In this research, propolis from A. mellifera was used to determine the in vitro antiproliferative/cytotoxic activity on five human cancer cell lines. Although there are many bee species that can produce propolis, especially stingless bees, such as Melipona fasciculate:  and Tetragonula carbonaria , A. mellifera was chosen since it is commonly cultured for honey, is an easy to manage species in apiaries and so makes access to propolis on a commercial, as well as environmentally sustainable, scale feasible. In addition, the bioactivities of propolis are reported to depend on the geographical regions , seasons  and other external factors. Thus, the propolis of A. mellifera from Thailand, a floral biodiversity hotspot, is of interest since it has never been reported previously yet maybe different from the propolis of this species reported previously from other regions. The selection of Nan province was based upon the diverse flora still present in this region of Thailand, and so the potential for novel compounds in the propolis. This native and remote area of the country is dry, mountainous and full of deep forests with unique plants, such as Bretschneidera sinensis Hemsl.
Propolis was initially sequentially extracted with MeOH (high-polar solvent), then CH2Cl2 (medium-polar solvent) and finally hexane (non-polar solvent). Both the hexane (CHE) and CH2Cl2 (CDE) extracts revealed a good antiproliferative/cytotoxic activity against the five selected human cancer cell lines, as determined by the MTT assay. Thus, in general the antiproliferative/cytotoxic compounds in this propolis from A. mellifera in Nan, Thailand, are unlikely to be highly polar. This notion is supported by Castro et al.  who reported the best antiproliferative activity against HeLa tumor cells was from prenylated benzophenone (hyperibone A), which is found in the CHE of Brazilian propolis, with an IC50 value of 175.6 nM (91 ng/ml).
Both MeOH and water/EtOH, two polar solvents, could be used to extract the antioxidant activity from propolis from Portugal , whilst other optimal extraction solvents were reported to be chloroform for the antimicrobial activity against oral pathogens  and ethanol for the anti-influenza A virus activity . Thus, the bioactivities of crude propolis extracts, and so the frequently, albeit incorrectly, inferred propolis bioactivities, depend also on the extraction solvents used as well.
The different cell line sensitivities and IC50 values for the antiproliferative/cytotoxic activity before and after fractionation by adsorption chromatography could represent the removal of inhibitory components that exert an antagonistic effect, or the separation of different components with different activities. Comparing the IC50 values of compounds 1 and 2 (Table 5 and Figure 3), compound 2 (cardol) looked to be a promising agent for anti-cancer treatment in terms of its lower IC50 values for antiproliferation/cytotoxicity compared to compound 1 (cardanol), assuming that (i) the same IC50 values observed against the non-transformed Hs27 cell line reflects an antiproliferative activity only and not a cytotoxic activity and that (ii) a specific delivery system could be used to target the cancer cells or tumor area rather than systemic delivery, so as to avoid or minimize side affects. Moreover, consumption of the crude form of propolis should be warned against because Aliboni et al.  reported that propolis can cause an allergic reaction to sensitive individuals due to the presence of the two allergenic esters, benzyl salicylate and benzyl cinnamate.
Both compounds 1 and 2 (cardanol and cardol) are phenolic lipids with an amphiphilic character  derived from the hydrophilic hydroxyl group and the hydrophobic long chain hydrocarbon . These compounds are found in tropical plants in the family Anacardiaceae, both in native and cultivated cultures . Economic plants in this family include cashew nut, mango and ginkgo , whilst the diversity of both compounds is high, such as in the form of anacardic acid, catechol, resorcinol and gingkolic acid . Indeed, members of these groups have previously been reported to exhibit diverse bioactivities, such as antibacterial , antiplasmodial , antioxidant  and antifungal activities . However, the diversity of chemical structures in the cardanol and cardol groups may account for the diverse bioactivities , rather than a few pluripotent compounds.
Wang et al.  reported that they could purify CAPE from propolis, and that it showed an antiproliferative activity on the human colorectal cancer cell line (CRC) in a dose- and time-dependent manner. The IC50 value of CAPE after 72 h treatment was 22.7 μM (6.47 μg/ml). Comparing compound 2 (cardol) from our research with that for CAPE, the antiproliferative/cytotoxic activity IC50 value of compound 2 on the SW620 cell line (< 3.13 μg/ml; < 6.8 μM), which is also a human colorectal cancer cell line, was over 3.3-fold lower than the IC50 value of CAPE on CRC (in terms of molarity). Thus, subject to the risk of side effects, compound 2 (cardol) purified from Thai A. mellifera propolis could be a better antiproliferative agent against human colorectal cancer cells.
CAPE is also reported to have an effect on breast cancer cells, with a similar IC50 value on the ER- and ER+ MDA-231 and MCF-7 cell lines, respectively, of 15 μM (4.26 μg/ml) . Thus, the IC50 value reported for CAPE is broadly similar in terms of mass, but some 1.5-fold higher in terms of molarity, to that seen here for compound 2 (cardol) against the breast cancer cell line BT474 (4.41 μg/ml; 9.61 μM), again indicating that cardol purified from Thai A. mellifera propolis could be an interesting antiproliferative agent against human breast cancer cells.
CAPE has been reported to display a broad target range inhibiting the growth of many cancer cell lines, such as C6 glioma cells  and human leukemia (HL-60) cells , and also to be cytotoxic to the neck metastasis of gingiva carcinoma (GNM) and tongue squamous cell carcinoma (TSCCa) cells . Moreover, CAPE showed a strong inhibitory effect on the matrix metalloproteinase (MMP-9), which is related to the invasion and metastasis ability of hepatocellular carcinomas . In the future, the effect of compounds 1 (cardanol) and 2 (cardol) from this Thai A. mellifera propolis should be evaluated accordingly.
Since many cancer drugs or chemotherapy agents used nowadays cause adverse side effects through being cytotoxic to normal cells, it is necessary to find new compounds that will not cause such adverse side effects and not be cytotoxic to normal cells. Therefore, the apparent absence of cytotoxicity of compounds 1 (cardanol) and 2 (cardol) to the non-transformed Hs27 cell line in vitro is of interest, but requires conformation in a broader range of non-transformed cell lines. However, against that was the observed antiproliferative affect noted on the Hs27 cell line, which may well then result in strong adverse side affects and so the requirement for more localized drug delivery systems. This is because although compounds 1 (cardanol) and 2 (cardol) affected some cancer cell lines in vitro with lower IC50 values than that against the non-transformed Hs27 cell line, this small difference is unlikely to be sufficient to allow safe systemic administration without side affects, but may be sufficient when targeted local delivery is performed [50, 51].
Propolis and its phenolic compounds have been reported to induce the death of cancer cells either by necrosis  or by apoptosis, the latter of which might be by mitochondria mediated-  or death signal mediated-  apoptosis. Thus, the in vitro effects of compounds 1 and 2 upon the cell morphology and DNA fragmentation of the cell lines was observed.
A change in the cell morphology with a decrease in the cell number was observed for SW620 cells when cultured in vitro with compounds 1 (cardanol) or 2 (cardol), which is consistent with a cytotoxic effect. In contrast, no change in the cell morphology was observed with the Hs27 cells under the same conditions. It is likely that compounds 1 (cardanol) and 2 (cardol) affected the SW620 cancer cells by necrosis, not by apoptosis, whereas they induced an antiproliferation response and not cell death in the Hs27 cells. In contrast, Vatansever et al.  reported that CEE from Turkey induced the death of the human breast cancer cell line (MCF-7) by the induction of apoptosis. Although the morphology of the MCF-7 cells was not visibly changed, the number of cells was decreased. In addition, whilst Umthong et al.  found that CWE and CME from Trigona laeviceps (stingless bee) in Samut Songkram province, Thailand, had a similar effect upon SW620 cells as that reported here (change in the cell morphology, loss of cell adhesion and cell death), in contrast, they found evidence of DNA fragmentation, unlike in this study with compounds 1 (cardanol) or 2 (cardol). Moreover, Chen et al.  reported that propolins A and B extracted from Taiwanese propolis could induce apoptosis of human melanoma A2058 cells, in addition to inducing the morphological changes in the cells, chromatin condensation and cell shrinkage. However, since we did not screen the crude extracts for changes in the cell morphology and DNA damage, but only the two purified compounds that were not propolin A or B, then it is unclear if this represents the diversity of bioactivity within different propolis components or between propolis samples.
Cancer can be caused by the misregulation of, and so its treatment can be targeted at inhibition of, phosphatidylinositol-specific phospholipase Cγ1 (PI-PLCγ1), since it plays a key role in the proliferation and progression of human cancer . Thus, an inhibitor of PI-PLCγ1 would be a useful tool for development of anticancer agents. Lee et al.  reported the isolation of a cardanol from the chloroform extract of Ginko biloba that exhibited inhibitory effects against PI-PLCγ1 in a concentration-dependent manner. They also found that the structure of the cardanol could influence the inhibitory effect. Cardanol with unsaturated long carbon chains (cardanol C15:1 and cardanol C17:1) showed more potent activities than those with saturated long chains (cardanol C13:0 and cardanol C15:0). Other than the inhibition on PI-PLCγ1, cardanol is reported to be cytotoxic in vitro to human cancer cell lines, such as HCT-15 (colon), MCF-7 (breast), A-549 (lung), HT-1197 (bladder) and SKOV-3 (ovary), but was not found to be cytotoxic to the normal colon cell line, CCD-18-Co.
In addition, Kubo et al.  reported that the cardol (C15:0) isolated from Anacardium occidentale was moderately cytotoxic to the murine B16-F10 melanoma cells in a dose-dependent manner with an IC50 value of 24 μM (8.352 μg/ml) and complete lethality at 40 μM (13.92 μg/ml), which in terms of molarity is some two- to 3.5- fold higher than that observed here for compound 2 (cardol) from the Thai A. mellifera propolis (albeit subject to the caveat of on different cell lines). Since cardol is an amphipathic molecule, the cytotoxicity is potentially facilitated by its ability to act as a surfactant.
The two potentially new compounds isolated here from Thai A. mellifera propolis (a cardanol and a cardol) could be alternative antiproliferative agents for future development as anti-cancer drugs.