Hyperactivation of p53 using CRISPRa kills human papillomavirus-driven cervical cancer cells

Clinical and pre-clinical work for a number of cancer types has demonstrated relatively positive outcomes and effective tumour regression when the level and function of p53, a well-established tumour suppressor, is restored. Human papillomavirus (HPV)-driven cancers encode the E6 oncoprotein, which leads to p53 degradation, to allow the carcinogenic process to proceed. Indeed, there have been several attempts to revive p53 function in HPV-driven cancers by both pharmacological and genetic means to increase p53 bioavailability. Here, we employed a CRISPR activation (CRISPRa) approach to overcome HPV-mediated silencing of p53 by hyperexpressing the p53 gene promoter. Our data show that CRISPRa-mediated hyperexpression of p53 leads to HPV+ cervical cancer cell killing and the reduction of cell proliferation. This proof-of-concept data suggest that increasing p53 bioavailability may potentially be a promising therapeutic approach for the treatment of HPV-driven cancers.


Introduction
Human papillomavirus (HPV) is the causative agent in over 99% of cervical cancer cases, with high-risk HPV types 16 and 18 accounting for majority of cases [1]. Current gold standard lines of treatment (chemotherapy, surgery, or both) for these cancers has not improved survival rates in the past two decades and desperately warrants novel treatment approaches [2]. Indeed, HPV E6 and E7 oncoproteins drive the carcinogenic process in HPV-driven cancers [3]. Importantly, the E6 oncoprotein directly binds and dysregulates p53 function [4], a tumour suppressor protein resulting in loss of control of the mitotic cell cycle, allowing cells to proliferate uncontrollably. Previous efforts have been made to reverse the biological effects of p53 deficiency directly or indirectly in HPV cancers using both pharmaceutical and genetic means [5]. However, all the attempts to date have not resulted in the permanent, constitutive expression of p53. In this study, we aim to overcome p53 dysregulation in HPV + cervical cancers by increasing the bioavailability of p53 using a novel CRISPR activation (CRISPRa)-based approach. Here, we engineered a doxycycline (DOX)-inducible CRISPRa system using a "deficient" cas9 (dCas9) fused with heterologous activator domains [6], VP64-p65-Rta (VPR), to hyperexpress the p53 gene promoter.

Chemicals
To promote gRNA expression in cell lines, doxycycline hyclate (DOX) (#D9891, Sigma-Aldrich, St Louis, MI) was dissolved in sterile water at a stock concentration of 10 mg/ ml and added to cell culture media at a final concentration of 10 µg/ml.

Lentivirus production and transducing HeLa and CaSki dCas9 cells
Low passage HEK293T cells (human embryo kidney cell line) were co-transfected with pMDG2.G, pRSV-Rev, pMDLg/pRRE, and the p53 gRNA containing FgH1tUTG plasmid using the Lipofectamine 3000 transfection reagent (#L3000015, Thermo Fisher Scientific, Waltham, MA). After 48 h post-transfection, viral supernatant was concentrated in Amicon ® Ultra-15 Centrifugal Filter units (Merck, Germany) and viral titre determined by flow cytometry (BD LSR FORTESSA cell analyser (BD bioscience, San Jose, CA)). HeLa and CaSki dCas9 cells were then infected with lentivirus bearing targeting gRNAs for 24 h, before cells were subjected to cell sorting on the BD FACSAria™ III Cell Sorter (BD bioscience, San Jose, CA).

Colony forming assay
Cells were seeded at 300 cells per well in a 6-well plate and left overnight to adhere before adding media containing 10 µg/ml of DOX. DOX-supplemented media were changed daily over 8 days before media were removed and cells stained with crystal violet. Plates were visualised and imaged captured on a Chemidoc XRS Visualiser using a white light filter setting (BioRad, Hercules, CA).

Immunoblotting
Protein from cells were extracted in RIPA buffer (Thermo Scientific, Waltham, MA) containing 1 × Halt Protease Inhibitor (Thermo Scientific, Waltham, MA). Immunoblots were probed with antibodies against p53 (Cell Signaling Technologies, Danvers, MA) and s6 (Cell Signaling Technologies, Danvers, MA). Rabbit and mouse secondary antibodies (Cell Signaling Technologies, Danvers, MA) and ECL were used to detect protein signals on a Chemidoc XRS Visualiser (BioRad, Hercules, CA).

Fig. 1 Hyperexpression of p53 is toxic to HPV + cervical cancer cells.
A Diagrammatic representation of the doxycycline (DOX) inducible gRNA system. dCas9 is constitutively expressed in cells. Treatment with DOX rapidly induces the sgRNA, which activates dCas9 and directs it to target the p53 gene promoter sequence. B HeLa and CaSki dCas9 control (-) or gRNA-bearing (gRNA#1 or #2) cells were either treated with sterile water (− DOX) or with DOX (10 μg/ ml) (+ DOX) before performing an MTT assay at the indicated times. Data representative of one out of three independent experiments. Bars denote mean percentage viability to dCas9 control cells treated with its respective treatment groups. Error bars denote SEM of technical quadruplicate treatments. *p < 0.05, **p < 0.01, ***p < 0.005, ****p < 0.0001, One-way ANOVA test. C HeLa and CaSki dCas9 control (-) or gRNA-bearing (gRNA#1 or #2) cells were either treated with sterile water (− DOX) or with DOX (10 μg/ml) (+ DOX) over 8 days before performing a colony forming assay. Data repre-sentative of one out of three independent experiments. D HeLa and CaSki dCas9 control (-) or gRNA-bearing (gRNA#1 or #2) cells were either treated with sterile water (− DOX) or with DOX (10 μg/ ml) (+ DOX) before extracting RNA at the indicated times. RNA was then subjected to RT-PCR to measure p53 mRNA expression. p53 expression was measured relative to that of the housekeeping gene, GAPDH. Fold p53 expression to p53 expression in dCas9 control cells treated with its respective treatment groups is shown. Error bars denote SEM of technical triplicate treatments. Data representative of one out of three independent experiments. *p < 0.05, **p < 0.01, ****p < 0.0001, One-way ANOVA test. E HeLa and CaSki gRNAbearing (gRNA#1 or #2) cells were either treated with sterile water (− DOX) or with DOX (10 μg/ml) (+ DOX) over 72 h before proteins were extracted and immunoblotted for p53. s6 protein was used as a loading control. Data representative of one out of three independent experiments

Statistical analysis
Differences between treatment groups were done using an ordinary one-way ANOVA test on GraphPad Prism v9.

Results and discussion
Several approaches to increase p53 bioavailability in HPVdriven cancers have been previously attempted. A human recombinant adenovirus gene therapy drug that expresses wild-type p53 protein, Gendicine®, when tested in combination with radiotherapy resulted in better clinical response and overall survival rates than with radiotherapy alone in advanced cervical cancer patients [8,9]. Previous preclinical attempts to overexpress p53 using p53 expressing plasmids [10], chemotherapeutics [11][12][13], miRNA targeting [14], and indirectly by targeting pathways or interrupting oncoprotein function [15,16] have all resulted in favourable HPV + cervical cancer cell killing in vitro. However, none of these approaches result in the long-lasting, permanent overexpression of p53. To achieve this, we employed a CRISPRa approach with the aim of directly hyperactivating p53 gene expression via its promoter. HPV + cervical cancer cell lines, HeLa and CaSki, overexpressing dCas9 fused with the VP64-p65-Rta (VPR) transcriptional activating machinery (HeLa dCas9 and CaSki dCas9) were engineered with a doxycycline (DOX)-inducible gRNA expressing system [7], whereby DOX will induce the expression of p53-targeting gRNAs (Fig. 1A). We hypothesise that hyperexpression of p53 would result in cell killing. Indeed, DOX treatment of cells bearing two independent p53-targeting gRNAs (gRNA#1 and #2) resulted in significant time-dependent cell killing from 48 h onwards (Fig. 1B) and reduced cell proliferation (Fig. 1C) for both cell lines. Importantly, this cell killing trend is consistent with the observed significant time-dependent increase in p53 gene (Fig. 1D) and p53 protein (Fig. 1E) expression. Overall, we show that elevated p53 expression leads to cervical cancer cell killing.
Unlike other cancer types [17], p53 is not mutated in HPV-driven cervical cancers [18,19]. It is important to note the challenges expected with increasing p53 bioavailability for HPV cancers. In HPV cancers, a switch from minute double murine 2 protein (MDM2)-p53 binding to E6-mediated degradation of p53 is an important hallmark of HPV-driven cancers [4]. MDM gene encodes E3 ubiquitin-ligases, which are involved in the p53 negative feedback loop, degrading p53 rapidly when there are no signals for DNA stabilisation (e.g. DNA damage). Increasing levels of p53 in cells will not only compete with E6 binding but also its own negative feedback loop involving MDM2. Furthermore, the p53-p21-RB-like, E2F and multi-vulval class B (DREAM) complex is a target for the E7 oncoprotein and is directly involved in the p53-p21 activation pathway [20]. Hence, increasing p53 expression alone may not be sufficient to overcome this and may require a combinatorial therapeutic approach (e.g. standard chemotherapy). Of concern, hyperexpression of p53 can lead to prion formation in yeast [21] and this could pose a problem if this occurs in mammalian cells. Directly targeting E6 and E7 using CRISPR has been shown to be successful in clearing HPV + cervical cancer tumours in vivo [22]. It is likely that combination therapy of targeting HPV oncogenes and increasing p53 bioavailability can result in an additive effect as recently demonstrated by Xiong et al., [23]. Nonetheless, the positive effect of increasing p53 bioavailability in HPV-driven cancers is clear and reflects the potent anti-tumour effects of p53. Overall, using a novel CRISPRa approach, our data support the notion that hyperactivating p53 for HPV-driven cancers and that this approach could be applied for cancer types carrying the wild-type p53 gene.