Targeting Sarcoplasmic/Endoplasmic Reticulum Ca²⁺-ATPase 2 by Curcumin Induces ER Stress-Associated Apoptosis for Treating Human Liposarcoma

Liposarcoma is the most common soft tissue sarcoma and accounts for at least 20% of all sarcomas in adults (1). The 2 major locations of liposarcoma are the extremities and the retroperitoneum (2). Surgery serves as the main effective therapy for localized liposarcoma so far and there are few effective treatment options (3). Patients with liposarcoma often suffer from a high risk of relapse locally in cases of incomplete resection. In some cases, the local relapse of liposarcoma may usually accompany dedifferentiation and metastasis. The key chemotherapeutic drugs for soft tissue sarcomas such as doxorubicin and its combinations (4) have been used as the mainstay over decades despite their substantial toxicity. On the other hand, the newly developed molecular-targeted agents have been hardly applied for the treatment of liposarcoma in the present time. A basic research of the targeted drug sorafenib found that liposarcoma cell lines were completely resistant to sorafenib treatment (5).

Curcumin is a polyphenol that is derived from the dietary spice turmeric. The compound shows wide-ranging antiinflammatory and anticancer properties (6). The abilities of curcumin to induce apoptosis of cancer cells (7, 8), as well as to inhibit angiogenesis and cell adhesion (9), have been reported to contribute to its efficacy in the treatment of cancer. The phase I and phase II clinical trials indicate that curcumin is quite safe and exhibits therapeutic efficacy in patients with progressive advanced cancers (10). However, to our knowledge, the underlying mechanism of the anticancer effect of curcumin remains to be investigated.

In this study, to find an effective drug therapy for liposarcoma, we report for the first time that curcumin induced endoplasmic reticulum (ER) stress in human liposarcoma cell line SW872 cells via inhibiting the activity of sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase2 (SERCA2), which then resulted in CHOP-mediated apoptosis. The elevated expression of SERCA2 in SW872 as well as human liposarcoma tissues may be considered as a target of curcumin.

Curcumin (>98% purity) was purchased from Shanghai R&D Centre for Standardization of Traditional Chinese Medicine and the stock solution was prepared with dimethyl sulfoxide. Nutlin-3 was purchased from Sigma and dissolved in dimethyl sulfoxide at the concentration of 20 mmol/L to produce a stock solution. Doxorubincin and cisplatin were purchased from Sigma and dissolved with PBS at 5 mg/mL and 1 mmol/L, respectively. Sorafenib and sunitinib (Pfizer) were dissolved in DMSO to a concentration of 20 mmol/L as stock solution.

Six-week-old female homozygous ICR SCID mice were purchased from the Shanghai Laboratory Animal Center. They were maintained in pathogen-free condition at 22 ± 2°C and kept on a 12-hour light–dark cycle. Animal welfare and experimental procedures were carried out strictly in accordance with the Guide for the Care and Use of Laboratory Animals (The Ministry of Science and Technology of China, 2006) and the related ethical regulations of our university. All efforts were made to minimize the animals' suffering and to reduce the number of animals used.

Human liposarcoma cell line SW872 cells and K562 cells were purchased from the American Type Culture Collection. The SW872 cell line was initiated by A. Leibovitz in 1974 from a surgical specimen of a patient. The histopathology evaluation reported an undifferentiated malignant tumor consistent with liposarcoma. Cells were maintained in L-15 Medium supplemented with 10% fetal bovine serum (FBS, Life Technologies), 100 U/mL penicillin, and 100 μg/mL streptomycin and incubated at 37°C. The K562 cell line was a human-immortalized myelogenous leukemia line that was maintained in Iscove's Modified Dulbecco's Medium supplemented with 10% fetal bovine serum (FBS, Life Technologies), 100 U/mL penicillin, and 100 μg/mL streptomycin and incubated at 37°C in a humidified atmosphere containing 5% CO₂. Fat tissues from donors were washed extensively with ice-cold PBS to remove debris and blood cells. Then the tissues were treated with 0.15% collagenase (type I) containing 0.1% BSA in PBS for 30 minutes at 37°C with gentle agitation. The adipose-derived cells digested from tissues was then resuspended in Low Glucose Dulbecco's Modified Eagle's Medium supplemented with 10% FBS, 100 U/mL penicillin and 100 μg/mL streptomycin, 1 ng/mL human recombined bFGF, and incubated at 37°C in a humidified atmosphere containing 5% CO₂. The authentication of adipose-derived cells was determined by Oil Red staining after differentiated for 8 days (Supplementary Fig. S1).

Full length of human SERCA2b was amplified with the primers as follows: sense: 5′-GCGGCTAGCATGGAGAACGCGCACACC-3′, antisense: 5′-GCGGGTACCTCAAGACCAGAACATATCG-3′, which was subcloned into pcDNA3.1(+) (Promega). After confirming the sequence of SERCA2b, SW872 cells were transfected with pcDNA3.1 and pcDNA3.1-SERCA2b for 48 hours, respectively. Then the cells were passaged to a 100-mm dish and geneticin (G418 sulfate; Sigma Chemical Co.) was added to final concentration of 400 μg/mL. Resistant cells were allowed to grow for 2 weeks.

Chemically synthesized 21-nucleotide sense and antisense RNA oligonucleotides were obtained from Invitrogen. SW872 cells were plated on 6-well plates at 3 × 10⁵ cells per well and transfected with 100 pmol of siRNA duplex per well using Oligofectamine (Invitrogen). CHOP siRNA sequences were as follows: siRNA-1: UUCAUCUGAAGACAGGACCUCUUGC; siRNA-2: UUGAGCCGUUCAUUCUCUUCAGCUA. SERCA2 siRNA sequences were: 5′-CAAAGUUCCUGCUGAUAUA(dTdT)-3′. Luciferase siRNA was used as described before (11).

RNA samples were treated by DNase and subjected to semiquantitative RT-PCR. First-strand cDNAs were generated by reverse transcription using oligo (dT). The cDNAs were amplified by PCR for 28 cycles (94°C for 30 seconds, 58°C for 30 seconds, and 72°C for 30 seconds) using TaqDNA polymerase (Promega Corp.,). The PCR products were electrophoresed on a 2% agarose gel and visualized by ethidium bromide staining. The Gel Imaging and Documentation DigiDoc-It System (version 1.1.23; UVP, Inc.) was used to scan the gels and the intensity of the bands was assessed using Labworks Imaging and Analysis Software (UVP, Inc.). Quantitative PCR was done with the ABI Prism 7000 sequence detection system (Applied Biosystems) using SYBR Green I dye (Biotium, Inc.), and threshold cycle numbers were obtained using ABI Prism 7000 SDS software version 1.0. Conditions for amplification were 1 cycle of 94°C for 5 minutes followed by 35 cycles of 94°C for 30 seconds, 59°C for 35 second, and 72°C for 45 seconds. The primer sequences used in this study were as follows: GAPDH forward, 5′-AACAGCGACACCCACTCCTC-3′; GAPDH reverse, 5′-GGAGGGGAGATTCAGTGTGGT-3′; SERCA1 forward, 5′-GTGATCCGCCAGCTAATG-3′; SERCA1 reverse, 5′-CGAATGTCAGGTCCGTCT-3′; SERCA2a forward, 5′- CTGTCCATGTCACTCCACTTCC-3′, SERCA2a reverse, 5′- AGCGGTTACTCCAGTATTGCAG-3′; SERCA2b forward, 5′-CGCTACCTCATCTCGTCCA-3′; SERCA2b reverse, 5′-TCGGGTATGGGGATTCAA-3′; SERCA2c forward, 5′- CTG- GAACCTGTTCTTAGCTCAG-3′; SERCA2c reverse, 5′- TCTAGAGCAGCAGAGCAGGAGCCTT -3′;SERCA3 forward, 5′-GATGGAGTGAACGACGCA-3′; SERCA3 reverse, 5′-CCAGGTATCGGAAGAAGAG-3′.

The Western blot was done as described before (11). Briefly, the cells were collected and lysed (50 mmol/L Tris, pH 8.0, 150 mmol/L NaCl, 1% NP-40, 0.1% SDS, 5 mmol/L EDTA, 0.1 mmol/L PMSF, 0.15 U/mL aprotinin, 1 μg/mL pepstatin, and 10% glycerol). The proteins were fractionated by SDS-PAGE and electrotransferred to nitrocellulose membranes. Antibodies to PARP, cleaved PARP, caspase-9, caspase-8, caspase-3, Fas, FADD, phosphorylated elf-2a, elf-2a, CHOP (Cell Signaling Technology, Inc.), antibody to α-tubulin, GAPDH, ATF4, and SERCA2 (Santa Cruz), respectively, were used for blotting, and detection was done by enhanced chemiluminescence (Amersham Pharmacia Biotech).

TUNEL assay was done to detect apoptotic cells using the In situ Cell Death Detection kit from Roche Applied Science according to the manufacturer's instructions. Immunostaining of SERCA2 and CHOP was done using Real Envision Detection kit from Gene Tech Company according to the manufacturer's instructions. H&E staining in tumor tissues were done following manufacturer's protocol.

Cultured SW872 cells were washed with and resuspended in ice-cold PBS. Portions of the suspension (3 × 10⁶ cells in 0.1 mL) were injected into the left inguinal area of SCID mice. Four weeks after injection, the mice-bearing tumors were distributed into 2 groups. Curcumin, dissolved in olive oil, was given daily for 40 days by intraperitoneal injection at dose of 100 mg/kg. Tumor volumes were measured every 3 days and calculated by the following formula: 0.5236 × L1 × (L2)², where L1 and L2 are the long and short diameters of the tumor mass, respectively.

Data were expressed as mean ± SEM. Student's t-test was used to evaluate the difference between 2 groups. Kaplan–Meier method was used to evaluate the survival test. P < 0.05 was considered to be significant.

The structure of curcumin is presented in Fig. 1A. Human liposarcoma cell line SW872 cells and adipose-derived cells were exposed to various concentrations of curcumin, chemotherapeutic agents doxorubicin and cisplatin, molecular-targeted drugs sorafenib and sunitinib, and the MDM2 antagonistic agent nutlin-3, respectively, for 48 hours. The results revealed that doxorubicin, cisplatin, and sunitinib showed a dose-dependent inhibition of both liposarcoma cells and adipose-derived cells whereas sorafenib and nutlin-3 did not show any inhibition of both cells at the concentrations used. However, unlike the others, curcumin showed a dose-dependent inhibition of human liposarcoma cells, whereas it did not affect adipose-derived cells (Fig. 1B).

In order to detect the mechanism of the unique inhibition of curcumin toward human liposarcoma cells, the cell extracts of SW872 cells were prepared as described in Materials and Methods for immunoblotting after 24 hours incubation with various concentrations of curcumin. As shown in Fig. 2A, curcumin increased the active form of caspase-8, which then cleaved its substrate bid and caspase-3 in both dose- and time-dependent manners. Consistently, we also detected a dose-dependent increase in both caspase-3 and caspase-8 activities of SW872 cells after curcumin treatment (Fig. 2B). Caspase-8 is a key mediator of apoptotic signals triggered by death receptors such as Fas, TNFR1, and TRAIL-R1/TRAIL-R2. In SW872 cells, however, no change in either Fas or FADD at the protein level was found after curcumin treatment (Fig. 2C). Interestingly, we detected a dramatic increase in DR5 (TRAIL-R2) but not DR4 (TRAIL-R1) at the protein level (Fig. 2C). Consistently, the cell surface expression of DR5 but not Fas could also be induced by curcumin (Fig. 2C, lower panel). The apoptotic changes in morphology in SW872 cells were also observed by curcumin treatment (Fig. 2D).

It has been reported that CHOP acted as a trigger transcript factor of DR5 under the condition of ER stress. After 24-hour exposure to curcumin, a significant increase in the phosphorylation of elf-2α was found. The expression of ATF4 and CHOP in SW872 cells was also elevated in both dose- and time-dependent manners at both protein (Fig. 3A) and mRNA levels (Fig. 3B), indicating that curcumin induced ER stress in SW872 cells. Knocking down of CHOP expression with siRNA reversed curcumin-induced apoptosis as well as the inhibition of cell survival (Fig. 3C and D). ER stress could be induced by unfolded protein response or imbalance of Ca²⁺ transfer. However, we found no obvious changes of GRP78 or PDI expression at the protein level (Supplementary Fig. S2). Interestingly, as a noncompetitive inhibitor of SERCA, thapsigargin also caused the cleavage of PARP, capsase-8, caspase-3, and bid and increased the CHOP expression in SW872 cells as curcumin (Fig. 3E). Like the apoptosis induction mode of curcumin, thapsigargin treatment caused an increase in the expression of DR5, but not Fas, in SW872 cells by flow cytometric analysis (Fig. 3F).

For the similar function mode of both curcumin and thapsigargin, we measured the activity of Ca²⁺-ATPase in SW872 cells after incubation with various concentrations of curcumin for 24 hours. As shown in Fig. 4A, curcumin significantly inhibited the activity of Ca²⁺-ATPase in a dose-dependent manner. There was no detectable SERCA1 and SERCA2a in adipose-derived cells or SW872 cells, compared with cDNA from human normal muscle tissue, and there was also no detectable SERCA3 expression in human adipose-derived cells or SW872 cells compared with cDNA from k562 cells. Interestingly, the expression of SERCA2b was dramatically elevated in SW872 cells, at both mRNA and protein levels (Fig. 4B). To confirm the possibility of SERCA2b as the most likely target of curcumin in SW872 cells, a docking study followed by a 2.5 ns molecular dynamics simulation suggests that curcumin has its 2 carbonyl oxygen atoms chelating the calcium cation, together with D59 and 4 water molecules. These water molecules also nearly bridge the interactions between calcium and some residues, namely D254, R264, and Q56 (Fig. 4C). Furthermore, we found that as a compound with fluorescence, curcumin located on the endoplasmic reticulum in SW872 cells, which was well merged with the reported localization of SERCA2b (Fig. 4D). Moreover, knocking down the expression of SERCA2 blocked the elevation of CHOP and cleaved PARP induced by curcumin (Fig. 4E, top panel). Stable overexpression of SERCA2b in SW872 cells reversed at least a part of the cell growth inhibition (Fig. 4E, middle panel). The CHOP elevation as well as the caspase-8 of the 3 cleavages induced by curcumin could also be reduced by SERCA2b overexpression (Fig. 4E, lower panel).

SW872 cells (3 × 10⁶) were inoculated s.c. into the right flank of SCID mice. Four weeks later when the tumors began to enlarge (about 50 mm³), the mice were randomized into 2 groups for treatment of olive oil control and curcumin (day 0). As shown in Fig. 5A, curcumin-treated tumors grew to 1.583 cm³ as the average size against the 3.697 cm³ in control 40 days after the daily treatment. The protein level of CHOP in the tumor tissues removed was greatly elevated in the curcumin treatment group, and the cleaved forms of PARP, caspase-3 and caspase-8 were also detected (Fig. 5B). Curcumin treatment could significantly inhibit Ca²⁺-ATPase of SW872 cells in vivo (Fig. 5C). Consistent with the data in the Western blot, the immunohistochemical and tunnel staining assay also showed the elevated CHOP expression and apoptosis induction in tumor tissues from the curcumin-treated group against those from olive oil-treated group, respectively (Fig. 5D and E).

Paraffin section of human liposarcoma tissues were collected from 24 patients who underwent surgical resection for liposarcoma between 2008 and 2009. Of the cases, SERCA2 expression was detected in 16 patient samples (66.7%) by immunohistochemical assay with brown cytoplasmic, but not nuclear staining. The positive expression rate in high malignant liposarcoma including pleomophic liposarcoma (Fig. 6A), round cell liposarcoma, (Fig. 6B) and dedifferentiated liposarcoma (Fig. 6C) was significantly higher (83.3%) than that in low malignant liposarcoma samples (50.0%) such as well-differentiated liposarcoma (Fig. 6D). The percentage data are shown in Fig. 6E.

So far, there is no effective therapy for human liposarcoma except surgery, and the molecular events involved in the pathogenesis remain unknown. Although Singer and colleagues have reported that MDM2 could be a potent drug target for human liposarcoma by using gene expression profiling between well-differentiated and dedifferentiated liposarcoma tissues (3), in this study, no effect of MDM2 antagonist nutlin-3 on the SW872 cell growth was observed (Fig. 1B). Comparatively, the main chemotherapeutic drugs doxorubicin and cisplatin, as well as the multiple tyrosine kinases inhibitor sunitinib, showed a dose-dependent inhibition of SW872. However, these drugs also inhibited adipose-derived cells significantly (Fig. 1B), suggesting the predicable toxicity of normal tissue cells. Interestingly, we have for the first time highlighted a strong inhibition of SW872 cells by curcumin without any influence on the normal adipose-derived cells (Fig. 1B).

Curcumin has been reported to show wide-ranging anti-inflammatory (12) and anticancer properties (13). Several phase I and phase II clinical trials indicate that curcumin is quite safe (14) and exhibits therapeutic efficacy in patients with progressive advanced cancers. Curcumin upregulates the proapoptotic proteins of the Bcl-2 family Bax, Bim, Bak, Puma, and Noxa and downregulates the antiapoptotic proteins, Bcl-2 and Bcl-xl, in various cancer cells (15). Moreover, curcumin is able to decrease the number of aberrant crypt foci in an azoxymethane-induced rat colon cancer model through apoptosis induction via the mitochondrial pathway (16). Curcumin is also known for its antioxidant properties and acts as a free radical scavenger by inhibiting lipid peroxidation and oxidative DNA damage (17).

ER participates in the initiation of apoptosis by at least 2 different mechanisms (18). The first is the unfolded protein response, which can be induced by many kinds of proteasome inhibitors (19). Proteins that are unable to fold properly in ER are ubiquitinated and degraded by the 26S proteasome (20), and inhibition of the activity of proteasome will then lead to induction of a terminal unfolded protein response in tumor cells, leading to ER stress (21). The other is Ca²⁺ signaling. Ca²⁺ is taken up from the cytosol by the sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase (SERCA) and released through the inositol-1, 4, 5-trisphosphate receptor/Ca²⁺ channels or ryanodine receptor/Ca²⁺ channels. Disruption of the function of SERCA by thapsigargin will also lead to ER stress (22).

To find which mechanism is involved in the curcumin-induced ER stress, we detected the protein levels of GRP78 and PDI, the crucial markers of unfolded protein response (23), in curcumin-treated SW872 cells. However, there was no obvious change in both proteins (Supplementary Fig. S2). Then, we exposed SW872 cells and adipose-derived cells to various concentrations of thapsigargin, the inhibitor of SERCA (24), for 48 hours. As shown in Supplementary Fig. S3, thapsigargin showed the same specific inhibition effect as curcumin, where thapsigargin inhibited the growth of SW872 cells in a dose-dependent manner without affecting adipose-derived cells. Consistent with the results obtained by Wootton and Michelangeli (25) and Bilmen and colleagues (26), curcumin inhibited the activity of Ca²⁺-ATPase in a dose-dependent manner like thapsigargin in vitro (Fig. 4A). These results revealed that curcumin-induced ER stress in SW872 cells through inhibiting the activity of SERCA but not inducing unfolded protein response. Although there could be a relationship between unfolded protein response and Ca²⁺ signaling, disruption of calcium homeostasis by thapsigargin may induce GRP expression. Curcumin, a well-known compound with multiple targets, may also have some inhibitory effect on the crosstalk between Ca²⁺ signaling and unfolded protein response.

In order to detect whether SERCA could be the most likely target of curcumin in SW872 cells, first, we compared the level of different isoforms of SERCA expression in SW872 cells and adipose-derived cells. As shown in Fig. 4B, there was a dramatic increase in SERCA2b but not the other SERCAs expression in SW872 cells. The docking study supported the interaction between curcumin and SERCA2 (Fig. 4C). On the other hand, we discovered for the first time a specific localization on the endoplasmic reticulum in SW872 cells (Fig. 4D), which is different from the membrane and nuclear localization of curcumin reported before (27). This finding also supported the interaction between curcumin and ER-located SERCA2 in SW872 cells. We then developed SW872 cells with SERCA2b stably overexpression. As expected, the SW872-SERCA2b cells became less sensitive to the curcumin treatment than SW872-pcDNA3.1 cells, both in the inhibition of cell survival and in the induction of apoptosis (Fig. 4E). This could be explained by the fact that overexpressed SERCA2b competed with endogenous SERCA2b when it interacted with curcumin.

SERCA has been reported to reside in the sarcoplasmic reticulum within muscle cells. It acts as a Ca²⁺-ATPase that transfers Ca²⁺ from the cytosol to the lumen of the sarcoplasmic reticulum at the expense of ATP hydrolysis during muscle relaxation (28). As one of the family members, SERCA2 is known as an important regulator of normal calcium homeostasis and signaling (29, 30). Alterations in calcium-dependent signaling are involved in cell proliferation and differentiation, apoptosis, and disruption of calcium homeostasis, which may contribute to cancer development. Chung and colleagues have reported that elevated SERCA2 mRNA was detected in 90% of human colorectal cancer tissues, indicating the possible role of SERCA2 in the development and progression of human colorectal cancer (31). On the other hand, a null mutation in 1 copy of the SERCA2 has been shown to lead to squamous cell tumor in mice, suggesting that the proper amount of SERCA2 is crucial for the destiny of cells (32). Curcumin has been reported to be a compound with multiple targets, such as NF-κb, PI3K/Akt, Src/Stat3, and so on (33), in fact, many of which are crucial for normal cell survival and tissue development. The phase I and phase II clinical trials indicate that curcumin is quite safe and exhibits therapeutic efficacy in patients with progressive advanced cancers. The inhibition effect of curcumin on tumors but not normal tissue is still being studied, and one explanation is higher curcumin uptake in tumor cells than normal cells (27). By means of curcumin, a new role of SERCA2b in the formation and development of human liposarcoma has been found. Further investigation should be focused on the role of SERCA2b as the first target in the therapy of human liposarcoma.

No potential conflicts of interest were disclosed.

We want to sincerely consecrate this paper to Prof. Ting Chen who had been our colleague and had struggled with retroperitoneum liposarcoma for 15 years, and who is also the donor of some of the samples used in this study.

This work was supported by funds from the National Natural Science Foundation of China (No. 30730107 and 90913023), the Science Fund for Creative Research Groups (No. 30821006), and the Provincial Natural Science Foundation of Jiangsu (No. BK2008022).

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