Primary, non-invasive and invasive breast carcinomas following radical mastectomy at different tumor stages (T1-4N0-1M0) of female human subjects (27 ductal carcinoma and 3 lobular carcinoma cases) belonging to varying age groups from 40–70 years were used for the study. Normal tissue adjacent to the tumor from the same breast was used as control. The nature of the tissue was confirmed by histopathological studies. Prior written consent from patients was obtained before tissue samples were collected following mastectomy and the samples were not exclusively used for this study and not meant for commercial use. The study and the method of tissue collection were approved by the “Institutional Ethics Committee for Biomedical Research”, Hyderabad, India and Osmania University.

After surgical resection (mastectomy), samples were collected in ice-cold Tris- buffered saline (10 mM Tris, pH 7.2 and 0.9% NaCl) and immediately processed for future analysis. Tissues were ground in Remi homogenizer on ice in tissue lysis buffer (40 mM Tris buffer, pH 7.2, 1% Triton X-100, 10% glycerol, 1 mM β-ME, 2 mM EGTA, 1 mM EDTA, 130 mM NaCl, 50 mM NaF, 1 mM Na₃VO₄ and protease inhibitor cocktail) and centrifuged at 16000×g for 20 min at 4°C. The supernatant obtained was used for western blotting analyses.

1 g of breast tissue (tumor and normal) derived after mastectomy was cut up manually into small pieces, approximately 0.5 cm cubed, collected and washed in ice-cold Kreb’s Ringer buffer (116 mM NaCl, 4.2 mM KCl, 2.5 mM CaCl₂, 1.6 mM NaH₂PO₄, 1.2 mM MgSO₄, 22 mM NaHCO₃, 11 mM glucose, pH 7.4). The tissue was digested overnight at 37°C in a shaking incubator at 50 rpm with 0.5 mg/ml collagenase (Type IV, sigma) in 5 ml RPMI medium (RPMI 1640 media containing 2 mM L- glutamine, 100 U of penicillin/ml, 100 µg streptomycin/ml and 10% fetal calf serum). Following enzyme digestion, the fat layer was decanted by centrifuging at 1000×g for 5 min at room temperature (RT) and washed several times with the buffer. The cell-rich supernatant was backflushed through 60 µ and 50 µ meshes to isolate breast organoids from red blood cells, fibroblasts and endothelial cells. The cells were disaggregated by pipetting in media containing 0.25% trypsin-EDTA followed by filtration through 40 µ meshes. The cell pellet was collected and washed three more times to yield a predominantly single cell suspension [20]. The final pellet was resuspended in 1 ml media and the cells were counted on cytometer after trypan blue staining. The cells were lysed in M2 lysis buffer (20 mM tris, pH 7.6, 0.5% NP-40, 250 mM NaCl, 3 mM EGTA, 3 mM EDTA, 5 mM NaF, 5 mM Na₃VO₄ and 5 µg/ml protease inhibitor cocktail) and the lysate was cleared by centrifuging at 12000×g for 15 min at 4°C.

MCF7 cells were purchased from NCCS (National Center for Cell Science), Pune, India (previously published in [21]). MCF7 cells were grown in MEM supplemented with 0.01 mg/ml Insulin, 10% fetal bovine serum and 1% antibiotic-antimycotic solution. MCF7 cells transfected with profilin 1 and mutants were maintained in the above media supplemented with 800 µg/ml G418. Insulin and G418 were purchased from Sigma-Aldrich, USA.

Total protein in clear supernatants obtained from breast tissue and cell lysates was estimated by Bradford’s method according to the manufacturer’s protocol (Sigma-Aldrich). 50 µg protein from tissue lysates and cell lysates was subjected to 10% tricine SDS-PAGE [22] and transferred onto nitrocellulose membrane. The membrane was blocked with 3% gelatin in TTBS buffer (10 mM Tris, 154 mM NaCl and 0.1% Tween-20, pH 7.5) for 1 hr, then washed with TTBS (4×10 ml) and incubated with rabbit anti-profilin 1 antibodies (1∶1000 dilution, Novus biologicals) or rabbit anti-PKCζ antibodies (1∶500, in-house) and anti-actin antibodies (1∶3000, Sigma) for 2 hr. After washings with TTBS, the blot was incubated with horse radish peroxidase conjugated secondary anti-rabbit/anti-mouse goat IgG for 2 hr. The blot was then colorimetrically developed using NBT/BCIP (Amersham Pharmacia Biotech).

GFP-fusion constructs of profilin were constructed. GST clones of profilin-1 WT, R74E and H133S (characterized by [23], [24] were subcloned into EGFP-C1 vector using forward primer: 5′-AGC CAA GCT AGT GCC GGG TGG AAC GCC-3′ and reverse primer: 5′-CAG GGA TCC TCA GTA CTG GGA ACG CCG AAG-3′. Similarly, S137A from pcDNA3 construct [25] was subcloned into EGFP-C1 vector. All inserts were cloned between HindIII and BamH1 restriction sites (forward primer: 5′-CCA AGC TTC GAT GGC CGG GTG GAA CG-3′ and reverse primer: 5′-GCGGATCC CTA GTA CTG GGC ACG CCG A-3′). All the constructs including empty vector (eGFP-N1) were confirmed by plasmid sequencing (Invitrogen). These constructs were then overexpressed by transfection in MCF7 cells.

MCF7 cells were transfected with a control EGFP-N1 vector or following profilin1 expressing constructs: GFP-PFN1-WT (Wild-Type, human), GFP-PFN1-S137A (serine 137 phosphorylation mutant), GFP-PFN1-R74E (actin mutant) or GFP-PFN1-H133S (PLP binding mutant). For transfection, 0.5 µg plasmid DNA was mixed along with 1 µl of transfection reagent (chloroform:DT-DEAC lipid micelles made in laboratory) [26] in a total volume of 200 µl serum-free MEM media. The transfection mix was incubated at room temperature for 30 min, following which it was added to 20,000 MCF7 cells pre-incubated in serum-free MEM media for 1 hr. After 5 hr of transfection, media was changed to complete MEM media and cells were allowed to grow for 24 hr for protein expression. Selection media with 800 µg/ml G418 was added and polyclonal populations were selected over a period of two weeks.

To determine cell growth of transfected clones, 2000 cells of each clone were plated in triplicate in 96-well plates and cell viability was assessed by MTT (3-[4,5-dimethylthiazol-2-yl]- 2,5-diphenyltetrazolium bromide) assay at 24 hr, 48 hr, 72 hr and 7 days. MTT reagent (diluted from a 4 mg/ml solution in PBS) was added to all the wells at a final concentration of 0.8 mg/ml and the cells were further incubated for 4 hr at 37°C. The reaction was terminated by adding 200 µl/well dimethyl sulfoxide (DMSO) and absorbance was read at 570 nm in an ELISA plate reader. Wells containing complete media alone was used as blank and the average values subtracted from the average values of all clones. The assay was repeated three times and analyzed using two-sided Student’s t test. Data points represent the mean±s.d and p<0.005 was considered significant.

Anchorage independent growth was assayed by the soft agar growth assay as described elsewhere [27]. The first step involved plating a bottom layer of 0.6% agar in serum-free media in 12-well plates. The plates were incubated at room temperature for 30 min to solidify the agar. Cells were harvested by trypsinization and 1×10⁴ cells were mixed with 0.3% agarose (made in complete media) and layered carefully on the top of existing 0.6% agarose. The plates were covered with 1 ml of medium supplemented with 10% FBS and incubated at 37°C in a 5% CO₂ incubator for 10days. Complete media was added every 2 days to the test wells. At the end of 10 days, cell colonies with >100 µm in diameter were imaged and counted under a microscopic field at 2.5× magnification. Mean colony count was based on numbers from 5 different fields imaged under 2.5× magnification from two independent experiments done in duplicate for each treatment condition and was analyzed using two-sided Student’s t test.

5×10⁴ cells were plated in a 6-well plate in triplicate and allowed to grow to confluency. Cells were washed and left in serum-free media for 5 hr. A scratch was made and media was added that was either only serum-free or with 200 nM PDBu (phorbol 12,13-dibutyrate) or with 20 µM LY294002/200 nM PDBu. Images were captured at 10× magnification using Leica Microscope at zero hour and again at 24 hr. The scratch area at zero hour and empty area not covered by migrated cells after 24 hr was analyzed with ImagePro software. The assay was repeated three times. Data points represent the mean±s.d. from five fields of uncovered scratch area from three independent experiments and p<0.05 was considered significant.

Invasion assay was performed as per the protocol published by Dasgupta et al [27]. MCF7 overexpressing profilin 1 and its mutant clones were harvested by trypsinization and 1×10⁵ cells were placed in the top chamber of boyden filters as untreated/treated with 200 nM PDBu or a combination of 20 µM LY294002 and 200 nM PDBu in serum-free media. The filters were pre-coated with 100 µg/ml ECM (BD Biosciences) and incubated for 4 hr at 37°C before adding cells and the assay was performed according to the protocol [27]. Invaded cells were stained with hematoxylin and images for 6 different fields per treatment were captured using 40× magnification using a Leica microscope. The assay was repeated twice in duplicate and analyzed using two-sided Student’s t test. Data points represent the mean±s.d. and p<0.05 was considered significant.

Asynchronously growing MCF7 clones overexpressing profilin and its mutants were treated with hypotonic buffer (1 mg/ml sodium citrate and 0.3% Igepal) for 20 min. The buffer digests the cytoplasm and nuclei detach from the plate. Nuclei are gently pipetted a few times and collected in an eppendorf for Green Fluorescence acquisition [28]. Similarly, 10⁵ cells of each clone were plated and serum starved for 24 hr. Cells were then treated with DMSO control/200 nM PDBu/a combination of 20 µM LY294002 and 200 nM PDBu for 24 hr. Nuclei were harvested and subjected to flow cytometry for Green Fluorescence acquisition. The data was acquired with flowcytometer (Millipore Guava 8 HT easycyte). 50000 events were acquired with flow rate of 0.59 µl/s. Data was analyzed by FlowJo software. Experiments were done three times in duplicate.

MCF7 clones (40,000cells) were plated on cover slips (Corning) in a 6-well plate and cultured for 24 hr to visualize localization of overexpressed profilin 1 and its mutants. The cells were fixed in 2% para-formaldehyde for 15 min. The nuclei were stained with DAPI. Cells were mounted on glass slides. Images were captured at 63× oil immersion magnification with a Leica confocal microscope [29]. Images are representative of three independent experiments.

5×10⁵ cells were plated in a T25 flask and allowed to grow for a day. Cells were placed in serum-free media for 5 hr and then treated with either DMSO control media or 200 nM PDBu for 24 hr. Conditioned media were collected and centrifuged to remove debris. Samples were concentrated 10times using a centricon (Millipore, Bedford, MA) from 2 ml to 0.2 ml. Zymography was performed as described previously by Toth M et al., [30] with a little modification. Samples were loaded on a 10% SDS-polyacrylamide gel incorporated with 0.1% gelatin (Sigma) for electrophoresis. MMP2 and MMP9 zymographic standards were used as positive controls (Chemicon). After electrophoresis, MMPs were allowed to renature by washing the gel twice in 50 ml of 2.5% Triton X-100 for 30 min, incubated for 18 hr at 37°C in developing buffer, and stained for 1 hr with 0.5% Coomassie brilliant blue G250. After destaining gelatinolytic activity was visualized as a transparent band against a blue background. Gels were imaged using a digital imaging analysis system (Bio-Rad). The assays were repeated three independent times.

All data was analyzed statistically by two sided student’s t test in Microsoft Office 2007 and statistical significance was estimated. Densitometric analysis of western blots was carried out with ImageJ software. Flow cytometry data was analyzed using FlowJo software.

Published results on levels of profilin in different cancer tissues/cell lines are contradictory. Western blotting analyses was performed on lysates prepared from patient samples for breast cancer and adjacent normal breast tissues. Profilin 1 levels were significantly higher in breast cancer tissues compared to their adjacent normal counterparts (Figure 1A). Densitometric analysis of the western blots revealed profilin 1 was 3.6±2.5-fold (*p<0.01) elevated in cancer tissues (Figure 1B). Simultaneously, primary epithelial cells were isolated from the same patient samples and lysates were subjected to western blot analysis for profilin 1. A representative western blot is presented in Figure 1D and densitometric analysis of the western blots show a 1.6±0.3- fold (*p<0.03) increase in profilin 1 levels in cells isolated directly from the breast cancer tissues when compared to adjacent normal tissues (Figure 1E).

Post-translational modification such as phosphorylation can alter the function of cytoskeletal-associated proteins in tumorigenesis; since profilin 1 is known to undergo phosphorylation at Ser137 preferentially by PKCζ, an enzyme with potential role in breast cancer cell chemotaxis and tumor invasion [31]–[33]; we also examined PKCζ levels by western blotting in breast cancer and adjacent normal tissues (Figure 1A, middle panel). Figure 1C is a densitometric representation of increased PKCζ expression (1.6±0.5-fold, *p<0.05) in breast cancer tissues. Elevated protein levels of PKCζ in breast cancer tissues were indicative of a phosphorylation event targeting profilin 1 too. Therefore, we studied the probable role of profilin 1 phosphorylation at serine residue (S137) in breast cancer.

Since profilin 1 levels were high in patients with breast cancer, we overexpressed profilin 1 and some of its mutants in MCF7 cells to study their effect on cancer progression. MCF7 cells were transfected with each of the following vectors individually: EGFP-N1 (Empty Vector, designated as EV), GFP-Profilin 1-wild type (PFN-WT), GFP-Profilin 1-S137A (mutant for serine 137 phosphorylation site; PFN-S137A), GFP-Profilin 1-R74E (mutant for actin-binding site, PFN-R74E) and GFP-Profilin 1-H133S (mutant for poly-L-proline (PLP) sequence binding, PFN-H133S). Expression of profilin clones was studied by confocal microscopy. Left panel depicts images of GFP-Profilin and its mutants, middle panel shows DAPI images and right panel shows merged images for each clone (Figure 2). Each clone showed varying localization of GFP-Profilin. EV expression was observed in the nucleus and cytoplasm. PFN-WT, PFN-S137A and PFN-H133S showed cytoplasmic and nuclear expression. Of these, PFN-H133S was predominantly cytoplasmic and PFN-R74E was mostly nuclear with little cytoplasmic expression. These results indicate that interaction of profilin with actin is essential to maintain profilin in the cytoplasm. PFN-H133S showed the least expression of nuclear profilin suggesting that interaction with PLP-containing proteins could be essential for its nuclear translocation. Simultaneously, lysates were prepared from MCF7 cells expressing different clones of profilin and subjected to western blotting with anti-profilin antibodies. EV did not show over-expressed profilin whereas all the clones showed a band corresponding to GFP-profilin (41 kDa). Endogenous cellular profilin 1 (14 kDa) was observed in all the lanes including EV and actin was used as loading control (42 kDa) (Figure 3A). Next, we determined the nuclear expression alone of GFP-profilin and mutants in asynchronously growing MCF7 cells by subjecting them to hypotonic solution that bursts the cytoplasm but leaves the nuclei intact and detached from the plate. The nuclei are collected as a single cell suspension and subjected to flow cytometric analysis (Figure 3B). Untransfected MCF7 cells were used to gate the positive nuclei from the clones. All the clones showed similar numbers of nuclei that were positive for profilin 1. Figure 3C–3G show the flow cytometry dot plots of GFP-positive nuclei from asynchronously growing EV (96.1%), PFN-WT (85.9%), PFN-S137A (92.8%), PFN-R74E (94.6%) and PFN-H133S (91.3%) respectively. These clones were then selected in neomycin (G418) to obtain a 100% positive polyclonal population overexpressing profilin 1 and its mutants respectively.

We first assessed the growth ability of transfected cells in culture at 24 hr, 48 hr, 72 hr and 7days of plating in complete media by MTT assay (Figure 4A). PFN-WT demonstrated the highest viability at all time points studied followed by PFN-R74E and PFN-H133S. PFN-S137A showed the least growth rate and was comparable to EV transfected MCF7 cells (***p<0.005). Anchorage independency is one of the hallmarks of cancer and this feature allows tumor cells to invade adjacent tissues and lead to metastasis [34]. To determine which of the profilin 1 mutants might interfere with the tumorigenic property of MCF7 cells, soft agar assay was performed and colonies above 100 µm where counted and plotted. Figure 4B shows representative images of colonies in soft agar of different mutants of profilin 1 and figure 4C is a graphic representation of the average number of colonies for each clone from 5 different fields under 2.5× magnification. We observed that MCF7 cells expressing PFN-WT formed significantly more and larger colonies in soft agar compared to EV-expressing MCF7 cells (**p<0.01). PFN-S137A formed smaller and less number of colonies than EV and 42 percent less number of colonies (***p<0.005) than PFN-WT, thereby indicating that (S137) phosphorylation of profilin 1 facilitates anchorage independency in breast cancer cells. Actin mutant (PFN-R74E) formed 70 percent more colonies (**p<0.01) than PFN-WT, consistent with previous reports demonstrating that profilin 1 defective in actin binding ability promotes tumorigenesis [17]. Defect in PLP binding ability of PFN (PFN-H133S) did not allow any colony formation in soft agar, though the viability of these cells in monolayer cultures was fairly comparable to PFN-R74E till 72 hr and decreasing by 7days. These results were in corroboration with Wittenmeyer’s study [17] reporting the inability of PFN-H133S to form tumors in mice.

Whether phosphorylation of profilin 1 affects migration and invasion of breast cancer cells expressing various profilin 1 mutants was next assessed by wound healing assays and boyden chamber assays respectively. PKCζ activity can be reduced by the pharmacological inhibitors of PI3K. Previous in vivo studies from our lab have demonstrated that phorbol 12,13-dibutyrate (PDBu), a protein kinase C activator (PKC), phosphorylates profilin 1 at Serine 137 and the effect is reversed by LY294002, an inhibitor of PI3K and subsequently PKCζ [12], [35]. We studied their effect on profilin 1 mediated migration and invasion. MCF7 cells overexpressing profilin 1 and its mutants were grown to confluency and a scratch was made with a pipet tip. Images were captured at 20× magnification of the scratch at zero hour for each treatment. After 24 hr following treatments, images were captured and non-migrated area was analyzed. Figure 5 shows representative images of the wound and migrated cells into the wound for all the clones. All clones migrated to some extent into the wound at 24 hr in serum-free media. PDBu induced significant migration in all clones but not in the PFN-S137A mutant clone, therefore demonstrating that phosphorylation of profilin 1 at Serine 137 is an important event in promoting migration of MCF7 cells. PFN-WT and PFN-R74E expressing cells migrate more than EV expressing cells upon treatment with PDBu (**p<0.05). Pre-treatment of cells with LY294002 significantly abrogated migration of MCF7 cells overexpressing PFN-WT, PFN-R74E and PFN-H133S (**p<0.05). PFN-S137A did not show much change in migration at 24 hr in either serum-free media or PDBu or LY/PDBu treated cells (***p<0.005) (Figure 6A). This result reinforces the fact that serine phosphorylation of profilin 1 facilitates migration of breast cancer cells.

Similarly, invasion assays using ECM (extracellular matrix) coated boyden chamber filters revealed that PFN-WT and PFN-R74E cells invade more than EV cells in the presence and absence of PDBu. PFN-R74E showed more invasiveness than PFN-WT (**p<0.05). PFN-H133S invades the least (**p<0.05). However, PDBu- treated PFN-H133S showed 3 times more invasion when compared to serum-free media, indicating that though PFN-H133S is defective in binding to any PLP containing proteins, phosphorylation at Ser137 made it more aggressive. Pre-treatment with LY249002 inhibits invasion of all clones and in fact the invasiveness is reduced further than serum-free media. PFN-S137A demonstrates significantly reduced invasion even in the presence of PDBu (**p<0.05) (Figure 6B).

Prior studies establish that Profilin 1 promotes MMP2 (Matrix metalloproteinase) secretion, ECM degradation and 3D morphogenesis in vascular endothelial cells [36]. Studies have also been reported showing colocalization of profilin with MT1-MMP (MMP14) at the leading edges of activated monocytes migrating on a layer of activated endothelial cells [37]. MT1-MMP14, in turn, activates MMP2 and MMP9. In the present study, we make evident that phosphorylation of profilin 1 at S137 plays a crucial role in promoting migration and invasion by activating MMP2 and MMP9. Conditioned media from clones growing in serum-free media or 200 nM PDBu containing media were subjected to zymography to analyze MMP2 and MMP9 activity. We found that serum-free clones produce very low MMP2 only, whereas PDBu-treated cells produced significant amounts of MMP9 and to a certain extent MMP2 also. Among the clones, PFN-WT showed maximum MMP2/9 activity. PFN-R74E also expressed increased MMP9 activity comparable to PFN-WT. PFN-H133S by itself was not very migratory and invasive in nature but PDBu-treatment makes this clone sufficiently pro-invasive. Consistent with that data, we found MMP2/9 activity increasing in this clone in response to PDBu. PFN-S137A had the least MMP2 activity and remained same under all conditions tested. MMP9 activity in PFN-S137A was comparable to EV and considerably less than PFN-WT (Figure 7A).

Though profilin 1 has been shown to be localized partially in the nucleus, very few studies are available on its role in the nucleus. Recent studies carried out in dipteran Chironomas tentans have shown that a fraction of profilin in located in the nucleus; is highly concentrated in the nucleoplasm and nuclear periphery. It associates with protein-coding transcriptionally active genes and is independent of actin [38]. Similarly, we observed PFN-WT and its mutants to be localized in the nucleus, especially PFN-R74E, mutant for actin-binding was predominantly nuclear. We also find that increasing amounts of profilin localizes to the nucleus upon phosphorylation by PDBu as early as 5 min and upto 18 hr when compared to serum starved cells (Figure 7B). PFN-WT showed the most nuclear concentrations of profilin 1 when compared to serum starved cells. However, PFN-S137A, mutant for phosphorylation at serine 137, did not show any change in profilin localization with or without PDBu (Figure 7B). These results suggest that phosphorylation of profilin is importing profilin into the nucleus and probably facilitating in the binding of profilin to transcriptionally active genes relative to promoting tumorigenesis. This observation will need more confirmatory studies and identification of nuclear proteins or genes regulated by profilin.