Elimination of Fkbp51 ameliorates CCl 4 -induced liver injury.
To determine the role of FKBP51 in liver injury, wildtype (WT) and KO mice were treated with CCl4 three times per week for two weeks. In the WT mouse liver, CCl4 treatment resulted in a robust color change, rough surface, and stiffness with multiple fibrotic nodules (Fig. 1A, arrows). However, the CCl4-treated livers of KO mice exhibited fewer of these pathophysiological phenotypes, suggesting a protective effect of in Fkbp51 KO (Fig. 1A). Serum levels of aspartate transaminase (AST) and alanine aminotransferase (ALT) (Fig. 1B and C) were low at baseline, however, serum AST and ALT levels were significantly more elevated in WT mice than in KO following CCl4 treatment (Fig. 1B and C), indicating less liver damage in the KO mouse group.
Masson’s trichrome staining (Fig. 1D) was used to evaluate histopathologic alterations in the tissue. Both WT and KO livers in the control group contained cells possessing a normal morphology and regular collagen distribution. After CCl4 treatment however, fatty degeneration (*), hepatocellular necrosis (arrows), and extensive collagen deposition, were observed in WT mice (Fig. 1D, CCl4 panel). Notably, collagen bundles surrounding hepatic nodules were significantly attenuated in KO mice when compared to WT mice (Fig. 1D). Consistent with these findings, immunohistochemical (IHC) labeling showed that hepatic collagen I deposition was much lower in KO relative to WT mice after CCl4 treatment (Supplemental (S)-Fig. 1A). The grade of liver fibrosis was scored following published guidelines and was shown to be significantly lower in KO mice relative to WT mice following CCl4 treatment (P < 0.0001) with no differences observed at baseline (Fig. 1E). (Dai et al., 2014). Other fibrogenesis-associated markers upregulated to a greater degree in WT mice after CCl4 injection, included tissue inhibitor of metalloproteinase 1 (TIMP-1) (Fig. 1F and G) and connective tissue growth factor (CTGF) (S-Fig. 1B). Alpha smooth muscle actin (α-SMA), a marker of hepatic stellate cell (HSC) activation, was also expressed at higher levels in CCl4-injured WT liver than in KO (S-Fig. 1C). These data indicate that KO mice are less susceptible to CCl4-induced liver injury-related fibrosis, matrix remodeling, and HSC activation relative to WT mice. Additionally, more apoptotic cells (arrows), a key feature of liver injury, were found in CCl4-treated WT mice compared to KO mice as determined by terminal deoxynucleotidyl transferase deoxyuridine triphosphate nick end (TUNEL) staining (Fig. 1H and I). The observed morphological and cellular maker alterations demonstrate that loss of Fkbp51 confers protection from hepatic injury.
RNA transcriptome profiling identifies pathways and genes important to liver and mitochondrial function.
RNA-seq was applied to profile gene expression differences in the livers of KO and WT mice with or without CCl4 treatment. Pairwise comparisons identified differentially expressed genes (DEG). Using a fold change cut off > 2.0 and an adjusted p-value (adj p) of < 0.05, the number of DEG between KO and WT included genes specific for either CCl4 injection (745 genes), control solvent (105 genes), or both (48 genes) (Fig. 2A). The 745 unique DEG between KO and WT for CCl4 treatment became our main interest and were used as input for ingenuity pathway analysis (IPA). The bar-chart represents the most significant pathways altered between KO and WT mice after CCl4 injury (Fig. 2B). Among those, we observed significant differences in pathways relevant to liver function, such as lipopolysaccharides (LPS) / interleukin 1 (IL1)-mediated inhibition of retinoid X receptor (RXR), hepatic fibrosis/HSC activation, and liver X receptor (LXR)/RXR activation fatty acid β-oxidation I, demonstrating a distinctive response to CCl4 injury in Fkbp51 KO mice. Molecules of interest in the IPA pathway are included in S-Table 1. Similarly, the identified Kyoto encyclopedia of genes and genomes (KEGG) pathways and involved proteins suggested PPAR signaling, metabolic pathways and protein process in endoplasmic reticulum (ER), and many others are significantly different between KO and WT mice (Table 1). These signaling pathways directly regulate liver function and are well established in the progression of liver disease pathology, further supporting a critical role of Fkpb51 in liver injury response (Geisler and Strazzabosco, 2015).
Table 1
Pathway description
|
Observed gene count
|
FDR
|
Matching proteins in the data
|
PPAR signaling pathway
|
9
|
7.56E-06
|
ACSL3,ACSL4,APOA2,APOA5,CYP7A1,CYP8B1,FABP4,FABP5,LPL
|
Metabolic pathways
|
27
|
0.00416
|
ACSL3,ACSL4,AGXT2L1,ALAS1,ALDH3A2,AMDHD1,ATP6V0C,B3GALT1,CSAD,CYP2E1,CYP7A1,CYP8B1,GCLC,GLUL,HAL,HMGCR, HSD3B2,HSD3B7,HYI,ISYNA1,LPIN1,LPIN2,MOCS1,OAT,RRM2,SPHK2,SQLE
|
Protein processing in endoplasmic reticulum
|
9
|
0.00416
|
DNAJC3,HSP90AA1,HSP90B1,HSPA5,HYOU1,PDIA3,PDIA4,SSR3,SYVN1
|
Bile secretion
|
6
|
0.00453
|
ABCG5,ABCG8,CYP7A1,HMGCR,NR0B2,SLC10A2
|
Amoebiasis
|
7
|
0.00506
|
CD14,COL1A1,COL1A2,COL3A1,COL4A1,COL5A3,IL1R1
|
ECM-receptor interaction
|
6
|
0.00897
|
COL1A1,COL1A2,COL3A1,COL4A1,COL5A3,SPP1
|
Protein digestion and absorption
|
6
|
0.00897
|
COL1A1,COL1A2,COL3A1,COL4A1,COL5A3,SLC3A1
|
Platelet activation
|
7
|
0.0104
|
COL1A1,COL1A2,COL3A1,COL5A3,FGA,FGB,FGG
|
Primary bile acid biosynthesis
|
3
|
0.0181
|
CYP7A1,CYP8B1,HSD3B7
|
PI3K-Akt signaling pathway
|
11
|
0.0181
|
CDKN1A,COL1A1,COL1A2,COL3A1,COL4A1,COL5A3,HSP90AA1,HSP90B1,IL6R,NR4A1,SPP1
|
Glutathione metabolism
|
4
|
0.0355
|
GCLC,GSTA2,GSTM4,RRM2
|
Transcriptional misregulation in cancer
|
7
|
0.0355
|
CD14,CDKN1A,CEBPA,CEBPE,ID2,TMPRSS2,ZBTB16
|
Glycerolipid metabolism
|
4
|
0.047
|
ALDH3A2,LPIN1,LPIN2,LPL
|
Regulator effects analysis featuring the top predicted upstream regulators were plotted by their activation Z-score (Fig. 2C). Immuno-responsive substances (LPS, transforming growth factor β1 (TGFβ1), IL6, and IL1B), particularly, inactivation of IL6 and IL1B in KO mice indicate decreased liver inflammatory response (Fig. 2C). We also measured immune factors and consistently found lower levels of interferon γ (IFNγ), IL6, TGFβ1, and nuclear factor-κB (NFκB) in KO mice relative to WT mice at baseline (S-Fig. 2A-D). Following treatment with CCl4, the levels of these factors remained significantly lower in KO mice compared to WT (S-Fig. 2A-D). The levels of IL10 and tumor necrosis factor α (TNFα) were similar between KO and WT mice at baseline but were lower in KO mice compared to WT mice after CCl4 injury (S-Fig. 2E and F). Additionally, hormone regulators (Medrol, dihydrotestosterone (DHT), and leptin (LEP)), D-glucose, and CCl4 were inactivated, while PPARγ, acyl-coA oxidase 1 (ACOX1), and aryl hydrocarbon receptor (AHR) were activated in KO mice after CCl4 injury, suggesting that FKBP51 plays a role in hormone nuclear translocation, metabolism of glucose, and other functions in response to CCl4-induced injury (Fig. 2C). It was confirmed that glucocorticoid (GC) and fibroblast growth factor (FGF) levels were lower in KO than WT (S-Fig. 2G and H). Further analysis on downstream effect networks identified the engulfment of myeloid cells, endocytosis, and activation of antigen presenting cells (Table 2). Thus, our data support differential responses between KO and WT mice in the recruitment of myeloid cells to the liver and the secretion of inflammatory cytokines through the innate immune in response to liver injury.
Table 2
Identified disease and function categories and involved regulators when compare Fkbp51 KO and WT after CCl4-induced liver injury.
Top Regulator Effect Networks
|
|
|
Regulators
|
Disease and Function
|
Consistency Score
|
Akt,APOE,C5,CSF1,CSF2,Ige,IL1,IL17A,IL1A,IL2,KIT
|
Engulfment of myeloid cells
|
43.7
|
Alpha catenin,AR,CSF2,EGFR,FOXA1,GLIS2,IL17A,IL2
|
Endocytosis, Engulfment of cells
|
21.1
|
ACOX1,CREBBP,CSF2,IFI16,IL17A,IL2,OSM,PRKCA,ROCK2
|
Arteriosclerosis, Endocytosis, Engulfment of cells
|
12
|
Akt,Brd4,CSF2,F2,IL17A,KITLG,PRKCD,TLR4
|
Activation of antigen presenting cells, Endocytosis
|
20
|
Table 3
Predicted functional alterations in Fkbp51 KO after CCl4-liver injury.
Function
|
Category
|
p-value
|
|
z-Score
|
Cellular Movement
|
Cell movement
|
3.75E-16
|
Decreased
|
-4.673
|
Cell-To-Cell Signaling and Interaction
|
Activation of cells
|
5.45E-10
|
Decreased
|
-4.262
|
Cellular Function and Maintenance, Inflammatory Response
|
Phagocytosis
|
2.21E-06
|
Decreased
|
-3.891
|
Inflammatory Response
|
Immune response of cells
|
1.28E-07
|
Decreased
|
-3.799
|
Amino Acid Metabolism, Small Molecule Biochemistry
|
Metabolism of amino acids
|
4.33E-07
|
Increased
|
2.185
|
Lipid Metabolism, Small Molecule Biochemistry
|
Conversion of lipid
|
5.63E-07
|
Increased
|
2.312
|
Organismal Survival
|
Morbidity or mortality
|
1.12E-13
|
Increased
|
2.442
|
Organismal Survival
|
Organismal death
|
4.81E-13
|
Increased
|
2.514
|
The most significant DEG expression level with fold changes (FC) > 4 (or Log2 FC > 2) included in S-Table 2 were further analyzed for their protein–protein interactions using the Search Tool for the Retrieval of Interacting Genes (STRING) database. Five robust functional gene hubs centered by HSP90aa1, FKBP51, and Parkin were identified. Genes in each cluster with their gene expression log2FC between KO and WT are listed in Fig. 2D. Prominently, a group of Cyp450 family members were upregulated in KO mice, while DEG related to cell cycle, lipid biosynthesis, fibrogenesis, and immune-function groups were downregulated (Fig. 2D). Further analysis points to FKBP51 as a key player in affecting associated genes and resulting in unique changes in disease and development, physiological systems, signaling pathway, and toxicity following CCl4 injury (S-Fig. 3A). Enrichment of biological process, cellular component, and molecular function further highlighted functionally-relevant changes between KO and WT mice following liver injury (S-Fig. 3B), which include response to organic substances and chemicals (biological process); extracellular, membrane-bounded vesicle, endoplasmic reticulum, and endocytic vesicle lumen (cellular component); and protein folding, glycosaminoglycan binding and anion binding, protein binding, and enzyme activity (molecular function). These data suggest that FKBP51 plays a critical role in response to liver injury by altering metabolism, immune response, and other biological processes.
Elimination of Fkbp51 affects mitochondrial function-related gene and capacity for ATP production.
As Parkin and Hsp90 are involved in mitochondrial function, we studied DEG related to mitochondrial function. We observed that DEG related to mitochondrial function are overrepresented (more than 10%) (p < 0.05, FC > 2) in the data, and the top DEG were further analyzed (Fig. 3A). STRING analysis was utilized to cluster the genes based on their functional connections. Genes related to triacylglyceride (TAG) synthesis were downregulated (Lipin 1 (Lpin1), acyl-CoA synthetase long chain family member 3 (Acsl3), and Acsl4), translocator genes were upregulated (translocase of outer mitochondrial membrane 40 like (Tomm40l), solute carrier family 8 member b1 (Slc8b1)), and the mitophagy-related gene Parkin (Park2) was upregulated in KO following CCl4 treatment (Fig. 3A). Additionally, sulfite oxidation-related genes (sulfite oxidase (Suox) and ethylmalonic encephalopathy protein 1 (Ethe1)) and amino acid degradation-related genes (proline dehydrogenase (Prodh), and glutaminase 2 (Gls2)) were upregulated (Fig. 3A). Gene expression fold-changes of these mitochondrial function-related DEG between KO and WT mice are shown in Fig. 3A. The identification of a functional connection between Parkin and FKBP51 (Qiu et al., 2022), along with the established relationship between Pink1 and FKBP51 (Boonying et al., 2019), strengthens the argument that FKBP51 plays an important role in mitochondrial regulation. Our study further pointed out those DEG with higher FC between KO and WT and their involved functions.
Based on the data demonstrating that mitochondrial function-related DEG were altered, we compared primary cultured WT and KO mouse embryonic fibroblasts (MEFs) and examined whether they exhibit an intrinsic difference in energy production. Indeed, KO MEFs possess a significant increase in ATP production compared to WT (Fig. 3B). To understand if cellular ATP production differences are attributable to differences in mitochondrial respiration, the Seahorse XF Cell Mito Stress Test was conducted to assess mitochondrial respiration differences between cultured KO and WT MEFs. Following sequential treatments of oligomycin, carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP), rotenone (Rot), and antimycin A, the oxygen consumption rate (OCR) was measured at four parameters of mitochondrial function: basal respiration, ATP turnover, proton leak, and spare respiratory capacity (Fig. 3C). The parameter calculations from the above measurements demonstrated that basal respiration, ATP production, maximal respiration, spare respiratory capacity, and non-mitochondrial respiration, were all increased in KO MEFs, while H+ (proton) leak did not significantly differ between WT and KO MEFs (Fig. 3D). Together, these findings suggest that the loss of Fkbp51 induces higher mitochondrial function in energy metabolism as measured by ATP production.
Co-localization of Parkin with FKBP51 protein in mitochondria
Our above RNA-seq data indicate that Fkbp51 KO possess upregulated Park2 expression (Fig. 3A), a gene crucial for mitochondrial function and cytoskeleton formation (Huynh et al., 2000; Vergara et al., 2015; Williams et al., 2015; Xiao et al., 2017). We confirmed a higher Parkin mRNA expression in KO mice compared to WT by real-time PCR (Fig. 4A). While CCl4 treatment decreased the mRNA expression in both groups, Parkin was higher in KO mice compared to WT (Fig. 4A). Based on these findings, Parkin protein expression in KO was studied using IHC and western blotting. Liver IHC revealed that Parkin expression is increased in KO mice when compared to WT, regardless of CCl4 treatment (Fig. 4B and C). Consistent with these findings, western blotting demonstrated that Parkin protein levels are higher in KO mice versus WT both at baseline and following CCl4 treatment (Fig. 4D and E). Moreover, Parkin protein expression is higher in both the enriched mitochondrial fraction and cytosolic fraction in KO mice (Fig. 4F-I).
To understand these differences at the cellular level, mitochondria (MitoTracker labeling) and Parkin were tracked in WT and KO MEF cells. Parkin was expressed strongly in the cytoskeleton, more densely in KO MEFs than WT MEFs, particularly in the perinuclear area (Fig. 4J). A robust increase in mitochondrial network signal intensity was observed in KO (Fig. 4J, magnified panel), a result consistent with findings in western blot. Because Fkbp51 KO demonstrates a significant increase in Parkin expression and Parkin plays a critical role in mitochondrial function, identification of the co-localization of Parkin and FKBP51 could be an indication of their functional interaction within the cell. The HepG2 liver cell line was used for co-transfection experiments. Flag-FKBP51 was transfected into HepG2 cells and was detected using an anti-Flag antibody. As shown in Fig. 4K, FKBP51 was co-labeled with Parkin or MitoTracker; however, due to reduced expression of Parkin activity due to increased FKBP51, Flag-Fkbp51 and Parkin merged images show only weak Parkin labeling, a result consistent with our previously published paper (Fig. 4K) (Qiu et al., 2022). Our result is in line with studies that demonstrate FKBP51 presence in mitochondria and that FKBP51 and Parkin are found in a complex (Gallo et al., 2011; Qiu et al., 2022). In HepG2 cells, FKBP51 and Parkin were found to co-localize within mitochondria networks, with more intense labeling observed in the perinuclear area (Fig. 4K, middle and right). The co-localization of FKBP51 and Parkin in the mitochondria could partially explain how Fkbp51 affects mitochondrial function by serving as a Parkin regulator.
Ablation of FKBP51 increases mitochondrial size and induces autophagy/mitophagy in the liver.
To evaluate how the loss of FKBP51 affects mitochondrial function and Parkin-associated MQC at the subcellular level, mitochondrial morphology was assessed by electron microscopy (EM) (Vernucci et al., 2019). Interestingly, KO liver cells were initially observed to possess larger mitochondria compared to WT mouse livers both at baseline and after CCl4 treatment (Fig. 5A). Quantification of random representative microphotographs measuring 250–300 mitochondria revealed that KO possess significantly larger mitochondria than WT. Although CCl4-induces a decrease in mitochondrial size in both, the average size of mitochondrial size is still bigger in KO mice versus WT (Fig. 5B); however, no difference in mitochondria number was observed between KO and WT in either condition (Fig. 5C).
No obvious ER morphological difference was observed between KO and WT in control treatment (Fig. 5D). However, we observed more mitochondrial damage and ER expansion (arrow-head) with wrinkled and broken mitochondrial membranes (Figs. 6A-C) in WT mice, while the ER appeared normal in KO (Figs. 5D KO panel, 6D and E) after CCl4 treatment. The formation of mitochondrial-derived vesicles (MDVs) serves as a defense mechanism to remove harmful mitochondrial components and as a mechanism of immune tolerance and immune response (Popov, 2022; Towers et al., 2021). MDVs were observed both but more often in KO compared to WT mice after CCl4 treatment (Fig. 6C in WT vs Fig. 6E, F and K in KO). Laminated bodies (LB) and MDVs, to be released to the lysosome (L) were identified more prominently in KO mice (Fig. 6D, F, G, H, I, and L). Remarkably, more occurrences of mitophagy (Fig. 6J-K, red arrows) were observed in KO mice versus WT after CCl4 treatment. The enhanced whiter appearance of KO mitochondria was observed following CCl4 treatment which may link to it’s unique adaptation to insult (Fig. 5D and S-Fig. 4). These EM images provided morphological evidence that KO mitochondria were better protected from CCl4 injury in KO than in WT. To further identify molecular evidence to support these observations, protein-related to mitochondrial damage clearance (mitophagy) and MDV processes were studied.
Mitophagy is an autophagy process for the removal of damaged mitochondria and is regulated by several important proteins including Parkin, dynamin-related protein 1 (DRP1), and PINK1 (Scarffe et al., 2014). As shown above, a significant increase in Parkin expression was found in the cytoplasmic, mitochondrial, and total protein fractions of KO mice relative to WT (Fig. 4F-I). Relative to WT, KO exhibited higher p-DRP1/DRP1 ratios in both mitochondrial and total protein fractions, but no change in PINK1 expression was detected (Fig. 7A-D). KO livers demonstrated a reduction in levels of mitochondrial mitofusin 1 (MFN1) but increases in total MFN1 (Fig. 7A-D). Initialization of the mitophagosome requires microtubule-associated protein light chain 3B (LC3B) to degrade damaged mitochondria (Wei et al., 2017). LC3B protein was measured in livers from untreated and CCl4-treated WT and KO mice. The ratio of LC3BII/I significantly increased after CCl4 treatment, particularly in KO mice, indicating increased autophagosome activity (Fig. 7E-G). The in vivo and in vitro data support a critical role for FKBP51 affecting mitochondrial dynamics, including increased occurrences of mitophagy and MDVs and less ER swelling in KO mice. Enhanced autophagy/mitophagy may explain the amelioration of CCl4-induced liver injury in the KO liver (Vernucci et al., 2019).
Fkbp51 KO MEF cells demonstrate increased autophagy/mitophagy marker expression and reduced ROS after CCCP treatment.
To directly test whether FKBP51 plays a role in mitophagy, proton ionophore carbonyl cyanide m-chlorophenylhydrazone (CCCP), a mitochondrial uncoupler, was used to induce mitochondrial damage in primary cultured MEFs, and LC3B and LC3-binding protein ubiquitin-binding protein p62, an autophagy receptor, were measured to assess mitochondrial stress (Ichimura et al., 2008). Consistent with previous results, KO cells possess enhanced MitoTracker signal compared to WT, with no visual difference following DMSO control treatment (Fig. 8A). LC3B signals of both WT and KO MEFs were enhanced following CCCP treatment, with greater labeling intensity observed in KO MEFs. Merged images of MitoTracker and LC3B indicate greater mitophagy in KO MEFs, as signified by a stronger yellow signal (Fig. 8A). Western blotting data support that the ratio of total LC3II/I is significantly higher in KO MEFs compared to WT at baseline in the mitochondrial fraction and after CCCP treatment in both the mitochondrial and total protein fractions (Fig. 8B-D). Following CCCP treatment, p62 was increased in both fractions of KO and WT MEFs, though it was significantly higher in KO than in WT (Fig. 8B-D). Additionally, the mitochondrial p-DRP1/ DRP1 ratio was higher in KO MEFs before and after CCCP treatment but the total protein p-DRP1/ DRP1 ratio was lower in KO, suggesting that KO MEFs are more protected from CCCP-induced shifts in mitochondrial dynamics (Fig. 8B-D). The above data support an enhancement of autophagy/mitophagy function in Fkbp51 KO MEF cells, protective mechanisms for preventing mitochondrial damage.
Although we demonstrated that more apoptotic cells are present in CCl4-treated WT liver than in KO (Fig. 1H and I), a direct difference of apoptosis was assessed resulting from mitochondrial damage induced by CCCP. In addition to confirming the presence of fewer apoptotic cells in KO, a significant increase of apoptosis was found in WT cells after CCCP treatment (Fig. 8E). Additionally, a CCCP-induced increase in the level of ROS was detected solely in primary cultured WT MEFs as indicated by peak shifting (Fig. 8F). This provided indirect evidence that KO cells produce fewer ROS during mitochondrial insult, which could partially explain the protective effect observed in KO mice following CCl4 treatment. Thus, these data support better MQC in Fkbp51 KO through the regulation of mitophagy/autophagy processes, mitochondrial morphology and dynamics, and ATP production.
Inhibition of FKBP51 with SAFit2 increases Parkin and ameliorates CCl4-induced liver injury
We applied a highly specific selective FKBP51 inhibitor (SAFit2) to test its efficacy in preventing liver injury. SAFit2 has previously been applied in other studies and no toxicity was demonstrated with long-term treatment (Gaali et al., 2015; Hartmann et al., 2015; Konig et al., 2019). Concurrent injection of CCl4 and SAFit2 reduced indicators of liver injury, including lower intensity of trichrome staining (Fig. 9A) and lower quantified hepatic fibrosis scores (Fig. 9B). Serum AST and ALT levels were reduced in CCl4 + SAFit2-treated WT mice (Fig. 9C-D). Additionally, SAFit2 inhibition result in decreased IL6 and NFκB in serum which is consistent with KO mice has lower level of IL6 and NFκB, however, SAFit2 inhibition showed less effect on IFNγ (Fig. 9E). SAFit2 inhibition was associated with a commensurate increase in Parkin (Fig. 9F-G). We further studied mitochondrial functional alteration after FKBP51 inhibition by measuring ATP production, in vitro. WT and KO MEFs were treated with SAFit2, and ATP production was increased only in WT MEFs, KO MEFs already benefitting from a lack of FKBP51 function (Fig. 9H). The data support that inhibition of FKBP51 by SAFit2 prevents CCl4-induced liver injury, potentially via regulation of ATP production in the mitochondria.