Skip to main content
NCBI home page
Springer logoLink to Springer
. 2025 Aug 21;99(12):4941–4953. doi: 10.1007/s00204-025-04152-7

Transcriptomic profile of TiO2 particle overload in rat alveolar macrophages: examining responses between P25 and two pigment grade particles in vitro

Laeticia Perez 1,✉, Jérôme Ambroise 2, Bertrand Bearzatto 2, Yousof Yakoub 1, Mihaly Palmai-Pallag 3, Caroline Bouzin 4, Laurence Ryelandt 5, Cristina Pavan 6, François Huaux 1, Dominique Lison 1
PMCID: PMC12534340  PMID: 40841704

Abstract

The industrial uses of titanium dioxide (TiO2) are extensive, with pigment grades (particle size > 100 nm) being the most common forms produced. Nanoforms (particle size < 100 nm) of TiO2 (e.g., P25) are also produced for specialist applications such as photocatalysts. P25 induced inflammatory and carcinogenic lung responses in rats at high exposure doses inducing lung overload. We previously identified an in vitro transcriptomic signature (18 genes) associated with overload of P25, in rat alveolar macrophages. The objective of the present study was to determine whether this signature also applies to other, more commonly used pigment grades of TiO2. Using high-throughput sequencing, we examined the transcriptomic responses of three TiO2 grades (the photocatalyst P25 and two pigment grades, G3-1 and Bayertitan T) in primary rat alveolar macrophages exposed in vitro at non-overload (4 µg/mL for P25; 2 µg/mL for G3-1 and Bayertitan T) or overload (40 µg/mL for P25; 20 µg/mL for G3-1 and Bayertitan T) doses. At an equivalent internalized volume of particles, the response to P25 overload was significantly higher than those to pigment particles. However, a consistent modulation of the 18-gene signature was observed across all three TiO2 grades, albeit with a markedly lower magnitude of response for the pigment particles. Although the gene signature of particle overload in rat alveolar macrophages seems to apply across different TiO2 particle grades, our results suggest that unrealistic and extreme exposure concentrations to the pigment TiO2 particles would be required to induce in vivo toxicity similar to that observed with P25 particles.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00204-025-04152-7.

Keywords: Gene signature, Rat pulmonary toxicity, Particle comparison, Particle reactivity, Impaired lung clearance

Background

Titanium dioxide (TiO2) is an odorless powder discovered in 1791. At the beginning of the twentieth century, the production and use of TiO2 emerged as a safer alternative to existing white pigments (e.g., white lead), and TiO2 became one of the most widely used pigments worldwide (TDMA; Racovita 2022). In addition to its use as an ultra-white colorant, TiO2 has the ability to scatter light and to absorb UV without degrading. Thanks to these properties, TiO2 has been used in various industrial and consumer products including paints, plastics, papers, cosmetics, pharmaceuticals, and food (Racovita 2022).

Two primary crystalline forms of TiO2 particles are commercialized; rutile and anatase. Rutile, accounting for approximately 90% of global production, is the most widely used TiO2 form, primarily as a pigment in paints and plastics, due to its hardness and opacity. Anatase, while less dense and opaque than rutile, offers a softer, bluer tone with reduced UV absorption. Anatase is primarily employed in applications such as food additives, medicines, and specialty papers (Macwan et al. 2011).

To effectively scatter visible light, pigment applications typically require TiO2 particles with a median volume-based size of 150–300 nm (or 100–250 nm by number). TiO2 can be produced through various methods, with the sulfate and chloride processes being the most common for pigment grade particles (Middlemas, Fang, and Fan 2013). For specialized applications, smaller TiO2 particles can be produced using methods such as flame spray pyrolysis, sol–gel techniques, and hydrothermal treatments (Uddin et al. 2020). Importantly, the production method significantly influences the physicochemical properties of TiO2, including particle size, morphology, crystallinity, and surface structure. For example, flame hydrolysis of TiCl4 results in a near-perfect crystalline habit substantially different from samples derived from wet-chemistry pathways (Uddin et al. 2020).

Flame pyrolysis is used in the production of P25 TiO2, an 80% anatase/ 20% rutile mix nanoform widely employed in toxicological studies but less used commercially. The unique production process of P25 results in specific physicochemical characteristics that may not be representative of all commercially available TiO2 forms. Chronic inhalation of high levels of TiO2, including P25, can lead to lung overload, which may in turn lead to inflammatory and carcinogenic lung responses in the rat (Heinrich et al. 1995; Bermudez et al. 2004; Lee et al. 1985). These adverse outcomes in rats have prompted considerable debate in the scientific and regulatory communities regarding the relevance of rat lung overload to human health (Warheit et al. 2016; Driscoll 2022). The International Agency for Research on Cancer (IARC) has conservatively classified all TiO2 particles as “possibly carcinogenic to humans” (category 2B) (IARC 2010). The European Commission has also classified all powdered forms of TiO2 containing at least 1% of particles with aerodynamic diameter ≤ 10 µm as a suspected human carcinogen when inhaled (category 2). The latter decision was, however, annulled by the General Court of Justice of the European Union at the end of 2022 (“CWS Powder Coatings GmbH and Others v European Commission” 2022) in part because the adverse outcomes observed in experimental rat studies were not related to the intrinsic properties of TiO2 particles but to the more general phenomenon of lung overload (Relier et al. 2017).

Lung overload is mainly related to particles defined as poorly soluble low toxicity (PSLT) such as TiO2 and carbon black (ECETOC 2013). Due to their poor solubility, at low doses, inhaled PSLT are almost exclusively cleared from the lower airways by phagocytosis and translocation by alveolar macrophages (AM). However, at high levels of exposure, AM can become overloaded, leading to a decrease in their phagocytic and mobile activities. This results in the rapid accumulation of PSLT particles in the lungs, triggering localized inflammation and oxidative stress that persists due to failed clearance. In 1988, Morrow hypothesized that lung overload is initiated when the pool of AM has accumulated an average volume of PSLT particles that exceeds 6% of their total cellular volume (Morrow 1988). In experimental rats, lung overload induced by PSLT may induce chronic alveolitis, with significant recruitment of neutrophils, oxidative stress, fibrosis, type II epithelial cell hyperplasia, metaplasia, and, in extreme cases, lung cancer (Bos et al. 2019; Driscoll and Borm 2020). To date, epidemiological studies have not demonstrated an increased risk of lung cancers in occupational cohorts exposed to TiO2 particles. Moreover, rat AM seem to be more sensitive than human AM to particle overload (Bos et al. 2019; Perez et al. 2025).

We have previously identified, in rat AM in vitro, an 18-gene signature of overload induced by P25 and Printex 90 (a type of carbon black), two forms of PSLT known to cause lung overload and severe adverse outcomes in vivo in rats (Perez et al. 2025). This signature can be functionally related to the in vivo outcomes observed in experimental studies and may, thus, represent a common response to all PSLT particles at overload doses or may reflect more nuanced changes related to specific physicochemical properties of some PSLT particles (Wang and Fan 2014; Shin et al. 2015). Various TiO2 grades, including pigment TiO2, which are more representative of occupational exposure might, thus, induce different responses under overload conditions. Here, we examined the genomic expression profile in rat AM exposed at overload doses (> 6% AM vol. on average) to three grades of TiO2 particles: the photocatalyst P25, a nanoform mix of 80% anatase and 20% rutile, and G3-1 and Bayertitan T, both 100% rutile pigment grades.

Methods

Isolation and culture of primary rat AM

The procedure for collection of rat AM was approved by the Ethical Committee for Animal Experimentation at the Health Science Sector, UCLouvain, Brussels, Belgium (No LA1230312). Eight-week-old male F344/HanZtmRj rats were purchased from Janvier Labs (St Bertevin, France) and were housed in individually ventilated cages in a controlled environment (21.3 ± 0.4 °C, 53.4 ± 8.9% relative humidity, 16 h light/8 h dark cycle, with water containing 4 ppm chlorine (hydropac-pouches from Plexx, Gelderland, Netherlands) and food ad libitum) at the IREC animal facility (UCLouvain, Brussels, Belgium). AM were isolated from bronchoalveolar lavages collected from healthy untreated rats using the methods described in Perez et al. (2025). For each experiment, cells collected from a group of rats were pooled in complete medium consisting of Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% antibiotic–antimycotic (AA), and 100,000 cells/cm2 culture well surface area were pre-cultured in a 96- or a 48-well culture plate in a CO2 incubator (37 °C, 5% CO2) for 5 days prior to TiO2 exposure. Each experiment included specific plates for determining non-overload and overload doses, cytotoxicity dose–response, and collecting RNA (see below).

Selection and characterization of TiO2 particles

P25 (nanoform, 80% anatase/ 20% rutile), G3-1, and Bayertitan T (pigment 100% rutile) particles were provided by the Titanium Dioxide Manufacturers Association (TDMA, Brussels, Belgium). Morphology and specific surface area (SSA) of the particles were assessed on the powder product by scanning electron microscopy (ultra 55 FegSEM, Carl Zeiss, Oberkochen, Germany) and N2 adsorption (BET method), respectively. Prior to each experiment, TiO2 particles were weighed, heated for 2 h at 200 °C to inactivate any microbial contaminants, and suspended in complete cell culture medium to achieve a stock concentration of 5 mg/mL. As particle size is an essential aspect of inhalation and in vitro toxicology, we developed a dispersion protocol for each test particle to achieve size distribution in the relevant respirable size range (e.g., median diameter < 5 µm). Particle suspensions were dispersed with a VCX-750 probe sonicator (Sonics & Materials, Connecticut, USA) at 750W, 40% amplitude and 0 pulse during 20 s. Suspensions of P25 and G3-1 particles were next vortexed for a few seconds to avoid large agglomerates which was not required for Bayertitan T. Particle size and effective density (density in medium) of the dispersed particles were measured by HELOS laser diffraction (Sympatec GmbH, Etten-Leur, Netherlands) and by the volumetric centrifugation method developed by DeLoid et al. (2017), respectively.

Cytotoxic activity of TiO2 particles

Rat AM were cultured for 5 days in 96-well culture plates (Greiner Bio-One, Kremsmünster, Austria) and then exposed for 4 days to 150 µL of increasing concentrations (from 0 to 100 µg/mL) of P25, G3-1, or Bayertitan T particles. The dose–response relationship for cell viability was assessed using a water-soluble tetrazolium salt (WST)-1 assay (5%, Roche, Basel, Switzerland) following the manufacturer’s guidelines (N = 1 experiment; n = 3 technical replicates). The data were analyzed by a two-way ANOVA followed by a post hoc Dunnett test (treatment vs control) using GraphPad Prism 9.4.1. Differences with p value < 0.05 were considered statistically significant.

Selection of the non-overload and overload doses

The volume of particles internalized by AM was assessed in black 96-well culture plates with clear bottom (Greiner Bio-One, Kremsmünster, Austria) using contrast microscopy (ZEISS LSM 800, Jena, Germany) (N = 2 experiments; n = 3 technical replicates). Following culture and particle exposure, cells were fixed in 100µL formaldehyde 3.6% for a minimum of 20 min. After rinsing twice with PBS, pictures of cells containing particles were taken by contrast microscopy (ZEISS LSM 800, Jena, Germany) using a 63X water immersion objective. Pictures were then processed via the ImageJ software 1.53t. A custom macro was created to highlight the intracellular particles in black using a thresholding method and to calculate the % of the cell surface area occupied by particles ('Introduction into Macro Programming ‘). To estimate particle uptake, this 2D surface area was used as a proxy for volume, acknowledging the limitations of this approach, since adherent macrophages in vitro are not perfectly flat and have a 3D morphology. Nonetheless, this method provides a comparative estimate of intracellular particle load. For each tested dose, more than 130 cells were analyzed and the mean and median % volume of intracellular particles were calculated.

RNA extraction

Primary rat AM cultured in 48-well culture plates were exposed for 4 days to 500 µL of control, non-overload or overload doses of P25, G3-1 or Bayertitan T. RNA-seq analyses were performed on three technical replicates of pooled cells for each tested condition, in two separate experiments. For the first experiment, total RNA was isolated using a phenol–chloroform extraction method with TriPure Isolation Reagent according to manufacturer’s instructions (Roche, Basel, Switzerland). For the second experiment, a silica membrane-based on-column extraction was performed using the RNeasy Mini Kits (Cat. 74,106, Qiagen, Hilden, Germany) following the manufacturer’s instructions. Residual DNA was removed by on-column DNase digestion using the RNAse-Free DNase Set (Cat. 79,254, Qiagen, Hilden, Germany).

High-throughput RNA sequencing

RNA extracts corresponding to untreated controls, non-overload, and overload doses were quantified and prepared following the methods described in Perez et al. (2025). The high-throughput RNA-sequencing (HTS) analyses were performed on three technical replicates of pooled cells for each tested condition, in two separate experiments (N = 2), on a Novaseq 6000 Illumina platform. A minimum of 25 M paired-end reads (2 × 100 bp) were generated per sample.

Bioinformatics

All sequencing data were analyzed using the Automated Reproducible MOdular workflow for preprocessing and differential analysis of HTS data (ARMOR v1.5.4) pipeline (Orjuela et al. 2019). In this pipeline, reads underwent a quality check using FastQC (Babraham Bioinformatics). Quantification and quality control results were summarized in a MultiQC report before being mapped using Salmon (Cunningham et al. 2019) to the transcriptome index which was built using all Ensembl cDNA sequences obtained in the Rattus_norvegicus.mRatBN7.2.cdna.all.fa. Then estimated transcript abundances from Salmon were imported into R version 4.1.1 using the tximeta package (Soneson et al. 2015) and analyzed for differentially expressed genes (DEG) with edgeR (Robinson et al. 2010). Genes that showed differences in expression between the different exposure conditions (control, non-overload, and overload) were identified for each of the three particle types separately. The absolute expression difference between two conditions was calculated for each gene as a log2-fold change (logFC). The statistical significance of this difference was then calculated as a false discovery rate (FDR). A multiple test correction of the p value was performed to control the FDR at a level of 0.05. When gene modulations were compared between the three test materials, Pearson correlation coefficient R and intraclass correlation coefficient ICC between the corresponding logFC were computed to assess the correlation and the concordance, respectively (Liu et al. 2016). The pathways over-represented by the DEG were analyzed by over-representation analysis (ORA) using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. Raw and processed HTS data were deposited and made publicly available on the Gene Expression Omnibus (First rat experiment: GSE271469, Second rat experiment: GSE272350).

Results

Physicochemical properties of TiO2 particles

All three test samples exhibited high purity (≥ 99.5%). P25 consisted of aggregated nanoparticles. G3-1 and Bayertitan T had a higher effective density and lower specific surface area (SSA) compared to P25 (Table 1). Despite the nanoscale nature of P25 particles, the median sizes of the agglomerated particles in culture medium were similar to G3-1 and Bayertitan T, which were comparable to aerosol particle sizes used in rat studies (~ 1–2 µm) recording chronic inflammation and lung cancer following TiO2 exposure(Lee et al. 1985; Heinrich et al. 1995). Particle size distributions following dispersion are shown in Supplementary data Fig. S1.

Table 1.

Main physicochemical characteristics of the three tested TiO2 particles

Test particle name1 P25 G3-1 Bayertitan T
Type1 Photocatalyst Pigment Pigment
Crystallinity1 Anatase/rutile (80/20) Rutile Rutile
Morphology graphic file with name 204_2025_4152_Figa_HTML.gif graphic file with name 204_2025_4152_Figb_HTML.gif graphic file with name 204_2025_4152_Figc_HTML.gif
% TiO2 purity 99.51 99.71 99.52
Skeletal density (g/cm3) 4.11 4.11 4.32
Effective density (g/cm3) 1.4 2.8 2.8
Specific surface area (m2/g) 51.4 6.7 6.7
Number median diameter of primary particles (nm)1 22 147 150
Mass median diameter of agglomerated particles (µm) 1.52 1.04 1.35
Open in a new tab

1 Communicated by the supplier (TDMA), 2 FromJackson et al. (2012)

Selection of non-overload and overload doses

The transcriptomic responses were assessed after 4 days of exposure to doses inducing control, non-overload, or overload conditions of particles (Fig. 1a). The data for P25 particles presented in this study are identical to those previously reported in Perez et al. (2025). Based on Morrow’s hypothesis (Morrow 1988) considering that lung overload is initiated when the average percent volume of cellular particles is above 6%, non-cytotoxic exposure doses of 4 and 40 µg/ mL (corresponding to 1.8 and 18 µg/cm2 surface well area) for P25 and 2 and 20 µg/mL (corresponding to 0.9 and 9 µg/cm2 surface well area) for G3-1 and Bayertitan T were selected as the non-overload and overload doses, respectively (Fig. 1b–d and Table S1). Comparing both rat experiments, the fractions of particles internalized by rat AM were quite similar, except at high exposure doses of Bayertitan T particles (Fig. 1c and Table S1).

Fig. 1.

Fig. 1

Open in a new tab

Rat AM exposure to P25, G3-1, or Bayertitan T particles. a Experimental design of particle exposure. b Rat AM were exposed to increasing doses of P25, G3-1, or Bayertitan T particles for 4 days. The number of live cells was determined with the WST-1 assay. Results are expressed as percentage of the control (0 μg/mL). Values are mean ± SD from three technical replicates in a single experiment (N = 1; n = 3). For each particle, the data were compared by a two-way ANOVA followed by a post hoc Dunnett’s test (treatment vs control). c Violin plot and box plot representing the percentage volume distribution of particles accumulated in rat AM, 4 days after exposure to non-overload or overload doses of P25 (4 and 40 µg/mL), G3-1 (2 and 20 µg/mL), or Bayertitan T (2 and 20 µg/mL). The boxplots represent the 25th and 75th percentiles with the median in the middle. The dotted line represents a volume of 6%. d Pictures by contrast microscopy of rat AM exposed to non-overload or overload doses of P25 (4 and 40 µg/mL), G3-1 (2 and 20 µg/mL), or Bayertitan T (2 and 20 µg/mL)

Differentially expressed genes (DEG) upon overload of P25, G3-1, or Bayertitan T particles

In all tested conditions, about 15,000 genes were expressed in the rat AM cultures. Hundreds to thousands of genes were significantly (FDR < 0.05) differently expressed upon P25 particle overload (vs control or non-overload conditions) (Fig. 2a), with 128 genes being specifically modulated under overload conditions (vs control and non-overload conditions) in both experiments (Fig. 2b). Although Bayertitan T particles caused significant gene expression changes under overload conditions in the second experiment when comparing to control, none of these genes were, however, significantly modulated when comparing overload to non-overload conditions, in either the first or second experiments.

Fig. 2.

Fig. 2

Open in a new tab

Differentially expressed genes in rat AM exposed to P25, G3-1, and Bayertitan T. a Comparison of the total number of significant differentially expressed genes (DEG), upregulated or downregulated, for each condition comparison in rat (first (HTS 1) or second (HTS 2) experiment) AM. CTL = control, NO = non-overload, O = overload. b Venn diagram of unique and shared DEG under overload conditions (comparing to control and non-overload conditions) in rat AM from the first and second experiment (HTS 1&2) exposed to P25, G3-1, or Bayertitan T particles

Similarly, no genes were significantly differentially expressed under overload conditions (compared to control and non-overload) of G3-1 particles in both experiments (Fig. 2b). Absence of statistically significant modulation does, however, not mean that genes had not been modulated at all by G3-1 or Bayertitan T particle overload. Therefore, the comparison of responses between P25, G3-1, and Bayertitan T next focused on amplitudes of gene variation (logFC).

DEG specifically modulated by P25 particle overload

In both experiments, overload of P25 particles in rat AM induced a significant modulation of 128 genes, 38 upregulated and 90 downregulated. Over-representation analysis revealed that phagosome, complement and coagulation cascade, as well as pathways related to T lymphocyte activation and differentiation, were significantly associated with the downregulated genes (Fig. 3c).

Fig. 3.

Fig. 3

Open in a new tab

Differentially expressed genes in rat AM under overload of P25 particles. a Over-representation analysis of the genes significantly downregulated or upregulated by P25 particle overload. The dotted line indicates an enrichment ratio of 1. b, c Correlation tables representing the Pearson coefficient R (b) or the intraclass correlation coefficient (c) reflecting the correlation or concordance of the logFC (overload vs non-overload conditions) between P25, G3-1, and Bayertitan T, in the first (HTS 1) and second (HTS 2) experiment, of the 128 DEG by P25 particle overload in rat AM. Stars indicate significant correlation: *p < 0.05; **p < 0.01; ***p < 0.001. d Heatmap representing the logFC (overload (O) vs non-overload (NO) conditions) of the 128 DEG by P25 particle overload in rat AM, comparing P25, G3-1, and Bayertitan T particles in the first (HTS 1) and second (HTS 2) experiment

The amplitude and direction of gene expression changes (logFC) between overload and non-overload conditions were used to analyze and compare gene modulations across experiments and particle types. The logFC, under overload conditions, of the 128 genes were significantly positively correlated between the three particles tested, in the two experiments, except comparing the first and second experiment with Bayertitan T (Fig. 3b). However, the amplitude of the genes modulated in the same direction was markedly lower (logFC on average 1.5 lower) under G3-1 and Bayertitan T particle overload than under P25 particle overload (Fig. 3c, d).

In vitro signature of P25 and Printex90 particle overload: comparison between TiO2 particles

We have previously identified 18 genes, included in the 128 DEG mentioned above, as potential markers of particle overload in vitro, as these were specifically modulated by overload (vs control and non-overload) of both P25 and Printex90 particles in rat AM. In addition, the known functions of many of these 18 genes can be related to the mechanisms of the in vivo rat adverse outcomes (Fig. 4a) (Perez et al. 2025). As for the 128 DEG, the logFC of these 18 genes under overload conditions (vs non-overload) were strongly positively correlated between the three particles tested and in the two experiments, except comparing the first and second experiment with Bayertitan T (Fig. 4b). The amplitude of the genes modulated in the same direction was, however, markedly lower (logFC on average 1.6–1.7 lower, except for G3-1 s experiment which was on average 1.47 lower) (Fig. 4c–e).

Fig. 4.

Fig. 4

Open in a new tab

Comparison of the modulation of a potential in vitro 18-gene signature of particle overload in rat AM. a From Perez et al. (2025). Relationship between the proposed rat adverse outcome pathway (Bos et al. 2019) and genes modulated by P25 and Printex90 particle overload in rat AM. b, c Correlation tables representing the Pearson coefficient R (b) or the intraclass correlation coefficient (c) reflecting the correlation or concordance of the logFC (overload vs non-overload conditions) between P25, G3-1, and Bayertitan T, in the first (HTS 1) and second (HTS 2) experiment, of the 18-gene signature of P25 and Printex90 particle overload in rat AM. Stars indicate significant correlation: *p < 0.05; **p < 0.01; ***p < 0.001. d, e Heatmap (d) or bar plot (e) representing logFC (overload (O) vs non-overload (NO) conditions) of the 18-gene signature of P25 and Printex90 particle overload in rat AM, comparing P25, G3-1, and Bayertitan T particles in the first (HTS 1) and second (HTS 2) experiment

Discussion

Nanoform P25 TiO2 particles have been shown to induce rat lung overload in vivo resulting in pro-inflammatory, genotoxic and/or carcinogenic responses in rat models (Heinrich et al. 1995; Relier et al. 2017; Chezeau et al. 2018; Loret et al. 2018; Koltermann-Jully et al. 2020). While the nanoform P25 is almost exclusively used as photocatalytic agent, the majority of TiO2 used in other industries are pigment grades, mostly rutile with larger primary particle sizes. Clarifying the toxic potential of pigment grades relative to nanoform photocatalytic grades of TiO2 particles is, therefore, highly relevant for human health.

The present study compared the genome expression profile of rat AM exposed to overload of the photocatalyst P25, the pigment G3-1, or the pigment Bayertitan T, using untargeted transcriptomic analyses. At similar % volume of internalized particles, P25 particles were largely more active than the two other TiO2 grades, as overload conditions (> 6% AM vol.) induced high numbers of DEG for P25, with a far more muted responses to the two pigment grades of TiO2 and none that were specific to overload (i.e., vs control and non-overload).

This difference in both the amplitude as well as lack of genes associated specifically with overload conditions suggests a markedly different response to P25 as compared to the pigment grades tested. There was considerable variability between the number of DEG induced by P25 and Bayertitan T particles between the two experiments. The median value of the standard deviation (SD) of technical replicates within a condition among the two rat experiments and the three particles was similar, and technical variability can, therefore, not explain these differences, suggesting a possible role of extraction methods used and/or batch effects (Scholes and Lewis 2020; Leek et al. 2010). In addition, differences related to the % volume of particles internalized by rat AM were observed when comparing the two experiments, particularly at overload doses of Bayertitan T particles, where a 5% difference was observed. Despite these differences in DEG and % volume, the amplitude and direction of gene expression changes of the 128 genes significantly modulated by P25 overload correlated extremely well between both rat experiments.

The function/activity of some of these 128 genes was related to a decrease in phagocytic and efferocytic (phagocytosis of apoptotic cells) activity of the AM, indicating that AM were likely functionally overloaded, as well as potential impairment of T lymphocyte activation and differentiation. The functions of these genes are highly similar to those of the 18 genes specifically modulated by P25 and carbon black (Printex 90) particle overload in previous work (Perez et al. 2025), suggesting that these functions are critical in the responses of overloaded rat AM. In addition, some genes significantly modulated by P25 overload suggest a metabolic reprogramming of the AM towards a pro-inflammatory M1 phenotype, similar to changes induced by lipopolysaccharide (LPS) stimulation (Fig. S2). Given that heat treatment was used to inactivate potential LPS contaminants, the observed changes in gene expression point to a macrophage activation specifically induced by P25 overload. LPS-activated macrophages switch from oxidative phosphorylations to glycolysis (Kelly and O’Neill 2015). This switch is partly mediated by the increased expression of the hypoxia inducible factor (HIF)-1α. The transcription of this factor was not directly modulated by P25 overload, but Tnxip, which codes for a protein driving the degradation and export of the HIF-1α protein (Zhu et al. 2018), was downregulated by P25 overload. In addition, Slc2a1 coding for the glucose transporter GLUT1 and Pfkfb3 are target genes of H1F-1α and were both significantly upregulated by P25 overload, thus favoring a glycolytic switch. We also observed that the expression of Coq8a, coding for a kinase involved in the biosynthesis of coenzyme Q, was significantly reduced by P25 overload, limiting oxidative phosphorylations (Crane 2001). LPS-activated macrophages also increasingly produce nitric oxide from arginine (Kelly and O'Neill 2015). Following P25 particle overload, the expression of the gene coding for the transporter of arginine (Slc7a2) was upregulated in rat AM. LPS also reduces the metabolism of fatty acid in macrophages (Kelly and O'Neill 2015) that was, again, observed here under P25 overload (Alox5, Alox15, Hsd17b10, and Abcd2). It, thus, appears that the responses associated with P25 overload are similar to those induced by LPS exposure, which in turn leads to inflammatory-related macrophage activation consistent with a M1 phenotype.

Most inflammation-related genes were downregulated under P25 particle overload (Pycard, Nlrc4, Cx3cr1, Csf1r, Ccr5) with the exception of Cxcr4, Gdf15 and the chemokine Cxcl2, also known as macrophage inflammatory protein 2 (MIP2)-alpha. The increase of Cxcl2 in BAL or lungs of rats exposed to P25 or pigment TiO2 particles has already been observed in several studies (Sager et al. 2008; Driscoll et al. 1993; Chezeau et al. 2018). These three genes are also upregulated following LPS macrophage stimulation (Tian et al. 2019; Luan et al. 2019; Wang et al. 2021). Overall, the modulations of some genes specifically induced upon P25 overload were, thus, similar to gene modulations induced by inflammatory stimuli, suggesting a shift in AM toward a pro-inflammatory, M1 phenotype.

While G3-1 or Bayertitan T overload did not induce statistically significant modulations of the genes linked to P25 overload (128 genes), including those linked to P25 and Printex 90 overload (18 genes), a good correlation was found between the logFC under overload (when compared to non-overload) of P25 and, G3-1 or Bayertitan T. This similitude of response of the three particles to overload indicates a unique signature of PSLT overload in rat AM in vitro. However, the amplitude of these modulations was markedly lower for G3-1 and Bayertitan T, possibly explaining why these genes did not appear modulated in the same way in the first and second experiments upon Bayertitan T and, to a lesser extent, upon G3-1 overload.

The physicochemical properties of particles and, in particular, their size, surface area, morphology, coating, aggregation state or crystallinity are key parameters influencing their reactivity and toxicity (Wang and Fan 2014; Shin et al. 2015). In the present study, there was no coating for any of the three TiO2 grades, and the sizes of the agglomerated particles were very similar.

Inhaled particles, including TiO2, with smaller constituent particle size, and thus higher surface area, are generally more reactive and more prone to induce acute lung inflammatory responses than their larger counterparts (Warheit et al. 2007; Hussain et al. 2009; Sager et al. 2008; Rushton et al. 2010; Lison et al. 1997). In the present study, the criteria of particle overload were based on Morrow’s hypothesis of volume dose, directly related to mass dose (Morrow 1988). This hypothesis has been supported by several studies (Oberdorster et al. 1992; Cullen et al. 2000) but has never really been demonstrated. In contrast, Tran et al. (Tran et al. 2000) hypothesized that the overload paradigm might be more related to surface area dose than volume dose. Here, the specific surface area of P25 particles was eight times greater than that of G3-1 and Bayertitan T. By converting the mass doses (μg/mL) used in this study into surface area doses (cm2/ml), the overload doses of G3-1 and Bayertitan T particles would approximately correspond to the non-overload dose of P25 particles (Table 2). It would, therefore, appear interesting to analyze the responses to G3-1 and Bayertitan T particles at higher doses, to strictly compare the particles at equivalent surface area doses. The mass doses of G3-1 and Bayertitan T needed to reach the surface dose inducing overload of P25 are, however, extremely high (318 µg/mL) and likely to induce high cytotoxicity and/or analytical interferences. Furthermore, while macrophages have considerable capacity for particle uptake, it may be volumetrically impossible to achieve the required SA dose as the volume of particles needed may exceed the volumetric capacity of the cell.

Table 2.

Conversion of mass administered doses leading to similar internalized % volume of particles in rat alveolar macrophages exposed to P25, G3-1, or Bayertitan T particles into surface doses

Non-overload Overload
P25 G3-1 Bayertitan T P25 G3-1 Bayertitan T
Mass dose (µg/ml) 4 2 2 40 20 20
Surface dose (cm2/ml) 2.06 0.13 0.13 20.6 1.3 1.3
Open in a new tab

Also, various in vitro and in vivo studies have observed that anatase or the mix anatase/rutile particles were more bioactive than their rutile counterpart (Johnston et al. 2009). In vivo, Warheit et al. (Warheit et al. 2007) observed that a mix of 80% anatase / 20% rutile, following intratracheal instillation of 1 or 5 mg/kg, induced adverse lung outcomes (inflammation, cytotoxicity, and parenchymal cell proliferation) in rats that were close to the effects observed at the same doses with quartz silica particles. These effects were not observed after the instillation of the same mass doses of pure rutile. The authors concluded that the differences in responses were not solely attributed to differences in specific surface area but were also related to the crystallinity, or the inherent pH of the particles. In vitro, anatase induced more reactive oxygen species and cytotoxicity than rutile in human fibroblast and epithelial cells (Sayes et al. 2006; Gerloff et al. 2012). Differences in reactivity may be related to the affinity of the particles to interact with different cell components (Yu et al. 2017). Rutile particles have a greater affinity for phospholipids, while anatase particles have a high affinity for proteins. Therefore, the crystallinity of P25 particles, mainly composed of anatase, could also explain the higher responses observed to these particles in the present study. Other factors may also contribute to the higher bioactivity of P25 particles, particularly their unique production process. Specifically, P25 is produced through flame hydrolysis of TiCl4. This process results in sintered particles, distinct from the milled TiO2 particles obtained through wet-chemistry methods (Uddin et al. 2020).

In our previous publication, we identified an 18-gene signature of P25 or Printex 90 particle overload in rat AM (Perez et al. 2025). The amplitudes of modulation of these genes positively correlated between rat and human AM exposed to P25 or Printex 90 particle overload, although the magnitude of response in human AM was markedly lower and a unique additional 16 gene expression signature was observed in human AM, but not in rat AM. In this study, we observed that the 18-gene signature also correlated between P25, G3-1 and Bayertitan T in rat AM. These results further support the identification of a signature of PSLT overload in rat AM in vitro. The species and particle characteristics may affect the sensitivity of these responses, as a markedly lower amplitude of gene modulation was observed in human AM exposed to P25 or Printex 90 compared to rat AM (Perez et al. 2025) and in rat AM exposed to G3-1 or Bayertitan T compared to P25. We do not know yet whether this signature is also induced in vivo in rats and/or humans. Our results suggest, however, that an extremely high and probably unrealistic exposure dose (expressed as mass dose) of pigment TiO2 would be required to induce AM overload and, in turn, adverse lung outcomes in rats and possibly humans. To validate this, additional in vivo studies may be carried out.

Of course, this study has its strengths and weaknesses. The primary challenge was the variability in the number of DEG observed between the two experiments, a point that has been discussed earlier. Despite these differences, the 18 genes, specifically modulated by overload induced by P25 and Printex 90, showed robust and consistent regulation across both experiments. Furthermore, the modulation of the 18 genes discussed here was previously validated by RT-qPCR analyses on AM isolated from five rat biological replicates (Perez et al. 2025). This study analyzed the in vitro responses in a single cell type, AM which is at the center of the early phenomena associated with lung overload. An in vitro system obviously differs from the dynamic and interactive in vivo responses which also involve epithelial cells, fibroblasts, inflammatory cells, and many more. In addition, AM were isolated from Fischer 344 rats, which are known to exhibit higher inflammatory responses compared to other rat strains (e.g., Sprague–Dawley), both in vivo and in vitro in AM (Zhang et al. 2011). We used a single duration of particle exposure (4 days) to avoid an acute response associated with bolus dose exposure, which would not have been relevant to in vivo overload. However, the temporal pattern of AM responses to overload may differ over time. This was not captured by the present study as only a single time point (4 days of exposure) was examined.

Conclusion

The present study revealed that, at similar % volume of internalized particles, P25 was more reactive than G3-1 and Bayertitan T in terms of gene transcription. The specific responses to P25 overload, which can be related to outcomes observed in vivo in rats, were also observed upon G3-1 and Bayertitan T overload, indicating an in vitro PSLT overload signature. However, the amplitude of the responses to G3-1 and Bayertitan T particles was markedly lower, which may be related to the differences in specific surface area and/or crystallinity between the particles. Overall, these findings suggest that extreme exposure to rutile pigment particles would be required to induce in vivo toxicity similar to that observed with P25 particles.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary file1 (PDF 767 KB) (766.5KB, pdf)

Acknowledgements

The authors thank the BRIGHTcore team at VUB/UZ Brussels for their help with the RNA-seq analyses. They also thank the Titanium Dioxide Manufacturers Association (TDMA) and the International Carbon Black Association (ICBA) teams for their advice and expertise. Finally, they also would like to acknowledge the contribution of Rita Vanbever team from the Louvain Drug Research Institute (LDRI) at UCLouvain for their expert guidance in laser diffraction analyses.

Author contributions

L.P. performed the experiments, data analyses, co-supervised the experimental design, and wrote the manuscript. J.A. performed most of the bioinformatic analyses. B.B. helped with RNA-sequencing analyses. Y.Y. and M.P. helped with the animal experiments. C.B. helped with confocal microscopy analyses. L.R. performed scanning electron microscopy analyses. C.P. performed N2 adsorption analyses. F.H. and D.L. conceived, designed, and supervised the work and wrote the manuscript. All authors read and approved the final manuscript.

Funding

This study was supported by the Titanium Dioxide Manufacturers Association (TDMA).

Data availability

Raw and processed HTS data were deposited and made publicly available on the Gene Expression Omnibus (First rat experiment: GSE271469, Second rat experiment: GSE272350).

Declarations

Conflict of interests

The authors declare that they have no conflict of interest.

Ethical approval

The procedure for collection of rat AM was approved by the Ethical Committee for Animal Experimentation at the Health Science Sector, UCLouvain, Brussels, Belgium (No LA1230312).

Footnotes

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

  1. Bermudez E, Mangum JB, Wong BA, Asgharian B, Hext PM, Warheit DB, Everitt JI (2004) Pulmonary responses of mice, rats, and hamsters to subchronic inhalation of ultrafine titanium dioxide particles. Toxicol Sci 77:347–357 [DOI] [PubMed] [Google Scholar]
  2. Bos PMJ, Gosens I, Geraets L, Delmaar C, Cassee FR (2019) Pulmonary toxicity in rats following inhalation exposure to poorly soluble particles: the issue of impaired clearance and the relevance for human health hazard and risk assessment. Regul Toxicol Pharmacol 109:104498 [DOI] [PubMed] [Google Scholar]
  3. Chezeau L, Sebillaud S, Safar R, Seidel C, Dembele D, Lorcin M, Langlais C, Grossmann S, Nunge H, Michaux S, Dubois-Pot-Schneider H, Rihn B, Joubert O, Binet S, Cosnier F, Gate L (2018) Short- and long-term gene expression profiles induced by inhaled TiO(2) nanostructured aerosol in rat lung. Toxicol Appl Pharmacol 356:54–64 [DOI] [PubMed] [Google Scholar]
  4. Crane FL (2001) Biochemical functions of coenzyme Q10. J Am Coll Nutr 20:591–598 [DOI] [PubMed] [Google Scholar]
  5. Cullen RT, Tran CL, Buchanan D, Davis JM, Searl A, Jones AD, Donaldson K (2000) Inhalation of poorly soluble particles. I. Differences in inflammatory response and clearance during exposure. Inhal Toxicol 12:1089–1111 [DOI] [PubMed] [Google Scholar]
  6. Cunningham F, Achuthan P, Akanni W, Allen J, Amode MR, Armean IM, Bennett R, Bhai J, Billis K, Boddu S, Cummins C, Davidson C, Dodiya KJ, Gall A, Giron CG, Gil L, Grego T, Haggerty L, Haskell E, Hourlier T, Izuogu OG, Janacek SH, Juettemann T, Kay M, Laird MR, Lavidas I, Liu Z, Loveland JE, Marugan JC, Maurel T, McMahon AC, Moore B, Morales J, Mudge JM, Nuhn M, Ogeh D, Parker A, Parton A, Patricio M, Abdul Salam AI, Schmitt BM, Schuilenburg H, Sheppard D, Sparrow H, Stapleton E, Szuba M, Taylor K, Threadgold G, Thormann A, Vullo A, Walts B, Winterbottom A, Zadissa A, Chakiachvili M, Frankish A, Hunt SE, Kostadima M, Langridge N, Martin FJ, Muffato M, Perry E, Ruffier M, Staines DM, Trevanion SJ, Aken BL, Yates AD, Zerbino DR, Flicek P (2019) Ensembl 2019. Nucleic Acids Res 47:D745–D751 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. "CWS powder coatings GmbH and others v European commission." In. 2022. ECLI:EU:T:2022:725. General court (Ninth Chamber, Extended Composition)
  8. DeLoid GM, Cohen JM, Pyrgiotakis G, Demokritou P (2017) Preparation, characterization, and in vitro dosimetry of dispersed, engineered nanomaterials. Nat Protoc 12:355–371 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Driscoll KE (2022) Review of lung particle overload, rat lung cancer, and the conclusions of the Edinburgh Expert Panel-it’s time to revisit cancer hazard classifications for titanium dioxide and carbon black. Front Public Health 10:907318 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Driscoll KE, Borm PJA (2020) Expert workshop on the hazards and risks of poorly soluble low toxicity particles. Inhal Toxicol 32:53–62 [DOI] [PubMed] [Google Scholar]
  11. Driscoll KE, Hassenbein DG, Carter J, Poynter J, Asquith TN, Grant RA, Whitten J, Purdon MP, Takigiku R (1993) Macrophage inflammatory proteins 1 and 2: expression by rat alveolar macrophages, fibroblasts, and epithelial cells and in rat lung after mineral dust exposure. Am J Respir Cell Mol Biol 8:311–318 [DOI] [PubMed] [Google Scholar]
  12. ECETOC 2013 'Poorly soluble particles/lung overload', Technical Report No. 122
  13. Gerloff K, Fenoglio I, Carella E, Kolling J, Albrecht C, Boots AW, Forster I, Schins RP (2012) Distinctive toxicity of TiO2 rutile/anatase mixed phase nanoparticles on Caco-2 cells. Chem Res Toxicol 25:646–655 [DOI] [PubMed] [Google Scholar]
  14. Heinrich U, Fuhst R, Rittinghausen S, Creutzenberg O, Bellmann B, Koch W, Levsen K (1995) “Chronic inhalation exposure of wistar rats and two different strains of mice to diesel engine exhaust carbon black, and titanium dioxide.” Inhalation Toxicol 7:533–556 [Google Scholar]
  15. Hussain S, Boland S, Baeza-Squiban A, Hamel R, Thomassen LC, Martens JA, Billon-Galland MA, Fleury-Feith J, Moisan F, Pairon JC, Marano F (2009) Oxidative stress and proinflammatory effects of carbon black and titanium dioxide nanoparticles: role of particle surface area and internalized amount. Toxicology 260:142–149 [DOI] [PubMed] [Google Scholar]
  16. IARC. 2010. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans—Carbon Black, Titanium Dioxide, and Talc (Lyon, France) [PMC free article] [PubMed]
  17. 'Introduction into Macro Programming '. Accessed 12/12. https://imagej.net/scripting/macro#example-macros. Accessed 12 Dec
  18. Jackson P, Hougaard KS, Vogel U, Wu D, Casavant L, Williams A, Wade M, Yauk CL, Wallin H, Halappanavar S (2012) Exposure of pregnant mice to carbon black by intratracheal instillation: toxicogenomic effects in dams and offspring. Mutat Res 745:73–83 [DOI] [PubMed] [Google Scholar]
  19. Johnston HJ, Hutchison GR, Christensen FM, Peters S, Hankin S, Stone V (2009) Identification of the mechanisms that drive the toxicity of TiO(2)particulates: the contribution of physicochemical characteristics. Part Fibre Toxicol 6:33 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kelly B, O’Neill LA (2015) Metabolic reprogramming in macrophages and dendritic cells in innate immunity. Cell Res 25:771–784 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Koltermann-Jully J, Ma-Hock L, Groters S, Landsiedel R (2020) Appearance of alveolar macrophage subpopulations in correlation with histopathological effects in short-term inhalation studies with biopersistent (nano)materials. Toxicol Pathol 48:446–464 [DOI] [PubMed] [Google Scholar]
  22. Lee KP, Trochimowicz HJ, Reinhardt CF (1985) Pulmonary response of rats exposed to titanium dioxide (TiO2) by inhalation for two years. Toxicol Appl Pharmacol 79:179–192 [DOI] [PubMed] [Google Scholar]
  23. Leek JT, Scharpf RB, Bravo HC, Simcha D, Langmead B, Johnson WE, Geman D, Baggerly K, Irizarry RA (2010) Tackling the widespread and critical impact of batch effects in high-throughput data. Nat Rev Genet 11:733–739 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Lison D, Lardot C, Huaux F, Zanetti G, Fubini B (1997) Influence of particle surface area on the toxicity of insoluble manganese dioxide dusts. Arch Toxicol 71:725–729 [DOI] [PubMed] [Google Scholar]
  25. Liu J, Tang W, Chen G, Lu Y, Feng C, Tu XM (2016) Correlation and agreement: overview and clarification of competing concepts and measures. Shanghai Arch Psychiatry 28:115–120 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Loret T, Rogerieux F, Trouiller B, Braun A, Egles C, Lacroix G (2018) Predicting the in vivo pulmonary toxicity induced by acute exposure to poorly soluble nanomaterials by using advanced in vitro methods. Part Fibre Toxicol 15:25 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Luan HH, Wang A, Hilliard BK, Carvalho F, Rosen CE, Ahasic AM, Herzog EL, Kang I, Pisani MA, Yu S, Zhang C, Ring AM, Young LH, Medzhitov R (2019) GDF15 is an inflammation-induced central mediator of tissue tolerance. Cell 178(1231–44):e11 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Macwan DP, Dave PN, Chaturvedi S (2011) A review on nano-TiO2 sol–gel type syntheses and its applications. J Mater Sci 46:3669–3686 [Google Scholar]
  29. Middlemas S, Zak Fang Z, Fan P (2013) A new method for production of titanium dioxide pigment. Hydrometallurgy 131–132:107–113 [Google Scholar]
  30. Morrow PE (1988) Possible mechanisms to explain dust overloading of the lungs. Fundam Appl Toxicol 10:369–384 [DOI] [PubMed] [Google Scholar]
  31. Oberdorster G, Ferin J, Morrow PE (1992) Volumetric loading of alveolar macrophages (AM): a possible basis for diminished AM-mediated particle clearance. Exp Lung Res 18:87–104 [DOI] [PubMed] [Google Scholar]
  32. Orjuela S, Huang R, Hembach KM, Robinson MD, Soneson C (2019) ARMOR: an automated reproducible modular workflow for preprocessing and differential analysis of RNA-seq data. G3 Genes|genomes|genetics 9:2089–2096 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Perez L, Ambroise J, Bearzatto B, Froidure A, Pilette C, Yakoub Y, Palmai-Pallag M, Bouzin C, Ryelandt L, Pavan C, Huaux F, Lison D (2025) Unique transcriptomic responses of rat and human alveolar macrophages in an in vitro model of overload with TiO(2) and carbon black. Part Fibre Toxicol 22:8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Racovita AD (2022) Titanium dioxide: structure, impact, and toxicity. Int J Environ Res Public Health. 10.3390/ijerph19095681 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Relier C, Dubreuil M, Lozano Garcia O, Cordelli E, Mejia J, Eleuteri P, Robidel F, Loret T, Pacchierotti F, Lucas S, Lacroix G, Trouiller B (2017) Study of TiO2 P25 nanoparticles genotoxicity on lung, blood, and liver cells in lung overload and non-overload conditions after repeated respiratory exposure in rats. Toxicol Sci 156:527–537 [DOI] [PubMed] [Google Scholar]
  36. Robinson MD, McCarthy DJ, Smyth GK (2010) edgeR: a bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26:139–140 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Rushton EK, Jiang J, Leonard SS, Eberly S, Castranova V, Biswas P, Elder A, Han X, Gelein R, Finkelstein J, Oberdorster G (2010) Concept of assessing nanoparticle hazards considering nanoparticle dosemetric and chemical/biological response metrics. J Toxicol Environ Health A 73:445–461 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Sager TM, Kommineni C, Castranova V (2008) Pulmonary response to intratracheal instillation of ultrafine versus fine titanium dioxide: role of particle surface area. Part Fibre Toxicol 5:17 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Sayes CM, Wahi R, Kurian PA, Liu Y, West JL, Ausman KD, Warheit DB, Colvin VL (2006) Correlating nanoscale titania structure with toxicity: a cytotoxicity and inflammatory response study with human dermal fibroblasts and human lung epithelial cells. Toxicol Sci 92:174–185 [DOI] [PubMed] [Google Scholar]
  40. Scholes AN, Lewis JA (2020) Comparison of RNA isolation methods on RNA-Seq: implications for differential expression and meta-analyses. BMC Genomics 21:249 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Shin SW, Song IH, Um SH (2015) Role of physicochemical properties in nanoparticle toxicity. Nanomaterials (Basel) 5:1351–1365 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Soneson C, Love MI, Robinson MD (2015) Differential analyses for RNA-seq: transcript-level estimates improve gene-level inferences. F1000Res 4:1521 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. TDMA, Titanium dioxide manufacturers association-. 'What is titanium dioxide?'. https://tdma.info/what-is-titanium-dioxide/. Accessed 4 May 2022
  44. Tian X, Xie G, Xiao H, Ding F, Bao W, Zhang M (2019) CXCR4 knockdown prevents inflammatory cytokine expression in macrophages by suppressing activation of MAPK and NF-kappaB signaling pathways. Cell Biosci 9:55 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Tran CL, Buchanan D, Cullen RT, Searl A, Jones AD, Donaldson K (2000) Inhalation of poorly soluble particles. II. Influence Of particle surface area on inflammation and clearance. Inhal Toxicol 12:1113–1126 [DOI] [PubMed] [Google Scholar]
  46. Uddin MJ, Cesano F, Chowdhury AR, Trad T, Cravanzola S, Martra G, Mino L, Zecchina A, Scarano D (2020) Surface structure and phase composition of TiO2 P25 particles after thermal treatments and HF etching. Front Mater. 10.3389/fmats.2020.00192 [Google Scholar]
  47. Wang J, Fan Y (2014) Lung injury induced by TiO2 nanoparticles depends on their structural features: size, shape, crystal phases, and surface coating. Int J Mol Sci 15:22258–22278 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Wang G, Huang W, Wang S, Wang J, Cui W, Zhang W, Lou A, Geng S, Li X (2021) Macrophagic extracellular vesicle CXCL2 recruits and activates the neutrophil CXCR2/PKC/NOX4 axis in sepsis. J Immunol 207:2118–2128 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Warheit DB, Webb TR, Reed KL, Frerichs S, Sayes CM (2007) Pulmonary toxicity study in rats with three forms of ultrafine-TiO2 particles: differential responses related to surface properties. Toxicology 230:90–104 [DOI] [PubMed] [Google Scholar]
  50. Warheit DB, Kreiling R, Levy LS (2016) Relevance of the rat lung tumor response to particle overload for human risk assessment-update and interpretation of new data since ILSI 2000. Toxicology 374:42–59 [DOI] [PubMed] [Google Scholar]
  51. Yu Q, Wang H, Peng Q, Li Y, Liu Z, Li M (2017) Different toxicity of anatase and rutile TiO(2) nanoparticles on macrophages: involvement of difference in affinity to proteins and phospholipids. J Hazard Mater 335:125–134 [DOI] [PubMed] [Google Scholar]
  52. Zhang Y, Lin X, Koga K, Takahashi K, Linge HM, Mello A, Laragione T, Gulko PS, Miller EJ (2011) Strain differences in alveolar neutrophil infiltration and macrophage phenotypes in an acute lung inflammation model. Mol Med 17:780–789 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Zhu G, Zhou L, Liu H, Shan Y, Zhang X (2018) Microrna-224 promotes pancreatic cancer cell proliferation and migration by targeting the TXNIP-mediated HIF1alpha pathway. Cell Physiol Biochem 48:1735–1746 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary file1 (PDF 767 KB) (766.5KB, pdf)

Data Availability Statement

Raw and processed HTS data were deposited and made publicly available on the Gene Expression Omnibus (First rat experiment: GSE271469, Second rat experiment: GSE272350).

Articles from Archives of Toxicology are provided here courtesy of Springer

RESOURCES