Beyond Platinums: Gold Complexes as Anticancer Agents

Present and Future Perspectives

Based on the encouraging results obtained for the first-generation compounds, we focused our attention on improving their bioavailability and their cellular uptake, exploiting the possibility of functionalizing the dithiocarbamato moiety with small oligopeptides to be recognized by specific transporters. Since cancerous cells are known to overexpress specific biomarkers and receptors needed for carcinogenesis and tumor growth, targeted chemotherapies actually aim to block cancer cells proliferation exploiting such specific up-regulated biomolecules. This led to the development of efficient and innovative delivery systems in which conjugated drugs incorporating a tumor-targeting group can selectively reach the tissue of interest and deliver the cytotoxic agent directly into the tumor cells (13). In particular, specific peptide transporters (PEPTs) were selected as potential targets for our compounds. They are integral membrane proteins expressed in mammals in two different isoforms, PEPT1 and PEPT2, which are known to promote cellular uptake of potentially all physiologically-occurring di- and tripeptides with different sequence, charge and hydrophobicity (14). They are present predominantly in epithelial cells of small intestine, bile duct, mammary glands, lung, choroid plexus and kidney, but are also localized in other tissues (such as pancreas, liver, gastrointestinal tract) and are up-regulated in different types of cancer (15). Owing to their specificity in recognition and transport, they represent an excellent target for the delivery of some pharmacologically active peptidomimetics [e.g. β-lactam antibiotics and angiotensin-converting-enzyme (ACE) inhibitors], and might also have a central role in the uptake of our gold(III) compounds (16). Indeed, we recently designed and reported on the synthesis, characterization and biological evaluation of new gold(III)-dithiocarbamato complexes functionalized with short oligopeptides, of the type [Au^IIIX₂(dct-Sar-AA)] (X=Br, Cl; Sar=sarcosine, N-methylglycine; AA=different amino acids). Among the library of tested compounds, three complexes (Figure 1) with greater in vitro antiproliferative activity were selected, namely AuD6 [Au^IIIBr₂(dct-Sar-Gly-OtBu)], AuD8 [Au^IIIBr₂(dct-Sar-Aib-OtBu)] and AuD9 [Au^IIICl₂(dct-Sar-Gly-OtBu)]. These complexes were tested on a large panel of cancer cell lines representatives of different cancer types, recording IC₅₀ in the low micromolar range, with values up to about four-times lower than the reference drug cisplatin, showing no cross-resistance with cisplatin (17, 18).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Chemical structure of cis-diaminodichloroplatinum(II) (cisplatin) and three of our gold(III) peptidomimetics (32): dibromido[1-(1,1-dimethylethyl)N-dithiocarboxy-κS,κS’-N-methylglycylglycinato]gold-(III) (AuD6), dibromido[1-(1,1-dimethylethyl)N-dithiocarboxy-κS,κS’)-N-methylglycyl-2-methylalaninato]gold(III) (AuD8) and dichlorido[1-(1,1-dimethylethyl)N-dithiocarboxy-κS,κS’)-N-methylglycylglycinato]gold-(III) (AuD9).

After this preliminary screening was carried out on several human tumor cell lines, we decided to perform a deeper biochemical evaluation in order to better-understand their biological activity. Thus, we have taken into account two specific types of adenocarcinoma, namely human triple-negative breast cancer (TNBC) and two androgen-resistant types of cancer, breast and prostate cancer these being among the most frequent causes of tumor death in women and men, respectively (19).

In particular, for further in-depth studies, we exploited the breast cancer MDA-MB-231 cell line and PC3 and DU145 prostate cancer cell lines. The choice of these tumor subtypes was not accidental. In fact, the MDA-MB-231 cells belong to the so-called TNBC subtype as they do not express the genes for estrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor 2 (HER2/neu), thus preventing the use of traditional hormonal or monoclonal antibody therapies (20, 21). Furthermore, these cells are highly metastatic, invasive and resistant to cisplatin, and lead patients to have frequent and early gastrointestinal metastases, with adverse prognosis.

Regarding the PC3 and DU145 prostate cancer cell lines, it is worth noting that neither is androgen-responsive, thus ruling out the benefits of androgen ablation in the early stages of the disease. Therefore, these types of androgen-independent tumors require alternative therapeutic strategies (22).

For all the cell lines, we performed a number of biological tests aimed at elucidating the mechanism of action and identifying the major targets of our gold(III) compounds (23).

As far as the epithelial breast adenocarcinoma is concerned, the complexes AuD6 and AuD8 (Figure 1) were tested. First, the 3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide (MTT) assay was used to assess their in vitro cytotoxicity after 24 or 72 h treatment. Both complexes inhibited tumor cell growth in a dose-dependent manner, AuD8 being about three times more potent than AuD6,with IC₅₀ values in the low micromolar range. It should be noted that MDA-MB-231 cells were resistant to cisplatin under the same experimental conditions. In fact, after 24 h treatment, the IC₅₀ value was not reached using concentrations ranging from 25 to 100 μM, in agreement with literature data. After 72-h treatment with the reference drug at 100 μM, about 23% cell viability was recorded, in agreement with the awareness that its antitumor activity occurs slowly (23).

In the past 10 years, after FDA approval of bortezomib (N-acyl-dipeptidyl boronic acid; Velcade®) in 2003 for the treatment of relapsed multiple myeloma (MM) and mantle cell lymphoma, several families of organic and inorganic compounds were tested as proteasome inhibitors (5, 24). The 26S proteasome is an ATP-dependent multicatalytic protease, playing a role of paramount importance in cytoplasmic and nuclear homeostasis via hydrolysis of a large collection of proteins previously labeled with poly-ubiquitin chains. The ubiquitin-proteasome system (UPS) is indeed involved in several intracellular processes including the removal of abnormal, misfolded or improperly assembled proteins, cell cycle progression, apoptosis, drug resistance, cell differentiation, angiogenesis and stress response (25). From the structural point of view, 26S proteasome is made up of a central barrel-shaped complex termed the 20S proteasome (consisting of 28 subunits), capped at each end by a regulatory component referred to as 19S complex. 19S units recognize ubiquitinated client proteins and regulate their unfolding and translocation into the 20S proteasome proteolytic chamber. Eukaryotic 20S core contains only three proteolytically active subunits (double present and placed within the complex according to a C2 symmetry), namely β1, β2 and β5, which recognize distinguishing side chains of different peptides, thus resulting in specific activity for each catalytic site (peptidyl glutamyl peptide hydrolyzing-like, trypsin-like and chymotrypsin-like, respectively).

Interestingly, our compounds were shown to trigger cell death via a different mechanism of action if compared to clinically-established platinum drugs as they exhibited proteasome-inhibitory activity both in vitro and in vivo (23). Tumor cell proteasome targeting is critical as the homeostasis of the UPS is unbalanced and some regulatory proteins either circulate longer due to decreased degradation (e.g. anti-apoptotic proteins) or lost due to accelerated degradation (e.g. pro-apoptotic proteins).

In order to elucidate to what extent cellular proteasome is inhibited by our compounds, and to provide insights into the mechanism of action and cell death type, we carried out dose-dependent and time-dependent studies. MDA-MB-231 cells were treated with either i) AuD6 or AuD8 for 24 h at different concentrations (5, 10, 15 and 20 μM), or ii) each complex at 20 μM for 4, 8, 16 or 24 h. Overall, upon western blot analysis of the cell extracts, we observed accumulation of ubiquitinated proteins, the proteasomal target proteins IκBα (nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, alpha) and p27 (pro-apoptotic), thus highlighting proteasome inhibition in vitro. Unlike AuD6, treatment with AuD8 also led to caspase-3 activation, poly (ADP-ribose) polymerase (PARP) cleavage and increased levels of the pro-apoptotic form of Bcl-2-associated X protein (BAX) (p36) with consequent decrease or complete disappearance of p21/BAX and p18/BAX fragments. AuD6 was less potent than AuD8, in agreement with its higher IC₅₀ value. Indeed, we observed bands as intense as the control for some proteins, while weaker effects on the apoptotic proteins PARP and caspase-3 were detected after treatment for longer times, or at higher concentrations (23).

Based on these results obtained in vitro, we addressed our attention on cell-free systems, investigating the effects of both compounds on the proteasomal chymotrypsin (CT)-like, trypsin (T)-like and peptidyl-glutamyl-peptide-hydrolizing (PGPH)-like activities of MDA-MB-231 cell extracts. Both complexes were able to inhibit all the three enzymatic activities in a concentration-dependent manner (23). Thus, we incubated purified human 20S proteasome with each compound at different concentrations in order to provide direct evidence for this inhibition activity. AuD8 seemed to be slightly more selective toward the CT-like activity of the purified proteasome core. In this context, it has been reported that chymothrypsin-like pocket targeting is associated with apoptosis induction in cancerous cells (26). In terms of proteasome inhibition, both complexes proved to be about an order of magnitude more potent than the first generation compounds (10, 27).

The previously described western blot analysis highlighted that our gold agents seem to induce cell death via apoptosis. To confirm this type of cell death commitment, we carried out an Annexin V/Propidium Iodide assay. It should be noted that the amount of cells undergoing non-apoptotic cell death was comparable to the solvent control. After 16h treatment with AuD8 at 20 μM, cell death occurred mainly by apoptosis (62.9%, late stage) and this population of cells further increased after 24 h (74.1%). Both compounds caused similar percentages of cells in early-stage apoptosis (about 7-9%), with AuD6 being less potent than AuD8, triggering apoptosis at the late stage in only 49% of cells after 24 h treatment (23).

Subsequently, we investigated the antiblastic activity of both complexes in vivo exploiting female athymic nude mice bearing human breast cancer MDA-MB-231 cells. Unexpectedly, our gold complexes showed similar antiproliferative activity on this xenograft breast cancer model. In fact, contrary to the results collected in vitro, AuD6 turned out to be as active as AuD8 both in the short-(13 days) and in the long-term (27 days) treatment at 1.0 mg kg⁻¹ d⁻¹, thus highlighting the importance of animal models when considering the various factors involved in the tumor tissue physiology. In this study, we observed 53% inhibition of xenograft growth compared to control after 27 days, while after a short-term treatment some mice displayed 85% inhibition and, in some cases, tumor shrinkage. Remarkably, mice looked healthy and very active during the treatment showing neither signs of fatigue nor weight loss, thus ruling-out the onset of systemic toxicity (23).

Tumors were collected and weighed, followed by western blot analysis after two weeks and after 27 days of treatment. Our compounds showed similar levels of p27 expression and PARP cleavage in both the short- and the long-term treatments. Conversely, AuD8 was slightly more potent than AuD6 toward accumulation of IκBα and Bax fragmentation. To validate the apoptosis-inducing and proteasome-inhibitory activities in vivo, we performed two immunohistochemical stainings after 27 days of treatment, showing enhanced p27 levels in both AuD6- and AuD8-treated xenografts, thus indicating proteasome inhibition, and increased TUNEL positivity, representing apoptotic cell death in vivo (23).

Regarding the PC3 and DU145 androgen-resistant prostate adenocarcinomas (28), the cytotoxic effect of two among the most active gold(III) peptidodithiocarbamato derivatives, AuD8 and AuD9 (Figure 1), was evaluated in vitro after 72 h treatment considering cisplatin as the reference drug. Additionally, cell growth recovery experiments were performed to check whether tumor cells are able to proliferate again upon gold compound removal. In particular, following 24-h incubation with increasing concentrations of both derivatives, the medium was replaced with drug-free one and cells were incubated for further 72 h, recording IC₅₀ values comparable to those observed after 72 h treatment (in the low-micromolar range). These results pointed out that the cytotoxic activity is not reversed, thus ruling-out a cytostatic effect of our compounds. Notably, both complexes showed higher anticancer activity than cisplatin, with IC₅₀ values about 2- to 4-fold lower than the reference drug against these androgen receptor-negative prostate cancer cells.

Afterwards, the biological investigation was focused on PC3 cells as they are more invasive and more resistant to chemotherapy than DU145 cells (28).

Based on an annexin-V/PI (propidium iodide) staining, our complexes caused cell death by inducing apoptosis after 24 h incubation at 10 μM (for both compounds about 60% early apoptosis and about 20% late apoptosis were detected), with cisplatin inducing no death effect under the same experimental conditions. Since BCL-XL, BCL-2 and BAX proteins have a critical role in regulating the intrinsic apoptotic pathway, we analyzed whether our complexes are able to affect their expression. Thus, we treated PC3 cells for 24 h with either AuD8/AuD9 or cisplatin (10 μM). Both gold complexes significantly reduced the expression of the pro-survival BCL-XL protein while increasing the levels of the pro-apoptotic BAX protein. However, only AuD8 treatment led to down-regulation of the expression of the anti-apoptotic BCL-2 protein. Under these experimental conditions, cisplatin reduced BCL-XL expression but had no effect on BAX and BCL-2 counterparts (28).

Moreover, cell-cycle progression and DNA fragmentation were evaluated by PI staining and flow cytometric analysis. After 12 h of incubation, both complexes caused an increase of the S phase population, with comparable losses (with respect to the control) from both the G₂/M and G₀/G₁ phases. Interestingly, an S phase population comparable to that of the control (upon percentage reduction with respect to 12 h incubation) and a block in G₀/G₁ step, due to a loss of the G₂/M phase population with respect to the medium, were detected after 24 h. Such a situation (viz. G₀/G₁ block) remained steady, with comparable percentages in the different phases, after 48 and 72 h incubation. DNA fragmentation was significant (about 20%) only after 48 or 72 h treatment. Conversely, after 72 h of incubation, an equivalent concentration of cisplatin caused a blockade in the G₂/M phase of the cell cycle, with no significant DNA fragmentation, thus suggesting a different mechanism of action for our compounds compared to Pt-based drugs (28).

It has been reported that the epidermal growth factor receptor (EGFR) is overexpressed in androgen-refractory and metastatic prostate cancer. Furthermore, EGFR phosphorylation affects the migration and proliferation of prostate cancerous cells (29). Contrary to cisplatin, treatment of PC3 cells with our compounds (at 10 μM for 24 h) resulted in a remarkable decrease of the surface expression of EGFR and its phosphorylated form (pEGFR) as evaluated by flow cytometry. The effect of both complexes on PC3 cells migration was then studied by means of a 12- and 24-h cell scratch assay (30). Treatment with each gold complex at 4 μM for 24 h reduced PC3 cell migration rate measured as the area of the scratch covered by the migrating cells after 12 and 24 h by imaging software, resulting in failure to form a confluent monolayer (28).

Remarkably, similarly to the previously described breast cancer cells and xenografts (10, 23, 27), our gold(III) derivatives also significantly inhibited proteasome activity in PC3 cells (analysis of cytosolic extracts after treating cells at 10 μM for 12 h at 37°C). Under the same experimental conditions, cisplatin was much less active than both compounds (28).

Similarly to breast cancer described above, we investigated the anti-neoplastic activity of AuD8 against prostate tumor xenografts in nude mice. Following s.c. injection of 2 mg/kg AuD8 every other day for 19 days of treatment, we observed a reduction of 70% in tumor volume with neither decreased activity nor weight loss, contrary to mice treated with cisplatin. Furthermore, no organ damage was histologically-detected in the spleen, liver and kidney of animals except for the observation of focal lung lymphocyte infiltration (28).

Recently, we have been investigating acute toxicity in 80 mice by two administration routes (endovenous and oral) taking into account different doses, much higher than those used in chemotherapy (xenografts treatment at only 1-2 mg/kg). So far, we have recorded outstanding results (unpublished data). To sum up, no deaths, nor signs of toxicity have been observed during the study in any of the treated animals. There were no clear clinical findings in any of the survivors. It is worth highlighting that body weight for any mouse was constant or increased during the 14-day observation period. On necropsy, no alterations or gross anatomical findings in the main organs and tissues were observed.

In conclusion, the selected gold(III)-dithiocarbamato complexes were shown to be able to inhibit the proliferation of each studied cancer cell line, both in vitro and in vivo. We have herein shed some light on the mechanisms of action and the biological targets of our gold(III) compounds. In spite of initial skepticism of the scientific community, proteasome inhibition has been emerging as a winning strategy to treat malignancies and we have here reported that our compounds proved to be proteasome inhibitors, both in vitro and in vivo, in two distinct human tumor cell lines.

Overall, the anticancer activity and toxicological data described highlight that the design of our potential drugs is very successful both in stabilizing the transition metal center and in avoiding the non-specific reactivity of each complex towards healthy tissues (4, 6, 31).

At present, we are completing a transcriptomic study on MDA-MB-231 and MCF7 breast cancer cells upon treatment with two compounds in order to determine differentially expressed genes and pathways.

All the obtained results are very encouraging and account for the interest in ongoing advanced pre-clinical studies on these compounds, evaluating the possibility for them to approach phase I clinical trials (32).