Lipid-Derived Electrophiles Mediate the Effects of Chemotherapeutic Topoisomerase I Poisons

Summary

Topoisomerase 1 (Top1) reversibly nicks chromosomal DNA to relax strain accumulated during transcription, replication, chromatin assembly and chromosome condensation. The Top1 poison camptothecin targets cancer cells by trapping the enzyme in the covalent complex, Top1_cc, tethered to cleaved DNA by a tyrosine-3′ phosphate bond. In vitro mechanistic studies point to interfacial inhibition, where camptothecin binding to the Top1-DNA interface stabilizes Top1_cc. Here we present a complementary covalent mechanism that is critical in vivo. We observed that camptothecins induce oxidative stress, leading to lipid peroxidation, lipid-derived electrophile accumulation and Top1 poisoning via covalent modification. The electrophile 4-hydroxy-2-nonenal (HNE) can induce Top1_cc on its own and forms a Michael adduct to a cysteine thiol in the Top1 active site, potentially blocking tyrosine dephosphorylation and 3′ DNA phosphate release. Thereby, camptothecins may leverage a physiological cysteine-based redox switch in Top1 to mediate their selective toxicity to rapidly proliferating cancer cells.

eTOC Blurb

The camptothecins are important chemotherapy agents that poison Top1 by trapping the protein covalently bound to cleaved DNA. Here, we show camptothecins and other Top1 poisons induce oxidative stress and lipid peroxidation, resulting in modification of Top1 by lipid-derived electrophiles. This covalent mechanism of action may determine cytotoxicity in vivo.

Graphical Abstract

Introduction

The type IB topoisomerase enzyme Top1 serves a critical role in the cell nucleus by nicking chromosomal DNA to relieve torsional strain that accumulates during transcription, replication, and chromatin assembly (Gentry et al., 2011, Pommier et al., 2016). Top1 activity also facilitates chromosome condensation and separation during mitosis (Delgado et al., 2018).

In the Top1 catalytic cycle, the enzyme binds to overwound DNA, cleaves a phosphodiester bond on the DNA strand, and transfers the 3′ phosphate to phosphorylate an active site tyrosine, leaving the 5′ end free. Once rotation relieves the strain, the covalently bound Top1 releases the phosphate and religates the DNA before dissociating (Pommier et al., 2010). Human Top1 protein consists of an N-terminal domain, a core domain, a linker domain, and a C-terminal domain. The core domain reorganizes to form a DNA binding pocket, guiding phosphodiester bond cleavage and transfer of the phosphate to tyrosine Y723 (Lesher et al., 2002). Cleavage and religation are modulated by conserved residues in the active site (Stewart et al., 1996) including cysteine 630 (C630, (Lesher et al., 2002)). While Top1 covalent complexes (Top1_cc) are normally transient, they occasionally fail to resolve on their own, as during repair of misincorporated genomic ribonucleotides (Cho et al., 2013). Then, the Top1_cc adduct is recognized and processed as DNA damage. In concert with other factors, tyrosyl-DNA phosphodiesterase (Tdp1, (Pommier et al., 2014)) can release the tethered peptide, facilitating religation of the cleaved DNA strand.

Given the greater potential to accumulate topological strain in proliferating cells, Top1 offers an attractive cancer target. While catalytic inhibitors that prevent DNA scission have not found significant clinical application, Top1 poisons that leave the enzyme covalently linked to cleaved DNA have long been important chemotherapy agents (Pommier et al., 2015) reflecting the toxicity of persistent Top1_cc to proliferating cancer cells (Nitiss and Wang, 1996, Pommier et al., 2010). The first candidate Top1 poison, camptothecin (CPT; (Beretta et al., 2013, Capranico et al., 2017)) is an alkaloid isolated from the Happy Tree Camptotheca acuminata. CPT displays problematic toxicity, instability, and poor bioavailability (Venditto and Simanek, 2010, Martino et al., 2017) but semisynthetic analogs derived from CPT such as topotecan (TPT) and irinotecan (IRN, a pro-drug for the active metabolite SN-38) that overcome these limitations have become core components of common regimens for lung, ovarian, and other solid tumors (Moukharskaya and Verschraegen, 2012, Delgado et al., 2018, Liang et al., 2019). Structural analysis of CPT and its analogs in ternary complexes with Top1 and DNA established interfacial inhibition as a mechanism of action, where drug binding stabilizes Top1_cc, delaying release of the 3′ DNA strand (Pommier et al., 2015). Nonetheless, the structural diversity of agents found to poison Top1 (Cinelli, 2019) is difficult to reconcile with the constraints of interfacial inhibition, suggesting alternative mechanisms may mediate Top1_cc formation in vivo.

Reactive oxygen species (ROS), including superoxide, hydroxyl radicals and peroxides, damage cellular macromolecules and organelles, compromising homeostasis, signaling, survival and proliferation (Poli et al., 2004, Ray et al., 2012). Cytochrome p450-dependent drug metabolism may produce ROS and multiple studies have described oxidative stress in cells treated with Top1 poisons, raising the question whether the resulting damage participates in their mechanism of action. While much attention goes to direct effects of ROS on DNA (Cadet and Davies, 2017), this ignores the toxic end-products of lipid peroxidation (LPO), particularly oxidative free radical damage to polyunsaturated fatty acids (Ayala et al., 2014). Among these are lipid-derived electrophiles (LDEs), the most toxic of which include α-, β-unsaturated carbonyl compounds such as 4-hydroxy-2-nonenal (HNE), 4-hydroperoxy-2-nonenal (HPNE), and 4-oxo-2-nonenal (ONE) (Sayre et al., 2006, Schaur et al., 2015, Sousa et al., 2017). LDEs readily modify protein lysine, histidine, and cysteine residues as well as nucleotide bases, and bifunctional LDEs such as HNE and ONE can participate in secondary reactions, inducing intra- or intermolecular protein and/or nucleic acid crosslinks (Sayre et al., 2006). A key cellular defense relies on LDE modification of cysteine thiols on Keap1, releasing transcription factor Nrf2 to induce expression of antioxidant-response element regulated genes such as glutathione synthesis enzymes (Dinkova-Kostova et al., 2017), superoxide dismutases (SOD; (Sheng et al., 2014)), NAD(P)H quinone dehydrogenase 1 (NQO1; (Ross and Siegel, 2010)), and aldehyde dehydrogenase (ALDH; (Singh et al., 2013)) enzymes. Strikingly, CPT may target Nrf2 to suppress gene activation and sensitize cancer cells to oxidative stress (Chen et al., 2017). When their levels overwhelm glutathione and other defenses, LDEs can induce apoptosis (Awasthi et al., 2003, Csala et al., 2015), cell cycle arrest (Chaudhary et al., 2013, Camarillo et al., 2016), and premature senescence (Zimniak, 2011, Shoeb et al., 2014).

Here, we show that camptothecin drugs induce ROS, LPO, and LDEs in cancer cells in vitro, while depleting glutathione and activating the cellular antioxidant response. Treating cells with LDE scavengers can block CPT from inducing Top1_cc while exposing cells to LDEs is sufficient to induce Top1_cc. Stabilizing Top1_cc by knocking out the repair enzyme Tdp1 can sensitize cells to LDEs. Suggesting a direct covalent mechanism, treating cells with CPT induces HNE adducts on Top1, while treating Top1 with HNE in vitro results in Michael addition at the active site cysteine C630. Drawing on the paradigm of reversible inactivation of tyrosine phosphatases by LDE modification of catalytic cysteines (Ostman et al., 2011), we propose that modification of C630 by LDEs may block Y723 dephosphorylation, trapping Top1_cc by preventing release of the bound 3′ DNA strand.

Results

Camptothecins induce proteomic signatures of redox stress in tumor cells

Reports over the past three decades, including recent studies by our group (Flor et al., 2016, Flor et al., 2017), have described increased cellular ROS and/or LPO in cells and animals treated with topoisomerase poisons. To directly examine oxidative stress induced by camptothecin drugs, we conducted label-free LC-MS/MS quantitative proteomics analysis of whole cell lysates from A549 human lung adenocarcinoma cells treated in triplicate with CPT or TPT for 72 h (Fig. 1a). Of 3117 total proteins identified in samples taken from cells treated with either drug (0.1 μM) or vehicle control (VEH; 0.1% DMSO), 1882 (60.3%) were identified across all samples (Fig. 1b). Of proteins significantly down-regulated or up-regulated (>1.5 fold) in CPT or TPT compared to control, most were similarly changed in the drug-treated samples (Fig. 1c).

Figure 1. Camptothecins induce proteomic signatures of redox stress in tumor cells.

A549 human lung adenocarcinoma cells were treated in triplicate with vehicle (VEH), camptothecin (CPT) or topotecan (TPT) at 0.1 μM for 72 h. Cells were then lysed and proteins separated by PAGE for proteomics analysis. (a) After PAGE, gel lanes were separated and segmented into three sections as shown. Each section was then digested and prepared for LC-MS/MS analysis. (b) Venn diagram indicates total proteins identified per treatment condition and shared between conditions. Of 3,117 proteins identified overall, 1882 were detected in all three conditions. (c) Heat map showing proteome-wide label-free quantitation of relative expression among the three conditions. Green, downregulated. Red, upregulated. Black, no significant change. Overall, patterns of expression were similar between CPT/VEH and TPT/VEH while the CPT/TPT comparison reveals few significant differences. (d) Volcano plots of proteins from the GO pathway “Oxidation-Reduction Process” (GO:0055114) quantified in the CPT/VEH and TPT/VEH comparisons, displayed as Log₂ fold change in expression vs. −Log(P-value) of confidence. Green, downregulated. Red, upregulated. Gray, insignificant fold change (≤ 1.5) and/or P-value (< 0.05).

Multiple proteins linked to cellular redox homeostasis were significantly changed in abundance by CPT or TPT treatment (Fig. 1d). Up-regulated proteins included aldehyde dehydrogenase 1A3 (ALDH1A3; CPT +2.09, TPT +1.24 Log₂ fold change vs. CTRL), superoxide dismutase 1 (SOD1; CPT +2.39, TPT +0.79), and all-trans-retinol 13,14 reductase (RETSAT; CPT + 3.91, TPT +0.93). Enrichment of multiple redox-related pathways was also observed (Table 1). Redox pathways included glutathione metabolism (KEGG pathway; P-value CPT = 8.5 X 10⁻³, TPT = 1.0 X 10⁻³), oxidation-reduction process (GO Biological Process; P-value CPT = 8.2 X 10⁻¹³, TPT = 1.6 X 10⁻⁷), oxidoreductase activity (GO Molecular Function; P-value CPT = 4.3 X 10⁻⁹, TPT = 1.3 X 10⁻⁵), and others. By comparison to oxidative stress, DNA damage response GO pathways displayed a surprising lack of enrichment in the CPT and TPT-treated samples (Table 1), potentially reflecting the 72 h time point.

Table 1.

Oxidative stress pathway activation, yet lack of DNA-damage related pathway activation, after 72 h treatment with camptothecins.

# proteins		1663		1690
		CPT vs. CTRL		TPT vs. CTRL
Database	Term	# genes	P-Value	# genes	P-Value
KEGG_PATHWAY	Glutathione metabolism	16	2.30E-03	16	2.20E-03
GO_BP_DIRECT	Oxidation-reduction process	109	5.70E-11	118	8.60E-15
GO_BP_DIRECT	Cell redox homeostasis	23	2.40E-06	26	2.10E-08
GO_BP_DIRECT	Response to oxidative stress	19	1.90E-02	21	3.40E-03
GO_MF_DIRECT	Oxidoreductase activity	38	8.20E-05	44	2.80E-07
UP_KEYWORDS	Oxidoreductase	108	1.20E-15	120	8.20E-22
UP_KEYWORDS	Redox-active center	17	9.70E-07	17	7.80E-07
KEGG_PATHWAY	DNA replication	27	1.60E-08	15	1.20E-04
KEGG_PATHWAY	Base excision repair	-	ns	-	ns
KEGG_PATHWAY	Mismatch repair	-	ns	-	ns
KEGG_PATHWAY	Non-homologous end joining	-	ns	-	ns
GO_BP_DIRECT	Nucleotide-excision repair	9	9.90E-04	8	4.40E-03
GO_BP_DIRECT	DNA damage checkpoint	-	ns	-	ns
GO_BP_DIRECT	DNA damage response	-	ns	-	ns
GO_BP_DIRECT	DNA repair	-	ns	-	ns
GO_MF_DIRECT	Nucleotide binding	76	1.80E-11	77	3.70E-12
GO_MF_DIRECT	DNA topoisomerase activity	-	ns	-	ns
GO_MF_DIRECT	DNA topoisomerase binding	-	ns	-	ns
UP_KEYWORDS	Nucleus	589	1.30E-19	578	5.50E-19
UP_KEYWORDS	DNA damage	-	ns	-	ns
UP_KEYWORDS	DNA repair	-	ns	-	ns

To confirm the persistent oxidative stress response, we examined protein expression of six antioxidant proteins identified by proteomics as upregulated in both CPT and TPT treated cells (Fig. S1). Western blotting demonstrated increased levels of ALDH1A1, ALDH3A1, SOD1, SOD2, and NQO1 in lysates from cells treated for 72 h with CPT or TPT. Similar results were observed after 72 h treatment with the camptothecin 7-ethyl-10-hydroxy-camptothecin (SN-38) as well as the non-camptothecin Genz-644282 (GENZ).

Top1 poisons induce lipid peroxidation

To extend our study to a larger panel of Top1 targeting drugs, A549 cells were treated in triplicate with a panel of Top1 poisons, including CPT, TPT, 9-amino-camptothecin (9-CPT), (S)-10-hydroxy-camptothecin (10-CPT), the prodrug IRN, its active metabolite SN-38, the non-camptothecin GENZ, the non-canonical inhibitor SW044248 (SW), along with the oxidizing agent cumene hydroperoxide (CHP), each at 0.1 μM for 72 h. The cells were then probed with the cell-permeable fluorescent ROS sensor CellROX Deep Red and imaged by fluorescent microscopy (Fig. 2a) or analyzed by flow cytometry (Fig. S2a). Cells treated with the active camptothecins CPT, TPT, 9-CPT, and 10-CPT, displayed elevated ROS and decreased proliferation. Notably, the camptothecin prodrug IRN was similar to vehicle while its active metabolite SN-38 was comparable to CPT or TPT. Like the active camptothecins, the non-camptothecin GENZ and non-canonical inhibitor SW also induced ROS after 72 h. CHP displayed no persistent ROS or impact on cell proliferation at this concentration after 72 h.

Figure 2. Top1 poisons induce lipid peroxidation.

A549 lung adenocarcinoma cells were treated with vehicle, camptothecins CPT, TPT, 9-CPT, or 10-CPT, the prodrug IRN, its active metabolite SN-38, non-camptothecin GENZ, non-canonical inhibitor SW, or LPO inducer CHP, each at 0.1 μM for 72 h. (a) Active Top1 poisons induce ROS. Live cell imaging of CellROX Deep Red (red) and brightfield (grey). Scale bar = 10 μM. (b) Induction of ROS and LPO by Top1 poisons are linked. Plot displays mean fluorescence intensity for CellROX analysis of ROS (MFI CellROX, mean ± SD) and the fraction of cells probed with BODIPY-C₁₁ LPO sensor that display high fluorescence (% LPO⁺ cells, mean ± SD) for each Top1 poison (R² = 0.85). (c) Top1 poisons induce cellular accumulation of HNE. Competitive ELISA of cell lysate with anti-HNE-adduct antibody (HNE adducts (μM), mean ± SD, n = 2). *, P ≤ 0.05 (Kruskal-Wallis). (d) Top1 poisons deplete glutathione (GSH). Assay of GSH in deproteinated cell lysate (GSH (μM), mean ± SD, n = 2). *, P ≤ 0.05 (Kruskal-Wallis). (e) Top1 poisons induce ALDH enzyme activity. AldeRed flow cytometry ALDH activity assay (610 nm). Percent ALDH^HI cells shown. See also Figures S1, S2, and S5.

To assess levels of LPO, the cell-permeable probe BODIPY-C₁₁ was added to A549 cells treated as above and fluorescence analyzed by flow cytometry, setting a gate so that < 10% of untreated control cells fell into LPO⁺ (Fig. S2b). When median fluorescence intensity of CellROX Deep Red as determined by flow cytometry (MFI CellROX, Fig. S2a) was plotted vs. the percentage of LPO⁺ cells for each Top1 poison, linear regression analysis revealed a moderately strong correlation (R² = 0.85) between ROS and LPO (Fig. 2b). Confirming increased LPO and production of LDEs, cells treated with 0.1 μM CPT, TPT, SN-38 or GENZ for 72 h revealed 2.1 to 24.1 fold increases in HNE protein adducts over untreated control (Fig. 2c), and 1.8 to 4.0 fold depletion of GSH (Fig. 2d). A flow cytometry assay to detect aldehyde dehydrogenase (ALDH) activity with the cell-permeable probe AldeRed 610 revealed increased ALDH activity in cells treated with each Top1 poison (Fig. 2e). Percentage of ALDH^HI cells was < 5% in vehicle-only treated cells, vs. 53.5% for CPT, 82.5% for TPT, 78.2% for SN-38, and 52.8% for GENZ.

Camptothecins induce Top1 modification by HNE

The pattern of LPO induced by Top1 poisons raised the question whether the resulting lipid derived aldehydes might form adducts to Top1 during drug treatment. After treating A549 cells in triplicate with 0.1 μM CPT for 72 h, we used proximity ligation assay (PLA, Fig. 3) to detect HNE-protein adducts on Top1 protein in situ. Probing the cells with rabbit anti-Top1 and goat anti-HNE protein adduct antibodies and detection with fluorescent anti-species secondary antibodies revealed distinct patterns in untreated cells and increased staining in treated cells (Fig. S3). Relocalization of Top1 to the cytoplasm after CPT treatment has been reported previously (Sonavane et al., 2018). To evaluate HNE modification of Top1, cells were stained with both anti-Top1 and anti-HNE protein adduct antibodies and then with anti-species secondary antibodies labeled with PLA oligonucleotides. After ligation and rolling circle amplification (RCA), a far-red fluorescent hybridization probe was used to detect the RCA product, marking sites of HNE-protein adduct and Top1 proximity. Compared to PLA analysis of cells treated with vehicle (Fig. 3a) or mock PLA of CPT-treated cells without primary antibodies (Fig. 3b), PLA of CPT-treated cells revealed HNE modification of Top1 protein (Fig. 3c). Enlarged images (Fig. 3d) show punctate Top1-HNE PLA signals both in the cytoplasm and the nucleus, suggestive of HNE-modified Top1_cc.

Figure 3. CPT and TPT induce HNE modification of Top1 in treated cells.

After treating A549 cells with 0.1 μM CPT for 72 h, cells were analyzed by in situ rolling circle amplification (RCA) proximity ligation assay (PLA) after dual staining with anti-Top1 and anti-HNE-adduct antibodies to detect HNE modification of Top1. Blue, DAPI stained nuclei. Red, positive signal for PLA reaction. (a) Vehicle-treated cells (VEH) display no background while a strong PLA signal is detected in CPT treated cells, indicating HNE modification of Top1 protein. (b) Image zoom showing punctate PLA staining in nucleus and cytoplasm of CPT treated cells. Scale bars, 10 μm. Samples were analyzed in biological triplicate and images taken in three regions per microwell; representative cells are shown. (c) IP-Western analysis confirms HNE modification of Top1 after treatment with TPT. Anti-Top1 Western analysis of cell lysates after 72 h treatment with vehicle or 0.1 μM TPT demonstrates equivalent Top1 levels (left). Immunoprecipitation with anti-HNE-protein adduct antibody, then anti-Top1 Western analysis indicates increased HNE-modified Top1 in TPT-treated lysate over VEH control (right). See also Figures S3 and S4.

To further confirm HNE modification of Top1, HNE-protein adducts were immunoprecipitated from cell lysates following treatment with vehicle or 0.1 μM TPT for 72 h and analyzed by Western blotting for Top1 (Fig. 3c). Although total Top1 was similar in each lysate, anti-HNE adduct antibody immunoprecipitated more Top1 from TPT-treated cells compared to vehicle-only control. When anti-HNE adduct antibody immunoprecipitation was performed over a time course (Fig. S4a), Top1-HNE appeared to peak at 72 h. Western blotting for antioxidant proteins ALDH1A3 and SOD1 (Fig. S4b) also displayed a peak at 72 h.

Delayed elevation of ROS after treatment with topoisomerase poisons could be due to the induction of apoptosis or onset of accelerated senescence. To investigate this, we performed a triple-staining flow cytometry assay for cell viability (Calcein Violet 450 AM, violet fluorescence), apoptosis (YO-PRO-1 stain, green fluorescence), and senescence (DDAO-galactoside, red fluorescence). Representative data demonstrate that after 72 h treatment with TPT, most A549 cells remain viable but many shift into senescence, with a smaller fraction entering apoptosis (Fig. S5a–b). A similar pattern to TPT was observed for each of the other active camptothecins, CPT, 9-CPT, 10-CPT, SN-38, and for GENZ (S5c). The inactive pro-drug IRN was similar to vehicle while the non-canonical poison SW yielded a smaller shift toward senescence. In each case, the fraction of senescent cells exceeded the dead and apoptotic cells.

LDEs are necessary and sufficient to induce Top1_cc

These results raised the possibility that cellular production of HNE or other LDEs and resulting modification of Top1 may mediate the effects of CPT or other Top1 poisons. If so, a reasonable expectation is that treating cells with HNE might recapitulate the effects of CPT. Indeed, indirect immunofluorescence detection of Top1_cc with an adduct-specific monoclonal antibody (Patel et al., 2016) revealed similar levels of nuclear staining after treatment with 0.2 mM CPT or HNE for 1 h (Fig. 4a). Using the Top1_cc adduct-specific antibody for flow cytometry (Fig. 4b), we examined formation of Top1_cc in A549 cells treated for 1 h with 0.2 mM CPT, HNE or the related 9-carbon aldehyde LPO end-products HPNE or ONE. Flow cytometry analysis revealed that each LDE increased the fraction of Top1_cc⁺ cells above vehicle control (HNE, 35.4%; HPNE, 26.7%; ONE, 44.5%), albeit less than CPT (62.7%).

Figure 4. LDEs are sufficient to induce Top1_cc.

(a) Detection of Top1_cc induced after treating A549 cells with 0.2 mM CPT or HNE for 1 h. Immunofluorescence with Top1_cc adduct-specific antibody (magenta) and DAPI counterstain (blue). Scale bar, 20 μm. Insets: Zoomed image to show punctate nuclear staining. (b) Aldehyde end-products of LPO induce Top1_cc in A549 cells. Flow cytometry using anti Top1_cc adduct-specific antibody in cells treated with CPT, HNE, HPNE, or ONE (red filled histograms) compared to VEH control (gray dashed lines). Shown are representative data from three replicates per condition. Percent Top1_cc⁺ cells indicated for each treatment. (c) Flow cytometry detection of Top1_cc in cells treated for 1 h with lipid antioxidants CYS2, GSH, or BHT prior to CPT for 1 h. Each antioxidant reduced formation of Top1_cc by CPT. (d) CPT and HNE are comparable Top1 poisons. Slot-blot RADAR assay of Top1 immunoreactivity coprecipitating with A549 cell genomic DNA indicates similar Top1_cc formation after treatment with 20 μM to 0.4 mM CPT or HNE for 30 min. (e) HNE inhibits relaxation activity of recombinant Top1. SC, supercoiled DNA marker lane. REL, relaxed DNA marker lane. Standard reaction conditions or adding vehicle to Top1 prior to adding plasmid DNA yield complete relaxation within 30 min. Treating Top1 with 0.1 to 1.0 mM HNE for 5 min prior to addition to supercoiled plasmid DNA results in incomplete relaxation. See also Figure S6.

To investigate whether LPO may mediate the activity of CPT, we pretreated cells with antioxidants before exposure to CPT. A549 cells were treated for 1 h with reduced glutathione (GSH), the GSH precursor cystine (CYS2), or butylated hydroxytoluene (BHT). CPT was then added for 1 h and induction of Top1_cc was determined by flow cytometry (Fig. 4c). Each lipid antioxidant suppressed formation of Top1_cc by CPT, reducing Top1_cc⁺ cells to 25.7% for GSH, 8.5% for CYS2, or 31.1% for BHT. These results implicate LPO and LDEs in the mechanism of action of this canonical Top1 poison.

Given concerns about the Top1_cc adduct-specific antibody, as a complementary test, we performed a standard RADAR assay (Kiianitsa and Maizels, 2013) to directly detect formation of Top1_cc (Fig. 4d). Here, probing for Top1 adducts bound to ethanol-precipitated genomic DNA isolated from A549 cells treated for 1 h with CPT or HNE over the range of concentrations from 20 to 400 μM, we observed comparable levels of Top1 immunoreactivity for the two agents. A similar RADAR assay using SJCRH30 cells (Fig. S6a) demonstrated a detectable increase in Top1_cc after 4 h treatment with as little as 10 μM HNE, followed by 24 h of recovery (Fig. S6b).

A critical question remained whether the modification of Top1 by HNE might be sufficient to induce formation of poisoned Top1_cc irreversibly bound to chromosomal DNA. Thus, we performed a conventional topoisomerase enzyme activity assay using recombinant human Top1 and a supercoiled plasmid DNA substrate (Fig. 4d). 10 U of recombinant Top1 was sufficient to completely relax the supercoiled substrate within 30 min at 37 °C under standard conditions. Treating the Top1 with vehicle did not affect formation of relaxed plasmid, but exposure to 0.1 to 1.0 mM HNE for 35 min substantially inhibited Top1 activity, as evidenced by the persistence of supercoiled substrate in each reaction. Similarly, treating Top1 with 0.2 mM HPNE or ONE almost completely inhibited activity while glutathione conjugated-HNE had no inhibitory effect (Fig. S6c).

Tdp1 regulates cellular sensitivity to HNE

Tyrosyl-DNA phosphodiesterase (Tdp1) limits the toxicity of Top1 poisons by releasing the 3′ DNA phosphate trapped by Top1_cc, allowing religation to restore the phosphodiester bond (Pommier et al., 2014). Treating an HEK-293 Tdp1 CRISPR knockout (TDP1KO) cell line (Li et al., 2017c) with LDEs revealed moderate sensitization compared to HEK-293 Tdp1 wild type (WT) (Fig. 5a–b). HNE and ONE were selectively toxic to the Tdp1 deficient cells, with significant differences in toxicities in the range from 4 to 32 μM (Mann-Whitney test, P ≤ 0.05). The aldehyde scavenger hydralazine HCl (HYD) suppressed HNE toxicity in TDP1KO cells (Fig. 5c) while glutathione-conjugated HNE (HNE-GSH) displayed no toxicity (Fig. 5d).

Figure 5. Top1_cc resolving enzyme Tdp1 regulates cellular sensitivity to LDEs.

HEK-293 cells bearing a Tdp1 CRISPR knockout (TDP1KO) display enhanced sensitivity to both (a) HNE and (b) ONE. (c) Sequestering aldehydes with hydralazine (HYD) suppresses HNE toxicity. (d) Glutathione conjugation of HNE (HNE-GSH) eliminates the toxic effects. WST-1 proliferation assay, mean ± SD (n=3). *, P ≤ 0.05 (Mann-Whitney U test).

Top1 active site residue C630 is modified by Michael addition of HNE

Top1 encodes eight cysteine residues, all of which are located within the conserved core and C-terminal domains (Montaudon et al., 2007), and several of which are predicted to be solvent-accessible and remain as thiols in the highly reducing intracellular milieu. To detect cysteine thiols and their modification by HNE, 100 U of recombinant human Top1 was incubated with 0 to 1000 μM HNE for 30 min and the fluorogenic reporter Thiol Reactive Probe IV was added (Fig. S6d). Treatment with HNE decreased probe fluorescence, indicating covalent modification of thiols on the Top1 protein.

To map sites of HNE modification, recombinant Top1 was treated with 0.2 mM HNE, digested with trypsin, and the peptides subjected to LC-MS/MS analysis. HNE modifications can be detected by mass spectrometry via a characteristic mass shift of +156 Da, and can also be detected in a reduced state (+158 Da) or dehydrated form (+138 Da). For the recombinant Top1 sample treated with HNE, the characteristic shift upon Michael addition of HNE was observed on peptides containing the lysine residues K216, K299, K326, and K558 along with peptides containing cysteine C630 (Fig. 6a, b). Statistical analysis by A-score indicated > 99% probability of accuracy for each of the identified HNE modification sites. To confirm HNE modification of Top1 in vivo, A549 cells were treated with 0.1 μM CPT or TPT for 72 h and then lysed to obtain a whole cell extract. After separation on PAGE and in-gel trypsinization, peptides were analyzed by LC-MS/MS using an inclusion list to enhance selection of predicted HNE-modified Top1 peptide ions for fragmentation. Thereby, we identified Top1 peptides with HNE modification of C630 from lysates of cells treated with CPT or TPT (Fig. S7).

Figure 6. HNE modifies Top1 by Michael addition to C630, potentially stabilizing Top1_cc and mediating the effects of camptothecins.

(a) Proteomic analysis of recombinant Top1 after treatment with 0.2 mM HNE identifies an adduct on C630. Shown is MS/MS fragmentation spectrum of modified Top1 625-634 tryptic peptide AVAILC^HNENHQR marked to indicate the 641.33 m/z parent ion (M+2H+HNE) and 156.1 Da mass shift after neutral loss for y₅ to y₁₀ ions. (b) C630 is found in Subdomain Ill of the CAT domain within the DNA-binding active site of human Top1. HNE adducts to K216, K299, K326, K558 and C630 were each statistically significant as determined by A-score. (c) Human Top1_cc atomic model (PDB 1A36, left) indicating CAP (blue), C-Term (orange), CAT (red), Linker Domain (grey) and DNA (green scissile strand/pink complement). Zoom-in view of the active site (white) on the right indicates C630 (cyan) along with catalytic residues R488 (red), R590 (violet), H632 (mauve) and 3′ phospho-acceptor Y723 (yellow). Note close proximity of C630 thiol to the phosphotyrosine in the bound conformation. See also Figure S7. (d) In its normal function to relieve torsional strain, Top1 associates with and covalently binds to overwound DNA. The catalytic tyrosine Y723 attacks a 3′ phosphate, creating a transient Top1 covalent complex, Top1_cc. The free 5′ end can freely rotate to relieve strain. The transient DNA nick is then religated and Top1 is released. In the interfacial inhibitor model, CPT binds between Top1_cc and the DNA to sterically prevent religation. (e) An alternative, covalent mechanism of action for Top1 poisoning by CPT, mediated by oxidative stress, LPO, and LDEs. Treatment with CPT or its analogs can increase LDEs, which covalently modify Top1 at cysteine 630 (C630). This does not prevent formation of TOP1_cc, but blocks dephosphorylation and religation, trapping Top1_cc crosslinked to nicked DNA. If the damage overwhelms DNA repair, the resulting inhibition of transcription, replication and mitosis drives cells toward apoptosis or senescence.

Mapping C630 in the human Top1 catalytic domain crystal structure (Lesher et al., 2002) reveals proximity to Y723, the catalytic tyrosine that covalently binds to the 3′ DNA-phosphate in Top1_cc (Fig. 6c). The structural model points to several potential mechanisms. For example, thiol oxidation by HNE may prevent C630 from serving a role similar to a tyrosine phosphatase active site cysteine in facilitating cleavage of the 3′ DNA-phosphotyrosyl bond. Alternatively, following Michael addition to C630, the HNE carbonyl may react with other active site nucleophiles and the resulting crosslinks may constrain 3′ DNA phosphate release.

Discussion

The prevailing mechanism of action for topoisomerase poisons such as CPT is interfacial inhibition (Pommier and Marchand, 2005). For Top1, this elegant model has been validated by biochemical analysis and crystal structures that reveal camptothecin drugs can intercalate at the site of DNA cleavage, trapping Top1 in the otherwise transient covalent complex formed upon DNA strand cleavage and transfer of the 3′ phosphate to Y723 (Pommier and Marchand, 2005, Thomas and Pommier, 2019). Beyond interfacial inhibition, a substantial literature also exists that describes a diverse collection of natural and synthetic agents that can covalently modify topoisomerases (particularly Top2) in a redox-dependent manner (e.g. (Wang et al., 2001, Bandele and Osheroff, 2008, Ketron and Osheroff, 2014) and others). Unlike interfacial poisons, covalent poisons generally act outside of the enzyme active site (Dalvie and Di, 2019), can be abrogated by reducing agents (Jacob et al., 2011), and cause adducts that are typically irreversible (Vann et al., 2015).

For the camptothecin drugs, mechanism of action may be more complex than just interfacial inhibition, given their divergent structure-activity relationships and IC₅₀’s when examined in vitro versus in cells or tumor models. A broad literature over the past three decades has described oxidative stress in cells treated with camptothecins, with many authors finding that this “off-target” activity is critical to cytotoxicity. A common feature of Top1 poisons and their metabolites are redox active moieties including quinones, napthoquinones, quinolines, and quinolones (Elsea et al., 1997, Byler et al., 2009, Kennedy et al., 2011, Smith et al., 2014). Drug metabolism can lead to quinone and semi-quinone radicals that generate ROS including hydrogen peroxide (Watanabe et al., 2004) and superoxide (Song and Buettner, 2010). NADPH oxidases may be induced (Hiraoka et al., 1998). Once metabolically activated, drugs including camptothecins can conjugate to glutathione (Gamcsik et al., 2001), further continuing to increase oxidative stress. Other lines of evidence also point to a role for oxidative stress in mediating toxicity in cells. CPT suppresses expression and transcription of the antioxidant enzyme Nrf2 in several cancer cell lines (Chen et al., 2017). TPT alters gene expression of numerous antioxidant enzymes in cancer cells, including glutathione reductases and glutathione peroxidases (Sinha et al., 2020), while increasing enzymatic activity of glutathione peroxidase, superoxide dismutase, and catalase (Timur et al., 2005). The GI₅₀ of CPT was found to be positively correlated with glutathione-S-transferase activity in a screen of cancer cell lines (Bracht et al., 2007). Li et al. (Li et al., 2017b, Li et al., 2017a) have even proposed that in vivo efficacy of camptothecins may be independent of their effects on Top1.

This study may help reconcile these apparently conflicting views. Here, we have provided evidence of an indirect mechanism of action for camptothecins that implicates oxidative stress in poisoning Top1. We find that camptothecins induce LPO, driving generation of LDEs such as HNE that can covalently modify Top1 active site cysteine C630 and prevent resolution of the covalent complex, Top1_cc (Fig. 6d–e). Thereby, the Top1 poisons can indirectly mediate both their beneficial and adverse effects.

In LPO, a chain reaction of peroxides with polyunsaturated fatty acids yields diverse LDEs as end-products that may display one or two terminal aldehyde groups linked to a fatty acid hydrocarbon chain of 3 to 20 carbons (Gueraud et al., 2010) and serve as key toxic mediators of oxidative stress (Gaschler and Stockwell, 2017). Unsaturated LDEs such as HNE and ONE that bear a reactive hydroxyl or carbonyl group can rearrange after Michael addition to a cellular nucleophile and/or react again, forming intra- or intermolecular crosslinks (Sousa et al., 2017). Although proteins modified by LDEs have been surveyed (e.g. (Castro et al., 2017, Barrera et al., 2015)), these studies had not previously identified Top1 as a target. In turn, while LPO has been observed in cells treated with Top1 poisons (Yokoyama et al 2017, Timur et al., 2005, Singh et al., 2015), links to formation of Top1_cc were not explored.

Conserved cysteine residues located in active sites or situated nearby are a common feature of diverse enzymes (Verma et al., 2016, Rhee and Kil, 2017) where they may serve a direct role via the thiol serving as a nucleophile, which can affect catalysis or allosteric regulation (Sousa et al., 2019). Although the reducing environment of the intracellular milieu generally protects exposed cysteine thiols, structural context, such as hydrogen bonds, can lower the pKa and favor oxidation (Poole, 2015) and reactions with cellular electrophiles can directly inhibit catalysis or trap enzymes in bound states (Vazquez-Torres, 2012). However, thiol modification is often reversible and the resulting regulation considered redox sensing rather than protein damage (Vazquez-Torres, 2012). For example, reversible inhibition of tyrosine phosphatases upon modification of catalytic cysteines (Ostman et al., 2011) may mediate increased tyrosine phosphorylation in response to oxidative stress.

A striking feature of human Top1 is the number of solvent-accessible, reactive cysteine thiols. Of eight cysteine residues in the catalytic domain, at least C504, C505, and C630 are subject to covalent modification (Montaudon et al., 2007, Sharma et al., 2015). N-ethyl maleimide and phenylarsine oxide modify C504 and C505 to inhibit DNA cleavage (Montaudon et al., 2007). The Top1 active site residue C630 is subject to modification by nitric oxide (Sharma et al., 2015), suggesting a role in redox sensing. In tracing a pathway from camptothecins via LDEs to Top1_cc, we found that HNE acts as a Top1 poison both in cells and in vitro. Mass spectrometry analysis detected Michael addition of HNE to C630, suggesting a role in to trapping Top1_cc. Potentially, crosslinks between C630 and active site residues or pyrimidine bases may contribute to stabilizing Top1_cc. Finally, while high doses of Top1 poisons such as the camptothecins or LDEs may be toxic, increased persistence of Top1_cc at lower levels of oxidative stress may serve as a redox sensor, helping activate DNA damage surveillance as part of the antioxidant response.

STAR Methods

RESOURCE AVAILABILITY

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Amy Flor (flora@uchicago.edu).

Materials Availability

This study did not generate new unique reagents.

Data and Code Availability

The proteomics datasets generated during this study are available at PRIDE (Accession # PDX015181). Original data for figures in the paper are available at Mendeley (http://dx.doi.org/10.17632/gf8pnt826y.1).

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Cell lines

The authenticated human lung adenocarcinoma cell line A549 (male; RRID: CVCL_0023) was purchased from American Type Culture Collection (ATCC) and maintained at low passage in culture. Culture medium consisted of Dulbecco’s Minimal Essential Medium (DMEM, 4.5 g/L glucose) supplemented with 4 mM L-glutamine (Life Technologies), 1X penicillin/streptomycin (Life Technologies), and 10% fetal bovine serum (FBS, Thomas Scientific). Cells were plated at 10 X 10³ per cm² for each experiment and incubated for 18 - 24 h in a humidified 5% CO₂ environment at 37 °C prior to treatment in situ or harvest by monolayer dissociation using a cell scraper. TDP1KO cells were derived from HEK-293 kidney cells obtained from ATCC (gender unknown (embryonic); RRID: CVCL_0063). SJCRH30 cells verified by STR analysis (male; RRID: CVCL_0045) were used in certain RADAR assays. HEK-293 and SJCRH30 cells were also cultured in DMEM supplemented with 10% FBS and maintained in a humidified 37°C incubator with 5% CO₂.

METHOD DETAILS

Proteomics analysis

For LC-MS/MS analysis of Top1 poison effects, A549 cells plated in triplicate at 10 X 10³ cells/cm² overnight were treated with CPT (0.1 μM), TPT (0.1 μM), or DMSO (0.05%, vehicle) for 72 h in complete culture medium in a humidified 5% CO₂ incubator at 37 °C. Cell monolayers were rinsed once with PBS to remove dead cells, harvested by scraping, pelleted by centrifugation, and then lysed in 100 μL of an ice-cold lysis buffer containing protease inhibitors (Halt, Thermo), EDTA (5 mM, Thermo), and DNAse I (1 μg/mL, Sigma-Aldrich) for 20 min on ice with frequent vortexing. Lysates were clarified by centrifugation at 16500 x g (4 °C, 15 min). The supernatant was transferred to new tubes and used for proteomics analysis.

In experiments to identify putative HNE reactive amino acid residues, recombinant human Top1 (1 μg, Topogen) was incubated with HNE (200 μM, Cayman Chemical) in 100 μL of PBS at 37 °C for 30 min to HNE modify the protein.

Two technical replicates from all protein samples were then boiled at 95 °C in Laemmli sample buffer for 5 min and then subjected to electrophoresis (SDS-PAGE) using 10 μg of protein per lane as determined by BCA protein assay (Thermo). Proteins in gel were visualized with a gel stain (Imperial, Thermo) and gel regions of interest were excised by sterile razor blade. Gel sections were then chopped into ~ 1 mm³ pieces, washed in dH₂O and incubated in 100 mM NH₄HCO₃ pH 7.5 in 50% acetonitrile. A reduction step was performed by addition of 100 μL NH₄HCO₃ (50 mM, pH 7.5) and 10 μL of tris (2-carboxyethyl) phosphine hydrochloride (TCEP HCl, 200 mM) at 37 °C for 30 min. Gel sections were washed in purified water, then acetonitrile, and vacuum dried. Trypsin digestion was carried out overnight at 37 °C with 1:100 enzyme–protein ratio of sequencing grade-modified trypsin (Promega) in NH₄HCO₃ (50 mM, pH 7.5) and CaCl₂ (200 mM). Peptides were extracted with 5% formic acid and vacuum dried.

For HPLC, all samples were resuspended in HPLC-grade water (Burdick and Jackson) containing 0.2% formic acid (Fluka), 0.1% TFA (Thermo), and 0.002% Zwittergent 3-16 (EMD Calbiochem). The peptide samples were loaded to a 0.25 μl C₈ OptiPak trapping cartridge custom-packed with Michrom Magic C₈ (Optimize Technologies), washed, then switched in-line with a 20 cm X 75 μm C₁₈ packed spray tip nano column packed with Michrom Magic C₁₈AQ, for a two-step gradient. Mobile phase A was water/acetonitrile/formic acid (98/2/0.2) and mobile phase B was acetonitrile/isopropanol/water/formic acid (80/10/10/0.2). Using a flow rate of 350 nL/min, a 90 min, two-step LC gradient was run from 5% B to 50% B in 60 min, followed by 50% – 95% B over the next 10 min, hold 10 min at 95% B, back to starting conditions and re-equilibrated.

Electrospray tandem mass spectrometry (LC-MS/MS) was performed with a Q-Exactive Orbitrap mass spectrometer (Thermo), using a 70,000 RP survey scan in profile mode, m/z 340–2000 Da, with lockmasses, followed by 20 MS/MS HCD fragmentation scans at 17,500 resolution on doubly and triply charged precursors. Single charged ions were excluded, and ions selected for MS/MS were placed on an exclusion list for 60 sec. For Top1-HNE experiments, an inclusion list was generated with in-house software (Mayo Clinic Proteomics Core), and used with expected tryptic Top1 peptide ions, either unmodified or with potential 4-HNE amino acid modifications (C,H,K) during the LC-MS/MS runs.

All MS/MS samples were analyzed using MaxQuant (Max Planck Institute of Biochemistry, version 1.6.5, (Tyanova et al., 2015)) and X! Tandem (The GPM, version CYCLONE (Craig and Beavis, 2004)) and post-search processed with Perseus (Max Planck Institute of Biochemistry, version 1.6.7, (Tyanova et al., 2016)). MaxQuant and X! Tandem were set up to search the 181010_Top1_Human.fasta database with contaminants assuming the digestion enzyme strict trypsin with max miss cleavage set to 2. MaxQuant and X! Tandem were searched with a fragment ion mass tolerance of 20 PPM and a parent ion tolerance of 20 PPM.

Using Scaffold (Proteome Software Inc.), peptide identifications were accepted if they could be established at greater than 89.0% probability to achieve an FDR less than 1.0% by the Scaffold local FDR algorithm. Protein identifications were accepted if they could be established at greater than 99.0% probability to achieve an FDR less than 1.0% and contained at least 5 identified peptides. Protein probabilities were assigned by the Protein Prophet algorithm (Nesvizhskii et al., 2003). Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony. Ratios were reported based on the LFQ Intensities of protein peak areas determined by MaxQuant and reported in a proteinGroups.txt file. The proteingroups.txt file was processed in Perseus. Proteins were removed from this results file if they were flagged by MaxQuant as “Contaminants”, “Reverse” or “Only identified by site”. LFQ peak intensities were log₂ transformed, median normalized in CPT or TPT samples vs vehicle-only control, and missing values were imputed via default settings in Perseus. Gene Ontology analysis was performed with a significance cutoff of ± 1.5-fold via the DAVID bioinformatics resource (NIH; https://david.ncifcrf.gov/, (Huang Da et al., 2009)).

For Top1-HNE data analysis, HNE-Delta:H(2)O (C,H,K), HNE of (C,H,K), and HNE+Delta:H(2) (C,H,K) were specified in MaxQuant and X!Tandem as variable modifications, along with deamidation (NQ), oxidation (M), formylation (N-term), acetylation (N-term), carbamidomethylation (C), Glu->pyro-Glu of the N-terminus, ammonia-loss of the N-terminus, and gln->pyro-Glu of the N-terminus. LFQ quantification was conducted with a minimum of 1 high-confidence peptide. 1% false discovery rate (FDR) cutoff was selected for peptide, protein, and site identifications. Scaffold PTM (Proteome Software) was used to annotate PTM sites derived from MS/MS sequencing results obtained using Scaffold, and all HNE modification sites were confirmed by manual validation.

The full proteomic data sets (Top1 poisons systems biology) and (Top1-HNE modification sites) were uploaded to the ProteomeXchange repository (https://www.ebi.ac.uk/pride/archive/) with identifier PXD015181.

ROS and LPO assays

ROS assays were conducted using CellROX Deep Red probe (Life Technologies) according to manufacturer instructions. Briefly, A549 cells were incubated in triplicate with treatment compounds at 0.1 μM for 72 h in complete medium. At 72 h, cell monolayers were washed with sterile phosphate buffered saline (PBS) and media was replaced, followed by incubation with 5 μM of CellROX Deep Red probe for 30 min at 37 °C. For microscopy, cell monolayers were again washed with PBS and then a live cell imaging buffer (Life Technologies) was added to preserve sample viability during imaging. Microscopy was conducted using an epifluorescent microscope (Axiovert CFL 40, Zeiss, 20x magnification, Cy5 filter set). For flow cytometry analysis of ROS MFI, following CellROX staining, cells were harvested by cell scraper, pelleted by centrifugation at 1000 x g (4 °C, 5 min), washed once in PBS, resuspended in 0.3 mL of DMEM culture media (without FBS) and briefly placed on ice until flow cytometry analysis (within 30 min).

LPO assays were conducted in triplicate using BOPIPY-C₁₁-591 probe (Life Technologies) according to manufacturer instructions. A549 cells were cultured and treated with agents as described above, followed by harvesting and staining in DMEM media containing 10 μM of LPO probe for 30 min at 37 °C. Following staining, cell samples were washed once in PBS and resuspended in 0.3 mL of DMEM culture media (without FBS) for flow cytometry analysis.

For flow cytometry analysis of both ROS and LPO, all probe-stained samples were analyzed using a Fortessa flow cytometer (Becton Dickinson) equipped with excitation lasers and various emission detectors appropriate for the probe used. 10000 cells were acquired per sample. Data were exported in list mode (.fcs) and analyzed using FlowJo software (TreeStar). Dead cells and/or cell debris were eliminated from analyses by standardized scatter gating procedures.

Quantitation of HNE and GSH

A competitive ELISA kit (Cell Biolabs) was used to quantify cellular HNE-protein adducts according to manufacturer’s instructions. To prepare cell lysate samples for the assay, A549 cells were treated in triplicate with 0.1 μM of Top1 poisons for 72 h, rinsed once with PBS to remove dead cells, harvested by scraping, pelleted by centrifugation, and then lysed in 100 μL of ice-cold RIPA buffer (ThermoFisher) containing protease inhibitors (Halt, Thermo), EDTA (5 mM, Thermo), and DNAse I (1 μg/mL, Sigma-Aldrich) for 20 min on ice with frequent vortexing. Lysates were clarified by centrifugation at 16,500 x g (4 °C, 15 min). The supernatant was transferred to new tubes, and the protein concentration was assayed by BCA kit (ThermoFisher). 10 μg lysate in 50 μL of PBS was added to duplicate wells of an ELISA 96 well plate that had been precoated with HNE conjugate according to kit directions. Alongside the lysate samples, quantitative HNE-BSA standards were loaded into the plate in duplicate. 50 μL of anti-HNE antibody was added to each well for 1 h at 24 °C (ambient). The plate was washed 3X with the kit wash buffer. 100 μL of secondary antibody-HRP conjugate was added to all wells for 1 h at 24 °C, followed by another 3X wash cycle. 100 μL of HRP substrate was added to all wells for 5 min followed by 100 μL of stop solution. The absorbance of each well was read at 450 nm using a microplate reader (Synergy Neo, Tecan). Duplicate reads were obtained for all microwells. Absorbance data were exported to Excel (Microsoft) for quantitative determination of HNE in lysate samples as determined against the standard curve.

For quantitation of cellular GSH, a detection assay kit was used (Abcam). Cells were treated and lysates prepared as described for HNE assay above. Following protein content assay, equal amounts of treated cell lysates were deproteinated using trichloroacetic acid. Deproteinated lysate samples were then loaded into a 96 well microplate alongside GSH standard curve samples, both in triplicate. A GSH assay mixture was prepared according to kit directions and added to each microwell for 60 min at 24 °C in the dark. The fluorescent signal of each well was determined at 490 nm excitation and 520 nm emission using a microplate reader (Synergy Neo, Tecan). Duplicate reads were obtained for all microwells. Fluorescence data were exported to Excel software (Microsoft) for quantitative determination of GSH in lysate samples as determined against the standard curve.

Fluorescent Western blotting

To prepare samples for fluorescent Western blotting, A549 cells were treated with 0.1 μM of Top1 poisons for 72 h, harvested by scraping, pelleted by centrifugation, and then lysed in 100 μL of ice-cold RIPA buffer (ThermoFisher) containing protease inhibitors (Halt cocktail, ThermoFisher) for 20 min on ice with frequent vortexing. Lysates were clarified by centrifugation at 14,000 x g (4 °C, 15 min). The supernatant was transferred to new tubes, and the protein concentration was assayed by BCA kit (ThermoFisher). Lysates were then diluted into 4X SDS-PAGE sample buffer (40% glycerol, 8% SDS, 0.04% bromophenol blue, 20% beta-mercaptoethanol, and 240 mM Tris HCl pH 6.8, all components from Sigma-Aldrich) plus 1X protease inhibitor cocktail (‘Halt’, ThermoFisher) and boiled for 5 min at 95 °C. Equal amounts of protein per lane (20 μg) were loaded on to a 4-12% bis-tris acrylamide gel, along with a fluorescent molecular weight ladder (Life Technologies). Electrophoresis was conducted at 200 V for 40 min in MES buffer 1X. The gels were electrotransferred to low-fluorescence PVDF membranes (Bio-Rad) which had been preactivated in 100% methanol (Alfa Aesar) for 5 min. Electrotransfer proceeded at 30 V for 60 min in Novex transfer buffer 1X (Life Technologies) + 10% methanol. Transfer was confirmed by staining for total protein (Memcode, ThermoFisher) according to kit directions. The PVDF membranes were then blocked in fluorescent blotting blocking buffer (BlockerFL 1X, ThermoFisher) for 30 min at 24 °C. Primary antibodies were added at 1:1000x dilution overnight in 1% BSA-PBS. Primary antibodies against antioxidant proteins included anti-SOD1 (ProteinTech, rabbit polyclonal), anti-SOD2 (Abcam, rabbit clone EPR2560Y), or anti-NQO1 (Cell Signaling Technology, mouse clone A180). Primary antibodies targeting ALDH proteins included anti-ALDH1A1 (ProteinTech, rabbit polyclonal), ALDH2 (Novus, mouse clone 4G6A3), and ALDH3A1 (ProteinTech, rabbit polyclonal). The following day, unbound primary antibodies were washed away from membranes using a wash buffer of TBS 1X + 0.05% Tween-20 detergent, 3 X 10 min each. Fluorescent secondary antibodies included anti-mouse Alexa Fluor 680 Plus and anti-rabbit Alexa Fluor 800 Plus (Life Technologies), and were added at 1:40000 dilution in 1% BSA-PBS for 45 min at 4 °C. Membranes were washed 3 X 10 min each to remove unbound secondary antibody. Imaging was conducted using a fluorescent imager (iBright, Life Technologies) using ≤ 2 min camera exposure times. Image files were exported as high resolution TIFFs and minimally post-processed.

Aldehyde dehydrogenase flow cytometry assay

For flow cytometry analysis of ALDH activity, the AldeRed ALDH detection assay kit (Millipore) was used. First, A549 cells were treated in triplicate with 0.1 μM of Top1 poisons for 72 hr, harvested by scraping and pelleted by centrifugation. Cell samples were then resuspended in AldeRed 588-A probe staining buffer for 60 min at 37 °C, pelleted by centrifugation, and resuspended in 0.5 mL of kit assay buffer supplemented with Verapamil to prevent efflux of probe reaction product. Samples were placed on ice and immediately analyzed using a Fortessa flow cytometer (Becton Dickinson) equipped with a 532 nm excitation laser and a 610/20nm emission detector. 10000 cells were analyzed per sample, from each of two technical replicates per condition. Data were exported in list mode (.fcs) and analyzed using FlowJo software (TreeStar). Dead cells and/or cell debris were eliminated from analysis by standardized scatter gating procedures.

Proximity ligation assay

A549 cells were seeded in triplicate in 96 well glass bottom plates at 10 X 10³ cells per cm² and incubated for 18 h to allow cells to adhere. Cells were treated with CPT or 0.1% DMSO (vehicle) for 72 h in 100 μL of complete culture medium, washed once with 200 μL of PBS, fixed to 10 min in 4% paraformaldehyde, and washed twice with PBS. Cells were then permeabilized with 0.1% Triton-X-100 in PBS for 10 min and washed once with PBS. A proximitiy ligation assay kit (Sigma-Aldrich DuoLink PLA, Far Red) was used to visualize HNE modified Top1, in combination with a Top1 antibody (ProteinTech, from rabbit) and an HNE-protein adduct antibody (Abcam, from goat). Fixed and permeabilized cells were incubated in DuoLink blocking buffer for 60 min at 37 °C. Top1 and HNE antibodies were diluted to 10 μg/mL in 100 μL of DuoLink antibody dilution buffer and incubated on cells for 18 h at 4 °C. Cells were washed 2 X 5 min in 100 μL of DuoLink Buffer A. Anti-rabbit “Minus” and Anti-goat “Plus” probe-labeled secondary antibodies were diluted 1:5 in antibody diluent and applied for 1 h at 37 °C. Cells were washed again in DuoLink Buffer A. DuoLink ligase was diluted 1:40 in ligase buffer 1X and added for 30 min at 37 °C, followed by a wash step. DuoLink polymerase was diluted 1:80 in “Amp” buffer 1X and added for 100 min at 37 °C, followed by a wash step. All samples were finally washed 2 X 10 min in DuoLink Buffer B and then counterstained with DAPI (1 μg/mL in PBS) for 15 min. Stained samples were stored at 4 °C overnight and then imaged at 40X magnification using an Axiovert 40 CFL microscope using the same exposure times and settings for all sample wells in the microplate. A DAPI filter set was used to image DAPI stained nuclei and a Cy5 filter set was used to image the far-red fluorescent signal from the PLA probe (ex/em: 644/669 nm).

For immunofluorescence validation of Top1 and HNE antibodies used for proximity ligation assay, A549 cells were plated in coverglass based, sterile cell culture 96 well plate at 10 X 10³ per cm² for 18 h. The following day, cells were treated with CPT (1 μm) or DMSO (0.1%, vehicle) at 37 °C for 72 h, washed twice in PBS, fixed for 10 min in 4% paraformaldehyde, and permeabilized with 0.1% Triton-X-100 for 10 min. Cells were blocked in DuoLink blocking buffer from the PLA kit (Sigma-Aldrich) for 60 min at 37 °C. Top1 antibody (ProteinTech, from rabbit) and HNE antibody (Abcam, from goat) were diluted to 10 μg/mL in PLA kit antibody dilution buffer and added to cells, in parallel with PLA-stained cells, for 18 h at 4 °C. Cells were washed 2 X 5 min in 100 μL of DuoLink Buffer A. Anti-rabbit Alexa 647 secondary antibody (Life Technologies, 1:200 dilution) or anti-goat Alexa 647 (Jackson Immunological, 1:500 dilution) were diluted in 5% BSA-PBS and added to cells for 60 min at 4 °C. Cells were washed in DuoLink buffer B for 2 X 10 min and counterstained with DAPI at 1 μg/mL for 15 min at 24 °C. Samples were then imaged alongside PLA labeled cells at 40x magnification using the same epifluorescent microscope (Axiovert CFL 40, Zeiss) filter sets (DAPI and Cy5), monochrome CCD camera (mRM, Zeiss), and imaging software (Zen, Zeiss). Three images per microwell were taken. Images were saved as high-resolution TIFFs and minimally processed using ImageJ software (Version FIJI, NIH; (Schindelin et al., 2012)) post-acquisition. Image manipulations were applied equally to all images.

Top1 inhibition agarose gel assay

To study inhibition of Top1 by LDEs, recombinant Top1 (Topogen) was incubated with or without LDEs and then substrate DNA (pHOT1, Topogen) in a relaxation assay, before analysis by agarose gel electrophoresis. Duplicate Top1-LDE-DNA relaxation assays were performed in Tris-Glycine buffer (Topogen). 10 U of recombinant Top1, where one unit can relax 0.25 μg supercoiled plasmid in 30 min at 37 °C, was pre-incubated with VEH or HNE, HPNE, ONE, or HNE-GSH (Cayman Chemical) for 5 min at 37 °C and then 0.5 μg of pHOT1 supercoiled DNA plasmid was added for 30 min at 37 °C in a 20 μL reaction. PBS was used as the vehicle for diluting LDEs. Reactions were stopped with 6.6 μL of 5X stop buffer (Topogen), and then each reaction was separated on a 1% agarose (Lonza) gel prepared in 1X Tris-Borate EDTA (TBE; National Diagnostics). Electrophoresis was carried out for 90 min at 50V constant using 1X TBE as the running buffer. DNA was visualized in gels using 1X Sybr-Safe stain (Life Technologies) in 1X TBE for 30 min at 24 °C. Stained gels were imaged using a fluorescent gel imager (iBright; Life Technologies). Images were exported as high-resolution TIFFs, and were not adjusted post-acquisition (raw data are shown).

Senescence and apoptosis assays

A triple staining flow cytometry assay was used to determine accelerated senescence and apoptosis in viable cells. First, A549 cells were plated in duplicate at 10 X 10³ cells per cm² in 6 well sterile culture dishes for 18 h, and then treated (or not) with Top1 poisons at 0.1 μM for 72 h. Then, cell monolayers were detached using trypsin-EDTA, cells were pelleted by centrifugation for 5 min at 1,000 x g, and resuspended in 1 mL of DMEM (no FBS) with bafilomycin A1 (1 μM; Research Products International) for 30 min at 37 °C to adjust lysosomal pH in order to detect senescence-associated beta-galactosidase (SA-βGal). At 30 min, the SA-βGal senescence stain DDAO-galactoside (Life Technologies, 10 μg/mL) was added for 60 min at 37 °C. The apoptosis stain YO-PRO-1 (Life Technologies, 0.2 μM) was added for the last 30 min. Then, cells were pelleted by centrifugation, washed twice with PBS, and resuspended with viability stain Calcein Violet 450 AM (eBioscience, 1 μM) in DMEM (no FBS) on ice for 15 min. Triple stained cell samples were then immediately analyzed using a flow cytometer (Fortessa, Becton Dickinson) with appropriate lasers and detectors for the stains used (Calcein Violet, 405 nm excitation laser and 450/50 nm emission detector; YO-PRO-1, 488 nm laser and 525/50 nm detector; DDAO, 633 nm laser and 660/40 nm detector). 10,000 events were recorded for each sample using FACSDiva software (Becton Dickinson) and exported to FlowJo software (TreeStar) for gating and analysis. Single cells and then viable cells were gated equally across all samples. Apoptotic and senescent cell gating was determined for single-viable cell populations based on signal from vehicle-only samples and these gates were then applied equally across all samples. Resulting percentages of viable, senescent, and apoptotic cells in each sample were exported from FlowJo to Excel for graphing.

RADAR assay

RADAR assays were performed generally following the methods of (Kiianitsa and Maizels, 2013). SJCRH30 or A549 cells were treated with HNE, CPT, or DMSO (vehicle only, < 0.1%) in 1 mL of PBS or DMEM (without FBS) for 1-4 h at 37 °C. In some experiments, cells were returned to culture in DMEM + 10% FBS for 24 h to recover. Cells were pelleted by centrifugation, washed once with PBS, and 1.0 mL of cell lysis reagent (DNAZol, Life Technologies) was added for 10 min. The lysate was vigorously pipetted to mix, and stored at −80 °C before continuing the protocol.

Lysates were thawed on ice and 50% volume of EtOH (0.5 mL per 1.0 mL DNAZol) was added for 5 min at −20 °C to precipitate DNA-protein complexes. Precipitated samples were centrifuged at 4 °C for 10 min at 16,500 x g and the supernatants were removed. The DNA pellet was dissolved in 120 μL of TBS 1X and placed on ice. DNA quantitation was performed using a dsDNA assay kit (PicoGreen, Life Technologies) and a fluorescent microplate reader (Tecan Synergy Neo) against a standard curve of lambda DNA. Tubes were snap frozen in liquid nitrogen and again stored at −80 °C for several days prior to slot blotting for Top1.

For Top1 slot blotting, DNA samples were thawed on ice and DNA concentrations were adjusted to 1 μg/mL using TBS 1X. The slot blotter apparatus (Bio-Rad) was assembled using a presoaked nitrocellulose membrane-filter paper stack and connected to a house vacuum line. A priming load of TBS (200 μL) was added per well and passed through the device prior to loading DNA samples. DNA samples were then loaded at 200 μL per slot and passed through using a low vacuum setting. The membrane was removed and washed once with TBS-Tween 0.05% (TBS-T) and then blocked for 1 h. The membrane was incubated in the primary antibody solution (rabbit polyclonal Top1 antibody, ProteinTech; or mouse monoclonal Top1_cc antibody, clone 1.1A, Millipore) for 1 h at 24 °C, followed by 3 X 10 min washing in TBS-T. The membrane was incubated in HRP-conjugated secondary antibody solution (HRP anti-rabbit IgG, GE Life Sciences; or HRP anti-mouse IgG, Dako) for 30 min at 4 °C, followed by 3 X 10 min washing in TBS-T. Chemiluminescent HRP substrate (ECL Femto, Thermo) was added for 5 min and the membrane was imaged. The images were saved as TIFF files and minimally processed following acquisition.

Top1_cc imaging and flow cytometry

For immunofluorescence imaging of Top1_cc, a mouse monoclonal antibody specific to the phosphotyrosine bond in Top1cc-DNA complexes (clone 1.1A, Millipore) was used with a procedure adapted from (Patel et al., 2016). In summary, A549 cells were plated in triplicate in 6 well culture dishes at 20 X 10³ per cm² overnight. The following day, cells were treated with 0.2 mM CPT or HNE in PBS for 60 min at 37 °C, washed twice in PBS, fixed for 15 min in 4% paraformaldehyde at 4 °C, washed twice in PBS, and permeabilized with 0.25% Triton-X-100 for 15 min at 4 °C. Cells were incubated in 1% SDS at 24 °C for 5 min to stabilize Top1_cc. Cells were then washed 3X with PBS and blocked in a buffer consisting of 10% nonfat milk and 5% normal goat serum in 150 mM NaCl + 10 mM Tris HCl, pH 7.4, for 30 min at 24 °C. Top1_cc antibody was incubated overnight at 10 μg/mL in PBS. The following day, cells were washed three times with PBS plus 0.05% Triton-X-100 and incubated with goat anti-mouse-Alexa Fluor Plus 594 secondary antibody (Life Technologies) for 60 min at 4 °C, washed 3X, and counterstained with DAPI at 1 μg/mL for 15 min at 24 °C. Samples were then imaged in three regions per well at 20x magnification using an epifluorescent microscope (Axiovert CFL 40, Zeiss) with a long distance objective suitable for imaging in culture dishes (LD-A-Plan), appropriate filter sets, a monochrome CCD camera (mRM, Zeiss), and imaging software (Zen, Zeiss). Images were saved as high-resolution TIFFs and minimally processed using ImageJ software (NIH) post-acquisition. Image manipulations were applied equally to all images.

For flow cytometry analysis of Top1_cc, as above the staining procedure was adapted from (Patel et al., 2016). Cells were harvested in log phase growth (40 – 80% confluence), treated with 0.2 mM of each LDE in PBS for 60 min at 37 °C, washed once in PBS, and subjected to Top1_cc staining. For LDE scavenger assays, cells were pretreated with CYS2 (1 mM), GSH (1 mM), or BHT (0.1 mM) or no scavenger for 60 min and then subjected to CPT treatment at 1 μM for 60 min, and then subjected to Top1_cc staining.

For Top1_cc staining, all treated cell samples were fixed for 15 min in ice-cold 4% paraformaldehyde at 4 °C, and permeabilized with 0.25% Triton-X-100 for 10 min at 4 °C. and incubated in 1% SDS at 24 °C for 5 min to stabilize Top1_cc. Cells were then washed once with PBS and blocked in 10% nonfat milk (US Biological) plus 5% goat serum (Jackson Immunoresearch) in 10 mM Tris HCl, pH 7.4, for 30 min at 24 °C. 10 μg/mL clone 1.1A anti-Top1_cc was added for 60 min at 24 °C in PBS. Cells were then washed and incubated with goat anti-mouse Alexa Fluor Plus 594 secondary antibody (Life Technologies) for 60 min at 4 °C, washed, and resuspended in 1% BSA-PBS for flow cytometry analysis. All samples were analyzed using a Fortessa flow cytometer (Becton Dickinson) equipped with excitation lasers and emission detectors appropriate for the fluorophores used. At least 1,000 cells were acquired per sample. Raw data were exported from FACSDiva software in list mode (.fcs) and analyzed using FlowJo software (TreeStar). Dead cells and/or cell debris were eliminated from analysis by gating.

Preparation of Tdp1 CRISPR/Cas9 knockout cell line

A Tdp1 CRISPR/Cas9 knockout cell line (TDP1KO) was prepared as described previously (Li et al., 2017c). Briefly, Tdp1 targeting plasmids were constructed by ligating the gRNA targeting sequences 5′-TCTTTGGGCAGTGCCGTCAT-3′ or 5′-GACATCTCTGCTCCCAATGA-3′ into the pX330-U6-Chimeric_BB-CBh-hSpCas9 plasmid (Addgene 42230). The gene targeting plasmids were then transfected into HEK-293 cells along with the pcDNA3.1.hygro plasmid (Life Technologies) for 2 d. Transfected cells were then selected in hygromycin (200 μg/mL) for 3 d. Clonal expansion was conducted in 96 well plates to identify knockout clones by Western Blot as well as PCR amplification followed by sequence analysis to confirm disruption of Tdp1 coding sequences.

Cellular proliferation assays

HEK-293 wild type or TDP1KO cells were seeded in 96-well plates at 3,000 cells per well and incubated for 24 h to allow cells to adhere. Microwells were seeded with equal numbers of cells in triplicate. Cells were treated with LDEs from 0 to 100 μM and incubated for 72 h. DMSO vehicle concentrations for all treatments were < 1 % of final well volume (100 μL). At treatment endpoint, cellular proliferation was detected using a WST-1 assay kit (Sigma-Aldrich) according to manufacturer’s instructions. Absorbance at 450 nm was detected using a Synergy 2 Multi-Mode Reader (BioTek). Duplicate reads per microwell were taken. Resulting data were analyzed using Excel (Microsoft).

Immunoprecipitation of Top1-HNE

To prepare cell lysates for immunoprecipitation, A549 cells were plated at 10 X 10³ cells per cm², incubated for 18 h, and treated with CPT (1 μM), TPT (1 μM), or DMSO (0.1%, vehicle) for 0 - 120 h. Cells were washed once with PBS, scraped into fresh PBS, and pelleted by centrifugation for 5 min at 1,000 x g. The supernatant PBS was removed and 100 μL of RIPA lysis buffer (Thermo) was added, including protease inhibitor cocktail (Halt, Thermo). Cells were lysed for 20 min on ice with vortexing every 5 min. Lysates were clarified by centrifugation for 15 min at 16,500 x g (4 °C). Protein content assay was conducted for lysates using a BCA kit (Thermo) with a BSA standard curve and a standard spectrometer (Beckman Coulter) at 562 nm absorbance.

For immunoprecipitation, 50 μL of DynaBeads Protein G beads from an IP kit (Life Technologies) was added per tube. 10 μg of monoclonal HNE antibody (Abcam) was added to DynaBeads in 200 μL of PBS-Tween 0.02% and incubated on rotator for 10 min at 24 °C. Tubes were placed on magnetic stand for 2 min and supernatant was removed. Cell lysates (40 μg) were then added to antibody-linked DynaBeads and incubated for 30 min at 24 °C. Tubes were placed on magnetic stand for 2 min and supernatant was removed, followed by washing DynaBeads 3 X 200 μL of wash buffer from the IP kit (Life Technologies). Bead suspensions were transferred to new tubes and 20 μL of elution buffer from kit was added, along with 6.6 μL of 4X SDS-PAGE sample buffer (for 1X final). Samples were heated for 10 min at 95 °C, cooled for 5 min on benchtop, placed on magnetic rack to separate used beads, and the supernatants were transferred to new tubes and stored at −20 °C until Western blotting (approximately one week).

For Western blotting of HNE IP samples using Top1 antibody, eluted samples were thawed on ice and then loaded into lanes of a 4-12% bis-tris acrylamide gel (Life Technologies). Electrophoresis was conducted in MOPS buffer 1X for 50 min at 200 V constant. The gel was then transferred to a nitrocellulose membrane for 75 min at 30 V constant. The membrane was washed once in TBS-Tween 0.05% (TBS-T) and blocked for 60 min in 5% BSA-PBS. Top1 antibody was diluted 1:1000 in 1% BSA-PBS and the membrane was incubated in the antibody solution for 60 min at 24 °C. The membrane was washed 3 x 10 min in TBS-T. Anti-rabbit-HRP (GE) was added at 1:40000 for 30 min at 24 °C, and the membrane was washed again for 3 x 10 min in TBS-T. Chemiluminescent HRP substrate (ECL Femto, Thermo), was added for 5 min and the membrane was imaged using a digital imager (iBright, Thermo). The images were saved as TIFF files and minimally processed following acquisition.

Thiol reactive probe assay

To assess thiol reactivity of hTop1 protein ± HNE, Thiol Reactive Probe IV (EMD Millipore) was used. Recombinant hTop1 protein (Topogen) was stored at 300 U/μL in a buffer consisting of 20 mM NaH₂PO₄ (pH 7.4), 300 mM NaCl, 50 μg/mL BSA, 50% glycerol, and 50 mM imidazole. Just prior to reaction, 100 U of hTop1 was diluted into 100 μL of PBS-only to avoid potential off-target interactions between HNE and buffer components. HNE was added at 0 to 1000 μM for 30 min at 37 °C, and then Thiol Reactive Probe IV was added at 1 μM for 5 min. Reactions were conducted in individual wells of a black clear bottom microplate (Greiner) with a light blocking seal applied to the plate lid. Fluorescent signal was measured using a microplate reader (Synergy Neo, Tecan) using 400 nm excitation and 440 nm emission wavelengths. Readings were taken in duplicate and exported to Excel (Microsoft) for further analysis.

QUANTIFICATION AND STATISTICAL ANALYSIS

For ROS vs. LPO analysis, linear regression analysis was performed. Kruskal-Wallis tests were used to analyze differences between sample medians for the HNE adducts ELISA assay and GSH content assay. For microplate-based proliferation assay data, Mann-Whitney tests were used to compare two data points. A P-value of ≤ 0.05 was considered significant. Statistically significant conditions are marked with an asterisk (*) throughout the figures.

To perform statistical analysis of proteomics data, Scaffold PTM software was used, incorporating a site localization algorithm (Beausoleil et al., 2006) to re-analyze MS/MS spectra identified as modified peptides. Ambiguity score (A-score) values and site localization probabilities were calculated to assess the level of confidence in each PTM localization. The A-score is a probability-based metric that measures the likelihood that a difference in site-determining ions between two site positions was matched by random chance. Scaffold PTM then combines localization probabilities for all peptides containing each identified PTM site to obtain the best estimated probability that a PTM is present at that particular site. We calculated site localization P-value as follows: (1/(10^(A-score/10))). Any PTM site with an A-score > 19 represents a probability of > 99%.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-ALDH1A1, polyclonal	ProteinTech	Cat #15910-1-AP
Anti-ALDH1A3, polyclonal	ProteinTech	Cat #25167-1-AP
Anti-ALDH2, clone 4G6A3	Novus	Cat #NBP2-37397
Anti-ALDH3A1, polyclonal	ProteinTech	Cat #15578-1-AP
Anti-HNE protein adducts, polyclonal	Abcam	Cat #ab46544
Anti-HNE protein adducts, clone HNEJ-2	Abcam	Cat #ab48506
Anti-SOD1, polyclonal	ProteinTech	Cat #10269-1-AP
Anti-SOD2, clone EPR2560Y	Abcam	Cat #ab68155
Anti-NQO1, clone A180	Cell Signaling Technology	Cat #3187
Anti-Top1, polyclonal	ProteinTech	Cat #20705-1-AP
Anti-Top1cc, clone 1.1A	Millipore	Cat #MABE1084
Bacterial and Virus Strains
< none >
Biological Samples
< none >
Chemicals, Peptides, and Recombinant Proteins
Camptothecin (CPT)	Cayman Chemical	Cat #11694
Topotecan (TPT)	Cayman Chemical	Cat #29082
9-amino-camptothecin (9-CPT)	Cayman Chemical	Cat #17232
(S)-10-hydroxy-camptothecin (10-CPT)	Cayman Chemical	Cat #14635
Irinotecan (IRN)	Hospira Oncology	NDC 61703-349-09
SN-38 (SN-38)	Cayman Chemical	Cat #15632
Genz-644282 (GENZ)	Cayman Chemical	Cat #24193
SW044248 (SW)	Cayman Chemical	Cat #23459
4-hydroxy-2-nonenal (HNE)	Cayman Chemical	Cat #32100
4-hydroperoxy-2-nonenal (HPNE)	Cayman Chemical	Cat #10004413
4-oxo-2-nonenal (ONE)	Cayman Chemical	Cat #10185
4-hydroxy-2-nonenal glutathione salt (GSH-HNE)	Cayman Chemical	Cat #10627
Glutathione, reduced (GSH)	Sigma-Aldrich	Cat #G6013
Cystine (CST)	Sigma-Aldrich	Cat #C6727
2, 6-Di-tert-butyl-4-methylphenol (BHT)	Sigma-Aldrich	Cat #B1378
Recombinant human topoisomerase 1 protein	Topogen	Cat #TG2005H
Critical Commercial Assays
CellROX Deep Red Probe	Life Technologies	Cat #C10422
BODIPY 581/591 C11 Lipid Peroxidation Sensor	Life Technologies	Cat #D3861
Glutathione Assay Kit	Abcam	Cat #ab138881
HNE Adduct Competitive ELISA Kit	Cell Biolabs	Cat #STA-838
AldeRed Assay Kit	Millipore	Cat #SCR150
DuoLink Proximity Ligation Assay Kit, Far Red	Sigma-Aldrich	Cat #DUO092013
PicoGreen dsDNA Assay Kit	Life Technologies	Cat #P7589
WST-1 Cell Proliferation Reagent	Sigma-Aldrich	Cat #11644807001
DynaBeads Protein G Immunoprecipitation Kit	Life Technologies	Cat #10007D
Yo-Pro-1 iodide (apoptosis stain)	Life Technologies	Cat #Y3603
DDAO-galactoside (senescence stain)	Life Technologies	Cat #D6488
Thiol Fluorescent Probe IV	Millipore	Cat #595504
Deposited Data
Proteomics (Fig. 1, 6, S7)	PRIDE	Accession #PXD015181
Flow cytometry, microscopy, etc.	Mendeley	http://dx.doi.org/10.17632/gf8pnt826y.1
Experimental Models: Cell Lines
Human: A549 lung adenocarcinoma cells	ATCC	Cat #CCL-185 RRID: CVCL_0023
Human: HEK293 embryonic kidney cells	ATCC	Cat #CRL-1573 RRID: CVCL_0045
Human: SJCRH30 rhabdomyosarcoma cells	ATCC	Cat #CRL-2061 RRID: CVCL_0041
Oligonucleotides
TCTTTGGGCAGTGCCGTCAT	IDT	Custom primer (gRNA targeting sequence #1 for CRISPR knockout of Tdp1)
GACATCTCTGCTCCCAATGA	IDT	Custom primer (gRNA targeting sequence #2 for CRISPR knockout of Tdp1)
Recombinant DNA
pHOT1 supercoiled DNA plasmid	Topogen	Cat #TG2030
pX330-U6-Chimeric_BB-CBh-hSpCas9 plasmid	Addgene	Cat #42230
pcDNA3.1.hygro	Life Technologies	Cat #V87020
Software and Algorithms
MaxQuant software	Max Planck Institute of Biochemistry	Version 1.6.5, https://www.maxquant.org/
Perseus software	Max Planck Institute of Biochemistry	Version 1.6.7, https://maxquant.net/perseus/
X! Tandem software	The GPM	Version CYCLONE, https://www.thegpm.org/TANDEM/instructions.html
Inclusion list software	University of Chicago (Kron Lab) in-house software	Version 1.4.9 (available upon request)
Scaffold software with local FDR algorithm	Proteome Software Inc.	Version 4.10.0, http://www.proteomesoftware.com/products/scaffold/download/
DAVID Bioinformatics Resource	National Institute of Health (NIH) Laboratory of Human Retrovirology and Immunoinformatics	Version 6.8, https://david.ncifcrf.gov
ImageJ software	National Institute of Health (NIH)	Version “FIJI”, https://imagej.net/Fiji.html#Downloads
FACSDiva software	Becton Dickinson (BD)	https://www.bdbiosciences.com/en-us/instruments/research-instruments/research-software/flow-cytometry-acquisition/facsdiva-software
FlowJo software	TreeStar	Versions 9-10, https://www.flowjo.com/solutions/flowjo/downloads
Prism software	GraphPad	Version 8.4.3, https://www.graphpad.com/demos/
Excel software	Microsoft	Version 16, https://www.microsoft.com/en-us/microsoft-365/excel
Other
< none >

Significance.

Despite the emergence over the last decades of new targeted therapies, topoisomerase poisons such as the camptothecin analogs are likely to remain critical drugs for chemotherapy, reflecting their high benefit, low cost and broad availability. Here, we show that increased oxidative stress upon treating cells with camptothecins is not an “off-target” effect, but a mechanism of action. We find that reactive end-products of lipid peroxidation are both necessary and sufficient to poison Top1, likely mediated by covalent modification at the active site. Interfacial inhibition may still be critical as camptothecin binding to stabilize the covalent complex of Top1 and DNA may promote specificity in vivo. Beyond cancer therapy, these results implicate Top1 in a new role beyond maintaining proper DNA topology during cell division. Increased accumulation of Top1 stably bound to cleaved DNA may serve as a sensor for oxidative stress, inducing a DNA damage signal in parallel to the anti-oxidant response.

Highlights.

• Top1-targeting camptothecin drugs induce oxidative stress and lipid peroxidation

• Lipid-derived electrophiles modify a Top1 active site cysteine by Michael addition

• Electrophile-modified Top1 forms covalent complexes with cleaved DNA

• Oxidative stress may mediate cytotoxicity of Top1 poisons in vivo