Anticancer Effects of New Disulfiram Analogs

Abstract

Disulfiram (DSF), an irreversible aldehyde dehydrogenase 2 (ALDH2) inhibitor, is an U.S. Food and Drug Administration (FDA)-approved drug for the treatment of chronic alcoholism. Recent studies have reported an interesting antitumor activity of DSF against a wide range of human malignancies, while a growing number of ongoing and completed clinical trials have provided evidence supporting the repurposing of DSF as an anticancer treatment. Nevertheless, despite its current clinical indications and potential future therapeutic applications for treating various diseases, DSF is associated with serious side effects attributed to the inhibition of ALDH2. We have recently synthesized DSF analogs (2a–v) with limited inhibition of ALDH2. Here, we report the anticancer activity of these molecules by highlighting their effects on cell signaling in Jurkat cells. DSF and two DSF analogs, 2g and 2r, all stimulated apoptotic signaling pathways, although the 2g analog activated more apoptosis-related genes and induced higher levels of apoptosis in vitro. Differential gene expression data suggested that compounds 2g and 2r specifically reprogram target cells to downregulate pathways related to cell growth and division, while upregulating pathways related to apoptosis or differentiation. Interestingly, both compounds 2g and 2r had more differentially expressed genes related to DNA damage pathways (including those related to apoptosis) when compared to DSF, which may offer insights into their mechanisms of action.

INTRODUCTION

Disulfiram (1) (DSF; Fig. 1) is an irreversible inhibitor of aldehyde dehydrogenase-2 (ALDH2) used in the treatment of alcoholism. By inhibiting ALDH2, disulfiram blocks alcohol metabolism at the acetaldehyde stage, leading to the accumulation of acetaldehyde in the blood. This, in turn, produces a highly unpleasant reaction that includes low blood pressure, tachycardia, facial flushing, vomiting, and vertigo when a patient treated with disulfiram ingests even small amounts of alcohol.¹⁾ DSF has been approved by the U.S. Food and Drug Administration (FDA) as an effective treatment for chronic alcoholism since 1951. However, recent studies have reported that DSF has interesting antitumor activity against a wide range of human malignancies, such as bladder,²⁾ intestinal,³⁾ breast,⁴⁾ colorectal,⁵⁾ ovarian,⁶⁾ and kidney⁷⁾ cancers. Furthermore, the complex formed between the disulfiram metabolite diethyldithiocarbamate and copper ions (CuET) has been reported to possess potent anticancer activity through the inhibition of nuclear factor-kappaB (NF-κB)⁸⁾ and the ubiquitin-proteasome system,⁹⁾ as well as by generating reactive oxygen species in cancer cells, leading to apoptosis.^10,11) Additionally, CuET inhibits several molecular targets associated with drug resistance, stemness, angiogenesis, and metastasis.¹²⁾ An accumulating number of ongoing and completed clinical trials have provided evidence that DSF has potential for repurposing as an anticancer treatment.¹³⁾

Fig. 1. Chemical Structure of Disulfiram (1)

Notwithstanding its current clinical indications and potential future therapeutic uses in treating different diseases, DSF administration is associated with serious side effects.^14–16) This is because ALDH2 inhibition, the main mechanism underlying the alcohol-aversive effects of DSF, is associated with a number of human pathologies, including cardiovascular diseases, neurodegenerative disorders, and cancer.¹⁷⁾ Inhibition of this enzyme therefore has the potential to hamper any repurposing of disulfiram for managing chronic diseases like cancer. In our research, we have recently developed bulkier analogs of DSF (2a–v) (Table 1) that can avoid the side effects caused by DSF inhibition of ALDH2.^18–21) These compounds are too large to pass through the small ALDH2 substrate tunnel²²⁾; therefore, they show negligible inhibition of ALDH2.

Table 1. Chemical Structures of DSF Analogs (2a–v)

In the present study, we investigated the anticancer potential of DSF and its analogs (2a–v) in a human Jurkat cancer cell line and explored their mechanism of action by performing RNA-seq.

MATERIALS AND METHODS

Chemistry

The commercially available compounds 2a–c and 2u–v were purchased from BOC Sciences, Shirley, NY 11967, U.S.A. The synthesis of compounds 2d–t was described elsewhere.^18–21)

Cell Culture

Jurkat cells (ATCC^® TIB-152™) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% (v/v) fetal bovine serum (Lonza, Verviers, Belgium), 50 U/mL penicillin and 50 µg/mL streptomycin (Sigma-Aldrich, St. Louis, MO, U.S.A.). Cell culture was maintained in humidified incubator at 37 °C in 5% CO₂ atmosphere.

Apoptosis Assay

Cells were seeded in 96-well flat-bottom plates (Corning^® Costar^®, Cat. No. CLS3596) in 0.2 mL at 10 × 10³ cells/well. Cell treatment was started 2 h after seeding. Investigated compounds were tested using two concentrations (10 and 50 µM). As controls, cells treated with vehicle (0.5% dimethyl sulfoxide (DMSO)) were used. After 3 h of exposure, Apoptosis rate was then assessed by flow cytometry (Guava EasyCyte Plus, EMD Millipore Corporation, Billerica, MA, U.S.A.), using propidium iodide (PI) (Miltenyi Biotec, Paris, France) and annexin V-fluorescein isothiocyanate (FITC) (Immunotools, Friesoythe, Germany), following the manufacturer’s recommendations. A minimum of 5 × 10³ cells were acquired per sample and analysed on the InCyte software (Guava/Luminex, CA, U.S.A.).

Statistical Analysis

Data, presented as bar graphs, were expressed as means ± standard error of the mean (S.E.M.) of at least three independent experiments. Statistical evaluation was performed with the one-way ANOVA test followed by the post-hoc Bonferroni test using GraphPad Prism software (Prism version 5.04 for Windows, GraphPad Software, CA, U.S.A.).

RNA-Seq

Library Construction and Sequencing

Total RNAs were purified from subconfluent Jurkat cells using standard methods. Libraries of template molecules suitable for strand-specific high-throughput DNA sequencing were created using a TruSeq Stranded Total RNA with Ribo-Zero Gold Prep Kit (RS-122-2301; Illumina), as previously described.²³⁾ The libraries were sequenced on Illumina HiSeq 4000 sequencer as single-end 50-bp reads following Illumina’s instructions. Reads were preprocessed in order to remove adapter, polyA and low-quality sequences (Phred quality score below 20). After this preprocessing, reads shorter than 40 bases were discarded for further analysis. These preprocessing steps were performed using cutadapt²⁴⁾ version 1.10. Reads were mapped to ribosomal RNA (rRNA) sequences using bowtie²⁵⁾ version 2.2.8, and reads mapping to rRNA sequences were removed for further analysis. Reads were mapped onto the hg38 assembly of Homo sapiens genome using STAR²⁶⁾ version 2.5.3a.

Gene Expression Analysis

Gene expression quantification was performed from uniquely aligned reads using HTSeq-count²⁷⁾ version 0.6.1p1, with annotations from Ensembl version 106 and “union” mode. Only non-ambiguously assigned reads to a gene have been retained for further analyses. Read counts have been normalized across samples with the median-of-ratios method proposed by Anders and Huber,²⁸⁾ to make these counts comparable between samples. Comparisons of interest were performed using the test for differential expression and p-value adjustment method proposed by Love et al.²⁹⁾ and implemented in the Bioconductor package DESeq2 version 1.16.1. Comparisons of interest were performed using the test for differential expression and p-value adjustment method proposed by Love et al.²⁹⁾ and implemented in the Bioconductor package DESeq2 version 1.16.1.

DNA Repeat Expression Analysis

Repeat analysis was performed following a methodology derived from Fadloun et al.³⁰⁾ Briefly, read were aligned to repetitive elements in two passes. In the first pass, reads were aligned to the non-masked Homo sapiens reference genome (hg38) using BWA v0.6.2.³¹⁾ Reads which sense were of the same sense as overlapping transcript were removed. Prior to this step, genomic coordinates of transcripts were extended 3kb upstream of TSS and 10kb downstream of TTS to remove reads arising from transcriptional readthrough. Positions of the reads uniquely mapped to the Homo sapiens genome were cross-compared with the positions of the repeats extracted from UCSC (rmsk table in UCSC database for Homo sapiens hg38) and reads overlapping a repeat sequence were annotated with the repeat family. In the second pass, reads not mapped or multi-mapped to the Homo sapiens genome in the previous pass were aligned to RepBase³²⁾ v21.08. Reads mapped to a unique repeat family were annotated with the family name. Finally, we summed up the read counts per repeat family of the two annotation steps. Data were normalized based upon library size. Significance of the difference of repeat read counts between RNA samples was assessed using the Bionconductor package DESeq2.

RNA-Seq Analysis Methods

Transcript counts were normalized and further analyzed using DESeq2.^29,33) Genes with zero counts in all samples, genes with an extreme count outlier, and genes with a low mean normalized counts were excluded from the analysis. Shrink gene-wise dispersion estimates were performed using the ‘apeglm’ model.³⁴⁾ Gene Ontology terms were extracted using the ‘ensembledb’ R package.³⁵⁾ Data were analyzed and plotted using R v4.3.3.

RESULTS

The in vitro anticancer potential of DSF and its analogs (2a–v) was investigated in a human Jurkat cancer cell line by conducting annexin V-FITC binding assays to determine the induction of programmed cell death by these compounds (Fig. 2). A 3 h treatment with DSF at 50 µM resulted in 44% of the treated Jurkat cells becoming annexin V-FITC positive, indicating the induction of apoptosis. Among the analogs, compounds 2g and 2r demonstrated the most potent anticancer properties, inducing apoptosis by 73 and 52%, respectively.

Fig. 2. A. Apoptosis Induced by DSF (1) and Its Analogs 2g and 2r in Jurkat Cancer Cells

Cells were treated for 3 h with each compound (50 µM) and then analyzed by capillary cytometry (annexin-V and PI staining). Dot plots represent living (lower left panel: annexinV⁻ & PI⁻), early apoptotic (lower right panel: annexinV⁺ & PI⁻), late apoptotic (upper right panel: annexinV⁺ & PI⁺), and necrotic (upper left panel: annexinV⁻ & PI⁻) cells. B. Percentage of apoptotic cells induced by DSF and its analogs 2g and 2r. Values are mean ± S.E.M. of at least three independent experiments; statistically significant: ***, p < 0,01.

The mechanism of action of DSF and its analogs was explored by performing RNA-seq on Jurkat cancer cells exposed to DSF and compounds 2g and 2r at 50 µM for 3h. A total of 8.7 × 10⁷, 1.0 × 10⁸, and 9.9 × 10⁷ gene counts, corresponding to 39468, 36688, and 37798 genes, were collected for cells treated with DSF, 2g, and 2r, respectively. Examination of the distributions of read counts across replicates treated with DMSO or the three selected drugs revealed that the variance was greater than the mean, indicating that the Negative Binomial model (Fig. 3A) should be used to fit the data. Although several genes did not seem to follow the modeling assumptions and had greater variability due to biological factors (e.g., mitochondrial DNA), the gene-wise dispersion estimates among the genes in all samples revealed an overall good fit with the expected dispersion values (Fig. 3B). This closeness in fit suggested that identification of the differentially expressed genes (DEGs) was likely to be accurate and to lack false positives or false negatives. A further correlation of gene expression for all pairwise combinations of samples showed that the replicates all clustered within their respective conditions, ensuring replicate validity (Fig. 3C). A comparison between the expression profiles of the top 500 most variable genes reported for each compound revealed a higher similarity between compound 2r and either DSF or compound 2g than between DSF and compound 2g (Fig. 3D). Initial data filtering and quality assessment suggested that the reads were good for determining which genes were differentially expressed among the different treatments.

Fig. 3. RNA-Seq Quality Control and Replicate Similarity

A. Dispersion estimates for DMSO-treated cells are shown. B. A plot with the curve fit to the estimated gene-wise dispersion values is displayed. C. A hierarchical clustering heatmap shows the pairwise correlation of gene expression profiles between all samples. D. A principal components analysis plot for normalized gene counts is shown, with each condition (i.e., treated with a drug or DMSO) represented by a different color.

RNA-seq differential gene analysis provides a list of statistically significant DEGs for identifying biological processes altered by a given condition (e.g., a drug). First, the DEGs of DSF and the two analogs with respect to the expression profiles of the DMSO-treated cells were independently analyzed. Because numerous genes were detected in the analysis after filtering, stringent p-values (<0.001) and log₂ fold changes (<−0.5 or >0.5) were selected. With DSF, 5135 genes were upregulated and 2135 were downregulated; with 2g, 3883 were upregulated and 2261 were downregulated; and with 2r, 4184 were upregulated and 2373 were downregulated. As shown in Figs. 4A, 4C, and 4E, more DEGs were upregulated than downregulated after treatment with any of the compounds. Interestingly, PPP1R15A, FOS, and FOSB were among the ten DEGs with the lowest reported p-values for all three compounds, while JUN and HSPA1B were among the top ten for compounds 2g and 2r. Similarly, for the downregulated genes, RP6-159A1.4 (MIR223HG), CCDC26, C1orf132 (MIR29B2CHG), RP5-1171I10.5 (microRNA 142), and ARHGAP19-SLIT1 had the lowest p-values for all three compounds, while CAD and RIMS3 were among the top ten downregulated DEGs for compounds 2g and 2r. Notably, for all three compounds, many of the top 100 significantly downregulated DEGs were non-coding RNAs, while the upregulated DEGs were primarily protein-coding genes.

Fig. 4. Significant Differentially Expressed Genes and Enriched Pathways between the Compounds and DMSO

Volcano plots depict the differential expression of cells treated with DSF (A), compound 2g (C), and compound 2r (E) versus DMSO. Each point on the volcano plots represents one gene. Upregulated genes are colored in red, downregulated in blue, and not significant in gray. The ten genes with the lowest p-value for both the upregulated and downregulated genes are labeled with text in the corresponding color. The gray vertical dotted lines represent the fold₂ cut-off of 0.5 and the horizontal dotted line for the p-value cut-off of 0.01. Dot plots show the top 25 differentially enriched “Biological Process” GO terms for DSF (B), compound 2g (D), and compound 2r (F) versus DMSO. Mixed violin and boxplots depict the distribution and density of log₂ fold changes between the three compounds and DMSO for DEGs annotated to be involved in apoptosis (G) and negative and positive regulation of the cell cycle (H and I, respectively).

Utilizing the available Gene Ontology (GO) annotations for genes in the human genome, the biological processes that are enriched in DEGs for each condition can be analyzed to shed light on the functional properties of the compounds. As shown in Figs. 4B, 4D, and 4F, all three compounds significantly affected metabolic pathways related to DNA/RNA (e.g., “purine ribonucleotide metabolic process”) and protein (e.g., “regulation of protein catabolic process”) synthesis or degradation/recycling. Diverse anti-cancer drugs have been shown to induce remodeling of energetic and metabolic pathways. Interestingly, all three compounds affected the expression of genes in the “intrinsic apoptotic signaling pathway,” and DEGs in compound 2g also mapped to the “regulation of apoptotic signaling pathway.” There were 71, 77, and 75 DEGs involved in apoptosis for cells treated with DSF, compound 2g, and compound 2r, respectively. Among the DEGs in apoptotic signaling pathways, 7, 9, and 5 were unique to DSF, compound 2g, and compound 2r, respectively. Overall, the differential expression profiles of genes involved in apoptosis were similar among all three compounds (Fig. 4G), although the slight increase in the number of total and unique DEGs for cells treated with compound 2g could explain the cytometry results. Furthermore, in support of their potential use as anti-cancer compounds, DEGs corresponding to “negative regulation of cell cycle” were found among the most highly enriched pathways in cells treated with DSF and compound 2r. Interestingly, cells treated with DSF had more downregulated DEGs that both positively and negatively regulate cell cycle processes (Figs. 4H, 4I). Further comparative analyses between compounds may reveal drug-specific responses.

Comparing the detected statistically significant DEGs among the compounds revealed a large number of shared genes. Notably, DSF treatment resulted in a much higher number of unique upregulated DEGs and also shared the largest number of upregulated genes with compound 2r (Fig. 5A). Interestingly, as shown in Figs. 5B and 5D, five of the top ten upregulated genes (CD69, ATF3, NR4A1, MT1F, and MT1E) affected by compounds 2g and 2r are shared when compared to DSF.

Fig. 5. Significant Differentially Expressed Genes and Enriched Pathways between the DSF and Its Analogs

(A) A Venn diagram comparing the number of DEGs shared among the three compounds is shown. Volcano plots depict the differential expression between cells treated with DSF and compound 2g (B), DSF and compound 2r (D), and compound 2g and compound 2r (F). Each point on the volcano plots represents one gene. Upregulated genes are depicted in red, downregulated in blue, and not significant in gray. The ten genes with the lowest p-value for both the upregulated and downregulated genes are labeled in text in the corresponding color. The gray vertical dotted lines represent the fold₂ cut-off of 0.5 and the horizontal dotted line for the p-value cut-off of 0.01. Dot plots show the top 25 differentially enriched “Biological Process” GO terms for DSF and compound 2g (C), DSF and compound 2r (E), and compound 2g and compound 2r (G). (H) The bar chart compares the number of genes involved in DNA damage response pathways.

Three significantly downregulated genes (FLNA, TLN1, and CAD) were found in common in cells treated with compounds 2g and 2r when compared with DSF treatment. Comparison of the top DEGs between the compound 2r and compound 2g treatments revealed a lower p-value for many of the downregulated genes than for upregulated genes (i.e., higher expression for compound 2g than compound 2r treatment). These DEGs suggest that compounds 2g and 2r specifically reprogram target cells to downregulate pathways related to cell growth and division and to upregulate pathways related to apoptosis or differentiation.

Notably, comparison of the enriched signaling pathways between the compounds (Figs. 5C, 5E, 5G) revealed high enrichment of biological processes related to DNA damage repair (e.g., “signal transduction in response to DNA damage”) after compound 2r treatment compared with DSF treatment, potentially offering insights into the mechanism of action. Interestingly, cells treated with compound 2g or compound 2r had more DEGs in DNA damage pathways (including those related to apoptosis) when compared with cells treated with DSF (Fig. 5H). DNA damage repair pathways are often targeted and considered for cancer drugs.³⁶⁾

DISCUSSION

DSF has an important antitumor activity against a wide range of human cancers. However, despite its current clinical indications and potential future therapeutic repurposing, DSF is associated with serious side effects attributed to the inhibition of ALDH2. To address those drawbacks, we have recently synthesized DSF analogs with limited inhibition of ALDH2. In the present study, we report the anticancer activity of these molecules by highlighting their effects on cell signaling in Jurkat cells. Two derivatives, 2g and 2r, showed interesting apoptotic activity, higher than that displayed by DSF. DEGs analysis was conducted to elucidate the pathways involved in the anticancer activity. Several genes among the topmost upregulated such as CD69, ATF3, NR4A1, MT1F, and MT1E were shared by all the three compounds. Interestingly, CD69 has been identified as a lymphocyte activator and enhances differentiation.³⁷⁾ Activation of ATF3 induces differentiation of immune cells, which may prevent cancer development.³⁸⁾ A genome-wide analysis has identified NR4A1 as an inducer of T cell dysfunction, particularly in the context of the tumor microenvironment, and NR4A1 has been implicated in negative regulation of the cell cycle.³⁹⁾ MT1 are metalloproteins that play a crucial role in metal homeostasis, alleviating heavy metal toxicity and protection against oxidative stress and DNA damage.^40,41) MT1 have multifaceted roles in cancer, e.g., MT1F is considered an immunosuppressor in colon cancer.⁴²⁾

On the other hand, three significantly downregulated genes (FLNA, TLN1, and CAD) were found in common in cells treated with compounds 2g and 2r when compared with DSF treatment. Silencing of TLN1 can prevent proliferation of cancerous lymphocytes,⁴³⁾ and overexpression of FLNA has also been associated with cancer.⁴⁴⁾ The CAD gene is part of the pyrimidine biosynthesis pathway; thus, its downregulation may lead to an inhibition of DNA synthesis and cell proliferation.⁴⁵⁾ Notably, more DEGs involved in DNA damage repair pathways which are often targeted and considered for cancer drugs,³⁶⁾ were found in cells treated with the analogs than DSF or DMSO.

CONCLUSION

In conclusion, compounds 2g and 2r showed interesting in vitro anticancer activity. The RNA-seq data showed that apoptotic signaling pathways were affected by DSF and by both analogs 2g and 2r. Overall, 2g activated a slightly larger number of related genes and induced higher levels of apoptosis in vitro. The DEGs data suggest that compounds 2r and 2g specifically reprogrammed the tested cells to downregulate pathways related to cell growth and division and to upregulate pathways related to apoptosis or differentiation. Many of the genes with differential expression profiles, such as NR4A1, were immune cell specific. Others, such as TLN1, are related to cancerous lymphocytes. Interestingly, cells treated with either DSF analog had more DEGs related to DNA damage pathways compared to cells treated with DSF, which suggests potential mechanisms of action that differ from those of DSF. These differences highlight how small structural changes might significantly affect drug promiscuity. Further work to discover the specific binding partners of these analogs may shed light on their mechanisms of action and identify opportunities for lead compound optimization.

Acknowledgments

The authors are grateful to Dr. Ali Hamiche, Dr. Abdulkhaleg Ibrahim and Dr. Fathi Salama from University of Strasbourg, France, for their help in RNA-Seq experiments. The authors also acknowledge the support provided by the Research, Development, and Innovation Authority (RDIA), Kingdom of Saudi Arabia (Grant Number: 12840-KSAUHS-2023-KAIMR-R-2-1-HW-). This work was funded by King Abdullah International Medical Research Center (KAIMRC), Kingdom of Saudi Arabia (Grant Number: NRC23J/423/05).

Conflict of Interest

The authors declare no conflict of interest.

REFERENCES

• 1) Bell RG, Smith HW. Preliminary report on clinical trials of antabuse. Can. Med. Assoc. J., 60, 286–288 (1949).
• 2) Kita Y, Hamada A, Saito R, Teramoto Y, Tanaka R, Takano K, Nakayama K, Murakami K, Matsumoto K, Akamatsu S, Yamasaki T, Inoue T, Tabata Y, Okuno Y, Ogawa O, Kobayashi T. Systematic chemical screening identifies disulfiram as a repurposed drug that enhances sensitivity to cisplatin in bladder cancer: a summary of preclinical studies. Br. J. Cancer, 121, 1027–1038 (2019).
• 3) Hu H, Cui L, Lu J, Wei K, Wei J, Li S, Zou C, Chen T. Intestinal microbiota regulates anti-tumor effect of disulfiram combined with Cu(2+) in a mice model. Cancer Med., 9, 6791–6801 (2020).
• 4) Wang R, Shen J, Yan H, Gao X, Dong T, Wang P, Zhou J. The evolving role of disulfiram in radiobiology and the treatment of breast cancer. Onco Targets Ther., 13, 10441–10446 (2020).
• 5) Hu Y, Qian Y, Wei J, Jin T, Kong X, Cao H, Ding K. The disulfiram/copper complex induces autophagic cell death in colorectal cancer by targeting ULK1. Front. Pharmacol., 12, 752825 (2021).
• 6) Guo F, Yang Z, Kulbe H, Albers AE, Sehouli J, Kaufmann AM. Inhibitory effect on ovarian cancer ALDH+ stem-like cells by Disulfiram and Copper treatment through ALDH and ROS modulation. Biomed. Pharmacother., 118, 109371 (2019).
• 7) Yoshino H, Yamada Y, Enokida H, Osako Y, Tsuruda M, Kuroshima K, Sakaguchi T, Sugita S, Tatarano S, Nakagawa M. Targeting NPL4 via drug repositioning using disulfiram for the treatment of clear cell renal cell carcinoma. PLOS ONE, 15, e0236119 (2020).
• 8) Guo X, Xu B, Pandey S, Goessl E, Brown J, Armesilla AL, Darling JL, Wang W. Disulfiram/copper complex inhibiting NFkappaB activity and potentiating cytotoxic effect of gemcitabine on colon and breast cancer cell lines. Cancer Lett., 290, 104–113 (2010).
• 9) Kona FR, Buac D, Burger MA. Disulfiram, and disulfiram derivatives as novel potential anticancer drugs targeting the ubiquitin-proteasome system in both preclinical and clinical studies. Curr. Cancer Drug Targets, 11, 338–346 (2011).
• 10) Lewis DJ, Deshmukh P, Tedstone AA, Tuna F, O’Brien P. On the interaction of copper(II) with disulfiram. Chem. Commun. (Camb.), 50, 13334–13337 (2014).
• 11) Tawari PE, Wang Z, Najlah M, Tsang CW, Kannappan V, Liu P, McConville C, He B, Armesilla AL, Wang W. The cytotoxic mechanisms of disulfiram and copper(ii) in cancer cells. Toxicol. Res. (Camb.), 4, 1439–1442 (2015).
• 12) Kannappan V, Ali M, Small B, Rajendran G, Elzhenni S, Taj H, Wang W, Dou QP. Recent advances in repurposing disulfiram and disulfiram derivatives as copper-dependent anticancer agents. Front. Mol. Biosci., 8, 741316 (2021).
• 13) Zhang S, Zong Y, Chen L, Li Q, Li Z, Meng R. The immunomodulatory function and antitumor effect of disulfiram: paving the way for novel cancer therapeutics. Discov. Oncol., 14, 103 (2023).
• 14) McConchie RD, Panitz DR, Sauber SR, Shapiro S. Disulfiram-induced de novo seizures in the absence of ethanol challenge. J. Stud. Alcohol, 44, 739–743 (1983).
• 15) Erensoy H, Sargin AE. Epileptic seizure due to disulfiram treatment. Korean J. Fam. Med., 40, 406–408 (2019).
• 16) Chick J. Safety issues concerning the use of disulfiram in treating alcohol dependence. Drug Saf., 20, 427–435 (1999).
• 17) Chen CH, Ferreira JC, Gross ER, Mochly-Rosen D. Targeting aldehyde dehydrogenase 2: new therapeutic opportunities. Physiol. Rev., 94, 1–34 (2014).
• 18) Omran Z. New disulfiram derivatives as MAGL-selective inhibitors. Molecules, 26, 3296 (2021).
• 19) Omran Z. Development of new disulfiram analogues as ALDH1a1-selective inhibitors. Bioorg. Med. Chem. Lett., 40, 127958 (2021).
• 20) Omran Z. Novel disulfiram derivatives as ALDH1a1-selective inhibitors. Molecules, 27, 480 (2022).
• 21) Omran Z, U. S. Patent US11753371B2 (2023).
• 22) Moore SA, Baker HM, Blythe TJ, Kitson KE, Kitson TM, Baker EN. Sheep liver cytosolic aldehyde dehydrogenase: the structure reveals the basis for the retinal specificity of class 1 aldehyde dehydrogenases. Structure, 6, 1541–1551 (1998).
• 23) Ors A, Papin C, Favier B, Roulland Y, Dalkara D, Ozturk M, Hamiche A, Dimitrov S, Padmanabhan K. Histone H3.3 regulates mitotic progression in mouse embryonic fibroblasts. Biochem. Cell Biol., 95, 491–499 (2017).
• 24) Martin M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet. J., 17, 10–12 (2011).
• 25) Langmead B, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nat. Methods, 9, 357–359 (2012).
• 26) Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, Gingeras TR. STAR: ultrafast universal RNA-seq aligner. Bioinformatics, 29, 15–21 (2013).
• 27) Wang L, Wang S, Li W. RSeQC: quality control of RNA-seq experiments. Bioinformatics, 28, 2184–2185 (2012).
• 28) Anders S, Huber W. Differential expression analysis for sequence count data. Genome Biol., 11, R106 (2010).
• 29) Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol., 15, 550 (2014).
• 30) Fadloun A, Le Gras S, Jost B, Ziegler-Birling C, Takahashi H, Gorab E, Carninci P, Torres-Padilla ME. Chromatin signatures and retrotransposon profiling in mouse embryos reveal regulation of LINE-1 by RNA. Nat. Struct. Mol. Biol., 20, 332–338 (2013).
• 31) Li H, Durbin R. Fast and accurate short read alignment with Burrows–Wheeler transform. Bioinformatics, 25, 1754–1760 (2009).
• 32) Jurka J, Kapitonov VV, Pavlicek A, Klonowski P, Kohany O, Walichiewicz J. Repbase update, a database of eukaryotic repetitive elements. Cytogenet. Genome Res., 110, 462–467 (2005).
• 33) Piper M, Mistry M, Liu J, Gammerdinger W, Khetani R. hbctraining/scRNA-seq_online: scRNA-seq lessons from HCBC (first release). Zenodo, https://doi.org/10.5281/zenodo.5826256 (2022).
• 34) Zhu A, Ibrahim JG, Love MI. Heavy-tailed prior distributions for sequence count data: removing the noise and preserving large differences. Bioinformatics, 35, 2084–2092 (2019).
• 35) Rainer J, Gatto L, Weichenberger CX. Ensembldb: an R package to create and use Ensembl-based annotation resources. Bioinformatics, 35, 3151–3153 (2019).
• 36) Wang Y, Duan M, Peng Z, Fan R, He Y, Zhang H, Xiong W, Jiang W. Advances of DNA damage repair-related drugs and combination with immunotherapy in tumor treatment. Front. Immunol., 13, 854730 (2022).
• 37) Cibrián D, Sanchez-Madrid F. CD69: from activation marker to metabolic gatekeeper. Eur. J. Immunol., 47, 946–953 (2017).
• 38) Filén S, Ylikoski E, Tripathi S, West A, Bjorkman M, Nystrom J, Ahlfors H, Coffey E, Rao KV, Rasool O, Lahesmaa R. Activating transcription factor 3 is a positive regulator of human IFNG gene expression. J. Immunol., 184, 4990–4999 (2010).
• 39) Liu X, Wang Y, Lu H, et al. Genome-wide analysis identifies NR4A1 as a key mediator of T cell dysfunction. Nature, 567, 525–529 (2019).
• 40) Si M, Lang J. The roles of metallothioneins in carcinogenesis. J. Hematol. Oncol., 11, 107 (2018).
• 41) Dai H, Wang L, Li L, Huang Z, Ye L. Metallothionein 1: a new spotlight on inflammatory diseases. Front. Immunol., 12, 739918 (2021).
• 42) Yan DW, Fan JW, Yu ZH, Li MX, Wen YG, Li DW, Zhou CZ, Wang XL, Wang Q, Tang HM, Peng ZH. Downregulation of metallothionein 1F, a putative oncosuppressor, by loss of heterozygosity in colon cancer tissue. Biochim. Biophys. Acta Mol. Basis Dis., 1822, 918–926 (2012).
• 43) Cui D, Cui X, Xu X, Zhang W, Yu Y, Gao Y, Mei C, Zheng W. Identification of TLN1 as a prognostic biomarker to effect cell proliferation and differentiation in acute myeloid leukemia. BMC Cancer, 22, 1027 (2022).
• 44) Tello-Lafoz M, Srpan K, Sanchez EE, Hu J, Remsik J, Romin Y, Calo A, Hoen D, Bhanot U, Morris L, Boire A, Hsu KC, Massague J, Huse M, Er EE. Cytotoxic lymphocytes target characteristic biophysical vulnerabilities in cancer. Immunity, 54, 1037–1054.e7 (2021).
• 45) Evans DR, Bein K, Guy HI, Liu X, Molina JA, Zimmermann BH. CAD gene sequence and the domain structure of the mammalian multifunctional protein CAD. Biochem. Soc. Trans., 21, 186–191 (1993).