Self-transfecting GMO-PMO chimera targeting Nanog enable gene silencing in vitro and suppresses tumor growth in 4T1 allografts in mouse

Abstract

Phosphorodiamidate morpholino oligonucleotide (PMO)-based antisense reagents cannot enter cells without the help of a delivery technique, which limits their clinical applications. To overcome this problem, self-transfecting guanidinium-linked morpholino (GMO)-PMO or PMO-GMO chimeras have been explored as antisense agents. GMO facilitates cellular internalization and participates in Watson-Crick base pairing. Targeting NANOG in MCF7 cells resulted in decline of the whole epithelial to mesenchymal transition (EMT) and stemness pathway, evident through its phenotypic manifestations, all of which were promulgated in combination with Taxol due to downregulation of MDR1 and ABCG2. GMO-PMO-mediated knockdown of no tail gene resulted in desired phenotypes in zebrafish even upon delivery after 16-cell stages. In BALB/c mice, 4T1 allografts were found to regress via intra-tumoral administration of NANOG GMO-PMO antisense oligonucleotides (ASOs), which was associated with occurrence of necrotic regions. GMO-PMO-mediated tumor regression restored histopathological damage in liver, kidney, and spleen caused by 4T1 mammary carcinoma. Serum parameters of systemic toxicity indicated that GMO-PMO chimeras are safe. To the best of our knowledge, self-transfecting antisense reagent is the first report since the discovery of guanidinium-linked DNA (DNG), which could be useful as a combination cancer therapy and, in principle, can render inhibition of any target gene without using any delivery vehicle.

Graphical abstract

Cell-permeable morpholino, by incorporation of guanidinium linkages within the PMO backbone, which eradicates its conjugation with separate delivery agents, can be used to silence any gene. Here we have shown the consequences of GMO-PMO chimera-mediated NANOG knockdown. In vivo studies indicated tumor regression along with no added toxicity in vital organs.

Introduction

Phosphorodiamidate morpholino oligonucleotides (PMOs; Figure 1A) are a unique class of antisense oligonucleotides (ASOs) because their ribose ring is replaced by a morpholine ring and the anionic phosphate backbone is replaced by a neutral phosphorodiamidate backbone.^1,2 PMOs are routinely used as gene-silencing reagents as they bind to mRNA and block translation by a steric blockade mechanism when designed to be complementary to the 5′ leader sequences or to the first 25 bases 3′ to the AUG translational start site.^3,4 Notably, four PMO-based drugs (eteplirsen golodirsen, viltolarsen, and casimersen)^2,5 have been approved by the US Food and Drug Administration (FDA) for Duchenne muscular dystrophy (DMD). These drugs demonstrate the safety and effectiveness of PMO-based antisense therapy and they are drivers of the future development of such therapies for the treatment of other diseases. However, similar to other oligonucleotides,⁶ the main bottleneck thwarting PMO’s clinical application is its intracellular delivery.⁷ At present guanidinium-rich cellular transporters are used for PMO delivery via the conjugation of PMO either with cell-penetrating peptides (CPPs) or cationic dendrimer to give PPMOs or vivo PMOs (Figure 1A).⁸ Unfortunately, these oligos are associated with some level of toxicity due to the presence of too many cationic groups. Also, the correlation between transfection efficiency and cytotoxicity of such transporters has been unsatisfactory.^9,10,11 CPPs are peptide-based transporters and sensitive to peptidases. To overcome the peptidase stability of conventional CPPs, modified versions have been reported to improve the bioavailability in PMO delivery.^12,13 Generally, eight to 12 cationic guanidinium groups are required for cell transfection, and hence they are prone to having non-specific interactions with negatively charged biomolecules, leading to toxicity.

Figure 1.

Chemical structures of GMO-PMO, MTT assay and dose dependent Gli1 gene silencing

(A) Chemical structures of (a) PMO, PPMO, and vivo PMO; (b) DNG; (c) GMO; (d) GMO-PMO; and (e) PMO-GMO chimeras. (B) Cell viability assay by MTT in Shh Light2 cells after treatment with different doses of GGP Scr MM. (C) Western blot analysis of Gli1 in ShhLight2 cells with different doses of multi-mutated (GGP Scr MM) and Gli1 antisense GMO-PMO (GGP ASO). (D) Comparison of protein repression potential of GGP Scr MM and GGP Scr SM; western blot analysis of the expression of Gli1- and Gli1-related proteins with GGP SCR SM (0.75 μM), GGP SCR MM (0.75 μM), GGP Scr ASO (0.75 μM), GP ASO (0.75 μM), and cyclopamine after 36 h of treatment. The densitometric analysis has been provided along with (E). (F) Western blot analysis of Gli1 at 0.75μM dose of GPG ASO and GGP ASO along with its densitometric analysis. Data presented as mean ± SEM. ∗p < 0.05.

To overcome the shortcoming associated with PMO delivery, we have previously developed different types of non-peptidic di-substituted internal guanidinium delivery moieties, which were found to work successfully in PMO delivery with as few as four guanidinium linkages. Also, it was observed to be superior to vivo PMO concerning both efficacy and toxicity.^14,15 In continuation of our research on PMO delivery, our goal is to modify PMO in such a way that it will have effective antisense properties and at the same time it would be cell-permeable without any help from transporter or transfection reagents. Guanidinium-rich molecules have been known to facilitate cell penetration. Bruice and his group reported guanidinium-linked DNA analogues (DNG) (Figures 1A and 1B) to be nuclease resistant with high binding affinity toward DNA^16,17; however, no report about their gene-silencing ability is available to date except its cell-penetrating property.¹⁸ While initially working on PMO, we introduced guanidinium linkages into the morpholino-T nucleoside backbone to obtain guanidinium-linked morpholino oligomer (GMO) (Figures 1A(c)).¹⁹ This initial penta-thymidine GMO, when conjugated with regular Gli1 PMO (25-mer), was found to efficiently inhibit Gli1 expression.¹⁹ By this encouraging result, we became interested in incorporating guanidinium linkages into the backbone of a morpholino ASO sequence either at the 5′ end or 3′ end to get 20- to 25-mer GMO-PMO or PMO-GMO chimera (Figures 1A(d,e) and S1). We hypothesized that, despite the replacement of phosphorodiamidate linkages by guanidinium linkages, the corresponding bases might participate in Watson-Crick base pairing in order to actuate their ASO activity. In this case, no further conjugation with any transporter is required with PMO.

Considering the importance of NANOG, which is highly expressed in multiple types of cancer and is responsible for cancer stem cell (CSC) maintenance in the tumor milieu, we aimed to knock it down against the backdrop of breast cancer. An edge in targeting cancer stemness factors such as NANOG, Sox2, Oct4, and Klf4 in anticancer research lies in the fact that these are exclusively expressed in either embryonic tissues or adult germ line stem cells.²⁰ These proteins are transcription factors that act in concert to maintain the pluripotency of a cell. Their over-expression has been associated with the tumorigenesis and metastasis of a variety of cancers, including breast cancer. Since the intricate functioning of these proteins aids in multiple hallmark events of cancer prognosis, including increased cell proliferation via gearing up of the cell-cycle machinery, metastasis through increased migration-invasion, oxidative stress tolerance, and lastly chemoresistance, we looked into the effect of NANOG knockdown by GMO-PMO on all the above-mentioned cancer attributes. Furthermore, we looked into the chemosensitivity of Taxol in breast cancer cells as a function of NANOG repression because NANOG expression influences chemoresistance in ovarian cancer.²¹ Since tumor drug resistance, especially the cellular drug efflux machinery, seems to be closely associated with many acquired or intrinsic characteristics of CSCs,²² we also evaluated its modulation after NANOG knockdown.

In the course of our study, as the ASO activity of GMO-PMO chimera became evident through protein expression profiling experiments and the said ASO showed promising results as anticancer agent, we were intrigued to know about their cellular import mechanism. Additionally, as ASOs need to be present in the cytoplasm to function, we looked into their cytoplasmic localization subsequent to cellular entry. For this, we conjugated BODIPY dye with the GMO-PMO chimeras, which enabled us to conduct flow cytometry and fluorescence imaging techniques.

Last, to assess whether our GMO-PMO chimeras have the potential to suppress gene expression in complex multicellular organisms, we checked the anti-tumor efficacy of NANOG GMO-PMO ASO in 4T1 allograft tumors of BALB/c mice. Here, it should be notified that, initially, we confirmed the gene-silencing activity of a GMO-PMO chimera in a simpler zebrafish model.

Hence in this manuscript, for the first time, we report the self-transfecting and gene-silencing potential of GMO-PMO chimeras. Additionally, we provide insight into how these modified ASOs can be developed into potential anticancer drugs.

Results

We synthesized the GMO-PMO and PMO-GMO chimeras using a solid-phase method by the incorporation of guanidinium linkages into PMO at the 5′ and 3′ end, respectively (Figures S1 and S2). The PMO part of the chimeras was synthesized following our published protocol.^23,24 Primarily, we wanted to determine which modification could function better, so we generated both GMO-PMO and PMO-GMO chimeras for initial testing. Also, to explore whether these GMO parts could participate in ASO activity or were merely responsible for cellular transfection, we generated scrambled GMO-PMO chimeras with a single mismatch (SM) and four mismatched GMO-PMO (MM) in the GMO part and checked their efficacy as gene-silencing agents. We have seen previously that regular Gli-PMO (25-mer) and pentameric GMO conjugate exhibited antisense effects against Gli1 in ShhL2 cells (derived from NIH3T3 mouse embryonic fibroblast cell line, stably transfected with Gli-dependent firefly luciferase reporter); hence, as a proof of principle, we made a sequence of GMO-PMO targeting Gli1 and tested in the same cell line. Next we proceeded to target NANOG in breast cancer cells and, accordingly, NANOG GMO-PMO chimera was synthesized. Similarly, GMO-PMO chimera targeting the no tail (ntl) gene of zebrafish was synthesized as PMOs are routinely used in the zebrafish model for comparative study. Thus, to serve all the above-mentioned purposes, we generated oligonucleotides (GMO-PMO/PMO-GMO/PMO) accordingly as listed in Table 1 (Figure S2; Table S1). Note that their respective acronyms used in this manuscript are provided in Table 1.

Table 1.

Sequences of GMO-PMO chimeras, PMO-GMO chimeras, and PMOs used in gene silencing study^a

Sl. no.	Chimeras	Acronyms	Sequence	Target gene and model
1	Gli1 GMO-PMO ASO	GGP ASO	5′OH-T∗T∗G∗G∗ATTGAACATGGCGTCT-3′	Gli1 in ShhL2 cells (derived from NIH3T3 mouse embryonic fibroblast)
2	Gli1 PMO-GMO ASO	GPG ASO	5′OH-TTGGATTGAACATGGC∗G∗T∗C∗T-3′
3	Gli1 GMO-PMO Scrambled-Single mismatch	GGP Scr SM	5′OH-T∗T∗G∗T∗ATTGAACATGGCGTCT-3′
4	Gli1 GMO-PMO Scrambled-Multiple mismatches	GGP Scr MM	5′OH-G∗G∗T∗T∗ATTGAACATGGCGTCT-3′
5	Gli1 PMO ASO#	GP ASO	5′OH-TTGGATTGAACATGGCGTCT-3′
6	NANOG GMO-PMO ASO	NGP ASO	5′OH-G∗T∗G∗A∗GTTGCCTGCATAATAACATGA-3′	Nanog in human cancer cell lines (e.g., MCF7)
7	NANOG GMO-PMO Scrambled	NGP Scr	5′OH-T∗G∗A∗G∗GTTCGCTGCATAATAACATGA-3′
8	NANOG PMO ASO#	NP ASO	5′-GTGAGTTGCCTGCATAATAACATGA-3′
9	Ntl GMO-PMO ASO	Ntl GP ASO	5′OH-G∗A∗C∗T∗TGAGGCAGACATATTTCCGAT-3′	Ntl in zebrafish
10	Mice NANOG GMO-PMO ASO	mNGP ASO	5′OH-A∗A∗G∗G∗AAGACCCACACTCATGTCAGT-3′	Nanog in mouse cell line: 4T1 and 4T1 derived allograft tumor in mice
11	Mice NANOG GMO-PMO Scrambled	mNGP Scr	5′OH-A∗A∗G∗G∗AACACCCACAGTCATGTCACT -3′
12	Mice NANOG PMO ASO	mNP ASO	5′OH-AAGGAAGACCCACACTCATGTCAGT-3′

^aBold letters indicate GMO with the asterisks “∗” representing guanidinium linkages. Underlined letters indicate the mismatched nucleotides. ^#NP ASO stands for naked PMO from Gene Tools, and “m” is used to denote mice.

The GMO part of the GMO-PMO chimera participates in Watson-Crick base pairing and subsequent ASO activity

Here it is critical to understand that we used scrambled sequences for initial cell viability assays as we wanted to assess the cytotoxicity as a function of the GMO-PMO chimera but not due to their ASO activities. MTT assay of GGP Scr MM (not having ASO activity, as evident from Figure 1B) revealed that GMO-PMO chimeras are non-toxic until a concentration of 70 μM is reached (Figure 1B). Thus a western blot analysis for comparing the efficacy of GGP ASO and GGP Scr (MM) was conducted at much lower doses (0.5 and 0.75 μM). The results indicated that GGP ASO could deplete Gli1 protein level by nearly 60% at 0.5 μM and by ≈ 80% at 0.75 μM. Here, as expected, GGP Scr MM did not show any protein repression activity (Figure 1C).

Next, we were intrigued to examine whether the GMO part participates in actuating the ASO activity or is just responsible for the cellular entry, so we synthesized the scrambled sequence with only one mismatch in the GMO part (GGP Scr SM) and compared its effect on Gli1 expression with that of GGP Scr MM; i.e., the scrambled sequence having four mismatches (MM) in the GMO part. We used cyclopamine, a known inhibitor of the sonic hedgehog (Shh) signaling pathway, to show Gli1 repression as a positive control (Figure 1D). The results revealed that, while GGP Scr MM had lost its gene-repressing ability, the scrambled oligomer with SM in the GMO part retained it as it exhibited nearly 40% Gli1 downregulation at the 0.75 μM dose, which was even more than that found in case of cyclopamine.

To further consolidate our claim of GGP ASO being an Shh pathway blocker via Gli1 repression and GGP Scr SM to be partly effective as the same, we evaluated their effect on the level of Gli1-related proteins, such as NANOG, CXCR4, c-Myc, Snai1, Sox2, and Twist^25,26 (Figure 1D). Their pattern of expression clearly reflected that of Gli1, indicating their expressional dependence on the latter. As Gli1 repression reduces the viability of a cell,²⁷ we found that Bcl2 also followed the same pattern of expression that Gli1 exhibited, and, as expected, Bax followed an exact opposite pattern of expression (Figure 1D).

Now, knowing that the GMO part of the GMO-PMO chimera participates in Watson-Crick base pairing to actuate the ASO activity, we became curious to know the role of the position of the GMO part in the ASO sequence. Subsequent western blotting revealed that GPG ASO (GMO-PMO) and GGP ASO (PMO-GMO) at an equal dose of 0.75 μM each had similar Gli1-repressing ability (Figure 1F).

NGP ASO effectively downregulated NANOG and NANOG-related proteins in MCF7 cells

After confirming that GMO-PMO scrambled with multiple mismatches can serve as negative experimental control, we now proceeded to target NANOG in breast cancer. Here too, just like in the case of GGP Scr MM, we conducted an initial cytotoxicity test with NGP Scr in MTT assay, which was found to be non-toxic until a 70 μM dose (Figure 2A). This was in stark contrast with the MTT assay in the case of NGP ASO exposure, which possessed significant cytotoxicity from 0.5 μM onward due to the antisense effect (Figure 2B).

Figure 2.

Efficacy of NANOG GMO-PMO ASO in downregulating NANOG and related proteins in MCF7

(A and B) (A) Cell viability assay of MCF7cells after treatment with different doses of NGP Scr and (B) with different doses of NGP ASO. (C) Protein expression of NANOG in MCF7 cells treated with 0.5, 1, and 2 μM NGP ASO compared with that in cells treated with 2 μM NANOG ASO IGT-PMO. (D–F) (D) Comparison among NGP Scr, NGP ASO, NP ASO, and Taxol concerning their effect on the expression of NANOG and NANOG-related proteins (E) in MCF7 along with their densitometric analysis (F). Data presented as mean ± SEM. ∗p < 0.05.

Next, we performed protein expression profiling studies with different doses (0.5, 1, and 2 μM) of NGP ASO. Here, considering our previous report about the efficacy of IGT-conjugated NANOG PMO ASO (2 μM),¹⁴ we compared it with NGP ASO. The results revealed that NGP ASO at all three doses showed better antisense activity than the IGT-conjugated NANOG PMO ASO (≈30% repression). NGP ASO showed nearly 52% and 75% repression at 0.5 μM and 1 or 2 μM dose, respectively (Figures 2C and S6B). This prompted us to use the lower dose, between 1 and 2 μM, for further studies.

Now we compared NGP ASO with NP ASO; i.e., the naked morpholino NANOG ASO purchased from Gene Tools (without any guanidinium linkages). Here too, we used NGP Scr as negative control. The results revealed that NGP ASO showed dose-dependent repression of NANOG with 0.5 μM causing ≈50% repression and 1 μM causing approximately 70% repression. In contrast, NP ASO (i.e., the naked morpholino-treated cells) showed very less downregulation of NANOG (≈20%). Taxol alone did not affect the protein level of NANOG at all (Figure 2D). This confirmed that its mechanism of action does not involve the said protein. The protein expression pattern of NANOG-related genes, viz. Sox2, pGSK3β, CXCR4, EpCAM, c-Myc, N-cadherin, vimentin, Snai1, Twist, and MMP9,^14,28,29 all reflected that of NANOG with dose-dependent downregulation in 0.5 and 1 μM NGP ASO-treated cells (Figures 2E and 2F). As expected, there was no effect in NGP Scr-treated cells, and NP ASO showed minimum activity. E-Cadherin intuitively exhibited the exact opposite expression pattern with dose-dependent upregulation due to NGP ASO exposure. Similar downregulation of NANOG and related proteins by NGP ASO was also confirmed in PC3 (prostate cancer; Figures S3 and S4) and MDA MB 231 (triple-negative breast cancer; Figure S5) cell lines.

In addition, a similar level of inhibitory effect on NANOG was seen after 36 and 72 h of incubation under both 0.5% and 10% serum concentrations (Figure S6A), which indicated that the presence of internal guanidinium linkages does not involve serum interaction. It is an interesting observation just like our previously reported IGT-PMO,¹⁴ whereas serum dependency was reported in the case of other guanidinium-based cellular transporters where guanidinium groups are attached flexibly.^30,31 We also confirmed that there were no off-target effects of NGP ASO, NGP Scr, and NP ASO, as these compounds elicited no effect on expression of acetylated α tubulin, α-tubulin, β-tubulin, and PDL1 in MCF7 cells (Figure S7) (vide infra for in vivo study).

Combination of NGP ASO and Taxol exerts a more profound anticancer effect in MCF7 cells than that caused by their individual treatment

NGP ASO-mediated reactive oxygen species generation, DNA damage, and apoptosis aggravates the anticancer impact of Taxol

For all the experiments regarding the composite effect of Taxol and NGP ASO, the total exposure duration was 72 h. MCF7 cells were treated with NGP Scr (1 μM)/NGP ASO (0.5 and 1 μM)/NP ASO (1 μM) for first 36 h and then Taxol was added. For the initial cell viability assay, i.e., to determine the effective exposure concentration of Taxol, these already treated cells (for the first 36 h) were subjected to Taxol treatment (at different concentrations; viz., 12.5, 25, 50, and 100 nM) for the next 12 and 36 h (total exposure duration 48 and 96 h, respectively) (Figure 3A). The results indicated that NGP Scr (red trend line; Figure 3A) and NP ASO (orange trend line; Figure 3A) did not affect the cytotoxicity of Taxol (blue trend line; Figure 3A), which individually caused dose-dependent depletion of MCF7 cell viability. Interestingly, NGP ASO (green and violet trend lines; Figure 3A) was found to aggravate the cytotoxic potential of Taxol at both time points in a dose-dependent manner.

Figure 3.

Aggravated chemosensitivity of MCF7 toward Taxol due to NANOG knockdown via NGP ASO in terms of ROS formation and cell-cycle arrest

(A) Cell viability assay of MCF7 cells treated with NGP Scr (1 μM), NGP ASO (0.5 and 1 μM), and NP ASO (1 μM) after they were subjected to 12.5, 25, 50, and 100 nM doses of Taxol for 48 and 96 h, respectively. (B and C) The histogram plots (B1 and C1) and respective bar diagrams (B3 and C2) depict potentiation interaction between NGP ASO and Taxol in increasing both cellular and mitochondrial ROS in MCF7. (D) Similar pattern was observed in the downregulation of Mn-SOD, evident through western blot analysis. (E) Representative distribution of MCF7 cells in different phases of the cell cycle as functions of different compound treatments. (F) Immunofluorescence analysis and subsequent confocal microscopy of Cyclin D1 (shown in green) in control and treated cells after 72 h. Nuclei were stained with DAPI (blue) (F1). Quantification of Cyclin D1 intensity per nucleus was calculated for 20–25 cells from confocal images (mean ± SEM). Data are presented as percentages relative to the non-treated MCF-7 cells (F2). (G) Protein expression of p-p53 as obtained from immunoblotting along with its densitometric analysis. Data presented as mean ± SEM. ∗p < 0.05.

Since Taxol exposure can cause oxidative stress-mediated apoptosis in cancer cells,³² we examined whether NANOG knockdown can aggravate this (Figure 3B). We observed that NGP ASO individually caused nearly 8-fold increase in cellular reactive oxygen species (ROS) generation with respect to untreated control (Figures 3B1 and 3B3), whereas it caused nearly twice of that in combination with Taxol. To check the integrity of our experiment, we subjected NGP ASO + Taxol exposed cells to N-acetylcysteine (NAC), a cysteine prodrug that replenishes intracellular reduced glutathione (GSH) levels and hence pacifies oxidative stress, and found that cellular ROS was curbed down to near-control level (Figures 3B2 and 3B4). In case of mitochondrial ROS, NGP ASO alone caused nearly 3.5-fold rise in ROS generation, which was a bit less than that caused by NGP ASO + Taxol (4-fold increase with respect to control) (Figures 3C1 and 3C2). This can be corroborated with the expression profile of mitochondrial superoxide dismutase (SOD) or Mn-SOD, which exhibited downregulation in both NGP ASO-treated (≈42% downregulation) and NGP ASO + Taxol-treated cells (≈70% downregulation) (Figure 3D).

The cell-cycle profile revealed through propidium iodide (PI) staining and subsequent flow cytometric analysis that NGP ASO alone increased the sub-G1 cell population, a phenomenon that was aggravated due to co-exposure of NGP ASO and Taxol. This indicated the occurrence of extensive cellular demise in both NGP ASO-treated and NGP ASO + Taxol-treated MCF7 cells (Figure 3E). The immunofluorescence of Cyclin D1 also indicated depletion of the said protein in both the above-mentioned treatment groups with a more profound effect in the latter (Figure 3F).

As the cell-cycle analysis indicated extensive cell death in both NGP ASO- and NGP ASO + Taxol-treated cells, we evaluated the externalization of phosphatidyl serine in the outer leaflet of the plasma membrane to confirm the onset of apoptosis in both the treatment groups. In Figure 4A, the lower-left quadrant depicts the viable cell population, which got depleted from 98.76% ± 4.5% of cells in control to 81.64% ± 4.2% in the case of only-NGP ASO-treated cells and 88.45% ± 2.5% in Taxol-treated cells. The lower right quadrant and the upper right quadrant, which represent the early apoptotic population and late apoptotic population, showed significant rise in cell population with 8.26% ± 1% and 6.9% ± 0.90% in NGP ASO-treated cells and 5.25% ± 0.71% and 4.2% ± 0367% in NGP ASO + Taxol-treated cells, respectively. Here the upper left quadrant represents the necrotic population that was found to be 3.2% ± 0.30% in the NGP ASO treatment group and 2.1% ± 0.50% in NGP ASO + Taxol group (Figure 4A). It is to be noted that Taxol alone was found to increase the percentage of cell death to approximately 5%, which was further aggravated in NGP ASO + Taxol cells exhibiting ≈11% cell death, hence indicating the sensitization of MCF7 cells to Taxol toxicity due to NANOG knockdown (Figure 4A).

Figure 4.

Increased apoptosis due to the combined effect of NANOG suppression and Taxol in MCF7

(A) Representative scattered dot-plot images of flow cytometric analysis show the percentages of MCF-7 cells undergoing necrosis and early and late phases of apoptosis. Here, the lower-left quadrant stands for viable cell population, the lower right quadrant represents the early apoptotic population, the upper right quadrant represents the late apoptotic population, and the upper left quadrant represents the necrotic population. (B) Effect of NAC in Taxol + NGP ASO-treated cells. (C) Change in mitochondrial membrane potential depicted as scattered dot plots after flow cytometric analysis. The upper quadrant shows the cell population with no loss of mitochondrial membrane potential and the lower quadrant shows the cells with loss of membrane potential. (D) Effect of NAC on membrane depolarization of Taxol + NGP ASO-treated cells. (E) Confocal images after BrdU incorporation assay where the nuclei were stained by PI and the incorporation of BrdU was detected by the Alexa 488 tagged (green) anti-BrdU antibody. (F) Western blot analysis of Bax and Bcl2 in Taxol-, NGP ASO-, NGP ASO + Taxol-, NP ASO-, and NP ASO + Taxol-treated MCF7 cells. (G) Bar diagram representing active caspase 9 levels in the above-mentioned treatment groups. Data presented as mean ± SEM. ∗p < 0.05.

Mitochondrial membrane depolarization, which is a hallmark phenomenon in apoptosis, detected through JC1 staining, revealed that the percentages of cells with altered membrane potential in only-Taxol-treated and only-NGP ASO-treated cells were 3.92% ± 0.61% and 3.92% ± 0.67%, respectively. This percentage was elevated to a level of 7.78% ± 1.67% in cells subjected to NGP ASO + Taxol, again indicating chemo-sensitization of MCF7 cells to Taxol due to NANOG knockdown (Figure 4C). In both apoptosis detection and membrane depolarization, NAC was found to mitigate the effect of NGP ASO + Taxol to a basal level, hence clearly indicating the occurrence of apoptosis due to ROS-induced damage (Figures 4B and 4D). The Bax/Bcl2 ratio serves as a marker of apoptosis, and the results of immunoblot clearly supported the findings of JC1 staining. Bax, which was found to be upregulated as a function of individual Taxol exposure by ≈ 1.6 fold, was found to have its expression induced by 2.6-fold and 3.3-fold in the case of NGP ASO- and NGP ASO + Taxol-treated cells, respectively (Figure 4F). Considering p-p53 as a transcriptional inducer of Bax, these results can be correlated as NGP ASO caused nearly twice the increment of p-p53 level, and NGP ASO + Taxol exposure resulted in its ≈3.6-fold upregulation (Figure 3G). BCl2, as expected, exhibited the exact opposite pattern.

Caspase 9 activity showed a marked increase in the only-Taxol group (by ≈1.5-fold), only-NGP ASO-treated cells (by ≈2.4-fold), and cells subjected to the combination of NGP ASO and Taxol (by ≈3.7-fold) (Figure 4G).

One of the factors that connect oxidative stress to apoptosis is extensive DNA damage. The same pattern that was evident in the case of apoptotic detection, 5,5′,6,6′-tetrachloro-1,1',3,3′-tetraethylbenzimidazolylcarbocyanine iodide (JC-1) staining and p-p53 and Bax immunoblotting, was also found to be reflected in the bromodeoxyuridine (BrdU) staining-imaging experiment (Figure 4E). Here, the TUNEL index plot again revealed more BrdU intensity in NGP ASO + Taxol-treated cells than in the cells that were treated individually with NGP ASO and Taxol.

Silencing of NANOG aggravates the effect of Taxol on stemness and the epithelial to mesenchymal transition pathway of MCF7 cells

Carcinoma cell migration is often associated with the stemness property of cancer. In the wound-healing assay, which is a regular and frequently used method for examining cell migration, ≈20% and ≈50% inhibition of wound closure was evident in the case of cells individually treated with Taxol and NGP ASO, respectively. Additive effect was evident as their combined treatment caused 70% inhibition of the said process (Figure 5A). This was also reflected in case of Transwell invasion assay, as the combination treatment resulted in least invasion of the cells to the lower chamber. Just like migration assay, Taxol and NGP ASO individually caused ≈20% and ≈60% inhibition of said process, whereas the composite exposure of both the treatments caused 80% inhibition of Transwell invasion (Figure 5B). However, the Taxol + NP ASO group exhibited nearly 30% and 25% inhibition in the case of migration and invasion, respectively. Thus, it denotes the incapability of NP ASO in increasing the chemotherapeutic efficacy of Taxol. The obliteration of MCF7 cell’s invasive property in NGP ASO and NGP ASO + Taxol-treated cells was confirmed through profiling the protein levels of EpCAM, Twist, vimentin, MMP9, N-cadherin, and E-cadherin (Figure 5C). EpCAM expression, which was found to be unaffected by both Taxol and NP ASO, was reduced significantly by a margin of ≈63% in NGP ASO-treated cells. In NGP ASO + Taxol- and NP ASO+ Taxol-treated cells, EpCAM level was downregulated by ≈ 82% and ≈15%, respectively. Twist expression was found to be approximately 42% and ≈60% when treated with NGP ASO alone and with Taxol, respectively. While NP ASO could not exert any effect on Twist expression, NP ASO with Taxol managed to reduce the protein level of Twist by only ≈11.02%. Just like EpCAM and Twist, the expression of vimentin was unaffected by Taxol and demonstrated nearly 72% repression in only-NGP ASO-treated cells. It was further downregulated to show ≈91% repression in NGP ASO + Taxol-treated cells. The expression of MMP9 in Taxol-, NGP ASO-, and NGP ASO + Taxol-treated cells exhibited the same pattern of expression that was evident in the case of EpCAM, perhaps suggesting a common regulator. N-cadherin, the protein responsible for metastatic behavior of cancer cells, was found to exhibit approximately half of the basal level expression due to NANOG knockdown. Taxol, in cells with NANOG repression, resulted in ≈69% downregulation of N-cadherin. However, there was mild repression of N-cadherin observed in only-Taxol-, only-NP ASO-, and NP ASO + Taxol-treated cells. It was evident that NP ASO did not affect Taxol’s action. E-cadherin downregulation is a marker of epithelial to mesenchymal transition (EMT). NGP ASO, both individually and in combination with Taxol, caused massive upregulation of its expression. Snail1 expression evaluated through immunofluorescent studies revealed a pattern that was similar to that shown by EpCAM, Twist, and N-cadherin, again denoting a common mode of regulation for all (Figures 5D and S8).

Figure 5.

NGP ASO and Taxol synergistically prevent EMT in MCF7 cells

(A) Wound-healing assay where the width of wound closure in MCF7 cells at 0 h was set to 100% (A1). (A2) Percentage of wound closure plotted graphically with respect to control. Wound closure was most inhibited in the NGP ASO + Taxol group. (B) Transwell migration assay in MCF7 cells after combined compound treatment at indicated doses (B1). The number of cells that invaded to the lower chamber through Matrigel in the control sample is taken as 100%. (B2) Bar diagram shows the percentage of Transwell migration in MCF-7 cells of treated groups with respect to untreated control. Scale bars, 60 μm. (C) Western blot analysis for expression profiling of EpCAM, Twist, vimentin, MMP9, N-cadherin, and E-cadherin (C1) with their densitometric data (C2). (D) Confocal immunofluorescence microscopic analysis of Snai1 (shown in green) in control and treated cells after 72 h. Nuclei were stained with DAPI (blue). Quantification of Snai1 intensity per nucleus was obtained from confocal immunofluorescence microscopy and was calculated for 20–25 cells. Data presented as mean ± SEM (n = 3). ∗p < 0.05.

Next, with the notion that NANOG promotes highly aggressive and anchorage-independent cell proliferation, we were intrigued to examine the effect of the combined treatment on the anchorage-dependent colony formation ability of MCF7 cells in monolayer culture. The clonogenic potential of NGP ASO + Taxol-treated cells decreased to a greater extent compared with Taxol-alone-treated cells (Figures 6A1 and 6A2). The inhibition of stemness was also indicated by the reduction in mammosphere formation as evident from the spheroid assay (Figure 6A3). There was more inhibition of both spheroid size (Figure 6A4) and number (Figure 6A5) in the case of NGP ASO + Taxol treatment than that observed in their individual treatment. In mammary carcinoma, the subpopulation identified with CD44+/CD24– marker makeup is considered to be CSCs.³³ This population was found to be 36% ± 3.1% and 35.4% ± 4.1% in untreated control and Taxol-treated MCF7 cells, respectively. However, NGP ASO reduced its population to 6.04% ± 1.2%. Taxol combined with NGP ASO further brought it down to 3.87% ± 0.9, indicating the promulgated effect of NGP ASO by Taxol in abrogating stemness (Figure 6B). To find out the impact of Taxol treatment on cancer pluripotency factors and other NANOG-related proteins in NANOG-silenced cells, the expression of NANOG, Sox2, c-Myc, p-GSK3β, and p-Akt was evaluated through the western blot technique (Figure 6D1), and protein levels of Oct4, p-STAT3, and focal adhesion kinase (FAK) (Figure 7) were determined through flow cytometry. In the case of NANOG expression, Taxol alone was found to be ineffective as expected. Considering the control’s protein level as 100%, NANOG expression level was found to be 24.8% ± 4.7%, 8.1% ± 3.9%, 87.37% ± 4.6%, and 70% ± 7.8% in only-NGP ASO-, NGP ASO + Taxol-, NP ASO-, and NP ASO + Taxol-treated cells, respectively. Exactly reflecting this pattern, Sox2 was found to show 26.5% ± 3.8%, 10.3% ± 2.1%, 92.8% ± 6.6%, and 89.62% ± 5.3% expression in only-NGP ASO-, NGP ASO + Taxol-, NP ASO-, and NP ASO + Taxol-treated cells, respectively. The same pattern was evident in c-Myc too, with 29.31% ± 4.1% expression in only-NGP ASO-treated cells, 12.8% ± 3.7% in the NGP ASO + Taxol group, 85.3% ± 6.4% in only-NP ASO cells, and 77.2% ± 8.1% in cells subjected to NP ASO + Taxol treatment (Figure 6D2). These data clearly suggested the downregulation of stemness factors is due to NGP ASO treatment. Thus, we now looked into the PI3K-Akt pathway, which has been found to be crucial in maintaining pluripotency of the stem cells.³⁴ While Taxol individually could not actuate any significant perturbation in p-Akt level, the said protein in its phosphorylated form was downregulated in both NGP ASO-treated cells and NGP ASO + Taxol-treated cells, with 76.1% ± 8.3% and 55.8% ± 4.3% expression, respectively. Again, NP ASO was found to be ineffective both individually and in combination with Taxol. Akt, after activation, is known to cause the inactivation of GSK3β. In our study, p-GSK3β was found to show the opposite pattern of that exhibited by p-Akt (Figure 6D). Corroborating results were obtained from expression profiling of Oct4, pSTAT3, and FAK through flow cytometry, which revealed more downregulation of all the three proteins in Taxol-treated NANOG-silenced cells than in only-Taxol-treated or only-NANOG-silenced cells (Figure 7A). The same expression patterns for Oct4 (Figures 7B1, 7B, and S9) and FAK (Figure 7C) were evident from immunofluorescence studies too. As the physical interaction of NANOG and Oct4 is a crucial requisite for the transcriptional induction program of several stemness and EMT genes, we checked for their intracellular localization along with their abundance. The results of immunofluorescence and subsequent confocal studies demonstrated maximum co-localization of NANOG and Oct4 in the control cells of MCF7 cells. This co-localization was absent in all NANOG-silenced cells irrespective of whether they were exposed to Taxol or not (Figures 7B and S10).

Figure 6.

Synergistic effect of NGP ASO and Taxol on the stemness property of MCF7 cells

(A) Clonogenic assay; the percentage of cells forming colonies for each compound treatment as indicated are compared, respectively, with that of parental MCF7 cells. Scale bars, 4 μm (A1). (A2) Bar diagram represents the number of colonies with respect to untreated control cells. Spheroid formation assay showing representative microscopic images (A3) of sphere formation for the indicated cell treatments along with bar diagrams depicting the size (A4) and number (A5) of spheroids, respectively. (B) Identification of a CD44+/CD CD24− subpopulation in MCF7 cells by flow cytometry in different experimental treatment groups. (C) Western blot analysis of Gli1 as a function of treatment with Taxol, NGP ASO, NGP ASO + Taxol, NP ASO, and NP ASO + Taxol in MCF7 cells with densitometric analysis. Immunoblot images of NANOG and other related proteins (D1) with their densitometric analysis (D2). Data presented as mean ± SEM. ∗p < 0.05.

Figure 7.

Synergistic effect of NGP ASO and Taxol on the stemness property of MCF7 cells

(A) Expression of Oct4, p-STAT3, and FAK detected through flow cytometry analysis in Taxol-, NGP ASO-, NGP ASO + Taxol-, NP ASO-, and NP ASO + Taxol-treated MCF7 cells depicted through histogram (A1) and bar diagrams (A2). (B) Immunofluorescence and confocal microscopic analysis of co-localization of NANOG (green fluorescence) and Oct4 (red fluorescence) in control and other experimental types of treated cells (B1). Nuclei were stained with DAPI (blue fluorescence). Bar diagrams representing relative fluorescence intensity of NANOG (B2) and Oct4 (B3). (C) Confocal images (C1) after immunofluorescence of FAK (red fluorescence) in control and all treatment groups after 36 h with relative fluorescence intensity of FAK. Quantification of NANOG, Oct4, and FAK intensity per nucleus was calculated from confocal images of 20–25 cells each. Data presented as mean ± SEM. ∗p < 0.05.

NGP ASO-mediated chemosensitivity to Taxol is mediated through downregulation of multidrug resistance

One of the major challenges in cancer treatment that is accountable for poor treatment outcomes and tumor relapse is the emergence of multidrug resistance. The molecular mechanisms involved in such chemoresistance are the over-expression of ATP binding cassette (ABC) transporters, mainly that of MDR1 and ABCG2, which is specifically over-expressed in breast CSCs.^35,36 Since NANOG has been reported to confer chemoresistance in cancer cells, this study was done to elucidate whether NGP ASO-mediated NANOG inhibition can enhance the drug sensitivity of cells to Taxol. The experiment was designed in the same pattern as designed for all the experiments regarding composite exposure of NGP ASO and Taxol. Briefly, in a total exposure duration of 72 h, cells were treated with either NGP Scr (1 μM) or NGP ASO (0.5 and 1 μM) or NP ASO (1 μM) at the start, and, after completion of 36 h, 100 nM Taxol was added to each treatment type and incubated for 36 h. We evaluated the expression of drug-resistance proteins MDR1 and ABCG2 through immunoblot. The results showed significant downregulation of MDR1 and ABCG2 in NANOG-silenced cells (Figure 8B). These results were further confirmed by immunocytochemistry, where a significant reduction in the expression of MDR1 was obtained (Figures 8A, S11, and S12). From the images, it can be concluded that, in untreated control cells and NP ASO-treated cells, MDR1 (red fluorescence signal) was found at the cell periphery, indicating its localization in the plasma membrane, which is typical for a normal unstressed cancer cell. In the case of Taxol-treated cells and NP ASO + Taxol-treated cells, MDR1 was located both at the plasma membrane and in the cytoplasm, indicating a response that upregulates MDR1 production in response to Taxol exposure. NGP ASO and NGP ASO + Taxol caused overall downregulation of the protein.

Figure 8.

Cause of NGP ASO induced increased chemosensitivity

(A) Confocal images after immunofluorescence analysis of MDR1 (shown in red) in control and treated cells after 72 h of combined treatment with antisense and Taxol. Nuclei were stained with DAPI (blue). Quantification of MDR1 intensity per nucleus obtained from confocal immunofluorescence microscopic images (A2) was calculated for 20–25 cells. (B) Western blot analysis reveal MDR1 and ABCG2 downregulation in NGP ASO-treated cells. Data presented as mean ± SEM. ∗p < 0.05.

Effect of NGP ASO on MCF10A cells, a “normal” counterpart of MCF7 cells

MCF10A is an immortalized, non-transformed epithelial mammary cell line that can be used as the normal counterpart of MCF7 cells. Even though Qu et al. have suggested it to be dubious to consider MCF10A cells as a normal mammary epithelial cell line concerning the expression of stemness factors Oct4 and Sox2, NANOG was not found to be expressed in the said cell line.³⁷ We therefore chose this cell line to assess any off-target effect of NGP ASO. MCF10A cells too were treated with the same doses of NGP Scr to assess whether there is any cytotoxic effect. From the cell viability assay, it was found that NGP Scr did not affect MCF10A cells up to 100 μM dose (Figure S13A). We have also not observed any cytotoxic effect due to ASO activity in the MTT assay when MCF10A cells were treated with 0.5, 1, and 2 μM NGP ASO (Figure S13B). However, NANOG is reported to be not expressed in MCF10A cells, and, probably due to this, NGP ASO could not actuate any stemness deregulation, as evident from the immunoblot results of other stemness factors; viz., Oct4, c-Myc, and Sox2 (Figure S13D). We also observed minimal expression of NANOG in MCF10A cells (Figures S13C and S13D).

NGP ASO chimeras undergo clathrin-mediated endocytosis in MCF7 cells to exhibit time and dose-dependent uptake

In all the previous experiments concerning comparison between NGP ASO and NGP Scr, while the former exhibited NANOG knockdown potential, the latter failed completely to do so. Its sequence is designed to have multiple mismatched base pairs (Table 1), and thus it is expected to show no ASO activity. However, the caveat that remained is what if the null gene-repressing ability of NGP Scr is an outcome of its poor cell permeability and not abrogated ASO activity? To address this, we conjugated Bodipy-based fluorophore to our antisense oligomers and checked their cellular uptake through flow cytometry. The results after 4 h of exposure clearly indicated that the uptake of NGP ASO and NGP Scr were similar at all three comparable doses; viz., 250, 500, and 1,000 nM (Figures 9A1 and 9A2). At these three exposure concentrations, NP ASO was found to show no uptake at all. Additionally, the only Bodipy fluorophore did not show detectable uptake until a concentration of 1,000 nM. Only the cells exposed to an exposure concentration of 10 μM Bodipy itself showed significant uptake. However, this was much less than that evident in the case of NGP ASO and NGP Scr (Figures 9A1 and 9A2). Also, the same study suggested a time-dependent (Figure 9C) and dose-dependent (Figures 9B1 and 9B2) increment in the uptake of NGP ASO. Briefly, the incorporation of NGP ASO-BODIPY was significantly increased with the increasing time of incubation, and, importantly, 100% cell transfection was observed after 1 h of incubation (Figures 9C1). To identify the minimum concentration of NGP ASO-BODIPY required to get 100% transfection, cells were treated with 50–500 nM NGP ASO-BODIPY for 4 h and 100% cell transfection was observed with 50 nM compared with control (Figures 9B1). Similar dose-dependent uptake of NGP ASO was observed in another breast cancer cell line, MDA MB 231, along with two prostate cancer cell lines, DU145 and PC3 (Figures S14–S19).

Figure 9.

Uptake efficiency and mode of cellular entry

(A) Uptake efficiency of NGP ASO, NGP Scr, NP ASO, and Bodipy. (B) NGP ASO shows dose-dependent increase in cellular entry to reach 100% saturation at 50 nM. (C) Time-dependent increase in uptake efficiency; 100% saturation after 1-h exposure duration at both 500 and 1,000 nM. (D) Flow cytometry results of NGP ASO-BODIPY uptake after use of chloroquine, chlorpromazine, amiloride, and genistein. (E) Western blot analysis of NANOG in MCF7 cells after treatment with NGP ASO/NPASO alone and in combination with chloroquine. (F) Immunoblot analysis of NANOG in MCF7 cells after treatment with NGP ASO/NPASO alone and in combination with amiloride, genistein, and chlorpromazine. Data presented as mean ± SEM. ∗p < 0.05.

Next, we became interested in NGP ASO’s mode of entry. Three major endocytosis pathways are known to operate in the MCF7 cells; viz., clathrin mediated, caveolin mediated, and macropinocytosis.^38,39 In order to determine the mode of cytosolic delivery of NGP ASO, MCF7 cells were treated with NGP ASO (for western blot studies) as well as with its BODIPY-conjugated form (for flow cytometry and confocal imaging) separately with or without chloroquine, chlorpromazine, genistein, and amiloride. In flow cytometry, after 4 h of exposure at 500 nM, chlorpromazine was found to inhibit the cellular uptake of NGP ASO (Figure 9D). Western blot studies also suggested the same through reticence of NGP ASO (1000 nM)-mediated NANOG repression by chlorpromazine (Figure 9F), hence indicating the involvement of the clathrin-mediated endocytosis pathway in the cellular uptake of antisense GMO-PMO chimera. Mild NANOG repression was found in MCF7 cells treated with NP ASO alone (13.05% ± 2.31%). Interestingly, co-treatment with chloroquine, an endosomal escape-enhancing agent, increased the effect of NP ASO in terms of NANOG suppression (44.21% ± 4.27%) (Figure 9E), indicating enhanced endosomal escape. A similar effect of chloroquine was also found in the case of NGP ASO, where chloroquine co-treatment increased the suppression of NANOG from 76% to more than 93% (Figure 9E). Our hypothesis of chloroquine-mediated enhanced endosomal escape of NGP ASO is further supported as chloroquine treatment did not increase the uptake (fluorescence intensity) of NGP ASO in MCF7 cells but conferred more ASO potential to it (Figure 9D). Henceforth, we suggest that chloroquine facilitated the endosomal escape, and, thus, the same concentration of NGP ASO could show better inhibition in the presence of chloroquine. It is worth mentioning that the dose of chloroquine, amiloride, genistein, and chlorpromazine that was applied to establish the mechanism did not show any cytotoxicity (Figures S20A and S20B) or inhibition of NANOG (Figure S3C) at the respective concentrations. Cumulatively, it was evident from these experiments that NGP ASO undergoes clathrin-mediated endocytosis. Now we looked into the sub-cellular localization of NGP ASO-BODIPY after cellular entry.

For this, the cells were separately counterstained with CellLight Early Endosomes RFP, BacMam 2.0 (Endotracker), which stains early endosomes, and LysoTracker red, which stains both late endosomes and lysosomes. Confocal microscopy revealed that, after 5 h of incubation in 10% serum, Bodipy-tagged GMO-PMO (NGP ASO) prominently co-localized with Endotracker, which therefore indicated their localization in early endosomes at the initial phase (Pearson’s co-localization coefficient = 0.735 ± 0.055) (Figure 10A). Endosomal escape is a rate-limiting step in the therapeutic efficacy of any drug that performs its activity in the cytoplasm. In order to know the endosomal escape of NGP ASO-BODIPY (0.5 μM), the images were also captured after 18 h of incubation. Although co-localization of NGP ASO was observed with Endotracker at 18 h, it was much less in comparison with that observed after 5 h of incubation (Pearson’s co-localization coefficient = 0.52 ± 0.041) (Figure 10B). While NGP ASO’s co-localization with lysosomal vesicles at this same time point was close to that with early endosome (Pearson’s co-localization coefficient = 0.0.60 ± 0.025) (Figure 10C), after 18 h the major distribution of fluorescence was found in the cytosol, which denoted NGP ASO’s escape from endosomal compartments within 18 h of treatment (Pearson’s co-localization coefficient = 0.0.38 ± 0.065) (Figure 10D). To confirm that this fluorescence signal in the cytosol was not due to uncoupling of fluorescent label (i.e., Bodipy), we ensured the stability of NGP ASO-BODIPY after 18 h of incubation in both MCF7 cell lysate and 10% serum-containing DMEM medium through MALDI-TOF mass analysis according to Amantana et al. (Figures S21 and S22)⁴⁰ where the mass of NGP ASO-BODIPY was intact. Additionally, we also looked into the intracellular localization of NGP ASO-BODIPY in other cell lines (viz., MDA MB 231, PC3, and DU145), which confirmed the flow cytometry results regarding uptake (Figures S15. S17, S19).

Figure 10.

Intracellular localization of NGP ASO

(A and B) Localization of NGP ASO concerning early endosome after (A) 5- and (B) 18-h incubation. (C and D) Localization of NGP ASO concerning late endosome/lysosome after (C) 5- and (D) 18-h incubation.

GMO-PMO chimeras show promising ASO activity in in vivo models

GMO-PMO mediates knockdown of ntl in zebrafish (Danio rerio)

Since PMOs are routinely used for gene silencing in zebrafish, we also chose them for our initial in vivo study.^41,42 The rationale behind targeting this gene is that ntl mutants are easily demarcated with reproducible ntl-dependent phenotypes.⁴¹ As per a previous report,¹⁵ 9 ng (0.25 mM stock, 3–4 nL) of morpholino was injected into zebrafish embryos at 16 cell (1.6 hours post fertilization), 32 cell (1.75 hpf), and at the 64-cell stage (30–33 embryos were injected in each set). In all cases, zebrafish ntl mutant phenotype was observed with a complete loss of vacuolated notochord cells, posterior structures, and U-shaped somites rather than normal V-shaped somites (Figure S23). Zebrafish ntl mutant phenotypes were compared with wild-type embryos and imaged at 30 hpf under the microscope. Regular PMO injected after the 32-cell stage did not show any phenotypes that indicated that self-penetrating Ntl GP ASO is capable of gene silencing even after four to eight cell stages. To visualize the distribution of Ntl GP ASO into zebrafish embryos, BODIPY-conjugated Ntl GP ASO was also synthesized and injected into zebrafish embryos just like Ntl GP ASO. Under a fluorescence microscope, BODIPY-conjugated Ntl GP ASO-injected embryos showed a widespread green fluorescence, which implies its ubiquitous distribution in the in vivo zebrafish model (Figure S23C).

Suppression of NANOG leads to regression of 4T1 allograft tumor in BALB/c mice

Allograft tumors were generated in mice following 14 ± 3 days of subcutaneous injection of 4T1 cells (Figure S24). Two intra-tumoral injections of mNGP ASO at 5 mg/kg body weight (BW) within a week led to reduction of tumor volume to nearly half of that seen in control. In contrast, tumors were found to grow with time in control, mNGP Scr-, and mNP ASO-treated mice (Figures 11A and 11B). After dissection, measuring the weight of tumor exhibited a similar pattern (Figure 11D). Before assessing the ASO efficacy of mNGP sequence as an ASO in mice, we checked it in vitro in 4T1 cultures, where the results revealed approximately 70% downregulation of NANOG at 1 μM treatment concentration (Figure 11E). In tumor homogenate, the protein level of KLF4 was unaltered by NGP ASO, NGP Scr, and NP ASO. NANOG and N-cadherin showed repression in the only-NGP ASO-treated group (Figure 11F). The histological analysis of NGP ASO-mediated regressed tumor revealed that there was presence of necrotic regions containing numerous cancer cells with pycnotic nucleus (Figure 11G).

Figure 11.

Effect of mNGP ASO in 4T1 mice allografts

(A) Representative images depicting the formation and growth of tumor in Balb/c mice; right-most panel contains representative images of intact tumors subsequent to dissection. (B) Graphical representation of changes in tumor volume with time in different groups. (C) Graph representing BW of mice in all treatment groups across experiment duration. (D) Bar diagram representing weight of tumors after dissection. (E) Western blot analysis of NANOG in mice 4T1 cell line following treatment with mNGP ASO at 1 and 2 μM treatment concentration. (F) Western blot analysis of KLF4, NANOG, and N-cadherin from 4T1 allograft tumor after 6 days of intratumoral treatment (5 mg/kg BW; twice in 6 days at interval of 3 days) with mNGP Scr, mNGP ASO, and mNP ASO. (G) Histological images of tumor from untreated mice and mice treated with mNGP Scr, mNGP ASO, and mNP ASO. The red dotted boxes represent the area that is demonstrated in magnification in the following row. Regions lined with dotted black line indicate areas with necrotic tissue. Red arrowheads in the lowest panel depict cells with pycnotic nuclei. Data presented as mean ± SEM. ∗p < 0.05.

The results so far adequately indicated that NGP ASO (mNGP ASO for mice) can be used as an anticancer agent against mammary carcinoma. Henceforth, we looked into the effect of its intra-tumoral injection on vital mouse organs, including liver, kidney, and spleen, through histopathological analysis. Initial calculation of organo-somatic indices (OSIs) (Figure S25E) revealed the occurrence of splenomegaly, a phenomenon that is associated with tumor in mice due to induction of leukemoid reaction by 4T1 murine mammary carcinoma.⁴³ Splenomegaly was found to be associated with the increase in white pulp area in spleen of tumor-bearing mice (Figures S25B and S25D), with restoration to some extent in case of mice with NGP ASO-mediated tumor regression. In liver, the frequency of hepatocytes with vacuoles, karyolysis, pycnotic nuclei, and regions with cellular degeneration was increased in liver of mice bearing untreated 4T1 allograft. Similar effects were evident in NGP ASO-treated mice too, but at much lower frequency (Figure S25A). It is notable that these changes in liver histoarchitecture have been associated with the occurrence of oxidative stress.⁴⁴ According to Hojo et al. there is massive oxidative stress in the hepatic tissue of 4T1-bearing mice.⁴⁵ We hypothesized that a similar rise in hepatocellular oxidative stress has led to such changes in the liver of mice with both untreated tumors and those treated with NGP ASO. The restoration of histoarchitectural damage in treated mice is an outcome of regressed tumor and therefore less oxidative stress. This result can be further correlated with our findings regarding serum parameters of hepatotoxicity as both SGPT (serum glutamate-pyruvate transaminase) and SGOT (serum glutamic oxaloacetic transaminase) were high in all mouse groups with tumors in comparison with those in healthy mice (Figure S26). Interestingly, the activity of both SGPT and SGOT were near basal level in mice treated with mNGP ASO. In renal tissue, the same phenomenon was reflected with structural anomalies such as the atrophy of glomerulus, reduced Bowman’s capsular space, and reduction of tubular luminal space with their distortion (Figure S27A). This is also supported by the results regarding serum levels of albumin, urea, and creatinine. Serum albumin levels were found to be depleted in all mice groups with 4T1 allografts. Again, the serum urea nitrogen/creatinine ratio was found to decrease in mice with tumors to an extent of nearly half of that exhibited by healthy mice. Interestingly, NGP ASO caused elevation of the said parameter compared with other treatment groups (Figure S26). Histological analysis of lungs and testes did not reveal any histoarchitectural change in any treatment group (Figures S27B and S27C). Cardiac tissue from the healthy mice demonstrated normal histoarchitecture. Examination of heart from mice with untreated tumor and mice treated with NGP Scr, NGP ASO, and NP ASO revealed noticeable histological alterations in the form of myofibrillar loss (Figure S27D). However, it was noticeable that, although NGP ASO did not cause restoration of tumor induced damages, it did not add to it. In contrast to this, restoration by NGP ASO-mediated tumor regression was evident from serum CK-MB activity that showed elevation in untreated control mice, mice treated with NGP Scr, and NP ASO but near-basal-level activity in mice with regressed tumor; i.e., those treated with NGP ASO (Figure S26).

Apart from the above-mentioned organs, we evaluated the histoarchitecture of different parts of the alimentary canal. Esophageal structure was intact in all treatment groups, with well-organized mucosal arrangement and inner epithelial lining (Figure S28A). The presence of allograft tumor seemed to induce histological damage in stomach, as all mice groups with tumor exhibited disorganized mucosal layer (Figure S28B). Similar tumor-induced damage (viz., disorganized villi arrangement in small intestine) was evident in control mice with untreated tumor and mice treated with NGP Scr, NGP ASO, and NP ASO. No restoration due to NGP ASO-mediated tumor regression was observed (Figure S29A). The structures of large intestine, caecum, proximal colon, and distal colon were intact in all treatment groups and were found to be similar to that of healthy mice (Figures S29B, S29C, S30A, and S30B).

The serum concentration of tumor necrosis factor alpha (TNF-α) and interferon gamma (IFN-γ) were monitored through ELISA. All mouse groups with tumor exhibited depleted serum TNF-α titer compared with healthy mice. No significant difference was found among control mice with untreated tumor and those treated with NGP Scr, NGP ASO, and NP ASO (Figure S31A). Serum IFN-γ titer was found to be reduced in control tumor mice and mice treated with NGP Scr in comparison with healthy mice. Interestingly, it was found to be elevated in NGP ASO-treated mice (Figure S31B).

Discussion

ASOs are rapidly emerging as a therapy for genetic diseases.^46,47 Among ASOs, PMOs are the most promising ones concerning sequence specificity and nuclease stability.² Additionally, their neutral phosphorodiamidate backbone prevents several off-target effects, unlike ASOs with phosphorothioate or phosphodiester backbones. Moreover, since 2016, 11 ASO-based drugs have been approved by the FDA, four of which are based on PMO.⁵ However, in order to improve the therapeutic efficacy, a delivery vehicle is essential to facilitate their intracellular entry.^7,46,47 In the course of our research journey on PMO chemistry, we reported here the modification of PMOs by incorporation of guanidinium linkages into the PMO backbone to get GMO-PMO or PMO-GMO chimera as self-transfecting reagents with efficient gene-silencing ability. Our self-transfecting GMO-PMO or PMO-GMO chimera concept is novel in that it can remove the need for any delivery vehicle. Moreover, unlike fluorescent-labeled vivo PMO/PPMO/IGT-PMO synthesis, its fluorophore conjugation at the 3′ end can be achieved easily.

Despite the presence of multiple reports in both in vitro and in vivo systems about PMO-mediated inhibition of specific genes in malignancy⁴⁸ and successfully conducted phase I study, PMO therapy has not been approved for cancer treatment. In the present study, we have achieved greater feasibility of easy intracellular delivery of PMO than previously reported methods, which perhaps was the rate-limiting step for bringing PMO therapy into cancer treatment. Since NANOG plays a crucial role in the stemness, proliferation, and chemoresistance of various types of cancers, we have knocked down NANOG with our self-penetrating GMO-PMO chimera. Prior to that, we confirmed that the GMO part of a GMO-PMO/PMO-GMO can participate in Watson-Crick base pairing to cause gene repression through repressing Gli1 in a Shhl2 murine cell line. We observed that an SM in the GMO part of the scrambled GMO-PMO resulted in ASO activity, although with low efficacy. Here, we have also shown that the position of GMO in GMO/PMO chimera has no influence on its inhibitory efficacy, as similar Gli1 inhibition was observed at comparable doses. In our study, repressing Gli1 led to the downregulation of NANOG and c-Myc, both of which are the regulators of cancer stemness.⁴⁹ CXCR4, a chemokine receptor in cancer cells and transcriptional target of Gli1²⁶ was also downregulated due to Gli1 knockdown. Our results also indicated significant downregulation of the EMT-associated molecular cascade along with spiking up of the Bax/Bcl2 ratio as an outcome of the Gli1 knockdown. It is important to mention here that GMO-PMO showed higher downregulation of Gli1 at a much lower dose (i.e., at 750 nM) than our previously reported IGT-PMO (5 μM) ¹⁵; hence, it is 6.6 times more effective than later. To the best of our knowledge, among the CPP-conjugate PMOs reported to date, Pip6a-PMO was the most efficient conjugate, found to work at nanomolar concentration.¹³ However, in the case of GMO/PMO chimeras, no such delivery vehicle is required. It would be interesting to use these chimeras for exon-skipping studies such as PPMO¹³ and vivo PMO.⁵⁰ Although backbone-modified PMO and their chimeras,^51,52 PMO-DNA chimera,⁵³ or chimeric oligonucleotides containing morpholino thymidine analogues^54,55 or with a single guanidinium morpholino unit incorporated DNA⁵⁶ have been reported earlier, to the best of our knowledge, GMO-PMO/PMO-GMO chimera is the first report from our group.

Next, considering stemness as a notorious reason behind progression, metastasis, and recurrence of malignancy, we embarked on our main work regarding NANOG suppression in breast cancer. Our initial evaluation regarding the efficacy of GMO-PMO targeting NANOG in MCF7 cells using NANOG GMO-PMO (NGP ASO) against the scrambled GMO-PMO (NGP Scr) and naked PMO (NP ASO) (i.e., unmodified PMO, purchased from Gene Tools) yielded promising results as we got the expected dose-dependent repression of NANOG by NGP ASO. The modulation in expression of NANOG-associated genes further consolidated these results. c-Myc acts in concert with other pluripotency factors to endow self-renewal capability to CSCs. In the present study, NANOG knockdown in MCF7 cells by NGP ASO resulted in downregulation of c-Myc. A similar result was reported by Han et al. in the case of RNA interference-mediated NANOG silencing.⁵⁷ Sox-2, another important pluripotency factor, was also evidently downregulated due to NANOG suppression, which can be explained by the Sox2 promoter being known to bind with NANOG, which in turn results in Sox2 induction.⁵⁸ The downregulation of CXCR4 in NGP ASO-treated MCF7 cells can be attributed to the fact that the CXCR4 promoter also contains NANOG regulatory elements.²⁹ In cancer cells including MCF7, NANOG actuates the EMT through modulation of the EMT markers. The expression of Snai1, a zinc-finger transcription factor that attenuates E-cadherin directly through binding to E-boxes present on the E-cadherin promoter⁵⁹ is highly correlated with NANOG.⁶⁰ Another similar transcription factor, Twist, also prompts the repression of E-cadherin and upregulation of N-cadherin.⁵⁹ Twist, in turn, is reported as being induced by NANOG.³⁶ The above-discussed crosstalk was well reflected in our study as NANOG knockdown led to the downregulation of Snai1, Twist, and N-cadherin along with upregulation of E-cadherin. Snai1 is also reported to induce the transcription of MMP9, whose protein product is considered crucial for progression of breast carcinogenesis.⁶¹ Thus, intuitively, the protein level of MMP9 was found to be reduced in NANOG knockdown MCF7 cells. The expression of vimentin was also found to be reduced in cells treated with NGP ASO, similar to a report on repression of vimentin by Liu et al.⁶² in ovarian cancer cell lines as a function of NANOG inhibition by RNA interference.

Since neoplastic stemness is intricately responsible for the most dreaded aspect of malignancy (i.e., its resistance to multiple drugs) and breast CSCs (BCSCs) have been reported to evade Taxol therapy,²¹ this prompts the requirement of combination therapy against breast cancer. Accordingly, we evaluated the effect of NANOG knockdown on Taxol’s action in MCF7 cells. For this, we looked into the hallmark aspects of cancer related to either Taxol’s mode of action or NANOG’s regulatory network.

NANOG is known to suppress the basal level of ROS generation through the inhibition of oxidative phosphorylation in order to sustain stemness and drug resistance. In the present research, the repression of NANOG by NGP ASO increased both mitochondrial and cellular ROS generation. The cause was found to be that NGP ASO affected Mn-SOD downregulation, as it has been reported to be transcriptionally dependent on NANOG and Oct4.⁶³ The effect on cellular/mitochondrial ROS level and Mn-SOD repression was far more in the case of composite exposure to NGP ASO and Taxol. According to Meshkini and Yazdanparast, Taxol induces oxidative stress through glutathione (GSH) depletion.³² We hypothesize that NANOG knockdown caused Mn-SOD downregulation and resulted in an increment of the total intracellular ROS level to exhibit synergism that was actually formed as an additive function of Taxol and NGP ASO exposure in the case of composite treatment. Cell-cycle profiling through PI staining revealed the spiking up of the sub-G1 cell population due to NANOG knockdown. This event was promulgated by the composite application of Taxol in NANOG-suppressed cells. The sub-G1 cell population signifies apoptotic cells as they have reduced DNA-associated fluorescence due to typical diminished DNA content and morphological changes. Apoptosis was further confirmed by dual PI-Annexin V staining, which detects the presence of phosphatidyl serine in the outer leaflet of plasma membrane, and JC1, which demarcates cells with perturbed mitochondrial membrane potential. We observed that application of NAC dramatically curbed the percentage of apoptotic cells in Taxol-treated NANOG-silenced cells, hence indicating that composite exposure of NGP ASO and Taxol led to ROS-mediated damage, which in turn resulted in apoptosis. Immunofluorescence with anti-BrdUTP antibody revealed significant damage due to NANOG knockdown and even more BrdU intensity in NGP ASO + Taxol-treated cells, hence corroborating the results of our previously discussed apoptosis detection assays. Oxidative stress-mediated apoptosis is generally actuated through the intrinsic apoptotic machinery. This was also evident in our study as NANOG inhibition led to a higher p53 level and its pattern of expression exactly reflected the extent of apoptosis in different treatment groups. Here, it should be noted that, though p-p53 is reported to inhibit the Gli1-mediated upregulation of NANOG,²⁵ no report suggests any feedback phenomenon from NANOG to p-p53. We hypothesize that, in our study, p-p53 upregulation is solely a mark of ROS-mediated DNA damage and apoptosis. This was further confirmed as the expression patterns of Bax and Bcl2 were found to be modulated according to the expression of p-p53 in different treatment groups. Last, the activity of caspase 9, which was more in the case of NGP ASO + Taxol-treated cells than NGP ASO-treated cells, can be correlated with all our findings regarding cellular demise due to NANOG knockdown. Importantly, the whole above-mentioned regulation involving DNA damage, p53, Bax, Bcl2, and caspase activity was found to remain unaffected with NP ASO (i.e., naked morpholino).

The activation of the EMT pathway that in turn is promoted by CSCs has been reported to play an important role in neoplastic metastasis. Since taxol is generally known to impart no effect on stemness-EMT machinery, we probed whether Taxol exposure along with NANOG silencing can elicit more anticancer effects. The effect of only-NGP ASO on the expression of Ep-CAM, Twist, vimentin, MMP9, N-cadherin, and E-cadherin was further amplified when used in conjunction with Taxol. This can be corroborated with the findings of Ding et al.⁶⁴ about more Taxol chemosensitivity in terms of EMT due to NANOG disruption in HeLa cells. Since, there is strong involvement of the EMT pathway in acquired Taxol resistance of human breast cancers, our report about inhibition of EMT by Taxol due to NANOG knockdown in breast cancer is a significant addition to anti-mammary carcinoma research.

Clonogenic assay is considered a measure of cancer stemness, signifying how a single CSC can generate colonies. In our study, NGP ASO showed significant inhibition of colony formation, which was further diminished in the case of NGP ASO + Taxol-treated cells. The same pattern was evident in the inhibition of spheroid size and the number of spheroids in the case of spheroid formation assay, which also serves to assess cancer stemness. This pattern in reduction of stemness was also evident in the case of depletion of the breast CSCs population (i.e., CD44+/CD24− MCF7 cells) due to NANOG knockdown.³³

Although Gli1is an upstream regulator of NANOG, NANOG repression in MCF7 cells interestingly caused downregulation of Gli1. We have reported a similar phenomenon in our previous report about IGT-mediated delivery of NANOG PMO.¹⁴ According to Zbinden et al., there exists a positive feedback loop that enables NANOG activity to induce Gli1 expression.²⁵ However, reports about Gli1 regulation by NANOG are scarce and will require elaborate research. Another upstream regulator of NANOG is GSK3β, which has a controversial role in cancer progression and tumorigenesis. It is known to promote the growth and development of cancers such as human colorectal cancer while also functioning as a tumor suppressor for certain types of tumors, including mammary carcinomas. In the present study, it exhibited upregulation following NANOG knockdown, suggesting negative feedback from NANOG to GSK3β. We hypothesize that NANOG knockdown might have caused the repression of Oct4,⁵⁸ which in turn is a transcription factor of the Akt1 gene.⁶⁵ The downregulation of Akt, as also evident in our study, leads to the withdrawal of its inhibitory effect on GSK3β, thus leading to the accretion of its phosphorylated form, hence signifying its role as a tumor suppressor in breast cancer.⁶⁶ The activity of NANOG involves STAT3 as a downstream effector.⁶² The activation of STAT3 to p-STAT3 is vital for normal cells, but the activating signal is stringently regulated. Conversely, the anomalous expression of downstream genes (such as Snail and Twist) due to out-of-control STAT3 signaling endorses cell proliferation while thwarting apoptosis.⁶⁷ NANOG has also been found to bind to the promoter of FAK to upregulate its expression.⁶⁸ Thus, being the beneficiaries of NANOG regulation, the downregulation of both STAT3 and FAK is intuitive and expected in NANOG knockdown cells. Since Meshkini and Yazdanparast reported that Taxol induces oxidative stress in cancer cells,³² which leads to downregulation of stemness via repression of NANOG/Oct4,⁶⁹ we hypothesize that Taxol-induced oxidative stress might have worked additively with the action of NGP ASO in MCF7 cells subjected to the composite treatment of both NGP ASO and Taxol.

Unfortunately, most chemotherapeutic drugs have limited efficacy due to drug resistance, a phenomenon for which plasma membrane-embedded drug efflux ABC transporters are often held responsible. We report in our study that silencing of NANOG leads to the decrease in the expression of MDR1 and ABCG2 proteins, hence increasing the chemosensitivity of MCF7 cells toward Taxol. Similar results were also reported by Zhou et al. in liver cancer cells subjected to siRNA-mediated NANOG knockdown.⁷⁰

After confirming the anticancer effect of NGP ASO, we demonstrated that the cellular uptake of GMO-PMO chimera follows the endocytosis pathway, particularly through the clathrin-mediated uptake. This was established through multiple pieces of evidence in this study. First, the inhibition of NGP ASO’s uptake as well as its ASO activity by chlorpromazine, an inhibitor of clathrin-mediated endocytosis, confirmed the pathway. Additionally, we confirmed that chlorpromazine or any other inhibitor used in this study does not have any effect on NANOG expression (Figure S17). Second, gene silencing efficacy was improved in the presence of chloroquine, a drug used for endosomal escape study for PPMO.⁷¹ Third, the co-localization analysis revealed that endosomal escape of NGP ASO-Bodipy was occurring before 18 h of incubation.

The above-mentioned encouraging results prompted us to venture into testing GMO-PMO chimera in multicellular in vivo models. Our preliminary study in zebrafish embryos targeting Ntl gene showed increment in the frequency of ntl-dependent phenotypes. It is worth mentioning that GMO-PMO works even after the eight-cell stage, which was not possible to achieve earlier using regular PMO except using photocaged PMO,^72,73,74 or perhaps our IGT-PMO¹⁵ could be useful for regeneration biology in the zebrafish model.

Finally, we assessed the anti-tumoral effect of NANOG knockdown in 4T1 allograft tumors of BALB/C mice. For this we designed an ASO sequence targeted against murine NANOG (mNGP ASO). Initially, we checked its efficacy in 4T1 cell lines that indicated ≈70% NANOG knockdown in vitro. Intra-tumoral injections of mNGP ASO twice a week led to significant tumor regression, whereas similar administration of mNGP Scr and mNP ASO did not elicit any effect. After the treatment duration, we evaluated the level of NANOG; its upstream regulator, KLF4; and a downstream target, N-cadherin. KLF4 is identified as an upstream orchestrator of an intricate, feedforward loop of transcription factors including Oct4, Sox2, c-Myc, and NANOG.⁷⁵ The results clearly indicated that the KLF4-NANOG-EMT axis was downregulated from NANOG onward, hence serving as a proof of our principle. Histological analysis of mNGP ASO-treated tumor in comparison with control tumor indicated the incidence of large necrotic areas in the former. Zooming in into these areas revealed the occurrence of cancer cells with pycnotic nucleus, hence indicating cellular demise. These results can be corroborated with similar findings by Chen et al. in 4T1 mammary carcinoma regressed due to anti-tumoral effect of 2-dodecyl-6-methoxycyclohexa-2, 5-diene-1, 4-dione.⁷⁶ No such changes were evident in NGP Scr or NP ASO-treated mice.

Additional to these evaluations, we checked for any systemic toxicity that can occur due to the intra-tumoral administration of GMO-PMO. Histopathological analysis of vital organs including liver, kidney, heart, and spleen from mNGP ASO-treated mice, untreated mice with tumor, and healthy mice indicated that the extent of histoarchitectural damage in all these tissues due to mice bearing 4T1 allograft tumor was ameliorated as mNGP ASO caused tumor regression. Histopathology of lungs, testes, and different parts of the alimentary canal indicated that none of the treatment molecules added to any additional toxicity. Next, we also monitored the serum concentrations of TNF-α and IFN-γ, which are two of the important pro-inflammatory cytokines present in the tumor microenvironment of breast cancer. However, several reports suggest their dubious roles in maintaining tumor burden. TNF-α, at its discovery, was found to destroy tumor vasculature and promote apoptosis upon exogenous administration.⁷⁷ In our study, all the mouse groups with tumor showed depleted serum TNF-α titer compared with healthy mice, showing that its low level promotes tumor formation. However, it was found that all the treatment groups had similar serum TNF-α titer to that of control mice with untreated tumor. This indicates that none of the treatment molecules could elicit an inflammatory response in terms of TNF-α. We also monitored IFN-γ, a member of the type II interferon family that is known to play multiple roles such as effecting antiviral, anti-tumor, and immunomodulatory effects.⁷⁸ In our experiment, control mice with untreated tumor and those treated with NGP Scr had depleted serum IFN-γ denoting its tumor-resisting potential. Corroborating this, the concerned cytokine was found to be elevated in mouse group treated with NGP ASO, indicating alleviation of the pathological condition associated with tumor regression. Considering the simplicity, the possibility of rational design, relatively low expense, and easy synthesis, GMO-PMO self-transfecting chimera can be developed as useful tools to target any gene (for instance, those in cancer) toward the development of antisense therapy. Additionally, their high level of antisense efficacy at a low dose (500–750 nM) could overcome the long-standing problems associated with off-target effects of antisense reagents.⁴⁷ Here, we have shown how targeting pro-cancerous factors such as NANOG through self-transfecting GMO-PMO chimeras can cause efficient inhibition of malignancy in both in vitro and in vivo models.

Materials and methods

Chemicals and reagents

MCF-7, PC3, MCF10A, and MDA MB-231 cells were purchased from NCCS, Pune, India. The antibodies used in immunoblot and immunofluorescence study were procured from Cell Signaling Technology (CST) and Sigma-Aldrich. DMEM, fetal bovine serum (FBS), and antibiotic cocktail were purchased from Gibco. The Caspase-Glo 9 Assay Kit was purchased from Promega. The APO-BrdU TUNEL Assay Kit was obtained from Thermo Fisher Scientific. All the oligos used were >95% pure in high-pressure liquid chromatography (HPLC) (provided in supplemental information along with their MALDI-TOF spectra; Figures S21, S22, and S32–S53).

Cell culture and treatment

Shh-Light2 cells (derived from mouse embryo NIH3T3 cell lines stably transfected with a Gli-dependent firefly luciferase and constitutive Renilla luciferase reporters) were obtained as a gift from Professor J. K. Chen, Stanford University. 4T1 murine breast cancer cells (ATCC CRL-2539) were obtained as a gift from Professor S. S. Jana, IACS, Kolkata. The culture of Shh-Light 2, MCF7, PC-3, and MDA MB 231 cells was performed in 10% bovine calf serum (Gibco)-DMEM supplemented with 100 μg/mL streptomycin and 100 units mL⁻¹ penicillin at 37°C in a 5% CO₂ humidified incubator. The 4T1 cells were maintained in 10% bovine calf serum (Gibco)-RPMI in the same conditions. The Shh-N cells were cultured to 80% confluency in DMEM supplemented with 10% FBS, 100 μg mL⁻¹ streptomycin, and 100 units mL⁻¹ penicillin. The media was replaced with 2% FBS-DMEM containing 100 μg mL⁻¹ streptomycin and 100 units mL⁻¹ penicillin, cultured for another 24 h, and the medium was collected, centrifuged, and filtered through 0.22-μm filter. The Shh-N conditioned medium was stored at −20°C and used to activate Shh pathway in Shh-Light 2 cells.¹⁹

Cell viability assays

For MTT assay, the cells were seeded in 96-well plates at a density of 1 × 10⁴ cells/well for 24 h before treatment with different concentrations of oligonucleotides. After 72 h of treatment, 100 μL of MTT solution was added to each well. After 4 h, the medium containing MTT was removed and the purple formazan product was dissolved in 100μL/well of DMSO and quantified by plate reader at 570 nm.¹⁴

For chemosensitivity assay regarding combinatorial treatment of Taxol and antisense oligomers, MCF7 cells were seeded in 96-well plates at a density of 1.0 × 10⁴ cells/well for 24 h. Before Taxol addition at concentrations of 12.5, 25, 50, and 100 nM, cells were separately treated with NGP ASO, NGP Scr, and NP ASO for 36 h. After addition of different concentrations of Taxol, the cell viability was evaluated by MTT assay at different time points (48 and 96 h, respectively).¹⁴

Western blotting analysis

Protein lysates from compound treated cells were resolved by 8%–10% SDS polyacrylamide gel electrophoresis, subsequently transferred onto nitrocellulose membranes, and treated with primary antibodies against the proteins. The membranes were then subsequently probed with a horseradish peroxidase-conjugated secondary antibody, developed with Femto, and visualized in Bio-Rad Chemidoc.¹⁴

Immunocytochemistry

Cells were grown on coverslips (pre-treated with poly-L-lysine), fixed with 4% paraformaldehyde for 10 min, permeabilized with 0.5% Triton X-100 for 5 min, then blocked in 1% BSA in PBS for 1 h and incubated with primary antibodies overnight at 40°C. Cells were then stained with Alexa 488 anti-rabbit immunoglobulin G (IgG) or Alexa 594 anti-mouse for 1 h at room temperature and washed twice in PBS; nuclei were stained with DAPI and were then mounted on slides with Slow Fade Gold anti-fade reagent. The cells were visualized under Carl Zeiss confocal microscope (Holdtecs Technology, Chengdu, China).¹⁴

Counterstaining with Endotracker and LysoTracker

To check the sub-cellular localization of NGP ASO-BODIPY, NGP Scr-BODIPY, and NP ASO-BODIPY, the compounds were added at 0.5 μM concentration to MCF7 cells and incubated for 5 and 18 h. CellLight Reagents ∗BacMam 2.0∗ Endotracker was added 15 h before visualization according to manufacturer’s protocol for localization analysis concerning early endosomes. For localization analysis concerning late endosomes/lysosomes, LysoTracker red was added 1 h before visualization. The cells were observed and imaged under Carl Zeiss confocal microscope.¹⁴

Flow cytometry

For measuring the uptake of NGP ASO-BODIPY, NGP Scr-BODIPY, NP ASO-BODIPY, and only BODIPY, the compounds were added at respective concentrations 1, 4, or 8 h before flow cytometry.

To measure intracellular ROS production, the treated cells were trypsinized and washed with medium containing serum followed by a wash with PBS. DCFH-DA was added to each experimental group at the required concentration. The cells were incubated for 5 min and then analyzed by flow cytometry (520-nm emission).

JC-1 stain was used to measure mitochondrial membrane potential (Ψm). Briefly, cells were incubated in medium containing JC-1 (2.5 μg/mL) for 30 min at 37°C, suspended in PBS, and subjected to flow cytometry (530-nm emission).

For assessment of phosphatidyl serine externalization in the outer leaflet of the membrane, cells after 4 h of compound treatment were treated with Annexin-APC/PI for 15 min at room temperature.

To check the protein expression, the cells were fixed in 4% paraformaldehyde in PBS for 20 min at room temperature and permeabilized in 0.1% Triton X-100 in PBS with 0.1% FBS for 5 min. After washing twice in PBS with 3% FBS, the permeabilized cells were incubated with primary antibody on ice for 2 h and washed twice in PBS. The cells were then incubated with Alexa 488/594-conjugated goat anti-rabbit/rabbit anti-mouse IgG as a secondary antibody for 30 min on ice and washed twice in PBS. The stained cells were acquired and analyzed against isotype in flow cytometry.

For cell-cycle analysis, the cells were fixed in 70% ethanol at 4°C and left overnight. After treatments with 2 mg mL⁻¹ RNaseA, cells were stained with 50 mg mL⁻¹ PI containing 0.1% Triton X-100 and EDTA (0.02 mg mL⁻¹).

For each case, the fluorescence of 10,000 cells was recorded using a flow cytometry (BD FACS ARIA III). The data were analyzed using FlowJo 8.0 software.

Microinjections into zebrafish embryos and imaging

Embryos used in these studies were obtained by natural mating of adult zebrafish (wild type) and cultured in E3 embryo medium at 28.5°C according to standard procedures. Microinjection of zebrafish was done using the Eppendorf Femtojet express microinjection setup. Compounds were microinjected in the embryos at the cell-yolk interface in different stages. Nine nanograms (0.25 mM stock, 3–4 nL) of morpholino was injected into zebrafish embryos at the 16-cell stage (1.6 hpf), 32-cell stage (1.75 hpf), and 64-cell stage (30–33 embryos were injected in each set). H₂O-injected embryos were considered as control. Injected embryos were then transferred to Petri dishes and cultured at 28.5°C. Fluorescence intensities and distribution in the larvae were observed under an inverted fluorescence microscope (Olympus IX51).¹⁵

Colony formation assay

Briefly, cells were grown in six-well plates, treated with compounds, and kept at 37°C in an incubator for 15 days with intermittent medium changes. After the stipulated time, cells were then stained with 0.005% crystal violet for 1 h and images were taken under a microscope.¹⁴

Wound-healing assay

Cells were cultured to completely confluent state in 12-well plates and a scratch was made across the cell monolayer of each sample well with a 10-μL pipette tip. Then the cell monolayer was washed with PBS and incubated with compounds in 0.5% FBS DMEM at 37°C with 5% CO₂ for 36 h. The widths of the scratches were measured and the percentages of relative wound closure were compared at 0 and 36 h, respectively.¹⁴

Caspase 9-Glo assay and TUNEL assay

Apoptosis-associated DNA damage and caspase 9 activity was measured by Caspase-Glo 9 Assay Kit (Promega) and APO-BrdU TUNEL Assay Kit (Thermo Fisher), respectively according to manufacturer’s protocol.

Tumor growth and in vivo studies

All mice were maintained according to the guidelines of the Institutional Animal Ethics Committee. The 4T1 cells (>85% viability) were harvested from cultures using trypsin-EDTA (Invitrogen), and 1 × 10⁶ cells were diluted in PBS and injected subcutaneously in the rear of male BALB/c mice aged 5–6 weeks. The tumors grew and became palpable after 14 ± 3 days (70% incidence rate). Tumor growth was calculated morphometrically using calipers, and tumor volumes were calculated according to the formula V (mm³) = length × width²/2.⁴³ mNGP ASO/mNGP Scr/mNP ASO/PBS (in the case of the control) was administered through intra-tumoral injection (volume not exceeding 50 μL) twice in 6 days at a 3-day interval (experimental design provided in Figure S24). After 6 days, mice were euthanized, sacrificed, and dissected for the collection of tumors along with other organs, including liver, kidney, spleen, and heart. All the procedures during rearing and experimentation with mice were conducted by stringently following the rules of Animal Ethical Committee, IACS. OSI was calculated from organ weights as per Bhowmik et al.⁴⁴ For histological slide preparation, tissues were fixed in 10% neutral buffered formalin (NBF) overnight and then provided to a local clinical testing facility. Following slide preparation, H&E staining was conducted and slides were visualized under bright-field microscope. For western blot analysis of tumor tissue, homogenate was prepared in radioimmunoprecipitation assay (RIPA) buffer, kept at 4°C for 4 h, and then subjected to centrifugation at 10,000 rpm for 15 min. Blood was collected via intracardial syringe action and, upon clotting, was centrifuged at 3,000 rpm for 20 min to obtain serum for biochemical assays including SGOT, SGPT, albumin, serum urea, creatinine, and CK-MB activity. Assays for the above-mentioned parameters were conducted using Transasia, ERBA commercial kits by following the manufacturer’s protocol. ELISAs for TNF-α and IFN-γ were conducted from mouse serum using the respective manufacturer’s protocol (Ray Bio Mouse TNF-alpha ELISA Kit and Ray Bio Mouse IFN-gamma ELISA Kit).

Statistical analysis

The numerical values of results were given as mean ± standard error of the mean (SEM). One-way analysis of variance (ANOVA) with Tukey’s post hoc test was done for statistical evaluation of the data and for the determination of level of significance in various groups using the software Origin Lab 8.0. In all cases, a value of p < 0.05 was considered significant. The significance of differences was calculated by comparing all experimental groups with control. The symbol "∗" indicates significant difference between control and treatment groups until and unless it is mentioned in the figure legend.

Data availability

Detailed experimental procedures and characterization data, including the spectra for all new compounds, can be found in the supplemental information. Also, the datasets used in the current study are available from the corresponding author on reasonable request.