Cell culture and Cipa treatment
Cissampelos pareira L. whole plant extract was obtained commercially in lyophilized form, from a GMP certified manufacturer. Crude water extracts were reconstructed from powdered Cipa30. Fresh extracts were prepared right before each experiment. The chemical profiling of the extract was done using UPLC. The details are described in supplementary methods and the identified constituents are provided in supplementary Table S2. For subsequent experiments, guidelines set by CSIR, India were followed.
MCF7 (triple positive breast cancer) cell line was obtained from NCCS, Pune. It was maintained in DMEM high glucose media, supplemented with 10% FBS,1 M HEPES and 1X antibiotic antimycotics. The cells were kept at 37˚C and 5% CO2. MycoAlert (Cat no. LT07-318) was used to test the culture for mycoplasma.
MCF7 cells were treated with 1 µg, 10 µg, 100 µg, 500 µg and 1000 µg per ml of the aqueous extract from Cissampelos pareira L. The extract was prepared outside the cell culture hood and then filtered using 0.22 µm syringe filter before its use in cell culture. The MCF7 cells were seeded at a confluency of 60–70% 18–24 h. prior to treatment. The cells were treated with Cipa for 24 h for each experiment. A fresh extract was made for each set of experiments.
RNA isolation and whole transcriptome analysis
MCF-7 cells were seeded in 6 well plates at 60–65% confluency and then treated with vehicle, 1 µg, 10 µg, 100 μg, 500 µg and 1000 µg of Cipa for 24 h. The cells were then trypsinized and washed with PBS. Total RNA was isolated using TRIzol (Ambion, Cat no. 15596026) extraction method and its integrity was checked on 1% agarose gel which was followed by Nanodrop quantification (ND1000, Nanodrop Technologies, USA). Genome-wide transcriptome data was generated using 250 µg of the total RNA for each sample following the manufacturer’s protocol. GeneChip (Affymetrix) Human Transcriptome Array 2.0 cartridges were used for this experiment. HTA 2.0 chip can capture 245,349 protein coding transcripts and 40,914 non protein coding transcripts in a 64-format. Briefly, the total RNA was prepared with poly A controls and first strand cDNA followed by second strand cDNA were synthesized. This was followed by in vitro transcription to form cRNA which was used as a template for the formation of single strand cDNA or ss-cDNA. Finally, the ss-cDNA was fragmented and labelled. At each step, purification and quantification of the samples were ensured. Approximately 200 µl of the sample, containing about 5.2 µg of the labelled ss-cDNA was loaded into the cartridges, which were then kept in the Affymetrix Hybridization Oven, set at 45 °C and RPM 60, overnight. The cartridges were registered on AGCC (Affymetrix GeneChip Command Console) and the fluidics of the experiment (washing and staining) was done using Affymetrix GeneChip Fluidics Station 450. The stained chips were then scanned using GeneChip Scanner 3000. The preliminary images were quality checked and .cell files were generated.
The generated CEL files were background normalized using the RMA method. Batch effects were removed and differential gene expression analysis was done using the limma package in R. After background correction and RMA normalization, the probes that exceeded the p value < 0.05 were annotated and analyzed for differential expression. Since no probe could qualify the cut off at 1 µg, 10 µg and 100 µg, the differential gene expression was calculated using the dataset from 500 and 1000 µg treatments with the vehicle used as control. The genes with absolute log twofold change ≥ 1 and p value < 0.001 were taken as differentially expressed. The raw files and data have been submitted to GEO with the accession number GSE156445.
cDNA synthesis and qPCR
cDNA was prepared from 1 μg DNase-treated RNA using Applied Biosystems High-Capacity cDNA Reverse Transcription Kit (Cat no. 4368814) following manufacturer’s instructions. RT-qPCR was performed with using 2X SYBR Green I master mix (Kapa Biosystems, Cat no. KK3605) and the reaction was carried out in Roche Light Cycler 480. GAPDH (FP: CGACCACTTTGTCAAGCTCA, RP: CTTCCTCTTGTGCTCTTGCTG) and ATF6 (FP: GATGAAGATTGGGATTCTGC, RP: AGAAAGTGGCTGAGGTTCTG) were used as normalizing controls. ESR1: (Forward Primer: CAAGGAGACTCGCTACTGT and Reverse Primer: TTTCGTATCCCACCTTTCAT), RPL7: (Forward Primer: AAGCTCAACAAGGCTTCGAT and Reverse Primer: CCAAGAGATCGAGCAATCAA). The experiments were carried out in triplicates and fold change was calculated using 2 − ΔΔCT method.
Functional enrichment of the differentially expressed genes
For functional analysis we used g: profiler31 (https://biit.cs.ut.ee/gprofiler/gost). After correcting for FDR < 5%, enrichments were analyzed for GO: Molecular Function, Cellular Compartment, Biological Processes, KEGG and Reactome.
In order to identify the specific sets of genes modulated by Cipa, we used Gene set enrichment analysis. We carried out a pre-ranked analysis for GSEA (Gene set enrichment analysis), in which the gene list was ordered from the highest positive gene expression to the lowest negative gene expression. The latest versions of gene sets databases were selected for query. Datasets with more than 500 and less than 15 genes were excluded from the query. The output was set to a minimum cut off of p value < 0.05 and corrected for FDR < 25%. An additional filter of FDR < 5% was applied for identification of significant enrichments.
Estrogen response elements (ERE) analysis
Sequence of 5 KB upstream region of genes were downloaded from UCSC genome browser for Human genome GRCh38 build. ERE sites were mapped to 5 KB upstream sequence using promo tool32 with similarity cut off value of 8.7. High ERE density in 5 KB upstream leads to down regulation of genes as per regression analysis. ERE density’s coefficient of slope of regressed line (β = − 0.0044) is negative and p-value is significant (0.018). Pair wise ERE density comparison of differentially expressed genes and unchanged genes was done using R software (version 3.2).
Virtual screening of constituent compounds for estrogen receptor binding affinity
Virtual screening of Cipa ligands was performed against the estrogen receptor crystal structure (PDBID: 3OS8) using Autodock Vina, a more accurate and faster version of Autodock 4. The Cipa ligands available in the PubChem database were downloaded from there, the 3D structures of the rest of the ligands were drawn using Marvin Sketch, a computational tool for drawing 3 and 2 dimensional chemical structures. The structures were randomized and minimized prior to docking. A blind docking study was performed for each ligand wherein the possible search-space was the complete receptor molecule. For each drug molecules, the docking parameters were as follows: center_x: 9.43, center_y: 22.811, center_z : 23.418, size_x: 60, size_y: 60, size_z: 60, exhaustiveness 5000, num_modes 50,000, energy_range: 20. Furthermore, for each ligand, 50 such runs were performed and the conformation with the minimum binding energy from among all the stable conformations of the 50 runs (a total of 1000 conformational possibilities) was selected for the cluster analysis.
The analysis was performed using ADT, UCSF Chimera and LigPlot + .
Connectivity map analysis for identifying similarities with known perturbations
The connectivity map dataset currently contains 1,319,138 gene expression profiles, resulting in 473,647 signatures, generated against approximately 27,000 perturbations in 9 cell lines including MCF7, HEPG2, A549, A375, PC3, HCC515, HT29, HA1E and VCAP. The differentially expressed genes were ranked according to fold change and a signature of up and downregulated genes was generated. The gene signature containing genes which were valid i.e., having a valid HGNC symbol or Entrez ID and were also present in the LINCS gene space (represented in the L1000 data as landmark or well inferred) was used to query the connectivity map using clue.io touchstone database.
Cipa treatment in DENV infection
For these experiments, institutional Biosafety clearance was obtained at Translational Health and Science and Technology Institute, Faridabad, Haryana, India. MCF-7 were seeded at 100,000 cells per well in 24 well plate and were maintained for 24 h (37 degrees; 5% CO2). Virus challenge (at 10 MOI, see supplementary) was given for 1 h, here the media was supplemented with 2% FBS. After 1-h of viral adsorption, cells were washed with PBS and DMEM high glucose with 10% FBS with or without Cipa (50 μg and 100 μg) was added. DENV titers in the supernatant were determined by plaque assay at 24 h pi. DENV plaque assay was set up in BHK-21 cells (C-13), purchased from ATCC (Cat. no. CCL-10). 50,000 BHK-21 cells were plated per well in 24-well plate. Serial dilutions of virus were added in triplicates and allowed to adsorb for 1 h, followed by overlay with 0.5% carboxymethyl-cellulose (CMC; Sigma). After 72 h, cells were fixed in 3.7% formalin and plaques were visualized by staining with crystal violet.
siRNA-mediated knockdown of ESR1 in MCF7 cells
1 μM concentration of siRNA targeting ESR1 gene and non-targeting control (NTC) were mixed with Opti-MEM (Life Technologies) and 1 μl of Lipofectamine RNAiMax to a total volume of 100 μl in a 24-well plate. Cells were trypsinized and volume made up so as to contain 60,000 cells in 400 μl antibiotic-free medium. After 20 min incubation of the transfection complex, cell suspension was added into each well. Knockdown efficiency was determined by qRT-PCR at 48 h post transfection. Cells were infected with DENV-2 at 48 h post transfection. Cipa was added at a concentration of 50 µg after 1 h of viral adsorption. DENV-2 titers were measured by plaque assay at 24 h pi.
Statistical analysis
All data represents mean ± SEM; n = 3–7 each group; *p < 0.05, **p < 0.01, ***p < 0.001. p-value > 0.05 is considered non-significant (NS). Statistical significance of the differences was determined with a one/two-tailed Student’s t test as per appropriate. Any additional steps are mentioned in the text.
