Nanoquinacrine induced apoptosis in cervical cancer stem cells through the inhibition of hedgehog-GLI1 cascade: Role of GLI-1

To improve the pharmacokinetics and to study the anti-cervical cancer and anti-stem cells (CSCs) mechanism of Quinacrine (QC), a spherical nano particle of QC (i.e. NQC) was prepared and characterized. QC and NQC showed higher cytotoxicity in multiple cancer cells than the normal epithelial cells. NQC exhibited more toxicity in cervical cancer cells and its CSCs than QC. A dose-dependent decreased expression of Hedgehog-GLI (HH-GLI) components were noted in NQC treated HeLa cells and its CSCs. NQC increased the expressions of negatively regulated HH-GLI components (GSK3β, PTEN) and caused apoptosis in CSCs. Reduction of GLI1 at mRNA and promoter level were noted after NQC exposure. The expressions of HH-GLI components, GLI1 promoter activity and apoptosis were unaltered in NQC treated GLI1-knockdown cells. In silico, cell based and in vitro reconstitution assay revealed that NQC inhibit HH-GLI cascade by binding to the consensus sequence (5′GACCACCCA3′) of GLI1 in GLI-DNA complex through destabilizing DNA-GLI1 complex. NQC reduced the tumors size and proliferation marker Ki-67 in an in vivo xenograft mice model. Thus, NQC induced apoptosis in cancers through inhibition of HH-GLI cascade by GLI1. Detail interaction of QC-DNA-GLI complex can pave path for anticancer drug design.


Western Blot
Described in the main document.

Caspase-3 Immunofluorescence
The expression of caspase 3 on cervical cancer stem cell was done by immunofluorescence assay. For this assay briefly, 10,000 PEMT cells were seeded in the well of culture plates on the surface of cover slip and allowed to grow for 24 h 1 . Then the cells were treated with different concentrations of NQC for 48 h. Then cells were washed with 1X PBS and fixed with fixer acetone: methanol (1:1) at -20ºC for 20 min. After 20 min cells were blocked with 2% BSA and 0.2% triton X-100 in PBS for 20 min in 37°C. After that cells were incubated with primary antibody for 2 h. Then the cells were washed with 1X PBS and incubated with secondary antibody conjugated with TRITC and incubated for 1 h. After that cells were washed with 1X PBS and stained with DAPI for 10 min.
Again cells were washed and images were taken using fluorescence microscope (AMG Evos Fluorescence Microscope, Thermo Fisher Scientific, MA, USA).

Proteasome mediated inhibition study
To study the impact of proteasome meditated degradation of GLI1 by NQC in PEMT cells, approximately 1×10 6 cells were seeded in a 6 well cell culture plate. After 80% confluency cells were pre-treated with MG-132 (5µM) for 3h followed by NQC treatment for 48 h prior to harvest 2 . Then whole cell lysates were prepared and processed for western blot experiment.

UV-Vis spectroscopic study
To check the drug -DNA interaction we have used UV-VIS spectrophotometric based assay.
For this a fixed concentration of DNA (50µg/mL) was incubated to varied concentrations of NQC (upto 1 µM) in solution of 50 mM Tris-HCl/NaCl (pH 7.5) for 1h at 37ºC 1 . We have used two types of DNA i.e. wild type APC which do not contain GLI-DNA binding consensus sequence and PGL2GLI-LUC plasmid (which contain the GLI-DNA binding consensus sequence). After end of the incubation an absorption spectra was taken within wavelength 200-320 nm.

Molecular docking studies
Protein Structure Preparation: The 3 dimensional crystal structure of GLI-DNA complex is available in the Protein Data Bank (PDB ID: 2GLI) at a resolution of 2.60 Å 3 . The GLI-DNA complex, GLI and DNA structures were further prepared for the molecular modeling studies using Protein Preparation Wizard module of Maestro 9.3 package 4 . The missing hydrogens were added and right bond order was assigned. For optimizing the orientations of hydroxy group (in Ser, Thr and Tyr), amino group (in Asn and Gln) and ionization state (His), protassign utility and impref utility of the Protein Preparation Wizard were used. Glide docking score 6 . The generated DNA-GLI-QC ternary complex was found to be satisfactory in terms of interactions and score. The QC is a DNA intercalator occupying the regions having GC base pair. Therefore, a manual intercalation of the DNA was performed between the GC base pairs present near the docked pose (dG10···dC13 and dG11···dC12). The cationic tail was facing towards the minor groove and the acridine ring was inserted at the base pair interfaces. This complex was further optimised using molecular dynamics simulation.

NQC induces apoptosis in cervical cancer cell line by targeting HH-GLI axis
In order to confirm further the effect of NQC on HH-GLI pathway, protein

Proteasomal degradation of GLI1
The GLI1 expression was found to be decreased after NQC exposure (Figure 6b main text) in PEMT cells. In order to evaluate whether the decrease was due to proteasomal degradation of GLI1 after being destabilized from the DNA-QC-GLI1 complex, proteasomal degradation assay was carried out using proteasomal degradation inhibitor MG-132. PEMT cells were pre-treated with MG-132 and then exposed to increasing concentration of NQC. It was observed that blocking the proteasome using MG-132 restored GLI1 levels even after NQC treatment ( Supplementary Fig.S3a). GAPDH was used as a loading control.

Auto-activation of GLI1
GLI1 has been previously reported to bind specifically to a consensus promoter sequence of 5'GACCACCCA3' 7 . In order to check whether GLI1 binds to the promoter sequence of the GLI1 gene and activates itself, the presence of the consensus sequence in the promoter sequence of GLI1 gene was evaluated. Supplementary Fig.S3b shows the presence of the consensus sequence 5'CGGGTGGTC3'which is complimentary to 5'GACCACCCA3' 7 . In order to check whether NQC has any specificity towards the binding of this sequence an in vitro DNA-drug binding assay was carried out (Supplementary Fig.S3c). In principle, pure DNA gives a peak of absorbance spectra at 260nm wavelength, but when a drug binds to it the absorbance spectra shifts left as the binding of the drug decreases the absorbance spectra 2 . pGL2-GLI1 promoter was used for the in vitro study. It was observed that pure DNA gives an exact spectrum of 260 nm but with increasing concentrations of NQC, there is a shift in the absorption spectra indicating the possible binding of the NQC to the DNA sequence ( Supplementary Fig.S3c). Supplementary Fig. S3d gives the K d value i.e. half saturation value of binding isotherm and was found to be 2.5×10 -6 M -1 (Supplementary Fig.S3d). Supplementary   Fig. S3e is the spectral analysis of NQC to a wild type APC DNA and the spectral scan indicated that NQC has no significant role in binding to APC DNA. So, this result confirms that NQC very specifically binds to the GLI DNA sequence.

Electrostatistical analysis of QC and 2GL1 (DNA-GLI complex)
The available information on QC interaction with DNA states that the acridine ring of QC intercalates between the GC base pairs and the cationic tail interacts with the DNA backbone 8 . To understand the electrostatistical basis of DNA-GLI complexation, surface potential of GLI was visualised using Pymol. As expected, the DNA-GLI interface is complementary in terms of surface potential i.e. DNA has highly electronegative sugar-phosphate backbone whereas the GLI surface at the DNA-GLI interface is highly electropositive (Supplementary Fig. S4a). This confers the high stability to the DNA-GLI complex.
The electro-statistic nature of QC surface (visualised using Jaguar module of Maestro9.3) was analysed to understand the reported facts about QC-DNA interaction 9,10 . The QC has acridine ring which can intercalate in the GC base pair rich region. The cationic tail can interact with the DNA sugar-phosphate backbone suitably (Supplementary Fig. S4b). Due to similar electro-statistics of QC tail and GLI i.e. electropositive nature, GLI and QC cannot interact efficiently. Figure S4: Surface electrostatic potential of (a) DNA-GLI complex and (b) QC.

Supplementary
Red colour indicates electronegative potential whereas blue colour indicates the electropositive potential.

Cavity Analysis in the DNA, GLI and GLI-DNA Complex
The 3D macromolecular structure (PDB ID: 2GLI) was analysed for the possible binding sites in the crystal structure using CastP server 11 . For this purpose, three structures were analysed i.e.
DNA-GLI complex, DNA and GLI. The analysis revealed the possible binding sites in the macromolecular structures (Supplementary Fig.S5). The cavity volume and QC volumes were found to be complementary to the cavities in DNA-GLI binary complex. The molecular volume of QC is 9 291.2 Å 3 whereas the cavity volumes were 564.7 Å 3 and 440 Å 3 in the binary complex. The cavity volume in DNA and GLI were found to be 895.3 Å 3 and1051 Å 3 respectively. These results indicate that the binding sites formed due to the complexation of DNA and GLI can be considered as the binding pocket for the QC. These sites were further targeted for the molecular docking of QC in DNA-GLI complex. Further, the molecular docking of QC in DNA and GLI were performed in the possible binding sites.  Fig. S8b) so as to accommodate the NQC acridine ring. Supplementary Fig.S9 shows the surface potential of 2GLI (the crystal structure of DNA-GLI complex. The production run of 20 ns was analyzed for the stability of the system. The    (Table S1) to complex stabilization in ternary complex, the surface potential was visualized. The results were astonishing and unveiled a very important effect of QC intercalation on the DNA-GLI complex formation. Supplementary Fig. S10a, binary complex of DNA-GLI after molecular dynamics simulation of 20 ns. Supplementary Fig. S10b and ternary complex of DNA-GLI-QC after molecular dynamics simulation of 20 ns. The results indicate that the surface potential for crystal structure and the binary complex after dynamics simulation were similar. At the DNA-GLI interface, GLI have electropositive surface potential which is complementary to the electronegative DNA sugar-phosphate backbone. Therefore the DNA-GLI complexation is facilitated. In the ternary complex ( Supplementary Fig. S10c), the DNA is distorted due to QC intercalation. This further has led to the 15 GLI movements bringing electronegative surface close to the DNA backbone and thus producing the destabilizing effects on the ternary complex.