Mitochondrial Dysfunction in the Liver and Antiphospholipid Antibody Production Precede Disease Onset and Respond to Rapamycin in Lupus-Prone Mice

Abstract

Objective: Antiphospholipid antibodies (aPL) constitute a diagnostic criterion of systemic lupus erythematosus (SLE), and aPL have been functionally linked to liver disease in patients with SLE. Since the mechanistic target of rapamycin (mTOR) is a regulator of oxidative stress, a pathophysiologic process that contributes to the development of aPL, this study was undertaken in a mouse model of SLE to examine the involvement of liver mitochondria in lupus pathogenesis.

Methods: Mitochondria were isolated from lupus-prone MRL/lpr, C57BL/6.lpr, and MRL mice, age-matched autoimmunity-resistant C57BL/6 mice as negative controls, and transaldolase-deficient mice, a strain that exhibits oxidative stress in the liver. Electron transport chain (ETC) activity was assessed using measurements of oxygen consumption. ETC proteins, which are regulators of mitochondrial homeostasis, and the mTOR complexes mTORC1 and mTORC2 were examined by Western blotting. Anticardiolipin (aCL) and anti-β₂ -glycoprotein I (anti-β₂ GPI) autoantibodies were measured by enzyme-linked immunosorbent assay in mice treated with rapamycin or mice treated with a solvent control.

Results: Mitochondrial oxygen consumption was increased in the livers of 4-week-old, disease-free MRL/lpr mice relative to age-matched controls. Levels of the mitophagy initiator dynamin-related protein 1 (Drp1) were depleted while the activity of mTORC1 was increased in MRL/lpr mice. In turn, mTORC2 activity was decreased in MRL and MRL/lpr mice. In addition, levels of aCL and anti-β₂ GPI were elevated preceding the development of nephritis in 4-week-old MRL, C57BL/6.lpr, and MRL/lpr mice. Transaldolase-deficient mice showed increased oxygen consumption, depletion of Drp1, activation of mTORC1, and elevated expression of NADH:ubiquinone oxidoreductase core subunit S3 (NDUFS3), a pro-oxidant subunit of ETC complex I, as well as increased production of aCL and anti-β₂ GPI autoantibodies. Treatment with rapamycin selectively blocked mTORC1 activation, NDUFS3 expression, and aPL production both in transaldolase-deficient mice and in lupus-prone mice.

Conclusion: In lupus-prone mice, mTORC1-dependent mitochondrial dysfunction contributes to the generation of aPL, suggesting that such mechanisms may represent a treatment target in patients with SLE.

Figure 1

O₂ consumption rates and respiratory control ratios at complex II of the electron transport chain (ETC) in mitochondria isolated from the livers of 4‐week‐old C57BL/6 (B6), C57BL/6.lpr (lpr), MRL, and MRL/lpr mice. A, Cumulative analyses of O₂ consumption rates through ETC complexes I, II, and IV. During the assay of each ETC complex, 150 μM ADP and 150 μM Pi was added to attain state 3 respiration, and when the ADP had been exhausted, state 4 respiration was attained. After achievement of a stable rate for state 4 respiration, 2 μM carbonylcyanide m‐chlorophenylhydrazone was added to measure uncoupled O₂ consumption, which is an indicator of maximal ETC capacity 22. In all experiments, O₂ consumption was normalized to that of the autoimmunity‐resistant B6 control strain (set as 1.0 for each ETC complex), which was studied in parallel with the lupus‐prone strains. B, Respiratory control ratio (state 3:state 4 respiration) at ETC complex II of mitochondria isolated from the livers of 4‐week‐old mice. C, Change in mitochondrial transmembrane potential (ΔΨm) over mitochondrial mass as a measure of mitochondrial hyperpolarization in the livers of MRL/lpr mice relative to the MRL and lpr parental strains. The ΔΨm was detected using a JC‐1 carbocyanine dye fluorescent probe, while mitochondrial mass was assessed using MitoTracker green (MTG) and nonylacridine orange (NAO) fluorescent probes. Results are the mean ± SEM in ≥4 mice per group. P < 0.05 versus B6 controls (∗) or between individual mouse strains (brackets), based on 2‐tailed unpaired t‐tests.

Figure 2

Increased expression of Rab4A and depletion of dynamin‐related protein 1 (Drp1) in MRL/lpr mice. Western blot analyses were performed to assess Rab4A expression (A) and the expression of Drp1, pDrp1^S616, and pDrp1^S637 (B) in the livers of 4‐week‐old C57BL/6 (B6), MRL, C57BL/6.lpr (lpr), and MRL/lpr mice. Left, Representative blots are shown. Right, Cumulative analyses of the fold change in expression relative to β‐actin are shown. Results are the mean ± SEM of 5 mice per strain. P values are versus B6 controls.

Figure 3

Increased activation of mechanistic target of rapamycin complex 1 (mTORC1) and diminished activation of mTORC2 in the livers of 4‐week‐old MRL/lpr mice. The mTORC1 and mTORC2 signature substrates 26, unphosphorylated S6 kinase (S6K) and phosphorylated S6K (pS6K^T389) and unphosphorylated Akt and phosphorylated Akt (pAkt^Ser473), were quantified by Western blotting, relative to β‐actin, in liver protein lysates from 4‐week‐old C57BL/6 (B6), C57BL/6.lpr (lpr), MRL, and MRL/lpr mice. Left, Representative blots are shown. Right, Cumulative analyses of the fold change in expression relative to β‐actin are shown. Results are the mean ± SEM in 5 mice per strain. P values are versus B6 controls.

Figure 4

Blockade of mechanistic target of rapamycin complex 1 (mTORC1) activity in the livers of MRL/lpr mice by treatment with rapamycin, administered in vivo at ages 4–14 weeks. Western blot analyses of the expression of unphosphorylated S6 kinase (S6K) and unphosphorylated Akt, as well as phosphorylated S6K (pS6K^T389) and phosphorylated Akt (pAkt^Ser473), were performed in mice at age 14 weeks, after treatment with rapamycin (n = 8) or with a solvent control of 0.2% carboxymethylcellulose (CMC; n = 3). Left, Representative Western blots are shown. Right, Cumulative analyses of the fold change in expression relative to β‐actin are shown. P values are versus CMC‐treated controls.

Figure 5

Increased production of anticardiolipin antibodies (ACLA; aCL) and anti–β₂‐glycoprotein I (anti‐β₂GPI) antibodies dependent on activation of mechanistic target of rapamycin complex 1 in lupus‐prone mice. A, Production of aCL and anti‐β₂GPI antibodies in 4‐week‐old MRL, C57BL/6.lpr (lpr), and MRL/lpr mice and 14‐week‐old MRL/lpr mice relative to C57BL/6 (B6) controls (n = 4–8 animals per strain). Results are the mean ± SEM fold change in OD at 630 nm relative to that in wells developed with secondary anti‐mouse antibodies alone and normalized to the values in B6 controls (set as 1.0). P < 0.05 versus B6 controls (∗) or between individual strains (horizontal lines), by 2‐tailed t‐test. B, Blockade of the production of aCL and anti‐β₂GPI antibodies by rapamycin (RAPA) treatment in MRL/lpr mice. Mice were treated 3 times weekly with intraperitoneal injections of 0.2% carboxymethylcellulose (CMC) (a solvent control for rapamycin; n = 3) or 1 mg/kg rapamycin (n = 8). Treatment was started at age 4 weeks and antibody production was tested at age 14 weeks. Results are the mean ± SEM fold change in OD at 630 nm relative to CMC‐treated control MRL/lpr mice (set as 1.0). ∗ = P < 0.05 versus CMC‐treated controls, by 2‐tailed t‐test.

Figure 6

Reversal of dynamin‐related protein 1 (Drp1) depletion and selective blockade of the expression of NADH:ubiquinone oxidoreductase core subunit S3 (NDUFS3) of electron transport chain (ETC) complex I by rapamycin (Rapa) treatment in the liver mitochondria of MRL/lpr mice. Female mice were treated 3 times weekly with intraperitoneal injections of 0.2% carboxymethylcellulose (CMC) (a solvent control for rapamycin; n = 3) or 1 mg/kg rapamycin (n = 8). Treatment was begun at age 4 weeks and continued through age 14 weeks. Western blot analyses (left) and cumulative analyses of the fold change in expression relative to β‐actin (right) were performed to assess the expression of Rab4A, Drp1, pDrp1^S616, and pDrp1^S637 (A) and ETC complex I (subunits NDUFS3 and NDUFS1), ETC complex II (succinate dehydrogenase complex flavoprotein subunit A [SDHA]), and ETC complex IV (mitochondrial cytochrome c oxidase subunit 1) (B). P values are versus CMC‐treated controls, by unpaired 2‐tailed t‐test.