N-Acyloxymethyl-phthalimides deliver genotoxic formaldehyde to human cells

Formaldehyde is a pollutant and human metabolite that is toxic at high concentrations. Biological studies on formaldehyde are hindered by its high reactivity and volatility, which make it challenging to deliver quantitatively to cells. Here, we describe the development and validation of a set of N-acyloxymethyl-phthalimides as cell-relevant formaldehyde delivery agents. These esterase-sensitive compounds were similarly or less inhibitory to human cancer cell growth than free formaldehyde but the lead compound increased intracellular formaldehyde concentrations, increased cellular levels of thymidine derivatives (implying increased formaldehyde-mediated carbon metabolism), induced formation of cellular DNA-protein cross-links and induced cell death in pancreatic cancer cells. Overall, our N-acyloxymethyl-phthalimides and control compounds provide an accessible and broadly applicable chemical toolkit for formaldehyde biological research and have potential as cancer therapeutics.


Synthesis
N-acyloxymethyl-phthalimides were synthesised using the following procedure.Carboxylic acid (1.1 mmol) was added to a solution of (N-bromomethyl)-phthalimide (240 mg, 1 mmol) in EtOAc (5 mL) in a glass vial.DIPEA (192 µL, 1.1 mmol) was then added dropwise and the mixture was stirred at room temperature (8-12 hours).CH2Cl2 (10 mL) was then added and the mixture was transferred to a round-bottomed flask.Removal of the solvent in vacuo gave a white solid, which was redissolved in CH2Cl2 and purified by flash chromatography on silica gel (conditions: 95 % CH2Cl2, 5 % MeOH, 10 CV).The product was isolated after evaporation of the solvent in vacuo.
For spectroscopic characterisation of all compounds, as well as the procedures for the syntheses of PGlycAc and POxoBut, please see below.The purity of all compounds was ≥95 %.

Stability Studies
N-acyloxymethyl-phthalimide or control compound (50 µL of a 100 mM stock in DMSO), TSP (10 µL of a 10 mg/mL solution in H2O), D2O (65 µL) and Dulbecco Modified Eagle Medium media with 10 % v/v heat-inactivated foetal bovine serum (DMEM with FBS, 600 µL) was mixed in a microcentrifuge tube, transferred to a 5 mm diameter NMR tube and incubated at 37 °C for 24 hours.Due to the poor solubility of PFProp, 575 µL of DMEM with FBS containing 25 µL of DMSO was used.Degradation of the compounds was observed by 1 H NMR spectroscopy with water suppression (pre-saturation) using a Bruker AV500 NMR spectrometer equipped with a BBO probe and installed with TopSpin 4.1 software.After 24 hours, dimedone (7.5 µL of a 1 M stock in H2O), was added to each mixture and the samples were re-analysed by 1 H NMR spectroscopy.Characterisation of the PFAc degradation product were conducted on a sample containing PFAc (50 µL of a 100 mM stock in DMSO-D6), TSP (10 µL of a 10 mg/mL stock in D2O), D2O (150 µL), and 100 mM sodium phosphate buffer in D2O pH 7.5 (390 µL).After mixing in a microcentrifuge tube, the sample was transferred to a 5 mm NMR tube and incubated at 37 °C for 24 hours. 1 H-13 C-HSQC and 1 H-13 C-HMBC Electronic Supplementary Material (ESI) for Chemical Science.This journal is © The Royal Society of Chemistry 2023 analyses were then conducted on a Bruker Avance NEO 800 MHz spectrometer equipped with a TCI cryoprobe.

NMR analyses with porcine esterase
Samples for NMR analyses (600 µL total volume) contained porcine esterase (10 µL of a 60 mM stock in 100 mM phosphate buffer in H2O pH 7.5), an N-acyloxymethyl-phthalimide or control compound (0.6 µL or 50 µL of a 100 mM stock in DMSO), DMSO (0 µL -30 µL), either H2O (0.6 µL) or dimedone (0.6 µL of a 100 mM stock in H2O), D2O (150 µL), and 100 mM phosphate buffer in H2O pH 7.5 (355 -434 µL).After mixing in microcentrifuge tubes, the samples were transferred to 5 mm diameter NMR tubes and monitored by 1 H NMR with water suppression (pre-saturation) using a Bruker AV500 NMR spectrometer equipped with a 5 mm BBO probe and installed with TopSpin 3.6.1 software.Each 1 H NMR experiment consisted of 16 transients and 224 seconds of acquisition time.For time-course studies, each analysis consisted of 29 1 H NMR experiments running consecutively, giving a total acquisition time of 109 -120 minutes.Reaction rates were determined by measuring the intensities of the 1 H resonances corresponding to the carboxylate products over time and comparing them to the intensity of the 1 H resonance of the TSP standard.Rates are reported in µM s -1 .

Growth inhibition assays
Human U2OS (osteosarcoma) cells were seeded in 10 cm diameter petri dishes and incubated in 10 mL of DMEM with FBS for 24 hours at 37 °C.The dishes were then treated (10 µL) with either DMSO, HCHO in DMSO, an N-acyloxymethyl-phthalimide in DMSO or control compound in DMSO (5-100 µM final concentration).The dishes were then incubated for 12-16 days at 37 °C (5 % CO2) to allow growth of colonies.The dishes were then drained of media and washed with phosphate-buffered saline (PBS).Coomassie Blue stain (0.1 % Coomassie Brilliant Blue R-250, 7 % glacial acetic acid, 50 % methanol in H2O) was added to stain colonies before draining and washing with H2O.Colonies were counted using an Oxford Optronix Colcount cell counter.

Mass spectrometry analyses
U2OS cells were grown to confluence in 25 cm 2 flasks at 37 °C containing Dulbecco Modified Eagle Medium containing 10 % v/v heat-inactivated foetal bovine serum before addition of HCHO or PFAc (100 µM final concentration).After two hours incubation at 37 °C, the media was removed and the cells were washed with PBS.A 4:1 CH3CN:H2O mix containing 1 mM dimedone (500 µL) was then added to each flask to induce cell lysis and scavenge HCHO.The mix was pipetted into a microcentrifuge tube and centrifuged at 13000 rpm for 30 minutes.The supernatant was then removed and filtered through a 10 kDa cut-off syringe filter.Samples were stored at -80 °C until the day of analysis.Each sample was thawed and then analysed by ion-exchange chromatography coupled directly to high-resolution mass spectrometry using the method previously described [1] .P values (here and elsewhere) were determined by the ANOVA significance test using GraphPad Prism 9.3.1.

Detection of DPCs
DPCs were measured as previously described [2] .Briefly, 5 x 10 5 U2OS cells either untreated or treated for 2 hours with either DMSO (carrier control), HCHO, an N-acyloxymethylphthalimide, or control compound, and were then washed twice with DMEM with FBS and lysed by scraping in 400 μL of denaturing lysis buffer (2 % SDS, 20 mM Tris.HCl pH 7.5).The cell lysate was then frozen at -80 ºC.Sample processing involved thawing at 55 ºC for 5 minutes and sonication at maximum attenuation for three cycles of 20 seconds on, 20 seconds off (on ice).This lysate was then pelleted at 15000 x g.The supernatant was then removed and saved as the soluble DNA fraction.The pellet fraction was washed three times by being resuspended in 400 μL of TK buffer (20 mM Tris pH 7.5, 200 mM KCl), shaken at 55 ºC for 5 minutes, cooled on ice for 5 minutes and then pelleted at 15000 x g for 5 minutes.The samples were resuspended in TK buffer and proteins digested with 0.2 mg/mL Proteinase K at 55 ºC for 45 minutes.Bovine serum albumin (BSA, 500 μg) was added to each sample, which was then cooled on ice for 5 minutes before undergoing final centrifugation.The supernatant containing the initially cross-linked DNA was then treated with RNAse A (0.2 mg/mL) for 30 minutes at 37 ºC.DNA concentrations of the soluble fraction and the cross-linked fraction (20 μL each) were determined using the Qubit dsDNA HS assay according to the manufacturer's instructions (ThermoFisher Scientific) and presented as a ratio.With the exception of the experiment with P + F, all experiments were conducted at least twice on different occasions.Averaged values and standard deviations are calculated from all replicates (n = 2-6).

Cytotoxicity assays
Cell viability was determined by a colorimetric MTT assay.Cell suspension (100 µL, 1 x 10 5 cells/mL) was added each well of a 96-well plate and incubated overnight at 37 °C, 5 % CO2 to allow cells to adhere.Media was aspirated and cells were treated with varying concentrations of HCHO or phthalimide-containing compound in fresh media (200 µL with 1 % DMSO), and then incubated for 72 hours at 37 °C, 5 % CO2.After treatment, 20 µL of MTT solution (5 mg/mL in PBS) was added to each well and incubated for a further 4 hours.The media was then aspirated and the resultant formazan crystals were dissolved in DMSO (50 µL).The absorbance intensity at 570 nm was then measured using a Hidex Sense microplate reader.All experiments were performed in triplicate and the relative cell viability (%) was expressed as a percentage relative to the untreated control cells.No colour change was observed in control experiments containing MTT and either HCHO or PFAc (no cells, 4 hours incubation at 37 °C).

Figure
Figure S3. 1 H NMR spectra of samples containing PFFor, PFAc, PFProp, PGlycAc, POxoBut and HCHO in DMEM with FBS after addition of dimedone. 1H resonances corresponding to the dimedone-HCHO adduct are highlighted.Note the sample with PFFor was incubated for 27 hours before addition of dimedone, whereas the samples with PFAc, PFProp, PGlycAc, POxoBut and HCHO were incubated for 24 hours.Note HCHO scavenging by dimedone is unlikely to be quantitative in these complex mixtures due to competing reactions with residual phthalimide (as is present in the sample with PFFor) and media components.The sample with HCHO was with a different DMEM with FBS batch of marginally higher concentration.

Figure S5 .
Figure S5.LC/MS analysis of a sample of PFAc in 1:1 water:acetonitrile after 24 hours.The presence of both PFAc and its N-acyloxymethyl-phthalamic acid derivative is evidenced by peaks after 1.64 minutes and 1.17 minutes respectively.Note degradation of PFAc is less proficient in water/acetonitrile than in water, 100 mM phosphate buffer pH 7.5, or DMEM with FBS.

Figure S6 .
Figure S6.LC/MS analysis of a sample of PFProp in 1:1 water:acetonitrile after 24 hours.The presence of both PFProp and its N-acyloxymethyl-phthalamic acid derivative is evidenced by peaks after 1.82 minutes and 1.36 minutes respectively.Note degradation of PFProp is less proficient in water/acetonitrile than in water, 100 mM phosphate buffer pH 7.5, or DMEM with FBS.

Figure S10 .
Figure S10.(A) Graph showing time-dependent hydrolysis of PGlycAc catalysed by porcine esterase.(N-hydroxyethyl)-phthalimide and acetate are produced during catalysis.(B) Graph showing time-dependent hydrolysis of PFAc catalysed by porcine esterase in the presence of excess dimedone.Formation of acetate and hydroxymethyl-dimedone are highlighted.(Nhydroxymethyl)-phthalimide was only observed at low levels.(C) 1 H NMR spectra showing time-dependent hydrolysis of PFAc catalysed by porcine esterase in the presence of excess dimedone. 1H resonances corresponding to acetate and hydroxymethyl-dimedone are highlighted.

Figure S12 .
Figure S12.MTT cytotoxicity assays showing concentration-dependent toxicity of HCHO, PFAc, PGlycAc and P + C in MIA-PaCa-2 pancreatic cancer cells.Significant toxicity was only observed at 1 mM in all samples.Error bars represent standard deviations (n = 3).

Table S1 .
Raw abundance values for hydroxymethyl-dimedone and thymidine derivatives from the metabolomics analyses.Five biological replicates were analysed for each experiment.

Table S2 .
DPC ratios calculated for each replicate from the DPC detection analyses.With the exception of P + F, experiments were conducted on at least two separate occasions.