Profiling ribonucleotide and deoxyribonucleotide pools perturbed by remdesivir in human bronchial epithelial cells

Remdesivir (RDV) has garnered much hope for its moderate anti-COVID-19 effects, but its limited amelioration of survival in hospitalized patient causes a huge controversy over the applicability of RDV to COVID-19 treatment. Developing strategies to improve its antivirus efficacy is urgently required. As anticipated, RDV exhibits similar behavior with other nucleotide analogs to disrupt the metabolism of natural endogenous ribonucleotides (RNs) and deoxyribonucleotides (dRNs). Alterations in endogenous RNs and dRNs play a critical role in virus replication as well as other key cellular functions. Thus elucidation of the disturbances of RDV on RNs and dRNs could help to understand its exact mechanism of action. Here, the metabolic profiling determined by liquid chromatography–mass spectrometry method showed a general increase in the abundance of nucleotides and a more than 2-fold increase for specific nucleotides. However, the variation of pyrimidine ribonucleotides was relative slight or even contrary, resulting in obvious imbalance between purine and pyrimidine ribonucleotides, which implied the obstacle of RDV to pyrimidine synthesis and could further block the transcription and replication of viral RNA. Additionally, the extreme disequilibrium between cytidine triphosphate (CTP) and cytidine monophosphate might result from the inhibition of CTP synthase and provide a metabolic target for the treatment of COVID-19 infection. Since nucleotides metabolism pathways are vulnerable to nucleotide analogues and are liable to be the regulation targets, it is promising to enhance the efficacy of RDV through co-administration with CTP synthase inhibitors or de novo pyrimidine synthesis inhibitors to exacerbate the imbalance of nucleotide pools.


RDV induced S phase arrest in BEAS-2B cells
On account of the significant inhibitory effect on cell viability and proliferation, we investigated the effect of RDV treatment on the distribution of cells in cell cycle under different time point. BEAS-2B cells were treated with or without 10 μM RDV for 12, 24 and 48 h, and subsequently analyzed by flow-cytometry. As shown in Figure 1 B, an altered pattern of cell cycle was observed in BEAS-2B cells exposed to 10 μM RDV and with respect to control. With the incubation time increased, the proportion of cells in S phase significantly increased while the percentage of cells in G2/M phases obviously decreased in comparison to untreated cells. After incubated for 24 h, the percentage of cells in S phase was 35.3 ± 0.75% in control, which gradually increases to 48.69 ± 1.8 % in RDV group (P < 0.01). The number of cell in G2 phase decreased from control 29.67 ± 1.59% to 20.81 ± 1.92% of RDV (P < 0.05). The resembled results at 48 h was obtained. In brief, RDV can arrest the cells in S phase.

RDV inhibited RNA and DNA synthesis
In order to detect the effects of RDV on RNA and DNA synthesis in proliferating cells, we performed EU and EdU staining method on the basis of click chemistry. EU and EdU are the structural analogues of uridine and deoxyuridine, respectively. Their triphosphate metabolites compete with UTP and TTP to incorporate into newly synthesized RNA and DNA, respectively, and subsequently reacted with Azide -modified fluorophores. The fluorescence intensity was proportional to the amount of the incorporated EU and EdU in nascent RNA and DNA. As shown in Figure 2 A and B, after incubation with RDV for 14 h, the fluorescence intensity of Alexa594-azide decreased significantly in comparison to control group, indicating the reduction of RNA and DNA synthesis, and the inhibition of proliferation of BEAS-2B cells. Interestingly, not all DAPI stained cells were labeled with EdU. The reason for this phenomenon is that the incorporation of EdU only occurs in S phase during DNA replicating, while DAPI is a nonspecific fluorescent dye with the strong binding ability to the existing or nascent DNA [37].

Perturbation of RNs and dRNs pool size by RDV in BEAS-2B cells
To examine metabolic reprogramming events that influence the cellular response to virus, we used targeted LC/MS-MS via selected reaction monitoring (SRM) to examine changes in the steadystate metabolomics profile of BEAS-2B cells after exposure to 10 μM RDV with 12 h, 24 h and 48 h. The specific nucleotide levels were shown in Supplementary Table 1 and Table 2. The fold change of the nucleotides were evaluated by comparison of their concentrations in cells treated with RDV and in the parallel controlled RDV-free cells at the same time points. Significant differences in the metabolite profiles of cells with or without RDV were observed. In general, RDV increased the abundance of the majority of RN and dRN species after 24 h incubation, including a greater than 2fold increase in AMP, GTP, dAMP, dGDP, dGTP, dCTP and TMP levels, then decreased to the normal levels at 48 h (Figure 2 C-F). A rational interpretation was that RDV significantly inhibited the synthesis of nascent RNA and DNA, and arrested the cell cycle in S phase, inevitably resulting in the accumulation of (deoxy)nucleoside triphosphates and subsequently the increase of their respective di-and monophosphates [38]. However, it was notable that most of the pyrimidine ribonucleotides remained unchanged or even declined, among which the significant decrease of CTP was in stark contrast to the 3-fold greater increment of CMP after incubation for 24 h (Figure 2 D). CTP is synthesized from UTP by CTP synthase, which is the rate-limiting step of de novo CTP biosynthesis and probably a practical target just as in the treatment of leukemia [39] and parasitic infections [40][41][42]. In this study, the ratio of CTP/UTP was calculated and shown a significant decrease after 24 incubation (Figure 3 D), implying the inhibition effect of RDV on CTP synthase. Besides of the de novo pathway, the salvage pathway plays an important role in cellular nucleotides metabolism as well.
The relative low level of CTP might allosterically activate the recycle of free bases and nucleosides to promote the production of CMP, resulting in the abnormal elevation of CMP ( Figure 4).  The alterations in nucleotide pools were also evaluated through comparing the percent of each NTP in the whole nucleotide pools. It showed that RDV exposure (10 μM) stimulates an increase in GTP and a decrease in CTP (Figure 3 A). Consequently, a significant increment of GTP/CTP was observed (Figure 3 C), indicating the huge disequilibrium in RN pools. Although there were no statistically significant differences, the level of ATP reduced and UTP increased slightly ( Figure 3A), resulting in the elevated ratio of ATP/UTP (Figure 3 C). From the aspect of drug disposition, RDV was hydrolyzed to its corresponding nucleoside monophosphate (RDV-MP) in cell, and furtherly metabolized to RDV-TP. Due to the structural similarity of RDV-MP to AMP, the further phosphorylation of RDV-MP was achieved through the competitive inhibition of adenylate kinase, which inevitably resulted in the accumulation of AMP and the decrease of ADP and ATP. Meanwhile, the accumulation of AMP might inhibited the activity of adenylosuccinate synthase and the whole purine biosynthesis pathway in a negative feedback mode, which would simultaneously decrease the production of GMP, and ultimately GDP and GTP [43]. This speculation was proved by the relative high levels of the AMP/GMP ratio and the reduced ATP/GTP and ADP/GDP ratios at 24 h and 48 h (Figure 3 C and D). The relative percent of dNTPs pools were shown in Figure 3 B. The change of dNTPs percent were just contrary to that of NTPs, and there was obvious hysteresis, which was probably because of the allosteric regulation of NTPs to riboreductase. In summary, RDV exerted the antiviral activity partly via aggravating the imbalance of nucleotide pools, especially through reducing CTP.

RDV upregulated the riboreductase R2 expression
The remarkably evaluated dNTP pools in cell are probably related with the dNTP synthesis enzymes, especially the ribonucleotide reductase (RR) that catalyzes the formation of deoxyribonucleotides from ribonucleotides [44,45]. Mammalian RR is comprised of three subunits including RRM1, RRM2 and p53R2, which are expressed in a cell cycle-dependent manner [46]. In cycling cells, the RRM1 protein is metabolically stable throughout the cell cycle whereas the expression and degradation of RRM2 protein limit the S-phase-dependent activity of RR complex, leading to the high cellular dNTPs pools at S phase and low dNTPs pools outside S phase [47]. To further investigated whether the growth inhibitory activity of RDV resulted from the induction of RR, we determined the expression of RRM1, RRM2 and p53R2 by using western blot assay. From the results (Figure 1 C), there' no obvious difference of RRM1 level after the BEAS-2B cells incubated with RDV for 12, 24 and 48 h. However, the expression of RRM2 was significantly increased after 24 h exposure to RDV at 10 μM (P < 0.05), which caused by the S phase arrest. At the same time, the p53R2 level presented distinct down-regulated tendency (P < 0.05). In addition, there were also no changes in the levels of RRM2 and p53R2 after 48 h incubation with RDV. Taken together, it suggested that RDV inhibited the proliferation of BEAS-2B cells through the impact on RR expression.

Discussion
Although authorized for COVID-19 treatment, RDV still cannot meet the clinical needs due to the unsatisfied therapeutic outcome and high mortality [11]. It is therefore urgent to develop new treatment modalities with high efficacy, among which co-administration is a practical strategy. For this purpose, we investigated and found that RDV arrested BEAS-2B cells in the S phase, remarkably inhibited the biosynthesis of nascent RNA and DNA, perturbed RNs and dRNs pool size, and upregulated the expression of riboreductase. Further, based on the action mechanism of RDV, we tried to develop several substances to enhance RDV's antiviral efficacy via a combined strategy. As shown in this study, the RDV-induced decrease of CTP level probably represented a metabolic target for the treatment of COVID-19. The main source of CTP is from the conversion of UTP to CTP by CTP synthase in pyrimidine biosynthesis. Based on the biological property of CTP synthase, coadministration of remdesivir with cyclopentenyl cytosine, an analogue of CTP [48], would be an effective means of sensitizing human bronchial epithelium cells against virus infection. Another helpful strategy might be combination treatment with remdesivir and leflunomide, a clinically approved inhibitor of the de novo pyrimidine synthesis pathway, to reduce the CTP production as realized in the reversion of drug resistance of triple-negative breast cancer (TNBC) [32].

Cell culture and colorimetric MTT assay
BEAS-2B cells were purchased from the ATCC (Manassas, VA, United States). They were cultured in DMEM supplemented with 10% FBS, 100 U/mL penicillin-streptomycin in a humidified incubator at 37 °C and 5% CO2. Cell viability was determined by a modified colorimetric MTT assay [34]. Briefly, the logarithmic BEAS-2B cells at 6.0 × 10 4 cells/mL were seeded in a 96-well plate at 100 μL/well for 24 h at 37 °C, then treated with RDV at different concentrations (0-100 μM) for 24, 48 and 72 h, respectively. After the appropriate incubation time, 10 μL MTT solution (5 mg/mL) was added for another 4 h incubation, and 100 μL DMSO was added to dissolve formazan crystals for the further measurement at 570 nm using a microplate ultraviolet-visible spectrophotometer. Cell viability was calculated as follows: cell viability (%) = (absorbance of the test group/absorbance of the control group) × 100. The IC50 value was taken as the concentration that caused 50% inhibition of cell viability and was calculated by GraphPad Prism. (Graph-Pad Software, Inc., La Jolla, CA, USA).

EU and EdU detection using click chemistry
In order to label and visualize specifically newly synthesized DNA and RNA, the click chemistry method was used. The experiments were conducted based on previous publications with slight modification [35,36]. In short, BEAS-2B cells were grown in 6-well plate at 2.0 × 10 5 cells/well for 24 h and then incubated for 14 h with 1.0 mM EU or 10 μM EdU in the presence or absence of 10 μM RDV. After the labeling, cells were washed with PBS and fixed with 4% PFA for 30 min. The fixed cells were neutralized with 2 mg/mL glycine, rinsed with PBS and stained for 30 min at room temperature with a click reaction buffer including 100 mM Tris, 1 mM CuSO4, 10 μM Alexa594-azide and 100 mM ascorbic acid. After staining, cells were washed several times by using PBS with 0.5 mM EDTA, 1% TWEEN ® 20 and 0.1% Triton™ X-100, and then stained with 0.5 μg/mL DAPI for 30 min. Finally, the cells were imaged by IncuCyte ZOOM Live-Cell Analysis Platform.

Cell cycle analysis
BEAS-2B cells were seeded in 6-plate at 2.0 × 10 5 cells/well, cultured for 24 h and then treated with/without RDV for 12, 24 and 48 h, respectively. Then, the cells were harvested, re-suspended in ice-cold PBS and fixed with 70 % ethanol at -20 ℃ overnight. Subsequently, the fixed cells were washed again using ice-cold PBS and incubated with 500 μL PI containing 0.05 % RNase A for 30 min at room temperature in the dark circumstances. Finally, cell cycle distribution profile after the staining treatment was accessed by flow cytometry. The percentages of cells in G0/G1, S and G2/M phases were analyzed by using MODFIT software (Verity Sofware House, USA).

Western blot assay
BEAS-2B cells were treated with RDV at the indicated concentration for 12, 24 and 48 h, respectively. Then the cells were washed with cold PBS twice and lysed with RIPA buffer on ice. The lysates were centrifuged at 12000 g for 30 min at 4 ℃ in order to acquire the protein samples. The concentration of cellular total protein was measured by using the Bradford reagent at 595 nm according to the manufacturer's instructions. 30 μg protein samples were loaded on 10% SDS-PAGE gel and transferred to onto nitrocellulose membranes. The membranes were blocked with 5% skim milk for 1.5 h, followed by the incubation of primary antibodies diluted in Tris-buffered saline with Tween® 20 (TBST) buffer. (1:1000 for β-tubulin, RRM1, RRM2 and P53R2, Cell Signaling Technologies, Danvers, MA) overnight at 4 ℃. After that, the membranes were washed with TBST and incubated with secondary horseradish peroxidase-conjugated antibody anti-rabbit IgG (Cell Signaling Technologies, Inc. Danvers, MA, USA) for 1 h at room temperature. The immunoreactive protein bands were finally detected with an Amersham Imager 600 Western blotting system. Densitometry analysis of protein band was performed by Quantity One software (Version 4.6.2, Bio-Rad, USA).

Sample preparation and LC-MS/MS analysis
BEAS-2B cells were plated in 10 cm petri dishes with the density of 2.0 × 10 6 cells/dish, cultured with medium for 24 h and then treated with RDV for 12, 24 and 48 h, respectively. After that, the cells were re-suspended with ice-cold PBS. The number of cells was counted before centrifugation at 1200 rpm for 5 min, and the cell pellet was washed with 1.0 mL ice-cold PBS again and spun down at 1200 rpm for 5 min. Subsequently cell pellets were treated with 150 μL 80% methanol containing 4 μM AMP-13 C10, 15 N5 and 2 µM ATP-13 C10, 15 N5 as an internal standards (IS). The following sample preparation and the determination of endogenous RNs and dRNs were performed based on the method previously described [33]. The concentrations of cellular nucleotides were finally calculated according to dividing the absolute amount of each RN and dRN in each sample by the corresponding cell number.

Statistics analysis
Data analyses were performed using GraphPad Prism software and the values were expressed as mean ± standard deviation (SD) from three independent replicate experiments. The statistical significance of the comparison between control and treated groups was determined by Student's ttest or one-way ANOVA, which is indicated as *P < 0.05 and **P < 0.01.

Acknowledgments:
The authors also thank the Department of Science and Technology of Guangdong Province for the support of Guangdong-Hong Kong-Macao Joint Laboratory of Respiratory Infectious Disease.

Conflicts of Interest:
The authors declare no conflict of interest.