Subtoxic Alterations in Hepatocyte-Derived Exosomes: An Early Step in Drug-Induced Liver Injury?

Abstract

Drug-induced liver injury (DILI) is a significant clinical and economic problem in the United States, yet the mechanisms that underlie DILI remain poorly understood. Recent evidence suggests that signaling molecules released by stressed hepatocytes can trigger immune responses that may be common across DILI mechanisms. Extracellular vesicles released by hepatocytes, principally hepatocyte-derived exosomes (HDEs), may constitute one such signal. To examine HDE alterations as a function of drug-induced stress, this work utilized prototypical hepatotoxicant acetaminophen (APAP) in male Sprague-Dawley (SD) rats, SD rat hepatocytes, and primary human hepatocytes. HDE were isolated using ExoQuick precipitation reagent and analyzed by quantification of the liver-specific RNAs albumin and microRNA-122 (miR-122). In vivo, significant elevations in circulating exosomal albumin mRNA were observed at subtoxic APAP exposures. Significant increases in exosomal albumin mRNA were also observed in primary rat hepatocytes at subtoxic APAP concentrations. In primary human hepatocytes, APAP elicited increases in both exosomal albumin mRNA and exosomal miR-122 without overt cytotoxicity. However, the number of HDE produced in vitro in response to APAP did not increase with exosomal RNA quantity. We conclude that significant drug-induced alterations in the liver-specific RNA content of HDE occur at subtoxic APAP exposures in vivo and in vitro, and that these changes appear to reflect selective packaging rather than changes in exosome number. The current findings demonstrate that translationally relevant HDE alterations occur in the absence of overt hepatocellular toxicity, and support the hypothesis that HDE released by stressed hepatocytes may mediate early immune responses in DILI.

Drug-induced liver injury (DILI) is responsible for significant pre- and post-market drug failures and is the primary cause of acute liver failure in the United States and Europe (Bernal et al., 2010; Ostapowicz et al., 2002; Wilke et al., 2007). DILI, therefore, represents a major economic and clinical challenge for patients, healthcare providers, and pharmaceutical companies. Identification of underlying mechanisms has proved especially difficult as DILI has been associated with hundreds of drugs that elicit a broad spectrum of injuries (Hornby et al., 2014; Hussaini and Farrington, 2014). However, increasing evidence suggests that drug-induced hepatocellular death may often be mediated by immune responses (Adams et al., 2010; Ju and Reilly, 2012; Russmann et al., 2009). For example, human leukocyte antigen genotype is a strong risk factor for DILI associated with a wide range of drugs, suggesting a role for the adaptive immune system in drug-mediated hepatotoxicity (Urban et al., 2014). In addition, elevations in biomarkers that trigger an innate immune response such as high-mobility group box 1 (HMGB1), and those that are suggestive of immune cell activation including miR-155 and acetylated HMGB1, have been associated with DILI (Antoine et al., 2014; Bala et al., 2012; Fontana, 2014). Furthermore, immune reactions are triggered by hepatocellular danger signals, which can be released in the absence of hepatocyte death and communicate to other cell types as part of a stress response (Uetrecht and Naisbitt, 2013).

Extracellular vesicles (EVs) released by hepatocytes, principally hepatocyte-derived exosomes (HDEs), may constitute one such danger signal (Conde-Vancells et al., 2008; De Maio, 2011; Royo and Falcon-Perez, 2012; Royo et al., 2013). By virtue of their small size (<150 nm), HDE can readily cross through fenestrations in the sinusoidal endothelium and enter the bloodstream (Conde-Vancells et al., 2008), carrying liver-specific contents that are modified during DILI (Wetmore et al., 2010). Clinically relevant alterations in HDE content and quantity have been detected in multiple models of liver injury including viral hepatitis (Zhang et al., 2014), alcoholic hepatitis (Momen-Heravi et al., 2015b), nonalcoholic steatohepatitis (Hirsova et al., 2016), cirrhosis, and hepatocellular carcinoma (Bernal et al., 2010). Moreover, exosomes are immunomodulatory and can carry damage-associated molecular patterns (DAMPs), antigens, and cytokines to recipient cells (Beninson and Fleshner, 2014; De Maio, 2011; Liu et al., 2006; Robbins and Morelli, 2014). In the liver, HDE released by hepatitis C-infected hepatocytes activate dendritic cells and those released during lipotoxicity influence macrophage phenotype (Dreux et al., 2012; Hirsova et al., 2016). Recent experiments have demonstrated that one means by which HDE evoke immune responses is the delivery of functional genetic material such as microRNA (miRNA) and messenger RNA (mRNA). For example, HDE-based RNA species miR-122 and albumin mRNA (ALB) have been shown to activate recipient monocytes and fibroblasts, respectively (Momen-Heravi et al., 2015a; Royo et al., 2013).

HDE may play a critical role in the initiation of and response to DILI by transmitting signals of hepatocellular stress that directly or indirectly stimulate an immune response. For HDE to mediate early DILI events, we hypothesized that significant alterations in HDE content and/or number must occur prior to, or in the absence of, overt hepatocellular injury. To better understand the potential of HDE to modulate early immune-driven DILI events, we sought to establish that subtoxic drug-induced HDE alterations were consistent in vivo and in vitro, and relevant in human DILI. Acetaminophen (APAP) was selected as the model hepatotoxicant due to its well-defined in vivo and in vitro toxicity profiles. We utilized this prototypical intrinsic hepatotoxicant for verification that the exosomes released by hepatocytes following drug exposure were altered in the absence of overt injury. Although the role of immune reactions in APAP-induced liver injury is controversial, there is recent evidence of induction of immune tolerance in healthy adults that do not develop alanine aminotransferase (ALT) elevations after recurrent therapeutic doses of APAP (Fannin et al., 2015), further supporting selection of APAP as our model toxicant. This finding also suggests that some stress-induced HDE alterations may promote an adaptive response to toxicity rather than an exacerbation of injury.

We therefore evaluated the dose- and time-dependent effects of APAP on HDE-associated RNAs in rats, primary rat hepatocytes, and primary human hepatocytes. Because exosomes in the blood are derived from multiple cell types, exosome-associated liver-enriched ALB mRNA and liver-specific miR-122 were assayed as a means to confirm the presence of, and track alterations in, HDE across platforms (Bandiera et al., 2015; Miyamoto et al., 2008).

MATERIALS AND METHODS

Animal Exposures

Male Sprague-Dawley rats (Crl:SD), 7–9 weeks of age, were purchased from Charles River Laboratories and allowed to acclimate in-house for 7–10 days prior to study initiation. Water and food (pelleted PicoLab Rodent Diet 20 [Lab Diet]) were available ad libitum during this time. Food was removed during an 18 h fasting period immediately prior to compound administration. APAP (Sigma-Aldrich) was administered in 0.5% (w/v) methylcellulose (Sigma-Aldrich) at 0, 500, or 1400 mg/kg by oral gavage. After dosing, food was provided ad libitum until necropsy. Studies were conducted in accordance with the Guide for the Care and Use of Laboratory Animals, and were approved by the Mispro Institutional Animal Care and Use Committee (Mispro Biotech Services, Inc.).

Necropsy and Histology

Rats were euthanized by CO₂ inhalation and subsequent exsanguination at 0, 1, 2, 4, 8, 12, and 24 h post-dose (n = 6 per group). Blood was collected in K₂EDTA Vacutainer tubes (Becton Dickinson). Portions of the liver (median lobe) were minced and placed in RNAlater Stabilization Solution (Applied Biosystems) for gene expression analysis. Sections from the left and median lobes were fixed in 10% neutral buffered formalin followed by 70% ethanol. Paraffin-embedded blocks were sectioned (5 µm) and stained with hematoxylin and eosin (H&E). Histopathological examination was performed by a board-certified veterinary pathologist. Severity scores were assigned for centrilobular degeneration and necrosis (+1, minimal; +2, mild; +3, moderate; +4, marked). Degeneration was defined as prenecrotic hepatocellular change in centrilobular areas, including cytoplasmic hypereosinophilia, nuclear condensation, and cellular dissociation. Degeneration is assumed to precede fulminant necrosis such that, if given more time, degenerating cells likely would have progressed to necrosis. Necrosis was marked by areas of complete hepatocellular destruction and neutrophil infiltration.

Plasma Isolation and Plasma-Derived Exosome Enrichment

Within 15 min of collection, blood was centrifuged to obtain cell-free plasma. ALT activity in fresh plasma was determined using a CLC 720 clinical chemistry analyzer (Carolina Liquid Chemistries). Additional aliquots of plasma were frozen at −80°C until exosome isolation. Thawed plasma was treated with Purified Thrombin Plasma Reagent and exosomes were isolated with ExoQuick Exosome Precipitation Solution (System Biosciences) according to the manufacturer’s instructions. The post-ExoQuick supernatant was reserved for RNA isolation to analyze liver-specific RNAs in the extra-exosomal, protein-rich plasma fraction.

Primary Hepatocyte Culture and Medium-Derived Exosome Enrichment

Freshly isolated primary hepatocytes from human donors (n = 7; Table 1) and male SD rats (n = 3) were obtained from commercial sources (QPS Hepatic Biosciences and Triangle Research Labs). Individual batches of fresh hepatocytes were seeded into 12-well plates coated with collagen type I (Corning) at a density of 0.9 × 10⁶ cells per well in hepatocyte plating medium (William’s E medium, penicillin-streptomycin, GlutaMax, HEPES, sodium pyruvate, fetal bovine serum [Thermo Fisher Scientific], insulin, and dexamethasone [Sigma-Aldrich]). Culture plates were incubated at 37°C with 5% CO₂. After attachment, nonadherent cells were removed and fresh plating medium was added. Cultures were transitioned to hepatocyte maintenance medium without fetal bovine serum after 3–4 h. Following a 24-h acclimation period, hepatocyte cultures were exposed to APAP or a medium control (vehicle) for 24 h. All conditions were tested in triplicate culture wells for each rat or human donor.

TABLE 1.

Characteristics of primary human hepatocyte donors

Age (year)	BMI	Sex	Race
<1	15	M	African-American
3	17	M	Caucasian
14	24	F	Caucasian
26	32	M	Caucasian
30	33	F	African-American
48	24	F	Caucasian
74	23	F	Caucasian

Hepatocyte-conditioned medium (1–2 ml) was collected after 24 h and clarified by centrifugation at 3000 g for 15 min. Exosomes were precipitated from medium with ExoQuick-TC (System Biosciences) following the protocol for this sample type. Lactate dehydrogenase (LDH) activity was assayed with a Cytotoxicity Detection Kit (Roche) as described previously in Kia et al. (2015). Cellular ATP content was measured with CellTiter-Glo Luminescent Cell Viability Assay (Promega) according to the manufacturer’s instructions. Data were collected using a SpectraMax M3 microplate reader (Molecular Devices).

EV Characterization and Validation of Exosome Enrichment

Nanoparticle tracking analysis

Nanoparticle tracking analysis was performed on exosomes produced in vitro using a NanoSight NS500 (Malvern Instruments). The instrument was calibrated using polystyrene bead standards (100 nm) before each use (Malvern Instruments). Diluted samples were advanced through the detection chamber to obtain 5 readings per sample, each with a capture period of 60 s. Particle concentration and size were determined using NTA software version 3.0 (Malvern Instruments).

Electron microscopy

Exosomes were precipitated from filtered (0.2 µm PES) hepatocyte-conditioned culture medium using ExoQuick-TC. Exosome preparations were diluted 1:1000 in Dulbecco’s phosphate buffered saline to start the assembly reaction. At 10 min, samples were applied to glow-discharged 200 mesh Quantifoil R 2/1 grids (Electron Microscopy Sciences), blotted briefly, and plunged into liquid ethane as described previously (Dokland, 2006). The frozen grids were transferred to a Gatan 622 cryo-holder and imaged using an FEI Tecnai F20 transmission electron microscope, operated at 200 kV with magnifications from 5000 to 29 000× and defocus settings of −6.0 to −10.0 μm. Images were collected on a Gatan Ultrascan 4000 CCD camera.

Immunoblot analysis

Protein extracts were prepared from hepatocyte monolayers or exosomes using RIPA Lysis and Extraction Buffer (Thermo) with Halt Protease Inhibitor Single-Use Cocktail (Thermo), and Triton X-100 (hepatocytes only). Protein concentration was determined using the BCA Protein Assay Kit (Thermo). Equal protein quantities of each sample were separated on 4–12% NuPAGE Bis-Tris gels (Life Technologies) and transferred to Immobilon-FL PVDF membranes (Millipore). Membranes were blocked with either 5% (w/v) nonfat dry milk in PBS or Odyssey PBS Blocking Buffer (Licor Biosciences). The following primary antibodies were used for immunoblotting: Flotillin-1 (BD Transduction Laboratories), CD63 (Abcam), CD81 (AbD Serotech), Prohibitin-1 (Santa Cruz Biotechnology Inc.), and Grp78 (BD Transduction Laboratories). Proteins were detected using IRDye secondary antibodies (Licor Biosciences) or Alexa Fluor 680-conjugated AffiniPure IgG (Jackson Immunoresearch Laboratories). Blots were imaged using an Odyssey Classic Imaging System (Licor Biosciences).

Total RNA Isolation and Analysis by Absolute qRT-PCR

Total RNA was isolated from plasma- and culture medium-derived exosomes using the miRCURY Cell & Plant Isolation Kit (Exiqon) according to the manufacturer’s instructions for RNA isolation from blood. To analyze extracellular RNA in the protein-rich compartment of culture medium and plasma, total RNA was isolated from culture medium samples or post-ExoQuick plasma supernatants, respectively, using the miRNeasy Serum/Plasma Kit (Qiagen) and manufacturer’s protocol. Total RNA from rat liver, rat hepatocytes, and human hepatocytes was isolated with the miRNeasy Mini Kit (Qiagen) according to the manufacturer’s instructions. For all RNA isolations, a spike-in cocktail consisting of linear acrylamide carrier (Applied Biosystems) and exogenous RNA controls, Luciferase mRNA mimic (Promega) and Caenorhabditis elegans miR-39 (Qiagen), was added.

Synthesis of cDNA for mRNA analysis was performed with equal volumes of RNA using a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). RT-PCR was performed on a 7900HT Fast RT-PCR System (Applied Biosystems) using TaqMan Gene Expression Assays and Fast Advanced Master Mix. Absolute quantities of mRNAs were calculated using a species-matched standard curve generated from bacterial plasmids or PCR amplification of liver cDNA. Synthesis of cDNA for analysis of miR-122 copy number was performed using a TaqMan miRNA Reverse Transcription Kit (Applied Biosystems). Absolute qRT-PCR was performed on a 7900HT Fast RT-PCR System (Applied Biosystems) using TaqMan miRNA Assays (hsa-mir-122-5p) and Universal Master Mix II (no UNG). Absolute quantities of miR-122 in cells, exosomes, and the protein-rich fraction were computed from a standard curve of synthetic miR-122 cDNA.

Statistical Analysis

Analyses were conducted using GraphPad Prism 6 (GraphPad Software, Inc.) unless otherwise noted. For all tests, p < 0.05 was considered statistically significant. Mean values from plasma-based endpoints (i.e. exosomal RNA and ALT activity) were analyzed using a 2-way analysis of variance (ANOVA). Results from APAP-treated animals were compared with time-matched control groups with Fisher’s Least Significant Difference (Fisher’s LSD) test. In vitro data from rat hepatocytes were averaged across triplicate wells and a fold change relative to control was calculated for each rat. Fold changes across rats (n = 3) were averaged and compared by 1-way ANOVA. Results from each APAP concentration were compared with controls using Dunnett’s test for multiple comparisons. In vitro data from each human hepatocyte batch were averaged across triplicate wells and a fold change relative to donor-matched controls was calculated. From individual fold changes, a mean across all donors was obtained. Mean fold changes of control and APAP-treated groups were compared by Student’s 2-tailed t-test.

RESULTS

Exosomal Albumin Transcript Quantities Are Significantly Elevated Prior to APAP-Induced Hepatic Necrosis in Rats

Adult male SD rats were administered a low (500 mg/kg) or high (1400 mg/kg) dose of APAP or a vehicle control and were sacrificed at 0, 1, 2, 4, 8, 12, or 24 h post-dose. Liver injury was determined by histology and plasma ALT activity. Liver-specific RNAs isolated from circulating exosomes were quantified to determine the relationship between HDE alterations and liver injury. Histological findings of necrosis (Figs. 1A–D; Supplementary Table 1) and elevated plasma ALT activity (Figure 2A, Supplementary Figure 1A) were observed only in rats treated with the high dose of APAP at 24 h post-exposure. Exosomal albumin transcript (ALB) quantity, however, was significantly elevated in APAP-treated animals (relative to time-matched, vehicle-treated controls) prior to histological observations of necrosis or increases in plasma ALT (Figure 2B, Supplementary Figure 1B). Drug-induced elevations in mean exosomal ALB quantity were detected as early as 4 h post-dose; these elevations became dose-dependent at 8 h post-dose. By 12 h, mean exosomal ALB mRNA had increased 30- and 50-fold in the low- and high-dose groups, respectively (Figure 2B). Increases in exosomal ALB reached statistical significance in the high dose group at 12 h and in the low dose group at 24 h, without signs of liver injury (Figure 2B, Supplementary Figure 1B). Together, these data suggest that changes in HDE release occur as a result of subtoxic APAP exposure in vivo.

FIG. 1.

Histopathology of rat liver sections 24 h after APAP treatment. Representative photomicrographs of H&E stained liver sections from rats 24 h after administration of (A) 0.5% methylcellulose vehicle, (B) 500 mg/kg APAP, or 1400 mg/kg APAP with (C) moderate or (D) marked centrilobular necrosis. Bars indicate 200 µm.

FIG. 2.

In vivo elevations in ALT activity and exosomal ALB as a function of APAP dose and exposure time. APAP-induced elevations in (A) plasma ALT activity and (B) exosomal ALB mRNA. Data are presented as mean + SEM of fold change over time-matched controls (n = 5–6 rats/group). *p < 0.05, ****p < 0.0001; 2-way ANOVA with Fisher’s LSD.

Quantity and Distribution of Extracellular miR-122 Are Altered by Subtoxic APAP Exposure In Vivo

Unlike ALB mRNA which was not detected outside of the exosomal plasma fraction (data not shown), circulating miR-122 is stable in exosomes and in a ‘protein-rich’ Argonaute-bound form (Arroyo et al., 2011; Hornby et al., 2014). In order to determine if miR-122 followed a similar trend as exosomal ALB upon in vivo APAP exposure, we measured exosomal and protein-rich fractions of circulating miR-122. A trend toward increased exosomal miR-122 was observed in APAP-treated animals at 8 h post-dose (Figure 3A, Supplementary Figure 1C). This trend continued, with mean levels reaching 3- and 4-fold over controls in the low- and high-dose groups at 12 h, respectively. At 24 h post-dose, mean exosomal miR-122 quantities in the low-dose group reached 9-fold that of vehicle-treated controls, despite a lack of hepatocellular toxicity. MiR-122 in the protein-rich fraction showed trends similar to exosomal miR-122 but was significantly elevated in the low-dose group after 24 h (Figure 3B, Supplementary Figure 1D). Whereas the content of miR-122 in both exosomal and protein-rich compartments was increased by APAP exposure in our study, the distribution of miR-122 between these plasma compartments shifted over time. In the high-dose group, the percentage of circulating miR-122 within HDE decreased over time and reached statistical significance at 12 h, prior to overt liver injury (Figure 3C).

FIG. 3.

Alterations in exosomal and protein-rich fractions of miR-122 after in vivo APAP exposure. Fold change in (A) exosomal and (B) protein-rich miR-122 plasma compartments relative to time-matched vehicle controls (n = 5–6 rats/group). *p < 0.05, **** p < 0.0001; 2-way ANOVA with Fisher’s LSD. (C) Percentage of circulating miR-122 in HDE from 1400 mg/kg APAP-treated rats. *p < 0.05, ***p < 0.001; 1-way ANOVA with Dunnett’s post-test relative to t = 1 h. (D) Relationship between liver-specific RNAs measured in exosomes and liver tissue at 2 h post-dose. *p < 0.05; 1-way ANOVA with Dunnett’s post-test relative to vehicle. Data are presented as mean + SEM.

Unexpectedly, we observed previously undocumented dose-dependent decreases in exosomal and protein-rich miR-122 fractions after 2 h of APAP exposure, which were not accompanied by decreases in exosomal ALB (Figure 3D). To determine whether the decreases in exosomal miR-122 reflected selective packaging of RNA into HDE or decreased hepatocyte miR-122 content, we quantified miR-122 in liver tissue obtained from the 2 h post-exposure animals. There was a trend toward increased hepatic miR-122 copy number coincident with statistically significant dose-dependent decreases in exosomal miR-122 (Figure 3D). Exosomal ALB, however, was not significantly affected in circulating exosomes or liver tissue (Figure 3D).

EV Preparations From Primary Rat and Human Hepatocyte Cultures Are Highly Enriched for Exosomes

We selected a well-characterized exosome enrichment method for our in vivo studies, and evidence suggests that most EVs released by hepatocytes are exosomal in size (Giugliano et al., 2015; Koeck et al., 2014; Momen-Heravi et al., 2015a; Royo et al., 2013). However, multiple cell types and tissues contribute to the population of circulating vesicles, and therefore we could not study the nature of hepatocyte-specific vesicles isolated from the in vivo samples. To confirm the presence of exosomes in EVs isolated by ExoQuick, we analyzed the size, morphology, and protein enrichment of EVs produced in vitro by primary rat and human hepatocytes.

EVs were isolated from conditioned medium of untreated primary rat or human hepatocytes by ExoQuick-TC polymer precipitation. Nanoparticle tracking analysis of rat and human EV preparations revealed a majority of small exosome-sized particles <200 nm in diameter, with the greatest concentration of particles being <100 nm in size (Figs. 4A and B). These results were confirmed using a ZetaView instrument (ParticleMetrix) (data not shown). Cryo-electron microscopy verified exosome enrichment in precipitated rat and human EV preparations, which contained circular membrane-bound vesicles smaller than 100 nm (Figs. 4C and D). Western immunoblot analyses of rat EVs confirmed the presence of exosomal protein markers CD81 and CD63 (Figure 4E). Prohibitin-1, a mitochondrial membrane protein, was detected in rat hepatocytes but was not present in rat HDE, suggesting a lack of cellular contamination in exosomal samples (Figure 4E). Human EV preparations were positive for exosomal markers CD63 and Flotillin-1 and did not contain detectable cellular contamination, as measured by Prohibitin-1 and Grp78, a protein found in the endoplasmic reticulum (Figure 4F). These results confirm that our ExoQuick preparations contain mostly HDE.

FIG. 4.

Characterization of exosome enrichment in EVs secreted by primary rat and human hepatocytes. Representative nanoparticle tracking analyses illustrating the average size distribution (+ SEM) of EVs isolated from untreated (A) rat and (B) human hepatocyte-conditioned culture medium. Representative cryo-electron micrographs of (C) rat and (D) human hepatocyte-derived EV preparations. Bars indicate 100 nm. Western blot analysis of exosomal and cellular markers in (E) rat and (F) human HDE (‘Ex’) relative to hepatocytes (‘Cell’).

Subtoxic APAP Exposure Elicits Exosomal RNA Elevations in Primary Rat Hepatocytes

To further characterize HDE release from drug-treated hepatocytes, we explored whether APAP-dependent changes in exosomal RNA content observed in vivo could be reproduced in monocultures of primary rat hepatocytes. Primary rat hepatocytes were exposed to a concentration range of APAP (0–30 mM) for 24 h. Across 3 rat hepatocyte batches, statistically significant elevations in LDH leakage and decreases in cellular ATP were not observed at concentrations <30 mM APAP (Figure 5A). However, statistically significant increases in exosomal ALB mRNA were evident at concentrations as low as 1.1 mM APAP (Figure 5B). Hepatocellular ALB levels were not altered significantly at any APAP concentration, suggesting that increased exosomal ALB was not a result of increased expression at the cellular level. In contrast to our in vivo findings, miR-122 was not elevated significantly in the exosome- or protein-rich fractions prior to the onset of overt cytotoxicity (Figure 5C).

FIG. 5.

APAP-induced alterations in exosomes released by primary rat hepatocytes. Data are presented as mean + SEM (n = 3 rats) after 24 h APAP exposure. (A) Cytotoxicity, as indicated by cellular ATP and LDH activity in medium. (B) Exosomal and cellular levels of ALB in response to APAP. (C) miR-122 in exosomal, protein-rich, and cellular fractions in response to APAP. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; 1-way ANOVA with Dunnett’s post-test relative to vehicle. (D) Nanoparticle tracking analysis on HDE preparations from primary rat hepatocytes (n = 2 rats) after 24 h incubation with APAP. Data are presented as the mean + SEM of fold change in exosomes/ml relative to control. No statistically significant differences in exosome quantity between treated and untreated groups were observed; 1-way ANOVA with Dunnett’s post-test relative to vehicle.

We then tested the hypothesis that liver-specific exosomal RNA levels are indicative of HDE number, which cannot currently be assessed in vivo. We performed nanoparticle tracking analysis on HDE from primary rat hepatocyte cultures exposed to 0, 10, or 30 mM APAP for 24 h (n = 2 rats). No statistically significant changes in HDE quantity were observed at any concentration (Figure 5D).

Increased Secretion of Exosomal RNA in the Absence of Overt APAP Toxicity Translates to Primary Human Hepatocyte Cultures

To determine whether primary human hepatocytes would respond to APAP exposure by increasing production of exosomal RNAs, we separately exposed individual batches of freshly isolated primary hepatocytes (n = 7 donors) to a subtoxic APAP concentration (10 mM) for 24 h (Figure 6A). APAP treatment caused increases in mean exosomal ALB (not statistically significant) and miR-122 (p < 0.001) levels in the absence of overt injury (Figs. 6B and C). Quantities of ALB and miR-122 in cells and protein-rich medium fractions were not affected by APAP exposure (Figs. 6B and C).

FIG. 6.

Alterations in exosomes released by primary human hepatocytes with and without APAP exposure. (A) LDH leakage; (B) ALB mRNA; and (C) miR-122 after 24 h of exposure to APAP (10 mM). Data are presented as the mean + SEM across 7 donors. ***p < 0.001; Student’s t-test comparing treatment and control. (D) Nanoparticle tracking analysis on HDE preparations from primary human hepatocytes (n = 4 donors) after 24 h incubation with APAP. Data are presented as the mean + SEM of fold change in exosomes/ml relative to control. No statistically significant differences in exosome quantity between treated and untreated groups were observed; Student’s t-test.

To define the relationship between drug-induced changes in liver-specific exosomal RNA and HDE number in a human system, nanoparticle tracking analysis was performed on exosomes isolated from control and treated (0 vs 10 mM APAP) primary human hepatocyte cultures (n = 4 donors). As observed in the rat experiments, APAP exposure did not cause statistically significant changes in exosome quantity (Figure 6D).

Establishing Baseline Quantities of Liver-Specific RNAs in Exosomes From Primary Human Hepatocytes

Individual lots of primary human hepatocytes (n = 7) were cultured for 24 h with or without 10 mM APAP. Substantial donor-to-donor variability in the exosomal quantities of ALB and miR-122 was apparent. To identify donor characteristics that potentially influenced exosomal content, we related characteristics such as sex, age, BMI, and race to the quantity of liver-specific exosomal RNAs from medium control- and APAP-treated hepatocytes (Supplementary Figure 2A). No association reached statistical significance, but we observed a relationship between exosomal ALB and BMI (p = 0.074, regression analysis) in untreated hepatocytes (Supplementary Figure 2B).

In an effort to establish healthy ranges of exosomal ALB and miR-122, we examined absolute copy numbers of these exosomal RNAs across hepatocyte donors (n = 7) under control conditions. We found that quantities of exosomal miR-122 generally corresponded with levels of protein-rich miR-122, although this relationship did not hold true for every donor (Supplementary Figure 3A). We did not observe a consistent relationship between exosomal and cellular ALB. Absolute copy numbers of exosomal miR-122 ranged from approximately 10⁷–10⁹, while exosomal ALB ranged from approximately 10³–10⁶ (Supplementary Figs. 3A and B).

Additional Liver-Enriched Transcripts Are Present in Human HDEs

Finally we sought to confirm that additional liver-enriched mRNAs previously identified in rat EVs, and enriched in both rat and human liver, were packaged within human HDE (Miyamoto et al., 2008; Okubo et al., 2013; Royo et al., 2013; Yu et al., 2010). Exosomes were precipitated from the culture medium of untreated primary human hepatocytes (n = 3 donors), and exosomal transcripts were quantified by absolute qRT-PCR. All 7 liver-enriched transcripts were detected in human HDE under control conditions, and although the absolute amounts varied, the relative proportions of each mRNA across donors were consistent (Figure 7).

FIG. 7.

Analysis of additional liver-enriched transcripts within exosomes released by untreated primary human hepatocytes. Bars denote different donors (n = 3). Data represent the mean + SEM of exosomal mRNAs across triplicate culture wells for each donor. Transcript quantities are expressed as exosomal copy number per ml of culture medium used to isolate HDE.

DISCUSSION

To determine whether HDE could mediate the immune reactions that occur without overt hepatotoxicity following drug exposure, we tested the hypothesis that HDE content and/or number would be altered by drug-induced hepatocellular stress in the absence of cell death. We found that the RNA content of HDE was affected prior to, and in the absence of, overt hepatocellular injury in rats receiving both toxic and subtoxic doses of APAP. These results translated across experimental models and species as similar phenomena were observed in primary hepatocyte cultures from both rats and humans.

APAP was selected for this work largely for its defined toxicity profile in vivo and in cultured hepatocytes from both rats and humans. Although we are ultimately interested in the role of exosomes in immune signaling, this study does not address the consequences of the immune component of APAP-induced hepatotoxicity, a topic that remains controversial. Whether advantageous (Antoniades et al., 2012; Jaeschke et al., 2012; Ju et al., 2002; Williams et al., 2014) or damaging (Fannin et al., 2015; Ferreira et al., 2016; Huebener et al., 2015; Tujios and Fontana, 2011), there is no question that inflammatory and immune responses are associated with APAP-induced hepatotoxicity in humans. In a study of recurrent therapeutic APAP dosing, activation of immune tolerance was evident in the peripheral blood transcriptome of healthy adults without APAP-induced ALT elevations (Fannin et al., 2015). This suggests that immune responses after APAP exposure manifest prior to overt hepatocellular injury, a process that likely involves subtoxic hepatocellular release of stress signals such as HDE.

Although mouse models are generally considered to be more appropriate for the study of APAP-induced hepatotoxicity mechanisms (Jaeschke et al., 2014; McGill et al., 2012), the goal of the current work was to evaluate early changes in HDE that occur during APAP-induced stress, prior to hepatocellular injury. Therefore, the rat was intentionally selected as a relatively resistant preclinical species to extend the duration of APAP-induced stress in vivo so that HDE alterations could be observed prior to overt DILI. Evidence suggests that this latency may be more similar to humans than the latency to fulminant hepatotoxicity observed in mice (Jaeschke et al., 2014). Because mouse hepatocytes are difficult to culture, selection of a rat model also facilitated comparisons of data collected from primary hepatocytes relative to in vivo data from the same species.

In this study, we selected liver-specific exosomal contents ALB mRNA and miR-122 as translational markers of HDE alteration that also elicit immunologically relevant responses when transferred to other cell types (Momen-Heravi et al., 2015a; Royo et al., 2013). Circulating ALB mRNA and miR-122 have been identified previously as more sensitive and specific biomarkers of liver injury in both preclinical species and humans (Antoine et al., 2013; Okubo et al., 2013; Starckx et al., 2013; Wang et al., 2009; Zhang et al., 2014). In particular, the value of miR-122 in peripheral blood as a biomarker of APAP overdose is well established (Antoine et al., 2013). Therefore, the biomarker potential of miR-122 and HDE was not a focus of this work.

In our in vivo rat model of APAP-induced liver injury, the proportion of plasma miR-122 contained within exosomes decreased over time in the high-dose group, presumably as a function of increased hepatocellular damage. These data are in agreement with what is known about the active nature of exosome release and the passive, necrotic release of protein-rich miR-122 (Bobrie et al., 2011; Dear and Antoine, 2014; Kia et al., 2015; Turchinovich et al., 2011, 2013; Villarroya-Beltri et al., 2014). Similar shifts in miRNA localization between protein-rich and exosomal fractions have been observed in DILI, and the nature of these shifts may reflect different mechanisms of liver injury (Bala et al., 2012, 2015). We propose that shifts in miRNA localization, such as those shown here, may serve a biological signaling purpose. For example, miR-122 in HDE released by alcohol-treated hepatocytes can prime monocytes for activation, while biological activity has yet to be attributed to miRNA in the protein-rich form (Momen-Heravi et al., 2015a; Turchinovich et al., 2013). Similarly, the effects of inflammatory mediators such as Hsp70 are amplified when they are delivered to recipient cells via exosomes (De Maio, 2011). While it is unlikely that exosomal miR-122 is a driving factor or major inflammatory mediator in overt APAP hepatotoxicity, during which necrosis releases massive amounts of DAMPs, the present miR-122 results support the assertion that HDE alterations may be actively mediated and functionally relevant in the early stages of hepatocyte stress.

In vivo, parallel increases in HDE-associated ALB mRNA and miR-122 were generally observed. However, exosomal ALB mRNA did not decrease in vivo at early time points alongside exosomal miR-122, providing evidence for the selective packaging of RNA into HDE. Selective mRNA packaging into exosomes has been observed in circulating vesicles from rats treated with APAP or D-(+)-galactosamine (DGAL), which contained distinct transcript profiles suggestive of compound-specific exosomal content (Wetmore et al., 2010). We confirmed these results with HDE, as different trends in exosomal ALB mRNA and miR-122 were observed following APAP and DGAL exposures in primary rat hepatocytes (Supplementary Figure 4). Although subtoxic concentrations of DGAL elicited increases in both exosomal RNAs, neither reached statistical significance prior to overt cytotoxicity. Our nanoparticle tracking results also recapitulate the selective packaging hypothesis by indicating that the quantity of exosomal RNA does not correspond to HDE number. As such, the APAP-induced exosomal ALB mRNA elevations observed in vivo and in vitro are likely the result of increased ALB packaging within a similar number of HDE rather than an increase in overall HDE production with constant ALB mRNA content. Although these findings are novel, it is important to note that exosome quantification could be affected by the stability of HDE in culture, which is currently unknown. Future experiments to establish the half-life and clearance rate of HDE will be imperative for understanding the biological implications of HDE in DILI.

Although the intent of this work was not to study biomarkers, by defining the kinetics and dose-dependence of HDE release prior to overt DILI, this study could help to validate HDE as sensitive, translational, and clinically relevant biomarkers of DILI. We assayed 7 liver-enriched transcripts within primary human HDE from 3 separate donors. To our knowledge, this is the first characterization of the mRNA content of HDE from primary human hepatocytes. All 7 liver-enriched transcripts were measurable, and there was some consistency in the relative abundance of each transcript across the 3 donors, suggesting that these mRNAs, originally detected in rat HDE (Miyamoto et al., 2008; Okubo et al., 2013; Royo et al., 2013), may also be used to identify human HDE. In addition, we have begun to define normal levels of exosomal ALB mRNA and miR-122 across donors. Despite the small donor pool, potential relationships were identified between donor characteristics and exosomal content. Continuing to study donor characteristics that influence HDE RNA content and to define a healthy range of HDE-based RNA levels will be important for interpreting biomarker data in the future. These data will need to be considered in the context of disease, which is known to affect basal HDE quantity and content. For instance, recent evidence suggests that high hepatocellular lipid content and nonlcoholic fatty liver disease correspond to increased EV production by hepatocytes (Hirsova et al., 2016; Kakazu et al., 2015).

In summary, we detected significant alterations in the liver-specific RNA content of HDE at subtoxic doses of the prototypical hepatotoxicant APAP in rats and in primary rat and human hepatocytes. These alterations appear to reflect differential packaging of select RNAs in HDE rather than HDE number. The changes in HDE content observed prior to and in the absence of overt hepatocellular injury, together with growing data on the role of exosomes in stimulating immune responses, support the hypothesis that HDE released from drug-stressed hepatocytes may transmit signals that mediate critical early immune responses across multiple forms of DILI. While this hypothesis was substantiated using an intrinsic hepatotoxicant, future work will address the release kinetics and signaling potential of HDE released by hepatocytes exposed to idiosyncratic DILI compounds.

SUPPLEMENTARY DATA

Supplementary data are available online at http://toxsci.oxfordjournals.org/.

FUNDING

This work was supported in part by the National Center for Advancing Translational Sciences (NCATS), National Institutes of Health (5UL1TR001111). N.S.H. was supported in part by the National Institutes of Health, National Institute of Environmental Health Sciences (NIEHS) Toxicology Training Grant (T32-ES007126). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.