On the mechanism of tumor cell entry of aloe‐emodin, a natural compound endowed with anticancer activity

Abstract Aloe‐emodin (1,8‐dihydroxy‐3‐[hydroxymethyl]‐anthraquinone), AE, is one of the active constituents of a number of plant species used in traditional medicine. We have previously identified, for the first time, AE as a new antitumor agent and shown that its selective in vitro and in vivo killing of neuroblastoma cells was promoted by a cell‐specific drug uptake process. However, the molecular mechanism underlying the cell entry of AE has remained elusive as yet. In this report, we show that AE enters tumor cells via two of the five somatostatin receptors: SSTR2 and SSTR5. This observation was suggested by gene silencing, receptor competition, imaging and molecular modeling experiments. Furthermore, SSTR2 was expressed in all surgical neuroblastoma specimens we analyzed by immunohistochemistry. The above findings have strong implications for the clinical adoption of this natural anthraquinone molecule as an antitumor agent.


| INTRODUCTION
In the era of combinatorial chemistry and high-throughput screening of large compound libraries development of natural products has been neglected. However, despite steady progress in the above fields, cancer treatment modalities are unsatisfactory for many solid tumors and there remains an unmet need for new drug discovery. 1 Aloe-emodin (AE) is a natural anthraquinone (1,8-dihydroxy-3-[hydroxymethyl]-anthraquinone) produced by different species of well-known plants, such as Aloe and Rheum, 2 as well as the annelid worms of the Tomopteris genus. 3 AE shows a number of antimicrobial, metabolic, diuretic and immunosuppressive properties including, lastly, anticancer activities. [4][5][6] We have originally described the selective in vitro and in vivo killing of neuroblastoma cells by AE, without appreciable signs of acute toxicity. 7 The anticancer activity, in particular, was deemed to result from apoptotic cell death, cellular differentiation and antiangiogenic mechanism. [7][8][9] In addition, a number of studies showed that AE affected ERK 1/2 pathway in glioma and fibrosarcoma cells 10,11 ; it was already reported that different SSTR subtypes 1, 2, 5 and are able to affect cell proliferation of C6 glioma cells in vitro through the activation of the same intracellular pathway (inhibition of ERK1/2 phosphorylation). 12 We have also demonstrated that the anticancer activity of AE was most likely promoted by a tumor-cell-specific drug uptake process. 13 However, the mechanism for the selective uptake of AE in large cytoplasmic vesicles by tumor cells remains elusive.
Like AE, somatostatin (SST) and derivatives thereof affect a broad range of biological pathways resulting in metabolic, cathartic, cell proliferation, cell survival effects and, notably, antineuroectodermic tumor activity. [14][15][16][17] In particular, two biologically active forms have been identified in mammals, the cyclic peptide somatostatin-14  and the N terminally extended somatostatin-28 (SST-28). The biological effects of somatostatin are mediated through a family of G protein-coupled receptors (GPCR) which comprises five distinct subtypes (SSTR 1-5). 14 SSTRS are highly expressed in various cultured tumor cells and primary tumor tissues of neuroectodermal and nonneuroectodermal origin, including neuroendocrine tumors (NET). Generally, each tumor variably expresses more than one somatostatin receptor subtype. [18][19][20][21] In recent years, considerable interest has been placed on receptor-targeted cancer therapy since certain receptors are found aberrantly expressed in cancer cells at higher concentrations than in normal cells. 20,21 Our study, that focuses on the mechanism of AE-uptake, unveils how this process is mediated by two of the five receptors of somatostatin: SSTR2 and SSRT5.

| Drugs
Aloe-Emodin was provided by Angelini S.p.A. and it was dissolved in DMSO stock solution of 200 mM. Somatostatin14 was purchased from Sigma-Aldrich (Sigma-Aldrich, Milan, Italy). All drugs were stored at À20 C. (RRID:CVCL_1174) were from the National Cancer Institute (NCI, New York, NY). MRC5, HeLa and LoVo cells were grown in DMEM medium, DMS114 and IMR32 were grown in RPMI-1640 medium (Lifetechnologies Gibco, Milan, Italy). Culture medium was supplemented with 10% heat-inactivated fetal bovine serum (Lifetechnologies Gibco, Milan, Italy), 100 units/mL penicillin and 100 μg/mL streptomycin (Lifetechnologies Gibco, Milan, Italy). All cell lines were grown at 37 C with 5% CO 2 humidified atmosphere. All experiments were performed with mycoplasma-free cells.

| Cell culture
All human cell lines have been authenticated using STR profiling within the last 3 years.  PCR reactions comprised 40 cycles after an initial denaturation step (95 C, 10 minutes) according to the parameters: denaturation at 95 C, 15 seconds and annealing/extension at 60 C, 1 minute, in an ABI PRISM 7700 Sequence Detection system. To normalize the RNA amount of the extracted samples, real-time PCR analysis of the human GAPDH cDNA was carried out. Moreover, we optimized the method for relative quantification with sequence-specific DNA probes for sstr2 and sstr5 genes TaqMan reagents using kit TaqMan Gene Expression Assays according to the manufacturer's instructions (Applied Biosystem, Milan, Italy). The relative quantification of a target template in the samples was evaluated using the 2 ÀΔΔCT method. All experiments were conducted in triplicate (details on primers are reported in Appendix S1).

| Molecular modeling
Since no crystallographic structure of the somatostatin receptor remains available, two homology models were constructed, using as template the recently published crystal structure of the k-opioid receptor (KOP) (PDB ID: 6B73), given the high sequence similarity with both SSTRs subtypes considered in our study. The quality of models was evaluated using the SWISS-MODEL workspace (details are reported on Appendix S1).
Three-dimensional structures of the ligands under investigation were built and correctly prepared taking advantage of the MOE suite, following the protocol described in Appendix S1.
GOLD docking tool was selected as a conformational search program and PLP as a scoring function. 22 In total 20 docking runs were performed for each somatostatin receptor, searching in a sphere of 15 Å radius. Along with AE, the compound under investigation, docking simulations were conducted also for ligands L-779976 and L-817818, the references nonpeptidic potent and selective agonists respectively of SSTR2 and SSTR5. 2.7 | Cell viability modulation by sstr-2 and sstr-5 antibodies and SST14 IMR32 cells were seeded as described above. After 24 hours of incubation the medium was discarded and replaced by 100 μL of medium supplemented with AE (50 μM), SST14 (10 μM), 1:500 SSTR2 Ab, SSTR5 Ab (Santa Cruz, CA), or SST14 with AE (50 μM), or SSTR2 Ab and SSTR5 Ab together with AE (50 μM). Then the cells were evaluated for viability, as mentioned before.  Cells were incubated at 37 C in a CO 2 incubator until assayed for gene knockdown. Gene silencing was stable for 72 hours. After 48 hours cells were treated with AE (50 μM). Twenty-four hours after AE treatment cells were harvested and prepared for quantitative proliferation assay, as previously described. Nontargeting siRNA transfected cells were used as a negative control.

| Statistical analyses
Statistical analyses were performed with Student's unpaired t-test with *P < .05 and **P < .01.

| RESULTS AND DISCUSSION
Following our original report 7 on AE activity against neuroectodermal tumor cells, a number of publications have shown AE activity also on tumor cells of different origins, 4-11 a result that seems to contradict, at least in part, our previous report of AE selectivity. In order to understand these apparently conflicting observations, we approached the search for an AE cellular specific receptor.
We have already demonstrated that AE is incorporated at high concentrations in sensitive cells. In neuroblastoma cells AE was incorporated with an intracellular concentration of 570-fold over the cellular background. The fluorometric signal in MRC5 and HeLa cells did not allow to appreciate a relative increase in drug uptake. 13 In order to evaluate a reduction of AE uptake in the sensitive cells we assayed different ligand and/or antibodies of a number of neuroectodermal tumor cell receptors, such as nerve growth factor (NGF), acetylcholine, somatostatin (SST14), clonidine, adrenaline and noradrenaline.
We evaluated the relative amount of AE uptake by different analyses: two-photon excitation microscopy (TPE) and flow cytometry (data not shown). This screening showed that sst14, sstr2 and 5 antibodies significantly reduced cellular AE uptake. Hence the role of these receptors seems quite relevant in the above process although we cannot exclude the role played by other receptors such opioid and G-coupled receptors that are abundantly present in IMR32 cells. 24 Considering the similarity of the pleiotropic effects of both somatostatin peptides and AE, we investigated if and how somatostatin receptors were expressed on representatives of AE-sensitive and AE-no sensitive cell lines.

| RNA expression and western-blot analyses
The sstr5 gene expression, but not sstr2 gene expression, was significantly up-regulated in IMR32 cells by exposure to AE. At variance, sstr2 and sstr5 genes were both up-regulated by somatostatin-14, but not at a significant level ( Figure 1B). The sstr5 gene was not expressed in MRC5 cells but sstr2 gene was significantly up-regulated both by AE and SST14 ( Figure 1B).
Western blot analyses confirmed the above findings ( Figure 1C).
It is reasonable to conceive that SSTR5 must play an epistatic role over SSTR2 in IMR32 cells. The latter receptor, in turn, being the sole receptor expressed in MRC5 cells, is the only one undergoing positive feedback by AE. In order to find a plausible role for these receptors in AE-sensitive cells, molecular docking simulations were exploited using the two SSTR subtypes that were significantly expressed and up-regulated. Since AE is a molecule considerably smaller than the endogenous agonist SST or the reference nonpeptide compounds, the explorable space by the ligand within the orthosteric binding site is large enough to increase the variability of docking predicted poses.
The most reasonable binding modes of AE, respectively, within SSTR2 ( Figure 2A) and SSTR5 binding sites ( Figure 2B) are shown. AE interacts mainly with residues of the transmembrane segments TM3, TM5, TM6 and with the extracellular loop EL2, a portion of the binding site very similar to the one described for the reference agonist compound (more details are reported on Appendix S1).

| Cell viability modulation
In order to obtain a functional validation of the gene expression experiments and molecular dynamic prediction, cell viability modulation was studied using the natural ligand SST14 as well as antibodies F I G U R E 1 Panel A -Cell viability assay. Cells were treated with different concentrations of AE (1-10-100 μM) for 72 hours. The viability was compared to the untreated cells. The plate absorbance (ABS) was measured in a spectrophotometer, at 620 nm wavelength. Cell survival was expressed as a percentage calculated as follow [(ABS AE treated cells À ABS Background)/(ABS Vehicle treated cells -ABS Background)] Â 100 with MTT analysis. Panel B -Sstr2 and sstr5 mRNA gene expression. IMR32 and MRC5 cells were exposed to AE (50 μM) and SST14 (10 μM) for different time points (5, 30, 60 and 120 minutes). Sstr5 gene was not expressed in MRC5 cells. Panel C -Western-blot analyses: SSTR5 and SSTR2 protein expression after AE treatment (50 μM) in AE-sensitive cells (IMR32) and in AE nonsensitive cells (MRC5). Densitometric analysis of data from representative of three experiments was presented as fold increase relative control against SSTR2 and SSTR5. As explained in Figure 2C, SST14 was less cytocidal for IM32 cells than AE alone or in combination with AE. It is, therefore, conceivable that both molecules competed for the same receptors affecting AE-sensitivity in a nonsynergistic nonadditive mode. A significant increasing cell survival was observed in AE and SSTR2 and/or SSTR5 antibodies co-treated cells indicating the relevant role of these receptors for AE activity. It was previously described that AE strongly inhibited ERK1/2 signaling pathway affecting proliferative potential of C6 astrocytoma and fibrosarcoma cells. 10,11 In the experimental model of C6 glioma cells, cytostatic F I G U R E 2 Panel A -Molecular docking predicted binding modes for Aloe-emodin in complex with the SSTR2 and the SSTR5 homology model (Panel B); residues involved in molecular recognition are labeled while hydrogen bond between ligands and the receptor are depicted by the blue lines. Panel C -IMR32 cells viability modulation. After 24 hours of treatment with AE, SST14 or cotreatment with AE and SST14 or Abs the cells were evaluated for viability. Panel D -Confocal microscopy analyses. IMR32 and MRC5 cells were pretreated for 20 minutes with SSTR2, SSTR5 Abs or SST14 and relative fluorescence amount in the cells was evaluated in 3 minutes from AE administration, as mentioned in section "Materials and Methods." Panel E -SiRNA knockdown. IMR32 cells were transfected with siRNA for somatostatin receptor 5. Western blot analysis for SSTR5 protein expression shown that gene silencing occurred 48 hours after siRNA transfection and was stable for 72 hours. IMR32 transfected cells were treated with AE (50 μM) for 24 hours and evaluated for the proliferation inhibition [Color figure can be viewed at wileyonlinelibrary.com] effects of somatostatin were mediated by the SSTR2, and SSTR5, 12 the likely carriers of AE. One could therefore assume that the ERK 1/2 pathway could also be targeted by AE in our cell line.

| Confocal microscopy analyses
Being an autofluorescent molecule (λ ecc at 410 nm and λ em at 610 nm) AE accumulation was investigated in sensitive and nonsensitive cells after treatment with antibodies and the somatostatin natural ligand, SST14. Fluorescence was significantly downregulated in IMR32 cells after incubation with SSTR5 Ab; only a minor difference in AE accumulation was observed after SST14 or SSTR2 Ab pretreatment ( Figure 2D). Besides, neuroblastoma cells exhibited no significant variation in survival after 24 hours of coexposure to AE and SST14 ( Figure 2C). These findings would prompt us to suggest that AE, although strongly competing for the same receptor as SST14 is, with respect to this ligand, the primary cytotoxic agent having a diversified mechanism of action that also includes DNA binding, as previously shown by us. 13 AE accumulation fluorescence was not evident in MRC5 cells treated or not treated with SSTR2 and SSTR5 antibodies or SST14 ( Figure 2D).

| Gene knockdown
IMR32 cells, as a representative cell line highly expressing SSTR5, were also transfected with sstr5 siRNA. After gene silencing cells were significantly less sensitive to AE growth inhibition ( Figure 2E).
Since we proposed neuroblastoma neoplasia as possible targets for AE treatment, standing the relatively high sensitivity of neuroblastoma cell lines to the drug, 7,13 the in vivo function of SSTR 18,21,25 was the most intuitive one to explore. To this end we evaluated SSTR2 and SSTR5 immunoreactivity in neuroblastoma's surgical specimens.
While SSTR5 immunoreactivity was barely detectable in one out of the nine tumor specimens evaluated, a moderate to strong immunoreactivity for SSTR2 was detectable in all neuroblastoma sections examined ( Figure 3).

| DISCUSSION
The original contribution of this article is the demonstration that somatostatin receptors 2 and 5 are likely responsible for AE accumulation and cytotoxicity.
Our results of gene amplification, western blotting, and imaging techniques, indicate that SSTRs were expressed in a different way in neuroectodermal cells and in AE-sensitive and nonsensitive cells and that AE has the biological and molecular characteristic to be recognized by SSTR2 and SSTR5. In particular, these receptors are induced after AE treatment. Moreover, a different pattern of SSTR2 and SSTR5 expression is shown in AE-sensitive and nonsensitive cells, that correlates with the differential cytotoxic potential of AE, receptor knockdown and competition experiment. These data are corroborated by our preliminary molecular modeling studies showing how AE can be recognized by SSTR2 and SSTR5 receptors establishing, in the orthosteric binding sites, a network of stabilizing interactions ( Figure 2 and Appendix S1). In conclusion, we considered these findings as the molecular and biological explanation of the AE selectivity

CONFLICT OF INTEREST
The authors declare no conflicts of interest.

DATA AVAILABILITY STATEMENT
The data that support the findings of our study are available from the corresponding author upon reasonable request.