Printer center nanoparticles alter the DNA repair capacity of human bronchial airway epithelial cells

Abstract

Nano-enabled, toner-based printing equipment emit nanoparticles during operation. The bioactivity of these nanoparticles as documented in a plethora of published toxicological studies raises concerns about their potential health effects. These include pro-inflammatory effects that can lead to adverse epigenetic alterations and cardiovascular disorders in rats. At the same time, their potential to alter DNA repair pathways at realistic doses remains unclear. In this study, size-fractionated, airborne particles from a printer center in Singapore were sampled and characterized. The PM_0.1 size fraction (particles with an aerodynamic diameter less than 100 nm) of printer center particles (PCP) were then administered to human lung adenocarcinoma (Calu-3) or lymphoblastoid (TK6) cells. We evaluated plasma membrane integrity, mitochondrial activity, and intracellular reactive oxygen species (ROS) generation. Moreover, we quantified DNA damage and alterations in the cells’ capacity to repair 6 distinct types of DNA lesions. Results show that PCP altered the ability of Calu-3 cells to repair 8oxoG:C lesions and perform nucleotide excision repair, in the absence of acute cytotoxicity or DNA damage. Alterations in DNA repair capacity have been correlated with the risk of various diseases, including cancer, therefore further genotoxicity studies are needed to assess the potential risks of PCP exposure, at both occupational settings and at the end-consumer level.

1. Introduction

Nano-enabled toners are widely used in printing equipment, such as photocopiers and laser printers found in many indoors environments, including professional printer centers, offices, and homes (Bello et al. 2013; Martin et al. 2015, 2017; Zhang et al. 2019). During controlled operation in test chambers, laser printers have been shown to emit nanoparticles, termed printer-emitted nanoparticles (PEPs), at high particle concentration numbers (S. V. Pirela et al. 2014, 2015). Emerging occupational exposure studies in printer centers show high concentrations of potentially hazardous particles, termed printer center particles (PCP), linked to the type of printers/photocopiers and nano-enabled toners used in such settings (Martin et al. 2015; Setyawati et al. 2020). The potentially hazardous chemical composition, inhalable size, and ubiquity of PCP and PEPs have recently raised serious environmental health and safety concerns for workers at printer centers and home users alike (S. V. Pirela et al. 2017).

Previous studies from the authors and others have shown that PEPs are released at concentrations that can reach up to 1.3×10⁶ particles cm⁻³ with sub-100 nm mobility diameters [6]. Metal nano-additives, like chromium, nickel, copper, and zinc, as well as polycyclic aromatic hydrocarbons (PAH) were also detected in separate studies (S. V. Pirela et al. 2015, 2017), showing that PEPs’ chemical composition is potentially hazardous and their toxicological evaluation is warranted (Martin et al. 2015).

In vitro cellular studies on PEPs isolated in controlled printer chambers showed that these particles are highly bioactive, promote the release of pro-inflammatory cytokines, and lead to p53 up-regulation in THP-1 human derived cells (Khatri, Bello, Pal, et al. 2013). Exposure of primary human small airway epithelial cells (SAEC) to the same type of particles showed that PEPs deposited on the epithelial side can induce adverse cytotoxic effects to the underlying endothelium through secretion of cytokines (Sisler et al. 2015). Significant adverse effects on the epigenome and accumulation of hereditary genomic changes were also observed using the same in vitro model (S. V. Pirela, Miousse, et al. 2016).

The bioactivity of PEPs has also been documented in vivo as well as in clinical case studies. For example, acute exposure to PEPs by intratracheal instillation to BALB-c mice led to a clear increase in neutrophils and pro-inflammatory cytokines in their bronchoalveolar lavage, as shown in a study by Pirela et al. (S. Pirela et al. 2013). A study by Lu et al. using the same animal model detected alterations in DNA methylation, general lung cytotoxicity, as well as epigenetic effects in peripheral blood following exposure to PEPs (Lu et al. 2016). Furthermore, a sub-acute study (21 days) on Sprague Dawley rats confirmed the pro-inflammatory effects and oxidative stress brought on by exposure to PEPs (Guo et al. 2019), while a 72-day study on the same animal model and particle type detected impaired ventricular performance, repolarization, and induced arrhythmias and hypertension, among other adverse effects (Carll et al. 2020). Photocopier operators exposed to PCP exhibited chronic upper airway inflammation underpinned by systemic oxidative stress, as shown in a relevant clinical study (Khatri et al. 2017a). Moe recently, RNA sequencing of blood samples from workers in the printing industry showed transcriptional changes that could promote lymphocyte infiltration through the microvasculature, cardiovascular impairment, and cancer, among others (Guo et al. 2020).

The effects of PEPs on genome expression and stability as reported in the literature are particularly worrisome. In one such study, Sprague Dawley rats exposed to whole-body inhalation of PEPs exhibited transcriptomic changes to mRNA and miRNA related to carcinogenic events and onset of diabetes, among other diseases (Guo et al. 2019). Yet another study on intratracheally exposed Balb-c mice to PEPs revealed the particles’ capacity to instigate oxidative damage on genes involved in DNA repair in the absence of considerable lung cytotoxicity (S. V. Pirela, Lu, et al. 2016). DNA damage has long been identified as a serious result of increased oxidative stress induced by exogenous stressors, like ambient particles (Møller et al. 2014), engineered nanomaterials (ENMs) (Sotiriou et al. 2014; Watson et al. 2014), and ionizing radiation (Koturbash et al. 2006; Soren et al. 2019; Toprani and Das 2015b, 2015a). Occupational exposure studies have confirmed the genotoxic potential of PCP on the human genome, elevation of chromosomal aberrations, and increased frequency of micronuclei in the peripheral blood samples of workers occupationally involved in photocopying (Goud et al. 2004). Lastly, in vitro studies have shown how PEP exposure may alter DNA methylation patterns and expression of DNA damage-related genes in SAEC (S. V. Pirela, Miousse, et al. 2016) while their organic extracts have been genotoxic against A549 cells (Gminski et al. 2011).

It is well-known that cells can repair DNA damages and avoid cell death. Balance between DNA damage and DNA repair is key to the cell fate and has even been described as the etiology of several human diseases (Toprani and Das 2020). Even more so, altered DNA repair capacity (DRC) has been linked to carcinogenesis (Nagel, Chaim, and Samson 2014), but also drug resistance to cancer treatment (McFaline-Figueroa et al. 2015; Toprani and Das 2020). A few studies have already reported that environmental stimulants, such as ENMs, and e-cigarettes may dampen DRC (Lee et al. 2018; Singh et al. 2017). To our knowledge, while the general bioactivity of PCPs has been documented in numerous cellular and animal toxicological studies, it is currently unknown whether such particulate exposure has any impact on DNA repair. Here, a recently developed method was used to quantify for the first time the effects of PCP to the DNA-repair capacity of human bronchial epithelial and lymphoblastoid cells at occupationally relevant, sub-cytotoxic doses.

In this study, Calu-3 and TK6 human lymphoblastoid cells were exposed to the nano-sized fraction of PCPs sampled from a printer center in Singapore in order to assess the genotoxicity and specifically quantify any alterations in DRC induced at sub-cytotoxic levels of particle exposure. To this end, a battery of toxicological analyses was performed to verify the absence of acute cytotoxicity of PCP at the employed doses. In turn, a high-throughput fluorometric method recently developed by the authors (Toprani et al. 2021) was employed to screen for changes in DRC induced by PCP. In brief, fluorescent plasmid reporters were used to transfect live cells exposed to the collected PCP, thus allowing for the direct and simultaneous quantification of six distinct DNA repair pathways. ZnO ENMs were also assessed using the same techniques to compare the bioactivity of PCP against a model nanoparticle for which there is a breadth of genotoxicity and cytotoxicity literature available. This study is part of a greater occupational cohort study launched in Singapore to assess potential health implications of exposed workers. The PCP exposure assessment in multiple printer centers in Singapore was recently published by the authors (Setyawati et al. 2020).

2. Methods

2.1. Collection and fractionation (PM_0.1) of PCP

A printer center at an industrial complex area in Singapore was recruited for the purposes of this study. Detailed background information about this printer center (labelled as center #2) is available in a companion study recently published by the authors on six printer centers in Singapore (Setyawati et al. 2020). In brief, the printer center operated without windows and was equipped with a single door which remained shut for the duration of printing jobs. Given the circumstances, the ratio between required and real ventilation rates of the print room was as high as 8.2. The weekly indoor air quality indicators in the printer center were 0.74 ± 0.21 ppm of VOC, 0.04 ± 0.02 ppm of NO₂, 0.2 ± 1.0 ppb of O₃, and 494.2 ± 23.1 ppm of CO₂. Total printing volume from 3 color printers was 8,328 - 16,960 pages per day, averaging at 10,465 pages per day with a maximum daily printing production of 17,000.

Airborne particles from the printer center were collected and size-fractionated to PM_>2.5, PM_0.1-2.5, and PM_0.1 using the Harvard Compact Cascade Impactor, as described in previous studies (Bello et al. 2013; Demokritou et al. 2004). The PM_0.1 fraction of PCP was collected on a 47 mm diameter pre-cleaned polytetrafluoroethylene (PTFE) membrane of 2 μm pore size (Pall Corporation, Port Washington, NY, USA). PCPs were then extracted from the PTFE membrane as previously reported in detail (Pal et al. 2015). Briefly, the PTFE filters were placed in beakers containing a small volume of 75% vol/vol ultrapure ethanol. They were then sonicated in a bath sonicator for 30-60 s with the particle side facing up so that dislodged particles would not get re-trapped to the filters due to the turbulent flow inside the beaker. Extracted particles were concentrated with 3 rounds of rotary evaporation. Endotoxin-free water (First Base, Singapore) was employed during these repeated evaporation steps to minimize residual ethanol in the final PCP suspension. Bath sonication was employed throughout the evaporation process (~90 min) to keep the PM_0.1 well suspended. A clean filter was treated according to the procedure described above and the retrieved solution was used as background control when assessing biological endpoints (see section 2.2 below).

2.2. Physicochemical characterization, microbiological sterility, and endotoxin assessment of PCP

The physicochemical and morphological properties of PCPs are presented in detail in a previous publication by the authors (Setyawati et al. 2020). In summary, the shape and absolute size of PCP were determined with transmission electron microscope (TEM, Carl Zeiss Libra 120 Plus). Briefly the particles were dispersed in ethanol (100%) and dropped on carbon coated TEM grid with ample drying in a convection oven. The particles were then imaged under accelerating voltage of 120kV. A scanning mobility particle sizer (NanoScan SMPS, TSI Inc., USA) was used to determine their aerodynamic size distribution and number concentration. The elemental composition of PCP was quantified using inductively coupled plasma mass spectrometry (ICP-MS). Finally, the levels of PAHs were measured using a gas chromatograph (Agilent GC-6890 PLUS) and a mass selective detector (Agilent GC/MS 5973N). Endotoxin assessment was performed using the Recombinant Factor C assay (Lonza PyroGene^® kit) following the manufacturer's instructions. More details in the characterization methods employed for PCP can be found in a companion study by Setyawati et al. (Setyawati et al. 2020).

2.3. Physicochemical and biochemical assessment of ZnO ENMs

ZnO ENMs were used as a comparator ENM in this study due to the availability of plethora of historical toxicological data including data on genotoxicity (Chang et al. 2012; Demir, Creus, and Marcos 2014; Heim et al. 2015; Sotiriou et al. 2014). Therefore, dose-matched cytotoxicity and genotoxicity assays were used on both ZnO ENMs and PCP to put in perspective the biological effects of the latter. ZnO ENMs were procured by Sigma Aldrich and characterized at the Harvard T.H. Chan School of Public Health as part of the Nanotechnology Health Implications Research (NHIR) Consortium established by the National Institute of Environmental Health Sciences (NIEHS).

ZnO ENMs were characterized using multiple state-of-the-art techniques, including TEM, X-ray diffraction, and N₂ adsorption to measure absolute particle size distributions, crystal structure, and specific surface area and porosity, respectively. In brief, the shape and size of ZnO ENMs were determined by TEM following the same procedure described above for PCP. X-ray diffraction patterns of ZnO ENMs were collected in 2θ at 10–90° configuration with a 0.02° increment and 2s step and analyzed on EVA software. Finally, N₂ adsorption and desorption were analyzed using the Brunauer–Emmett–Teller (BET, Quantachrome Instruments, NovaTouch LX4) by calculation of a 5-point BET isotherm. Average primary particle diameter (d_BET, nm) was calculated according to the following formula: d_BET = 6000 SSA⁻¹ ρ⁻¹, where ρ is the material density in g cm⁻³. Importantly for toxicological studies, ZnO ENM powder was tested for microbiological sterility and endotoxin presence according to the U.S. Pharmacopeia protocol for sterility (WHO document QAS/11.413). Endotoxin assessment was performed using the Recombinant Factor C assay (Lonza PyroGene^® kit) following the manufacturer's instructions.

2.4. Modeling the deposition dose of PCP along the human respiratory tract using the Multiple-Path Particle Dosimetry Model

In order to estimate a realistic range of PCP doses to be employed for the in vitro experiments of this study, the deposition of airborne PCP in the human respiratory tract was modeled using the multiple-path particle deposition model (MPPD, v3.041) (Price et al. 2002). The assumed airway morphometry for the MPPD simulation was based on the age (21-y.o.) symmetric model while the aerosol and exposure conditions were sourced from a companion study (Setyawati et al. 2020). To use relevant in vitro PCP doses for the toxicological and genotoxic assessment presented in Section 2.7, the deposition mass fluxes (μg m⁻² min⁻1) in the conducting, transitional, and respiratory zones (excluding the head airway region) as predicted by MPPD were summed. The total deposited PCP mass per lung surface area was calculated using the MPPD model at 0.026 μg cm⁻² by integrating 40 hours of exposure (i.e., five 8-hour shifts). Two doses were used in the vicinity of the MPPD estimated dose: a low dose at 0.016 μg cm-2 and a high dose at 0.16 μg cm⁻² were used for the cytotoxic and genotoxic evaluation of PCP, as described in Section 2.6. All employed parameters for the MPPD modeling are summarized in Table S1. Finally, the particokinetics of PCP were calculated and used to achieve the desired deposited mass of PCP on the surface of SAEC as described in the next section.

2.5. Dispersion preparation, colloidal characterization, and in vitro dosimetric analysis of PCP and ZnO

The dispersion preparation and colloidal characterization of PCP and ZnO ENMs were done in accordance to previously established protocols (Cohen et al. 2018; DeLoid et al. 2017). In brief, one milliliter of PCP or ZnO ENM suspensions at 0.5 mg mL⁻¹ in deionized (DI) water were submitted to cup-horn sonication (Branson Sonifier S-450D, 400 W, with Branson 3-in. cup horn, power delivered: 1.26W), followed by 30 s of high-speed vortexing. Between each round of cup-horn sonication/vortexing, the hydrodynamic diameter (z-average, d_H) and polydispersity index (PDI) of suspended particles were measured by dynamic light scattering (Zetasizer Nano ZS, Malvern UK). The procedure was repeated until both d_H and PDI did not change significantly (< 5%), at which point the total delivered acoustic energy is termed critical delivered sonication energy (DSEcr).

Once sonicated at DSE_cr, PCP or ZnO ENMs were added to fully supplemented Eagle's Minimum Essential Medium (EMEM) and RPMI-1640 at final concentration of 0.1 mg mL⁻¹ (see section on cell culture below for more information on growth media composition). Colloidal properties including d_H, ζ-potential, and polydispersity index were measured in DI water and cell culture media for both ZnO ENMs and PCP.

The effective densities of PCP and ZnO ENMs in fully supplemented EMEM were measured according to the volumetric centrifugation method as previously developed by the authors (Deloid et al. 2014). Next, the distorted grid model was executed on MATLAB (MathWorks, Massachusetts, USA) to simulate the particokinetics of PCP or ZnO ENMs in EMEM and calculate the fraction of administered mass (f_D) delivered to the surface of Calu-3 as a function of exposure time, according to a method previously implemented by the authors (DeLoid et al. 2015, 2017).

2.6. Toxicological evaluation of PCP

Cell culture:

Human lymphoblastoid TK6 cells (American Type Culture Collection, USA) were also used here as a cell line typically employed in genotoxicity studies related to chemicals, ionizing radiation, and ENMs (Demir, Creus, and Marcos 2014; Sotiriou et al. 2014; Watson et al. 2014). TK6 cells were cultured in RPMI-1640 medium supplemented with L-glutamine and 10% horse serum and 100 U mL⁻¹ streptomycin-penicillin. Calu-3 cells were also used in this study to evaluate the effects of PCP or ZnO ENMs on cells relevant to inhalation exposure. Calu-3 cells were maintained in EMEM medium fully supplemented with 10% vol/vol fetal bovine serum and 100 U mL⁻¹ streptomycin-penicillin to suppress microbial growth. Both Calu-3 and TK6 cells were cultured under humidified atmosphere of 37°C and 5% CO₂.

It is worth noting that for the toxicological assessment and inter-comparison of biological effects of PCP and ZnO ENMs, Calu-3 cells were exposed to the same deposited particle doses of 0.016 or 0.160 μg/cm² over 24 hours according to their respective f_D, rather than matching administered particle dose. TK6 cells are grown in suspension and, as such, particle dosimetry was not taken into consideration. Instead, TK6 cells were exposed to administered doses of PCP at 1 μg mL⁻¹ for 24 hours based on the knowledge that single-ppm particle concentrations do not generally elicit acute cytotoxic effects in vitro (Sengul and Asmatulu 2020). To get a complete dose-response relationship, administered particle doses of 10 and 100 μg ml⁻¹ were included to reach the high end of the spectrum, where several types of nanoparticles have been shown to be bioactive (Attarilar et al. 2020). Comparator ZnO ENMs were administered to TK6 cells following the same dosing scheme. The toxicological evaluation of PCP and ZnO ENMs was then conducted by assessing the following biological responses:

Mitochondrial activity:

Cell viability and mitochondrial activity was assessed using the Invitrogen^™ PrestoBlue^® assay (Thermo Fisher) according to the manufacturer's protocol. Briefly, after exposure of Calu-3 or TK6 cells to PCP or ZnO ENMs, cells were washed with phosphate-buffered solution (PBS) and incubated with PrestoBlue^® reagent for 30 min. After incubation, fluorescence intensity was measured using a microplate reader (Molecular Devices MX5) at the excitation/emission wavelength pair of 570/610 nm. Cell viability of exposed cells was normalized against the signal obtained from untreated cells set as 100%. To assess potential interferences from particles in the PrestoBlue^® assay, culture media alone and particles suspended in culture media at the highest dose were also measured using the same assay.

Cellular membrane integrity:

The levels of lactate dehydrogenase (LDH) leaked into the cell culture medium were quantified using the Pierce^™ LDH cytotoxicity kit (Thermo Fisher) following the manufacturer’s guidelines. In brief, cell supernatant was aspirated after exposure of Calu-3 or TK6 cells to PCP or ZnO ENMs. The aspirated volume was then centrifuged at 3000xg for 30’ to remove any cells, cellular debris, and unbound particles, and the supernatant was mixed with an equal volume of reaction reagents. The reaction was stopped after 30 minutes, and the solution was applied to absorbance measurement at the wavelength of 490 nm (LDH activity) and 680 nm (background signal from instrument). The absorbance value at 680 nm was subtracted from 490 nm prior to calculation. Then, the value from treated samples was subtracted by the data from untreated wells. Relative LDH levels were expressed as the percentage to the LDH levels from cells treated with lysis buffer for maximum amount of released LDH. To assess potential interferences from particles in the Pierce^™ LDH assay, culture media alone and particles suspended in culture media at the highest dose were also measured using the same assay.

Reactive oxygen species generation:

Intracellular ROS generation was quantified using the CellROX^® Orange or CellROX^® Green assay (Thermo Fisher, Waltham MA). After a 4-hour exposure of Calu-3 or TK6 cells to PCP or ZnO ENMs, cells were washed by PBS and the reagents were added and incubated at 37°C for 30 minutes. The reagent was then washed away, and cells were re-suspended in PBS for fluorescence measurement at the excitation/emission wavelength of 545/565 nm (for CellROX^® Orange) or 485/520 nm (for CellROX^® Green). Cells treated with 100 μM menadione for 1 hour at 37 °C were used as a positive control. ROS levels in ENM-treated cells were expressed as percentage of ROS levels quantified in cells treated with menadione.

DNA damage assessment using the nano-CometChip assay:

An alkaline, single-cell gel electrophoresis assay developed in-house was used to quantify the DNA damage induced by PCP and the ZnO ENM exposure according to a protocol described by Watson et al. (Watson et al. 2014). Briefly, 10⁵ cells per well were loaded for each cell type in 96-well plates and were left to settle over 15 or 30 min for TK6 or Calu-3 cells, respectively. After excess cells were aspirated, the chips were rinsed once with PBS. Molten, low-melting point agarose was then overlaid onto the chip and allowed to set for 2 min at 4 °C. Following this step, the chip was submerged in ice-cold alkaline lysis solution (2.5 M NaCl, 100 mM Na2EDTA, 10 mM Tris, pH 9.5 with 0.5% Triton X-100). After overnight incubation, alkaline buffer (0.3 M sodium hydroxide and 1 mM Na2EDTA in distilled water) was used for the unwinding at 4 °C for 40 min. Electrophoresis was carried out at 4 °C for 30 min at 21 V and 300 mA. Neutralization of the chips was performed with 0.4 M Tris-HCl buffer at pH 7.5. SYBR gold nucleic acid gel staining (Invitrogen, cat. no. S11494) was used to stain the comets which were then fluorescently imaged at 4× magnification using an epifluorescence microscope (Nikon Eclipse 80i, Nikon Instruments, Melville, NY).

Measurement of DNA repair capacity by fluorescence multiplex-host cell reactivation (FM-HCR) assay:

The activity of six DNA repair pathways was measured using a method recently developed by the authors and based on an assay technology developed by Nagel et al. called fluorescence multiplex-host cell reactivation (FM-HCR) (Nagel et al. 2014; Nagel, Chaim, and Samson 2014; Toprani et al. 2021). In brief, several types of reporter plasmids that carry site-specific DNA lesions repairable by their respective DNA repair mechanisms are used to transfect Calu-3 or TK6 cells exposed to PCP. DRC is then assessed by a quantifiable change in the expression of fluorescent reporter proteins. Flow cytometry is used as a high throughput means to assess the ability of cells to carry out DNA repair across large numbers of samples.

In more detail, the fluorescent reporter plasmids were transfected into TK6 cells using electroporation (Gene Pulser MXCell^™ electroporation system, Biorad). Plasmids were introduced into Calu-3 cells using lipid-mediated transfection (Lipofectamine^™ 3000, Thermofisher Scientific). After 24 h, cells were analyzed using flow cytometry and the percent of fluorescent reporter expression was calculated to measure the relative fold-change in DRC after exposure. Six reporter plasmids containing specific DNA lesions were transfected to investigate the repair activity of the DNA repair enzyme or pathway. Reporters included an A:8oxoG lesion repaired primarily by mutY DNA glycosylase (MUTYH), an 8oxoG:C lesion repaired primarily by 8-oxoguanine DNA glycosylase (OGG1), a hypoxanthine lesion repaired primarily by methyl purine DNA glycosylase (MPG), a tetrahydrofuran lesion repaired by long patch base excision repair (LP-BER), a UV-induced lesion repaired by nucleotide excision repair (NER), and a double strand break repaired primarily by non-homologous end joining (NHEJ). In addition to repair capacity measurements, the impact of PCP and ZnO ENMs on the transfection efficiency and fluorescence detection were assessed to exclude the possibility of interference from the particles themselves.

2.7. Statistical analyses

For the FM-HCR and nano-comet chip assay, we used one-way ANOVA comparisons to test for statistical significance between cells exposed to ENMs and control cells. Specifically, we applied one-way ANOVA analyses with Dunnett correction for multiple comparisons for the nano-comet chip assay. Regarding the FM-HCR assay, the internal transfection control signals were used to normalize the fluorescent reporter expression for each plasmid. Then, the fluorescence from damaged plasmids in the ENM-treated cells were normalized against their undamaged counterparts. Finally, percent reporter expression from the ENM-exposed cells was normalized to percent reporter expression in untreated cells. These final normalized percent reporter expression values were compared using an unpaired, parametric t-test. Multiple comparisons between reporter plasmid expressions for different DNA repair pathways were considered independent, therefore statistical significance was set at p ≤ 0.05. The statistical significance of ROS signals was tested using One-Way ANOVA analyses coupled with Dunnett’s correction for multiple comparisons. PrestoBlue data were tested using One-Way ANOVA analyses coupled with Brown-Forsythe and Welch tests. LDH results were tested using One-Way ANOVA coupled with Šídák corrections for multiple comparisons. All statistical analyses were carried out on Prism 8.1.

3. Results

3.1. Collection and physicochemical characterization of PCP

PCP collection and physicochemical characterization has been described in detail in the companion publication by Setyawati et al. (Setyawati et al. 2020). Particle extraction efficiency was verified by weighing the filters pre- and post-sonication and was found to be 97.68% (range: 96.89%-99.01%). The physicochemical properties of PCP are summarized in Figures S1 and S2. In brief, PCP were collected from a printer center that produced high particle emissions (~430,000 particles cm⁻³) with an aerodynamic count median diameter of 30.6 nm. Weekly average of the released particles was 30,569 ± 2.12 cm⁻³ with a maximum of 426,001 cm⁻³. TEM observations showed the presence of irregularly shaped nanoparticle aggregates (Figure S1A). The weight percentage of PCP was 36.4% of the total particulate matter and its mass-weighted concentration was 15.1 μg m⁻³. Particles number concentrations reached during printing were as high as 6.36-fold relative to background levels.

In terms of chemical composition, the mass-weighted fractions of elemental carbon, organic carbon, and inorganic content was 12.3, 38.3 and 49.4%, respectively. A series of metals, metalloids, and non-metals including Mn, Mg, Cr, Ni, Al, Zn, Si, Sn, Ti, and Fe made up the inorganic composition of PCP (Figure S2A and S2B). Mass concentration of Zn reached up to 7 μg per mg PCP. Additionally, PCP was noted to contain 16 EPA priority PAHs at the level of 374 pmol mg⁻¹, with the carcinogenic high molecular weight PAHs (HMPAH, i.e., comprising of 4 or more rings) constituting up to 37 mol% of the total PAH composition (Figure S2C). In the context of organic components, benzo[a]pyrene equivalent of the total PAH in PCP was estimated at 3.8 ng mg⁻¹. Based on the ICP-MS and the elemental and organic carbon content results and according to the density of organic and elemental carbon at 0.8 and 2.0 g cm⁻³ as previously reported by others (Choi et al. 1995), the density of PCP was calculated at ~1.2 g cm⁻³ while endotoxin levels in the suspension were found to be ~6.4 EU mg⁻¹.

In addition, ZnO ENMs were employed as comparator nanoparticles given the available studies on their toxicological and genotoxic properties previously performed by the authors (Konduru et al. 2017; Sotiriou et al. 2014; Watson et al. 2014). A representative TEM picture of ZnO ENMs can be found in Figure S1B. Additional physicochemical and morphological properties for ZnO ENMs have been previously reported in detail by the authors: in brief, ZnO ENMs were found to have a low aspect ratio (1.230 ± 0.148) and near-spherical geometry with a circularity of 0.944 ± 0.025 and roundness of 0.823 ± 0.090 with average diameter of 46 ± 17 nm, average surface area of 15.95 ± 0.80 m² g⁻¹, and 86.2% crystallinity (hexagonal zincite) (Duan et al. 2020; Eweje et al. 2019).

3.2. Calculated lung deposition dose of PCP

The deposited PCP mass fraction (dimensionless) and deposited mass flux (in μg m⁻² min⁻¹) in the human respiratory tract as calculated using the MPPD model are presented in Figure 1A and Figure 1B, respectively. The main PCP aerosol parameters were sourced from a companion study by Setyawati et al. (Setyawati et al. 2020). Their count median diameter was ~31 nm with a geometric standard deviation of 1.29 and the aerosol particle mass concentration was 15.1 μg m⁻³. As shown in Figure 1C, the model predicts 47%, 15%, and 7% total mass fraction deposition in the pulmonary, tracheo-bronchial, and head region, respectively. This deposition pattern is expected due to the small diameter of aerosolized PCP. MPPD model-derived cumulative deposited mass rate in the conducting, transitional, and respiratory zones of the human lung was calculated at 1.08 μg cm⁻² min⁻¹. Assuming no clearance or translocation over a 40-hour exposure (one business week), such mass flux amounted to deposited PCP mass per lung surface area of 0.026 μg cm⁻². Application of MPPD enabled the performance of in vitro toxicological tests using doses (in μg cm⁻²) relevant to those expected to deposit on the alveolar epithelium based on the already measured properties of the PCP aerosol and the working conditions recorded for this particular printer center.

Figure 1:

(A) Modelled deposited PCP mass fraction along the respiratory tract generations. (B) Modelled deposited PCP mass rate per unit area (μg min-1 m-2) along the respiratory tract generations. Generation numbers: 0: trachea; 1-3: bronchi; 4: bronchioles; 5-16: terminal bronchioles; 17-19: respiratory bronchioles; 20-22: alveolar ducts; 23: alveolar ducts. (C) Modelled mass deposition fraction in total lung for head, tracheobronchial region (TB), and pulmonary (P) regions.

3.3. Dispersion, colloidal characterization, and dosimetric analysis of PCP and ZnO ENMs

The dispersion preparation of ZnO ENMs and PCP was carried out based on a protocol previously presented by the authors (Cohen et al. 2018). Regardless of the sonication energy delivered in the collected PCP suspension, their d_H and d.i. ranged from 194-199 nm and 0.213-0.228, respectively, thus setting their DSE_cr at 0 J mL⁻¹, as shown in Figure S3A. Their zeta-potential was strongly negative at −30.0 ± 0.2 mV, suggesting good colloidal dispersion due to electrostatic stabilization. In EMEM, d_H values remained stable (198 nm). Expectedly, the z-potential increased to −8 mV upon Debye screening effects induced by an abundance of ions in the culture medium. A summary of measured colloidal properties of PCP in DI H₂O and fully supplemented EMEM are presented in Table S2. Furthermore, Figure S3B presents the intensity-weighted d_H distributions of PCP in DI H₂O and fully supplemented EMEM.

The DSEcr for ZnO ENMs in DI H₂O is 420 J mL⁻¹, as previously reported (Eweje et al. 2019). Upon dispersion in DI H₂O, ZnO ENMs exhibited typically positive zeta-potential at 18 mV with a d_H of 274 nm. Once added in EMEM, ZnO ENMs remained well-dispersed as evidenced by a marginal increase of d_H up to 284 nm. Interestingly, the particles’ zeta-potential was reversed from 18 mV in water to −12 mV in EMEM, which could be attributed to adsorption of negatively charged serum proteins on the particle surface. A summary of measured colloidal properties of ZnO ENMs in DI H₂O and fully supplemented EMEM are summarized in Table S2. Figure S3C further presents the intensity-weighted d_H distributions of ZnO ENMs in DI H₂O and fully supplemented SAGM.

To enable the toxicological inter-comparison of ZnO ENMs and PCP on the basis of particle mass delivered to cells, the dosimetric analysis of both types of nanoparticles was considered as previously described by the authors (DeLoid et al. 2017). Owing to the minimal agglomeration of ZnO ENMs and PCP in EMEM, their effective density (ρ_eff) was considered to be same as that of the pristine particles at 6.155 and 1.2 g mL⁻¹, respectively. In combination with their volume-weighted d_H, their f_D is presented in Figure S3D. It is worth noting that ZnO ENMs deposited on cells at much faster rates than PCP due to two key properties. First, ZnO ENMs had a ρ_eff 5x higher than that of PCP. Second, once dispersed in EMEM, ZnO ENMs form ~50% larger agglomerates than PCP. Because of higher ρ_eff and agglomerate size, their in vitro deposition on cells occurs at a higher rate than PCP, as calculated based on our dosimetric calculations, described in detail by Deloid et al. (DeLoid et al. 2017). Specifically, about 70% of the administrated ZnO ENM mass deposited on cells within 2 hours whereas only ~12% of administered PCP mass was deposited on cells at the end of a 24-hour exposure period.

3.4. Toxicological assessment of PCP

To expose Calu-3 to realistic doses of PCP particles in vitro, data from the MPPD model were combined with their f_D, itself calculated by the distorted grid model. The cumulative deposited PCP mass per lung surface predicted by MPPD was 0.026 μg cm⁻² for a period 40 hours (equivalent to one business week). Calu-3 cells were exposed in vitro to 0.016 or 0.160 μg cm⁻² over 24 hours, following administration of PCP suspensions at 1.67 or 16.70 μg mL⁻¹, respectively. To match the deposited doses at 24 hours, ZnO ENMs suspensions at 0.24 or 2.40 μg mL⁻¹ were administered to Calu-3, following the dosimetric approach described above. Studies of the effects on suspended TK6 cells were based only on administered particle doses at 1, 10, or 100 μg ml⁻¹. Results of the toxicological and genotoxic assessment of PCP and ZnO ENMs against Calu-3 and TK6 cells are presented below. Of note, neither PCP nor ZnO ENMs exhibited autofluorescence or absorbance levels that interfered with the accuracy of the assays (data not shown).

Effects of PCP on cellular membrane integrity, metabolic activity, and ROS production of TK6 cells:

Figure 2A presents the LDH levels quantified in the extracellular medium of TK6 cells following a 24-hour exposure to either PCP or ZnO ENMs. Overall, results showed that PCP did not induce considerable cell death to Calu-3 whereas ZnO ENMs were significantly cytotoxic only at 100 μg ml⁻¹ and 24 hours of exposure. Figure 2B presents the results obtained by the PrestoBlue^® assay which indicates changes in mitochondrial enzymatic activity and overall cell viability. In summary, PCP did not suppress Calu-3 viability under any experimental conditions. On the contrary, ZnO ENMs were able to decrease cell viability only at the highest dose of 100 μg ml⁻¹ at 24 hours. Figure 2C shows the ROS levels as measured by the CellROX^® Orange assay following a 4-hour exposure of TK6 cells to PCP or ZnO ENMs. PCP elicited a substantial ~25% increase in ROS compared to untreated cells only at the highest administered dose of 100 μg ml⁻¹. In comparison, ZnO ENMs were bioactive even at the lower dose of 10 μg ml⁻¹ at which dose ROS levels increased by ~75%. Importantly, the filter extract did not induce any oxidative stress.

Figure 2. Acute cytotoxicity of PCP against TK6 cells.

(A) Extracellular release of LDH was quantified to assess the plasma membrane integrity of TK6 cells upon 24-hour exposures to PCP or the comparator ZnO ENMs. (B) The mitochondrial activity of TK6 cells was measured using the PrestoBlue assay after 24-hour exposures to PCP or the comparator ZnO ENMs. (C) Intracellular increase in ROS generation was measured using the CellROX Orange assay following 4-hour exposures to PCP or the comparator ZnO ENMs. Data represent an average of at least 3 independent replicates. Error bars represent mean ± SD. ***p<0.001 vs untreated control; “μg/ml” refers to the administered particle dose to TK6 cells which grow in suspension; “eq. to PCP at 100 μg/ml”: the volume ratio of filter extract-to-culture medium is the same as the PCP dispersion-to-culture medium for a 100 μg/ml PCP dispersion in cell culture medium.

Effects of PCP on cellular membrane integrity, metabolic activity, and ROS production of Calu-3 cells:

Figure 3A shows cytotoxicity results from Calu-3 cells exposed to PCP or ZnO ENMs for 24 hours. PCP were not cytotoxic at either 0.016 or 0.16 μg cm⁻². On the other hand, ZnO ENMs significantly compromised cellular membrane integrity, but only by 4% as compared to the untreated cells. As shown in Figure 3B, the metabolic activity of Calu-3 cells was suppressed following a 24-hour exposure to PCP at the highest dose of 0.160 μg cm⁻². Figure 3C shows that a 4-hour exposure of PCP did not induce any oxidative stress in Calu-3 cells at either deposited dose of 0.016 or 0.160 μg cm⁻², as measured by the CellROX^® Green assay. Of note, ZnO ENMs did not increase ROS generation at the same dosing scheme. For both TK6 and Calu-3 cells, menadione was used as a positive control and successfully increased ROS production by a factor of 1.75-2.

Figure 3. Acute cytotoxicity of PCP against Calu-3 cells following 24-hour exposure.

(A) Extracellular release of LDH was quantified to assess the plasma membrane integrity of Calu-3 upon dose-matched, 24-hour exposures to PCP or the comparator ZnO ENMs. (B) Mitochondrial activity of Calu-3 cells as measured using the PrestoBlue assay after 24-hour exposures to PCP or the comparator ZnO ENMs. (C) Intracellular generation of ROS as measured using the CellROX Green assay following 4-hour exposures to PCP or the comparator ZnO ENMs. Data represent an average of at least 3 independent replicates. Error bars represent mean ± SD. ***p<0.001 vs untreated control; “μg/cm²” refers to the particle mass deposited per cells surface area and was used for Calu-3 cells which adhere to the bottom of the well plate.

3.5. Effects of PCP on DNA damage and DNA repair capacity of Calu-3 and TK6 cells

The potential of PCP to induce DNA damage on TK6 and Calu-3 cells was assessed using the alkaline nano-CometChip Assay. Figure 4A shows that PCP were not able to induce considerable DNA damage at neither 10 nor 100 μg ml⁻¹ after 24 hours of exposure. ZnO ENMs could only be tested at the non-cytotoxic concentration of 10 μg ml⁻¹ at which they did not induce any considerable DNA damage on TK6 cells. Against Calu-3 cells, Figure 4B shows that PCP did not incur statistically significant levels of DNA damage, even at the highest delivered-to-cell dose of 0.160 μg cm⁻² over 24 hours. In contrast, ZnO ENMs at dose-matched conditions could severely compromise the integrity of the cells’ genome after 24 hours of exposure.

Figure 4. DNA damage results from the alkaline nano-CometChip assay.

(A) DNA damage quantitation in TK6 cells after exposure to PCP or comparator ZnO ENMs over 24 hours at administered doses of 10 or 100 μg ml⁻¹. (B) DNA damage quantitation in Calu-3 cells after exposure to PCP or comparator ZnO ENMs over 24 hours at deposited-to-cell doses of 0.160 μg cm⁻². Each dot represents ~250 cells; error bars: mean ± standard error (SE); ****p<0.0001 vs untreated control; “μg/ml” refers to the administered particle dose to TK6 cells which grow in suspension; “μg/cm²” refers to the particle mass deposited per cells surface area and was used for Calu-3 cells which adhere to the bottom of the well plate; “eq. to PCP at 100 μg/ml”: the volume ratio of filter extract-to-culture medium is the same as the PCP dispersion-to-culture medium for a 100 μg/ml PCP dispersion in cell culture medium; “eq. to PCP 0.160 μg/cm²”: the volume ratio of filter extract-to-culture medium is the same as the PCP dispersion-to-culture medium used for a 0.160 μg/cm² dose.

Following 24-hour exposures, fluorescence multiplex-host cell reactivation (FM-HCR) assays were used to quantify the DRC of Calu-3 and TK6 cells, according to procedures presented in our companion study (Toprani et al. 2021). Specifically, for Calu-3 cells, the doses were comparable to the PCP mass that would deposit after a week-long exposure to the aerosol settings described above. TK6 and Calu-3 cells were transiently transfected with six different reporter plasmids, each bearing different DNA lesions in order to assess the cellular repair capacity in multiple DNA repair pathways with and without exposure to PCP. A reporter with an enzymatically generated DSB was used to measure non-homologous end joining (NHEJ); a plasmid bearing a site-specific tetrahydrofuran abasic site analog was used to measure long patch base excision repair (BER); a reporter plasmid subjected to UV irradiation was used to measure the nucleotide excision repair (NER) pathway. DRC measured with these reporter plasmids is directly proportional to fluorescent protein expression, which is reported as a percentage of fluorescent protein expression from a corresponding damage free control plasmid.

Taking the advantage of transcriptional mutagenesis, we engineered three additional reporter plasmids where the DNA lesion induces transcriptional errors that lead to fluorescent protein expression unless repair occurs by the indicated pathway. Therefore, for these reporters, DRC is inversely proportional to fluorescent protein expression. A reporter plasmid with a site-specific hypoxanthine (Hx) adduct was used to measure the initiation of BER by methylpurine DNA glycosylase (MPG); a plasmid containing a site-specific 8oxoG lesion in the transcribed strand opposite cytosine (C) was used to measure the initiation of BER, which for this DNA lesion is catalyzed by several DNA glycosylases including OGG1, NEIL1, and NEIL2; a reporter plasmid with an undamaged adenine in the transcribed strand opposite 8oxoG was used to measure the repair capacity for MUTYH, which excises undamaged adenine when it is paired with 8oxoG in the opposing strand. Because the expression of the reporter plasmid is inversely proportional to DRC, we have reported the reciprocal of the fluorescent reporter expression after normalizing to the corresponding undamaged control plasmid. Table 1 presents a summary of the DNA lesions in the FM-HCR reporter plasmids. The original development and validation of these reporter plasmids and the method for calculating repair capacity from fluorescent reporter protein expression has been described in detail previously (Chaim et al. 2017; Nagel et al. 2014).

Table 1.

DNA repair proteins/pathways and the DNA lesions in the FM-HCR reporter plasmids

Reporter Plasmid	DNA repair protein/ pathway	DNA lesion
11	Non homologous end joining	Double strand break generated by enzymatic digest
2	Long patch base excision repair	Site specific Tetrahydrofuran abasic site analog
	Nucleotide excision repair	Ultraviolet (UV) irradiation (800J/m2)
	Initiation of base excision repair by methylpurine DNA glycosylase (MPG)	Site-specific hypoxanthine (Hx) lesion
	Initiation of base excision repair by several DNA glycosylases including 8-oxoguanine DNA glycosylase (OGG1) and Nei-like DNA glycosylases (NEIL1 and NEIL2)	Site-specific 8oxoG lesion in the transcribed strand opposite cytosine (C) base [8oxoG:C]
	Initiation of base excision repair by mutY DNA glycosylase (MUTYH)	Adenine (A) opposite a site-specific 8oxoG lesion [A:8oxoG]

Gating settings for untreated Calu-3 cells transfected with the FM-HCR plasmids are presented in Figure S4. Data presented in Figure S5 show that PCP dispersed in cell culture medium do not generate fluorescent signals that interfere with the reporter expression readout. ZnO ENMs were also assessed using the same method as a comparator nanomaterial with documented genotoxic potential (Toprani et al. 2021; Watson et al. 2014). Figure S6 shows that ZnO ENMs dispersed in culture medium in the presence of Calu-3 do not generate fluorescent signals and do not interfere with the FM-HCR assay parameters. Gating optimization for TK6 cells and interference exclusion from ZnO ENMs dispersed with cells in culture medium can be found at a companion study by Toprani et al. (Toprani et al. 2021).

Figure 5 shows that a 24-hour exposure of TK6 cells to PCP did not result in considerable alterations in the functionality of DNA repair pathways, which remained active at levels comparable to their respective controls. Similar results were obtained upon DRC quantification of TK6 cells following 24-hour exposure to ZnO ENMs (also presented in Figure 5). It has to be noted that due to pronounced toxicity at 100 μg ml⁻¹ over 24 hours as presented earlier (see Figure 2), ZnO ENMs were only administered to TK6 cells at 10 μg ml⁻¹. Results show that ZnO ENMs did not incur any considerable changes to DRC.

Figure 5. Effects of PCP or ZnO ENMs on the DRC of TK6 cells following 24 hours of exposure.

Quantification of alterations in DRC of TK6 cells following 24-hour exposure PCP or ZnO ENMs for 6 distinct DNA repair pathways related to namely (A) non-homologous end- joining repair (NHEJ) of double strand breaks; (B) long patch base excision repair (BER) of a tetrahydrofuran lesion; (C) Nucleotide excision repair (NER) of UV induced DNA damage; (D) Initiation of BER of a hypoxanthine lesion, ; (E) Initiation of BER of an 8oxoG:C lesion; and (F) Initiation of BER of an A:8oxoG lesion. Averages represent triplicate measurements; error bar represent mean ± SD. *p<0.05, **p<0.01 vs untreated control; “μg/ml” refers to the administered particle dose to TK6 cells which grow in suspension; “eq. to PCP at 100 μg/ml”: the volume ratio of filter extract-to-culture medium is the same as the PCP dispersion-to-culture medium for a 100 μg/ml PCP dispersion in cell culture medium.

Overall, PCP did not significantly alter the DRC of TK6 cells after 24 hours of exposure regardless of dose and any small changes in reporter expression were of similar magnitude to those caused by ZnO ENMs. The FM-HCR method was applied on Calu-3 cells to evaluate the potential genotoxicity of PCP on cells relevant to inhalation exposure. Figure 6 summarizes the DRC of Calu-3 cells following exposure to PCP. Firstly, it has to be noted that the filter extract did not modulate the screened DNA repair pathways other than a small decrease in repair of A:8oxoG which was not significantly different from the changes induced by PCP (Figure 5F). A general remark is that increased deposited doses of PCP on Calu-3 cells were accompanied by an increase in DRC for 5 out 6 DNA repair pathways (NHEJ, long-patch BER, NER, Hx (MPG), and A:8oxoG). In spite of this trend, only the NER pathway presented with a ~30% statistically significant increase in DRC, as shown in Figure 6C. In comparison, DRCs were increasingly suppressed with higher deposited doses of ZnO ENMs. Specifically, Figures 6A, 6C, 6E, and 6F show how NHEJ, NER, 8oxoG:C, and A:8oxoG reporters are respectively suppressed upon exposure to ZnO ENMs. Interestingly, ZnO ENMs had an inverse effect to nucleotide excision repair when compared to that brought upon by PCP. In particular, Figure 6C shows that NER activity was significantly suppressed by ~60% at a ZnO ENM dose of 0.160 μg cm⁻² over 24 hours whereas PCP significantly increased the activity of the same pathway by ~30%.

Figure 6. Effects of PCP or ZnO ENMs on the DRC of Calu-3 cells following 24 hours of exposure.

Quantification of alterations in DRC of Calu-3 cells following 24-hour exposure to administered doses of PCP or ZnO ENMs for 6 distinct DNA repair pathways related to namely (A) non-homologous end- joining repair (NHEJ) of double strand breaks; (B) long patch base excision repair (BER) of a tetrahydrofuran lesion; (C) Nucleotide excision repair (NER) of UV induced DNA damage; (D) Initiation of BER of a hypoxanthine lesion, ; (E) Initiation of BER of an 8oxoG:C lesion; and (F) Initiation of BER of an A:8oxoG lesion. Averages represent triplicate measurements; error bar represent mean ± SD. *p<0.05, **p<0.01 vs untreated control; “μg/ml” refers to the administered particle dose to TK6 cells which grow in suspension; “μg/cm²” refers to the particle mass deposited per cells surface area and was used for Calu-3 cells which adhere at the bottom of the well plate; “eq. to PCP 0.160 μg/cm²”: the volume ratio of filter extract-to-culture medium is the same as the PCP dispersion-to-culture medium used for a 0.160 μg/cm² dose.

In toto, 24-hour PCP exposure to Calu-3 cells revealed a dose-dependent upward trend in DRC which, however, was not statistically significant. On the contrary, oxidative damage repair by 8oxoG:C was significantly inhibited at the higher PCP dose. By comparison, exposure to ZnO ENMs at the same deposited particle mass per cell surface area had an inverse effect, with pathways associated with nucleotide excision and oxidative damage repair sustaining statistically significant inhibitions.

4. Discussion

This study examines the impact of PCPs on the genomic stability of the human bronchial epithelium following acute exposure to sub-cytotoxic doses similar to those an individual would experience if exposed to 40 hours in a printing center. The elevated HMPAH composition and other organic constituents of PCP suggests how they’re similar to PEPs (S. V. Pirela et al. 2014), but differentiate them from toner particles (Chalbot et al. 2017; Zhang et al. 2019). The known carcinogenic potential of HMPAH (Andersen et al. 2018; S. and J.M. 2013) and toxicity of metals like Mn and V, render the toxicological evaluation of PCP an urgent issue of occupational and end-consumer safety. PCP were administered to Calu-3 cells as one of the most commonly used tracheobronchial epithelial cell lines in studies on the biological effects engineered nanomaterials (Banga et al. 2012; He et al. 2020; Kroll et al. 2011). It is one of the few cell lines that form tight junctions in vitro and demonstrates the characteristics of differentiated, functional human airway epithelia (Foster et al. 2000; Kreft et al. 2015; Shen et al. 1994). Since p53 is mutated in these cells, p53-dependent stress responses may be altered, but this would not preclude detection of alterations in DNA repair capacity due to direct interactions between nanoparticles and the DNA repair machinery. Equally important, even though Calu-3 cells have a mutational pattern that involves DNA repair genes (Auslander, Wolf, and Koonin 2020), the pathways examined in this study were functional and it was thus possible to assess any effects nanoparticle exposure may have had on them. TK6 is a human lymphoblastoid cell line were used owing to their human origin and robust use with gene mutation tests (Lorge et al. 2016). It is widely used as a model cell line due to its high proliferation rate and stable karyotype (Auslander, Wolf, and Koonin 2020) which allows for comprehensive genotoxicity analyses, including in the field of nanotoxicology (Watson et al. 2014; Zhu et al. 2020).

The range of PCP doses were informed by their physicochemical properties and its aerosol characterization from an actual printer center in order to achieve “real world” relevancy (Setyawati et al. 2020). In more detail, these data were used with the MPPD model to calculate the total mass of PCP that would deposit per lung surface area in the human alveolar region (Price et al. 2002). Delivered particle doses of 0.016 and 0.160 μg cm⁻² bookend the dose of 0.026 μg cm⁻² calculated by MPPD as the dose that would deposit in the human deep lung following 40 hours of occupational exposure in the printing setting, assuming no particle translocation or clearance. It is also worth noting the importance of dosimetric analysis when comparing responses among particles with different particokinetics, like PCP and ZnO ENMs in this case. This was made possible with advances in the cellular dosimetry of ENMs (DeLoid et al. 2015, 2017) and using methods for measuring the particles’ effective density and colloidal properties in biological media (Cohen et al. 2018; Deloid et al. 2014), coupled with the computational modeling of their settling on cell monolayers in vitro (DeLoid et al. 2015, 2017).

Prior to genotoxic analyses, it was important to verify that PCP or ZnO ENMs at the employed doses and time-points would not be acutely cytotoxic. Should the particles compromise the membrane integrity of cells, bring about high levels of oxidative stress, or inhibit mitochondrial function, cells could become apoptotic (Redza-Dutordoir and Averill-Bates 2016). Such events could jeopardize the quantification of DNA repair by decreasing transfection efficiency, number of viable cells post transfection, or plasmid reporter expression because of increased endonuclease activity (Demir, Creus, and Marcos 2014). Even more so, it was important to define whether genomic alterations may occur at non-cytotoxic doses, which provide a more realistic inhalation exposure scenario.

At the deposited PCP mass per lung surface area doses, Calu-3 cells maintained their cellular membrane integrity. In fact, Calu-3 were not apoptotic even when exposed at the highest dose of ZnO ENMs, in accordance to previous findings that utilized the same particles against small airway epithelial cells (Wu et al. 2020). Despite the absence of acute cytotoxicity, the cells’ metabolic activity was considerably hampered after a 24-hour exposure to PCP which may signify p53 activation and onset of apoptotic mitochondrial dysfunction (Kreft et al. 2015). It was also interesting to see that neither PCP nor ZnO ENMs could considerably increase the production of intracellular ROS at doses comparable to those predicted by MPPD for 40 hours of workers exposure. It has to be noted that the current study examines the acute cytotoxicity of PCP in vitro. At the acute doses used here, it was expected -and confirmed- that PCP would not cause acute cell death, similarly to what has been shown in previous studies with similar particles (Chalbot et al. 2017; S. V. Pirela, Miousse, et al. 2016; Sisler et al. 2015). Still, this emergent contaminant is mostly related to chronic occupational exposures. Under chronic and sub-chronic conditions, it has been shown that PEPs can cause pulmonary alterations and inflammation in mice (Bai et al. 2010), increase the risk for adverse cardiovascular effects (Carll et al. 2020), and lead to the accumulation of epigenetic alterations (Lu et al. 2016). In fact, PEPs have been shown to compromise the function of the pulmonary microvascular endothelium, even in the absence of acute cytotoxicity (Sisler et al. 2015). More importantly, the findings reported here show that acute non-cytotoxic doses of PCP may alter the capacity of bronchial epithelial cells to repair their DNA, as discussed in more detail below.

The cytotoxicity assessment of TK6 cells showed that ZnO ENMs could significantly compromise their plasma membrane integrity after a 24-hour exposure at the highest dose (i.e., at 100 μg mL⁻¹). Such effects have been previously observed and can be partly attributed to Zn²⁺ ion shedding which has a pronounced cytotoxic potential in vitro (Chang et al. 2012). Both PCP and ZnO ENMs were non-cytotoxic at lower doses or time-points. In fact, the cells’ baseline metabolic activity tended to increase following exposure to either particle type, as measured by the PrestoBlue^® assay, which could be due to increased energy needs for clathrin- and caveolae-mediated particle endocytosis (Rejman et al. 2004). While increased mitochondrial enzyme activity may be due to pro-inflammatory increase in nitric oxide generation (Valdez, Zaobornyj, and Boveris 2006), it is also an indication that cells do not undergo apoptosis (Ren et al. 2015). At the same time, PCP elicited a dose-dependent increase in ROS production. In previous studies, PEPs from control print chamber studies, as well as several ENMs, were shown to induce oxidative stress at high enough doses (S. V. Pirela et al. 2017; S. V. Pirela, Miousse, et al. 2016; Zhang et al. 2019). In this study, the potential of PCP to induce oxidative stress in TK6 cells could be attributed to their high content in HMPAH, as previously shown by Chalbot et al. (Chalbot et al. 2017). By comparison, ZnO ENMs could instigate 3x ROS production, which is in congruence with numerous studies on the mechanism of in vitro toxicity of these engineered nanoparticles (Kreft et al. 2015).

The examination of the effect of PCP on DNA repair pathways is preceded by studies on the genotoxicity of PEPs performed by the authors as well as other groups. For example, epidemiological investigations revealed that photocopier workers exhibit higher levels of DNA damage than non-exposed subjects (Kasi et al. 2018) as well as chronic upper airway and systemic inflammation (Bello et al. 2021). Airborne PEPs have also been shown to be genotoxic against A549 cells at non-cytotoxic doses (S. V. Pirela, Miousse, et al. 2016; T. Tang et al. 2012) and may induce DNA damage, as previously shown by Pirela et al. (S. V. Pirela, Miousse, et al. 2016). In this study, it was found that PCP did not cause DNA strand breaks or oxidative damage at the employed doses and exposure time-points. Still, it has to be highlighted that genomic lesions and their repair is a multi-factorial and dynamic process. In other words, genomic lesions caused by other factors (e.g., from naturally occurring ROS or exogenous chemical factors) could be left partially unrepaired because of inhibited DRC (Carriere et al. 2017).

In this study, the DNA repair activity of Calu-3 and TK6 cells were assessed in six distinct pathways. To begin with, DNA glycosylase activity was assessed by quantifying the repair of 8oxoG:C, A:8oxoG, and hypoxanthine lesions. Then, long patch BER activity was also quantified using a reporter bearing a tetrahydrofuran lesion. Finally, NHEJ repair of double strand breaks and NER of UV-induced damage were also studied.

Depending on the damage type, cells can marshal the activity of several different pathways to repair genomic lesions (Nagel, Chaim, and Samson 2014). DNA damage is often the result of insults by environmental stressors. A variety of methods have been used to assess cell’s DNA repair capacity following exposure to environmental stressors, like PAHs and ionizing radiation (Cebulska-Wasilewska et al. 2007; Nagel et al. 2014; Toprani and Das 2020). Indirect assessments, like measuring changes in DNA damage and measuring aberrant gene expression involved in specific DNA repair pathways, have provided insights into cells’ DNA repair capacity in the past (Khatri, Bello, Pal, et al. 2013; Soren et al. 2019; Toprani and Das 2020). However, data obtained from such studies may not pinpoint specific DNA repair pathways and may not reflect functional repair activity. Furthermore, the relative participation of multiple DNA repair pathways cannot not be inferred from previous methods. To overcome these limitations, in this study, FM-HCR was used as a high throughput method to test the effects of PCP and ZnO ENMs on the simultaneous activity of six distinct DNA repair activities, using a platform developed in a recent study (Toprani et al. 2021). The underlying FM-HCR technology has been used to measure DRC in a variety of cell types, including both human and murine cell, primary tissues, as well as cancer cell lines (Chaim et al. 2017; McFaline-Figueroa et al. 2015). FM-HCR can be a new way to test DRC in a high-throughput manner in the field of nanotoxicology and more details about this recently developed methodology can be found a companion study by Toprani et al. (Toprani et al. 2021).

In our findings, higher PCP doses revealed a trend toward suppressing the DRC of TK6 cells, thus suggesting a dose-dependent effect. Results from DRC quantification of Calu-3 cells which are more relevant to inhalation exposure were even more alarming. PCP appeared to exert a variable effect on DRC of Calu-3 cells with some DNA repair pathways being inhibited while others being amplified compared to their untreated counterparts. Notably, nucleotide excision repair (NER), which is responsible for repairing bulky DNA adducts that could block DNA replication, was significantly increased, suggesting a possible stress-response. Such intensification in NER activity has been associated with increased resistance in chemotherapy of gastric and bladder cancer (Bellmunt et al. 2007; Kwon et al. 2007). Moreover, oxidative lesion repair of 8oxoG:C lesions (predominantly repaired by OGG1) was found to be inhibited by PCP. This effect is potentially worrisome given that inefficient repair by this pathway has been associated with lung cancer (Crosbie et al. 2012). In previous studies on samples collected from workers at print centers, urinary biomarkers of 8OHdG and 8OHG were increased (Khatri et al. 2017b; Khatri, Bello, Gaines, et al. 2013). Urinary 8-OHdG is thought to come from the repair of oxidative DNA damage (Graille et al. 2020) which is seemingly at odds with the observations of inhibited repair of 8oxoG:C lesions as presented in the current study. However, the origin of urinary 8-OHdG in vivo is not entirely worked out and oxidative stress in tissues other than those exposed to the PCPs could contribute to it. Another explanation could be the rapid resolution of transient oxidative damages following the clearance of PCPs, which could also cause increased values of urinary biomarkers of 8OHdG.

In contrast, ZnO ENMs at the same deposited particle mass per cell surface area over 24 hours had the inverse effect, suppressing NER activity, which has been linked to increased genomic instability to UV radiation (Busch et al. 1989). Consistent with genotoxic effects reported by others (Branica et al. 2016; Pati et al. 2016) and our previous work (Sotiriou et al. 2014; Toprani et al. 2021; Watson et al. 2014), DNA damage and alterations in DNA repair capacity were observed in Calu-3 cells following exposure to ZnO ENMs, included in this study as a comparator material for which genotoxicity is well established. Although the detailed molecular mechanisms underlying the genotoxicity of ZnO ENMs are incompletely understood, cations can potentially leach from ENMs and disrupt DNA repair pathways (S. Tang et al. 2013; Toprani et al. 2021). The differing biological effects between PCP and ZnO ENMs may be due to differences in cellular uptake, chemical and surface properties, the potential for ROS generation, or the fate of particles inside the cell (Faria et al. 2018). Increases in repair capacity as induced by PCP may reflect upregulation of DNA repair pathways as part of a stress response, whereas diminished repair capacity brought upon by ZnO ENMs could reflect direct inhibition of the DNA repair machinery. Potential mechanisms involved in particle-induced genotoxicity include epigenetic alterations, disruption of DNA repair machinery, direct induction of DNA damage, and reactive oxygen species that can be generated when mitochondrial function is disrupted (Carriere et al. 2017). Though beyond the scope of this study, these potential mechanisms merit deeper investigation.

In light of our findings, quantification of DRC appears to be an important insight to the genotoxic potential of PCP which are ubiquitous in occupational and even consumer settings. Our data on Calu-3 cells suggest that pneumocytes may be vulnerable to DNA repair capacity alterations following inhalation exposure to PCP, but in vivo extrapolations from in vitro data are confounded by concurrent environmental exposures and a multitude of physiological factors that are not replicated in vitro. Future studies should focus on employing additional in vitro cellular models, like primary small airway epithelial cells, and expand the application of FM-HCR on lung tissues in order to better understand the PCP genotoxicity at the organ level. Sub-cytotoxic chronic exposures are also necessary to simulate the constant deposition and long-term impact of PCP on the human lung. Furthermore, it is necessary to expand the use of FM-HCR on PCP collected from other print centers given their variability in PAH species and transition metals. PAHs and transition metals can be genotoxic, but have also been shown to induce synergistic genotoxic damage (Chalbot et al. 2017; S. and J.M. 2013; Valavanidis, Vlahoyianni, and Fiotakis 2005). Integrating DRC and DNA damage assessment should better inform us on the potential long-term pathogenicity that has been observed in the clinic with printer center workers and improve the safety measures that are put in place by regulators and key stakeholders for occupational and public health.

5. Conclusions

In summary, PCP significantly altered DNA repair pathways of Calu-3 cells at doses comparable to those anticipated in real-world occupational exposure scenarios. These doses do not elicit cytotoxicity, do not impair cell viability or mitochondrial metabolic activity, and do not increase the intracellular production of ROS. In spite of the absence of obvious cytotoxic effects, PCP exposure did inhibit the ability of Calu-3 cells to repair 8oxoG:C lesions. Inversely, PCP increased the activity of the NER pathway. Upon dose-matched comparison, PCP had comparable effects on oxidative lesion repair with ZnO ENMs which have long been identified as a potentially genotoxic particulate stressor in vitro and in vivo. To the best of our knowledge, this study is the first to show that DNA repair pathways can be differentially affected by PCP. These findings raise concerns about long-term exposure of workers and consumers to non-cytotoxic levels of these emerging particulate contaminants that could undermine the vital role of DNA repair in the human respiratory tract. Overall, this work flags the under-studied impact of PCP on DNA repair within the field of nanosafety. Finally, it is crucial to note that further in vitro and in vivo studies are necessary to understand the exact mechanisms by which PCP alter DNA repair pathways. Such studies should ideally quantify particle internalization, identify sub-cellular particle localization, and trace any particle biotransformations, like dissolution and ion leaching, that could potentiate the observed genotoxic effects.

Highlights.

• Airborne nanoscale (PM_0.1) printer center nanoparticles were sampled.

• Advanced genotoxic and toxicological assessment of printer center nanoparticles was performed against human lung epithelial and lymphoblastoid cells.

• Printer center nanoparticles at sub-cytotoxic doses significantly altered the capacity of exposed cells to perform nucleotide-excision repair and restore oxidative DNA lesions.

• Findings raise concerns for potential health effects for workers in such micro-environments.

Abbreviations

BET

Brunauer–Emmett–Teller

DI

deionized

DRC

DNA repair capacity

DSEcr

critical delivered sonication energy

EMEM

Eagle's minimum essential medium

ENM

engineered nanomaterial

f_D

fraction of administered mass delivered to the cell surface

FM-HCR

fluorescence multiplex-host cell reactivation

HMPAH

high molecular weight poly-aromatic hydrocarbons

ICP-MS

inductively-coupled plasma mass spectrometry

LDH

lactate dehydrogenase

lp-BER

long patch base excision repair

MPG

methyl purine DNA glycosylase

MPPD

multiple-path particle deposition

MUTYH

mutY DNA glycosylase

NER

nucleotide excision repair

NHEJ

non-homologous end joining

NHIR

Nanotechnology Health Implications Research

NIEHS

National Institute of Environmental Health and Safety

OGG1

8-oxoguanine DNA glycosylase

PAH

poly-aromatic hydrocarbons

PBS

phosphate-buffered saline

PCP

printer center nanoparticle

PDI

polydispersity index

PEP

printer-emitted particle

PTFE

polytetrafluoroethylene

ROS

reactive oxygen species

RPMI

Roswell Park Memorial Institute

SAEC

small airway epithelial cells

TEM

transmission electron microscopy