Alisol B, a Novel Inhibitor of the Sarcoplasmic/Endoplasmic Reticulum Ca²⁺ ATPase Pump, Induces Autophagy, Endoplasmic Reticulum Stress, and Apoptosis

Autophagy is an evolutionarily conserved mechanism by which cellular proteins and organelles are eliminated by the lysosomal degradation pathway (1). This is a multistep process that involves the formation, expansion, and fusion of an isolated membrane to form double-membraned vesicles known as autophagosomes, by which cytoplasmic materials are sequestered and subsequently fused with the lysosome for degradation (2). Autophagy is constitutively active at a low basal level in vivo and is important for cellular homeostasis by removal of damaged or superfluous organelles and proteins. Defects in the autophagic pathway is associated with DNA damage, chromosome instability (3), and increased incidence of malignancies (4), whereas an intact autophagic pathway is necessary for longevity (5). Under nutrient deprivation, growth factor deprivation, and hypoxia, autophagy acts as an alternative cellular energy source by degrading intracellular materials and generating free amino and fatty acids to fuel mitochondrial ATP production (2). Uncontrolled upregulation of autophagy might lead to autophagic cell death (type II programmed cell death; ref. 6).

Emerging evidence suggests that modulators of autophagy may have therapeutic potential. For instance, rapamycin, an inducer of mammalian target of rapamycin (mTOR)–dependent autophagy, is effective in treating fruit fly and mouse models of Huntington's disease through increased autophagic clearance of mutant huntingtin fragments (7). Recently, a small-molecule screen revealed additional new chemical entities that attenuate mutant huntingtin-fragment toxicity through an mTOR-independent mechanism (8). Modulators of autophagy may also play a role in cancer therapy. It has been proposed that for certain tumors undergoing protective autophagy in response to chemotherapy, inhibitors of autophagy may sensitize cancer cells to therapeutic agents. Moreover, autophagic inducers may promote autophagic cell death in tumors directly or augment the efficacy of chemotherapeutic agents when used in combination (9). Consistent with the latter hypothesis, several clinically approved or experimental antitumor agents induced autophagy-related cell death (9–12). However, whether autophagy is a prosurvival mechanism in response to the cytotoxicity of these agents or directly responsible for cell death in these cells remains to be characterized.

In this study, we set out to identify novel inducers of autophagy and evaluate their potential use as antitumor therapeutics. Through screening of our natural product collections and extracts from traditional Chinese medicinal herbs, we have identified alisol B as a novel enhancer of autophagy. We herein present evidence that alisol B increased autophagic flux in several cell lines derived from different tumors, leading to cell cycle arrest and cell death. Alisol B induced autophagy through mobilizing intracellular calcium and activating the CaMKK-AMPK-mTOR signaling cascade as a cell protective response toward alisol B–induced cytotoxicity. Furthermore, we show that alisol B induced unfolded protein responses (UPR) and apoptotic cell death. We also present evidence that the molecular target of alisol B is the sarcoplasmic/endoplasmic reticulum (ER) Ca²⁺ ATPase (SERCA) pump. Together, our work provides novel insights into the molecular mechanism of alisol B–induced cell death, which is essential for the further development of this compound into an antitumor agent.

All compounds were purchased from Sigma unless otherwise stated. Thapsigargin, compound C, BAPTA/AM, E64D, pepstatin A, staurosporine, and STO-609 were obtained from Calbiochem. Alisol B and alisol B 23-acetate were purchased from Wako Pure Chemical Industries. Alisol A and alisol A 24-acetate were from Herbstandard, Inc. Antibodies against p70S6 kinase, phospho-p70S6 kinase (Thr389), AMPKα, phospho-AMPKα (Thr172), eIF2α, phospho-eIF2α (Ser51), poly ADP ribose polymerase, beclin1, and PERK were purchased from Cell Signaling Technology, Inc. CHOP, GRP78, and ATF4 antibodies were purchased from Santa Cruz Biotechnology. Anti-LC3 antibody was from Medical & Biological Laboratories Co., Ltd. Anti-p27 antibodies were from DAKO. Anti-phospho-PERK (Thr980) antibodies were from BioLegend. Anti–β-actin antibody was from Sigma. pEGFP-LC3 reporter plasmid was a gift from Prof. Tamotsu Yoshimori (Osaka University, Osaka, Japan). pATF6-Luc was kindly provided by Dr. DY Jin (The University of Hong Kong, Hong Kong, China; ref. 13). SMARTpool small interfering RNAs targeting beclin1 and nontargeting control were obtained from Dharmacon.

All cells were obtained from the American Type Culture Collection unless otherwise specified. Immortalized wild-type and ATG7-deficient mouse embryonic fibroblasts were kindly provided by Prof. Masaaki Komatsu (Juntendo University School of Medicine, Tokyo, Japan). C666-1 cells and PC3 cells were gifts from Prof. KW Lo and Prof. KM Lau, respectively (The Chinese University of Hong Kong, Hong Kong, China). Alisol A, alisol B, alisol A 24-acetate, and alisol B 23-acetate, were dissolved in DMSO before adding to the culture medium.

Cells were fixed with 4% paraformaldehyde, permeabilized with methanol, and nuclei were stained with 4′,6-diamidino-2-phenylindole. To quantify autophagy, the percentage of cells with punctuate GFP-LC3 fluorescence was calculated by counting the number of the cells showing the punctuate pattern of GFP-LC3 in green fluorescent protein (GFP)–positive cells. A minimum of 150 cells from three randomly selected fields were scored per set of conditions per experiment.

Cell viability was measured using MTT as previously described (14).

Cells were fixed overnight with 2.5% glutaraldehyde followed by a buffer wash. Samples were postfixed in 1% OsO₄ and embedded in Araldite 502. Ultrathin sections were doubly stained with uranyl acetate and lead citrate, and analyzed using the Philips CM 100 transmission electron microscope at a voltage of 80 kV.

Cells treated with alisol B were analyzed after the indicated times by multiparametric flow cytometry using the Annexin V and 7-ADD assay (BD Biosciences) according to the manufacturer's instructions; to analyze DNA content, cells were stained with propridium iodide (Sigma-Aldrich). Flow cytometry was carried out using a FACSCalibur flow cytometer (BD Biosciences). Data acquisition and analysis were done with CellQuest (BD Biosciences).

MCF-7 cells were transiently transfected with pATF6-Luc and pCMV-Renilla (Promega), and treated with the test compounds. Reporter assays were done using the Dual-Luciferase kit (Promega). Luciferase and Renilla luciferase activity was measured with a Victor³ Multilabel Plate Reader (Perkin-Elmer).

Total RNA was extracted. First-strand cDNA was synthesized using the High-Capacity RNA-to-cDNA Master Mix (Applied Biosystems). To detect human spliced and unspliced Xbp-1 mRNA, PCR was done using primers 5′-CTGGAACAGCAAGTGGTAGA-3′ and 5′-CTGGGTCCTTCTGGGTAGAC-3′ as described in ref. (15). Spliced (398 bp) and unspliced (424 bp) Xbp-1 fragments were separated by electrophoresis in a 2% agarose gel with DNA bands stained with ethidium bromide and photographed.

Molecular docking was done using the ICM-Pro 3.6-1 program (Molsoft; ref. 16). The protein crystal structure of rabbit SERCA1A with bound thapsigargin (PDB code: 2AGV) was downloaded from the protein data bank as a basis for modeling. The complexes were evaluated with a full-atom ICM ligand binding score (17) that has been previously derived from a multireceptor screening benchmark as a compromise between approximated Gibbs free energy of binding and numerical errors. As a reference, molecular docking of the known SERCA pump inhibitor, thapsigargin, showed a score of −32.

Purified Ca²⁺ ATPase was prepared from female rabbit hind leg muscle (18). ATPase activity was determined using the enzyme-coupled method utilizing pyruvate kinase and lactate dehydrogenase as previously described (19). Porcine brain microsomes were prepared as described in ref. (20). Due to the relatively low level of Ca²⁺ATPase activity in these microsomal membranes, the rate of Ca²⁺-dependent ATP hydrolysis were measured using the more sensitive phosphate liberation assay, described in ref. (21). All SERCA inhibition data were fitted to the allostearic dose versus effect equation using Fig P (Biosoft):

To identify potential small-molecule therapeutics for cancer that targeted the autophagic pathway, we took advantage of the GFP-LC3 reporter system to screen for novel inducers of autophagy (17). We transiently transfected and expressed GFP-LC3 in MCF-7 cells and then incubated them with partially purified extracts from our natural product collection derived from traditional Chinese medicinal herbs. We also incubated the cells with DMSO alone and with tamoxifen as negative and positive controls, respectively. We analyzed the images in individual wells captured by fluorescent microscopy and quantified the levels of autophagy. We found that a partially purified fraction from the rhizome of Alisma orientale consistently upregulated the percentage of cells exhibiting GFP-LC3 puncta (Supplementary Fig. S1). The fraction was subjected to further bioactivity-guided purification using repeated flash column chromatography on silica gel and normal and reverse-phase preparative high performance liquid chromatography. Finally, the compounds in the purified active fraction were deduced by comparing their ¹H and ¹³C nuclear magnetic resonance spectroscopic data with those of known Alisma components (22, 23) and were identified to be alisol A 23-acetate and alisol A 24-acetate (Fig. 1A), compounds that are known to interconvert rapidly (24). Cytotoxicity was also observed by these compounds upon incubation with the cells.

To confirm the observed activity for inducing autophagy and cell death, four structurally related alisol derivatives, including alisol A 24-acetate, alisol B 23-acetate, alisol A, and alisol B (Fig. 1A), were obtained commercially and their cytotoxicities toward a panel of cancer cell lines were evaluated, respectively. Among these, alisol B 23-acetate and alisol B displayed similar cytotoxicities against tumor cells of different origins (Table 1). Alisol A 24-acetate exhibited a lower cytotoxicity, whereas alisol A is comparatively inactive. To evaluate the potency of these compounds for the induction of autophagy, MCF-7 cells were treated with each of the four compounds at their respective IC₅₀ and the formation of GFP-LC3 puncta were evaluated. GFP-LC3 puncta were significantly increased in the presence of all of the compounds except alisol A (Fig. 1B).

Among the three active compounds, alisol B was the most active while having the simplest pharmacophore required for the observed biological activities. Moreover, alisol B is relatively stable and is the most abundant alisol derivative in the alcohol extract of Alisma orientale (25). Therefore, it was selected as a model to characterize the molecular mechanism of action for this class of compounds.

Besides MCF-7 cells, alisol B also induced the formation of GFP-LC3 puncta in a variety of cell lines examined (Supplementary Fig. S2), demonstrating that this compound was able to induce autophagy in a broad spectrum of cell types. The ultrastructures of alisol B–treated MCF-7 cells were analyzed by transmission electron microscopy. Numerous autophagosomes characterized by double membrane structures were detected in cells treated with alisol B (30 μmol/L) and autophagic vacuoles containing degraded organelles were also identified (Fig. 2A). We then measured LC3-II formation in the presence of lysosomal protease inhibitors (26). Whereas alisol B or lysosomal protease inhibitors alone increased the level of LC3-II as expected, alisol B significantly increased the rate of LC3-II formation in the presence of the inhibitors, compared with the use of protease inhibitors alone (Fig. 2B), suggesting that alisol B induces autophagic activity as a result of enhanced autophagosome formation.

Next, the effect of alisol B on cell cycle progression was investigated. Alisol B induced a time-dependent accumulation of cells in the G₁ phase, with a concomitant reduction in the S and G₂-M phases, respectively (Table 2). Alisol B also leads to a gradual accumulation of p27 in several cell lines (Fig. 2C). Taken together, these data suggested that alisol B induces cell cycle arrest at the G₁ phase before inducing cell death.

Nutrient deprivation activates autophagy through an mTOR-dependent pathway, in which the 5′-AMP–activated protein kinase (AMPK) phosphorylates and activates TSC2, leading to the inactivation of mTOR (27). HeLa (data not shown) and MCF-7 cells treated with alisol B showed a time-dependent increase in AMPK phosphorylation at Thr172 (Fig. 3A), accompanied by a concomitant reduction in phosphorylated p70S6K, a downstream target of mTOR. On the other hand, a significant reduction in alisol B–induced GFP-LC3 puncta formation (Fig. 3B) was observed in cells pretreated with compound C, an AMPK inhibitor. Collectively, these data suggested that alisol B induces autophagy through the mTOR pathway in an AMPK-dependent manner.

Autophagy can be induced by Ca²⁺ mobilizing agents through the CaMKKβ-AMPK-mTOR signaling cascade (28). To elucidate the involvement of CaMKK in alisol B–induced autophagy, cells were treated with alisol B in the presence of STO-609, a CaMKK inhibitor (29). STO-609 inhibited the alisol B–induced phosphorylation of AMPK (Fig. 3C) and significantly reduced the percentage of cells exhibiting GFP-LC3 puncta (Supplementary Fig. S3A; Fig. 3D). Consistently, the addition of intracellular Ca²⁺ chelator (BAPTA/AM) to cells abolished the formation of GFP-LC3 puncta (Supplementary Fig. S3B; Fig. 3E). Importantly, BAPTA/AM also significantly inhibited alisol B–induced cell death (Fig. 3F). Taken together, these data suggested that alisol B induces autophagy by increasing the release of free cytosolic calcium ([Ca²⁺]c) from intracellular stores, and implicates the role of [Ca²⁺]c in alisol B–induced cell death.

Several anticancer therapeutics can induce autophagy in cancer cells (9), but it remains controversial as to whether autophagy is the driver of cell death or a prosurvival process in response to drug treatment. To determine the role of autophagy in alisol B–induced cell death, cells were either treated with 3-MA, a known class III phosphatidylinositol-3′-kinase inhibitor, or with RNAi against beclin1 to block autophagy before alisol B treatment. Although 3-MA significantly reduced cell death induced by alisol B (Fig. 4A), the gene knockdown of beclin1 failed to rescue cells from alisol B–induced cell death (Fig. 4B). These apparently contradictory findings prompted us to further investigate the role of autophagy in alisol B–induced cell death using ATG7-deficient mouse embryonic fibroblasts, which is resistant to autophagic induction (30). As shown in Fig. 4C, ATG7^−/− cells were more sensitive to alisol B–induced cell death compared with the wild-type cells, implicating that cells activate autophagy in response to alisol B as an adaptive mechanism for cell survival.

The apparent contradictory results obtained by pharmacologic inhibition and gene silencing/knockout of autophagic pathway could be explained by a recent report that showed that UPR can be blocked by 3-MA (30). Because our data showed that alisol B induces calcium mobilization, which is known to cause ER stress, UPR activation (31), and apoptosis (32), 3-MA might therefore promote cell survival through suppressing alisol B–induced UPR activation. To explore this possibility, the effect of alisol B on known components of the UPR process including the activation of PERK, IRE1, and the activation of transcription factor-6 (ATF6) signaling pathways was determined. Similar to the action of thapsigargin, a known autophagy inducer, ER stress inducer, and Ca²⁺ mobilizer, alisol B induced PERK and eIF2α phosphorylation in cells. This was accompanied by an induction of ATF4, ER molecular chaperone BiP/GRP78, and CHOP expression (Fig. 5A), suggesting that the PERK signaling pathway was activated by alisol B treatment. On the other hand, thapsigargin, but not alisol B, induced the splicing of Xbp-1 mRNA, suggesting that alisol B did not activate the IRE1 pathway (Fig. 5B). Finally, alisol B induced the activity of the ATF6 reporter gene similar to that of thapsigargin, suggesting that the ATF6 pathway was also activated (Fig. 5C). Together, these data suggested that alisol B activates UPR through the PERK and ATF6 signaling pathways. In addition, we confirmed that alisol B–induced cell death is mediated by apoptosis as measured by binding to Annexin V and 7-ADD (Table 3) and the cleavage of poly ADP ribose polymerase (Fig. 5D), respectively.

Because the biological action of alisol B resembled thapsigargin, a potent inhibitor of the SERCA pump, which induces autophagy and UPR through perturbing calcium homeostasis (28), we postulated that alisol B might also function as a SERCA pump inhibitor. We therefore carried out computational virtual ligand docking studies to evaluate the molecular interactions between alisol B and SERCA1A. Comparative analysis of the low-energy ligand conformations found the preferred site for alisol B to be within the transmembrane domain (Fig. 6A), similar to that found for thapsigargin (31), with a strong binding interaction as reflected by the score of −30.7. As a reference, molecular docking of the known SERCA pump inhibitor thapsigargin showed a score of −32. The top-scoring binding pose of alisol B is characterized by the protostane tetracycle residing in a hydrophobic pocket, being in close contact with amino acid residues Phe256, Val263, Ile765, Val769, whereas the OH group of alisol B is pointing toward Lys252 and Glu255 and having a direct hydrogen bond with the carboxylate of Glu255. An overlay of the low-energy pose of alisol B and thapsigarin with the SERCA pump is also shown (Fig. 6B). On the other hand, consistent with the biological activity, alisol A showed unfavorable interactions with the active site (binding score = −16.68; Supplementary Fig. S4).

To ascertain whether the SERCA pump was inhibited by alisol B, we measured its effect using both purified rabbit skeletal muscle sarcoplasmic reticulum (SR) membranes, which express the SERCA1A isoforms and porcine brain microsomes, which express mainly SERCA2B (33). SERCA2B is the most abundantly found SERCA isoforms in nonmuscle cells and therefore is likely to be a major isoform in the cells lines used. It should also be noted that SERCA1A and SERCA2B are highly conserved between mammalian species (i.e., SERCA1A has a 97% sequence identity between human and rabbit, whereas SERCA2B has a 99% sequence identity between human and pig). As shown in Fig. 6C, the SERCA1A pump (from rabbit skeletal muscle SR) was inhibited by alisol B in a dose-dependent manner, which was fitted to an allosteric dose versus effect equation (see Supplementary Table S1 for details; IC₅₀, 27 ± 5 μmol/L), whereas alisol A exhibited a much lower inhibitory effect (IC₅₀, 100 ± 30 μmol/L; full details of the curve fitting parameters are given in Supplementary Table S1). Figure 6D shows the effects of thapsigargin on Ca²⁺ ATPase activity when measured in rabbit skeletal muscle SR membranes and pig brain microsomes. Thapsigargin is an extremely potent SERCA inhibitor, exhibiting IC₅₀ values of 40 ± 7 nmol/L for SR membranes and 11 ± 1 nmol/L for brain microsomes, respectively. Consistent with our previous findings, the inhibition of activity approaches completion in SR membranes but plateaus in brain microsomes such that ∼30% of the measured Ca²⁺ ATPase activity remains unaffected (Supplementary Table S1; refs. 34, 35). We have since shown that this residual activity in brain microsomes is due to the presence of another Ca²⁺ ATPase known as SPCA, which is insensitive to thapsigargin as well as some other SERCA inhibitors (35, 36). Figure 6E shows the Ca²⁺ ATPase inhibition curves for alisol A and B when measured in brain microsomes. Again, both sets of data fitted to the allosteric equation, again showing that in these membranes, ∼30% of the activity was insensitive to the alisols as observed for thapsigargin (Supplementary Table S1). The results showed that alisol B was again a more potent SERCA2B inhibitor compared with alisol A (IC₅₀, 53 ± 6 μmol/L and 140 ± 25 μmol/L, respectively). The fact that alisol B is slightly more potent for SERCA1A compared with SERCA2B is consistent with the findings for thapsigargin, which is also more potent for SERCA1A than SERCA2B (37). The IC₅₀ values of alisol A and B toward both isoforms of the SERCA pump were concordant with the IC₅₀ values for their respective cytotoxicities and calculated binding scores. Collectively, our data strongly suggested that alisol B is a novel SERCA pump inhibitor.

Using an image-based screen, alisol A 23-acetate was purified and identified to be a modest inducer of autophagy and cell death. The alisols are triterpene compounds that belong to the protostane family and are known bioactive components of the rhizomes of Alisma orientale (38). Despite alisol derivatives having been reported to possess a range of biological activities, including the inhibition of nitric oxide synthesis (39), the inhibition of HBV replication (40), the induction of cell death in tumor cells (23, 41), and the reversal of multidrug resistance (42), the underlying mechanisms and the molecular targets of these compounds have remained elusive.

Here, using alisol B as a model drug for this family of compounds, we showed that it induces calcium mobilization from internal stores, leading to the activation of autophagy through the CaMKK-AMPK-mTOR pathway. We further showed that the disturbance of calcium homeostasis by alisol B activates UPR and apoptotic cell death. In contrast to some compounds that induce autophagic cell death (11), our findings suggested that autophagic induction in alisol B–treated cells is a prosurvival mechanism. Our conclusion is supported by the observations that genetic ablation of genes beclin1 and ATG7 in the autophagic pathway rendered cells more susceptible to the cytotoxicity of alisol B. Finally, we provided evidence that alisol B inhibited the activity of the SERCA pump, whereas alisol A is far less active.

Although alisol B 23 acetate has been found to be as effective as alisol B in the induction of autophagy and cell death, it is possible that in the cell-based assays, cell esterases catalyzed the hydrolysis of alisol B 23 acetate to generate alisol B; thus, alisol B is the actual active compound in both sets of experiments. This is corroborated by the docking studies that identified the OH group in alisol B acting as a hydrogen bond donor as one of the interactions within the binding site, which is not possible when the OH group is acetylated as in the alisol B 23 acetate. On the other hand, the higher activity of alisol B over alisol A could only be attributed to the presence of the epoxide in alisol B instead of a vicinal diol in alisol A. The epoxide functional group in alisol B contributes to the overall bioactivity by inducing a different orientation and conformation of the side chain, particularly at C23-C24-C25, and renders the terminus of alisol B more compact to fit into the binding site compared with alisol A. Although the reactivity of the epoxide allows the possibility of nucleophiles to attack and result in covalent bonding with SERCA, the examination of the pocket showed it to be a relatively hydrophobic region and did not reveal any obvious nucleophiles. Whether the binding of alisol B with SERCA is indeed irreversible could be determined experimentally by additional competition experiments in the future.

The SERCA pump has been explored as a potential target in cancer. Inhibition of SERCA triggers cell death in several cancer cells (43, 44). Thapsigargin is the most selective and potent inhibitor of the SERCA pump. It is a sesquiterpene lactone natural product extracted from the umbelliferous plant Thapsia garganica, which inhibits all the SERCA isoenzymes at subnanomolar potencies. Thapsigargin blocks the uptake of calcium into the sarcoplasmic and endoplasmic reticula, leading to calcium depletion in these stores and increase in cytosolic calcium. It is believed that the perturbation of calcium homeostasis by thapsigargin disrupts protein folding and activates all three arms of UPR, respectively (45, 46). Furthermore, thapsigargin-mediated increase in intracellular calcium also plays an important role in the activation of autophagy (28) and calpain-mediated apoptosis (43). Although thapsigargin is toxic toward both normal and tumor cells, it is particularly effective in targeting tumor cells with low proliferative indices such as prostate cancer. Accordingly, a thapsigargin analogue that is only activated by the prostate-specific antigen has shown efficacy in experimental prostate cancer models with no discernable toxicity (47).

Several compounds, including the cyclooxygenase-2 inhibitor celecoxib (48) and curcumin (36), are also known to possess antitumor activity as least partially ascribed to their inhibitory activity toward SERCA. Being a SERCA inhibitor, alisol B elicited a range of cellular responses resembling, but not identical, to that of thapsigargin. Our findings therefore suggest that alisol B can be further exploited as a cancer therapeutic by a prodrug strategy similar to that of thapsigargin. Although alisol B induced a disturbance of calcium homeostasis, autophagy, and ER stress, it activated only the PERK and ATF6 arms of the UPR pathway, while sparing the IRE1 pathway. It has been shown that although the PERK, ATF6, and IRE1 signaling pathways are activated by UPR, persistent activation of the PERK signaling pathway, coupled with the rapid attenuation of IRE1 and ATF6 activities, are required for the execution of apoptosis (49), whereas extended activation of IRE1 signaling promotes cell survival instead (50). Therefore, the unique pharmacologic property of alisol B may confer an advantage with regard to the induction of apoptotic cell death. Although the mechanistic basis behind this observation remains elusive, it is possible that alisol B may also be an inhibitor of the IRE1 signaling cascade. To our knowledge, alisol B is the only SERCA inhibitor known to selectively promote UPR induction. Although a relatively high dosage is required for alisol B to exert its effect, this could probably be improved by modification of the parental compound by rational drug design in the future. Accordingly, we have developed a novel purification procedure for the large-scale extraction of alisol B from Alismatis orientale recently.⁷

No potential conflicts of interest were disclosed.

We thank Janet Tam for the help in transmission electron microscopy, and Lai-To Leung, Alan Y. L. Chan, John C. K. Chu, and Mina M. T. Ip for help in the extraction and purification of the TCM herbal components.

Grant Support: University Grants Committee of the Hong Kong Special Administrative Region Areas of Excellence Scheme grant AoE/P-10/01 and University of Hong Kong Generic Drugs Research Program.

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.