Mesencephalic astrocyte-derived neurotrophic factor is an ER-resident chaperone that protects against reductive stress in the heart

Abstract

We have previously demonstrated that ischemia/reperfusion (I/R) impairs endoplasmic reticulum (ER)-based protein folding in the heart and thereby activates an unfolded protein response sensor and effector, activated transcription factor 6α (ATF6). ATF6 then induces mesencephalic astrocyte-derived neurotrophic factor (MANF), an ER-resident protein with no known structural homologs and unclear ER function. To determine MANF's function in the heart in vivo, here we developed a cardiomyocyte-specific MANF-knockdown mouse model. MANF knockdown increased cardiac damage after I/R, which was reversed by AAV9-mediated ectopic MANF expression. Mechanistically, MANF knockdown in cultured neonatal rat ventricular myocytes (NRVMs) impaired protein folding in the ER and cardiomyocyte viability during simulated I/R. However, this was not due to MANF-mediated protection from reactive oxygen species generated during reperfusion. Because I/R impairs oxygen-dependent ER protein disulfide formation and such impairment can be caused by reductive stress in the ER, we examined the effects of the reductive ER stressor DTT. MANF knockdown in NRVMs increased cell death from DTT-mediated reductive ER stress, but not from nonreductive ER stresses caused by thapsigargin-mediated ER Ca²⁺ depletion or tunicamycin-mediated inhibition of ER protein glycosylation. In vitro, recombinant MANF exhibited chaperone activity that depended on its conserved cysteine residues. Moreover, in cells, MANF bound to a model ER protein exhibiting improper disulfide bond formation during reductive ER stress but did not bind to this protein during nonreductive ER stress. We conclude that MANF is an ER chaperone that enhances protein folding and myocyte viability during reductive ER stress.

Keywords: cardiomyocyte; chaperone; endoplasmic reticulum stress (ER stress); heart function; ischemia; ischemia/reperfusion; mesencephalic astrocyte-derived neurotrophic factor (MANF); protein folding; reductive stress; unfolded protein response.

Figure 1.

Effect of MANF knockdown in the heart on expression of GRP94 and GRP78, expression of fetal genes, heart and lung weights, viability of isolated adult mouse ventricular myocytes, and cardiac contractility. A, MANF, GRP94, GRP78, and β-actin immunoblots of mouse heart extracts from 10-week-old female WT (n = 5) or MANF KD mice (n = 5). *, band of interest that was quantified in B–D. IgG_H indicates position of immunoglobulin heavy chain. B–D, densitometry of immunoblots shown in A. Band intensities were normalized to those for β-actin and displayed as -fold WT control level. E–G, fetal gene mRNA levels were determined by RT-qPCR. H and I, heart weights were normalized to tibia length (H) or body weight (I). J, lung weights were normalized to body weight. Echocardiography data and statistical analysis can be found in Table 1. K, myocytes were isolated from 10-week-old adult WT and MANF KD mice and then subjected to sI/R followed by determination of cell viability. To determine percentage cell viability, the number of calcein AM–positive cells per field was divided by the total number of cells in the same field. L–P, ex vivo I/R of WT and MANF KD mouse hearts. Hearts from female WT (n = 3) or MANF KD (n = 4) were subjected to ex vivo ischemia for 20 min, followed by 60 min of reperfusion (I/R). L and M, LVDP upon reperfusion was normalized to the LVDP obtained during equilibration, the latter of which was set to 100%. L, plot of individual LVDP time courses from different mouse hearts. M, average of the plots shown in L. N and O, heart sections were stained with TTC to assess the extent of myocardial damage (N); shown is the average infarct size divided by area at risk (O). P, samples of ex vivo heart perfusates were obtained after 45 min of reperfusion and then assayed for LDH activity relative to LDH activity in the equilibrium perfusate. *, statistically significant difference by Student's unpaired t test, p ≤ 0.05. Note that GRP78 and GRP94 immunoblotting was performed using an anti-KDEL antibody. Error bars, S.E.

Figure 2.

Effect of MANF re-expression in MANF KD mouse hearts on cardiac damage and contractility following I/R. AAV9-Con or AAV9-FLAG-MANF was administered to WT and MANF KD mice by tail vein injection. Seven days later, male hearts were extracted and subjected to MANF and β-actin immunoblotting (A), and female hearts were subjected to ex vivo I/R (B–F). For ex vivo I/R, hearts from WT mice injected with AAV-Con (n = 3) or MANF KD mice injected with AAV-Con (n = 4) or AAV-FLAG-MANF (n = 4) were subjected to 20 min of ex vivo global ischemia and then 60 min of reperfusion. B and C, LVDP upon reperfusion was normalized to the LVDP obtained during equilibration, the latter of which was set to 100%. B, plot of individual LVDP time courses from different mouse hearts. C, average of the plots shown in B. D and E, heart sections were stained with TTC to assess myocardial damage (D); shown is the average infarct size divided by area at risk (E). F, samples of perfusate were obtained after 45 min of reperfusion to assess LDH activity relative to LDH activity in the equilibrium perfusate. *, statistically significant difference from all other groups by two-way ANOVA followed by Tukey's post hoc analysis, p ≤ 0.05. These experiments were performed twice using separate cohorts of mice. Error bars, S.E.

Figure 3.

Effect of sI and sI/R on MANF expression and the effect of MANF knockdown on myocyte death, ROS generation, and expression of ER stress markers during sI/R. A, immunoblots of MANF and GAPDH from NRVMs subjected to 8 h of sI or 8 h of sI followed by 24 h of simulated reperfusion (sI/R). Band intensities were normalized to those for GAPDH and displayed as -fold control level. *, statistically significant difference from control by Student's unpaired t test, p ≤ 0.05. #, statistically significant difference from all other groups by one-way ANOVA followed by Newman–Keuls post hoc analysis. B, immunoblots of MANF and GAPDH from NRVMs transfected with siCon or siManf demonstrating MANF knockdown. Band intensities were normalized to those for GAPDH and displayed as -fold control level. *, statistically significant difference by Student's unpaired t test, p ≤ 0.05. C and D, siCon- or siManf-transfected NRVMs were subjected to 12 h of sI (C) or 6 h of sI followed by 24 h of reperfusion (sI/R) (D), and percentage cell death was assessed by staining cell cultures with Hoechst 33342 and PI and dividing the number of PI-positive cells by the number of Hoechst-positive cells in a given field. E, HMGB1 immunoblots of the culture medium indicating cell death and immunoblots of GRP94, GRP78, PDIA6, and GAPDH from NRVMs transfected with siCon or siManf and subjected to sI/R. F, densitometry of HMGB1 immunoblot shown in E. G, siCon- or siManf-transfected NRVMs were subjected to sI/R followed by ROS measurement with CellROX. H, densitometry of GRP94, GRP78, PDIA6, and GAPDH immunoblots shown in E. Band intensities of GRP94, GRP78, and PDIA6 were normalized to those for GAPDH and displayed as -fold control level. * and #, statistically significant difference from all other groups by two-way ANOVA followed by Tukey's post hoc analysis, p ≤ 0.05. Note that GRP78, GRP94, and PDIA6 immunoblots were performed using an anti-KDEL antibody. Error bars, S.E.

Figure 4.

Effect of MANF loss of function and pharmacological ER stressors on myocyte viability, effect of TG or DTT on FLAG-MANF and α1AT co-immunoprecipitation, and effect of rMANF on protein aggregation and folding. A–D, NRVMs were transfected with siCon or siManf, and after 72 h, cells were treated with H₂O₂ (A), DTT (B), TG (C), TM (D), or vehicle. Viability was then assessed by an MTT assay. * and #, statistically significant difference from all other groups by two-way ANOVA followed by Tukey's post hoc analysis, p ≤ 0.05. E, NRVM cultures were co-infected with adenoviruses encoding FLAG-MANF and α1AT ΔCT, as shown, and treated with TG or DTT for 1 h. Cell extracts were subjected to SDS-PAGE followed by immunoblotting for FLAG or α1AT ΔCT. The cell extracts were also subjected to FLAG IP followed by SDS-PAGE and then IB for FLAG or α1AT ΔCT, as shown. F, HeLa cell cultures were co-transfected with plasmid constructs encoding FLAG-MANF and/or α1AT-HA ΔCT, as shown. The cell extracts were subjected to SDS-PAGE followed by immunoblotting for HA or FLAG. The cell extracts were also subjected to FLAG IP followed by SDS-PAGE and then IB for FLAG or HA, as shown. G, effect of rMANF at low (7.8 μm) or high (23.4 μm) concentrations on the aggregation of insulin (113 μm). H, effect of rMANF at low (14 μm) or high (28 μm) concentrations on the aggregation of α-lactalbumin (14 μm). I, effect of recombinant GRP78 (1 μm) or MANF (1 μm) on activity of heat-denatured citrate synthase (1 μm). Citrate synthase activity is displayed as fold heat-denatured control level. * and #, statistically significant difference from all other groups by one-way ANOVA, p ≤ 0.05, followed by Newman–Keuls post hoc analysis. Error bars, S.E.

Figure 5.

Sequence homology between MANF and known chaperone and co-chaperone orthologs or homologues. A and B, Clustal sequence identity analysis of MANF and various chaperones (A) and co-chaperones (B). Error bars, 95% confidence intervals.

Figure 6.

Effect of mutation of the conserved cysteine residues in MANF on MANF redox status and chaperone function. A, alignment of MANF sequences from different species. Highlighted in yellow are the positions of cysteine residues, the positions of which are conserved across the species shown. B, diagram of FLAG-MANF_WT and FLAG-MANF_Mut. constructs indicating cysteine-to-alanine mutations. C and D, immunocytofluorescence of FLAG-MANF. NRVMs were transfected with siManf targeted to the 3′-UTR of the Manf transcript, followed by infection with AdV-FLAG-MANF_WT (C) or AdV-FLAG-MANF_Mut. (D) and then treated with tunicamycin to induce GRP78 expression. NRVMs were then examined by immunocytofluorescence for GRP78 (green) and FLAG-MANF (red). Nuclei are indicated by TOPRO staining (blue). E, duplicate HeLa cell cultures were co-transfected with plasmid constructs encoding FLAG-MANF_WT or FLAG-MANF_Mut. and α1AT-HA ΔCT and treated with DTT. The cell extracts were subjected to reducing SDS-PAGE followed by immunoblotting for α1AT-HA ΔCT (∼45 kDa) or FLAG-MANF (∼20 kDa). The cell extracts were also subjected to FLAG IP followed by nonreducing SDS-PAGE to maintain possible disulfide bonds between MANF and other proteins and then IB HA (bottom). (Note the FLAG-MANF/α1AT-HA ΔCT complex shown at ∼65 kDa). F, effect of recombinant MANF_WT (23.4 μm) or MANF_Mut. (23.4 μm) on aggregation of insulin (113 μm). G, effect of recombinant MANF_WT (21 μm) or MANF_Mut. (21 μm) on aggregation of α-lactalbumin (14 μm). * and #, statistically significant difference from all other groups by one-way ANOVA, p ≤ 0.05, followed by Newman–Keuls post hoc analysis. H, diagram depicting the function and mechanism of action of endogenous MANF in the heart resulting from this study. Error bars, S.E.