Mithramycin A Inhibits Colorectal Cancer Growth by Targeting Cancer Stem Cells

The pivotal role of cancer initiating stem cells (CSCs) in tumor initiation, growth, metastasis and drug resistance has led to the postulation of a ‘total cancer therapy’ paradigm, which involves targeting both cancer cells and CSCs for effective therapy. However, the progress in identifying drugs for total cancer therapy has been limited. Herein, we show for the first time that mithramycin A (Mit-A) can successfully inhibit CSC proliferation, in addition to inhibiting bulk cancer cells in a model of colorectal cancer (CRC), the second leading cause of death among men and women in the United States. To this end, a polymeric nanofiber scaffold culture system was established to develop 3D tumor organoids (tumoroids) from CRC cell lines such as HT29, HCT116, KM12, CT26 and MC38 as well as ex vivo mouse tumors. These tumoroids possessed increased expression of CSC markers and transcription factors, expanded the number of CSCs in culture and increased CSC functional properties measured by aldehyde dehydrogenase activity. Screening of an NCI library of FDA approved drugs led to the identification of Mit-A as a potential total cancer therapy drug. In both sphere and tumoroid culture, Mit-A inhibits cancer growth by reducing the expression of cancer stemness markers. In addition, Mit-A inhibits the expression of SP1, a previously known target in CRCs. Moreover, Mit-A significantly reduces growth of tumoroids in ex vivo cultures and CRC tumor growth in vivo. Finally, a dose-dependent treatment on CRC cells indicate that Mit-A significantly induces the cell death and PARP-cleavage of both CSC and non-CSC cells. Taken together the results of these in vitro, ex vivo and in vivo studies lead to the inference that Mit-A is a promising drug candidate for total cancer therapy of CRCs.


Results
Tumoroid culture of colorectal cancer cell lines expands CSCs. Previously, we reported that breast cancer cells cultured on 3D polymeric nanofiber scaffold (Fig. 1A) form tumoroids, which substantially (at least 5-fold) expand CSCs as determined by CSC biomarker expression and activity of aldehyde dehydrogenase enzyme (ALDH) 14 . Since CSC expansion of CRC tumoroids is hitherto unknown, we cultured three human CRC cells lines, HT29 (p53 mutant, K-RAS wild type, microsatellite stable), HCT116 (p53 wild-type, K-RAS mutant, microsatellite instable) and KM12 (p53 mutant, K-RAS wild type, microsatellite instable) 21 , and CT-26 murine cancer cells (p53 wild-type, K-RAS mutant, microsatellite stable) 22 on 3D scaffold for 6 days and examined tumoroids for stemness markers by qPCR and flow cytometry. HT29 cells formed tumoroids when grown on the scaffold for 6 days (Fig. 1B,C). The SEM image showed typical tumoroid formation with a smooth surface and tight cell junctions (Fig. 1B). Nuc-blue stained HT-29 tumoroids are shown in Fig. 1C. To determine whether tumoroids formed on scaffold could undergo the epithelial to mesenchymal transition (EMT), we compared the HT-29 cells grown on monolayer vs. scaffold for expression of E-cadherin (epithelial marker) and αSMA (α smooth muscle actin) (mesenchymal marker). Immunofluorescence (IF) staining showed that over six days of culture, HT-29 tumoroids showed robust expression of αSMA but not E-cadherin. On the contrary, monolayer culture expressed E-cadherin but not αSMA (Fig. 1D). In addition, expression of the mesenchymal EMT marker, Snail, was also increased at both RNA and protein level in scaffold culture of HT-29 and HCT-116 compared to cells grown on monolayer ( Fig. 1E-H). These results suggest that HT-29 tumoroids induced EMT when cultured on the scaffold.
Moreover, HT-29 tumoroid culture showed 50-and 60-fold higher expression of transcription factors NANOG and OCT4, respectively, compared to cells derived from normal colon, CCD841, and 5-and 2-fold compared to monolayer cells ( Fig. 2A). Also, HT-29 tumoroids showed a 12-and 5-fold increase in transcription of LGR5 and stem cell related transcription factor SOX2, respectively, compared to monolayer (Fig. 2B). Further, tumoroids derived from CT-26 cells exhibited a gradual increase in expression of Lgr5 and Sox2 transcripts, as the tumoroids were successively passaged from P1 to P3 (Fig. 2C) in the 3D scaffold environment. In addition, tumoroids derived from HCT116 human colon cancer cells showed a 5-fold increase in ALDH activity (63.4%), which is a hallmark of stemness measurable by the aldefluor assay, over monolayer cells (12.5%) (Fig. 2D). These studies suggest that the scaffold induces increased expression of CSC markers and transcription factors both in human and mouse CRC cells.

Identification of Mit-A as an inhibitor of CRC CSC and tumoroid growth.
In the search for drugs that enable 'total cancer therapy' , we screened an NCI Oncology Drugs Set of FDA-approved drugs using tumoroid platform in HT29 cells. Then, we selected the top nine drugs as candidate CSC inhibitors and assayed their ability to decrease viability of CRC CSCs using HT29 tumoroids. The tumoroids were treated for 48 hours with different drug candidates including cisplatin. Of these, Mit-A was found to have the lowest IC50 of all drugs tested (in the nanomolar range) (Fig. 3A). Thus, Mit-A was considered to be the most effective against HT29 cells.
In order to understand the mechanism of Mit-A inhibition of tumoroid growth, we cultured and treated tumoroids derived from HT-29, HCT-116 and KM12 cells. As shown in Fig. 3B-D, 48 hours of Mit-A treatment inhibited the growth of tumoroids. Both the number and size of HT-29, HCT116 and KM12 tumoroids were decreased in a dose-dependent manner. However, only 20% growth inhibition was observed when HT29 tumoroids were treated with a mixture of 25 µM of 5-FU and 0.625 µM of Oxaliplatin (FUOX) (Fig. 3E).
To compare whether these physiological features were due to changes in stemness, we analyzed the expression of stemness markers through qPCR, flow cytometry and/or Western blotting. Both HT-29 and HCT-116 tumoroids showed a decrease in stemness factors when treated with Mit-A. Interestingly, we found that Mit-A inhibited CD133 expression in a dose-dependent manner in both cells by qPCR and Western blotting (Fig. 4A,D,E). Both qPCR and flow cytometry analyses showed that Mit-A significantly decreased expression of LGR5 in HT-29 and HCT116 (Fig. 4A,B,C). It was already known that Mit-A can inhibit cell growth through inhibition of SP1, a stemness-related transcription factor 17 . We tested whether Mit-A can also inhibit SP1 in CRC cells. Mit-A treatment was found to inhibit the SP1 expression in HCT116 and HT29 cells, respectively (Fig. 4E,F). Lastly,  Fig. 1) in a dose-dependent manner. In order to compare the efficacy of Mit-A with standard of care chemotherapeutics in suppressing sphere growth, we chose the minimal concentration of Mit-A to significantly reduce the number or size of HCT116 spheres (30 nM) and found that Mit-A was potent in inhibiting sphere growth whereas the standard care drugs Carboplatin, Oxaliplatin and 5-FU could not (Fig. 5G). There was no significant change in the sphere area with these conventional drugs compared to control. Since the spheres are known to highly express the stem cell markers, we also tested whether the reduction of sphere formation ability is due to a decrease in the expression of these markers. qPCR analysis data showed that Mit-A inhibited the colonosphere formation through the inhibition of potential stemness marker OCT4 (Fig. 5H,I) in both HT29 and HCT116 cells, respectively.

Mit-A can inhibit the ex vivo and in vivo colon tumor growth.
We have previously reported that ex vivo culture of tumoroids from tumor biopsy-derived cells retained stromal cells, such as cancer-associated fibroblasts, immune cells and endothelial cells even after 6 days of culture 12,13 . Since tumor microenvironment modulates the responsiveness to drugs in vivo, we examined the growth inhibition potential of Mit-A in ex vivo MC-38 (murine CRC cells, p53 mutant, K-RAS wild-type, microsatellite instable) 23 and HT-29 tumoroid growth. To test this, MC38 CRC cells (10 6 cells/mouse) were implanted subcutaneously in C57BL/6 mice. A single cell suspension of tumors isolated from mice was cultured ex vivo on the scaffold, which formed tumoroids (Fig. 6A). Flow cytometry data showed a 2.8-fold increase in ALDH activity in the tumoroids derived from MC38 tumors compared to www.nature.com/scientificreports www.nature.com/scientificreports/ monolayer ( Fig. 6B; 17.6% in tumoroids vs. 6.7% in monolayer). Moreover, Mit-A treatment inhibited the growth of ex vivo MC-38 tumoroids in a dose dependent manner (Fig. 6C). Similarly, Mit-A also reduced the growth of ex vivo HT-29 tumoroids (Fig. 6D) in a dose-dependent manner.
To determine the antineoplastic effect of Mit-A in vivo, we implanted mouse tumor cells MC38 and CT26 into C57BL/6 and BALB/c mice, respectively. Intraperitoneal injection of Mit-A of 1 mg/kg/day significantly inhibited the tumor growth and tumor weight in both mice (Fig. 7A-D). The body weight did not change significantly www.nature.com/scientificreports www.nature.com/scientificreports/ between the groups during the course of the treatment (Supplemental Fig. 2), showing the negligible toxicity of Mit-A on these mice. qPCR analysis also confirmed that Mit-A inhibition of the CT26 tumor growth correlated to the reduced expression of CD133 (Fig. 7E) suggesting that Mit-A targets CSCs in vivo. www.nature.com/scientificreports www.nature.com/scientificreports/ We next sought to determine the mode of cell death induced by Mit-A. A dose-dependent treatment on HT29 and HCT116 monolayer cells for 48 hours followed by a cell death pathway analysis indicated that Mit-A significantly induces cell death in both cell lines (Fig. 8A,B). To analyze the cell growth suppression mechanism induced by Mit-A, we cultured CRC cells on scaffold and treated with Mit-A for 48 hours. As expected, Mit-A induced PARP1 mediated mitochondrial cell death in CRC cells. In both HT-29 and HCT-116 cell lines, Mit-A induced PARP cleavage in a dose-dependent manner (Fig. 8C,D). Also, in HT-29 cells Mit-A induced cleavage www.nature.com/scientificreports www.nature.com/scientificreports/ of Caspase-9 suggesting activation of mitochondrial cell death in these cells (Fig. 8D). To analyze whether Mit-A can induce cell death in CSCs, we cultured HCT116 cells in ultra-low attachment plate in suspension culture to induce stem cell growth and treated with Mit-A for 5 consecutive days. We found that Mit-A can kill CSCs by PARP mediated cell death pathway (Fig. 8E). Taken together, these studies suggest that Mit-A can induce apoptosis of both CSC and non-CSC CRC cells.

Discussion
CRC is often diagnosed at the advanced stage leaving the patient to be treated only by chemotherapy. The conventional treatment of CRC includes 5-FU, FOLFOX, and cisplatin-based treatments, which are not effective when used either alone or in combination 4 . Also, in the advanced stage patients rapidly develop chemoresistance and >50% of patients have a relapse of cancer. Therefore, it is an urgent need to establish new chemotherapeutic drugs that can advance 'total cancer therapy' , i.e., can effectively kill not only cancer cells, but also drug resistant CSCs that cause tumor relapse. To this end, we have identified Mit-A as a 'total cancer therapy drug' and characterized it to show for the first time that it can successfully inhibit CSC proliferation in addition to inhibiting the growth of bulk cancer cells.
Mit-A has been reported to inhibit growth of cancer cells of squamous cell carcinoma 24 , cervical 17 and prostate cancer 25 origin, however, its role in inhibiting CRCs has not been investigated heretofore. Additionally, the role of Mit-A in the emerging concept of CSC-targeted treatments has not been studied. A major finding of this study is that we show here for the first time that Mit-A can inhibit both human and mouse CRC growth, irrespective of the driver mutations they carry. Furthermore, Mit-A inhibits CRC growth through the suppression of proliferation of colon CSCs as revealed by reduced expression of CSC markers. Our results demonstrate that Mit-A effectively suppresses CSC growth through stem cell marker inhibition in conventional sphere culture and recently established CRC tumoroid culture on 3-D polymeric scaffold system 12-14 which rapidly expands CSCs.
CSCs play a pivotal role in CRC progression, metastasis, chemoresistance and cancer relapse. However, there is not a single stem cell marker for CRCs that can be regarded as the most important marker enabling tracking of these cells. Nonetheless, CD133 is considered as an important marker of CRCs since CD133 positive cells have been associated with tumor-initiating properties, whereas CD133 negative cells are not 26 . Additionally, there exist www.nature.com/scientificreports www.nature.com/scientificreports/ several other CSC markers including LGR5, SOX2, NANOG, OCT4 27,28 . Furthermore, SP1, a transcriptional activator, has been shown to play a crucial role in chemoresistance in CRC 20 . We and others have shown that there is a good correlation between inhibition of CSC marker expression and inhibition of tumor initiation in vivo in breast, lung and also CRC 12,29-32 .
Another major finding of our studies is that Mit-A can inhibit the number and growth of both human and mouse CRC cells in sphere cultures that are enriched in CSCs. This sphere growth inhibition was evident from the suppression of the stem cell markers like CD133, LGR5, ALDH. Of note, ALDH1 is one of the major CSC markers 33 . Polymeric 3D scaffold-based cultures provide an excellent tool to expand cancer stemness and amplification of CSCs and such culture could be used to discover new anti-CSC drugs. Cancer stemness can be maintained through repeated passaging of cells isolated from spheres 34 and similarly the results of our tumoroid cultures showed that cancer stemness increases through successive generations. In addition, we successfully established ex vivo tumoroid culture from MC38 and HT29 tumors grown in mice, which allows us to expand the rare CSC population from in vivo tumors and simultaneously screen for CSC targets. It is noteworthy that Mit-A was able to effectively suppress the ex vivo tumoroid growth of MC38 and HT-29 tumoroids. In addition, we have also shown that recurrent intraperitoneal injection of Mit-A can suppress CRC cell growth in mice. This reduction with marginal insignificant body weight change showed that Mit-A causes minimal toxicity and therefore can be a good therapeutic candidate. These results are consistent with previous reports of Mit-A inhibition of prostate, breast and cervical cancers 17,24,25 .
In conclusion, we have shown here for the first time that Mit-A induces apoptosis of both CSC and non-CSC CRC cells through inducing the cleavages of PARP and caspase 9 (Fig. 9). Mit-A can inhibit the growth of colorectal CSCs through the inhibition of transcription factors and stem cell markers including but not limited to SP1, CD133, LGR5 and OCT4. An in silico analysis by using the Ingenuity Pathway Analysis software by Qiagen Bioinformatics showed that Mit-A can also regulate P53, TGFβ receptor 2 (TGFR2), and MDM2 through SP1, which can have potential impact in cell survival pathway. Although it is still unclear how these proteins could be directly impacted by Mit-A and affect cell survival/death, more studies are warranted to understand the mechanism of Mit-A in suppressing cancer growth. These findings, taken together, suggest that Mit-A could be a potential drug for treating CRC patients. antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-cleaved PARP (5625 S), anti-cleaved caspase-9 (9501P) antibodies were from Cell Signaling. Anti-CD133 (ab19898) antibody was from Abcam (Cambridge, MA). Mithramycin A (Sigma, M6891) was purchased from Sigma and was dissolved in DMSO. Accutase (AT-104) was bought from Innovative Cell Technologies. E cadherin (sc-8426) was purchased from Santa CruzBiotechnology, Inc., and α-smooth muscle actin (αSMA) from Abcam.
Cell culture and drug treatments. HT29, HCT116, SW480 and CT26 were purchased from ATCC. HT29 and HCT116 cells were maintained in McCoy's 5A medium containing 2 mmol/L L-glutamine, 100 U/mL penicillin, 100 mg/mL streptomycin, and 10% FBS. SW480 and CT26.WT media were cultured in RPMI-1640 medium containing 2 mmol/L L-glutamine, 100 U/mL penicillin, 100 mg/mL streptomycin, and 10% FBS. MC38 cells were kindly provided by Dr. Shari Pilon-Thomas (Moffitt Cancer Center) and were cultured in complete medium as recommended 35 . KM12 cells were obtained from Dr. Bhaumik Patel, Richmond, VA. All cells were maintained in an atmosphere containing 5% CO 2 and at 37 °C. In addition, cells were routinely checked for mycoplasma contamination. C57BL/6 and BALB/c mice were purchased from The Jackson Laboratory. www.nature.com/scientificreports www.nature.com/scientificreports/ Tumoroid culture. Polymeric nanofiber scaffold was prepared as previously described 12 . Scaffolds were sterilized in ethanol and prepared for tumoroid culture as described 12 . Cell lines were obtained from ATCC. Cells were cultured in a humidified incubator at 37 °C in a 5% CO 2 atmosphere (Thermo Fisher). 3D Tumoroid formation was assessed using fluorescent microscopy (Olympus BX51) after nuclear staining with Nuc Blue dye (Thermo Scientific). For immunofluorescence (IF) staining the tumoroids were washed with PBS and incubated with E-cadherin (sc-8426) (1:500) and α-smooth muscle actin (SMA) primary antibodies (1:500) overnight separately and then with secondary antibodies (donkey anti-mouse (594) 1:1000 and donkey anti-goat (488) respectively). The monolayers were stained the next day of plating and were stained similarly. The images were taken in confocal microscope (Olympus FV1200 Laser Scanning Confocal Microscope).
Scanning electron microscope (SeM). SEM was performed using methods as described previously 12 .
Briefly, the day 4 tumoroids on scaffolds were fixed in 4% formaldehyde and then sodium cacodylate and osmium tetroxide were added to the scaffold with intermediate shaking at each step for 5 min at RT. Following the above procedure, 30% and then 70% ethanol was added after which the scaffolds were incubated overnight. Next day the scaffolds were submerged in HMDS and 50% ethanol and dried to air following which they were mounted for sputter coating.
Sphere assay. To measure the sphere formation ability, cells were trypsinized and were plated at a density of either 100 cells or 200 cells per well in a 96 well ultra-low attachment surface plate (Corning Inc. 3474) in serum free DMEM/F-12K and DMEM High Glucose (1:1) media containing 0.004 gm/ml BSA, epidermal growth factor (100 ng/ml), fibroblast growth factor (20 ng/ml), insulin (5 µg/ml) and B27. Cells were incubated in 5% CO 2 and at 37 °C and treated with Mithramycin A for 5 days. Sphere pictures were taken manually using a bright field microscope and total area of the spheres was calculated using Image J software. Each experiment represents a combination of three replicates (n = 3). The data was repeated at least twice. cell viability assay. Cell growth was quantified using CellTiter-Glo ® Luminescent Cell Viability (Promega, G7572) assay. For monolayer culture, cells were plated in a 96 well plate and treated the next day with Mithramycin A as indicated (n = 4). For scaffold culture, cells were plated in a 96 well plate in 50 µl volume to stabilize cells on the scaffold (n = 4). Next day, 150 µl fresh media was added. On day 4 of plating, cells were treated with Mithramycin A. In both cases, 48 hours after treatment, cell titer glow reagent was added with media (1:1 ratio) according to the manufacturer's protocol and the luminescence signal was read by an illuminometer in an opaque plate.
Western blot analysis. Whole cell lysates were extracted using lysis buffer and equal amounts of proteins were loaded in an 8-10% SDS-PAGE gel. Samples were transferred from the gel onto a nitrocellulose membrane (Bio-Rad, 162-0112) and incubated with 5% skim milk for an hour to block the membrane. Indicated antibodies were used for overnight incubation at 4 °C and the blots were developed using ECL substrates after HRP-conjugated secondary antibody incubation. β-Actin or α-Actinin blots were run for internal control. flow cytometry. To measure the cell marker expression associated with stemness, flow cytometry was performed. Cells from both monolayer and scaffolds were grown and treated with either vehicle or Mithramycin A as indicated. A single cell suspension was obtained by treating the cells with Accutase. Cell marker expression was observed by using Aldefluor kit (Stem Cell Technologies, 01700), anti-CD133 antibody (Miltenyil 130-113-188) and anti-leucine-rich repeat-containing G-protein coupled receptor (LGR5) (Abcam, 75850) antibody. Expression of the markers was analyzed with a BD FACS Canto II fluorescent activated cell sorter.