The objective of this study was to evaluate the silver clearance from tissues following the cessation of silver nanoparticle administration. The silver concentrations in the blood rapidly decreased during the first month of recovery and continued until 4 months, indicating continuous partitioning from the tissues to the blood. Other tissues, including the liver, spleen, ovaries, and kidneys, also showed a degree of clearance of the accumulated silver during the 4-month recovery period. However, the silver concentrations in the testes and brain did not decrease to the control levels, even after the 4-month recovery period, indicating that silver clearance is difficult across biological barriers, such as the blood–brain barrier or blood-testis barrier. The silver half-life in each tissue type was calculated using a zero-order elimination model (Table 1). However, since the R2 value was not significantly improved when compared with that calculated using a first-order elimination model, the zero-order elimination was only applied to the blood and liver treated with the 25 nm silver nanoparticles. A half-life estimation based on a zero-order elimination is primarily dependent on the initial concentration change. In previous literature, the systemic disposition of silver has also been described using first-order inter-compartment rate constants . Yet, a large inter-subject variability due to destructive sampling can result in a poor half-life estimation, and inter-individual variations of silver concentrations in biological samples are already known to be very high. Thus, the MRT, which represents the mean residence time of silver nanoparticles in tissue, would seem to be a better indicator of silver biopersistence. In the present study, the MRT showed similar values, regardless of the dose level, thereby indicating the biopersistence more clearly than the t1/2 values.
Although the clearance half-times differed according to dose and gender, the tissues with no biological barrier, such as the liver, kidneys, and spleen, showed a similar clearance trend. Furthermore, the MRT also showed differences in the silver biopersistence in various tissues. Thus, the silver concentration clearance was in the order of blood > liver = kidneys > spleen > ovaries > testes = brain. Therefore, the silver clearance from tissues containing biological barriers would appear to be differently regulated. As a result, the silver biopersistence in the testes and brain can complicate the risk assessment of silver nanoparticles. Silver biopersistence in the brain and testes was already observed in the case of 28 days of oral administration of silver nitrate (9 mg/kg bw/day), noncoated (<20 nm), and PVP-coated silver nanoparticles (90 mg/kg bw/day) followed by 8 weeks of recovery . However, the short-term (5 days) intravenous injection of various sizes (20, 80, and 110 nm) of silver nanoparticles followed by 17 days of recovery did not reveal any silver biopersistence in the brain and testes .
Wijnhoven et al.,  previously hypothesized that the toxic effect of silver is proportional to the free silver ions, yet how this relates to silver nanoparticles remains unclear. Notwithstanding, the results would seem to indicate two possible ways of clearing accumulated silver from the body: as silver ions or silver nanoparticles. However, there are clear difficulties in evaluating the silver ions originating from silver nanoparticles in vivo and estimating the fraction of ionization from silver nanoparticle surfaces in various tissues.
In the present study, the silver nanoparticles were dispersed in citrate after gas-phase synthesis, as described by Lee et al. . Stabilizing the layers prevents silver nanoparticle aggregation, enabling the nanoparticles to remain suspended in the water column. Silver nanoparticles are normally stabilized using a charged surface coating, such as citrate, that creates an electrostatic barrier to aggregation. In a natural environment, a stabilizing agent such as citrate, which is weakly complexed with the silver nanoparticle surfaces, can be rapidly displaced by ligands, such as sulfides . Surface-bound ligands, such as citrate, compete with oxygen for surface sites and thus decrease the rate of oxidation. Liu et al. provided evidence of this mechanism by demonstrating a decrease in the Ag + released from silver nanoparticles as the citrate surface coatings became more densely-packed . Nanoparticles in the GI tract show a different behavior depending on the pH in the stomach (pH 2–3) and intestines (pH 7.8), where unstable and agglomerated nanoparticles disperse well in an acidic and basic pH . Liu et al. recently proposed a conceptual model of ingested silver nanoparticles in the human body. In their model, argyrial silver deposits are not translocated engineered Ag-NPs, but rather secondary particles formed by partial dissolution in the GI tract followed by ion uptake, systemic circulation as organo-Ag complexes, and immobilization as zerovalent Ag-NPs by photoreduction in light-affected skin regions . The secondary Ag-NPs then undergo detoxifying transformations into sulfides and further into selenides or Se/S mixed phases through exchange reactions. The formation of secondary particles in biological environments implies that Ag-NPs are not only a product of industrial nanotechnology, but have also been long present in the human body following exposure to more traditional chemical forms of silver. Therefore, the lack of any difference in tissue distribution between the 10 and 25 nm silver nanoparticles after 28 days of oral administration indicated that the ingested silver nanoparticles were dissolved in the low pH gastric fluid environment, which then led to silver ion release. The silver ions and their complexes are brought into the bloodstream through ion or nutrient uptake channels and circulate systematically. Plus, the majority of silver in the circulation is predicted to be bound to thiol complexes, which have high binding affinities yet are easily exchangeable, giving Ag(I) a significant biomolecular mobility . Similarly, in vivo nanoparticle formation from silver ions has also been suggested following the detection of silver-containing nanoparticles in several tissues obtained from animals administered silver nitrate or silver nanoparticles orally for 28 days . Yet, the uptake of silver nanoparticles through the intestines cannot be excluded. In a previous 90-day silver nanoparticle (average 56 nm) oral administration study, a dose-dependent increase in the villi pigmentation was observed (presumably yet not confirmed as silver nanoparticle related) .
The formaline solution-based tissue preparation used in the present study did not allow the nanoparticles in the tissue samples to be examined by TEM. Yet, while TEM can be used to visualize silver nanoparticles in tissues, the particle shapes of the silver nanoparticles cannot be confirmed without the use of additional defining methodologies, such as EDX (energy dispersive X-ray analyzer). And even a TEM-EDX analysis may not provide exact composition information on target particles. Furthermore, the tissue presence of secondary particles generated from silver-ion complexes can also complicate whether particles are silver nanoparticles or secondary silver ion complexes. Thus, other recent technologies, such as hyperspectral images, need to be used in future studies.
In previous studies, certain forms of Mn (Manganese) have been reported to affect the brain. For example, in the case of rats, inhalation exposure to soluble forms of Mn, such as Mn sulfate and Mn phosphate, was found to increase the brain Mn concentration when compared to inhalation exposure to the less soluble Mn tetraoxide . After the cessation of welding-fume exposure (average diameter 100 nm), the absence of any further supply of soluble Mn to the lungs resulted in a reduction of the Mn concentrations in the blood and other tissues, including the brain, liver, and spleen . However, unlike Mn accumulation in the brain, the silver in this study exhibited a different pattern, showing poor clearance from brain tissue.
Furthermore, the different nanoparticle sizes used in this study did not have much effect on the ADME (absorption, distribution, metabolism, and excretion), although the size difference was relatively narrow. The distribution and clearance of silver from various tissues showed a similar pattern. Similarly, in other studies, the inhalation of 18–19 nm  and 12–15 nm  silver nanoparticles and oral dosing with 56–60 nm silver nanoparticles [4, 5] showed a similar tissue distribution. Therefore, this would seem to confirm the applicability of the 0-hypothesis for silver suggested by Wijnhoven et al.  and chemical transformation suggested by Lie et al. , which state that silver toxicity mainly originates from silver ions, in this case generated from the surface of the silver nanoparticles.
A certain level of liver toxicity was indicated by the increase of cholesterol (CHOL) in the male rats following the 28-day oral administration of the low and high dose of 10 nm silver nanoparticles and low dose of 25 nm silver nanoparticles, the increase of alkaline phosphatase (ALP) in the female rats following the low and high dose of 10 nm silver nanoparticles, and the increase of aspartate aminotransferase (AST) in the female rats following the high dose of 25 nm silver nanoparticles. While one case of bile duct hyperplasia was observed among the male rats treated with the 10 nm silver nanoparticles, the percent of inflammatory cell infiltration in the male rats was only mildly increased by the 10 nm and 25 nm silver nanoparticle treatments. The present study did not observe the more prominent bile duct hyperplasia reported in a previous 28-day oral study  and 90-day oral study . Thus, the development of bile duct hyperplasia may require a longer exposure duration than 28 days.
The present study also found that coating the silver nanoparticles, in this case with citrate, did not have any effect on the ADME when compared with the use of un-coated silver nanoparticles, including silver nanoparticles generated by evaporation/condensation, as used in other inhalation toxicity studies [6, 17], dry powder, as used in other oral toxicity studies [4, 5], and PVP-coated silver nanoparticles . When evaluating the toxicity of the citrate-coated silver nanoparticles used in this study, the extent of the toxicity was quite similar to that of the dry powder used in the previous studies by Kim et al. [4, 5].
Thus, when including silver biopersistence, especially in the brain and testes, this complicates the risk assessment of silver nanoparticles. Therefore, the NOAELs previously reported by Kim et al.  based on a subchronic oral toxicity study and Sung et al.  based on a subchronic inhalation toxicity study need to be adjusted in the light of biopersistence. Until now, histopathological evaluations of the testes and brain have not show any apparent toxicity, consequently, new tools are needed to evaluate biopersistent-relevant toxicity. Nonetheless, the current findings of minimized effects from silver nanoparticle coating and size differences on the ADME could simplify the risk assessment of silver nanoparticles.