Abstract
Background/Aim: Artemisinin and its derivatives are not only approved antimalarial drugs but also exert strong anticancer activity. Based on the clinical activity of artesunate (ART) that has been previously reported in cervix carcinoma, we investigated a panel of 12 different biomarkers and identified the Wilms Tumor 1 (WT1) protein as a potential target of ART. Patients and Methods: Matched biopsies of cervical carcinoma before, during, and after therapy from patients treated with ART were investigated for induction of apoptosis (TUNEL assay) and expression of Wilms Tumor protein 1 (WT1), 14-3-3 ζ, cluster of differentiation markers (CD4, CD8, CD56), ATP-binding cassette transporter B5 (ABCB5), glutathione S-transferase P1 (GSTP1), inducible nitric oxide synthase (iNOS), translationally controlled tumor protein (TCTP), eukaryotic elongation factor 3 (eIF3), and ADP/ATP translocase by immunohistochemistry. WT1 has been selected for more detailed analyses using molecular docking in silico, microscale thermophoresis using recombinant WT1, and cytotoxicity testing (resazurin assay) using HEK293 cells transfected with four different WT1 splice variants. Results: The fraction of apoptotic cells and the expression of WT1, 14-3-3 ζ, and CD4 increased upon ART treatment in tumors of patients. ART was bound in silico to a domain located at the DNA-binding site of WT1, while dihydroartemisinin (DHA) was bound with low affinity to a different site of WT1 not related to DNA-binding. The results were verified using microscale thermophoresis, where ART but not DHA bound to recombinant WT1. Transfectants overexpressing different WT1 splice variants exerted low but significant resistance to ART (≈2-fold). Conclusion: WT1 may represent a novel target of ART in cancer cells that contribute to the response of tumor cells to this drug.
Artesunate (ART) is an antimalarial agent belonging to the sesquiterpene trioxane lactone family (artemisinin-derived compounds). The lead compound artemisinin was isolated from the Chinese medicinal herb Artemisia annua that has been used for centuries in traditional Chinese medicine to treat malaria-related fever and other disorders (1). The combination of ART with piperaquine efficiently overcame multidrug resistance (MDR) of Plasmodium falciparum towards a wide range of antimalarial drugs (2, 3).
MDR is also a major hurdle in clinical oncology and hinders the success of many chemotherapeutic drugs used for cancer treatment. Various mechanisms account for MDR in cancer cells, e.g., ATP-binding cassette (ABC) transporters such as P-glycoprotein (ABCB1/MDR1), breast cancer resistance protein (ABCG2/BCRP), and ABCB5 that mediate drug resistance via ATP-dependent efflux of anti-cancer drugs (4, 5). Mutations and aberrations in genes involved in the control of apoptosis, e.g., the tumor suppressor protein (p53), also cause resistance (6, 7). Moreover, aberrations in oncogenes (e.g., of epidermal growth factor receptor, EGFR) suppress apoptosis and enhanced proliferation and angiogenesis of tumors resulting in treatment refractoriness (8, 9).
While the above-mentioned mechanisms confer resistance to established anticancer drugs, artemisinin derivatives exert cytotoxic activity in a wide range of cell lines obtained from diverse tumor types in vitro and in vivo (10, 11), and MDR cells exerted little or no cross-resistance to ART (12, 13). However, P-glycoprotein-expressing cells revealed cross-resistance to some other semisynthetic artemisinin derivatives (14, 15).
Despite artemisinin derivatives being used for many years to treat otherwise drug-resistant Plasmodium strains, recently, artemisinin-resistant malaria cases have been reported (16, 17). We also reported ART resistance in cancer cells, where the transcription factor NF-
B was a responsible resistance factor (18). Subsequently, the development of an ART-resistant cell line has been described (19).
The cellular and molecular modes of action of artemisinin-type drugs in cancer cells have at least in part been elucidated. ART arrests tumor cells in the G0/G1 or G2/M cell cycle phases (20), inhibited proliferation by affecting NF-
B DNA-binding and down-regulated VEGF (21). Furthermore, ART inhibits mammalian-target-of-rapamycin (mTOR) as well as the oncogene c-MYC mediated signaling pathways, which are dysregulated in many human tumors and lead to chemotherapeutic resistance (22, 23).
In the present study, we evaluated different tumor biomarkers in biopsies of patients, who were diagnosed with metastasized cervical carcinoma and received ART in a previous clinical pilot study (24), to find out the therapeutic effect of ART on different molecular targets that are vital for tumor progression. Among several other biomarkers, the expression of the Wilms Tumor 1 protein (WT1) was affected after treating patients with ART. Therefore, further in vitro analyses were performed to explore the ability of ART to affect WT1 as target protein either on the molecular level through binding experiments in silico or/and on the cellular level through cytotoxicity assays using WT1-transfected cell lines in vitro.
Patients and Methods
Patients. Matched biopsies of formalin-fixed and paraffin-embedded metastasized squamous cell carcinoma of the cervix uteri (stages 3 and 4) before, during, and after treatment with ART (Dafra Pharma Ltd., Turnhout, Belgium) were investigated. The clinical characteristics have been reported and biopsies of three patients were taken from a previously published open-label single-center study (24) and were involved in the present investigation. The patients were treated for 7 days with 100 mg/day and then 21 days with 200 mg/day. Those patients who experienced clinical relapse were treated for another 28 days with 200 mg/day (24). The patients’ characteristics were described in detail (24). All patients included in the study gave their written informed consent in the local common language of the region. The study protocol was reviewed and approved by the independent Ethics Committee, ‘Le Comité National d’ Ethique des Sciences de la Vie et de la Santé (CNESVS)’ of Ivory Coast as reported (24). Tissue biopsies were taken from each patient within three treatment stages, i.e., before, during, and after treatment with ART.
WT1-transfected HEK293 cell lines. Four cell lines (CW0, CW1, CW2, CW3) were generated by transfection of HEK293 cells (25). The cell lines were kindly provided by Prof. Dr. med. Manfred Gessler (Chair Developmental Biochemistry, Biocenter of the University of Würzburg, Germany). The four WT1 isoforms resulted from alternative splicing that differs in the presence or absence of 15 amino acids in exon 5 and three amino acids (KTS) at the end of exon 9. CW0 harbors a WT1 missing both splice sites (−E5/−KTS). Splice variant I (CW1) results from the absence of exon 5 (−E5/+KTS). Alternative splice variant II (CW2) lacks the three amino acids (KTS) between zinc fingers three and four (+E5/−KTS). The dominant-negative point mutation R394W in exon 9 of Denys-Drash Syndrome patients (DDS) was generated by PCR mutagenesis in CW3 (R394W/+E5/+KTS). All four cell lines were cultured in DMEM medium (Invitrogen) supplemented with 10% FBS and 1% penicillin/streptomycin (Invitrogen). All cell lines were incubated in a humidified 5% CO2 atmosphere at 37°C and passaged twice weekly.
Immunohistochemistry staining and evaluation. Ten commercially available antibodies were applied on paraffin-embedded tissue slides by using a polymeric labeling technique. Briefly, all slides were washed twice with xylene to remove paraffin. Primary antibodies used in immunohistochemistry staining were either monoclonal or polyclonal antibodies. Monoclonal antibodies included: WT1 (clone 6F-H2, Cat. No. MA1-46028, Thermo Fisher Scientific, Darmstadt, Germany, dilution 1:10), 14-3-3 zeta (14-3-3 ζ clone AT1A1, Cat. No. USBI144808, Acris, Heidelberg, Germany, dilution 1:100), CD4 (clone 4B12, Cat. No. MA5-12259, Thermo Fisher Scientific, dilution 1:10), CD8 (clone SP16, Cat. No. MA5-14548, Thermo Fisher Scientific, dilution 1:50), CD56 (clone 56C04, Cat. No. MA5-11563, Thermo Fisher Scientific, dilution 1:100), ATP-binding cassette transporter B5 (ABCB5, clone 5H3C6, Cat. No. MA5-17026, Thermo Fisher Scientific, dilution 1:200), glutathione S-transferase P1 (GSTP1, clone 3F2C2, Cat. No. MA5-15309, Thermo Fisher Scientific, dilution 1:500), inducible nitric oxide synthase (iNOS, clone K13-A, Cat. No. ab136918, abcam, dilution 1:150), translationally controlled tumor protein (TCTP, Cat. No. PA5-35332; Thermo Fisher Scientific, dilution: 1:300). Polyclonal antibodies included eukaryotic translation initiation factor e (eIF3e) (Cat. No. 10899-1-AP, Proteintech Germany GmbH, St. Leon-Rot, Germany, dilution 1:100) and ADP/ATP translocase (Cat. No. 17796-1-AP, Proteintech Germany, dilution 1:50). For the TUNEL assay, we used the Apoptag® Peroxidase Detection Kit (Cat. No. S7101, Millipore Sigma, Taufkirchen, Germany). Then, sample tissues were rehydrated through graded washes with isopropanol in water. Heat-induced epitope retrieval (HIER) was done using a pressure cooker as a heating device. Retrieval solutions used were citrate buffer and PBS. After heating, they were cooled down on ice for 25 min for all the antibodies. Ultravision protein block (Cat. No. TL-060-AL, Thermo Scientific GmbH, Karlsruhe, Germany) was added to block endogenous peroxidase and to avoid non-specific background staining. Overnight incubation at 4°C was done after the addition of primary antibodies. Then, horseradish peroxidase-labeled polymers conjugated with secondary antibodies were added. The final staining was performed with diaminobenzidine, and the slides were counterstained with hematoxylin.
The representative sections of three patients were immunostained. Each six areas of each section have been selected by two observers, scanned by using Panoramic Desk (3DHISTECH, Budapest, Hungary) and quantified by panoramic viewer software (NuclearQuant and MembraneQuant, 3DHISTECH), Quantification (percentage of positivity) was calculated by dividing the number of positively stained cells by the overall number of cells found each six independently annotated tumor areas (26).
Molecular docking. Molecular docking is a predictive method to evaluate the affinity and geometry of ligands with target macromolecules, before starting the docking essential hydrogens and Gasteiger charges were added to the macromolecules. We used the X-ray crystallography-based structure of the zinc finger domain of the WT1 protein (PDB ID: 2PRT) that resembles the splice variant, where three amino acids were assigned to positions 408, 409, and 410, respectively (+KTS). This zinc finger domain contains four canonical Cys(2)His(2) zinc fingers. The binding affinities were calculated for ART, artemether, arteether, DHA, and artemisinin. A grid box was constructed to define docking space. Docking parameters were set to 250 runs and 2,500,000 energy evaluations for each cycle. A Lamarckian Algorithm that was built on AutoDock 4.2.6 (AutodockTools-1.5.7rc1) was used to perform three times independent docking (27). The corresponding lowest binding energies and predicted inhibition constants were obtained from the docking log files (dlg). The mean values±SD of binding energies were calculated from three independent dockings. Visual Molecular Dynamics (VMD) was used to depict the docking poses of the artemisinin-type compounds on WT1.
Microscale thermophoresis. In vitro protein binding assays were performed to validate the in silico interaction between WT1 and ART). The recombinant WT1 (OPCD00233) was commercially obtained from Aviva Systems Biology (San Diego, CA, USA). This protein represents isoform 2 lacking the KTS motif (+E5/−KTS). Isoforms lacking KTS bind to DNA and act as transcription factors, while isoforms containing KTS bind to RNA and play a role in RNA metabolism. Therefore, we have chosen isoform 2 for the microscale thermophoresis studies. The protein was labeled according to the Monolith™ NT.115 Protein Labeling Kit RED-NHS (NanoTemper Technologies GmbH, Munich, Germany). Varying concentrations of ART (Saokim Co. Ltd., Hanoi, Vietnam) or DHA (Sigma-Aldrich, Taufkirchen, Germany) ranging from 1×10−2 to 1×106 nM were titrated with labeled WT1. The binding assay experiments were carried out using standard capillaries in the NanoTemper Monolith™ NT (NanoTemper Technologies GmbH, Munich, Germany) as previously described (28, 29).
Resazurin assay. Living cells are metabolically active and can reduce the non-fluorescent dye resazurin to the strongly-fluorescent dye resorufin (30). The resazurin (Promega, Mannheim, Germany) reduction assay was performed to assess the cytotoxicity of ART towards drug-sensitive and -resistant cell lines. The assay has been previously described (31, 32). IC50 values have been calculated from the dose-response curve and the resistance ratio was determined by dividing IC50 of the resistance cell line by IC50 of the parental cell line. For each assay three replicates were done.
Statistical analyses. The statistical analyses were done by using the ANOVA test followed by Bonferroni’s post hoc test to compare multiple groups as part of the SPSS Biostatistics software (IBM, Ehningen, Germany). p-Values <0.05 were considered as being statistically significant.
Results
Immunohistochemical evaluation of novel resistance markers in cervical carcinoma biopsies highlighted WT1. In the present clinical study, we evaluated a set of 11 protein markers by immunohistochemistry and apoptosis by the TUNEL assay at three intervals (before, during, and after treatment) to get an insight of ART’s effect. We selected examples from the fields of drug resistance (ABCB5, GSTP1), immunology (CD4, CD8, CD56), tumor-related transcription factors (WT1, eiF3e), apoptosis (TUNEL assay), cell growth regulators (14-3-3 ζ, TCTP, ADP/ATP translocase), and inflammation (iNOS). This allowed us to investigate, using selected examples, whether proteins of certain functional groups reflect therapy effects better than others.
The results of the immunohistochemical staining quantifications are shown in Figure 1. A strong increase of apoptotic cells (TUNEL assay) was observed in cervical cancer biopsies after ART treatment compared to biopsies from the same patients before treatment. Cell-mediated immune response was activated by CD4-positive cells upon ART treatment but not by CD8-positive T-cells, as indicated by increased CD4 immunostaining after ART treatment. An up-regulation of 14-3-3 ζ as a cell cycle regulatory protein and WT-1 as a tumor suppressor was also observed after treatment of cervical carcinoma patients with ART (Figure 1). However, the differential expression levels did not considerably vary for ABCB5, GSTP1, iNOS, TCTP, elFe3, ATP/ADP translocase, or CD56 (Figure 1). Since WT1 has not previously been studied in the context of the anticancer activity of artesunate, we focused on WT1 in subsequent experiments. Representative immunostaining photographs are shown in Figure 2.
Quantification of immunohistochemical staining of biomarkers. Wilms tumor protein 1 (WT1), 14-3-3 ζ, cluster of differentiation markers (CD4, CD8, CD56), ATP-binding cassette transporter B5 (ABCB5), glutathione S-transferase P1 (GSTP1), inducible nitric oxide synthase (iNOS), translationally controlled tumor protein (TCTP), eukaryotic elongation factor 3 (eIF3), and ADP/ATP translocase in matched cervical carcinoma biopsies taken before, during, or after treatment with artesunate (ART).
Immunohistochemical determination of tumor markers in formalin-foxed, paraffin-embedded cervical carcinoma tissue. Representative photographs of (A) apoptotic cells (TUNEL assay), (B) WT1, (C) 14-3-3 ζ, (D) CD4, (E) CD8, (F) ABCB5, (G) GSTP1, (H) iNOS, (I) TCTP, (J) eiF3e, (K) ADP/ATP translocase, (L) CD56, and (M) negative control (without primary antibody). One representative section of each from three patients were investigated. Six areas of each section were selected for quantification using Histoquant software. The methodology has been previously described in detail (26). Magnification: 20×.
ART interacts with WT1 in silico. We analyzed the interaction of ART to WT1 using WT1’s crystal structure. Since the whole human WT1 protein has not been crystallized to date, we obtained a predicted model of human WT1 from AlphaFold Protein Structure Database (https://alphafold.ebi.ac.uk/; entry number: AF-P19544-F1). The location of the zinc finger domain of WT1 is labeled in red in Figure 3A. To obtain optimal molecular docking results, we used the crystal structure of the zinc finger domain of human WT1 (PDB ID 2PRT) but not the homology model. As shown in Figure 3B, ART binds to a specific site that is involved in the interaction of WT1 with DNA. The full interface between WT1 and DNA is shown in Figure 3C. In addition to ART, we also docked other artemisinin-type drugs to WT1, i.e., artemisinin, dihydroartemisinin (DHA, artenimol), artemether, and arteether (Figure 3B). Artemisinin and arteether bind to the same site as ART (termed binding domain 1), indicating that these three compounds might hinder the binding of WT1 to DNA and subsequently WT1’s function as a transcription factor. In contrast, DHA and artemether interact with a different site of WT1 (termed binding domain 2) (Figure 3B). This binding site is not involved in the DNA binding of WT1 (Figure 3C), indicating that the latter two compounds may not inhibit the DNA binding of WT1.
Visual representation of the in silico molecular docking of artemisinin-type compounds to the zinc finger domain of WT1 result. The lowest binding energy positions of the artemisinins are shown. Each ligand is displayed with the interacting amino acids. (A) Homology model of human WT1 based on PDB ID: 2PRT. (B) Docking of artemisinin-type compounds to two different binding domains of the zinc finger domain of WT1. ART, artemisinin, and arteether bound to binding domain 1, while artenimol (dihydroartemisinin, DHA) and artemether bound to binding domain 2. (C) Binding of the zinc finger domain of WT1 to DNA. Binding domain 1 but not binding domain 2 is at the WT1-DNA interface. WT1 binds to the DNA sequence 5’-GCG(T/G)GGGCG-3’.
ART revealed lower free binding energy (FBE) and pKi values than the other compounds. The highest FBE and pKi values were measured for DHA, indicating that this compound might not interact with WT1 well. The FBE and pKi values are shown in Table I.
Molecular docking of different artemisinin derivatives on WT1. The lowest binding energies (LBE) and predicted inhibition constants (pKi) are shown below.
ART, but not DHA, binds to recombinant WT1 in vitro. To validate the in silico results, we performed in vitro binding assays by using microscale thermophoresis to measure the affinity of WT1 to ART (which was predicted in silico to bind to WT1) and to DHA (which was predicted in silico not to bind to WT1). The equilibrium dissociation constant (KD) obtained was 5.4 μM (Figure 4A), whereas DHA did indeed not bind to WT1 (Figure 4B).
Analysis of the interaction between ART (A) and DHA (B) with the zinc-finger domain of human WT1 by microscale thermophoresis. The recombinant protein was used at a concentration of 200 nM, while the concentration of the two compounds ranged from 1×10−1 to 1×10−6 nM. The migration of the fluorescent protein was determined upon local heating using a Monolith NT.115Pico with 20% laser power, 50% LED power at room temperature.
ART inhibited the viability of WT1-transfected HEK293 cells. Resazurin assays were performed in HEK293 cell lines transfected with four isoforms of WT1. Non-transfected HEK293 cells were used as a control. The IC50 values ranged from 3.09 to 6.27 μM. The WT1 isoform 1 which lacks exon 5 and three amino acids extension (KTS) at the end of exon 9 (−E5/−KTS) was more sensitive to ART than the other isoforms. All transfected cells were slightly more resistant than the non-transfected wild-type HEK293 cells (Figure 5), indicating that WT1 confers low-level resistance to ART.
Cytotoxicity of artesunate towards HEK293 cells transfected with cDNAs coding for four WT1 isoforms as determined by the resazurin assay. The four WT1 isoforms of the transfected HEK293 cells were: +E5/+KTS (CW0), −E5/+KTS (CW1), +E5/−KTS (CW2), and R394W/+E5/+KTS CW3 The results represent the mean±SD of three independent experiments, each with 6 parallel measurements. *p<0.05. The insert with the genomic organization of the four WT1 isoforms has been modified according to Thäte et al. (25).
WT1 expression was correlated with shorter survival times of cervical carcinoma patients. Finally, we were interested in clarifying whether WT1 expression is of prognostic relevance. For this reason, we mined the KM plotter (https://kmplot.com/analysis/) where the RNA-sequencing based WT1 mRNA expression of cervix carcinoma biopsies along with the clinical survival data of the patients are deposited. As can be seen in Figure 6, high WT1 expression was significantly correlated with short relapse-free survival times of patients with stage 3 tumors (log rank p=0.0054) with a hazard ratio of 4.39 (confidence interval=1.41-13.65). High WT1 expression was also associated with shorter overall survival times, but this relationship was only of borderline significance (logrank p=0.052) with a hazard ratio of 2.25 (confidence interval=0.97-5.22).
Kaplan-Meier survival statistics of cervix carcinoma. The WT1 mRNA expression was determined by RNA-sequencing of cervix carcinoma biopsies (stage 3) and deposited in the KM Plotter (https://kmplot.com/analysis/). Panel A shows the relapse free survival of patients with high and low WT1 expression, panel B the overall survival. HR, Hazard ratio.
Discussion
The anticancer activity of ART has been demonstrated in numerous studies in vitro and in vivo. Published preliminary data from clinical Phase I/II trials (24, 33-36) and compassionate uses in single patients (37, 38) also indicate clinical anticancer activity. In one of these clinical trials, we treated cervical carcinoma patients with ART and found that the expression of several tumor biomarkers decreased, i.e., the tumor suppressor p53, the oncogene EGFR, the proliferation marker Ki-67, the angiogenesis marker CD31, while the expression of transferrin receptor (CD71) increased (24). Therefore, we extended the immunohistochemical analyses and investigated the expression of 11 proteins as well as apoptosis by the TUNEL assay in matched biopsies of cervical carcinoma patients before, during, and after monotherapy with ART. We tested markers from different functional groups, i.e., resistance markers (ABCB5, GSTP1), immunological markers (CD4, CD8, CD56), proliferation markers (14-3-3 ζ, TCTP), transcription factors (WT1, eIFe3), the inflammation marker iNOS, and ADP/ATP translocase as the master regulator of mitochondrial energy production.
We found a strong increase of apoptotic cells by using the TUNEL assay. This result corresponds with our earlier findings that ART kills cancer cells by the induction of apoptosis (39-41). Here, we demonstrate that apoptosis was not only induced in vitro, but also in the clinical setting.
Among the panel of proteins investigated, WT1, 14-3-3 ζ, and CD4 expression increased during and after ART therapy. 14-3-3 ζ regulates CDC25C in the cell cycle and is a proliferation-associated protein (42), further confirming that ART represses proliferation and arrests the cell cycle as shown earlier in vitro (20, 43). Previously, we reported that the expression of the proliferation marker Ki-67 decreased during ART therapy in the same set of cervical carcinoma biopsies (24). This agrees with the present results on 14-3-3 ζ for cervical carcinoma biopsies from patients.
The increase of CD4 as a marker of T-cells, monocytes, and macrophages in ART-treated cervical carcinomas speaks for an involvement of the immune system in the anticancer activity of ART. Our data on clinical cervix carcinoma biopsies expand on in vivo results in mice. Cervix cancer-bearing mice were treated with ART. The drug did not only inhibit the growth of orthotopic tumors but did show increased CD4-positive T cells in the peripheral blood (44). In a murine ovarian carcinoma model, ART induced Th1 differentiation of peripheral CD4 cells (45).
The increased expression of WT1 during and after ART therapy in cervical carcinomas is a novel finding and piqued our interest. It is a transcription factor, which can act both as a tumor suppressor or as an oncogene, depending on its mutational status (46). The wild-type protein is a tumor suppressor, while the splice variant lacking 17 amino acids in exon 5 (−17AA) and 3 additional amino acids between zinc finger 3 and 4 (−KTS) induced aggressive tumor phenotypes (47). Alternative splicing produces multiple WT1 isoforms, designated WT1 (−E5/−KTS), WT1 (+E5/−KTS), WT1 (−E5/+KTS), and WT1 (+E5/+KTS) (Figure 5, insert) (48). Insertion of KTS between zinc fingers 3 and 4 leads to a reduced DNA binding affinity and altered subnuclear localization of WT1 (49). A point mutation in WT1 (R394W) was identified in patients with the hereditary Denys-Drash syndrome, which is related to urogenital anomalies and predisposes to the development of Wilms tumors (50).
High WT1 expression correlated with poor prognosis and response to treatment not only in breast cancer, osteosarcoma, and leukemia patients (51-53) as mainly highlighted but also in many other tumor entities. A meta-analysis of 3620 solid tumors of various origins revealed that positive WT1 expression significantly correlated with worse overall survival of patients (54). Furthermore, WT1 expression was significantly associated with resistance to chemotherapy (55-60), antihormonal therapy (61, 62), and radiotherapy (63).
The immunohistochemical staining results in the present investigation revealed an up-regulation of WT1 expression upon treating patients with ART. This observation prompted us to speculate whether oncogenic WT1 plays a role in the development of resistance towards ART. We hypothesized that ART may bind to and inhibit WT1. As a counter-reaction, tumors up-regulate WT1 and become resistant to ART. This assumption was based on comparable observations with artemisinin-type drugs and glutathione S-transferases, which are also target and resistance proteins of artemisinins, such as WT1.
Artemisinin-type drugs are known to inhibit glutathione S-transferase in malaria, schistosomiasis, and cancer (64-67). On the other hand, glutathione S-transferase and other glutathione-related enzymes are up-regulated in Plasmodia and cancer cells resistant to artemisinins (68-70). It can be speculated that WT1 behaves like glutathione-related enzymes. They are inhibited by artemisinin-type drugs, and their up-regulation represents an escape mechanism for developing drug resistance.
To test this hypothesis, we first performed molecular docking of ART to WT1 and indeed found reasonable binding energy of WT1 to ART but not to the same extent as to other artemisinin-type compounds. This in silico result was then biochemically verified by using microscale thermophoresis, which also showed that ART, but not DHA, was able to bind to recombinant WT1 protein. The consequences of this binding were further investigated in WT1-transfected cell lines. As shown by immunoblotting, incubation of transfected cells with ART for 72 h led to the down-regulation of WT1 expression.
There are two basic mechanisms whereby WT1 can be down-regulated (71); the first one concerns posttranslational down-regulation by proteasomal degradation. It can be assumed that the complexation of artesunate to WT1 leads to a non-functional protein that is subjected to ubiquitination and subsequent degradation. Other small molecules, e.g., trichostatin A (72), bortezomib (73) have recently been shown to promote proteasomal degradation of WT1. Another mechanism could be transcriptional down-regulation by epigenetic and transcription factor-based mechanisms. The histone deacetylation inhibitor vorinostat inhibited WT1 mRNA expression (73). The transcription factor NF-
B mediates transcriptional activation of WT1 (71), and artesunate is a known inhibitor of NF-
B (74, 75). Therefore, the artesunate-mediated down-regulation of WT1 may simultaneously occur via two independent pathways: artesunate-induced WT1 protein degradation and NF-
B-mediated or epigenetically induced down-regulation of WT1 mRNA expression.
Subsequent cytotoxicity assays revealed that different WT1-transfected cell lines were more resistant to ART than non-transfected control cells. We observed a low but statistically significant increase in ART resistance in the WT1-transfected cell lines (≈2-fold). The splice variants −E5 and −KTS have been described to induce cell cycle arrest and apoptosis (76) by inducing growth factor receptor-related signaling pathways (EGFR, insulin-like growth factor transforming growth factor β) (77, 78) and inhibiting VEGF-regulated angiogenesis (47). Interestingly, the immunohistochemical investigations of our previous cervical carcinoma trial also revealed decreased EGFR expression and micro-vessel density (24), which implies that ART may have affected tumor growth and angiogenesis via WT1.
Finally, we were interested to see whether WT1 is of prognostic relevance for cervix carcinoma patients. Indeed, we found that high WT1 mRNA expression was significantly associated to shorter relapse-free and with borderline significance to overall survival (Figure 6). This implies that inhibition of WT1 by ART might have beneficial effects for cervix carcinoma patients to improve their survival rates. This speculation needs to be substantiated in randomized clinical trials in the future.
In conclusion, WT1 is a novel determinant of resistance to ART. Further detailed investigations are warranted to gain more insight into such an adaptive response mechanism.
Acknowledgements
We thank Dr. Wynand Roos (Institute of Toxicology, University Medical Center, Mainz, Germany) for his critical reading of the manuscript. Candela Cives-Losada is supported by a predoctoral scholarship (FPU) and its complementary mobility grant funded by the Ministry of Science, Innovation and Universities, Spain. The project was funded with the intramural budget from the Johannes Gutenberg University, Mainz, Germany.
Footnotes
Authors’ Contributions
M.E.M.S. and C.C.-L. evaluated the immunohistochemical slides with the HistoQuant software, performed the microscale thermophoresis experiment and resazurin assays. C.C.-L. performed western blotting. T.E. designed and supervised the project and wrote and edited the paper.
Conflicts of Interest
T.E. discloses two patents related to artesunate, but not related to the results of the present investigation (US20050252304, ES200401396). The results of the present investigation are an independent continuation of a previously published study, which had been financed by Dafra Pharma, Turnhout, Belgium (24). The present investigation was solely funded by intramural grants of the Johannes Gutenberg University without the financial support of Dafra Pharma. The financial support of the former study (24) did not influence the design and performance of experiments and evaluation of the results of the present paper. C.C.-L. and M.E.M.S. have no conflicts of interest.
- Received May 27, 2022.
- Revision received September 1, 2022.
- Accepted September 15, 2022.
- Copyright © 2022 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.