Crystalline Silica Toxicology and Silicosis

RoseyFields Chemical-Specific Toxicology Profile

Crystalline Silica (SiO₂)

RoseyFields evaluates the toxicology of Crystalline silica using the following evidence base, drawn from the peer-reviewed toxicological and epidemiological literature on crystalline silica (c-silica) and, where informative for read-across, amorphous silica (a-silica). This profile synthesizes chemical identity, physicochemical determinants of toxicity, toxicokinetics, mode of action, the animal toxicology database organized by organ system, carcinogenicity findings, the relationship between experimental-animal findings and human epidemiology, and the principal data gaps that remain.

1. Chemical Identity and Physicochemical Properties Relevant to Toxicity

Silica is silicon dioxide (SiO₂) and exists in two broad classes: crystalline (c-silica) and amorphous (a-silica). The three principal naturally occurring crystalline polymorphs are quartz, cristobalite, and tridymite, each an ordered lattice of silicon–oxygen tetrahedra sharing the empirical formula SiO₂ and a molecular weight of 60.1. Quartz is by far the most abundant, occurring as at least a trace constituent of virtually all soils, sands, and many rocks. The physicochemical characteristics most relevant to toxicity are summarized in the table below.

Persistence and reactivity. Crystalline silica is chemically unreactive in the bulk environment and essentially insoluble in water, dissolving appreciably only in hydrofluoric acid (Zhuravlev 2000; Rimola et al. 2013). This low solubility is central to its toxicological behavior: because the particle does not dissolve, dissolution contributes negligibly to clearance from the lung, so deposited particles persist biologically for years. The persistence that follows from low solubility, rather than any classical metabolic property, is what drives the chronic, progressive nature of silica-associated lung disease.

Surface chemistry governs biological activity. Both crystalline and amorphous silica present surfaces composed of siloxane bridges (Si–O–Si) and silanol groups (Si–OH); contact with water cleaves surface Si–O bonds to generate additional silanols (Zhuravlev 2000; Rimola et al. 2013). The density and configuration of these silanol groups, and the population of surface radicals, determine hemolytic and inflammatory potency. Freshly fractured quartz—as produced during grinding or abrasive blasting—carries an abundance of reactive silanol and ionized-silanol radical sites and is markedly more cytotoxic than “aged” quartz; processing by heating, grinding, or chemical treatment can raise or lower surface reactivity (Murashov et al. 2006; Fenoglio et al. 2000; Donaldson and Borm 1998). Because toxicity tracks the surface rather than the bulk formula, two samples of the same polymorph can differ substantially in potency, which complicates read-across between studies and between industrial settings.

Bioaccumulation. Silica does not bioaccumulate in the conventional lipophilic sense (it has no meaningful log Kₒᵥ and is not lipid-partitioning). Instead, it exhibits pulmonary accumulation: because mechanical and dissolution clearance are slow, repeated inhalation leads to progressive retention of particles in the lung and draining lymph nodes, detectable in human lung tissue years after exposure ceases (Absher et al. 1992; Vacek et al. 1991). Particle size and crystallinity also modulate hazard—respirable particles (aerodynamic diameter ≤5 µm) reach the alveolar region, where the fibrogenic and inflammatory responses originate.

2. Toxicokinetics (Absorption, Distribution, Metabolism, Excretion)

2.1 Absorption

Inhalation is the toxicologically relevant route. Deposited particles are handled by three processes: mucociliary transport of larger (>5 µm) particles from the upper airways to the gastrointestinal tract; transport of respirable particles to thoracic lymph nodes; and macrophage phagocytosis with migration (Bailey et al. 2007). Dissolution—an important absorptive-clearance route for more soluble particles—is negligible for crystalline silica because of its very low solubility, and this is a principal reason quartz is cleared >10-fold more slowly than amorphous silica in rodents (Davis 1986; Kelly and Lee 1990; Reuzel et al. 1991; Schepers 1981). Lung retention is multiphasic: a fast mechanical phase and a slow phase (half-lives of years in humans) that reflects clearance of highly insoluble particles by phagocytosis. Among crystalline polymorphs, pure cristobalite is cleared more slowly than quartz and provokes a more intense inflammatory response (Hemenway et al. 1990).

A critical kinetic phenomenon is particle overload: at high lung burdens, macrophage-mediated clearance becomes saturated and clearance slows nonlinearly. The overload threshold is species-dependent and is lower in small mammals; rats overload at lower burdens than hamsters (Saffiotti et al. 1993; Snipes 1996; Morrow 1992). Overload produces a nonlinear relationship between inhaled dose and lung burden that materially affects low-dose and cross-species extrapolation. Oral absorption is minimal—gavage of 50 mg quartz to rats produced no detectable particle transfer out of the gut lumen (Gonzalez Huergo and Rojo Ortega 1991)—and dermal absorption is considered a minor pathway.

2.2 Distribution

Extrapulmonary distribution is limited but measurable. In rats, inhaled silica distributes chiefly to mediastinal lymph nodes and thymus, with only negligible amounts detected in blood, kidney, liver, and spleen, implicating lymphatic transport as the systemic route (Absher et al. 1992; Vacek et al. 1991). In humans, silica has been detected in kidney tissue and urine of exposed workers, consistent with systemic distribution and providing a mechanistic link to the renal effects seen epidemiologically (Ibrahim et al. 2011).

2.3 Metabolism

Absorbed silica is not metabolized. The only relevant molecular interaction identified is that silica solubilized from slate dust can bind serum albumin in vitro (Singh et al. 1984); this protein adsorption may be toxicologically important because denatured adsorbed protein has been proposed to acquire antigenic properties relevant to autoimmune and renal pathology.

2.4 Excretion

Silica absorbed from the respiratory tract is excreted in urine, as shown by detection of urinary silica in exposed ceramic workers (Ibrahim et al. 2011). Ingested silica is largely excreted unchanged in feces. Clearance studies in humans indicate that urinary silicon involves glomerular filtration and renal tubular secretion and correlates with urinary calcium, suggesting excretion as an orthosilicate complex; however, the quantitative relationship between urinary silica and cumulative exposure is unknown, so no validated exposure biomarker exists.

2.5 Species and sex differences relevant to interpreting animal data

The dominant interpretive issue is the species difference in macrophage overload. Rats are more susceptible to overload than hamsters, and it has been argued that lung-macrophage overload in rats may not be quantitatively relevant to humans (Saffiotti et al. 1993; Snipes 1996). This matters directly for carcinogenicity: crystalline silica is carcinogenic to rats by inhalation and intratracheal instillation but has not produced lung tumors in mice or hamsters, and the basis for this species difference is unresolved. Sex differences in the oral amorphous-silica database are modest and mostly confined to small organ-weight changes in females (e.g., decreased heart, liver, and spleen weights in female rodents at high dietary doses in Takizawa et al. 1988) without corresponding histopathology, so they are not considered toxicologically decisive. No PBPK model for crystalline or amorphous silica has been published, which limits formal cross-species dosimetric extrapolation.

3. Mechanism / Mode of Action

The unifying mechanism for the principal effects of crystalline silica is a self-perpetuating cycle of macrophage phagocytosis, cytotoxicity, and inflammation, driven by the reactive particle surface.

3.1 Initiation: phagocytosis, lysosomal damage, and the NALP3 inflammasome

Inhaled quartz depositing in the alveoli is phagocytosed by alveolar macrophages, an uptake mediated by scavenger receptors (notably MARCO and SR-B1) (Hamilton et al. 2008; Tsugita et al. 2017). Internalized particles destabilize the phagolysosome; the released contents and lysosomal enzyme activity trigger cytotoxicity and apoptosis (Hornung et al. 2008; Thibodeau et al. 2004). This activates the NALP3 (NLRP3) inflammasome—the assembly of NALP3 with the adapter ASC and pro-caspase-1—which is genetically required for experimental silicosis: NALP3-deficient animals are protected (Cassel et al. 2008; Hornung et al. 2008). Caspase-1 activation matures the pro-inflammatory interleukins IL-1β and IL-18, and reactive oxygen species (ROS) generated directly on freshly fractured surfaces or indirectly via the macrophage oxidative burst are required for full inflammasome activation.

3.2 Amplification and fibrosis (silicosis)

Dying macrophages release their silica load, recruiting fresh macrophages and repeating the cycle, producing chronic alveolitis. Released cytokines and chemokines (TNF-α, IL-1), ROS and reactive nitrogen species, and arachidonic-acid metabolites damage epithelium and matrix, while transcription factors NF-κB and AP-1 are up-regulated. A reparative phase driven by fibrogenic factors (TGF-β, IGF-1, EGF, IL-10) recruits and proliferates mesenchymal cells; persistent cycling between inflammatory and reparative phases produces excess extracellular-matrix deposition and fibrosis. TNF-α and IL-1 are required for silica-induced fibrosis in animal models, and miRNA-regulated increases in matrix proteins accompany lung-function decline in exposed workers (Rong et al. 2018). Gene-expression profiling of human lung epithelial cells shows silica up-regulates numerous immune and inflammatory pathways (Chan et al. 2017).

3.3 Carcinogenesis

Two non-exclusive pathways are proposed. The dominant view is a secondary, inflammation-driven genotoxicity: chronic inflammation generates ROS/RNS that damage DNA, while NF-κB/TNF-α signaling promotes survival and proliferation of transformed epithelial cells, expanding the preneoplastic pool; epigenetic silencing of the p16 tumor-suppressor via promoter hypermethylation has been reported in rat models. This pathway is threshold-dependent, requiring overload-driven inflammation. A secondary, non-threshold pathway proposes direct genotoxicity—silica particles interacting with DNA, disrupting mitotic chromosome segregation, and generating surface ROS; surface modification that blocks reactive sites abolishes oxidative DNA damage in vitro, underscoring the surface-dependence of genotoxic potential. Epidemiology showing lung cancer in workers without silicosis is cited in support of a component that does not strictly require fibrosis.

3.4 Autoimmune and renal mechanisms

Crystalline silica is a recognized immune adjuvant that non-specifically enhances antibody production. It is proposed to promote or accelerate autoimmune disease rather than cause it de novo, acting through chronic macrophage recruitment/destruction cycles in lymphatic tissue that stimulate T-helper and B-cell activity and, via Fas-dependent apoptosis, shift the CD4+CD25+ T-responder/T-regulator balance. The result is elevated autoantibodies (rheumatoid factor, antinuclear antibody, anti-Scl-70, ANCA). Renal injury arises either from direct accumulation of silica in the kidney with inflammation and fibrosis, or indirectly through autoimmune glomerular pathology (glomerulonephritis in ANCA-associated vasculitis and systemic sclerosis; immune-complex deposition in lupus). COPD in the absence of silicosis is attributed to ROS-mediated cellular damage and epithelial injury allowing particle penetration of small-airway walls (Hnizdo and Vallyathan 2003).

4. Animal Toxicological Evidence by Organ System

A key feature of the silica database is that adverse effects are essentially confined to the inhalation route; oral studies—almost all conducted with amorphous silica, with limited crystalline drinking-water data—generally identify no critical target organ. Because most repeat-dose and chronic animal work was performed with amorphous polymorphs (precipitated, pyrogenic/fumed, colloidal, and gel silica), the organ-system findings below draw heavily on that literature, with crystalline-specific data noted where available.

4.1 Respiratory (primary target)

The lung is the primary and most sensitive target. Acute and subchronic inhalation of synthetic amorphous silicas produces polymorph-dependent inflammation in the rat lung. In a 5-day study, precipitated and pyrogenic silica raised bronchoalveolar-lavage markers of cytotoxicity and inflammation and caused mild histopathology at ≥5 mg/m³, whereas silica gel required 25 mg/m³; minor lesions persisted 1–3 months after precipitated or pyrogenic (but not gel) exposure (Arts et al. 2007). In 13-week studies, Reuzel et al. (1991) reported increased lung weight, inflammation, granuloma formation, and increased collagen in Wistar rats, with focal interstitial fibrosis only after pyrogenic hydrophilic silica; fibrotic change persisted but did not progress up to one year, and other findings largely reversed. Subchronic pyrogenic-silica exposure at 50.4 mg/m³ produced septal fibrosis in F344 rats that regressed over an 8-month recovery (Johnston et al. 2000). Colloidal (Ludox) silica elevated inflammatory markers at ≥50 mg/m³ but not 10 mg/m³, and was less potent than precipitated silica (Warheit et al. 1991, 1995; Lee and Kelly 1992). Chronic single-concentration studies in monkeys showed early nodular pulmonary fibrosis and reduced lung function at 15 mg/m³ (Groth et al. 1981; Schepers 1962). The one natural-amorphous study (raw diatomaceous earth, 0% crystalline, in rabbits) produced alveolar-wall thickening and macrophage infiltration but no fibrosis (Tebbens 1957). By oral route, no respiratory effects were seen in rats at dietary pyrogenic silica up to 500 mg/kg/day for 6 months (Lewinson et al. 1994).

4.2 Cardiovascular

By the oral route, crystalline silica produced altered aortic endothelial vasoactivity: 3-month-old rats given 50 mg/kg/day as sodium metasilicate in drinking water for 8 days showed increased ex vivo contractile responses to phenylephrine and altered dilation responses, of uncertain toxicological significance (Öner et al. 2006). For amorphous silica by inhalation, cardiac hypertrophy in monkeys (15 mg/m³, 12 months; Schepers 1962) and hypertension with ventricular/auricular hypertrophy in rabbits (≥30 mg/m³; Schepers 1959) were reported, though limited animal numbers and cardiac-puncture artifacts temper interpretation (Schepers et al. 1957c). Dietary amorphous silica caused only small, histopathology-free heart-weight decreases in high-dose female mice (Takizawa et al. 1988).

4.3 Hepatic

No hepatic effects were identified for oral crystalline silica. For amorphous silica, inhalation produced no consistent liver clinical-chemistry or histological change in rats (Reuzel et al. 1991; Lee and Kelly 1992), though hepatocellular vacuolation was noted in monkeys at 15 mg/m³ (Schepers 1962). Orally, severe hepatic-epithelial atrophy occurred in rats at an extreme escalating pyrogenic-silica regimen (TWA 7,500 mg/kg/day, 8 weeks), with sporadic changes at 1,000 mg/kg/day in females only; the two-generation study saw no liver effects at gavage doses up to 1,000 mg/kg/day (Lewinson et al. 1994; Wolterbeek et al. 2015). In the 2-year dietary bioassay, only modest liver-weight decreases in high-dose females occurred without histopathology (Takizawa et al. 1988).

4.4 Renal

Renal effects are prominent in human crystalline-silica epidemiology (“silicon nephropathy,” glomerulonephritis, end-stage renal disease) and are supported by detection of silica in kidney and urine of workers (Ibrahim et al. 2011). Experimentally, oral crystalline silica as sodium metasilicate altered proximal-tubule mitochondrial function in rats (Öner et al. 2005), and early experimental work linked silicate to nephrotoxicity (Newberne and Wilson 1970; Dobbie and Smith 1982). Amorphous silica did not produce renal toxicity in the major oral or inhalation studies (Reuzel et al. 1991; Groth et al. 1981; Takizawa et al. 1988; Wolterbeek et al. 2015).

4.5 Immune / lymphoreticular

The human crystalline-silica literature associates exposure with a spectrum of autoimmune disorders (systemic sclerosis, rheumatoid arthritis, systemic lupus erythematosus, ANCA-associated vasculitis) and elevated autoantibodies, consistent with the adjuvant mechanism above. In contrast, amorphous silica produced no immunotoxicity in animal studies: no immune-system pathology in rats up to 30 mg/m³ (Reuzel et al. 1991) or in monkeys, rats, and guinea pigs up to 15 mg/m³ chronically (Groth et al. 1981; Schepers 1962), and only small high-dose spleen-weight decreases without histopathology in the dietary bioassay (Takizawa et al. 1988).

4.6 Hematological, gastrointestinal, and musculoskeletal

Amorphous-silica inhalation produced increases in neutrophils, hemoglobin, packed-cell volume, and erythrocytes in rats and rabbits at high concentrations, interpreted as an adaptive response to cardiopulmonary distress rather than direct hematotoxicity, and reversible on recovery (Reuzel et al. 1991; Schepers 1981; Lee and Kelly 1992). Oral amorphous silica caused no hematological change in dogs, rats, or mice across subchronic-to-chronic dietary studies (Newberne and Wilson 1970; Lewinson et al. 1994; Takizawa et al. 1988). No gastrointestinal or musculoskeletal toxicity was identified by either route (Reuzel et al. 1991; Groth et al. 1981; Lewinson et al. 1994).

4.7 Endocrine

No endocrine effects were identified. The two-generation study found no changes in adrenal, thyroid, pituitary, ovarian, or testicular weight or histology at gavage doses up to 1,000 mg/kg/day (Wolterbeek et al. 2015), and subchronic and chronic dietary pyrogenic-silica studies were likewise negative (Lewinson et al. 1994).

4.8 Neurological

Animal studies show no neurotoxicity: no changes in brain weight or nervous-tissue histology in rats inhaling amorphous silica up to 30 mg/m³ (Reuzel et al. 1991), and no clinical signs or brain changes across oral studies in rats and mice (Wolterbeek et al. 2015; Lewinson et al. 1994; Takizawa et al. 1988). Human oral (drinking-water) data are inverse or null: silica in drinking water was not associated with, and in some analyses was inversely associated with, cognitive impairment in the elderly (Jacqmin-Gadda et al. 1996; Gillette-Guyonnet et al. 2005).

4.9 Reproductive and developmental

The most robust study is a GLP two-generation reproduction study of precipitated amorphous silica in Wistar rats at gavage doses up to 1,000 mg/kg/day, which found no effect on mating, fertility, fecundity, gestation, live-birth or viability indices, sexual maturation, estrous cyclicity, or sperm parameters, and no developmental toxicity (no changes in pup weight, growth, sex ratio, or gross abnormalities) (Wolterbeek et al. 2015). A one-generation dietary study at 500 mg/kg/day was similarly negative for reproductive performance and offspring endpoints (Lewinson et al. 1994). For crystalline silica, a single drinking-water case-control study reported an association between the highest silica tertile and spontaneous abortion (Aschengrau et al. 1989), but multiple design limitations (surrogate community-water exposure, ad hoc exclusions, no consumption data) render it low-confidence; no animal reproductive data for oral crystalline silica exist.

5. Carcinogenicity Bioassay Findings

Crystalline silica. In experimental animals, crystalline silica (quartz) is carcinogenic to the lung in rats following inhalation or intratracheal instillation, and this finding underpins the classification of quartz/cristobalite as carcinogenic with sufficient evidence in experimental animals. Critically, no lung-tumor response has been demonstrated in mice or hamsters exposed to crystalline silica, and the mechanistic basis for this species divergence has not been established—consistent with the rat-specific overload phenomenon discussed above. The tumor type of concern is bronchoalveolar/lung carcinoma; the mode of action is predominantly the secondary, inflammation-driven genotoxic pathway secondary to chronic alveolitis and silicosis, with a possible direct genotoxic contribution.

Amorphous silica. Amorphous silica is not carcinogenic in the available bioassays. In a 2-year dietary bioassay, a-silica gel produced no neoplastic lesions attributable to treatment at doses up to 2,010 mg/kg/day in F344 rats or up to 6,010 mg/kg/day in B6C3F1 mice, although the study used small groups (18–21/sex/group) (Takizawa et al. 1988). A 24-month dietary pyrogenic-silica study in Wistar rats at 100 mg/kg/day likewise found no treatment-related neoplasia, but its reliability is low (small groups, no concurrent control, a single sub-MTD dose) (Lewinson et al. 1994). Chronic inhalation studies of amorphous polymorphs in multiple species also did not produce tumors (Groth et al. 1981; Schepers 1981). The weight of evidence indicates amorphous silica is not carcinogenic, though the database is limited by co-exposure to crystalline silica in occupational settings and sparse quantitative animal data.

6. Comparison of Animal Findings to Human Epidemiology

Human occupational epidemiology and the animal database converge on the lung as the critical target but diverge in important, mechanistically explicable ways.

Convergences. Both databases identify progressive pulmonary fibrosis (silicosis in humans; interstitial/nodular fibrosis in monkeys and rats) as the hallmark of chronic inhalation, and both support inflammation as the driving process. Human pooled and meta-analyses establish an exposure-response relationship for lung cancer: Steenland et al. (2001) reported a lung-cancer SMR of 1.2 across ~66,000 workers with a significant positive trend with cumulative exposure; Lacasse et al. (2009) estimated risk ratios of ~1.22 and ~1.84 at cumulative exposures of 1.0 and 6.0 mg/m³-year; and Finkelstein (2000) and Checkoway et al. (1997) reported comparable exposure-response slopes. Silicosis and renal-disease exposure-response relationships are likewise documented (Hnizdo and Sluis-Cremer 1993; Steenland, Sanderson, and Calvert 2001). The joint effect of silica and smoking on lung-cancer risk is more than additive and close to multiplicative (Liu et al. 2013). The rat carcinogenicity finding qualitatively supports the human lung-cancer signal.

Divergences and their toxicokinetic basis. The most consequential divergence is the species specificity of silica carcinogenicity—positive in rats but negative in mice and hamsters. This is widely attributed to the rat’s low threshold for pulmonary macrophage overload: overload-driven persistent inflammation, which the rat mounts readily and which may not scale to humans, is the presumed driver of the rat tumor response (Saffiotti et al. 1993; Snipes 1996; Morrow 1992). Consequently, the rat inhalation bioassay may overstate the inflammation-mediated component of human risk, and the nonlinear dose–burden relationship produced by overload complicates direct quantitative extrapolation. A second divergence is that the extensive oral animal database (largely amorphous) is essentially negative across organ systems, whereas the human hazard is entirely an inhalation, occupational phenomenon—reflecting the toxicokinetic facts that oral absorption is negligible and that hazard depends on delivery of persistent respirable particles to the alveolar macrophage compartment. A third is that human autoimmune and renal associations for crystalline silica have no clear counterpart in the amorphous animal studies, consistent with crystalline surface reactivity and lymphatic distribution being prerequisites that amorphous polymorphs and non-inhalation routes do not satisfy. Finally, the absence of any published PBPK model and of quantitative pulmonary-retention estimates means animal-derived points of departure cannot yet be placed on a rigorous human-dosimetry footing.

7. Data Gaps and Areas of Ongoing Research

Several gaps constrain quantitative risk characterization for crystalline silica:

Toxicokinetics. No PBPK model exists for crystalline or amorphous silica, and quantitative estimates of pulmonary deposition, retention, and clearance are lacking; well-controlled retention studies would clarify the toxic lung burden and improve cross-species extrapolation. Extrapulmonary distribution is described only qualitatively, and no validated biomarker links urinary silica to cumulative exposure (Ibrahim et al. 2011).

Surface/structure determinants. The mechanistic basis for differing fibrogenic and genotoxic potency among quartz, cristobalite, and tridymite—and between fresh and aged surfaces—remains incompletely defined; further work on silanol density, surface radicals, and particle size/shape is needed to explain why potency varies within a single polymorph (Murashov et al. 2006; Donaldson and Borm 1998; Fenoglio et al. 2000).

Carcinogenic mechanism and species relevance. The relative contributions of the threshold (inflammation-driven) and non-threshold (direct genotoxic) pathways are unresolved, as is the reason crystalline silica is carcinogenic in rats but not mice or hamsters; resolving the human relevance of rat overload is central to dose-response modeling (Saffiotti et al. 1993; Snipes 1996). The molecular events linking macrophage inflammasome activation to fibrogenesis and neoplasia continue to be an active research area (Cassel et al. 2008; Tsugita et al. 2017; Hornung et al. 2008).

Exposure-response for non-cancer endpoints. Quantitative exposure-response data adequate to derive NOAEL/LOAEL values are lacking for autoimmune disease and, to a lesser extent, renal disease and COPD; better-characterized occupational cohorts with reliable quantitative exposure metrics are needed (Steenland, Sanderson, and Calvert 2001; Hnizdo and Vallyathan 2003).

Amorphous silica. Because occupational amorphous-silica studies are confounded by crystalline co-exposure and lack quantitative data, and because in vitro genotoxicity results conflict, additional well-controlled studies are needed to confirm the apparent absence of chronic hazard and genotoxicity for amorphous polymorphs (Takizawa et al. 1988; Lewinson et al. 1994; Johnston et al. 2000).

Biomarkers of oxidative stress and inflammation (LDH, TNF, interleukins, Clara-cell proteins) are under investigation as early indicators of silica effect, but none is yet validated for exposure or effect surveillance.

Bibliography