Calpain 2 Regulates Akt-FoxO-p27^Kip1 Protein Signaling Pathway in Mammary Carcinoma

Background: Calpains are intracellular calcium-dependent proteases implicated in cancer.

Results: Calpain 2 knockdown in breast cancer cells correlated with reduced in vitro proliferation, migration, and in vivo tumorigenicity as well as reduced Akt activation, increased nuclear FoxO localization, and increased p27^Kip1 expression.

Conclusion: Calpain 2 promotes proliferation of cancer cells through Akt-FoxO-p27^Kip1 signaling.

Significance: Calpain 2 represents a potential therapeutic target in breast cancer.

Abstract

We investigated the role of the ubiquitously expressed calpain 2 isoform in breast tumor cell growth, migration, signaling, and tumorigenesis. RNAi-mediated knockdown of the capn2 transcript was used to manipulate expression of the catalytic subunit of calpain 2 in the AC2M2 mouse mammary carcinoma cell line. Stable knockdown of capn2 correlated with reduced in vitro proliferation rates, soft agar colony formation efficiency, and migration rates, indicating roles for calpain 2 in mitogenesis, survival, and motogenesis. Biochemical analysis showed increased levels of protein phosphatase 2A and reduced levels of activated Akt in calpain 2-deficient cells, and this correlated with increased levels of the FoxO3a target gene product p27^Kip1, a key regulator of cell proliferation. Calpain 2 deficiency in the AC2M2 cells correlated with enhanced nuclear localization of FoxO3a, consistent with it being in a derepressed state capable of regulating transcriptional targets. Orthotopically engrafted calpain 2 knockdown AC2M2 cells generated tumors with reduced growth rates and enhanced in vivo expression of p27^Kip1. In summary, calpain 2 deficiency correlated with reduced Akt activity, increased protein phosphatase 2A levels, derepression of FoxO3a, and enhanced expression of the p27^Kip1 tumor suppressor. These observations argue that calpain 2 promotes tumor cell growth both in vitro and in vivo through the PI3K-Akt-FoxO-p27^Kip1 signaling cascade. Inhibition of calpain 2 might therefore provide therapeutic benefits in the treatment of cancer.

Introduction

Calpains are a family of calcium-dependent intracellular thiol proteases that are implicated in a wide range of molecular, cellular, and physiological functions (for review, see Ref. 1). Of the 14 known mammalian isoforms, the ubiquitously expressed calpain 1 (μ-calpain) and calpain 2 (m-calpain) have been studied most extensively. These heterodimers consist of distinct large catalytic subunits of 80 kDa encoded by capn1 and capn2, respectively, and a small common regulatory subunit of 28 kDa encoded by capns1/capn4. Targeted disruption of capns1 in mice abolished both calpain 1 and calpain 2 activities and resulted in embryonic lethality (2–4). Capn1 disruption in mice resulted in more subtle phenotypes (5) whereas capn2 knockout was embryonic lethal (6). Thus, although some functional redundancy between calpain 1 and 2 may exist, knock-out phenotypes in mice, as well as observations from isoform specific knockdown experiments in fibroblasts, indicate that calpain 2 plays prominent and essential roles.

Disruption of calcium homeostasis is often associated with cell death. Although calpain activation may sometimes be a consequence rather than a cause of cell death, there is compelling evidence for pro-apoptotic calpain functions. Indeed, capns1 knock-out fibroblasts are resistant to many cytotoxic challenges (7–9). Paradoxically, there is also evidence for pro-survival functions of calpain (9–11). Thus, the ultimate contribution of calpain to survival or death may depend upon context in which the cell is challenged.

Calpain cleaves cytoskeletal proteins and may thereby contribute to the regulation cellular processes including migration and invasion. In human colon cancer cells, the calpain inhibitor ALLN was shown to block FAK cleavage, cell adhesion, and migration mediated by integrin α2β1 signaling (12). These observations implicate calpain in the regulation of cell adhesion and migration through interactions with integrins. In fibroblasts, disruption of calpain reduces cell migration (13), at least in part through inhibition of actin remodeling associated with lamellipodial dynamics at the leading edge (14).

Studies have shown that the calpain system is dysregulated in cancer. Calpain 2 is overexpressed in human colorectal adenocarcinomas (15). The endogenous inhibitor of calpain, calpastatin, is down-regulated in nasopharyngeal carcinoma (16). Calpain inhibitors blocked proliferation of cancer cell lines PC-3, HeLa, Jurkat, and Daudi cells (17). These observations suggest that calpain may be a useful therapeutic target in cancer.

The PI3K-Akt signaling cascade is important in regulating cell survival, and up-regulation of this pathway has been observed in various cancers. PTEN mutation has been found in 60–80% of prostate cancers, and this leads to a constitutive activation of the PI3K-Akt-mTor pathway. Increased Akt activity also correlates with reduced activity of the growth suppressive forkhead transcription factors FoxO1 and FoxO3, which are inhibited by Akt-mediated phosphorylation (18). A recent study demonstrated that depletion of the calpain small subunit correlated with up-regulated FoxO3a activity in mouse embryonic fibroblasts and human mammary carcinoma MCF-7 cells (19). The nuclear translocation of FoxO3a was associated with increased levels of p27^Kip1 and Bim, molecules that regulate cell cycle and cell death, respectively. Protein phosphatase 2A (PP2A)² was also implicated in calpain-mediated FoxO regulation. Calpain was shown to regulate the stability of the B56 α and γ regulatory subunits of PP2A in vitro (19). The association between PP2A and Akt was enhanced in capns1/capn4 knock-out embryonic fibroblasts, suggesting a potential mechanism for reduced Akt activity in calpain-deficient cells. Other studies have suggested that PP2A can regulate FoxO indirectly through dephosphorylation of Akt or directly through dephosphorylation of FoxO (20, 21).

In this study, we examined the role of calpain 2 in regulating cancer cells in vitro and in vivo. RNAi-mediated knockdown of calpain 2 in the AC2M2 mouse mammary carcinoma cell line correlated with reduced growth and migration in vitro and reduced in vivo tumor growth in a mammary fat pad engraftment model. Biochemical analysis suggested that calpain 2 regulates proliferation, survival, migration, and tumorigenesis of breast cancer cells through a PP2A-Akt-FoxO-p27^Kip1 signaling axis.

EXPERIMENTAL PROCEDURES

Cell Culture, Plasmids, and Reagents

The highly metastatic AC2M2 mouse mammary carcinoma cell line (22) was routinely cultured in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 10% fetal bovine serum (FBS; Sigma), 2 mm l-glutamine, and antibiotics/antimycotics (Invitrogen) and maintained at 37 °C with 5% CO₂ in a humidified incubator. The expression of calpain 2 was depleted using RNAi sequences designed using the pSUPER Design Tool from Oligoengine to target the mouse capn2 transcript, as described previously (23). The target sequence in capn2 was 5′-GGATGGCGATTTCTGCATC-3′, and a nonsilencing control sequence that lacked homology to known Mus musculus mRNAs was 5′-TTCTCCGAACGTGTCACGT-3′. These shRNA-encoding sequences were excised using the XhoI and EcoR I sites from the pSuper-retro vectors provided by Anna Huttenlocher (23) and inserted into the pWPXLd vector, provided by Didier Trono, at the SnaB I site. AC2M2 cells were then infected with the control (nonsilencing) shRNA or capn2 shRNA lentiviruses, and selection was performed to obtain clones that showed a significant level in calpain 2 knockdown. FoxO expression plasmids, pECE-HA-FoxO3a WT (wild type) and pECE-HA-FoxO3aTM (triple mutant T32A/S253A/S315A) were purchased from Addgene. FoxO3aWT and FoxO3aTM DsRed fusion protein expression plasmids were generated by subcloning the FoxO3a coding sequences into the pDsRed1N1 vector (Clontech).

Proliferation and Colony Forming Assays

Equal numbers of AC2M2 control and calpain 2 knockdown cells were cultured in 12-well plates. Cells were harvested and counted using a hematocytometer after 48 h. In the colony forming assay, the soft agar was composed of a base agar and a top agar. The bottom layer was made of 0.5% agarose with culture medium, whereas the top layer contained 20,000 cells in 0.25% agarose with culture medium. The assay was carried out using 6-well plates, and the plates were incubated at 37 °C for 2 weeks. Colonies grew in the soft agar were stained with 0.05% crystal violet solution and quantified using a light microscope.

Migration Assays: Wound Healing and Transwell Migration Assays

In wound healing assays, mouse mammary carcinoma AC2M2 cells were seeded on 6-well plates to form a confluent monolayer. Then a scrape wound was made on the cell monolayer using a pipette tip, and cell migration was allowed for the designated time. Pictures were taken at time 0 and 6 h. The cell migration was assessed by subtracting the width of the wound at each time point from that at time 0 to determine the percentage wound closure. In the Transwell migration assay, membrane inserts (8-μm pore size) coated with 10 μg/ml fibronectin were used. Fifty thousand cells were loaded into the top chamber and allowed to migrate through the membranes toward the bottom chamber containing medium with 10% FBS for 8 h at 37 °C. Cells were then fixed with 3% paraformaldehyde for 30 min at room temperature and stained with crystal violet for 1 h. The membranes were removed, mounted on glass slides, and the number of migrated cells was counted using a light microscope with a 20× objective and represented as the average of five random fields.

Engraftment Mouse Tumor Model and Metastasis Studies

AC2M2 cells (7,500 cells in 10 μl) were injected into the mammary fat pad of nu/nu mice (NCr-Foxn1^nu/nu, Taconic). Tumor growth was monitored every 2 or 3 days, starting on day 13 and up to day 27 using ultrasound imaging (VisualSonics^TM Vevo 770®). In metastasis studies, primary tumors were removed from the mice 22 days after tumor cell injection to promote the growth of lung metastatic lesions. On day 35, mice were euthanized, lungs were dissected, and GFP-expressing metastatic nodules were imaged using a Pan-A-See-Ya Panorama imaging system. Images were captured using a Hamamatsu B/W ORCA-ER digital camera with excitation filter at 470 nm/20 nm and an emission filter at 525 nm/20 nm. Mice were housed in the Animal Care Facility, and procedures were carried out according to the guidelines of the Canadian Council on Animal Care, with the approval of the institutional animal care committee.

Western Blotting and Casein Zymography

Cell or tumor lysates were prepared in buffer containing 150 mm NaCl, 10 mm Tris, pH 8, 10% glycerol, 1% Nonidet P-40, 10 μg/ml aprotinin, l0 μg/ml leupeptin, 100 μm PMSF, and 100 μm sodium orthovanadate. SDS-PAGE was carried out, and proteins were then transferred onto PVDF membrane. Membranes were blocked and incubated with primary antibodies against phospho-ERK, ERK, phospho-Akt, Akt, calpain 2, PP2A, cyclin D1, p27^Kip1, and Bim. Quantification of band density was performed using ImageJ software. Casein zymography was carried out to determine calpain activity in cell lysates. Equal amounts of protein (20 μg) were resolved on 8% nondenaturing polyacrylamide gels containing 2 mg/ml casein. The gels were incubated overnight with buffer containing 5 mm Ca²⁺ to activate calpain and then stained with Coomassie Brilliant Blue. Calpain activity was visualized as clear areas in the gels.

Statistics

All statistical analysis was done by using GraphPad Prism 5 software. All error bars represent S.E. p values were calculated by Student's t test and two-way analysis of variance (ANOVA) analysis. Data sets with p ≤ 0.05 were considered statistically significant.

RESULTS

Knockdown of Calpain 2 Expression in Mouse Mammary Carcinoma AC2M2 Cells

Mouse mammary carcinoma AC2M2 cells were transduced with control lentiviruses or lentiviruses encoding shRNA directed against capn2, and clones were selected. The capn2-encoded 80-kDa calpain 2 catalytic subunit was down-regulated by ∼80% in cells transduced with the capn2 shRNA compared with the cells transduced with the control shRNA. The capns1-encoded calpain 1/2 small subunit heterodimer partner was also reduced by ∼70% (Figs. 1A and 3). Zymogram analysis indicated a reduction in calpain 2 activity which correlated with the reduction in calpain 2 protein (Fig. 1B). Calpain 1 activity and expression of the capn1-encoded 80-kDa catalytic subunit were unchanged in capn2 knockdown cells (Fig. 1). Clone 5 was selected for subsequent analysis.

FIGURE 1.

Knockdown of calpain 2 in AC2M2 carcinoma cells: calpain expression and activity in the AC2M2 control (C) and calpain 2 knockdown cell lines (clones 2, 3, and 5). A, Western blot analysis of an equal amount of protein from each cell line was analyzed with SDS-PAGE and blotted with (upper) an antibody that detects the calpain 2 catalytic subunit (capn2), the small regulatory calpain subunit (capn1), and a nonspecific protein (ns) and with (lower) an antibody that detects only the calpain 1 catalytic subunit (capn1). B, zymogram analysis was performed in polyacrylamide gels containing casein as the calpain substrate. The gel was then fixed and stained with Coomassie Blue to visualize bands that represent calpain activity.

FIGURE 3.

Calpain 2 knockdown affects signaling properties in AC2M2 cells. AC2M2 control or calpain 2 knockdown cell lysates containing equal amounts of protein (20 μg) were analyzed by immunoblotting using antibodies directed against phospho-ERK, total ERK, phospho-Akt, total Akt, PP2A, cyclin D1, p27^Kip1, Bim, and calpain 2 (detects both capn2-encoded 80kDa and capns1-encoded 28-kDa subunits).

Depletion of Calpain 2 Correlated with Reduced Proliferation, Colony Formation, and Migration

We first examined the effect of calpain 2 depletion on proliferation, anchorage-independent growth, and migratory properties of AC2M2 cells. Calpain 2-deficient cells displayed significantly reduced proliferation relative to control cells (Fig. 2A, p = 0.017) and attenuated colony forming potential in soft agar (Fig. 2B, p = 0.011). In scrape wound assays, calpain 2-deficient cells showed a significant reduction in migration relative to the control cells (Fig. 2C, p < 0.05). A similar difference was observed in the Transwell migration assay when FBS was used as a chemoattractant (Fig. 2D, p < 0.05). We also attempted to investigate the signaling cascades involved in AC2M2 cell migration. Using PI3K inhibitor LY294002, we observed reduction in cell migration, suggesting that the PI3K-Akt pathway may play a key role in AC2M2 cell migration (data not shown). However, we have not yet ruled out the possibility that this is due to a reduction in cell survival. Thus, calpain 2 deficiency in the AC2M2 breast carcinoma cell line correlated with reduced proliferation, colony formation, and migration.

FIGURE 2.

Calpain 2 regulates proliferation, colony formation, and migration in AC2M2 cells. A, AC2M2 cells (calpain 2 knockdown or control) were plated at 5,000 cells/ml and cultured for 48 h. Cells were then harvested and counted using a hematocytometer. Cell proliferation is represented as the average of replicates of four (p = 0.017). B, colony formation potential was examined by mixing 20,000 cells in 0.25% agar and plating on a 0.5% agar layer with culture medium. Cultures were incubated for 2 weeks at 37 °C. Colonies were then stained with 0.05% crystal violet and quantified using a light microscope. The results shown are the sum of three individual experiments (p = 0.011) and are representative of a total of 15 experiments. C, wound healing assays were performed on confluent monolayers of AC2M2 cells. A scrape wound was induced using a pipette tip, and cells were allowed to migrate into the wound. Pictures were taken at t = 0 and t = 6 h. The differences in distance were calculated and expressed as percent of wound healing (p < 0.05). D, Transwell migration assays were carried out on fibronectin-coated membranes (8 μm pore size). 50,000 cells in serum-free medium were loaded on the top chambers, and 10% FBS-containing medium was placed in the bottom wells as a chemoattractant. Cells were allowed to migrate for 8 h. Cells that migrated to the bottom side of the membrane were fixed and stained with crystal violet. Migrated cells were counted using a light microscope. Cell migration is expressed as the average of cells in five 20× fields from each membrane, with replicates of four in each condition (p = 0.05). Error bars, S.E.

Calpain 2-deficient AC2M2 Cells Display Dysregulated Akt Signaling

We next compared the steady-state signaling properties of calpain 2-deficient and control cells (Fig. 3). Calpain 2 knockdown cells displayed a similar expression level and activation status of ERK compared with the control cells. However, the activation status of Akt was reduced by 80% in calpain 2 knockdown cells. This is in agreement with previous studies by us and others showing a reduction of Akt phosphorylation in calpain small subunit (capns1/capn4)-null embryonic fibroblasts (9, 19). PP2A has also been shown to contribute to the regulation of Akt (20, 21). Interestingly, we observed an 80% increase in PP2A in calpain 2-deficient cells, suggesting that it might contribute to suppression of Akt activity in AC2M2 cells.

FoxO transcriptional regulators are important targets of Akt. FoxO3a transcriptional activity is suppressed by Akt phosphorylation, 14-3-3 binding, and cytoplasmic sequestration (24). Activation of FoxO3a by dephosphorylation, 14-3-3 dissociation and nuclear translocation can lead to increased expression of target genes including p27^Kip1 and Bim (25), which inhibit proliferation and promote apoptosis, respectively. Transcriptional regulation of these and other FoxO target genes provides a mechanistic link for Akt-based regulation of cell survival, proliferation, and migration. When we compared the expression of the FoxO target genes p27^Kip1, Bim, and cyclin D1 in control and calpain 2-deficient AC2M2 cells, calpain 2 knockdown correlated with 2-fold higher levels of p27^Kip1, but only slightly elevated levels of Bim and cyclin D1 (Fig. 3). Collectively, these observations argue that calpain 2 deficiency could lead to dysregulation of the Akt-FoxO-p27^Kip1 signaling axis and thereby contribute to increased p27^Kip1 and reduced proliferation.

To explore further the role of calpain 2 in regulation of FoxO3a, AC2M2 cells were transfected with plasmids encoding DsRed-FoxO3aWT or DsRed-FoxO3aTM, and the subcellular distribution of the DsRed tagged proteins was evaluated. In the control AC2M2 cells, 68% of cells expressing DsRed-FoxO3aWT displayed exclusive cytoplasmic localization, whereas mixed distribution in both cytoplasm and nucleus was observed in 23% of the cells, and exclusive nuclear expression was found in 9% of the cells (Fig. 4A). As expected, the distribution of DsRed-FoxO3aTM, which lacks multiple negative regulatory Akt phosphorylation sites, shifted significantly to the nucleus. Only 10% of control cells showed exclusive cytoplasmic localization, whereas the remaining 90% of DsRed-FoxO3Atm-expressing control cells displayed localization either exclusively in the nucleus (46%) or in both the cytoplasm and the nucleus (44%). These data are consistent with the expectation that blocking Akt-mediated phosphorylation enhances FoxO3a nuclear translocation. In the calpain 2 knockdown AC2M2 cells, only 20% of DsRed-FoxO3aWT expressing cells showed cytoplasmic localization (Fig. 4B), whereas the remaining 80% displayed either exclusively nuclear localization (36%) or distribution in both cytoplasm and the nucleus (44%). Upon transfection with the DsRed-FoxO3aTM construct, only 1% of the calpain 2 knockdown cells displayed an exclusive cytoplasmic localization, whereas 64% showed exclusive nuclear localization and 35% displayed mixed nuclear/cytoplasmic localization. These data indicate that depletion of calpain 2 in AC2M2 cells correlates with a dramatic shift toward nuclear localization of FoxO3a. This shift phenocopies the effect of mutating Akt phosphorylation sites in FoxO3a, which further argues for an important role of calpain in regulating Akt-mediated suppression of FoxO3a functions in breast cancer cells.

FIGURE 4.

Calpain 2 knockdown correlates with enhanced nuclear localization of FoxO3a. Plasmids encoding DsRed-FoxO3aWT or Ds-FoxO3aTM fusion proteins were transfected into control AC2M2 cells (A) or calpain 2 knockdown AC2M2 cells (B). The localization of FoxO3a was assessed 48 h after transfection by fluorescence microscopy. Cells were grouped into three categories: exclusively cytoplasmic, exclusively nuclear, or in both cytoplasm and nucleus (C+N). Results are expressed as percentage of cells in each category. A minimum of 50 cells were analyzed per samples, with triplicates per condition. The experiment was repeated at least three times. Error bars, S.E.

Calpain 2 Knockdown Reduced in Vivo Tumor Growth Rate but Did Not Affect Metastasis

We next performed mouse engraftment experiments to investigate the in vivo role of calpain 2 in breast tumor growth and metastasis. The calpain 2 knockdown AC2M2 cells or the control cells were orthotopically injected into the mammary fat pads of nude mice, and tumor growth was assessed by longitudinal ultrasound measurements. A linear regression analysis of tumor volumes revealed that calpain 2-deficient tumors displayed significantly reduced growth rates relative to control tumors (Fig. 5; p = 4.25 × 10⁻⁵). These data indicate that calpain 2 deficiency correlates with slower in vivo tumor growth rates; which is consistent with the reduced in vitro growth rates observed the with calpain 2 knockdown AC2M2 cells. Immunoblotting analysis of the tumors harvested from these mice revealed 2-fold higher expression of p27^Kip1 in the tumors derived from calpain 2 knockdown cells (Fig. 6; p = 0.023), which was consistent with the in vitro analysis of the calpain 2-deficient cells. Interestingly, although in vitro migration behavior was compromised in calpain 2 knockdown AC2M2 cells, we observed no significant effect on metastasis to the lung in this in vivo tumor model (Table 1).

FIGURE 5.

Calpain 2 knockdown reduces tumor growth rates in a breast tumor engraftment model. Control or calpain 2 knockdown AC2M2 cells (7,500) were orthotopically injected into the mammary fat pad of nu/nu mice (n = 9). Ultrasound imaging was used to measure the sizes of tumors (mm³) on days 13, 15, 17, 20, 23, and 27 after the cell injection. Linear regression analysis was performed on the measured tumor volumes to determine the rate of change (27.33 ± 8.55 and 60.60 ± 15.76 mm³/day for calpain 2 knockdown and control tumors, respectively). A single-factor ANOVA test revealed a highly significant difference in tumor growth rates (p = 4.252 × 10⁻⁵). Error bars, S.E.

FIGURE 6.

Calpain 2 regulates in vivo p27^Kip1 expression in an orthotopic mammary carcinoma model. A, tumors were harvested from engraftment experiments and homogenized, and lysates containing equal amounts of protein were analyzed by immunoblotting using antibodies directed against p27^Kip1, tubulin, and calpain 2 (detects both capn2-encoded 80-kDa and capns1-encoded 28-kDa subunits). B, quantification of band density from A was performed using ImageJ software. The band intensity from p27^Kip1 was normalized with that of tubulin and was expressed as average. Error bars, S.E.

TABLE 1.

Lung metastasis was not affected by calpain 2 disruption

Cells injected	No. of mice engrafted	No. of mice with lung metastasis	Percent with metastasis
			%
Experiment 1
Control	5	2	40
capn2 knockdown	6	3	50
Experiment 2
Control	6	2	33
capn2 knockdown	8	4	50

DISCUSSION

In this study, we show that knockdown of calpain 2 correlates with reduced in vivo tumor growth in an orthotopic engraftment model of mammary tumorigenesis. Biochemical analysis revealed that Akt activation was compromised in calpain 2 knockdown tumor cells, and this correlated with increased expression of the cyclin-dependent kinase inhibitor p27^Kip1, which is a key target of the FoxO3a transcription factor. We further showed that FoxO3a subcellular location was skewed toward nuclear localization in calpain 2-deficient cells. This is consistent with a model where Akt-mediated FoxO3a phosphorylation correlates with nuclear exit of FoxO3a (for review, see Ref. 26). Increased in vivo p27^Kip1 expression was also observed in engrafted tumors generated with calpain 2 knockdown AC2M2 cells. These observations argue that calpain 2 promotes proliferation of mammary carcinoma cells through the Akt-FoxO3a-p27^Kip1 signaling pathway.

Previous studies have demonstrated roles for calpain in cell migration and invasion. Capns1 knock-out mouse embryonic fibroblasts, which are deficient in both calpain 1 and 2, have abnormal organization of the actin cytoskeleton and displayed reduced in vitro migration ability (13). Using an isoform-specific RNAi strategy, calpain 2 but not calpain 1 was found to regulate membrane protrusions and lamellipodia formation at the leading edge of migrating fibroblast cells (23). Others have suggested complex roles for calpain in regulating cell migration and invasion by influencing membrane-cytoskeletal remodeling and focal adhesion dynamics through cleavage of substrates including cortactin (14), Fak (27), talin (28), and paxillin (29). In the DU-145 prostate cancer cell model, an antisense knockdown approach implicated calpain 2 in promoting cell migration and invasion in vitro, and xenograft experiments suggested that calpain inhibition might also reduce the in vivo invasive potential of these prostate cancer cells (30). In our study, calpain 2-deficient AC2M2 cells displayed reduced migration in vitro, suggesting that calpain 2 might indeed promote in vivo metastatic potential. However, when metastasis from the orthotopic tumor site to the lung was evaluated in vivo, we saw no effect of calpain 2 knockdown (Table 1). These observations argue that calpain 2 is not required for metastasis in vivo, at least not in this particular model system. This was a surprising result because calpain 2 deficiency has been correlated extensively with reduced in vitro migration and invasion. However, Huttenlocher and colleagues have implicated calpain 2 in both inhibitory and stimulatory roles in cell migration and invasion (29, 31). We must also consider the possible involvement of calpain 1 or other calpain isoforms in vivo. It was recently shown that tissue-specific knock-out of both calpain 1 and 2 in T cells, through disruption of capns1, did not compromise integrin-mediated adhesion and migration in vitro (32). It will be interesting to see whether these capns1 knock-out T cells display in vivo migration defects. These observations underscore the importance of evaluating calpain isoform functions in vivo in appropriate animal model systems when exploring potential therapeutic applications of calpain inhibition.

In the AC2M2 cells used in this study, we achieved ∼80% knockdown of the capn2-encoded 80-kDa catalytic subunit. Zymography analysis showed a corresponding reduction in calpain 2 activity, but we could not precisely quantitate this. Unfortunately, calpain-specific in vivo activity assays are still lacking. We also failed to observe any compensatory up-regulation of calpain 1 (protein or activity) in the capn2 knockdown cells. In fact, the levels of the capns1-encoded 28-kDa regulatory subunit for calpain 1 and 2 were also reduced in capn2 knockdown cells, which is consistent with the regulatory subunit being unstable in the absence of stoichiometric levels of its partner catalytic subunits.

PI3K promotes the activation of Akt through generation of phosphatidylinositol 3,4,5-trisphosphate, which recruits pleckstrin homology domain-containing PDK1 and Akt to the membrane. PDK1 and mTORC2 then phosphorylate Akt at Thr-308 and Ser-473, respectively, resulting in its activation. Using capns1 knock-out embryonic fibroblasts, we had previously shown that calpain is important for Akt activation in response to a number of cytotoxic challenges (9). We had also observed that the protective effect of calpain against staurosporin-induced cytotoxicity was negated by PI3K inhibition (9). However, this did not tell us whether calpain was acting upstream or downstream of PI3K. Our attempts to determine whether there was less in vivo PI3K activity in calpain 2 knockdown cells using localization of an Akt-pleckstrin homology domain-GFP fusion protein as a surrogate indicator were inconclusive (data not shown). The activation status of PDK1, as assessed by PDK1 phosphorylation at Ser-241, did not appear to be affected by calpain 2 knockdown (data not shown). Thus, although it is clear that calpain contributes to Akt activation, we have not yet determined its specific molecular functions.

PP2A has been shown to inhibit Akt through dephosphorylation (33). Interestingly, PP2A expression was enhanced in calpain 2 knockdown cells compared with the control AC2M2 cells, suggesting that elevated levels of PP2A might play a role in reduced Akt activation. A previous study has demonstrated PP2A associates more tightly with Akt in capns1 knock-out mouse embryonic fibroblasts compared with wild-type cells, and this correlated with increased FoxO3a nuclear localization and expression of p27^Kip1 and Bim (19). Those authors also reported similar effects upon siRNA-based knockdown of capns1 in the estrogen-dependent MCF-7 breast cancer cells.

Inhibition of PI3K-Akt signaling correlates with increased expression of p27^Kip1. Herceptin-induced p27^Kip1 expression has been observed in Her2⁺ breast cancer cell lines SKBR3 and BT474 (34) whereas p27^Kip1 down-regulation has been correlated with herceptin resistance in breast SKBR3 cells (35). Interestingly, a recent report showed that calpain inhibition could resensitize herceptin-resistant variants of these HER2⁺ cells (36). Calpain knock-out in the chondrocyte lineage in mice has also been associated with decreased proliferation correlating with increased p27^Kip1 expression (37). Our observations in this breast tumor model system are consistent with a role for calpain in regulating PI3K-Akt-FoxO-p27^Kip1 signaling.

FoxO expression is dysregulated in several types of tumors including breast cancer (38) and prostate cancer (18). Nuclear exclusion of FoxO3 correlates with active Akt or IKK in primary tumors, and this seems to link with poor prognosis in patients with breast tumors (38). Overexpression of FoxO1 and FoxO3a in a prostate cancer cell line-induced apoptosis and enhanced TRAIL expression (18). Genetic studies from knock-out mice have also shown that disruption of all three FoxO genes (FoxO1, FoxO3a, and FoxO4) correlated with increased tumorigenesis, further indicating a tumor suppressor role of FoxO (39).

A recent study has suggested that PP2A interacts with FoxO3a (40). The 14-3-3 proteins bind to FoxO3a at phosphorylated Thr-32 and Ser-253, facilitating nuclear export of FoxO3a. However, inactivation of PI3K-Akt signaling leads to dephosphorylation of FoxO3a at these sites by PP2A, resulting in 14-3-3 dissociation, nuclear translocation, and transcriptional regulation by FoxO3a. Interestingly, PP2A is a proposed substrate of calpain, further suggesting a link between calpain and FoxO signaling (19). We also observed reduced Akt phosphorylation and elevated PP2A levels in calpain 2-deficient AC2M2 cells, and this correlated with increased nuclear localization of FoxO3a and up-regulation of p27^Kip1.

Another recent study using conditional deletion of capns1 showed a role for calpain in embryonic chondrocyte proliferation (37). Disruption of calpain 1 and 2 in chondrocytes reduced cell cycle progression at G₁/S transition, reduced cyclin D1 transcription, but enhanced the accumulation of cell cycle proteins cyclin D, cyclin E, and p27^Kip1. Our data indicate that knockdown of calpain 2 in breast cancer cells leads to reduction of Akt activation, and this correlates with enhanced p27^Kip1 expression. The phosphorylation status and nuclear/cytoplasmic translocation of p27^Kip1 regulate its function. Nuclear p27^Kip1 interacts with various cyclin-Cdk complexes to inhibit cell cycle progression whereas cytosolic p27^Kip1 facilitates cell motility through inhibition of RhoA activation (41). There appears to be a role of p27^Kip1 in human tumors. Low expression of p27^Kip1 is associated with enhanced cell proliferation and is linked to several types of human tumors including breast cancer (42). Targeted disruption of p27^Kip1 gene results in an increase of body size in mice, enlargement of organs, and development of pituitary tumors (43, 44). Reduced level or mislocalization of p27^Kip1 correlates with poor prognosis in tumor (45). Several types of cancers, including breast, colon, esophagus, and thyroid, show reduction of nuclear p27^Kip1 level (46). The p27^Kip1-null fibroblasts demonstrate a significant decrease in cell migration compared with wild-type fibroblasts (41). This regulation of migration appears to be independent of the cell cycle regulatory function of p27^Kip1. The p27^Kip1-null fibroblasts are found to have increased actin stress fibers, focal adhesions, and activation of RhoA.

Interestingly, calpain has been shown to degrade p27^Kip1 (47, 48). Calpain degrades p27^Kip1 during the mitotic clonal expansion of 3T3-L1 preadipocyte differentiation, and calpain activity is required during the early stage of adipocyte differentiation (48). In a choroidal melanoma tumor-derived cell line OCM-1, p27^Kip1 expression level was found to be degraded by calpains through a MAPK-dependent process (47).

In conclusion, we have shown that calpain 2 expression correlates with cell proliferation and migration in AC2M2 carcinoma cells. It appears that calpain 2 regulates phosphorylation of Akt, which then leads to transcription regulation of FoxO3a and its downstream target, p27^Kip1. Data from our mouse engraftment experiments also demonstrate a role of calpain 2 in tumor growth in vivo. Thus, our results provide further evidence that calpain may be a potential target in the development of novel cancer therapeutics.