The G-quadruplex DNA stabilizing drug pyridostatin promotes DNA damage and downregulates transcription of Brca1 in neurons

Pyridostatin is toxic for primary neurons

The G-quadruplex is a tetra-stranded secondary DNA structure formed by four stretches of guanines (Fig. 1A-C) [6]. Pyridostatin is a very selective G-quadruplex DNA-binding small molecule designed to form a complex with and stabilize G-quadruplex structures [9,18,19]. For cell lines, the drug is cytotoxic, and we, therefore, hypothesized that pyridostatin may be toxic for primary neurons as well. To test this, primary rat cortical cultures were transfected with the mApple construct (a red fluorescent protein). Pyridostatin or vehicle was added, and the mApple-expressing neurons were tracked for several days with an automated microscope. Loss of the red mApple fluorescence is a sensitive marker of neuronal death [20]. This method allows us to track large cohorts of individual neurons over their lifetimes and to sensitively measure their survival with statistical analyses used in clinical medicine [20]. Analyzing when each neuron lost its fluorescence allows us to measure neuronal survival with cumulative hazard statistics. By following neurons over their lifetimes, we can determine if the applied drug contributes positively, negatively, or even neutrally to neuronal fate (Fig. 1D). We found that treatment with pyridostatin substantially increased the risk of neuronal death in a dose-dependent manner (Fig. 1E). Notably, neurons exposed to pyridostatin retracted neurites before death (Fig. 1D), mimicking neurodegenerative processes commonly observed in neurons that express α-synuclein, mutant LRRK2 or mutant huntingtin [21,22].

Neurite retraction in pyridostatin-treated neurons suggests that neurons have lost their synapses as well. Cultures were treated with pyridostatin and fixed before neurons started dying. Expectedly, we found that pyridostatin reduced the density of synapses. Pyridostatin-treated neurons also had less neurites than vehicle-treated neurons (Fig. 1F–H).

Pyridostatin induces DNA DSBs

Putative G-quadruplexes are present on average once per 10 kb of the human genome [23]. In addition to inhibiting DNA replication, the pyridostatin–DNA complex may stall DNA polymerase during transcription [19,24]. DNA damage may then occur via the action of endonucleases, through a mechanism of so-called transcription–coupled–repair poisoning [25]. Since pyridostatin induces DNA DSBs in cell lines [9], we hypothesized that pyridostatin may induce DNA damage in primary neurons. To analyze DNA damage in neurons in our lab, we recapitulated a recent single–cell analysis technique based on the expression of a truncated 53BP1 DSB reporter in cell lines [26]. The mApple-53BP1trunc construct was expressed in two cohorts of primary neurons along with GFP, a marker of cell viability and morphology. The first neuronal cohort was treated with a vehicle; the second cohort was treated with pyridostatin. Both neuronal cohorts were then followed longitudinally (Fig. 2A). Expectedly, the mApple-53BP1trunc construct formed puncta in neurons exposed to pyridostatin, reflecting DNA damage (Fig. 2B).

Since overexpressed proteins sometimes aggregate in primary neurons, we decided to confirm our finding with the endogenous 53BP1 protein. Cultured neurons were treated with pyridostatin, fixed and stained for endogenous 53BP1 and MAP2c, and with the nuclear DAPI dye. Neurons exposed to pyridostatin had punctuated structures of 53BP1 localized to their nuclei (Fig. 2C,D).

Phosphorylated histone H2A variant X (γH2A.X) is commonly used as a read-out of DNA DBSs [14]. Therefore, to make sure that γH2A.X is phosphorylated when neurons are treated with pyridostatin, as seen in non-neuronal cells, neuronal cultures were treated with pyridostatin, fixed and stained for γH2A.X and MAP2c, and with the nuclear DAPI dye. Neurons exposed to pyridostatin had more staining of γH2A.X in the neuronal nuclei than control neurons (Fig. 3A,B). Our data suggest that stabilized G-quadruplexes contribute to DNA damage in neuronal cells.

Since DNA damage may, in principle, be a general phenomenon in degenerating neurons, we determined if other neurotoxic small molecules, bortezomib and cisplatin, promote DNA DSBs. Bortezomib, a chemotherapy drug that inhibits the ubiquitin-proteasome system (UPS), was used to promote neurodegeneration. Inhibitors of the UPS are highly neurotoxic [27]. Cisplatin, another anti-tumor agent, cross links DNA and, therefore, promotes cytotoxicity [28]. Cisplatin is also highly toxic for neurons [29]. Primary neurons were treated with bortezomib and cisplatin, and then fixed and stained for γH2A.X and MAP2c. Neither bortezomib nor cisplatin induced DNA DSBs (Fig. 3А,B). Our data demonstrate that DNA DSBs are not a general effect induced by cytotoxic agents, and pyridostatin-associated DNA damage likely depends on pyridostatin’s G-quadruplex-stabilizing properties.

Do other G-quadruplex-stabilizing drugs induce DNA DSBs? We tested if TmPyP4 [30], а cationic porphyrin, induces accumulation of γH2A.X-positive puncta in cultured primary neurons. Neuronal cultures were treated with TmPyP4 under conditions identical to the pyridostatin experiment, fixed and stained for γH2A.X and MAP2c, and with the nuclear DAPI dye. TmPyP4 stimulated the formation of γH2A.X-positive puncta in neurons (Fig. 3C).

Pyridostatin downregulates BRCA1 in neurons

The breast cancer type 1 susceptibility protein, BRCA1, repairs DNA DSBs in non-neuronal cells and neurons. In addition to well-studied BRCA1 dysregulation in cancer cells, BRCA1 protein is downregulated in the brains of patients with Alzheimer's disease. Amyloid-beta, a peptide associated with Alzheimer's disease, downregulates BRCA1 leading to the formation of DNA DSBs [31].

The Brca1 gene in Rattus norvegicus is a long gene that has two known alternatively spliced RNA products. We hypothesized that the Brca1 gene may contain putative G-quadruplexes. We analyzed the rat Brca1 gene sequence with the QGRP (putative quadruplex forming G-rich sequences) mapper and discovered that, in the gene, two putative sequences in the coding sequence can arrange into the G-quadruplexes. Additionally, its introns contain four putative G-quadruplex-forming sequences. There are also two putative G-quadruplex-forming sequences in the Brca1 gene promoter (Fig. 4A). Therefore, pyridostatin may stabilize these putative G-quadruplexes and inhibit the Brca1 gene transcription.

First, we tested if the levels of the BRCA1 protein are changed in neurons exposed to pyridostatin. Cultured cortical neurons were treated with pyridostatin, and then their extracts were analyzed by western blotting. Remarkably, the levels of the BRCA1 protein were indeed decreased (Fig. 4B,C). Levels of actin in samples were not affected, suggesting that the actin protein itself is not a target of pyridostatin. To confirm this finding, we investigated if TmPyP4 reduces the levels of BRCA1 in neurons. The BRCA1 protein levels were downregulated in TmPyP4-treated neurons as well (Fig. 4D,E). We, therefore, conclude that, in primary neurons, stabilizing G-quadruplex DNA downregulates BRCA1, at least at the protein level. As a result, DNA DSBs can accumulate due to the action of nucleases and to downregulated BRCA1, at least in part.

Pyridostatin was designed to selectively bind to G-quadruplex DNA and to stabilize the G-quadruplex structures [9,18,19]. Nevertheless, the drug can theoretically inhibit protein translation or even affect its degradation. To test whether transcription of the Brca1 gene is directly affected, we sought to investigate whether pyridostatin changes BRCA1’s mRNA levels in neuronal cells exposed to the drug. Neuronal cultures were treated with pyridostatin; mRNA was extracted and analyzed. Remarkably, the levels of BRCA1 mRNA were higher in pyridostatin-treated neurons, when normalized to the control actin mRNA (Fig. 4F).

An increase in BRCA1 mRNA levels was unexpected, but could reflect a change in the levels of actin mRNA itself in pyridostatin-treated neurons. To test that, we measured the levels of actin mRNA in vehicle-treated neurons and in neurons exposed to pyridostatin, and discovered that, indeed, actin mRNA was downregulated, indicating that pyridostatin targets the Actin gene as well. Interestingly, we did not observe a decrease in the levels of the actin protein (Fig. 4B), which indicates that actin is a long-lived protein.

Searching for other loading controls, we tested mRNA levels of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and TATA-binding protein (TBP). Interestingly, neurons exposed to pyridostatin had reduced levels of GAPDH, but the levels of TBP mRNA remained unchanged (Fig. 4G). Therefore, TBP mRNA was used as loading control, and in this case, BRCA1 mRNA was downregulated in neurons treated to pyridostatin (Fig. 4H).

We were surprised that pyridostatin downregulated BRCA1, actin, and GAPDH, but not TBP mRNAs. We, therefore, analyzed the sequences of the Actin, Gapdh and Tbp genes and their promoters with the QGRP mapper. Both Actin’s and Gapdh’s promoters contain two putative G-quadruplex-forming sequences, and Actin contains one putative G-quadruplex sequence. Remarkably, neither Tbp nor its promoter contains a putative G-quadruplex-forming sequence, indicating that our data generated with pyridostatin are physiologically relevant. Future studies will identify more genes regulated by the G-quadruplex.

Overexpressed BRCA1 mitigates pyridostatin-induced DNA damage

BRCA1 is required for DNA damage repair. We wondered if ectopically increasing BRCA1 protein levels would attenuate DNA damage associated with the exposure to pyridostatin. To test this, neurons were transfected with a plasmid that encodes GFP or GFP-BRCA1. Neurons were treated with a vehicle or with pyridostatin overnight and then fixed and immunostained with antibodies against γH2A.X or 53BP1 (Fig. 5A–D). The γH2A.X and 53BP1 puncta indexes in GFP-BRCA1-expressing neurons treated with pyridostatin were lower than in GFP-expressing neurons exposed to pyridostatin (Fig 5A–D). These results indicate that ectopically expressed BRCA1 mitigates DNA damage induced by pyridostatin treatment in neurons and demonstrate that BRCA1 is critical for maintaining neuronal genome integrity.

The BG4 antibody binds to a synthetic oligonucleotide that corresponds to a putative G-quadruplex in the Brca1 gene promoter

A single–chain BG4 antibody was identified through a phage display method with a library of 2.3 x 10¹⁰ different single–chain antibody clones with an in vitro selection for G-quadruplex structures. The BG4 antibody recognizes intramolecular and intermolecular DNA G-quadruplexes with a nanomolar affinity and does not bind to single-stranded or double-stranded DNA [32]. To determine if the putative G-quadruplexes folds into a four-strand structure that can be recognized by the BG4 antibody, we synthesized three Cy5-labeled oligonucleotides: the BRCA1-G1 and BRCA1-G2 oligonucleotide corresponded to the two putative G-quadruplex sequences in the Brca1 promoter, and the BRCA1-cont was a control oligonucleotide, which cannot fold into a G-quadruplex structure.

The G-quadruplexes form a stable structure in the presence of K⁺ and Na⁺ ions and do not properly fold into a G-quadruplex in a buffer that contains Li⁺ [33]. We incubated the BG4 antibody with the BRCA1-G1 oligonucleotide, the BRCA1-G2 oligonucleotide, and the BRCA1-cont oligonucleotide in the buffers that contained K⁺, Na⁺, or Li⁺ ions (Fig. 6A–C). The BG4 antibody did not form a complex with the BRCA1-cont oligonucleotide in all conditions tested (Fig. 6A–C). The BRCA1-G2 oligonucleotide also did not form a complex with the BG4 antibody. However, the BG4 antibody did bind to the BRCA1-G1 oligonucleotide, which was reflected by the slower migrating band on the gel, in both K⁺ and Na⁺ buffers. In the presence of Li⁺, the BRCA1-G1 oligonucleotide formed a weaker complex with the BG4 antibody (Fig. 6A–C).

We were surprised that the BG4 antibody recognized the BRCA1-G1 oligonucleotide as a G-quadruplex structure, but did not bind to the BRCA1-G2 oligonucleotide. In fact, however, the BRCA1-G1 oligonucleotide (AGGGCTCTCCCTTGGGGGGGGGGGGCAGGAAGGGA) contains more guanines and longer guanine–runs than the BRCA1-G2 oligonucleotide (AGGGAGGCCATGGGACGGGAAGACTTGGGT), indicating that the G-quadruplex formed by the BRCA1-G1 oligonucleotide may be significantly more stable.

Chemicals and plasmids

Pyridostatin (PDS) was obtained from Cayman Chemical (#18013). TmPyP4 was from Calbiochem (#613560). Hoechst dye was from Santa Cruz Biotechnology (#sc-394039). Antibodies against synapsin were from Synaptic Systems (#106002). Antibodies against BRCA1 (#ab191042) and 53BP1 (#ab172580) were from Abcam. Antibodies against γH2A.X (JBW301) were from EMD Millipore (#05636). Rabbit antibodies against MAP2c (H-300, #sc-20172), and mouse antibodies against MAP2c (A-4, #sc-74421) were from Santa Cruz Biotechnology. Antibodies against β-actin (8H10D10) were from Cell Signaling (#3700). Anti-rabbit Alexa Fluor 488–labeled (#A11008), anti-mouse Alexa Fluor 488–labeled (#A11001), anti-rabbit Alexa Fluor 546–labeled (A11010), and anti-mouse Alexa Fluor 546-labeled (#A11003) secondary antibodies were from Life Technologies. pCAG-mApple-hTP53BP1 (1220-1709aa) was synthesized by Vectorbuilder. pSANG10-3F-BG4 was from Addgene (#55756; deposited by Dr. Shankar Balasubramanian, the University of Cambridge). pDEST-FRT/T0-GFP-BRCA1 was from Addgene (#71116; deposited by Dr. Durocher, the University of Toronto).

Neuronal and astrocytic cultures and transfection

Primary cortical neurons and astrocytes were isolated from rat embryos (E17-E19), and plated on 24-well tissue-culture plates (4x10⁵/well) coated with poly-D-lysine (BD Biosciences), as described [14,52]. Neuronal cultures were maintained in Neurobasal Medium (Life Technologies) supplemented with B-27 (Life Technologies), GlutaMAX (Life Technologies), and penicillin-streptomycin (Life Technologies). Some primary cultures were transfected after 4 days in vitro with Lipofectamine 2000 (Invitrogen) and with a total of 1 μg of plasmid DNA per well, as described [14,52–54].

Fluorescence microscopy and image analysis

Live cell imaging was performed with the EVOS microscopy system (Life Technologies). Fixed and immunostained cells were imaged with the RFP (MAP2c, 53BP1 and γH2A.X) filter, the GFP (synapsin, 53BP1, γH2A.X and GFP-BRCA1) filter, and the DAPI filter (Hoechst). γH2A.X and 53BP1 fluorescence was analyzed by the puncta index, which is the standard deviation of the intensities measured among pixels within the neuronal nuclei. Low puncta index represents diffuse localization, whereas a high puncta index represents punctate localization.

Longitudinal fluorescent microscopy and survival analysis

For longitudinal imaging of DNA damage in living cells, primary cortical neurons were transfected with pCAG-mApple-hTP53BP1 (mApple-53BP1trunc) and pGW1-GFP, and then treated with a vehicle or 1 μM PDS. For survival experiments, primary cortical neurons were transfected with pGW1-mApple and treated with a vehicle or different concentrations of PDS (0.01, 0.1, 1 and 5 μM). Then, cultures were imaged every 24 h for 5 days. The plate was placed on the microscope stage, which automatically moves the 20x objective to the center of the first well and collects fluorescence images with the red filter (mApple, mApple-53BP1trunc) and the green filter (GFP), thereafter moving the stage to each adjacent field in the well. These steps are repeated until all required wells are imaged. For tracking the same group of neurons over time, an image of the fiduciary field with neurons on the plate was collected at the first time-point and used as a reference image. Each time the same plate was imaged thereafter, the fiduciary image was aligned with the reference image. Neurons that died during the imaging interval were assigned a survival time. These event times were used to obtain the exponential cumulative survival graphs and analyzed for statistical significance by Log-Rank test with JMP software (SAS Institute) as described [14,54,55].

Immunocytochemistry

Primary cortical neurons were fixed with 4% paraformaldehyde for 15 min at room temperature. Then, neurons were permeabilized with 0.1% Triton X-100 in PBS, and blocked with 1% bovine serum albumin in PBS for 1 h at room temperature. Cells were incubated with a primary antibody (anti-MAP2c, anti-synapsin, anti-53BP1, or anti-γH2A.X) diluted in blocking buffer at 4°C overnight. Cells were washed, and incubated with a secondary antibody (anti-rabbit-Alexa Fluor 546, anti-mouse-Alexa Fluor 546, anti-mouse-Alexa Fluor 488, anti-rabbit-Alexa Fluor 488) diluted in blocking buffer for 1 h at room temperature. Nuclei were stained with Hoechst dye in PBS for 2 min. Cells were washed and analyzed.

Immunoblotting

Neuronal cultures were collected and lysed in RIPA buffer (150 mM NaCl, 1% Nonidet P40, 0.5% sodium deoxycholate, 0.1% SDS and 50 mM Tris/HCl (pH 8.0), with phosphatase and protease inhibitors cocktail) on ice. Lysates were vortexed and cleared by centrifugation (14000 g, 10 min, 4°C). Supernatants were collected, and protein concentrations were determined by the Bicinchoninic Acid Protein Assay Kit (Thermo Scientific), according to the manufacturer’s protocol. Samples were analyzed by SDS/PAGE (4-12% gradient gels), and proteins were transferred on to nitrocellulose membranes using the iBlot2 system (Life Technologies). Membranes were blocked with 5% (w/v) non-fat dried skimmed milk for 1 h at room temperature, and they were incubated with anti-BRCA1 and anti-actin, as a loading control, overnight at 4°C. Membranes were then washed with TBS (Tris-buffered saline; 10 mM Tris/HCl and 150 mM NaCl (pH 7.4)) and incubated with anti-rabbit-HRP and anti-mouse-HRP for 1 h at room temperature. Chemiluminescent signal was visualized with Prometheus ProSignal Pico (Genesee Scientific) on Blue Devil autoradiography films (Genesee Scientific).

RNA extraction and qRT-PCR

Total RNA was extracted from primary culture using the RNeasy Mini kit (#74104, Qiagen), and then reverse transcribed using iScript Reverse Transcription SuperMix (#1708840, Bio-Rad), according to the manufacturer’s protocol. RT-qPCR was performed using a Bio-Rad CFX96 Touch machine using SSoAdvanced Universal SYBR Green (#1725275, Bio-Rad) for visualization and quantification according to the manufacturer’s instructions. Primer sequences were: BRCA1, forward: 5′-GCAGATGGGCTGACAGTAAA-3′, reverse: 5′-GCTTTCTACCACAGAGGGAATC-3′, TBP, forward: 5′-AGTGCCCAGCATCACTGTTT-3′, reverse: 5′-GGTCCATGACTCTCACTTTCTT-3′. β-actin, forward: 5′-CTTCACCACCACGGC-3′, reverse: 5′-CCATCTCTTGCTCGAAG-3’. GAPDH, forward: 5′-CAACTACATGGTCTACATGTTC-3′, reverse: 5′-CTCGCTCCTGGAAGATG-3′. PCR conditions were: 95°C for 3 min, followed by 40 cycles of 95°C for 10 s and 55°C for 30 s. Relative expression levels were calculated from the average threshold cycle number using the delta-delta Ct method.

Putative G-quadruplex analyses

The QGRS mapper (http://bioinformatics.ramapo.edu/QGRS/index.php) was used to determine the potential G-quadruplex structures contained in genes of interest. Search parameters: maximal length: 45; minimal G-group size: 3; loop size: from 0 to 10 [7,23].

BG4 binding assay

BG4 antibody expression and purification was carried out as described for the HF2 antibody [56] with the following modifications. E. coli BL21 cells were transformed with the pSANG10-3F-BG4 plasmid. The expression of the BG4 single–chain antibody was then induced by 100 mM isopropyl β-D-1-thiogalactopyranosid. Cells were centrifuged and pellets were resuspended in a lysis buffer (25 mM Tris-HCl, 100 mM NaCl, 10% glycerol, 1% NP-40 and 10 mM imidazole) and sonicated using the QSONICA sonicator. Purification of the 6XHis-tagged BG4 antibody was carried out using HisPur Ni-NTA resin (Thermo Scientific). The eluted protein was concentrated with the Amicon Ultra-4 Centrifugal Filter and stored at -20ºC in 50% glycerol. For the binding assay, 5′- Cy5-labelled oligoes (Sigma) were resuspended in 10 mM Tris-Cl containing 100 mM LiCl, NaCl, or KCl and denatured at 95°C for 5 min and then slowly cooled overnight to allow secondary structure formation. Annealed oligos were mixed with the purified BG4 antibody in 100 mM LiCl, NaCl or KCl, 20 mM HEPES pH 7.5, 0.01% NP40, 5% glycerol, 5 mM MgCl_2, and incubated at room temperature for 15 min before running on a 10% non-denaturing TBE-polyacrylamide gel with 0.5X TBE. Gel images were captured using the BioRad Chemidoc imager. Sequences of the oligos are: BRCA1-G1: Cy5-AGGGCTCTCCCTTGGGGGGGGGGGGCAGGAAGGGA; BRCA1-G2: Cy5-AGGGAGGCCATGGGACGGGAAGACTTGGGT; BRCA1-cont: Cy5-AGCTTCCAAAATAGCCAACTGCGAGCTAATTCTGT.

Statistical analyses

The statistical analyses and tests were performed with ImageJ software.

Ethics statement

Rats were maintained in accordance with guidelines and regulations of the University of Texas McGovern Medical School (the protocol number is #AWC-16-0081). All experimental protocols were approved by the University of Texas McGovern Medical School. The methods were carried out in accordance with the approved guidelines.