliferation of cocultured T47D. Cell count of cocultures with T47D, transfected with test or control (Neg) pre-miR, mean ± SD, n = 4 (A). Similar studies (n = 4) with anti-miR transfectants; P < 0.05 versus anti-miR (B). CXCL12 knockdown T47D (wt) or mut siRNA were transfected with test or negative control pre-miR. After 48 hours, cell proliferation was assessed (n = 4) and the data presented as above, *P < 0.05 versus wt siRNA; **, P < 0.05 versus Neg pre-miR (C). miRNAs were analyzed with the Ingenuity Pathway program for predicted and validated targets for miR-127, miR-222, and miR-223 (D) and miR197 for predicted targets (E). Cell cycle analyses (n = 3) were done in 48-hour cocultures with T47D, in the presence of 300 μmol/L 1-octanol (F) or 1 ng/mL CXCL12 (G). Each experiment used stroma from a different donor. *, P < 0.05 versus anti-miR.

Figure 6.

CXCL12 miRNAs in the proliferation of cocultured T47D. Cell count of cocultures with T47D, transfected with test or control (Neg) pre-miR, mean ± SD, n = 4 (A). Similar studies (n = 4) with anti-miR transfectants; P < 0.05 versus anti-miR (B). CXCL12 knockdown T47D (wt) or mut siRNA were transfected with test or negative control pre-miR. After 48 hours, cell proliferation was assessed (n = 4) and the data presented as above, *P < 0.05 versus wt siRNA; **, P < 0.05 versus Neg pre-miR (C). miRNAs were analyzed with the Ingenuity Pathway program for predicted and validated targets for miR-127, miR-222, and miR-223 (D) and miR197 for predicted targets (E). Cell cycle analyses (n = 3) were done in 48-hour cocultures with T47D, in the presence of 300 μmol/L 1-octanol (F) or 1 ng/mL CXCL12 (G). Each experiment used stroma from a different donor. *, P < 0.05 versus anti-miR.

Close modal

The next set of studies investigated if the miRNAs can cause off-target effects in the proliferation of T47D. We stably knocked down CXCL12 with wild-type (wt) or mutant (mut) siRNA and then transfected the cells with pre-miR or negative control pre-miR. As expected, after 48 hours, there was significantly (P < 0.05) reduced cell proliferation with wt siRNA as compared with mut siRNA (Fig. 6C). There was no off-target effect by miR-197, miR127, and miR-222 in the knockdown cells, since the outcome was similar to mut siRNA and pre-miR transfectants (Fig. 6C).

Since the miRNAs shown in Figure. 6A–C indicated roles for BCC quiescence, we determined if quiescence might involve other genes. IPA of a network with miR-127, -222, and -223 were established with predicted and validated targets. The addition of tags to highlight the molecules involved in cell cycle supported the roles of the miRs in cell cycle quiescence (Fig. 6D). miR-197 did not connect to the other miRNAs, but alone it was associated with cell cycle quiescence, proliferation, and growth (Fig. 6E).

To verify that the decrease in CXCL12 and GJIC caused cell cycle quiescence of cocultured T47D, we undertook 48-hour cocultures in the presence of 300 μmol/L 1-octanol or 1 ng/mL CXCL12. The latter was within the range of CXCL12, produced from untransfected T47D and anti-miR–treated T47D (Table 1). The coculture cells were deadhered with accutase and then colabeled with PI and FITC–anti-cytokeratin. While untreated T47D was mostly in the G1 phase (Fig. 1D), treatment with 1-octanol and CXCL12 caused cycling (Fig. 6F and G).

BCCs close to the endosteum of mice have long doubling times (34), consistent with dormant BCCs in BM (7, 8). We showed a role for miRNAs in BCC quiescence by demonstrating their passage through GJIC between BCCs and BM stroma. The passage of small nucleotides was shown, first with antisense and then with miRNA (Figs. 1G and H and 2). Array analyses indicated that miRNAs were significantly increased in cocultured T47D with concomitant decreases in stroma (Fig. 3A and B). This suggested that the miRNAs moved from stroma through GJIC to T47D (Fig. 3B). IPAs with validated and predicted functions indicated that the miRNAs could be involved in cell cycle quiescence (Fig. 3C and D). This prediction was expected because of the association of miRNAs with tumor suppressive and oncogenic functions (38). We identified 4 miRNAs that suppressed CXCL12 production in cocultured T47D to decrease cell proliferation and transitioned the T47D to G0 phase of the cell cycle. These findings explained our previous study that showed long doubling time by BCCs close to the endosteum of nude mice (34) and also explained why CXCL12 was decreased in coculture (11). The decreased CXCL12 in BCCs was relevant to BM homeostasis and also provided BCCs with an advantage to adapt dormancy. A decrease in CXCL12 prevents CXCR4-expressing BCCs to interact with endogenous HSCs; instead, the HSCs interact with the stromal microenvironment (39).

Cluster of the 4 miRs within the 3′-UTR of CXCL12 (Fig. 4A) supported their functional relevance (Fig. 5; Table 1). miR-197 appeared to be the most efficient with regard to reduction in CXCL12 translation (Fig. 5). However, miR-197 level was decreased in BCCs and stroma as separate cultures (Fig. 4A), both of which produced high levels of CXCL12. The role of miR-197 on CXCL12 was validated in transfection studies with pre-miR and anti-miR in T47D (Table 1) and stroma (not shown). The results on miR-197 indicated the relevance of this miR to HSC function. If miR-197 was increased in stroma, this would reduce CXCL12 to cause migration of HSCs from stroma (40, 41). Similarly, if miR-197 was increased in BCCs this would prevent the production of CXCL12 (11). We did not detect an increase in miR-197 in the array studies (Fig. 3B) because the value was greater than the upper level of detection. Among the 4 CXCL12 miRs, miR-197 level was the highest in cocultured T47D (Fig. 4C and D).

The role of CXCL12 miRNAs targeting CXCL12 in the quiescence of BCCs was studied in “gain of function” analyses with pre-miRs (Fig. 6A) and “loss of function” studies with anti-miR (Fig. 6B). Ingenuity network with miR-127, -222, and -223 predicted cells in G0/G1 phase (Fig. 6C). miR-197 did not link with the other 3 miRNAs, but alone it formed a network of genes with predicted functions of cell cycle quiescence and development (Fig. 6D). Together, the CXCL12-specific miRs showed relevance in decreasing BCC proliferation in coculture with stroma.

Taken together, we showed passage of miRNA via GJIC between T47D and BM stroma. We have identified a role for CXCL12 miRNAs in the quiescence of BCCs. Ongoing studies suggested that not all BCCs can form GJIC with stroma, but those with a primitive phenotype can. The studies partly showed how BCCs take advantage of the BM microenvironment for their survival. Moreover, the studies provided insights on the regulation of CXCL12 in stroma. miR-197 appeared to be most significant in decreasing CXCL12 translation. Therefore, undetectable miR-197, combined with the presence of 3 other miRs, indicated that this distribution of miRs in stroma could be significant for exacerbated CXCL12 production. An increase in CXCL12 could lead to mobilization of HSCs from BM. Future studies are under way to determine if a particular BCC subset can form GJIC with stroma for dormancy and whether miRNAs can be targeted to eradicate BCCs in BM without toxicity to endogenous BM cells. It appeared that stroma-derived exosomes could also be involved in the quiescence of T47D, similar to other studies that reported on reprogramming of hematopoietic progenitors (42). Targeting of miRNA is a valid form of therapy, based on a recent report of a phase I clinical trial with siRNA to patients with solid cancers (43). Although this study addressed dormancy with stroma, similar mechanisms could occur at other sites (44). Supplementary Figure S2 depicts the major findings, showing attainment of BCC quiescence by miRNA exchange and the possible contribution of exosomes. Also shown is the movement of CXCL12-specific miRNAs to BCCs for a decrease in proliferation and CXCL12 production. The cartoon also depicts the exchange of additional miRNAs with unidentified functions.

No potential conflicts of interest were disclosed.

We thank Dr. Sauyma Pant (Roche, Nutley, NJ) for her assistance on exosome isolation.

This work was supported by the Department of Defense (W81XWH-0810561 and W81XWH-0610689).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisementin accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Green
BB
,
Taplin
SH
. 
Breast Cancer Screening Controversies
.
J Am Board Fam Pract
2003
;
16
:
233
41
.
2.
Naume
B
,
Zhao
X
,
Synnestvedt
M
,
Borgen
E
,
Russnes
HG
,
Lingjaerde
OC
, et al
Presence of bone marrow micrometastasis is associated with different recurrence risk within molecular subtypes of breast cancer
.
Mol Oncol
2007
;
1
:
160
71
.
3.
Habeck
M
. 
Bone-marrow analysis predicts breast-cancer recurrence
.
Mol Med Today
2000
;
6
:
256
7
.
4.
Riethdorf
S
,
Wikman
H
,
Pantel
K
. 
Review: biological relevance of disseminated tumor cells in cancer patients
.
Int J Cancer
2008
;
123
:
1991
2006
.
5.
Karrison
TG
,
Ferguson
DJ
,
Meier
P
. 
Dormancy of mammary carcinoma after mastectomy
.
J Natl Cancer Inst
1999
;
91
:
80
5
.
6.
Willis
L
,
Alarcon
T
,
Elia
G
,
Jones
JL
,
Wright
NA
,
Tomlinson
IPM
, et al
Breast cancer dormancy can be maintained by small numbers of micrometastases
.
Cancer Res
2010
;
70
:
4310
7
.
7.
Braun
S
,
Pantel
K
,
Muller
P
,
Janni
W
,
Hepp
F
,
Kentenich
CRM
, et al
Cytokeratin-positive cells in the bone marrow and survival of patients with Stage I, II, or III breast cancer
.
N Engl J Med
2000
;
342
:
525
33
.
8.
Braun
S
,
Kentenich
C
,
Janni
W
,
Hepp
F
,
de Waal
J
,
Willgeroth
F
, et al
Lack of effect of adjuvant chemotherapy on the elimination of single dormant tumor cells in bone marrow of high-risk breast cancer patients
.
J Clin Oncol
2000
;
18
:
80
.
9.
Mansi
JL
,
Berger
U
,
McDonnell
T
,
Pople
A
,
Rayter
Z
,
Gazet
JC
, et al
The fate of bone marrow micrometastases in patients with primary breast cancer
.
J Clin Oncol
1989
;
7
:
445
9
.
10.
Demicheli
R
,
Terenziani
M
,
Bonadonna
G
. 
Estimate of tumor growth time for breast cancer local recurrences: rapid growth after wake-up?
Breast Cancer Res Treat
1998
;
51
:
133
7
.
11.
Moharita
AL
,
Taborga
M
,
Corcoran
KE
,
Bryan
M
,
Patel
PS
,
Rameshwar
P
. 
SDF-1apha regulation in breast cancer cells contacting bone marrow stroma is critical for normal hematopoiesis
.
Blood
2006
;
108
:
3245
52
.
12.
Kiel
MJ
,
Morrison
SJ
. 
Uncertainty in the niches that maintain haematopoietic stem cells
.
Nat Rev Immunol
2008
;
8
:
290
301
.
13.
Muller-Sieburg
CE
,
Deryugina
E
. 
The stromal cells' guide to the stem cell universe
.
Stem Cells
1995
;
13
:
477
86
.
14.
Dorshkind
K
. 
Regulation of hemopoiesis by bone marrow stromal cells and their products
.
Annu Rev Immunol
1990
;
8
:
111
37
.
15.
Chow
DC
,
Wenning
LA
,
Miller
WM
,
Papoutsakis
ET
. 
Modeling pO2 distributions in the bone marrow hematopoietic compartment. II. Modified Kroghian models
.
Biophys J
2001
;
81
:
685
96
.
16.
Chow
DC
,
Wenning
LA
,
Miller
WM
,
Papoutsakis
ET
. 
Modeling pO2 distributions in the bone marrow hematopoietic compartment. I. Krogh's model
.
Biophys J
2001
;
81
:
675
84
.
17.
Jang
YY
,
Sharkis
SJ
. 
A low level of reactive oxygen species selects for primitive hematopoietic stem cells that may reside in the low-oxygenic niche
.
Blood
2007
;
110
:
3056
63
.
18.
Sifri
ZC
,
Kaiser
VL
,
Ananthakrishnan
P
,
Wang
L
,
Mohr
AM
,
Hauser
CJ
, et al
Bone marrow failure in male rats following trauma/hemorrhagic shock (T/HS) is mediated by mesenteric lymph and modulated by castration
.
Shock
2006
;
25
:
12
6
.
19.
Guest
I
,
Uetrecht
J
. 
Drugs toxic to the bone marrow that target the stromal cells
.
Immunopharmacology
2000
;
46
:
103
12
.
20.
Taback
B
,
Giuliano
AE
,
Hansen
NM
,
Singer
FR
,
Shu
S
,
Hoon
DSB
. 
Detection of tumor-specific genetic alterations in bone marrow from early-stage breast cancer patients
.
Cancer Res
2003
;
63
:
1884
7
.
21.
Valastyan
S
,
Benaich
N
,
Chang
A
,
Reinhardt
F
,
Weinberg
RA
. 
Concomitant suppression of three target genes can explain the impact of a microRNA on metastasis
.
Genes Dev
2009
;
23
:
2592
7
.
22.
Ma
L
,
Young
J
,
Prabhala
H
,
Pan
E
,
Mestdagh
P
,
Muth
D
, et al
miR-9, a MYC/MYCN-activated microRNA, regulates E-cadherin and cancer metastasis
.
Nat Cell Biol
2010
;
12
:
247
56
.
23.
Liu
S
,
Goldstein
RH
,
Scepansky
EM
,
Rosenblatt
M
. 
Inhibition of Rho-associated kinase signaling prevents breast cancer metastasis to human bone
.
Cancer Res
2009
;
69
:
8742
51
.
24.
Yu
Z
,
Willmarth
NE
,
Zhou
J
,
Katiyar
S
,
Wang
M
,
Liu
Y
, et al
MicroRNA 17/20 inhibits cellular invasion and tumor metastasis in breast cancer by heterotypic signaling
.
Proc Natl Acad Sci U S A
2010
;
107
:
8231
6
.
25.
Claps
CM
,
Corcoran
KE
,
Cho
KJ
,
Rameshwar
P
. 
Stromal derived growth factor-1alpha as a beacon for stem cell homing in development and injury
.
Curr Neurovasc Res
2005
;
2
:
319
29
.
26.
Gazitt
Y
. 
Homing and mobilization of hematopoietic stem cells and hematopoietic cancer cells are mirror image processes, utilizing similar signaling pathways and occurring concurrently: circulating cancer cells constitute an ideal target for concurrent treatment with chemotherapy and antilineage-specific antibodies
.
Leukemia
2003
;
18
:
1
10
.
27.
Scherr
M
,
Eder
M
. 
Gene silencing by small regulatory RNAs in mammalian cells
.
Cell Cycle
2007
;
6
:
444
9
.
28.
Johnson
C
,
Sundaresan
V
. 
Regulatory small RNAs in plants
.
EXS
2007
;
97
:
99
113
.
29.
Joshi
DD
,
Dang
A
,
Yadav
P
,
Qian
J
,
Bandari
PS
,
Chen
K
, et al
Negative feedback on the effects of stem cell factor on hematopoiesis is partly mediated through neutral endopeptidase activity on substance P: a combined functional and proteomic study
.
Blood
2001
;
98
:
2697
706
.
30.
Ramkissoon
SH
,
Patel
PS
,
Taborga
M
,
Rameshwar
P
. 
Nuclear Factor-{kappa}B is central to the expression of truncated neurokinin-1 receptor in breast cancer: implication for breast cancer cell quiescence within bone marrow stroma
.
Cancer Res
2007
;
67
:
1653
9
.
31.
Boger
DL
,
Patterson
JE
,
Guan
X
,
Cravatt
BF
,
Lerner
RA
,
Gilula
NB
. 
Chemical requirements for inhibition of gap junction communication by the biologically active lipid oleamide
.
Proc Natl Acad Sci U S A
1998
;
95
:
4810
5
.
32.
Patel
SA
,
Heinrich
AC
,
Reddy
BY
,
Srinivas
B
,
Heidaran
N
,
Rameshwar
P
. 
Breast cancer biology: the multifaceted roles of mesenchymal stem cells
.
J Oncol
2008
;
2008:article 425895
.
33.
Greco
SJ
,
Rameshwar
P
. 
miRNAs regulate synthesis of the neurotransmitter substance P in human mesenchymal stem cell-derived neuronal cells
.
Proc Natl Acad Sci U S A
2007
;
104
:
15484
9
.
34.
Rao
G
,
Patel
PS
,
Idler
SP
,
Maloof
P
,
Gascon
P
,
Potian
JA
, et al
Facilitating role of preprotachykinin-I gene in the integration of breast cancer cells within the stromal compartment of the bone marrow: a model of early cancer progression
.
Cancer Res
2004
;
64
:
2874
81
.
35.
Betel
D
,
Wilson
M
,
Gabow
A
,
Marks
DS
,
Sander
C
. 
The microRNA.org resource: targets and expression
.
Nucleic Acids Res
2008
;
36
:
D149
53
.
36.
Corcoran
KE
,
Fernandes
H
,
Bryan
M
,
Taborga
M
,
Srinivas
V
,
Rameshwar
P
. 
Mesenchymal stem cells in early entry of breast cancer into bone marrow
.
PLoS One
2007
;
3
:
e2563
.
37.
Naumann
U
,
Cameroni
E
,
Pruenster
M
,
Mahabaleshwar
H
,
Raz
E
,
Zerwes
HG
, et al
CXCR7 functions as a scavenger for CXCL12 and CXCL11
.
PLoS One
2010
;
5
:
e9175
.
38.
O'Day
E
,
Lal
A
. 
MicroRNAs and their target gene networks in breast cancer
.
Breast Cancer Res
2010
;
12
:
201
.
39.
Nie
Y
,
Han
YC
,
Zou
YR
. 
CXCR4 is required for the quiescence of primitive hematopoietic cells
.
J Exp Med
2008
;
205
:
777
83
.
40.
Broxmeyer
HE
. 
Chemokines in hematopoiesis
.
Curr Opin Hematol
2008
;
15
:
49
58
.
41.
Dugan
MJ
,
Maziarz
RT
,
Bensinger
WI
,
Nademanee
A
,
Liesveld
J
,
Badel
K
, et al
Safety and preliminary efficacy of plerixafor (Mozobil) in combination with chemotherapy and G-CSF: an open-label, multicenter, exploratory trial in patients with multiple myeloma and non-Hodgkin's lymphoma undergoing stem cell mobilization
.
Bone Marrow Transplant
2009
;
45
:
39
47
.
42.
Ratajczak
J
,
Miekus
K
,
Kucia
M
,
Zhang
J
,
Reca
R
,
Dvorak
P
, et al
Embryonic stem cell-derived microvesicles reprogram hematopoietic progenitors: evidence for horizontal transfer of mRNA and protein delivery
.
Leukemia
2006
;
20
:
847
56
.
43.
Davis
ME
,
Zuckerman
JE
,
Choi
CH
,
Seligson
D
,
Tolcher
A
,
Alabi
CA
, et al
Evidence of RNAi in humans from systemically administered siRNA via targeted nanoparticles
.
Nature
2010
;
464
:
1067
70
.
44.
Demicheli
R
,
Biganzoli
E
,
Boracchi
P
,
Greco
M
,
Retsky
MW
. 
Recurrence dynamics does not depend on the recurrence site
.
Breast Cancer Res
2008
;
10
:
R83
.

Supplementary data