Abstract Proteomic investigations yield high‐dimensional datasets, yet their application to large‐scale toxicological assessments is hindered by reproducibility challenges due to fluctuating measurement conditions. To address these limitations, this study introduces an advanced tandem mass tag (TMT) labeling protocol. Although labeling approaches shorten data acquisition time by multiplexing samples compared to traditional label‐free quantification (LFQ) methods in general, the associated costs may surge significantly with large sample sets, for example, in toxicological screenings. However, the introduced advanced protocol offers an efficient, cost‐effective alternative, reducing TMT reagent usage (by a factor of ten) and requiring minimal biological material (1 µg), while demonstrating increased reproducibility compared to LFQ. To demonstrate its effectiveness, the advanced protocol is employed to assess the toxicity of nine benchmark nanomaterials (NMs) on A549 lung epithelial cells. While LFQ measurements identify 3300 proteins, they proved inadequate to reveal NM toxicity. Conversely, despite detecting 2600 proteins, the TMT protocol demonstrates superior sensitivity by uncovering alterations induced by NM treatment. In contrast to previous studies, the introduced advanced protocol allows simultaneous and straightforward assessment of multiple test substances, enabling prioritization, ranking, and grouping for hazard evaluation. Additionally, it fosters the development of New Approach Methodologies (NAMs), contributing to innovative methodologies in toxicological research. Keywords: adverse outcome pathway (AOP), mode of action (MoA), nanomaterials (NMs), new approach methodologies (NAMs), proteomics, Tandem mass tag (TMT) __________________________________________________________________ This work presents a cost‐effective, advanced TMT‐labeling protocol for large‐scale toxicological screenings, emphasizing its benefits in addressing the common limitations in proteomics, for example, time, cost, and biological material requirements. As proof‐of‐concept, the protocol is assessed by studying the cellular effects of nine well‐characterized benchmark nanomaterials on lung human epithelial A549 cells, with a comparative analysis to a traditional proteomic method. graphic file with name SMTD-8-2400420-g004.jpg 1. Introduction The rapid development and utilization of nanomaterials (NMs) across various fields has pushed forward significant advancements, transforming sectors ranging from electronics, to food and healthcare.^[ [38]^1 , [39]^2 ^] However, the increasing utilization of NMs in consumer products and industrial applications demands a greater awareness of their potential risks to human health and the environment. Considering the vast diversity of NMs, each with different shapes, sizes, coatings, and compositions, it is challenging to generalize toxicological properties from one NM to another, necessitating extensive testing for each form. Traditionally, hazard assessments rely on in vivo experiments, yet time and economical limitations, ethical considerations and the evolving regulatory landscape have underscored the urgency of seeking alternative methods.^[ [40]^3 , [41]^4 ^] As a result, there is an increasing need for new approach methodologies (NAMs) for predicting substance toxicity without animal experimentation.^[ [42]^4 , [43]^5 ^] These NAMs encompass a range of approaches, including in silico methods, in vitro techniques, and in chemico strategies, which serve as either alternatives or supplements to traditional animal testing.^[ [44]^4 ^] Most commonly used in vitro tests aim at investigating immediate (acute) effects. However, for comprehensive coverage of regulatory relevant endpoints, especially for chronic effects, it is crucial to understand the mechanism or mode of action (MoA) through which the test substance exerts its toxicity. A fundamental concept used in toxicology to illustrate mechanistic knowledge of the effect of a substance from molecular to organism levels is the Adverse Outcome Pathway (AOP). This concept serves as a solid framework that aids in making informed regulatory decisions.^[ [45]^6 , [46]^7 , [47]^8 ^] AOPs outline a sequence of key events (KEs) across molecular, cellular, organ, and organism levels from a molecular initiating event to an observed adverse outcome.^[ [48]^6 , [49]^8 ^] Each KE represents a measurable biological change.^[ [50]^8 ^] Omic techniques significantly contribute to the construction of AOPs frameworks,^[ [51]^9 , [52]^10 ^] due to their capacity to comprehensively explore toxicity mechanisms,^[ [53]^11 ^] enabling the generation of extensive datasets spanning various biological levels.^[ [54]^12 ^] The OECD recognizes the potential of these techniques as part of NAMs for risk assessment,^[ [55]^13 ^] as they enable a more refined, efficient and potentially animal‐free risk assessments. However, limitations arise due to the lack of standardized methods for regulatory requirements.^[ [56]^14 , [57]^15 ^] To this end, the OECD has developed the Transcriptomics and Metabolomics Reporting Framework,^[ [58]^16 , [59]^17 ^] and is now working on the advancement of the framework to cover proteomics, which is a valuable addition.^[ [60]^18 , [61]^19 ^] Among the omic methods for toxicological applications, transcriptomics is more advanced due to the availability of well‐established platforms, which facilitates the generation of standardized data. Additionally, transcriptomics can be utilized for extensive toxicological screenings since it is essentially a high‐throughput method, facilitating rapid and large‐scale assessments of potential toxicants.^[ [62]^20 , [63]^21 ^] However, the changes identified through transcriptomics, while valuable, often do not closely reflect the actual phenotype affected by a substance.^[ [64]^22 ^] Metabolomics studies provided as well important insights into MN toxicity,^[ [65]^23 , [66]^24 , [67]^25 ^] although for large‐scale toxicological studies involving a high volume of samples, metabolomics may present challenges in terms of scalability, cost, and data analysis complexity.^[ [68]^26 ^] Similarly as metabolomics, proteomics, has the potential to provide a more direct and accurate depiction of the toxicological effects.^[ [69]^27 , [70]^28 ^] The challenge of proteomics is the considerable variability in experimental approaches, instrumentation, and data analysis software.^[ [71]^29 , [72]^30 ^] This variability hampers the standardization of experimental results, posing a challenge to achieve comparable outcomes that are essential for widespread application and interpretation in the field. Consequently, the development of effective and standardized protocols for proteomics in toxicological applications is crucial. The complex nature of toxicology studies, involving different cell lines, various types of test substances, and multiple concentrations and time points, poses a significant challenge to the routine application of quantitative proteomics. This complexity makes it particularly difficult to consistently detect changes in protein abundance levels. Label‐Free Quantitation (LFQ) is a widely used technique known for its cost‐effectiveness; however, it can be time‐consuming due to the extended measurement times required for data acquisition^[ [73]^31 ^] during which reproducibility and accuracy limitations become increasingly apparent.^[ [74]^32 , [75]^33 ^] In contrast, labeled proteomics approaches, such as Tandem Mass Tag (TMT) and isobaric tags for relative and absolute quantification (iTRAQ), offer the advantage of multiplexing, allowing for the simultaneous quantification of multiple samples.^[ [76]^34 , [77]^35 ^] However, the convenience of labeling methodologies come at the expense of increased costs, making the choice between these quantitative proteomics strategies a crucial decision in proteomics research, balancing the trade‐off between time and effort while striving for accurate and reliable results. Hence, this study aims to develop and implement a cost‐effective advanced TMT labeling protocol for large‐scale toxicity assessments. Particularly addressing scenarios where limited biological material is available, for example, primary cells, and where simultaneous assessment is required, such as hazard priority ranking and grouping.^[ [78]^36 ^] Given the myriad of NM variants and the resulting challenge they pose to risk assessment, as a proof‐of‐concept, the advanced TMT labeling protocol was utilized to assess the effect of eight benchmark, commercially available NMs on human lung epithelial A549 cells, addressing inhalation as a primary exposure route. NMs selected from the JRC repository included four substances in two variants each – titanium dioxide (TiO[2]), cerium dioxide (CeO[2]), zinc oxide (ZnO), and silicon dioxide (SiO[2]) – along with copper oxide (CuO) as a positive control. Finally, results from the application of the advanced TMT protocol were compared to those obtained by traditional LFQ to assess which approach is more adequate for toxicological screening. 2. Results As the paradigm for hazard assessment is undergoing a transition, with NAMs aiming to replace traditional in vivo experiments, the need for a deeper understanding of toxicity mechanisms becomes crucial. In this evolving landscape, TMT labeling is emerging as a potent technique, recognized for its precision in quantifying peptides across multiple samples while preserving sample specificity. To facilitate this, we introduce an advanced, cost‐efficient, approach that adapts the TMT labelling process without compromising labelling efficiency. A series of experiments were initiated using HeLa digests, refining the vendor TMT labeling protocol and reducing the amount of TMT labels employed, all while maintaining a high labeling efficiency (≤99%), enabling reliable quantitative results. 2.1. Advanced TMT Protocol Development According to the manufacturer's TMT6‐plex protocol, 800 µg of TMT label is required to label 100 µg of protein, rendering the technique rather expensive. Therefore, to minimize both costs and reagent consumption, in a first step the amount of HeLa digest was reduced to 10 µg. Protein/TMT label ratios of 1:4, 1:6, and 1:8 were evaluated, corresponding to TMT concentrations of 10.6, 12.4, and 15.8 mmol L^−1 during the labeling process. Each HeLa digest was prepared in triplicates and analyzed (100 ng on column) per LC‐MS run. RAW files were processed using Proteome Discoverer 2.4.1.15 (Thermo Scientific) with TMT modifications (129.163 Da on Lys and peptide N‐terminus) set to dynamic. Label efficiency was determined by calculating the ratio of unlabeled peptide‐spectrum matches (PSMs) to the total PSM count. Approximately 10 000 peptides corresponding to ∼2500 proteins were identified (Figure [79]1 ). Figure 1. Figure 1 [80]Open in a new tab Peptide/protein identifications (FDR 0.01, database: Homo sapiens, UP000005640) of 10 µg HeLa digests (n = 3) labeled with 80, 60, or 40 µg TMT6‐plex label and measured via LC‐MS (DDA, three technical replicates, 100 ng on column). Labeling efficiency was calculated based on the number of labeled PSMs to the total PSM count. Values displayed as average ± SD. The initial label efficiency was measured at 96.5% on average for a 1:4 protein/TMT ratio but increased to 99.0% and 99.6% when applying ratios of 1:6 and 1:8, respectively. Due to the superior labeling efficiency, a TMT concentration of 15.8 mmol L^−1 was chosen for subsequent experiments, corresponding to a 1:8 ratio. Due to the required minimum TMT concentration, further downscaling of the overall approach was not feasible as this would reduce the total reaction volume below 20 µL. Although working at a lower volumetric scale is technically achievable, it presents challenges in handling samples due to viscosity and foaming. Nevertheless, achieving a protein amount of 10 µg may not always be feasible depending on the cell type and cell culture setup. Therefore, TMT label amount was kept constant and protein amount reduced to 5, 2, and 1 µg respectively. As a result, the protein/TMT label ratio increased to 1:16, 1:40, and 1:80 respectively. When assessing label efficiency, reducing protein amount does not result in a significant reduction, it stayed in a range of 99.2% to 99.8% (Figure [81]2 ), within the manufacturer's recommended range for labeling efficiency. Furthermore, protein and peptide identifications (≈2300/≈9500) remained unaffected (two‐way ANOVA and TUKEY´s test p>0.05) by the lower protein amounts. Regardless of utilized TMT6‐plex labels (129 and 131 Da) results stayed consistent and therefore proved label independency. Figure 2. Figure 2 [82]Open in a new tab Peptide/protein identifications (FDR 0.01, database: Homo sapiens, UP000005640) of 1, 2, or 5 µg HeLa digests (n = 3) labeled with 80 µg TMT6‐plex label (129 and 131 Da) corresponding to a 1:80, 1:40 and 1:16 protein/TMT label ratio and measured via LC‐MS (DDA, three technical replicates, 100 ng on column). Labeling efficiency was calculated based on the number of labeled PSMs to the total PSM count. Values displayed as average ± SD. A significant concern when working with such low protein concentrations is the potential for off‐target labeling, especially when an excess of TMT reagent is present. This can lead to the unintended labeling of amino acid residues like serine, threonine, and histidine, ultimately compromising the accuracy of quantification. Serine is the amino acid most commonly affected by off‐target labeling, and therefore, the manufacturer recommends that <5% of serine residues should be labeled with TMT. For protein amounts of 5, 2, and 1 µg, we did not observe significant off‐labeling. In fact, the off‐target labeling consistently remained below 4.7%. In summary, for the optimized TMT6‐plex workflow, it is advisable to use 1–10 µg of protein with 4.1 µg of TMT6‐plex label. These parameters guarantee a label efficiency >99% and negligible off‐target labeling. As a result, the vendor's approach could be downscaled by a factor of ten regarding TMT labels and by a factor of 20–100 for protein amount. 2.2. Proof‐Of‐Concept For the proof‐of‐concept of the advanced TMT protocol, we assessed the effects of eight, benchmark NMs sourced from JRC repository, representing four important, commercial relevant material classes, encompasses TiO[2], CeO[2], ZnO, and SiO[2], each represented in two different variants (Table [83]1 ). Additionally, CuO was included as a positive control for comparative purposes. Table 1. Primary characterization of the test materials. # Abbrev. JRC # info size[84] ^a) [nm] refs. 1 TiO[2] NM‐101 JRCNM01001a anatase 5–6 [[85]37] 2 TiO[2] NM‐105 JRCNM01005a rutile‐anatase 15–24 [[86]37] 3 CeO[2] NM‐211 JRCNM02101a precipitated, uncoated, cubic <10–20 [[87]38] 4 CeO[2] NM‐212 JRCNM02102a precipitated, uncoated 18–38 [[88]38] 5 SiO[2] NM‐200 JRCNM02000a synthetic amorphous silica, precipitated 14–23 [[89]39] 6 SiO[2] NM‐201 JRCNM02001a synthetic amorphous silica, precipitated 17–19 [[90]39] 7 ZnO NM‐110 JRCNM62101a uncoated 68 [[91]40] 8 ZnO NM‐111 JRCNM01101a coated with triethoxycapryl silane 76 [[92]40] 9 CuO[93] ^b) – – <50 [94]Open in a new tab ^^a) determined by SEM or TEM; ^^b) obtained from Sigma (Ref.# 544 868). By including this array of materials and variants, the investigation aims to uncover nuanced cellular responses that may be specific to both the type of NM and its particular variant. Human lung epithelial A549 cells, a commonly cell line used to investigate inhalation toxicity,^[ [95]^29 ^] were exposed to the different NMs at concentrations that remained below cytotoxic levels for a duration of 24 h (Alcolea‐Rodriguez et al., under review). Three distinct passage numbers, serving as biological replicates, were considered, with two corresponding technical replicates collected for each passage number, enabling measurements using either LFQ or TMT. 2.2.1. Application of Traditional LFQ Traditional LFQ measurements were performed using a conventional approach to investigate the cellular responses of A549 cells to the different NMs. Protein digests of treated cells were measured in three biological and three technical replicates for each sample, enabling a quantitative assessment of potential proteome alterations. MaxQuant results were imported into Perseus, where contaminants, decoy proteins, and proteins “identified by site only” were removed (see full workflow in Data [96]S1 (Supporting Information), Figure [97]1). Depending on the sample, within the Homo sapiens (UP000005640) proteome, up to of ≈30 200 peptides derived from ≈3590 proteins (FDR ≤ 0.01) were identified (Table [98]2 ). To ensure precise quantification results, a threshold of 70% valid values for LFQ intensities was implemented following log2 transformation. Depending on the sample up to ∼3360 proteins could be quantified with high confidence (Table [99]2; Data [100]S2 (Supporting Information), Table [101]1, [102]2). Significant differences in protein levels (t‐test, FDR ≤ 0.01, s0 = 0.1) were identified through a comparative analysis of the treated sample versus the untreated control. Samples treated with CuO showed significant alterations in the level of ten proteins. Conversely, in samples subjected to treatment with CeO[2] NM‐212 and ZnO NM‐110, two and five proteins, respectively, exhibited statistically significant changes in protein levels. Notably, no discernible effects on protein levels were observed in samples treated with TiO[2] NM‐101, TiO[2] NM‐105, CeO[2] NM‐211, or ZnO NM‐111. Considering that the positive control CuO, the most reactive particle, had no substantial effects on A549 cells, we omitted measuring samples from SiO[2] treatment due to their expected low reactivity. Table 2. Peptide and (quantifiable) proteins IDs (FDR 0.01, database: Homo sapiens, UP000005640) in A549 samples (n = 3) untreated or incubated with either TiO[2] NM‐101, TiO[2] NM‐105, CeO[2] NM‐211, CeO[2] NM‐212, SiO[2] NM‐200, SiO[2] NM‐201, ZnO NM‐110, ZnO NM‐111, CuO and quantified via LFQ or TMT6‐plex. ID protein IDs quantifiable proteins altered proteins[103] ^a) peptide IDs TMT LFQ TMT LFQ TMT LFQ TMT LFQ Control Ti 2671 3587 2001 3324 n.a. n.a. 15890 30216 TiO[2] NM‐101 2671 3591 2001 3363 97 0 15890 30213 TiO[2] NM‐105 2671 3593 2001 3330 217 0 15890 30143 Control Ce 2671 3579 2001 3073 n.a. n.a. 15890 29336 CeO[2] NM‐211 2671 3566 2001 3164 27 0 15890 29049 CeO[2] NM‐212 2671 3548 2001 3111 9 2 15890 28711 Control Si 2497 – 1847 – n.a. n.a. 14969 – SiO[2] NM‐200 2497 – 1847 – 3 – 14969 – SiO[2] NM‐201 2497 – 1847 – 3 – 14969 – Control Zn 2497 3151 1847 2773 n.a. n.a. 14969 22112 ZnO NM‐110 2497 3161 1847 2803 68 5 14969 22269 ZnO NM‐111 2497 3161 1847 2795 6 0 14969 22110 Control Cu 2678 3170 1841 2844 n.a. n.a. 15922 22530 CuO 2678 3151 1841 2811 127 10 15922 22355 [104]Open in a new tab ^^a) in comparison to control (FDR 0.01); n.a. not applicable. 2.2.2. Application of the advanced TMT protocol Although the above mentioned LFQ analysis enabled the quantification of a multitude of proteins, the treatment of A549 cells with various NMs minimally influenced their protein levels. Therefore, a more sensitive method is required for toxicity screening purposes. Hence the protocol was applied to the same experimental setup to prove its efficacy. During TMT labeling, a label switch was performed, ensuring that three different labels were used for the three biological replicates, guaranteeing label‐independent results. Given that the labels themselves are not 100% isotopically pure, a TMT6‐plex reporter mass was always skipped during the label switch, as M+1 has the largest isotope peak and most significantly influences the neighboring channels' intensity. An illustrative representation of the applied label switch setup is shown in Figure [105]3 . Figure 3. Figure 3 [106]Open in a new tab TMT 6‐plex labeling strategy employed for the proof‐of‐concept toxicity screening to minimizing potential biases introduced by specific labels. Each biological replicate is labeled with a unique TMT tag. Subsequently, the TMT‐labeled A549 samples were combined in an equal mass ratio and analyzed by LC‐MS/MS, with each sample measured in three technical replicates. Database search was conducted in MaxQuant, with “match between runs” and normalization based on the “weighted ratio to reference channel” enabled with all samples set as reference channels, according to Yu et al.^[ [107]^41 ^] The isotopic impurities of TMT labels were accounted for during database search by including correction factors provided by the manufacturer. Depending on the samples, up to 2670 proteins were identified based on approximately 15800 peptides. When considering a threshold of 70% valid values, slightly over 2000 proteins were quantified (Table [108]2; Data [109]S2 (Supporting Information), Table 3‐5). In contrast to LFQ results peptide and protein identifications as well as the quantifiable proteins were reduced by ≈30%. However, when comparing A549 controls with NM treated samples a higher number of TMT labeled proteins showed significant alterations in their abundance levels (Table [110]2). In samples treated with CuO, which served as positive control, a total of 127 altered proteins were identified. Similarly, treatment with TiO[2] NM‐101 or TiO[2] NM‐105 resulted in the altered levels of 97 and 217 proteins, respectively. In contrast, the treatment with ZnO NM‐110 affected 68 proteins, whereas only six proteins showed significantly affected levels in response to ZnO NM‐111 treatment. CeO[2] NM‐211 induced changes in protein abundance in 27 instances, while CeO[2] NM‐211 impact was minimal (nine proteins). As anticipated, treatment of A549 cells with SiO[2] NM‐200 or SiO[2] NM‐201 led to changes in the levels of only three proteins, showing almost no effect. After conducting a principal component analysis (PCA), it was observed that the technical replicates cluster more effectively when samples were analyzed with TMT instead of LFQ (Figure [111]4 ). Figure 4. Figure 4 [112]Open in a new tab Principal component analysis (PCA) of A549 (ATCC CRM‐CCL‐185) human lung epithelial cells untreated (●) or incubated (50 µg mL^−1) with TiO[2] NM‐101 (●), TiO[2] NM‐105 (●), CeO[2] NM‐211 (●) or CeO[2] NM‐212 (●) for 24 h (37 °C, 5% CO[2]) and analyzed (three biological and three technical replicates) via LFQ (A, B) or TMT6‐plex (C, D) in DDA mode via LC‐MS/MS. 2.3. Biological interpretation of NM toxicity Following the identification of significantly altered proteins using TMT measurements (see Data [113]S3 (Supporting Information), Tables 1–9), we conducted KEGG pathway enrichment analysis to gain insights into the cellular changes resulting from the various NM treatments. The summarized results are presented in Table [114]3 . Table 3. Effect of NM treatment on KEGG pathways in A549 cells, as observed from TMT‐proteomic analysis. NM KEGG pathways affected Count FDR TiO[2] NM‐101 Spliceosome 9 of 132 7.1E‐06 Ribosome 8 of 131 5.0E‐05 TiO[2] NM‐105 Folate biosynthesis 5 of 25 8.4E‐4 Pentose phosphate pathway 5 of 29 1.3E‐3 Glycolysis / Gluconeogenesis 8 of 64 1.2E‐4 Fructose and mannose metabolism 4 of 32 1.6E‐2 Biosynthesis of amino acids 8 of 73 1.5E‐4 Glutathione metabolism 5 of 52 1.1E‐2 CeO[2] NM‐211 Spliceosome 4 of 132 9.7E‐3 CeO[2] NM‐212 No alterations ZnO NM‐110 Ribosome 9 of 131 2.0E‐07 ZnO NM‐111 No alterations SiO[2] NM‐200 No alterations SiO[2] NM‐201 No alterations CuO Spliceosome 9 of 132 6.1E‐05 Ribosome 8 of 131 1.6E‐4 Necroptosis 7 of 147 1.9E‐3 RNA transport 7 of 161 2.9 E‐3 [115]Open in a new tab TiO[2] NM‐101 and TiO[2] NM‐105 are spherical particles with comparable sizes of 7 and 22 nm, respectively.^[ [116]^42 ^] The former is primarily composed of anatase, while the latter is a combination of anatase and rutile phases.^[ [117]^37 ^] The alteration of KEGG pathways – specifically the spliceosome and ribosome pathways – due to TiO[2] NM‐101 treatment indicates a significant effect on the cellular processes related to RNA splicing and protein synthesis. Changes in the spliceosome pathway may indicate alterations in the processing of pre‐mRNA transcripts, which can influence gene expression patterns. On the other hand, modification to ribosome pathway can affect the rate and accuracy of protein production. Of the different NM treatments evaluated, the exposure to TiO[2] NM‐105 stands out as the NM causing the most profound alterations. It causes changes across key metabolic and biosynthetic pathways (Table [118]3). Notably, TiO[2] NM‐105 significantly influenced three KEGG pathways tied to metabolism and cellular energy production, a distinction not shared by its counterpart, TiO[2] NM‐101, nor the other evaluated NMs. The pentose phosphate pathway plays a role in generating ribose‐5‐phosphate and NADPH, contributing to nucleotide biosynthesis and antioxidant defense. Glycolysis/gluconeogenesis pathways are central to cellular energy metabolism. The fructose and mannose metabolism pathway relates to the metabolism of fructose and mannose sugars, and any changes can have consequences for energy production and the availability of precursors for glycolysis and other vital metabolic pathways. CeO[2] NM‐211 treatment induced changes only in the spliceosome pathway. ZnO NM‐110 treatment affected ribosomal pathways, while ZnO NM‐111 appeared to cause no cellular alterations at the proteome level. Unlike TiO[2] NM‐101 treatment, cellular alterations observed for these two materials appear to be less pronounced. Treatment with CuO generate extensive cellular changes, as expected for a positive control.^[ [119]^43 , [120]^44 ^] In addition to the alteration of the KEGG pathways spliceosome and ribosome, CuO appears to affect proteins involved in the necroptosis pathway. Changes in this pathway could indicate an impact on the cellular fate, leading to a form of regulated cell death distinct from apoptosis. Changes in RNA transport pathway may affect the proper transport of RNA molecules and, as a result, gene expression regulation. 3. Discussion As hazard assessment evolves toward animal‐free testing, mechanistic understanding of substance toxicity is needed. In this context, omics techniques, particularly proteomics, emerge as valuable tools. This development aids in elucidating the MoA of substances and, in addition, contributes to the advancement of AOP and the development of NAMs. To evaluate the efficacy of our novel TMT labeling protocol, we employed a selection of benchmark NMs known for their comprehensive physico‐chemical characterization including SEM/TEM results (see references in Table [121]1). In addition, the nano‐safety aspects of