Peripheral nerves directly mediate the transneuronal translocation of silver nanomaterials from the gut to central nervous system

Abstract

The blood circulation is considered the only way for the orally administered nanoparticles to enter the central nervous systems (CNS), whereas non-blood route-mediated nanoparticle translocation between organs is poorly understood. Here, we show that peripheral nerve fibers act as direct conduits for silver nanomaterials (Ag NMs) translocation from the gut to the CNS in both mice and rhesus monkeys. After oral gavage, Ag NMs are significantly enriched in the brain and spinal cord of mice with particle state however do not efficiently enter the blood. Using truncal vagotomy and selective posterior rhizotomy, we unravel that the vagus and spinal nerves mediate the transneuronal translocation of Ag NMs from the gut to the brain and spinal cord, respectively. Single-cell mass cytometry analysis revealed that enterocytes and enteric nerve cells take up significant levels of Ag NMs for subsequent transfer to the connected peripheral nerves. Our findings demonstrate nanoparticle transfer along a previously undocumented gut-CNS axis mediated by peripheral nerves.

Peripheral nerves mediate Ag NMs translocation along gut-CNS axis.

INTRODUCTION

Silver nanomaterials (Ag NMs) are the most widely used engineered nanomaterials in consumer products, including medical dressings, food ingredients, food packaging (E174), food surface protectants, baby bottles, and health care products, due to their antimicrobial activity (1–4). The estimated daily consumption of Ag NM is about 20 to 80 μg from food intake (5, 6). For the occupational population, the content of Ag NMs is up to 15 mg kg⁻¹ in the feces of occupational workers (25 persons) (7). After entering the gut lumen, Ag NMs can cross multiple biological barriers and transport to distal organs, thus inducing adverse outcomes (8). The central nervous system (CNS) has been considered the predominant organ afflicted by orally administered Ag NMs (9, 10). Because the transport route determines the final target organ and subsequent biological effects of exogenous nanoparticles (NPs), the specific transfer route map of NPs is crucial to unveiling their nanotoxicological as well as nanomedical potential. However, there are still substantial knowledge gaps in understanding of the transport of NPs from the gut to the CNS, which hinders both our understanding of the gut-CNS axis communication and the development of new therapeutic strategies that target CNS-associated diseases.

On the basis of our current knowledge, NPs present in the brain are believed to have arrived in the tissue from the blood circulation, which is considered to be the only path by which exogenous NPs entering the gut can access the brain (11). A non-blood biological route for NP transport into the CNS has not yet emerged either theoretically and experimentally. Similar to the blood circulatory network, the complex neural network runs throughout the human body and mediates bidirectional communication between organs. For instance, the gut is directly connected to the brain and spinal cord via the vagus and spinal nerves, respectively (12). Functionally, the CNS communicates with the gut through the peripheral nerves, e.g., the vagus or spinal nerves, by releasing neurotransmitters (13–17). The gut, in turn, affects the brain by sending chemical signals or pathological proteins through the vagus nerves (18–20). Therefore, peripheral nerves are crucial mediators of the bidirectional communication and chemical signal transfer between the gut and CNS. This unique mode of interaction attracted our curiosity and led us to consider if there may exist neuronal routes responsible for the transport of NPs to the CNS from the gut after oral exposure.

In the present study, we attempted to identify transfer routes for Ag NMs along the gut-CNS axis in addition to blood circulation. We demonstrate, both theoretically and experimentally, the direct translocation of Ag NMs from the gut to the CNS (both the brain and spinal cord) via peripheral nerve fibers. Using laser ablation and field flow fractionation inductively coupled plasma mass spectrometry, we show that Ag NMs administered by oral gavage became enriched in a particle state in the brain and spinal cord of mice while not efficiently entering the blood. Impressively, the vagus and spinal nerves mediate the transneuronal transport of Ag NMs along the gut-brain axis and gut–spinal cord axis, respectively, with specific gut cell subsets (enterocytes and enteric nerve cells) taking up significant levels of Ag NMs for transfer to the connected peripheral nerves. Our study implicates transneuronal transport as a mechanism for the direct trafficking of NPs, filling a crucial gap in our quest of understanding the behavior of NPs in biological systems.

RESULTS AND DISCUSSION

Orally administered Ag NMs enter CNS of mice and rhesus monkeys

We rationalized that investigating the transport of NPs from the gut to CNS may shed light on the pathways underlying gut-CNS communication. To this end, we exposed mice to Ag NPs and Ag nanowires (Ag NWs) through oral gavage for 28 days and examined the distribution of the NPs in various organs (Fig. 1A) [The exposure dose of 10 mg kg⁻¹ was chosen to simulate the long-term exposure in humans (especially the occupational population). Because the size of Ag NMs in the food additives (E174) or pharmaceutical industry ranges from 15 to 500 nm (21–23), the median size Ag NPs (60 nm) and Ag NWs were set as model materials]. The size and morphology of Ag NPs and Ag NWs were characterized and are presented in fig. S1. There was no significant alteration in the general health status of the C57BL/6 mice treated with the Ag NMs (fig. S2). Inductively coupled plasma–mass spectrometry (ICP-MS) analysis revealed that the CNS, particularly the spinal cord, was the major target organ of Ag NPs and Ag NWs through oral route (Fig. 1B and fig. S3A). In contrast, the biodistribution pattern of Ag NMs was totally different in mice by intravenous administration (Fig. 1C). Notably, the proportion of silver in CNS to all tested tissues was higher than 90% via oral administration (fig. S3B), while this proportion was less than 1% in CNS through intravenous administration (fig. S3C), indicating that there are different pathways to mediate the transmission of Ag NMs to CNS between oral and intravenous administration.

Fig. 1. Biodistribution of Ag NMs in the CNS after oral gavage in mice and rhesus monkeys.

(A and B) Experimental design and the distribution of Ag NMs in mice and rhesus monkeys treated with Ag NPs (10 mg kg⁻¹) or Ag NWs (10 mg kg⁻¹) by oral gavage for 28 days. ID/g, ratio of Ag in the total dose per gram of the indicated tissue. The data are expressed as the mean ± SEM. (C) Amount of Ag in various organs of mice treated with Ag NPs or Ag NWs by intravenous administration for 14 days (n = 3). (D) Representative LA-ICP-MS images of Ag NPs and Ag NWs distribution in the horizontal (top) and coronal (bottom) brain sections of mice. CPu, caudate putamen (striatum); CA, hippocampus; LV, lateral ventricle; D3V, dorsal third ventricle; 3V, third ventricle; 4V, fourth ventricle; 3Cb, third cerebellar lobule; MD, mediodorsal thalamic nucleus. (E) Representative images of Ag NP and Ag NW distribution in the lumbar (top) and longitudinal (bottom) sections of spinal cord in mice. (F) Representative images of Ag NPs in the sagittal section of brain in rhesus monkeys. CC, cerebral cortex; ACC, agenesis of the corpus callosum; CE, cerebellum; PGO, ponto geniculo occipital. (G) Representative images of Ag NMs in the lumbar section of spinal cord in rhesus monkeys. The color bar unit of (D) to (G) is parts per million (ppm). (H) Representative SEM and EDX images of Ag NMs isolated from the gut, brain, and spinal cord of mice. Scale bars, 100 nm.

Next, we visualized the in situ distribution of Ag NPs and Ag NWs in the brain and spinal cord by laser ablation inductively coupled plasma–mass spectrometry (LA-ICP-MS), which indicated that Ag NPs and Ag NWs mainly enter the caudate putamen (striatum), hippocampus, ventricles (third and fourth), and third cerebellar lobule, in horizontal and coronal brain sections (Fig. 1D). As for the spinal cord, Ag NPs and Ag NWs accumulated largely in the gray matter of the lumbar (Fig. 1E, top; the iron element is used to indicate the shape of gray matter in spinal cord) and cervical spinal cord (fig. S3D), respectively. Notably, a significant amount of Ag NPs and Ag NWs was localized to the gray matter in whole spinal cord in longitudinal sections (Fig. 1E, bottom), suggesting that Ag NPs and Ag NWs tend to accumulate in the neuronal cell body rather than the axon in the spinal cord in mice. In addition to rodents, we used a primate model, rhesus monkeys, to verify the distribution of silver in the brain and spinal cord after oral gavage (10 mg kg⁻¹ per day) of Ag NPs or Ag NWs for 28 days. (Because of the limited supply of monkeys during the coronavirus disease 2019 (COVID-19) pandemic, we were only able to use one monkey for each group to verify the key results from mice, including the distribution of Ag NMs in brain and spinal cord.) Similarly, we found a large number of Ag NPs and Ag NWs in the sulcus and gyrus of the brain (Fig. 1F), as well as the gray matter of the spinal cord (Fig. 1G), which is composed of neuronal cell bodies. These results indicate at Ag NPs and Ag NWs tend to enter the CNS, especially neuronal soma, after oral exposure in both mice and rhesus monkeys.

Since the distribution of NPs has been suggested to relate to their chemical properties in target organs (24), we subsequently studied the chemical forms of silver in the brain and spinal cord of mice by synchrotron radiation x-ray absorption near-edge structures (XANES). The XANES spectra show similar characteristic spectral shapes for the silver in the brain and spinal cord, as that of native Ag NPs before administration; the form of silver was completely different from Ag₂S, AgNO₃, AgO, Ag₂O, and AgCl species (fig. S4A), indicating that these NPs remained in their original valence state. By using field flow fractionation–ICP-MS, we demonstrate that the Ag NPs and Ag NWs in the brain and spinal cord (homogenized solutions) are of a similar size to the original Ag NPs and Ag NWs in water, which is markedly different from the spectral peaks of silver ions (fig. S4B), suggesting that the Ag NPs and Ag NWs maintain their particle state inside the brain and spinal cord, without releasing detectable silver ions in the course of translocation from the gut to the CNS. Notably, scanning electron microscopy (SEM) analysis exhibited the original particle state of Ag NPs and Ag NWs in the brain and spinal cord (Fig. 1H). The overlapping of silver and carbon element indicated the binding of proteins on the surface of Ag NPs and Ag NWs illustrated by energy dispersive spectroscopy. The chlorine element indicated the generation of Ag-Cl bond that formed passivation layer on the surface of NP, as a consequence of the interaction of Ag NMs with gastric acid. These two mechanisms prevented the degradation of Ag NMs and dissolution of silver ions in gut and CNS, thus enabling Ag NMs to transport from the gut to the CNS with particle state.

The vagus nerves mediate the transneuronal transport of Ag NPs from the gut to the brain

We further investigated how NPs transport from the gut to the CNS. A clear picture of the routes underlying the gut-brain axis transmission of NPs has not been described to date. One suggested pathway is the transport of NPs to the brain from the blood circulation via the blood-brain barrier (BBB; only allows the essential molecules, such as oxygen, carbon dioxide, and lipophilic molecules smaller than 400 Da, to passively diffuse across the brain endothelium, but blocks the transfer of most particles from the blood to the CNS) (25–27). However, treatment with Ag NPs and Ag NWs by oral gavage did not influence the integrity of BBB in mice, as determined by the classical Evans blue (EB) staining assay (Fig. 2A), suggesting that Ag NPs and Ag NWs likely do not enter the brain through the BBB. Furthermore, the concentration of Ag NPs and Ag NWs was maintained at a background level in the blood of mice after either a single dose or multiple administrations by oral gavage at different times (days 0 to 28; Fig. 2B), suggesting that the NPs do not efficiently enter the blood circulation from the gut. This supposition is supported by a recent finding that the gut-vascular barrier strictly limits the entrance of substances ≥70 kDa into the bloodstream (28). In contrast to the effects of Ag NMs in blood, we noted that there were cumulative effects of Ag NPs and Ag NWs in the brain (Fig. 2C), implying that there must exist other pathways responsible for the NP transport to the brain.

Fig. 2. Ag NMs transport from the gut to the brain through the vagus nerves.

(A) Classical EB staining showing the integrity of the BBB of mice treated with Ag NM by oral gavage for 28 (n = 3). Ag content in the blood (B) and brain (C) of mice treated with Ag NPs or Ag NWs by oral gavage (10 mg kg⁻¹) with single/multiple administrations. Single dose: The mice were treated with Ag NPs or Ag NWs once. Daily dose: The mice were treated daily with Ag NPs and Ag NWs (n = 3). (D) Schematic illustration of the truncal vagotomy in mice. (E) Relative concentration of Ag NPs in the brain of mice treated with truncal vagotomy (n = 4). (F) Representative light sheet microscope images showing the Ag NP transmission inside the vagus nerve (green, Ag NPs). (G) Representative confocal microscopy images showing TRPV1 (green) in the vagus nerve of mice after subcutaneous injection of RTX for 3 days (1-day, 30 μg kg⁻¹; 2-day, 70 μg kg⁻¹; 3-day, 100 μg kg⁻¹). DAPI, 4′,6-diamidino-2-phenylindole. (H) Brain Ag NP content of the sham and RTX groups (n = 4). (I) Ag NP and Ag NW content in the vagus nerves of rhesus monkeys. (J) Representative transmission electron microscopy (TEM) images of Ag NPs and Ag NWs in the vagus nerves of rhesus monkeys. The data are expressed as the mean ± SEM. Statistical significance was tested using a two-tailed t test and one-way analysis of variance (ANOVA) analysis. ns, no statistical significance; ***P < 0.001; ****P < 0.0001.

We next focused on clarifying the alternative pathway mediating Ag NP and Ag NW transport from the gut to the brain. The vagus nerves originate from the medulla oblongata of the brain and are directly connected to the gastrointestinal tract along with the esophagus (Fig. 2D). Multiple lines of evidence indicate that immune, metabolic, and neural communication between the gut and brain take place along the gut-brain axis through the vagus nerves (14, 29–32). Therefore, we speculated that the vagus nerves may be a pathway for Ag NM transport from the gut to the brain. To test this experimentally, we performed a truncal vagotomy by cutting both left and right vagus nerve fibers proximal to the cardia (Fig. 2D and fig. S5), the mice were then administered Ag NPs (10 mg kg⁻¹ per day) by oral gavage for 28 days. As shown in Fig. 2E, there was a 75% reduction in the content of silver in the brain of vagotomy-treated mice compared to the sham group with intact vagus nerves, indicating that truncal vagotomy markedly prevented gut-to-brain transmission of Ag NPs and the vagus nerves are direct and central pathways mediating NP transmission from the gut to the brain.

To verify the transport of Ag NPs via the vagus nerves, we conducted tissue clearing and light sheet microscopy to visualize the transmission route of Ag NPs via the vagus nerves by labeling the NPs with the fluorescent molecule, dansyl-cys-NH₂ (Fluorescence labeled Ag NPs: Flu-Ag NPs). We injected the Flu-Ag NPs (1 μg) into proximal stomach vagus nerve explants by microinjection and then cultured the explants in Matrigel to track the migration of NPs. Calcein-Acetoxymethyl Ester/Propidium Iodide (AM/PI) staining analysis indicated good explant viability after incubation for 3 days (fig. S6A). The Ag NP clearly spread from the proximal stomach to the vagal ganglion (the proximal brain; the enlargement in the explants represents the vagal ganglion) after 3 days (Fig. 2F), revealing that the vagus nerves are capable of engulfing Ag NPs and transporting them to the brain along the nerve fiber. In support of this, the transport process of Ag NPs was also visualized through fluorescence micro-optical sectioning tomography (fig. S6B). Moreover, direct exposure of the vagus nerve explants to Flu-Ag NPs led to endocytosis of Ag NPs by the neuron and transport along the vagus nerve (fig. S6C). These findings implicate the vagus nerve as direct conduit to mediate the translocation of NPs from the gut to the brain. We propose to designate this previously uncharacterized transmission pathway of NPs as “transneuronal transport.”

To further elucidate the underlying mechanism for the vagus nerve–brain transmission of Ag NPs, we analyzed the highly expressed receptors on the nerves. Notably, gut-innervating nociceptors sense and regulate a variety of behavior in the gut and brain, including intestinal pain, diarrhea, and substance intake (33–35). Among these receptors, TRPV1, a nociceptive ion channel that detects heat and capsaicin, plays a central role in substance endocytosis and transmission (36). We speculated that this receptor may affect the transmission of NPs along the gut–vagus nerve–brain axis. To test this hypothesis, we subcutaneously injected resiniferatoxin (RTX), a potent TRPV1 agonist, for 3 days to denervate TRPV1⁺ neurons. RTX significantly reduced the expression of TRPV1 in the vagus nerves (tested after 3 weeks) and improved the pain threshold of mice, as determined by hot plate experiments (Fig. 2G and fig. S6D), indicating the successful establishment of the mouse model. After oral gavage of Ag NPs for 28 days, there was a sharp decrease in Ag NPs in the RTX-treated mice compared with sham controls (subcutaneous injection of saline; Fig. 2H), indicating that the TRPV1⁺ neurons were required for the transport of Ag NPs along vagus nerves from gut to brain, thus suggesting a mechanism distinct from the currently known transfer pathways.

The experiments described above establish a connection between the gut and brain by the vagus nerves in mice; we proceeded to explore whether vagus nerve–mediated transmission of Ag NPs is also common in primates. After oral gavage of Ag NPs or Ag NWs for 28 days, we found Ag NPs and Ag NWs in the vagus nerves of rhesus monkeys, as assessed by ICP-MS (Fig. 2I), transmission electron microscopy (TEM) (Fig. 2J), and TEM dark field imaging (fig. S7A). The content of Ag NPs and Ag NWs in the vagus nerves were ~1000-fold greater than that in the viscus (fig. S7B), again suggesting that the vagus nerves tend to take up large amounts of Ag NMs and are an important pathway for the transmission of Ag NMs in both rodents and primates. Furthermore, we also tested other type of NP, Au NPs (with a similar 60 nm size as Ag NPs), to verify the role of vagus nerves in gut-brain transmission in mice. Similarly, 82% less Au was present in the brains of vagotomy-treated mice than sham controls (fig. S7C), indicating that gold and silver NPs both can exploit the same vagal pathway along the gut-brain axis. Notably, the silver content was nearly 30-fold greater than Au in the brain of mice when using the same gavage dose (fig. S7D), indicating that neuron-enriched tissue may have higher affinity for silver than gold. In contrast to Ag NMs and Au NPs, TiO₂ (60 nm) and CeO₂ (60 nm) can hardly transport to the blood and brain by oral administration for 28 days at the same exposure dose (fig. S7, E and F). Apart from the element type, we further investigated the influence of size and surface modification on the transport behaviors of Ag NPs in mice. Compared to larger Ag NPs (200 nm), the smaller ones (20 and 60 nm) were more likely to transport to the brain of mice (fig. S8A), while the polyethyleneimine (PEI)–coated Ag NPs (60 nm) were more easily transport to the brain than the polyethylene glycol (PEG)–coated Ag NPs (fig. S8B), which may be because of the strong binding between positive surface charge of PEI and neuron cells. These results indicate that the intrinsic properties, including element type, size, and surface modification, can affect the transport and distribution of Ag NMs in the mice.

The spinal nerves mediate the transneuronal transport of Ag NPs from the gut to the spinal cord

Our results demonstrate the existence of vagus nerve–mediated translocation of Ag NPs from the gut to the brain. We proceeded to examine whether the NPs are also transported from the gut to the spinal cord, a region of the CNS arising from the medulla oblongata and communicating with other viscera, e.g., the gut (37). As shown in Fig. 2B, we observed a negligible amount of Ag NMs in the blood either after a single dose or multiple administrations by oral gavage, implying that the Ag NMs did not efficiently enter the blood circulation from the gut. In contrast, Ag NPs and Ag NWs accumulated in the spinal cord (Fig. 3A), suggesting again that alternate routes exist for the gut–to–spinal cord transmission of NPs. When Ag NMs were administered by intravenous injection, Ag NPs and Ag NWs were significantly enriched in the liver, spleen, and lung rather than the spinal cord (Fig. 1C and fig. S9), showing a totally different distribution pattern of Ag NMs provided to mice by oral gavage (Fig. 1B). By comparing the efficiency of Ag NMs entering spinal cord, it is clear that the Ag NMs given by oral gavage tend to enter the spinal cord at a higher rate than when given by intravenous administration (Fig. 3B). These findings again suggest that other as-yet unidentified pathways directly translocate NPs from the gut to the spinal cord.

Fig. 3. Ag NMs transport from the gut to the spinal cord via spinal nerves.

(A) Content of Ag in the spinal cord of mice treated with Ag NPs or Ag NWs by oral gavage (10 mg kg⁻¹) with a single dose or multiple doses. Single dose: The mice were treated with Ag NPs or Ag NWs once. Daily dose: The mice were treated daily with Ag NPs and Ag NWs (n = 3). (B) The efficiency of Ag NP entry into the spinal cord of mice through oral gavage and intravenous administration (n = 3). (C) Content of Ag in the spinal nerves and DRG after oral gavage for 28 days (n = 5). (D) Schematic illustration of the removal of spinal nerves by selective posterior rhizotomy (SPR) and the transfer of Ag NPs from the gut to the spinal cord via the spinal nerve. (E) Relative concentration of Ag NPs in the spinal cord of mice treated with SPR (n = 3). (F) Representative Raman image of Ag NPs (exposure dose, 5 μg ml⁻¹) showing the NP transmission inside dorsal root ganglion cells at different time points. Scale bars, 4 μm. (G) Ag NM content in the cervical, thoracic, and lumbar spinal nerve of rhesus monkeys. (H) Ratio of silver in the spinal nerves and viscera of rhesus monkeys. The data are expressed as the mean ± SEM. Statistical significance was tested using a two-tailed t test and one-way ANOVA analysis. ***P < 0.001; ****P < 0.0001.

Similar to the vagus nerves, spinal nerves directly link the gut wall to the spinal cord in mammalian anatomy (38). Therefore, we speculated that the spinal nerves can serve as a pathway to mediate the gut–to–spinal cord transmission of Ag NPs. We first evaluated the content of Ag NPs and Ag NWs in continuous sections between the gut and spinal cord, including spinal nerves and dorsal root ganglia (DRG) by ICP-MS. A relatively high levels of Ag NMs were present in the DRG and spinal nerves (Fig. 3C); this accumulation was 22- to 800-fold greater than that in the other viscera (heart, liver, spleen, lung, and kidney; fig. S10A), suggesting that Ag NPs preferentially enter the spinal cord, via the spinal nerves, over the major organs. To verify this pathway, we surgically removed the spinal nerves (L3 to L5) in mice via selective posterior rhizotomy (SPR), as shown in Fig. 3D (see fig. S10B for the surgical process), followed by oral gavage with Ag NPs (10 mg kg⁻¹ per day) for 28 days. Since the posterior root of the spinal nerve is an afferent neuron that controls visceral sensation (12), SPR does not influence the motor function of mice. After the spinal nerves were lesioned, there was a 90% decrease in Ag NPs in the spinal cord of SPR mice compared to the sham group (Fig. 3E), thereby indicating that the spinal nerves are key pathways mediating the NP transfer from the gut to the spinal cord.

As we found a large number of Ag NPs and Ag NWs in the DRG located in the intervertebral foramen along the spine, functionally, the DRG serves as the core hub to return sensory information from the spinal nerves to spinal cord (39). We next examined how the Ag NP trafficked inside DRG cells using a multimodal Raman microscope. Since the Raman signal of Ag NPs cannot be detected within cells, we generated rhodamine B–labeled Ag NPs to visualize the transport of NPs. The labeled Ag NPs were rapidly endocytosed by the cell body of the DRG in the early stages after treatment (0 to 6 hours; Fig. 3F). After endocytosis, the Ag NPs in the cell body gradually moved to the axon in the later stages of the experiment (12 to 24 hours). The NPs exhibited an identical transfer direction as the chemical signals (neurotransmitters) inside neurons (40, 41). We also treated DRG cells with sodium thiosulfate and potassium ferricyanide to rule out any interference of Ag NPs adhered to the cell membranes, as observed previously (2). Hence, the dynamic transport of Ag NPs we found indeed occurred inside DRG cells rather than on the membrane. With the large amount of Ag NPs found in the spinal cord (Fig. 1B), we propose a previously unknown transneuronal route of a spinal nerve–DRG–spinal cord axis for Ag NP transfer from the gut to the spinal cord. To verify the generalizability of this pathway, we examined the spinal nerves of rhesus monkeys treated with Ag NMs (10 mg kg⁻¹ per day) by oral gavage for 28 days. ICP-MS analysis confirmed the presence of Ag NPs and Ag NWs in the spinal nerves at different regions in rhesus monkeys, including cervical spinal nerves, thoracic spinal nerves, and lumbar spinal nerves (Fig. 3G), at higher levels than in viscera, such as the liver, lung, and spleen (Fig. 3H), which was consistent with the results in mice (fig. S10A). These results further strengthen our finding that the spinal nerve–mediated transneuronal transport of Ag NPs from the gut to spinal cord is a pathway common to rodents and primates.

After Ag NPs enter the spinal cord, the NPs may transport to the brain. We further explored the bidirectional transmission of Ag NPs between the spinal cord and the brain by injecting Ag NPs into the L5 segment of the spinal cord and the third ventricle of the brain in mice. After three injections (on days 1, 4, and 6, and sacrifice on day 8), only ~3% Ag NPs injected in the spinal cord were transported to the brain (fig. S11A). By contrast, ~28% Ag NPs in the third ventricle were transported to the spinal cord (fig. S11B). This inefficient transport of the NPs to the brain may explain the large number of Ag NMs retained in the spinal cord.

Ag NPs are transferred from the gut wall to neurons

So far, we have shown that the transneuronal transfer of Ag NPs from the gut to the CNS is mediated by previously uncharacterized pathways: vagus and spinal nerves. We next aimed to elucidate how these NPs are transported from the gut wall to neurons, the first step of NP transmission from the gut to the CNS. Our previous studies have shown that, once NPs enter biological fluids, various proteins rapidly bind to the surface of the NPs to form a “protein corona,” thus influencing their biological fates in the body (42–45). To simulate the uptake process of NP corona in the gastrointestinal tract, we extracted the contents of the stomach and gut lumen (the food residue, cell debris, and bacteria were removed by gradient centrifugation) and incubated them with Ag NP for 1 hour, after which Ag NP–corona complexes were obtained by centrifugation. Treatment of gut epithelial cells (Caco-2 cell line) with Ag NPs with or without a protein corona revealed a greater accumulation of the Ag NP–corona complex, as determined by ICP-MS analysis at 6 hours (Fig. 4A), indicating that biological molecules in the stomach and gut significantly promote the uptake of the NPs by epithelial cells. Because of the important role of the protein corona on cellular uptake (46, 47), we assessed the composition of the proteins adsorbed onto Ag NP by liquid chromatography–tandem mass spectrometry (LC-MS/MS; peptides were used to match their backbone fragments for protein identification). The 20 most abundant coronal proteins are presented in Fig. 4B. Most of the adsorbed proteins are involved in cytoskeleton synthesis, cell movement, and energy metabolism, all of which are essential processes in cellular uptake. Among these proteins, heat shock cognate 71-kDa protein (Hsc70) has been reported to be related to the vesicle-mediated transport (48, 49). We therefore examined the influence of Hsc70 on the cellular uptake of Ag NPs by incubating pure Hsc70 protein with Ag NPs (2:1) for 1 hour and then treating Caco-2 cells with the coated particles. The pure Hsc70 protein corona markedly enhanced the uptake of Ag NPs at 2 and 6 hours (Fig. 4C), thus indicating the crucial role of the protein corona on Ag NP uptake by gut epithelial cells.

Fig. 4. Ag NPs are transferred from the gut wall to neurons.

(A) Silver content in Caco-2 cells after exposure to Ag NPs or Ag NP–corona complexes at 5 μg ml⁻¹ at the indicated time points (four replicates per condition). The Ag NPs in cells were detected by ICP-MS. (B) The 20 most abundant proteins on the surface of Ag NPs, as identified by LC-MS/MS. GDP, guanosine diphosphate; NADP, nicotinamide adenine dinucleotide phosphate; PKM, pyruvate kinase M. (C) Silver content in Caco-2 cells after exposure to Ag NPs or complexes of Ag NPs and Hsc70 for 2 and 6 hours, as determined by ICP-MS (four replicates). (D) Representative confocal microscopy images of the transfer of Ag NPs from gut epithelial cells (Caco-2) to neurons (DRG). Scale bars, 4 μm. Single-cell mass cytometry showing (E) the proportion of total gut cells taking up Ag NPs and (F to H) the uptake of Ag NPs by different gut cell subsets. tSNE, t-distributed stochastic neighborhood embedding. (I) Schematic illustration of the cellular uptake and transfer process of Ag NPs from gut epithelial cells to peripheral neurons. The data are expressed as the mean ± SEM. Statistical significance was tested using a two-tailed t test and one-way ANOVA analysis. *P < 0.1; **P < 0.01; ***P < 0.001.

We next examined the behavior of NP transfer from gut epithelial cells (Caco-2) to primary peripheral neurons (DRG cells) using a coculture system (Fig. 4D; the two separate wells were seeded with Caco-2 or DRG cells. The Caco-2 cells were exposed to fluorescence-labeled Ag NPs (Flu-Ag NPs) at 5 μg ml⁻¹ for 24 hours, after which the culture medium was removed and the Ag NPs absorbed onto the cell membrane were visualized by sodium thiosulfate and potassium ferricyanide as described previously (2). After removing the insert and incubating the cells for an additional 24 hours, we observed the presence of Flu-Ag NPs in the distal DRG cells (Fig. 4D, right), indicating that the Ag NPs are capable of translocating from Caco-2 cells to DRG cells, thereby providing additional support for the presence of a transfer link between gut epithelial cells and peripheral neurons. We further compared the Ag NP uptake capacity of DRG and Caco-2 cells by quantitative analysis, which revealed that the neurons took up a greater amount of Ag NPs than the gut epithelial cells (fig. S12A). Compared with the other four types of NPs tested (Au, Pd, Cu, and TiO₂ NPs), which are all a similar size, neurons took up significantly more Ag NPs at the same exposure dose (5 μg ml⁻¹) in 24 hours (fig. S12B), thus revealing the specificity and selectivity of neurons for Ag NPs. A plausible explanation for this phenomenon is the argyrophilic property of nerve cells (the specific mechanism of which remains to be determined) (50), which is also supported by classical observations of nerve cells/fibers in the CNS labeled by silver staining methods (51, 52). In particular, the highly expressed heavy metal sulfides on the neurons might contribute to this effect (53).

We further performed single-cell mass cytometry (Cy-TOF) to discern the uptake behavior of Ag NPs by different cell subsets in the gut. To accomplish this, we stained, labeled, and sorted cells using metal isotope–tagged monoclonal antibodies. The Cy-TOF analysis showed that 2.45% of the cells in the gut are capable of taking up the Ag NPs (Fig. 4E). We then used graph clustering to partition these cells and mapped the cell subset containing Ag NPs by t-distributed stochastic neighborhood embedding (Fig. 4F). Among the different cell types, the immune cell population and the gut epithelial cell population, including leukocytes, neutrophils, monocytes, goblet cells, and enteroendocrine cells, exhibited weak Ag NP uptake (Fig. 4G). In contrast, enteric nerve cells in submucosal/myenteric plexus and enteric cells were the major two cell populations that accumulated Ag NPs, with enteric nerve cells taking up more Ag NPs than enteric cells (Fig. 4H). Thus, enteric nerve cells can receive the Ag NPs via transcellular transfer from enterocytes to the connected peripheral nerves (Fig. 4I).

After Ag NMs enter the CNS from the gut through peripheral nerves, they might induce adverse outcomes of CNS, we next evaluated the effects of Ag NMs on the CNS. The increase of glial fibrillary acidic protein (marker of astrocyte) indicated that Ag NMs induced the neuronal injury of brain (fig. S13A). The reduced secretion of nerve growth factors suggested that Ag NMs may inhibit the neuronal growth or development in substantia nigra (fig. S13B). Moreover, Ag NMs up-regulated the expression of acetylcholinesterase in cerebral cortex (fig. S13C), which play an important role in regulating the learning and memory through influencing the content of acetylcholine. We therefore assessed the behaviors of mouse using open-field test and elevated plus maze. Compared with control, Ag NMs decreased the number of entries to the central area and the average speed of mice (fig. S13D), implying that Ag NM–treated mice were in anxiety state and their motor function was impaired, which was supported by the reduction of OT% (percentage of staying in open arms) (fig. S13E). These combined data indicate that Ag NMs induced neurotoxicity and influenced the function of CNS after transporting to the CNS from gut.

In summary, the key findings of this study propose and verify a previously unknown route of silver NP translocation along the gut-brain axis and gut–spinal cord axis consisting of the transneuronal transport of NPs from the gut to the CNS via the vagus and spinal nerves (Fig. 5). Our results challenge the current wisdom mainly in two aspects: (i) The peripheral nerves (vagus and spinal nerves) are thought to serve only a route for chemical signal (e.g., neurotransmitters) transmission for the communications between the gut and CNS (16, 17); (ii) NPs use the blood circulation as the only way from the gut to CNS (11, 54), following oral administration. However, the gut-vascular barrier strictly restricts the entrance of substances ≥70 kD into the blood stream (28), thereby preventing Ag NPs in the gut from efficiently entering the blood (Fig. 2B). Through anatomic structure, the vagus and spinal nerves directly link the gut wall to the CNS (38), which provide direct pathways for NP transport along the gut-CNS axis confirmed by our data (Figs. 2D and 3D). Because of the argyrophilic property of neurons, Ag NPs preferentially enter nerve cells compared over other cell types (Fig. 4G and fig. S12). Our observation of efficient NP entry into the vagus and spinal nerves of both mice and the rhesus monkey (Figs. 2, I and J, and 3, G and H) indicates that peripheral nerve mediated transneuronal translocation of Ag NPs from the gut to the CNS, which is never been reported before and could be an alternative common pathway. These results definitely deepen our understanding of the communication between NPs in the gut and both the peripheral nervous system and the CNS. Building on these findings, new strategies for the therapy of CNS-associated diseases using transneuronal transport approach are expected to emerge, providing alternative delivery routes for nanocarriers/nanodrugs that do not efficiently cross the BBB or blood–spinal cord barrier.

Fig. 5. Working model of the transneuronal transport of NPs from the gut to the CNS.

When Ag NPs enter the gastrointestinal tract by oral gavage, they adsorb proteins to form a nano-corona that facilitates the uptake of NPs by various cell subsets of the gut, especially enterocytes and enteric nerve cells in submucosal/myenteric plexus. Thereafter, the vagus nerves act as direct conduits to mediate the transneuronal transport of Ag NPs along the gut-brain axis. Meanwhile, the spinal nerves mediate the transneuronal transport of Ag NPs along the gut–spinal cord axis.

MATERIALS AND METHODS

Materials

Ag NPs (100422, CAS 7440-22-4), Ag NWs (102884, CAS 7440-22-4), and Pd NPs (15 to 25 nm; 103010, CAS 7440-05-3) were purchased from XFNANO Materials Tech Co. Ltd. Au NPs (60 nm, Au010060) were purchased from Beijing Zhongkeleiming Daojin Technology Co. Ltd. TiO₂ NPs (60 nm) were purchased from Aladdin, and Ni NPs (69 nm) and Cu NPs (78 nm) were obtained from Nanjing Jicang Nanotechnology Co. Ltd (JCNO), China. Ag NPs (20 and 200 nm) were acquired from Zhongkekeyou Co. Ltd. Dopamine (CAS 2-31-7/D80661), PEI (CAS 002-98-6), and PEG (CAS 32130-27-1) were obtained from Macklin Inc. PEI- and PEG-modified Ag NPs were prepared as following procedure: 100 mg Ag NPs were dispersed in 100 ml of tris-HCl buffer (pH 8.5), followed by the addition of 25 mg of dopamine and incubation for 2 hours (Ultrasonication in water bath). After removing the excess dopamine by centrifugation, Ag NPs was redispersed in 100 ml of tris-HCl buffer. Next, 5 mg of PEI or PEG was added and incubated for 48 hours with stirring. Last, PEI- or PEG-modified Ag NPs were obtained through centrifugation.

Animal experiments

Male C57BL/6 mice (7 weeks old, ~20 g) were purchased from Vital River Co (Beijing, China). All mouse experiments were performed in accordance with the Institutional Animal Care and Use Committee of National Center for Nanoscience and Technology. The mice were bred in the cage in the specific pathogen–free (SPF) animal facility in a temperature-controlled (22°C) environment under a strict 12-hour light cycle and fed a standard, autoclaved chow diet and water ad libitum at 50 to 60% relative humidity, with the cages changed every 2 days. Before experiments, the mice were kept in an SPF animal facility for a 7-day acclimation period. The experimental design was conducted as follows: the animals were randomly divided into three groups: control (water), Ag NPs (10 mg kg⁻¹ per day), and Ag NWs (10 mg kg⁻¹ per day). Before oral gavage, the Ag NPs and Ag NWs suspensions were sonicated for 20 min. After oral gavage for 28 days, the mice were sacrificed for the various assays.

Male rhesus monkeys (3 years old, 4.0 kg) were purchased from the Beijing Sharing Institute of Biological Resources Co. Ltd. All monkey experiments were performed in accordance with the Institutional Animal Care and Use Committee of Beijing Sharing Institute of Biological Resources Co. Ltd. The monkeys were examined for any viral, bacterial, and parasite infection, and the general health of the animals was monitored during the experiment. The experimental design was conducted as follows: The monkeys were randomly divided into three groups: control (water), Ag NPs (10 mg kg⁻¹ per day), and Ag NWs (10 mg kg⁻¹ per day). After oral exposure through drinking water for 28 days, the monkeys were sacrificed for the various assays. Notably, monkey availability was severely limited. Therefore, the monkeys were only used to verify the key results from mouse experiments, including the distribution of Ag NMs in the brain and spinal cord, and the presence of Ag NMs in the vagus and spinal nerves. Only one monkey was used for each condition.

Preparation and characterization of Ag NPs and Ag NWs

Ag NWs were sonicated in sterilized deionized water for 2 hours at a dedicated concentration using an ultrasonic cell breaker (SCIENTZ-II D). The average diameter of Ag NPs was 60 to 150 nm, and the length of Ag NWs was 0.8 to 1.2 μm, as determined by TEM (TecnaiG2 20 S-TWIN; fig. S1, A and B). The hydrodynamic size of Ag NPs and Ag NWs was 159 and 277 nm, respectively, as detected by dynamic light scattering (fig. S1C). The zeta potentials of Ag NPs and Ag NWs showed a slightly anionic surface charge. The hydrodynamic sizes of Ag NPs and Ag NWs in simulated gastric juice and simulated intestinal fluid were also determined (fig. S1, D and E).

Inductively coupled plasma–mass spectrometry

After oral gavage for 28 days, the mice were sacrificed and the tissues were removed. The silver content of tested samples was measured by ICP-MS (Thermo Elemental X7). In detail, the blood of the mice was flushed out by cardiac perfusion with phosphate-buffered saline (PBS), and the different tissues were excised for ICP-MS detection. When measuring the silver content in the spinal nerves and DRG, the nervous tissues of three mice were combined into one group to obtain an accurate weight, since the tissue from a single mouse has very little mass. In the rhesus monkey measurements, 100 mg of each tissue was digested and analyzed by ICP-MS. Briefly, the tissues were transferred to conical flasks and digested in HNO₃ (Guaranteed Reagent) overnight. Next, they were heated to 180°C and held for 3 hours with the addition H₂O₂ until the samples were completely digested. Next, the solutions were heated up to 230°C to remove any residual HNO₃. Last, the remaining, colorless solutions were diluted to 3 ml with 2% HNO₃ and analyzed by ICP-MS.

In situ imaging of Ag in the brain and spinal cord

LA-ICP MS was used for in situ element mapping analysis of tissue sections from mice and rhesus monkeys. The LA system (PhotonMachines Analyte HE with 193-nm ArF excimer) combined with ICP-MS (Agilent 7900) was used in an atmosphere of UHP He (0.9 liter min⁻¹) mixed with Ar (0.87 liter min⁻¹), and the spot energy was 2.76 J/cm² with 20 Hz. NIST 610 was used as an external standard. Briefly, the brain and spinal cord were immersed in Optimal Cutting Temperature (OCT) compound and frozen at −20°C overnight. The tissues were then cut into 60-μm sections using a freezing microtome (Leica, CM1950). Last, LA-ICP-MS was used to analyze the distribution of silver (iron was used to visualize the shape of the gray matter in the spinal cord).

Evaluation of the integrity of the BBB by EB

The EB dye assay was performed as follows: After oral gavage, C57BL/6 mice were injected with EB dye (2%, 100 μl) through the caudal vein. After circulating for 1 hour, the mice were anesthetized and perfused transcardially with cold PBS. Subsequently, the brain was removed and the cerebellum and brain stem were discarded. The forebrains were divided into the ipsilateral and contralateral hemispheres. After weighing the tissues, the forebrains were homogenized in 500 μl of PBS, followed by the addition of trichloroacetic acid (500 μl), and centrifugation to collect the soluble fraction (supernatants). The supernatants were then diluted 4 in four volumes of ethanol and the fluorescence was measured (620 nm excitation, 680 nm emission). Last, the concentration of EB in the brain was calculated on the basis of a standard curve (0 to 500 ng ml⁻¹).

Truncal vagotomy

Male mice (7 weeks old) were anesthetized using isoflurane and held at a constant body temperature. After removing the hair from the abdomen, a midline laparotomy was performed to expose the stomach and esophagus (fig. S5). The bilateral vagus nerves were identified above ~2 to 3 mm of the cardia (the point where the esophagus and stomach connect to the diaphragm). The two truncal vagus nerves were separated and resected from the surrounding arteries and connective tissue under a stereomicroscope. Next, the peritoneum and skin were sutured separately, and the surgical site was dressed with antibiotics. After surgery, the mice were kept in isolation to prevent them from fighting and to monitor their health. Normally, the mice show a loss of appetite and weight in the first few days after vagotomy, but this behavior subsides with additional recuperation time. If necessary, 5% glucose was injected through the tail vein to help restore physical strength. After vagotomy (or sham operation), the mice were treated with Ag NPs (10 mg kg⁻¹ per day) for 28 days by oral gavage. The Ag content in the brain was then measured by ICP-MS.

Fluorescence labeling of Ag NPs

Ag NPs were dispersed in dimethyl sulfoxide (DMSO), assisted by ultrasonication, at a concentration of 1 mg ml⁻¹. The fluorescent molecules were synthesized by dansulfonic acid and cysteine (see fig. S6E for the structure of dansyl-cys-NH₂) and were then dissolved in DMSO at 1 mg ml⁻¹. Thereafter, the solutions of Ag NPs and dansyl cysteamine were mixed, and the pH was adjusted to 8.0 by NaOH in the dark. After 36 hours, the product was collected by centrifugation and washed three times in deionized water. Last, the as-prepared fluorescence-labeled Ag NPs were dispersed in deionized water. The excitation wavelength used to visualize the labeled Ag NPs was 488 nm.

Tissue clearing and light sheet microscope test

The vagus nerves with vagal ganglia were dissected from mice. Afterward, the fluorescence-labeled Ag NPs (1 μg) were immediately injected into the site of the proximal stomach by microinjection [The vagus nerves were immersed in Hanks’ balanced salt solution (HBSS)]. Next, the tissues were planted in a 1:1 mixture of Matrigel (Corning) and Neurobasal Plus medium (Gibco). After 3 days, the explants were immersed into 4% paraformaldehyde at room temperature. Tissue clearing reagents (CUBIC for animals) were purchased from Tokyo Chemical Industry (TCI), USA. The process of tissue clearing was performed according to the handbook. After tissue clearing, the images of Ag NPs in explants were captured by a light sheet microscope (LiTone XL) at the excitation wavelength of 488 nm.

Highly SPR

Highly SPR was performed according to a previous study (55). First, male mice (7 weeks old) were anesthetized with isoflurane, and the backs of the mice were shaved. After opening the skin, the vertebras in the L3 to L5 spinal cord were carefully exposed using micro tweezers and forceps (fig. S10B). The breath and heartbeat were monitored during the operation. Thereafter, the posterior roots of the L3 to L5 spinal nerves were separated using a glass dissecting needle and excised by microscopic scissors. Subsequently, the wound and skin were filled with medical Vaseline and sutured. Antibiotics was applied at the surgical site once a day. After complete recovery, the mice were treated with Ag NPs (10 mg kg⁻¹ per day) for 28 days by oral gavage. The Ag content in the spinal cord was then measured by ICP-MS. It should be noted that the SPR did not affect the movement ability of the mice.

DRG preparation and primary culture

DRG were dissected and dissociated as follows: Newborn mice were euthanized by CO₂ inhalation and immersed in 75% ethanol before dissection. Next, the spine was carefully collected, and the surrounding spinal muscles were removed. Both sides of the spinal processes were then cut with ophthalmic scissors. The spinal cord was exposed and clamped with microscopic tweezers, which contains bilateral DRG on both sides. The DRG were collected and placed into precooled F12/Dulbecco's modified Eagle's medium (DMEM) culture medium. Last, the DRG were transferred to 0.25% trypsin (Gibco) and incubated in water bath at 37°C for 15 to 30 min. Subsequently, the DRG were transferred to F12/DMEM medium, supplemented with 10% fetal bovine serum (FBS), and mechanically dissociated into single cells using Pasteur pipettes. The as-prepared DRG single cells were centrifuged to remove the supernatant and resuspended in Neurobasal Plus medium (Gibco) supplemented with B-27 (Gibco). Primary DRG cells (1.5 × 10⁴ cells/cm²) were seeded in culture dishes (half of the medium was replaced every 2 days).

Rhodamine B labeling of Ag NPs and Raman imaging

Ag NPs (1 mg ml⁻¹) were mixed with an aqueous rhodamine B solution (1 mg ml⁻¹) at a volume ratio of 50:1. The mixture was then stirred at room temperature for 30 min. Thereafter, free rhodamine B was removed by centrifugation. After washing three times with deionized water, the Ag NPs were redispersed in deionized water. After DRG cells (1.5 × 10⁵ cells per dish) were exposed to the Ag NPs (5 μg ml⁻¹), the Raman signal of the Ag NPs at different time points was obtained by Raman microscopy (LabRAM Soleil, Horiba).

Cell line

Human intestinal epithelial cells (Caco-2, American Type Culture Collection, USA) were cultured in minimum essential medium (WISENT, China) supplemented with 20% FBS (WISENT, China) and penicillin (100 U ml⁻¹) and streptomycin solution (100 μg ml⁻¹; WISENT, China).

Transport of Ag NPs from gut epithelial cells to neurons

A two-well culture insert was placed on 35-mm dishes with a glass bottom. The two wells were separately seeded with DRG or Caco-2 cells. After the Caco-2 cells were exposed to fluorescence-labeled Ag NPs (Flu-Ag NPs; 5 μg ml⁻¹) for 24 hours, the culture medium in the Caco-2 wells was removed, and the cells were washed with PBS containing sodium thiosulfate and potassium ferricyanide to etch the Ag NPs adsorbed onto the cell membranes. Last, the culture insert was removed, and the medium was replaced with fresh medium for additional culture for 24 hours. The fluorescence signal of Ag NPs was then detected by confocal microscopy.

Preparation of gut cell suspensions for mass cytometry

After oral administration of Ag NPs for 28 days, the small intestine of C57BL/6 mice was isolated and rinsed with cold HBSS supplemented with 2% FBS. The tissue was opened longitudinally and cut into small fragments approximately 2 mm in size. The tissue fragments were then transferred to 50-ml centrifuge tubes containing 20-ml predigestion buffer (5 mM EDTA, HBSS, and 5% FBS) and incubated at 37°C with gentle shaking for 20 min. Next, the samples were strained through a 100-μm strainer. The undigested tissues were collected, and the predigestion step was repeated. After the second straining, the cells were washed with HBSS to remove the remaining EDTA. After predigestion, the samples were incubated with prewarmed digestion medium [DMEM, WISENT; penicillin (100 U ml⁻¹) and streptomycin (100 μg ml⁻¹), Gibco; non-essential amino acid solution, Solarbio; 1 mM sodium pyruvate, Thermo Fisher Scientific] supplemented with 0.25% collagenase IV (Gibco) and deoxyribonuclease I (7.5 mg ml⁻¹; Sigma-Aldrich) at 37°C in a water bath for 30 min with shaking for 10 min. Last, the samples were placed on ice, dispersed by pulsing with air, using a 3-ml pipette, and strained through a 40-μm strainer. The cells were collected by centrifugation at 300g for 10 min. Last, the cell pellets were dispersed in DMEM medium supplemented with 5% FBS.

Mass cytometry barcoding, staining, acquisition, and clustering

Gut cell suspensions were washed twice with cell staining buffer (CSB; 201068, Fluidigm) and counted using a cell counter (JIMBIO). About 1 × 10⁷ cells per sample were obtained for staining. The antibodies used are listed in table S1. Metal conjugation with antibodies was achieved using metal-labeling kits (Fluidigm, USA). After centrifugation, the cells were resuspended in 150 ml of antibody cocktail solution I (the antibody dilution ratio was 1:200) per 10⁷ cells and incubated for 30 min at room temperature. The antibody cocktail solution I contained CLCA (goblet cells), Ly6C (monocytes), CD45 (leukocytes), Ly6G (neutrophils), CD3 (lymphocytes), and F4/80 (macrophages) antibodies to label the cell membrane markers. After washing with CSB, antibody cocktail solution II containing Pt-cisplatin (living cells), Ir-DNA (all cells), anti-beta III tubulin (enteric nerve cells), pan-keratin (enteric cells), and CHGA (enteroendocrine cells) antibodies was used to label the intracellular markers. Before adding solution II, the cells were permeabilized with Maxpar Perm-S buffer (Fluidigm) for 30 min. A washing procedure (1000 rpm for 3 min) was performed after every staining step. A total of 10% EQ Eight Element Calibration Beads (Fluidigm) were added to the samples before measurement with a Cy-TOF XT (Fluidigm, USA). Last, 2 × 10⁵ cells in each sample were analyzed. Data processing and clustering were performed using the Cytobank platform (http://cytobank.org/). All samples were normalized to the EQ beads. Clustering analysis was performed using Visualization of t-distributed stochastic neighbor embedding (viSNE). The number of iterations and perplexity was 1000 and 30, respectively.

Supplementary Materials

This PDF file includes:

Supplementary Methods

Figs. S1 to S13

Table S1

Correction (5 September 2023):

The original Supplementary Materials included two errors. The title for fig. S10 was originally “Concentration of silver in the nervous system of rodents and primates.” The title has been corrected, as the figure only presents the rodent model. In addition, the two panel labels in fig. S11 (A and B) originally did not correspond to the correct description in the caption. The labels have been corrected. The Supplementary Materials file and HTML have been updated. The original version is available here:

View/request a protocol for this paper from Bio-protocol.