Background
The 90-kDa heat shock protein HSP90AA1 is a chaperone protein associated with numerous client proteins that are highly expressed in many cancer cells [
1,
2]. It stabilizes several cancer-related client proteins including PIM1, AKT, and HIF1A, which are crucial for tumor progression [
1]. Thus, HSP90AA1 is an attractive target for cancer therapy [
3‐
5]. Consistently, small molecular inhibitors of HSP90AA1 such as 17-AAG and SNX-2112 show promising results as cancer therapies [
6‐
8]. These compounds bind HSP90AA1 and suppress its chaperone function, leading to degradation of its client proteins.
Copper ion (Cu) is a transition metal that participates in a wide range of cellular processes. As the disruption of copper homeostasis is a pathological feature of cancer cells [
9], copper complexes have been investigated for their potential applications as anti-cancer drugs [
10]. The anti-tumor mechanisms of copper complexes include cleavage of DNA, generation of oxidative stress, and proteasome inhibition [
10]. For instance, the copper complex NSC 689534/Cu exerted its anti-cancer effects by inducing oxidative stress, and counteracting ROS damage by addition of N-acetyl-l-cysteine (NAC) significantly decreased NSC 689534/Cu cytotoxicity. It is not known whether oxidative stress generated by copper complexes can regulate the expression of HSP90AA1 in cancer cells [
11]. However, it has been reported that ROS generated by vitamin C and K3 can induce degradation of HSP90AA1, and therefore contributes to inducing cancer cell death [
12].
Selenium (Se) is an essential trace element that has been widely studied because of its chemopreventive properties [
13]. Selenium compounds can efficiently induce cell death in various cancer cells [
14]. For example, metal complexes of selenosemicarbazones (selenium in the place of sulfur) induced apoptosis through the mitochondria pathway in cancer cells [
15]. Additionally, nickel (II) complexes of selenosemicarbazones efficiently inhibited metastasis and angiogenesis in breast cancer cells [
16].
Herein, we report a novel selenosemicarbazone compound 2-24a that induces cell-cycle arrest in cancer cells. Furthermore, the 2-24a complex with copper (2-24a/Cu) shows significantly increased cytotoxicity compared with 2-24a alone. Detailed analysis showed that 2-24a/Cu induced oxidative stress, accompanied by down-regulation of HSP90AA1 but not HSPA1A. Down-regulation of HSP90AA1 led to degradation of its client proteins, PIM1 and AKT1. These results suggest that 2-24a/Cu could serve as a potential candidate for anticancer therapy.
Material and methods
Synthesis of di-2-pyridyl ketone4,4-dimethyl-3-selenosemicarbazide (2-24a)
Di-2-pyridyl ketone 4, 4-dimethyl-3-thiosemicarbazone (4.6 g, 16 mmol) was dissolved in about 30 ml ethanol, and methyl iodide (CH3I) (2.84 g, 20 mmol) was added. The mixture was heated on reflux for 1 h, then mixed with the ethanolic solution of s-methyl-di-2-pyridyl ketone 4,4-dimethyl-3-thiosemicarbazide. A 100-mL, dry three-necked bottle was packed in an ice bath and 1.97 g (25 mmol) Se and 1.13 g (30 mmol) NaBH4 were added under a nitrogen atmosphere. These mixtures were heated up before adding 10 mL ethanol, and 1.06 g (10 mmol) NaCO3 after 1 h. The ice bath was then removed, and the ethanolic solution of s-methyl-di-2-pyridyl ketone 4, 4-dimethyl-3-thiosemicarbazide was added. The reaction continued for 20 h at room temperature and then 2 mL acetic acid was added. The exhaust was absorbed by an acetic acid lead solution (10%, 400 mL). The reaction solution was filtrated and the solvent was evaporated under reduced pressure. The residue was purified by column chromatography on silica gel using ethyl acetate - petroleum ether (1:2, v/v). The yield was 1.80 g (34%). 1H-NMR (400 MHz, DMSO) δppm: 3.471–3.537 (s,6H) 7.306–7.334 (m,1H) 7.380–7.388 (m,1H) 7.728–7.748 (d,1H) 7.816–7.840 (m,2H) 8.153–8.173 (d, 1H) 8.569–8.579 (d,1H) 8.693–8.705 (d,1H) 15.317 (s,1H) HLPC-MS m/z: 334.3[M + 1] + .IR(cm): 3050w, 2916 m, 1466s, 1430 m, 1325 m, 1305s, 1235s, 1189 m, 1121s, 1052s, 995 m, 892w, 802 s, 739 s, 709 m, 653 m.
Chemicals and antibodies
Chemicals and antibodies were purchased as follows: H2DCF (Beyotime, S0033), N-acetyl-L-cysteine (NAC, Sigma, A7250), propidium iodide (PI, Sigma, P4170), rabbit anti-HSP90AA1 (Bioworld, BS1181), rabbit anti-HSPA1A (Bioworld, BS6446), rabbit anti-PIM1 (Abcam, ab75776), rabbit anti-AKT1 (Bioworld, BS1978), rabbit anti-GAPDH (Bioworld, AP0063), goat anti-mouse HRP-linked antibody (ZSGB-BIO, ZB-2301), goat anti-rabbit HRP-linked antibody (ZSGB-BIO, ZB-2305), copper (II) chloride (CuCl2, Beijing Shiji, China). 2-24a/Cu was freshly prepared by mixing equal molar ratios of 2-24a and CuCl2, and diluted to the appropriate concentrations before treatment.
Cell culture
HeLa, U2os, and other cell lines were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai), and were cultured in Dulbecco’s modified Eagle’s medium (DMEM, HyClone, SH30022.01B) supplemented with 10% fetal bovine serum at 37°C in an atmosphere of 5% CO2. Cell lines were authenticated based on viability, recovery, growth, morphology, cytogenetic analysis, antigen expression, DNA profile, and isoenzymology by the provider.
Cell viability
Cell viability was measured by using the MTT assay (Sigma,M5655). About 3000 cells were plated in each well of a 96-well plate at 37°C in a humidified 5% CO2 incubator for 12 h before they were treated with the compounds. After treatment, 20 μL MTT (5 mg/mL) solution was added to each well, and the cells were incubated for 4 h at 37°C in a humidified 5% CO2 incubator. The medium was removed to stop the reaction and then 150 μL dimethylsulfoxide (DMSO) was added. The plate was measured using a Varioskan Flash reader (Thermofisher, USA) at 490 nm.
Cell-cycle analysis
About 1.4 × 105 cells were plated in each well of a 6-well plate for 18 h before treatment with Cu, 2-24a, or 2-24a/Cu. After treatment, the cells were collected by trypsinization, fixed in 70% ethanol overnight, washed in PBS three times, re-suspended in 500 mL of PBS containing 0.1 mg/mL Rnase, 50 mg/mL propidium iodide (PI), and 0.2% Triton X-100, and incubated in the dark for 30 min at room temperature. Then, the samples were placed in 12 × 75 Falcon tubes and read on a Becton Dickinson FACStarPLUS. The data were analyzed using Modfit software.
Analysis of oxidative stress
The intracellular accumulation of oxidative stress was detected using the probe H2DCF. Ten micrometers of H2DCF was added into the 2-24a-alone, 2-24a/Cu-, or mock-treated cells for 30 min. Cells were then harvested and analyzed by flow cytometry (Becton Dickinson).
RNA-Seq
Briefly, the cells were treated with 5 μM 2-24a/Cu or 0.1% DMSO for 8 h. Then the total RNA was extracted from both the 2-24a/Cu-treated cells and the control cells by using the RNeasy mini Kit (Qiagen, 74104) according to the manufacturer’s instructions. Total RNA was treated with DNase I (Qiagen, 79254) for 15 min at room temperature to remove residual genomic DNA. The purity and concentration of RNA was assayed by Nanodrop. The quality of RNA was further checked by running a sample of fragmented RNA on a RNA Pico 6000 chip in an Agilent 2100 Bioanalyzer. Total RNA (1 μg) was used to isolate mRNAs with poly(A) tails and then these mRNAs were converted to cDNA using the TruSeq DNA Sample Preparation Kit according to the manufacturer’s protocol. After generation of the target cDNA from U2os cells, sequencing adapters were ligated to short fragments after purifying with a QiaQuick PCR extraction kit, which were then used to distinguish different sequencing samples. Fragments with lengths from 200 to 700 bp were then separated by agarose gel electrophoresis and the fragments were subjected to 15 cycles of PCR amplification. Finally, the prepared libraries were sequenced using Illumina HiSeq™ 2000 before they were checked by q-PCR and analysis in the Agilent 2100 Bioanalyzer. The results obtained from each cell line were matched to the human genome (NCBI Build 36). Results were used for further analysis. Unambiguously mapped results were first used to generate gene counts. Feature counts were normalized using the RPKM (read per kilobases per million aligned reads) method. The RPKM method is able to eliminate the influence of different gene lengths and sequencing discrepancies on the calculation of gene expression. Therefore, the calculated gene expression can be directly used for comparing the difference of gene expression among samples. To detect different expression levels among difference stages, the P-value (one-tailed) corresponds to the differential gene expression test (two-sample t test with equal variances). Because differentially expressed gene analysis generates large multiplicity problems in which thousands of hypotheses (i.e., whether a particular gene is differentially expressed between the two groups) are tested simultaneously, corrections for false-positive (type I errors) and false-negative (type II) errors are performed using a false discovery rate method.
Western blot analysis
To analyze protein expression, western blotting was performed as described previously [
17].
Murine sarcoma S180 implanted mice study
Chinese Kun Ming (KM) mice (male and female in equal numbers) of 16–18 g were purchased from the Vital River Laboratories (China) and housed at the laboratory animal center of Peking University (AAALACi-accredited facility). Experiments were undertaken in accordance with the National Institute of Health Guide for Care and Use of Laboratory Animals, with the approval of the Peking University Laboratory Animals Center, Beijing. Murine sarcoma S180 cells were injected subcutaneously into the right oxter region of KM mice (1 × 107 in 200 μL) until the mice adapted to the new environment. After injection, tumors were allowed to develop for 2 days. We then randomly divided the 40 mice into four groups, treated with DMSO in 0.9% saline (control), 1 mg/kg 2-24a, 1 mg/kg CuCl2 or with 1 mg/kg of 2-24a/Cu. The mice in the four groups were intraperitoneally injected daily according to their weight. Tumor size was measured using calipers; tumor volume was estimated according to the following formula: tumor volume (mm3) = L × W2/2, where L is the length and W is the width. Tumor-bearing mice were sacrificed after 10 days. Xenograft tumors were harvested, weighed and then fixed in 4% formalin for histologic study.
Statistical analysis
Each experiment was repeated at least three times for calculation of standard deviations. The statistical significance of differences was assessed using the Student’s t test in GraphPad prism 5. A P < 0.05 was considered statistically significant.
Discussion
Compounds from the thiosemicarbazones family have shown promising anti-tumor activity
in vitro and
in vivo[
24]. Selenosemicarbazones, in which the sulfur atom is substituted by the selenium atom, also showed potential anti-tumor activity [
15,
16,
25‐
28]. Herein, we report that a novel selenosemicarbazone copper complex (2-24a/Cu) can efficiently inhibit cancer cell proliferation and induce cancer cell death, and can serve as the basis for designing other novel anti-cancer selenosemicarbazone compounds.
As expected, a selenosemicarbazone complexed with copper (2-24a/Cu) is significantly more cytotoxic than 2-24a alone in cancer cells because, as reported by other groups, copper can enhance anti-tumor effect of thiosemicarbazones by inducing oxidative stress [
20]. Consistently, 2-24a/Cu efficiently generated oxidative stress in cancer cells, and the stress could be efficiently inhibited by NAC. Dp44mT [
20], the thiosemicarbazone analog of 2-24a, forms a redox-active copper complex that is responsible for its anti-cancer activity [
20]. As copper is elevated in various cancer cells, it is reasonable that 2-24a could be more easily complexed with copper in certain cancer cells than normal cells, leading to cancer cell death. Further research is needed to investigate whether 2-24a could selectively induce cancer cell death via copper-mediated oxidative stress or directly inhibit tumor growth in
vivo.
We investigated the change of the transcriptome in the 2-24a/Cu-treated U2os cells. RNA-seq results showed that genes that participate in oxidative stress were significantly up-regulated. Additionally, transcription of the genes in the HIF1 signaling pathway were also significantly up-regulated, suggesting the HIF1 signaling pathway could play an important role in regulating copper-mediated cancer cell death.
Metal-thiolate cluster genes and stress-response genes were also up-regulated in the 2-24a/Cu-treated cells, which may play an antagonistic role in the induced cell death. Interestingly, although 2-24a/Cu induced a significant increase in the transcription of Hsp90 family genes (similar to the copper complex of thiosemicarbazone NSC 689534/Cu), HSP90AA1 protein was decreased in the 2-24a/Cu-treated cells. HSP90AA1 is critical for cancer cell metabolism and signal transduction pathways, and inhibition of HSP90AA1 is a promising strategy for cancer therapy [
3,
29]. As oxidative stress has been shown to induce HSP90AA1 cleavage in cancer cells [
12], it is possible that oxidative stress induced by the copper complex resulted in the degradation of HSP90AA1 in the cancer cells, whereas the transcriptional increase of HSP90AA1 served to compensate for the decrease of HSP90AA1 protein. By decreasing the abundance of HSP90AA1 in cancer cells, 2-24a/Cu could decrease the stability of HSP90AA1 client proteins, many of which are critical in tumor initiation and metastasis. Consistent with this hypothesis, PIM1 (a client protein of HSP90AA1 that affects sarcoma growth and bone invasion [
21,
22,
30,
31]) is rapidly decreased in the 2-24a/Cu-treated cells. Similarly, AKT1, which affects cell-cycle arrest and apoptosis [
22], is concomitantly decreased in the 2-24a/Cu-treated cells. Because AKT is also important for supporting angiogenesis signaling [
32] and 2-24a/Cu has been shown to down-regulate AKT, the latter could prove useful in inhibiting angiogenesis, but the possibility remains to be tested. Expression of other key proteins stabilized by HSP90AA1 such as Bcr-Abl, HER2/Neu (ErbB2), and mutated p53 protein [
1], could also be down-regulated by 2-24a/Cu-induced decrease of HSP90AA1, and these collectively contribute to the anti-cancer property of 2-24a/Cu.
Competing interests
The authors declare no competing financial interests.
Authors’ contributions
WZ, BG, and JL conceived and designed the experiments. HS, HZ, MS, YT, YH, HZ, and RC performed the experiments. HS analyzed the data. GB and WZ wrote the paper. All authors read and approved the final manuscript.