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

The G-quadruplex is a non-canonical DNA secondary structure formed by four DNA strands containing multiple runs of guanines. G-quadruplexes play important roles in DNA recombination, replication, telomere maintenance, and regulation of transcription. Small molecules that stabilize the G-quadruplexes alter gene expression in cancer cells. Here, we hypothesized that the G-quadruplexes regulate transcription in neurons. We discovered that pyridostatin, a small molecule that specifically stabilizes G-quadruplex DNA complexes, induced neurotoxicity and promoted the formation of DNA double–strand breaks (DSBs) in cultured neurons. We also found that pyridostatin downregulated transcription of the Brca1 gene, a gene that is critical for DSB repair. Importantly, in an in vitro gel shift assay, we discovered that an antibody specific to the G-quadruplex structure binds to a synthetic oligonucleotide, which corresponds to the first putative G-quadruplex in the Brca1 gene promoter. Our results suggest that the G-quadruplex complexes regulate transcription in neurons. Studying the G-quadruplexes could represent a new avenue for neurodegeneration and brain aging research.

In this study, we investigated if pyridostatin affects neuronal homeostasis. Pyridostatin decreased neuronal survival, damaged synapses, and promoted the formation of DNA DSBs. We found that, in neurons, pyridostatin downregulated a DNA DSB-repairing protein BRCA1 at the transcriptional level. Interestingly, in an in vitro gel shift experiment, a single-chain antibody raised against the G-quadruplex structures, binds to a synthetic oligonucleotide, which corresponds to the first putative G-quadruplex in the Brca1 gene promoter, indicating that our data are physiologically relevant. Neuronal G-quadruplexes in general and the G-quadruplex-dependent downregulation of transcripttion in neurons, therefore, might be a novel pathway of brain aging and a new target for mitigating ageassociated neurodegeneration.

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).

AGING
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). structure. Four guanine molecules (a single guanine is in red) can assemble into a square planar structure. The structure of a Gquadruplex is stabilized by hydrogen bonds between guanines and the interactions with a monovalent cation (Na + or K + ) resided in the central channel. (B) Repetitive guanine-rich DNA or RNA sequences have the potential to form G-quadruplex structures (C). (D) An example of survival analysis. Primary cortical neurons were transfected with mApple (a morphology and viability marker) and tracked with an automated microscope. Images collected every 24 h demonstrate the ability to return to the same field of neurons and to follow them over time. Each image is a montage of non-overlapping images captured in one well of a 24-well plate. AGING 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 AGING 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. Amyloidbeta, 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-quadruplexforming 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 Gquadruplex 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).  AGING An increase in normalized 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 down-regulated, 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 AGING that our data generated with pyridostatin are physiologically relevant. Future studies will identify more genes regulated by the G-quadruplex.

Overexpressed BRCA1 mitigates pyridostatininduced 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 Gquadruplex in the Brca1 gene promoter A single-chain BG4 antibody was identified through a phage display method with a library of 2.3 x 10 10 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 Gquadruplex 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].

AGING
We incubated the BG4 antibody with the BRCA1-G1 oligonucleotide, the BRCA1-G2 oligo-nucleotide, 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 oligo-nucleotide, 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 (AGGGCTCTCCCTTGGGGGGGGG GGGCAGGAAGGGA) contains more guanines and longer guanine-runs than the BRCA1-G2 oligonucleotide (AGGGAGGCCATGGGACGGGAAGACT TGGGT), indicating that the G-quadruplex formed by the BRCA1-G1 oligonucleotide may be significantly more stable.

DISCUSSION
Our results revealed a potentially new mechanism of neurodegeneration and transcriptional regulation in neurons (Fig. 7). Pyridostatin, a G-quadruplexstabilizing small molecule, causes neurite retraction, synaptic loss, and dose-dependent neuronal death. In cultured primary neurons, pyridostatin induces the formation of DNA DSBs. Remarkably, the drug downregulates the BRCA1 protein, a protein that guards and repairs the neuronal genome, at the transcriptional level. Using an in vitro gel shift experiment, we demonstrated that an antibody developed against the Gquadruplex structures binds to a synthetic oligonucleotide, which was synthesized to correspond to the first putative G-quadruplex sequence in the Brca1 gene promoter. Our results suggest that small molecules with pyridostatin-like properties promote neurodegeneration and identify a novel mechanism of transcriptional regulation in neurons.
Up to 700,000 G-quadruplex-induced polymerasestalling sites were detected in the human genome with a high-throughput sequencing method [34]. In mammalian cells, the putative G-quadruplex structures are frequent in oncogenes and regulatory genes [35,36]. The G-quadruplex in the promoter region negatively regulates the expression of the oncogenes c-MYC and KRAS. Small molecules that stabilize the G-quadruplex structures significantly decrease c-MYC and KRAS expression [37,38]. Notably, the G-quadruplex structures positioned as far as 3,500 base pairs away from the transcription start codon reduce gene expression [39]. An antibody-based chromatin immunoprecipitation and high-throughput sequencing The Cy5-conjugated BRCA1-G1 (G1), BRCA1-G2 (G2), and BRCA1-cont (cont) oligonucleotides (see Methods) were heat-denatured and then slow-cooled in the presence of K + (KCl) to allow the formation of a secondary structure. 1.5 pmoles of each oligonucleotides (oligo) and 0 (a buffer alone), 30 or 50 ng of the BG4 antibody were incubated in a buffer, which contained 100 mM KCl. Note the bands that correspond to the free BRCA1-G1 (at the bottom) and the BRCA1-G1 bound to the BG4 antibody (at the gel top), both indicated with arrow heads. An asterisk marks the faint bands, which resulted from a loading dye. (B) The Cy5-conjugated BRCA1-G1 (G1), BRCA1-G2 (G2), and BRCA1-cont (cont) oligonucleotides were prepared in the presence of Na + (NaCl). 1.5 pmoles of each oligonucleotides (oligo) and 0 (a buffer alone), 30 or 50 ng of the BG4 antibody were incubated in a buffer, which contained 100 mM NaCl. Note the bands that correspond to the free BRCA1-G1 (at the bottom) and the BRCA1-G1/BG4 complex (at the gel top), depictured with arrow heads. An asterisk marks the loading dye bands. (C) The Cy5-labeled BRCA1-G1 (G1), BRCA1-G2 (G2), and BRCA1-cont (cont) oligonucleotides were prepared in the presence of Li + (LiCl). 1.5 pmoles of each oligonucleotides (oligo) and 0 (a buffer alone), 30 or 50 ng of the BG4 antibody were incubated in a buffer, which contained 100 mM LiCl. Note the bands that correspond to the free BRCA1-G1 (at the bottom) and a weak BRCA1-G1/BG4 complex (at the gel top), depictured with arrow heads. An asterisk marks the loading dye bands.
AGING revealed about 10,000 G-quadruplex structures in the chromatin of human epidermal keratinocyte cells, mostly in nucleosome-free regions [40]. Our data suggest that neuronal cells have a neuron-specific set of "active" G-quadruplex structures. In the aged brain, in which G-quadruplex-stabilizing guanine oxidation is frequent, many genes involved in synaptic plasticity, vesicular transport, mitochondrial function, and protein turnover are downregulated [10]. Importantly, several genes involved in learning and memory have been studied with a great interest and are downregulated in aged brains. These are such as Egr1 and Arc, which whose promoters are abnormally methylated during brain aging [41][42][43]. The levels of synaptic plasticityrelated Agrn, Gap43, Syn1 and Vgf mRNAs also go down during aging [44,45]. All these genes or their promoters contain putative G-quadruplexes.
Interestingly, topotecan, an inhibitor of topoisomerase 1, which resolves DNA supercoiling during replication and transcription, downregulates the expression of a large number of long genes, including many neuronal genes [46,47]. The gene's length remarkably predicts whether inhibiting topoisomerase 1 downregulates gene transcription. Maab et al. speculate that topoisomerase 1 is the "Achilles heel" of neurons because its inhibition downregulates a number of long genes, which are critical for proper neuronal function [46]. Our data suggest that the G-quadruplex is the "chink in the armor" of neurons. The presence of a G-quadruplex may predict if a gene is sensitive to a potential Gquadruplex stabilizer. Intriguingly, there is a group of genetic disorders caused by mutations in the genes encoding DNA helicases, the enzymes that unwind DNA during replication and transcription [48]. Significantly, the G-quadruplexes are targeted by several DNA helicases, and mutations in these enzymes are associated with Cockayne syndrome, a rare fatal disease with a developmental and neurodegenerative phenotype [49][50][51].
In summary, we demonstrated that pyridostatin induces neurodegeneration, promotes DNA DSBs, and downregulates transcription in neurons. Our results are in an agreement with previous findings that stabilizing the G-quadruplexes is cytotoxic. Future studies in mice will determine if the G-quadruplex structures contribute to brain aging and neurodegeneration.

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, antisynapsin, 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, antimouse-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).

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.