Anchorage-independent growth (soft-agar colony formation assay) and migration and invasion assays were considered basic carcinogenic hallmark assays for almost all instances of NM-induced cell transformation when a human cell line was targeted. As exceptions, few studies chose only proliferation, migration, and invasion assays as proof of induced cell transformation [65,70,72]. Matrix metalloproteinases (MMPs), specifically MMP-2 and MMP-9, were used as biomarkers for NMNM-induced neoplastic transformation [37,39,42,43,60,72]. A special emphasis was given to the ‘assays’ performed, which not only confirmed the cancer phenotypic hallmarks in transformed cells but are also linked with other mechanistic markers to delineate the malignant transformation process, such as oxidative stress, inflammation, DNA damage/repair (genotoxicity), epigenetic modification, apoptosis resistance, global (transcriptomics or proteomics) or selected gene/protein expressions, etc.

Human bronchial epithelial cells (BEAS-2B, 16HBE, HBEC-3KT), human small airway epithelial cells (SAEC), human pleural mesothelial (MeT5A), mouse embryonic fibroblasts (MEF), mouse colon epithelial cells (IMEC), mouse embryonic (NIH 3T3), and normal rat mesothelial (NRM2) cells were mainly applied in NM-induced cell transformation assays. Besides those cell lines, Bhas 42 cell lines were used to perform OECD-test guideline-based CTAs. In particular, a co-culture cell model (macrophages-like THP-1 and mesothelial, MeT5A) was used for induced cell transformation by MWCNT [70]. Some studies also include cancer cell lines from the same tissue as a positive control counterpart for specific biomarker analysis. For instance, transformed BEAS-2B cells were compared with A549 cells (pneumocyte type I-like cell line derived from a lung cancer patient) [51], while transformed normal rat mesothelial (NRM2) cells were compared with rat mesothelioma (ME1) for OPN expressions [65].

The time required to induce cell transformation by nanomaterials ranges from 4 weeks to 26 weeks. Most studies, with few exceptions, did not demonstrate a link between the cell carcinogenic hallmark with the time of exposure, so it is difficult to discuss the optimum exposure time point needed for malignant transformation for particular nanomaterials to indicated cell lines, as it depends on the choice of NM and the in vitro cell model combination. Dynamic changes during the exposure stage were analyzed with related assays (ROS, cell apoptosis, cell cycle, as well as MMP expressions, etc.) at different exposure stages (0, 3, 7, 15, 30, and 60 days); nonetheless, the carcinogenic phenotypic assays (cell migration, invasion, wound healing, soft-agar colony formation, in vivo xenograft) were only performed at the end of the exposure (60 days) and recovery (30 or 60 days) phase [74]. Two time points (6th week and 10th week) were chosen for the assessment of the carcinogenic potentiality of cerium oxide (CeO₂-NP) and ferric oxide (Fe₂O₃-NP) NPs in SEAC. Clear time dependency was observed in Fe₂O₃-NP exposure with the cancer hallmark phenotypes, such as increased proliferation, invasion, and the ability to form colonies on soft agar [59]. In a similar pattern, the 8th and 12th weeks of exposure were selected for the cell transformation ability of pristine and functionalized MWCNTs. Nonetheless, time dependency was not clear in assessing cancer phenotypes among MWCNTs [69]. Time-dependent neoplastic phenomena were evident in AgNP exposed Caco-2 cells [43], Co-NP [37] and ZnO-NP [39] exposed MEF cells, and MWCNT (NM-400) and TiO₂-NP materials (NM62002) treated human bronchial epithelial cell line (HBEC-3KT) [57]. Kornberg et al. demonstrated oscillatory time effects on anchorage-independent colony formation on BEAS-2B cells exposed to Fe₂O₃-NP (uncoated and silica coated), gas metal arc mild steel-welding fumes (GMAMS), and TiO₂-NP [50]. Fe₂O₃-NP and GMAMS-exposed cells exhibited a significant increase (1.2- and 1.4-fold) in colony number at 83 days, gradually decreased by 111 days, and then became significantly elevated again (1.6-fold) at 138 days, which remained the same throughout the exposure period. In the case of TiO₂-NP exposure, a significant increase (1.3-fold) in colony formation occurred between 111 to 138 days, which then ablated and returned to be similar to the control cells by 174 days and remained at baseline levels until the end of the exposure period.

Mostly, a single sub-cytotoxic concentration was used to assess the NMs’ ability to induce cellular transformation; however, some studies also demonstrated a concentration-related carcinogenic potential by exposing NMs to various concentrations. A concentration-dependent increase in cell transformation ability was demonstrated in Balb/c 3T3 [41] and Caco-2 cells [43], as well as induced malignancies in BEAS-2B cells exposed to Ag-NP [42]. Synthetic amorphous silica nanoparticle (SAS) (NM-200, NM-201, NM-202, and NM-203)-exposed Bhas 42 cell transformation model showed concentration dependency [45,46]. Only the highest concentration (20 µg/mL) showed significant cell transformation ability in TiO₂-NP (NM102)-exposed BEAS-2B cells [54]. Cell transformation efficiency was found to be concentration-dependent in pristine and citrate-coated zirconium oxide (ZrO₂-NP) and citrate-coated TiO₂-NP [58] while not markedly concentration-dependent in Balb/3T3 cells exposed to MWCNTs [64]. In the same way, malignant transformation of BEAS-2B cells by MWCNT was not clearly concentration-dependent (i.e., significant soft-agar colonies were evident in the lowest (1 µg/mL) and the highest (20 µg/mL) but not in the medium concentration (10 µg/mL)) [55]. On the contrary, SWCNT exposure to MeT5A cells exhibited clear concentration dependency [73].

Some studies have also taken into account the combined effects of exposure to other environmental pollutants. For instance, cigarette smoke condensate (CSC) has been evaluated as a secondary environmental pollutant that is often co-exposed with NMs [60,61]. In another study, the carcinogenic effects of polystyrene nanoplastics were evaluated in combination with Arsenic (As^III), given that both agents are persistent water contaminants [63].

Carbon-based nanomaterials (CNT), MWCNT and SWCNT (functionalized, aged, pristine), followed by silica nanoparticles are among the most studied NMs for cell transformation. Mainly CNT-related studies have demonstrated the connection between nano-specific (physicochemical) characteristics, such as surface functionalization, coatings, shape, size, and carcinogenic potentiality (induced cell transformation ability). The surface functionalization of CNTs (-NH₂, -COOH, -OH, etc.) possesses a great impact on cell transformation efficiency, which may further differ with exposure concentrations. The –NH₂-functionalized MWCNT showed significant anchorage-independent growth ability in soft agar but minimal ability of invasion in comparison to pristine and –COOH MWCNTs [64,69]. The increased cell transformation efficiency observed in the –NH₂-functionalized MW-NHx may be attributed to its heightened reactivity with biomolecules, which could be a result of the reduced oxygen content and greater exposure of the carbon surface [69]. Shape as the determinant characteristic of MWCNT carcinogenic potentiality was documented when normal rat mesothelial (NRM2) cells were exposed to MWCNT, tangled vs. rigid. Exposure to the rigid, but not to the tangled, resulted in the transformation of NRM2 cells into an invasive phenotype. Furthermore, the study also postulated osteopontin (OPN) as a biomarker of in vitro cell transformation [65]. In general, the molecular mechanism of cell transformation, distinct from asbestos, shared similarities among CNTs (MWCNT vs. SWCNT). The studies with various carbon-based nanomaterials exhibited similar responses to malignant transformation hallmarks, such as migration and invasion [72], gene expression signatures [71], and cancer stem cell-like properties [67]. Size-related studies (diameter or aspect ratio) in CNTs did not obtain a clear outcome because the results variation is possibly related to the cell line of choice. MWCNT with a lower average diameter (NM-400) showed higher carcinogenic potentiality than that of a thicker diameter and high aspect ratio (NM-401), attested by soft-agar colony formation in the cell line HBEC-3KT (human bronchial epithelial cell) [57]. Conversely, mitsui-7, an MWCNT that possesses a morphological resemblance with NM-401, caused malignant transformation on human lung small airway epithelial cells (SAECs) [71].

Size-dependent cell transformation was observed when the HBEC-3KT cell line was exposed to the TiO₂-NP and MWCNT with a smaller diameter but not to that with a larger diameter, in the soft-agar colony formation assay [57]. SAS (NM-200 and NM-201, NM-202 and NM-203) were reported as tumor-promoter substances in the Bhas 42 cell transformation model, and a marked size-dependency was apparent, in which a higher the particle size indicated a higher transformation ability [45]. In a comparison between coated (either with citrate or silica) vs. uncoated TiO₂-NP and ZrO₂NP, it was shown that coating with silica seems to prevent Balb/3T3 morphological transformation induced by ZrO₂ NP [58]. In a similar line of evidence, silica-coated iron oxide nanoparticles (nFe2O₃) do not induce neoplastic transformation in BEAS-2B cells [50]. A single study with nanocellulose (CNC) showed type/form (gel vs. powder)-specific carcinogenic potentiality assessed by anchorage-independent growth (of soft-agar colony formation), as well as cell migration assays [62].

Nanoceria (CeO₂-NP) [59,60], nickel (NiNP) and nickel oxide (NiO-NP) nanoparticles [52] alone did not show clear cellular transformation, which was attested with in vitro cancer hallmark assays, such as soft-agar colony formation, cell migration, and invasion.

Contradictory results are reported for ZnO-NP, which did not induce cellular transformation in mouse embryonic fibroblast (MEF) [38,39] but was able to cause malignant transformation in mouse colon epithelial cells (IMECs) [40] and human embryonic kidney (HEK293) and mouse embryonic fibroblast [53] (NIH/3T3) cells [78], possibly through the CXCR2/NF-kB/STAT3/ERK and AKT pathways. An exact comparison between studies is difficult as these studies used different ZnO-NP with no similarity in physicochemical properties and no standardized protocol followed either. In a similar way of contradiction, while most studies reported positive cell transformation, some studies demonstrated no cell transformation in TiO₂-NP [58] and SiO2-NP-exposed [49,50] cells.

Oxidative stress, one of the main characteristics of carcinogens, is known to play a paradoxical role in cancer development. It can either support the transformation/proliferation of cancer cells by initiating/stimulating tumorigenesis or induce cell death [7,79]. Oxidative stress (such as reactive oxygen species (ROS) formation) associated with NM-induced neoplastic transformation was observed in various studies [37,39,50,53,54,55,59,62,74]. Nevertheless, none of the studies performed absolute phenotypic anchoring of carcinogenesis in connection with oxidative stress. For instance, whether low levels of ROS could alleviate NM-induced cell transformation capability by applying inhibitor/scavenger of ROS while NM treatment was not determined. Chronic inflammation predisposes to carcinogenesis at all stages of tumor formation [7,79]. BEAS-2B cells treated with cellulose nanocrystals (CNCs) (powder and gel form) cause the secretion of various pro- and anti-inflammatory cytokines, chemokines, and growth factors (IL-1β, IL-2, IL-4, IL-9, eotaxin, IL-1RA, IL-6, IL-8, G-CSF, IP-10, IL-15, TNF-α, PDGF-bb, RANTES, etc.), leading to neoplastic transformation [62]. A significant reduction in cytokine (IL-1β, IL-6, and IL-8) expressions was observed in correlation with MWCNT (NM403)-induced malignant transformation [55]. One study explicitly postulated the carcinogenicity of MWCNT through inflammation. The results from their long-term MWCNT-treated co-culture cell model (macrophages and mesothelial cells) indicated that IL-1β, secreted by macrophages, may significantly enhance the release of inflammatory cytokines (IL-8, TNF-α, and IL-6) from mesothelial cells. In particular, the NF-κB/IL-6/STAT3 pathway played a pivotal role in the malignant transformation of mesothelial (MeT5A) cells induced by MWCNTs [70].

Genotoxicity is among the most common and important characteristics of carcinogens [7]. It can be assessed by evaluating DNA damage, defects in the mechanisms of DNA damage repair, DNA damage response through several well-established in vitro assays, such as Ames, micronucleus and HPRT forward mutation assays (OECD genotoxicity test battery), comet assays, chromosomal aberration, altered expressions of DNA damage response/repair proteins, etc. These assays are continuously modified and adapted for safety assessments of nanomaterials [80,81]. The approaches adapted for genotoxicity evaluation in connection with carcinogenesis assessments are mainly the comet assay [37,39,52,54,55,58,62], the micronucleus assay [42,52,53,54,55,58,64,68], the chromosomal aberration assay [70], and the expressions of DNA damage response proteins (γ-H2AX and p53) [50,67]. The possibility of NM-induced oxidative DNA damage assessment cannot be ignored as NM-induced ROS formation and oxidative stress is a well-established phenomenon. DNA damage induced by oxidative stress and involvement of OGG1 gene, as well as the glycosylase of the base excision repair pathway which eliminates 8-oxoguanine lesions from DNA, were documented for cobalt (Co-NPs) and zinc oxide (ZnO-NP) nanoparticles, using the wild-type (M.E.F. Ogg1^+/+) and isogenic knockout (M.E.F. Ogg1^−/−) mouse embryonic fibroblast [37,39]. It was also shown that MTH1 could serve as a candidate biomarker to unravel NM (Co-NPs and ZnO-NP)-induced potential genotoxic and carcinogenic effects [38]. MTH1 is an important GTPase that helps to avoid the incorporation of oxidized nucleotides from the reservoir to the DNA by effectively degrading them. Nonetheless, most studies focus on DNA strand break but not on the mechanisms of genotoxicity, particularly how DNA damage repair is affected by NM exposure and in turn induces carcinogenesis. Cell-cycle aberrations as another potential mechanism of carcinogenesis have been reported in TiO₂-NP [53] and amorphous SiO₂-NP-induced [48] in vitro malignant transformation.

Epigenetic alteration is one of the critical characteristics of chemical carcinogens [7,82]; nonetheless, only a few studies have explored epigenetic alteration in connection with neoplastic-like transformation. Changes in epigenetic markers can be assessed by detecting global or gene-specific DNA methylation levels, histone modification (methylation, acetylation, phosphorylation etc.), or non-coding RNA (microRNA, lncRNA) expression levels [82]. Whole-genome DNA methylation microarray profiling delineated alteration of DNA methylation in cancer-related pathways as a potential underlying mechanism of post-chronic single-walled carbon nanotubes (SWCNTs) exposure-induced irreversible malignant transformation [74]. The integration of methylome and transcriptome data revealed altered genes expression, similar to lung adenocarcinoma and lung squamous cell carcinoma, for instance, promoter hypomethylation and upregulation of transmembrane serine protease 9 (TMPRSS9), proviral integration Moloney 2 (PIM2) genes or promoter hypermethylation and downregulation of calcium/calmodulin-dependent protein kinase II inhibitor 1 (CAMK2N1), and integral membrane protein 2A (ITM2A). Chronic exposure of nano silicon dioxide (Nano-SiO₂)-induced malignant cellular transformation was associated with global DNA (5mC) hypomethylation and reduced expressions and enzyme activities of DNMTs (DNMT1, DNMT3A, DNMT3B), as well as altered expressions of methyl-CpG binding proteins (MeCEP2, MBD2) in two types of human bronchial epithelial cells (16HBE and BEAS-2B cells) [51]. Moreover, the demethylation of NRF2 promoter activates the expression of NRF2, which plays a key role in Nano-SiO₂-induced carcinogenesis. A comparative carcinogenic potentiality in light of epigenetic modification (global DNA methylation) was carried out in Bhas42 mouse cell lines exposed to amorphous silica nanoparticles (NM-203) and crystalline silica particles (Min-U-Sil) [47]. Altered DNMT (DNMT1, DNMT3A, DNMT3B) expressions and global DNA (5mC) hypomethylation were evident in Min-U-Sil-exposed cells, but not in NM-203-exposed cells. In a similar line of evidence, increased histone (H4 but not H3) acetylation and modulated expressions of HDACs (HDAC1, HDAC2, HDAC3, and HDAC4) was observed only in Min-U-Sil-exposed cells. Furthermore, the transcriptional activation of the c-myc gene, a biomarker of carcinogenicity, through the regulation of epigenetic marks on its promoter was evident in both nano-silica-treated cells. Significant promoter modulation was observed for acetylated histone H3 lysine 4 (H3K4Ac), trimethylated histone H3 lysine 4 (H3K4me3), acetylated histone H3 lysine 9 (H3K9Ac), and acetylated histone H3 lysine 27 (H3K27Ac); however, no changes were evident for 5-methylcytosine (5-mC).

microRNA (miRNA) profiling, miRNA-mimic transfection, and gene and protein expressions analysis show that miR221 plays a critical role in MWCNT-induced neoplastic transformed cells. The miR221-ANNEXIN A1 axis was involved in the regulation of cell migration of the transformed cells [66]. Nanoceria CeO₂-NP exposure gave rise to cell transformation (invasion and tumorsphere induction) in association with the altered battery of miRNA expression [61]. A small set of five miRNAs (miR-23a, miR-25, miR-96, miR-210, and miR-502) were reported as biomarkers for NM-induced transformed cells, which were validated particularly for NMs (TiO₂NP, MWCNT, Co-NP, ZnO-NP, and CeO₂-NP) [56]. A recent bioinformatics-based study highlights lncRNAs (in particular four lncRNA, namely MEG3, ARHGAP5-AS1, LINC00174, and PVT1) and pseudogenes (specifically five pseudogenes, namely MT1JP, MT1L, RPL23AP64, ZNF826P, and TMEM198B) as candidate diagnostic biomarkers and drug targets for CNT-induced lung cancer [83].

Regarding the causal or associative role of biomarkers in NM-induced carcinogenesis, apoptosis resistance is considered one of the key characteristics of malignant transformation [84]. The role of p53 in the apoptosis-resistance process of SWCNT-induced neoplastic transformation of lung epithelial cells (BEAS-2B) was reported among the pioneering studies related to NM-induced in vitro carcinogenicity [77]. Several follow-up studies were carried out to elucidate the underlying mechanism of oncogenesis of the same SWCNT-transformed cell model. SWCNT-transformed cells did present an aggressive phenotype, including increased cell migration, invasion, anchorage-independent cell growth (in vitro), and tumor formation and metastasis (in vivo). It was observed that Slug, a key transcription factor that induces an epithelial–mesenchymal transition (EMT) was a central player in these mechanisms [85]. Overexpression of mesothelin (MSLN) in SWCNT-induced neoplastic cell model and the potential application of MSLN as a biomarker and therapeutic target for CNT-induced malignancies [86]. Global gene expression analysis of the same model delineated activation of the pAkt/p53/Bcl-2 signaling axis, Ras family proteins for cell-cycle control, Dsh-mediated Notch 1, and the downregulation of apoptotic genes BAX and Noxa [87]. Other global gene expression studies evidenced that CNT-induced neoplastic-like transformation in normal mesothelial cells (MeT5A) was associated with overexpression of cortactin and H-Ras-ERK1/2 signaling (in the case of SWCNT exposure) [73], activating the MMP-2 gene and its critical role in an invasive phenotypic trait (MWCNT and SWCNT exposures) [72]. Whole-genome microarray further demonstrated differential signaling pathways between CNT-induced vs. asbestos-induced malignant transformation in primary small airway epithelial cells (SAECs). Conversely, CNTs (MWCNT and SWCNT) shared similar signaling pathways as an underlying mechanism of in vitro cell transformation. For instance, CNT-induced neoplastic cells demonstrated altered cell death, proliferation, mobility, development signaling, inflammation-related signaling, lipid metabolic signaling, TNFR signaling, reduced immune response, and altered cancer-related canonical pathways [71]. Global gene expression analysis reveals that inactivated p53 and aberrant p53 signaling comprise major reasons for the amorphous silica nanoparticle (SiNP)-induced malignant transformation of BEAS-2B cells [48]. Other than the global gene expressions analysis, targeted gene/protein expressions or enzyme activities shed light on the underlying mechanism of NM-induced malignant transformation. The results from silver nanoparticle (Ag-NP)-induced transformed cells indicate cell migration/invasion and apoptotic resistance by complex regulation of MAPK kinase (p38, JNK, and ERK1/2) and p53 signaling pathways [42].

Research evidence indicates that cancer stem cells or stem-like cells (CSCs) are a subpopulation driving tumor initiation, progression, and metastasis. CSCs share characteristics with normal stem cells (NSCs); however, they have malignant phenotypic traits. CSCs can be identified based on stem cell surface markers, self-renewal capacity, potency for differentiation, resistance to apoptosis, unlimited proliferation, colony formation, formation of nonadherent spheroids, expression of epithelial–mesenchymal transition (EMT)-related transcription factors, and matrix metalloproteinase (MMP) secretion, xenograft tumor formation, etc. Particularly, CSCs are resistant to chemotherapy and sustain tumor growth and relapse after therapy [76,88,89]. Various CNTs (SWCNT, MWCNT, and ultrafine carbon black) showed CSC-like properties acquired through long-term exposure, as indicated by 3D spheroid formation, apoptosis resistance, and CSC marker expression through SOX2 and SNAI1 signaling [67]. SWCNTs caused the induction of CSC-like irreversible transformation with aberrant stem cell markers (Nanog, SOX-2, SOX-17, and E-cadherin and stem cell surface markers CD24^low and CD133^high) [75], and their mechanistic insights were reported, such as the role of SOX9 overexpression in CSC formation and tumor metastasis [76].

The EMT process refers to the process by which transformed epithelial cells acquire the abilities involved in cell migration, invasion, and eventual cancer metastasis [2,90]. EMT induction by NM exposure shed light on the carcinogenic potentiality and need for cancer hallmark assessment. BEAS-2B cells co-cultured with THP-1-derived macrophages exposed to SiNPs promote EMT via the AKT pathway by inducing the release of SDF-1α and TGF-α while combined with benzo[α]pyrene-7, 8-dihydrodiol-9, 10-epoxide [91,92]. TiO₂-NP can induce the EMT process in colorectal cancer (SW480) cells via the TGF-β/MAPK and WNT pathways [90]. MWCNT induces EMT in BEAS-2B cells via TGF-β-mediated Akt/GSK-3β/SNAIL-1 signaling pathway after extended (96 h) incubation at sub-cytotoxic concentrations [93]. A novel mechanism of CNT-induced carcinogenesis through the induction of cancer-associated fibroblasts, a critical tumor microenvironment component that provides the necessary support for tumor growth, has recently been described [94]. Moreover, the results also suggest the potential efficacy of podoplanin as a mechanism-based biomarker for rapid screening of carcinogenicity of CNTs and related NMs for their safer design [94]. RNA-seq analysis of silver nanoparticles exposed to BEAS-2B cells (1 µg/mL for 6 weeks) revealed fibrosis and the ‘epithelial–mesenchymal transition’ (EMT) pathway as a pivotal altered mechanism.