Lipofuscin is formed independently of macroautophagy and lysosomal activity in stress-induced prematurely senescent human fibroblasts

Free Radical Biology and Medicine 53 (2012) 1760–1769

Contents lists available at SciVerse ScienceDirect

Free Radical Biology and Medicine journal homepage: www.elsevier.com/locate/freeradbiomed

Original Contribution

Lipofuscin is formed independently of macroautophagy and lysosomal activity in stress-induced prematurely senescent human fibroblasts a,b ¨ Annika Hohn , Anna Sittig b, Tobias Jung a, Stefanie Grimm a, Tilman Grune a,n a b

Department of Nutritional Toxicology, Institute of Nutrition, Friedrich Schiller University Jena, 07743 Jena, Germany Department of Biofunctionality and Food Safety, Institute of Biological Chemistry and Nutrition, University of Hohenheim, Stuttgart, Germany

a r t i c l e i n f o

a b s t r a c t

Article history: Received 10 June 2012 Received in revised form 17 August 2012 Accepted 27 August 2012 Available online 1 September 2012

In the current literature, the lysosomal system is considered to be involved in the intracellular formation and accumulation of lipofuscin, a highly oxidized and covalently cross-linked aggregate of proteins that fills the lysosomal volume during aging. In contrast, our experimental results presented here suggest that both the autophagosomes and the lysosomal system are not mandatory for the formation of lipofuscin, since that material accumulates in the cytosolic volume if autophagy or lysosomal activity is inhibited. However, such an inhibition is accompanied by an enhanced toxicity of the formed protein aggregates. Furthermore, it could be proven that macroautophagy is responsible for the uptake of lipofuscin into the lysosomes. & 2012 Elsevier Inc. All rights reserved.

Keywords: Aging Oxidative stress Lipofuscin Macroautophagy Lysosome Stress-induced premature senescence (SIPS)

Introduction The intracellular accumulation of lipofuscin is one of the limiting factors in postmitotic cell aging. Understanding the basic mechanisms of lipofuscin formation and accumulation can contribute to ensuring the quality of life in old age and may identify ways to reduce the accumulation rate of protein aggregates in cells. Mammalian cells have several mechanisms for removing proteins when they are misfolded, become oxidized, or are damaged. One of the main degradation systems is the proteasome, a multisubunit protease. However, the proteasome only degrades unfolded monomeric proteins, so it cannot remove covalently cross-linked proteins from the cytosol [1,2]. The other basic degradation system is autophagy, which includes macro-, micro-, and chaperone-mediated autophagy. Macroautophagy functions as a stress response which is upregulated by oxidative stress, starvation, or other adverse conditions

Abbreviations: ATG/Atg, autophagy-related gene; ATP6V0A1, encoding gene of isoform a1 of vacuolar H þ -ATase V0 domain; H2DCFDA, 20 ,70 -dichlorodihydrofluorescein diacetate; LAMP1, lysosomal-associated membrane protein 1; LC3, microtubule-associated-protein-light-chain-3; LF, lipofuscin; MDC, monodansylcadaverine; MEFs, mouse embryonic fibroblasts; NH4Cl, ammonium chloride; PQ, paraquat; SIPS, stress-induced premature senescence. n Corresponding author. fax: þ 49 3641949672. ¨ E-mail addresses: [email protected] (A. Hohn), [email protected] (A. Sittig), [email protected] (T. Jung), [email protected] (S. Grimm), [email protected] (T. Grune). 0891-5849/$ - see front matter & 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.freeradbiomed.2012.08.591

and is controlled by a specific set of ATG genes and their protein products (Atg). One of the discussed options for disposal of aggregated proteins is the uptake into the lysosomes via macroautophagy. In 2007 the term ’’aggrephagy’’ was introduced [3]. This form of macroautophagy seems to be relevant for the uptake of large protein aggregates or lipofuscin particles, which were released into the cytosol due to rupture of lysosomes. The hallmark of macroautophagy is the formation of doublemembrane vesicles called autophagosomes that sequesters portions of the cells cytoplasm and delivers them to the lysosome. The formation of autophagosomes in mammalian cells requires two processes: Atg12 conjugation and LC3 (microtubule-associated-protein-light-chain-3) modification. Atg12 is ligated to Atg5 in a process assisted by the enzymes Atg7 and Atg10. The Atg12–Atg5 complex then forms larger oligomers with Atg16. This Atg12–Atg5–Atg16 complex is essential for elongation of the isolation membranes and a key regulator of the autophagic process [4]. LC3 also assists autophagosome formation, possibly by enhancing membrane fusion [5]. The soluble cytosolic LC3I becomes ligated to the lipid phosphatidyl ethanolamine in reactions assisted by Atg4, Atg7, and Atg3. The lipidated form is called LC3II and anchors to autophagosomal membranes. While the role of proteasome inhibition as a cause of agerelated increases in protein oxidation and increased lipofuscin amounts is established, the contribution of the lysosomal system and the role of macroautophagy are less clear. Genetically determined disorders of lysosomal degradation lead to severe

A. H¨ ohn et al. / Free Radical Biology and Medicine 53 (2012) 1760–1769

1761

defects with an extreme overload of cells with undegraded material [6], but also during aging a progressive decline of lysosomal function is postulated. Age-related impairments in lysosomal proteolysis including dysfunctional regulation of pH, impaired lysosomal stability, and targeting of proteins are observed in many cell types [7,8]. So, over time defective proteins and organelles as well as lipofuscin slowly but progressive accumulate especially in postmitotic aging cells. Our aim was to investigate the role of macroautophagy and the lysosomal system during aging in the formation and accumulation of lipofuscin. We used ATP6V0A1 siRNA for blocking of the lysosomal proton pump to interfere with the functionality of the lysosomes in an established ’’stress-induced premature senescence’’ (SIPS) model of human fibroblasts. SIPS is a model for inducing characteristic features of aged cells in young or ‘‘middleaged’’ cells by application of chronic oxidative stress [9]. In a next step we inhibited autophagic sequestration of protein aggregates and performed further experiments with a mouse embryonic fibroblast (MEF) ATG5–/– cell line and human dermal fibroblasts, treated with ATG5 siRNA.

protein concentrations of the supernatants were determined according to the Lowry method. The amount of 15–30 mg of total protein in reducing Laemmli buffer (0.25 M Tris (pH 6.8), 8% SDS, 40% glycerol, 0.03% Orange G) was denatured at 95 1C for 5 min and applied to SDS-PAGE of 12% (w/v) acrylamide followed by electrophoresis and blotting onto nitrocellulose membrane according to standard procedures. Immunodetections were performed with the following antibodies at dilutions recommended by the suppliers: mouse monoclonal LC3 antibody, mouse monoclonal Atg5 antibody (Nano Tools, Teningen, Germany), mouse monoclonal LAMP1 antibody, rabbit polyclonal GAPDH antibody, mouse monoclonal p62/ SQSTM1 antibody, and rabbit polyclonal ATP6V0A1 antibody (ABCAM, Cambridge, UK). The blocking buffer and fluorescentconjugated secondary antibodies were purchased from Li-Cor Biosciences (Lincoln, NE) and used according to protocols supplied by the manufacturer. The membranes were scanned and stained bands were quantified using an Odyssey Infrared Imaging System (Li-Cor Biosciences) according to the manufacturer’s instructions.

Materials and methods

Determination of cell viability using MTT

Reagents

Cell viability was determined using a standard tetrazolium salt assay [12]. The absorbance was measured at 590 nm with a microplate reader (Synergy 2, BioTek, Bad Friedrichshall, Germany).

All chemicals were obtained from Sigma (Deisenhofen, Germany). Cell culture materials were purchased from Biochrom (Berlin, Germany) unless otherwise indicated. Cell culture

Measurement of autophagic acitivity using MDC

Experiments were performed using human dermal fibroblasts obtained from skin tissue. The fibroblasts were a kind gift of Prof. Scharffetter-Kochanek (University of Ulm, Germany). Informed consent of the donor was achieved according to international rules. Fibroblasts were cultured in DMEM supplemented with 10% fetal calf serum (FCS), and 1% Glutamax. Mouse embryonic fibroblasts from WT and ATG5–/– embryos [10] were obtained from the RIKEN BRC cell bank (Tsubuka, Ibaraki, Japan) and maintained in DMEM with 10% FCS.

Cells were incubated with 0.05 mM MDC at 37 1C for 10 min [13]. Afterward cells were washed three times with PBS and collected in 10 mM Tris-HCl, pH 8, containing 0.1% Triton X-100. Intracellular MDC was measured using a fluorescence reader (Synergy 2, BioTek) at 360 nm excitation/530 nm emission. For normalization to the number of cells present in each well, DNA was stained with ethidium bromide at a final concentration of 0.2 mM and the DNA fluorescence was measured at 530 nm excitation/590 nm emission.

RNA interference Pooled small interfering RNA (siRNA) oligonucleotides against ATG5 and ATP6V0A1, as well as a control siRNA (nontargeting pool), were purchased from Thermo Scientific Dharmacon (Lafayette, CO, USA). Efficacy and specificity of knockdown were assessed by immunoblotting. Cells were transfected at 70% confluence with 50 nM of the pooled oligonucleotide mixture by using Dharmafect transfection reagent following the manufacturer’s protocols. After removal of the transfection media cells were allowed to recover for 10 h before further treatment. Stress-induced premature senescence (SIPS) SIPS was achieved by paraquat (PQ) treatment. The explicit procedure is described in Ref. [11]. PQ was applied to confluent cells for 10 days daily (3.0 mM per 106 cells). For further treatment (transfection, concanamycin A) PQ-pretreated cells were trypsinized and seeded subconfluent at Day 5 or 7 of PQ treatment. Immunoblot analysis Cells were lysed at 4 1C using 10 mM Tris-HCl (pH 7.5) buffer containing 1 mM Pefabloc, 0.9% Nonidet P-40, 0.1% SDS. The

Visualization of MDC-stained autophagosomes and autolysosomes Cells were incubated with 0.05 mM MDC at 37 1C for 20 min. After incubation cells were washed three times with PBS and fluorescence was microscopically investigated (Zeiss Axiovert 100 M running standard software, Zeiss, Jena, Germany) using an excitation filter of 360 nm and an emission filter of LP 420 nm.

Lipofuscin detection Lipofuscin was detected via its autofluorescence using fluorescence microscopy with an excitation wavelength of 360 nm and an emission filter of 420 nm.

Visualization of the lysosomal system and its colocalization with lipofuscin Lysosomes were labeled using ‘‘LysoTracker Blue’’ (Invitrogen, Karlsruhe, Germany) and according to the manufacturer’s instructions. After bleaching away the LysoTracker signal the autofluorescence of lipofuscin remained. The detailed procedure can be found in [14].

1762

A. H¨ ohn et al. / Free Radical Biology and Medicine 53 (2012) 1760–1769

Visualization of lysosomes via LAMP1 immunocytochemistry

Results

Cells were fixed with a mixture of diethylether and ethanol (1:1, v/v) for 8 min at a temperature of  20 1C. Afterward cells were washed using a PBS buffer containing 1% of fetal bovine serum (washing buffer), followed by an incubation for 30 min at 4 1C in this buffer to block unspecific binding sites. After this procedure, the LAMP1 antibody was applied (diluted 1:50 in washing buffer) for 2 h at 4 1C. Subsequently cells were washed followed by incubation with a secondary fluorescence-labeled antibody (FITC) in a dilution of 1:100 for 1 h at 4 1C, followed by another washing step and immediate fluorescence microscopical investigation.

Effects of PQ treatment on the lysosomal and autophagosomal system

Free radical detection via 20 ,70 -dichlorodihydrofluorescein diacetate (H2DCFDA) Free radicals were detected using the membrane-permeable ROS detecting agent H2DCFDA. At the end of each experiment cells were incubated at a final concentration of 5 mM H2DCFDA for 1 h. For detailed information see [11]. Quantitative analysis of fluorescence microscopic signals The intensity of fluorescence was analyzed using the Corel Photo-Paint X5 software from Corel Corporation. Regions of interest in the images were highlighted with Brush Mask Tool and evaluated via the histogram function. The exposure time was standardized and the means of the intensities were backgroundcorrected by subtracting mean intensity from reference regions outside the cells. Area and pixel intensity of the LysoTrackerlabeled lysosomes were determined using image analysis functions of Corel Photo Paint X5. The number of lysosomes per cell was determined manually after labeling with LysoTracker Blue. Ratio of lysosomal to whole cell area was determined after manually masking via image analysis functions of Corel PhotoPaint X5. After subtracting the autofluorescence from the image showing both autofluorescence and LysoTracker, the now separated fluorescences were colored artificially: green for LysoTracker and red for autofluorescence and overlayed again. The green parts of the overlay image were defined as lipofuscin-free lysosomes, the red parts were defined as free cytosolic lipofuscin, and the yellow parts as lysosomes containing lipofuscin. The procedure was the same with LAMP1-stained lysosomes, except that images of lipofuscin and lysosomes were taken in different channels (lysosomes were detected in the FITC channel and lipofuscin in the DAPI channel). The total amount of lipofuscin (lysosomal and cytosolic portion) was calculated via area and pixel intensity in relation to the particular cell area. The ratio of autophagosome area per whole cell area after MDC staining was determined after manually masking via image analysis functions of Corel Photo-Paint X5. Statistics Statistical analysis was performed using Prism 5.04 (GraphPad) software. The data presented in all figures are the mean values7SD from three different experiments. For microscopy at least 50 cells of each group per experimental approach were evaluated concerning signal intensity, number of lysosomes, and signal area. Differences between two groups were assessed by Student’s t test after Gaussian distribution of the data was verified. A P value of less than 0.05 was selected as the level of significance. For multiple comparisons one-way ANOVA was performed followed by Tukey’s post hoc test. Values of Po0.05 were accepted as significant.

Stress-induced premature senescence of human dermal fibroblasts induced an increase in LysoTracker dying intensity of 20% (after 5 days), 50% (after 7 days), to 160% (after 10 days). The number of lipofuscin-loaded lysosomes per cell increases significantly from about 270 for the control to 700 after 10 days of PQ treatment; furthermore, the percentage area of lysosomes in the microscopic image more than doubles from 25 to 53%. In both cases neither the number nor the area of lipofuscin-free lysosomes changed (Fig. 1A). Also no significant change in free cytosolic lipofuscin was detected (Fig. 1B). So all lipofuscin formed during SIPS is located in lysosomes. Whereas the pool of lipofuscin-free lysosomes remains constant, the number of lipofuscin-loaded lysosomes increased dramatically. This was confirmed by a nearly linear increase (by up to 110%) of LAMP1, a common used lysosomal marker. The concentration of free Atg5 did not change while the concentration of the Atg5–Atg12 protein complex increased to 80%. LC3II expression was only found to be increased significantly after 10 days of PQ treatment (140%) compared to control cells defined as 100% (Fig. 1C and D). Representative fluorescence microscopic images are shown in supplemental Fig. S1. The enhanced number of lipofuscin-loaded lysosomes is clearly visible at Day 10. Using another fluorescent probe, monodansylcadaverine (MDC), which marks autophagosomes and also autolysosomes, we could show that their number and area also increase in good correlation with the results of the LysoTracker experiments. MDC-mediated fluorescence increased to 235% after 10 days of PQ treatment compared to control cells defined as 100%. The autophagosomal area per cell area increased from 20 to 56% during PQ exposure (Fig. 2). siRNA mediated knockdown of ATP6V0A1 in human dermal fibroblasts After locating the SIPS-induced lipofuscin in the autophagosomal/lysosomal compartment we tested a role for lysosomal acidification in the process of lipofuscin accumulation. For alkalinization of lysosomes we decided to knockdown the encoding gene ATP6V0A1 of isoform a1 of the integral V0 domain of vacuolar H þ -ATPase. In a first step we tested the knockdown efficiency of ATP6V0A1 and the viability of human fibroblasts. ATP6V0A1 was significantly knocked down to about 22% remaining expression (3 days after transfection) and 31% (5 days after transfection) compared to untransfected control or negative control (nontargeted siRNA transfection) cells, respectively. In any case cells showed a viability of at least 77% (supplemental Fig. 2). Staining of lysosomes with LysoTracker was not possible, due to its ability to stain acidic compartments. Also this neutralization leads to swift loss of MDC staining and/or lack of MDC uptake. Therefore lysosomes were defined as organelles that contain the marker LAMP1 and were stained via immunocytochemistry. The number of lipofuscin-loaded lysosomes did not change in response to ATP6V0A1 knockdown compared to the PQ exposure alone, but the area significantly increased (Fig. 3A). Lipofuscin-free lysosomes did not significantly change, and neither did the amount of nonlysosomal lipofuscin (Fig. 3B). Quantification of lysosomal and autophagy-related marker proteins revealed no change due to loss of acidification compared to the SIPS control, whereas LC3II expression increased to 65% compared to the SIPS cells (Fig. 3C and D). Therefore, we concluded that lysosomal acidification is not important for lipofuscin formation and accumulation. For verification of the data and to further reveal the influence of lysosomal pH in protein aggregate formation we also used concanamycin A (at a final

A. H¨ ohn et al. / Free Radical Biology and Medicine 53 (2012) 1760–1769

80

LF loaded LF free

*#+

700 Number of lysosomes per cell [N]

LysoTracker intensity [%]

250

200 * 150

100

70

600 500

0.6

LF loaded LF free

*

400 300 200

0.5 *# +

60 50 * 40 *

30 20

50

250 *

Protein expression [%]

200

150

0.2

0.0

0 7 10 5 Days of PQ treatment

C.

C.

10 5 7 Days of PQ treatment

Control PQ treatment 5 days PQ treatment 7 days PQtreatment 10 days

C. 5 7 10 Days of PQ treatment

LAMP1 (120 kDa) GAPDH (40 kDa) *

*

Atg5-Atg12 (55 kDa)

*

*

0.3

10

0 C. 5 7 10 Days of PQ treatment

0.4

0.1

100 0

Non-lysosomal lipofuscin [%]

*

Area of lysosomes per cell area [%]

800

300

1763

*

*

Atg5 (33 kDa) GAPDH (40 kDa)

100

GAPDH (40 kDa) 50

LC3I (18 kDa) LC3II (16 kDa) C.

0 LAMP1

Atg5

Atg5-Atg12

LC3II

5 days PQ

7 days PQ

10 days PQ

Fig. 1. Quantification of lysosomes, lipofuscin, and both lysosomal and autophagy-related marker proteins in human dermal fibroblasts during PQ-induced SIPS. The signal intensity of LysoTracker is demonstrated in the left part of Panel A. Control (C., black column) is set to 100%. The middle part shows the number of counted lysosomes per cell, and the signal area is shown in the right part. In both parts it is distinguished between lipofuscin-free (light gray columns) and lipofuscin-loaded lysosomes (dark-gray columns). (B) The portion of free cytosolic lipofuscin in relation to the whole cellular amount is shown. (C) Relative protein expression of lysosomal and autophagy-related proteins is demonstrated. Protein expression of untreated cells is set to 100% (black column). (D) Representative immunoblots for Panel C. Statistical significance compared to the control is indicated by ‘n’, compared to treatment Days 5 and 10 by a ‘#’. ‘þ ’ indicates statistical significance between treatment Day 7 and Day 10, one-way ANOVA, P o0.05 (A). ANOVA was also applied to the data of Panel B., but no significant differences were returned. In Panel C, statistical significance is indicated by asterisks (Student’s t test, P o 0.05).

concentration of 250 nM), a specific inhibitor of the V-ATPase, to reduce again the formation of a pH gradient (data not shown). We obtained analogous results with chemical blocking of V-ATPase activity, for all analyzed parameters. Effects of PQ treatment on ATG þ / þ and ATG–/– mouse embryonic fibroblasts (MEFs) Since lipofuscin is almost exclusively located in the lysosomal compartment, we decided to block macroautophagy as the most probable mechanism for the uptake of protein aggregates into the endosomal-lysosomal compartment. Since Atg5 is essential for the macroautophagy process, Atg5–/– cells should lack this pathway completely. As in human fibroblasts in ATG5 þ / þ MEFs an increase of LysoTracker intensity to 228% (Fig. 4A) was found during SIPS. This effect was completely absent in ATG5–/– MEFs. A 10-day PQ treatment did not give any indication of an increase in LysoTracker intensity or number and area of lysosomes. Also an

effect on lipofuscin-free lysosomes could not be found in either wild-type or knockout MEFs (Fig. 4A). However, the free cytosolic lipofuscin does increase more than twofold in the case of ATG5–/– cells (from 0.51% for control cells to 1.24% after 10 days of PQ treatment) (Fig. 4B). The cellular expression of autophagy-related marker proteins in response to PQ treatment in the ATG5 þ / þ MEFs was comparable to the results of the human fibroblasts (see Fig. 1C and D), except of the LAMP1 signal which increased to a lower extent. In contrast the ATG5–/– MEFs naturally show no Atg5 expression and thus no formation of the Atg12–Atg5 complex. For LAMP1 no significant increase could be found; LC3 I shows a considerable and stable signal for all samples. Since ATG5–/– do not have a LC3II signal, because of a lack of the LC3lipidation machinery, we used an alternative method for detection of autophagic flux, and determined the levels of p62/SQSTM1. Our results revealed that the basal protein level of p62/SQSTM1 was increased in ATG5–/– compared to the wild-type MEFs. Also we found a decrease among samples for the wild-type MEFs,

1764

A. H¨ ohn et al. / Free Radical Biology and Medicine 53 (2012) 1760–1769

80 *

250 * 200 * 150 100 50

Autophagosome area per cell area [%]

MDC specific fluorescence [%]

300

Control

5 days PQ

7 days PQ

10 days PQ

70 *

60 50 *

*

40 30 20 10 0

0 C. 5 7 10 Days of PQ treatment

C. 5 7 10 Days of PQ treatment

Fig. 2. MDC-labeled autophagosomes and autolysosomes after SIPS. (A, left part) The MDC-specific activity. Intracellular MDC was measured with a fluorescence reader. The ratio of autophagosomal/autolysosomal area per whole cell area was determined by microscopy after manually masking via image analysis functions of Corel PhotoPaint X5 (A, right part). (B) Representative images of MDC-labeled cells are shown. Treatment of cells is denoted in each image. The control in A (left part) was set to 100%. nP o 0.05 versus control (C.), Student’s t test.

indicating an increase in autophagy. For the ATG5–/– MEFs the level was unaltered, indicating that p62 is a good indicator of autophagy suppression (Fig. 4C). As in human fibroblasts, also in MEFs the MDC fluorescence was strongly increased in the case of ATG5 þ / þ , whereas the ATG5–/– mutants did not show any increase. The same could be detected for the area of autophagosomes per cell area (supplemental Fig. S3).

siRNA mediated knockdown of ATG5 in human dermal fibroblasts Using the ATG5–/– MEFs we found a clear indication that the uptake of protein aggregates and the lysosomal localization of lipofuscin depend on macroautophagy. Since the effect of ATG5 knockdown most dramatically blocked macroautophagy-related SIPS-induced lysosomal lipofuscin accumulation in MEFs we decided to use siRNA to knockdown ATG5 also in human fibroblasts. First we tested the efficiency of ATG5 knockdown and the viability of human fibroblasts. ATG5 was significantly knocked down to about 18% remaining expression (3 days after transfection) and 27% (5 days after transfection) compared to untransfected control or negative control (nontargeted siRNA transfection) cells (supplemental Fig. 4A). In any case cells showed a viability of at least 80% (supplemental Fig. 4B). Knockdown of ATG5, starting at Day 5 or 7 of a 10-day PQ exposure, did induce a slight but significant decrease of PQ-mediated LysoTracker intensity (Fig. 5). It was found that the number of lipofuscin-loaded lysosomes per cell and the area of lysosomes per cell area decreased significantly (a decrease compared to cells treated only with PQ for 10 days; compared to PQ addition for 5 or 7 days, which represent the ‘‘base’’ values the intensity increased). Lipofuscin-free lysosomes showed a significant increase in number and area after 10 days of PQ exposure and transfection at Day 5 or 7 during the treatment period (Fig. 5A). The amount of cytosolic lipofuscin was considerably increased by ATG5 knockdown starting at both Day 5 and Day 7 (Fig. 5B). The results of quantification of the four different marker proteins (Fig. 5C and D) correlated again very well with the LysoTracker fluorescence results (Fig. 5A). Again analogous results were detected using MDC (supplemental Fig. S5).

Effects of macroautophagy on lipofuscin formation, protein aggregate toxicity, and oxidant formation By exploring knockout cells and siRNA we demonstrated clearly that macroautophagy is involved in lysosomal lipofuscin accumulation. Thus, whenever lipofuscin accumulation in lysosomes was inhibited, free cytosolic lipofuscin was increased. Therefore, we tested in the next series of experiments, whether blocking of macroautophagy is inhibiting the total lipofuscin formation. As demonstrated in Fig. 6A there is no change in the formation of SIPS-induced lipofuscin, regardless of whether macroautophagy or lysosomal activity is blocked by ATP6V0A1 siRNA, ATG5 siRNA, or in knockout cells. If lysosomal localization is not influencing the lipofuscin formation, this raises the question whether endosomal/ lysosomal sequestration is beneficial for cellular viability. In human dermal fibroblasts with nonimpaired macroautophagy, viability due to SIPS was reduced by some 16% after 10 days in comparison to untreated control cells. However, if macroautophagy is impaired the number of dead cells increases to some 34% (siRNA) and 30% (ATG5–/–) in comparison to the corresponding controls. However, ATP6V0A1 knockdown, reducing just lysosomal capability, did not influence the level of dead cells compared to SIPS cells (see Fig. 6B). In earlier work we could demonstrate that the presence of protein aggregates is accompanied by the induction of oxidative stress [11]. Therefore, we decided to test whether the increased cytosolic presence of lipofuscin is accompanied by an increased ROS production. Whereas SIPS increases the DCF fluorescence by some 25% compared to untreated controls (100%), the prevention of lipofuscin uptake by blocking of macroautophagy enhances this DCF production to 50–80% over the corresponding control (Fig. 6C). Thus, ATP6V0A1 knockdown preventing lysosomal acidification does not have such an effect. Again, the results from ATP6V0A1 knockdown could be repeated with concanamycin A (data not shown).

Discussion Autophagy is a critical catabolic process through which cytoplasmic components are degraded, including the removal of longlived and damaged proteins and organelles and recycling of essential

A. H¨ ohn et al. / Free Radical Biology and Medicine 53 (2012) 1760–1769

1765

Fig. 3. Quantification of lysosomes, lipofuscin, and both lysosomal and autophagy-related marker proteins in human dermal fibroblasts after PQ-induced SIPS and ATP6V0A1 knockdown. In the left part of Panel A the fluorescence intensity of stained LAMP1 is demonstrated. The control (C, black column) is set to 100%. The number of counted lysosomes via LAMP1 signal is shown in the middle part and the signal area in the right part. Lipofuscin-free lysosomes are represented by light gray columns and lipofuscin-loaded lysosomes by dark-gray columns. The percentage of free cytosolic lipofuscin in relation to the whole cellular amount is shown in Panel B. Panel C shows the relative protein expression of lysosomal and autophagy-rel proteins. Protein expression of untreated cells is set to 100% (black column). In D, representative immunoblots are shown. Statistical significance compared to the control (black columns) is indicated by ‘n’, compared to the PQ control (10-day PQ treatment) by ‘#’, oneway ANOVA, Po 0.05 (A and B). In C, statistical significance is indicated by ‘n’ regarding the specific control (black columns), Student’s t test, P o0.05.

anabolic building blocks. Several human diseases are associated with decreased or impaired autophagy like vacuolar myopathies. These muscular disorders are associated with a massive accumulation of imperfect autophagosomal or lysosomal vacuoles [15,16]. Also neurodegenerative disorders, such as Parkinson’s [17], Alzheimer’s [18], and Huntington’s [19] disease, are associated with reduced autophagic activity and the occurrence of intracellular protein aggregates (inclusion bodies) is characteristic of these neurodegenerative diseases. Cellular aging in general is characterized by a progressive accumulation of nonfunctional cellular components due to oxidative damage and a decline in turnover rates, which could be found in nearly all cellular and animal models analyzed so far [20]. Impaired macroautophagy and also chaperone-mediated autophagy are well documented [21,22]. The consequences of this reduced degradation rate, especially in postmitotic cells, are referred to as inadequate removal of abnormal proteins and slowing down of the turnover of functional ones [23]. Those abnormal, damaged proteins usually develop aberrant intermolecular interactions resulting in the formation of aggregates. These aggregates can be further oxidized and cross-linked resulting in undegradable lipofuscin [24]. This process is not considered to be pathologic, in contrast to the occurrence of inclusion bodies in

neurodegenerative diseases. According to the current doctrine the biogenesis of lipofuscin is a multistep process dependent on ironcatalyzed oxidation of autophagocytosed or heterophagocytosed material in the lysosomal lumen representing intralysosomal garbage [25,26]. It is, however, questionable whether lysosomes actually are essential for the formation of lipofuscin or whether lipofuscin is actually formed only within lysosomes. The results presented in this paper clearly show that lysosomes or especially the low lysosomal pH is not mandatory for lipofuscin formation. Furthermore, we could show that autophagy impairment via ATG5 knockout/knockdown leads to a massive cytosolic lipofuscin accumulation in SIPS fibroblasts. SIPS human dermal fibroblasts share similar characteristics with cells undergoing senescence. These similarities include the typical morphology of senescence [27], senescence-associated bgalactosidase activity, and growth arrest in the G1 phase of the cell cycle [28]. In the present paper we could also show the typical morphological change from a spindle shape to an enlarged and irregular shape (see supplemental Fig. S1). When autophagy is unaffected, lysosomal lipofuscin accumulation also occurs in SIPS fibroblasts (Fig. 1A) and the amount of free cytosolic lipofuscin did not change (Fig. 1B). This is associated with an increased number and area of lysosomes per cell and cell area, respectively

1766

A. H¨ ohn et al. / Free Radical Biology and Medicine 53 (2012) 1760–1769

Fig. 4. Quantification of lysosomes, lipofuscin, and both lysosomal and autophagy-related marker proteins in ATG5 þ / þ and ATG–/– MEFs after SIPS. In Panel A, the signal intensity of LysoTracker is demonstrated in the left part and each control (C, black column) is set to 100%. The middle part shows the number of counted lysosomes per cell, and the signal area is shown in the right part. In both parts it is distinguished between lipofuscin-free (light gray columns) and lipofuscin-loaded lysosomes (dark-gray columns). In Panel B, the percentage of free cytosolic lipofuscin in relation to the whole cellular amount is shown. (C) Relative protein expression of lysosomal and autophagy-related proteins. Protein expression of untreated cells is set to 100% (black column). For ATG–/– an exemplary immunoblot with LC3 antibody is shown. Statistical significance compared to the control is indicated by ‘n’, Student’s t test, Po 0.05.

(Fig. 1A), and consistent with earlier reports which showed that lysosomes also increase in number and size in senescent cells [29,30]. Lysosomes are key organelles in the aging process due to their involvement in both macroautophagy and other housekeeping mechanisms. Autophagosomes themselves have limited degrading capacity and rely on fusion with lysosomes. However activities of cathepsins, which are major proteolytic enzymes in lysosomes, are known to be decreased in senescent cells [31]. It has been recently shown that knockdown of isoform A1, part of the integral V0 domain of vacuolar proton-translocating ATPase, increases the lysosomal/endosomal pH, without altering the steady-state pH of the cytosol [32]. ATP6V0A1 knockdown or chemical inhibitors like concanamycin A (at nanomolar concentrations) [33] inhibit the acidification of lysosomes and thus also the intralysosomal degradation by lysosomal cathepsins with their optimum pH being at 4–5. One might, therefore, expect an increased amount of intralysosomal lipofuscin as a consequence of dysfunctional lysosomes due to high pH, as well as an increased number of lysosomes as a ‘‘compensation’’ process; or in contrast the formation of lipofuscin might be reduced if the low pH is

required for the chemical reactions necessary to build up lipofuscin. After knockdown of ATP6V0A1 as well as after concanamycin A treatment (data not shown), however, no significant changes in number of lysosomes (Fig. 3A) nor in the level of free cytosolic lipofuscin were observed (Fig. 3B). In summary these results allow two conclusions: (i) lysosomal cathepsins do not dramatically reduce the amount of fluorescent lipofuscin, and (ii) a low pH is not influential on lipofuscin formation and accumulation in lysosomes. But as noted above it was possible to modulate the colocalization of lipofuscin and lysosomes via inhibition of ATG5 and, therefore, the role of macroautophagy could be clarified in this context. ATG5 þ / þ MEFs after SIPS induction for 10 days (Fig. 4 and supplemental Fig. S3) showed effects comparable to the human dermal fibroblasts (Figs. 1 and 2). Concerning the ATG–/– MEFs, this is fundamentally different. SIPS treatment did not increase LysoTracker intensity, nor number and area of lysosomes, as well as LAMP1 expression (Fig. 4C), but compared to the ATP6V0A1 knockdown we found a high amount of free cytosolic lipofuscin for the ATG–/– MEFs per cell area (Fig. 4B). MDC fluorescence was

A. H¨ ohn et al. / Free Radical Biology and Medicine 53 (2012) 1760–1769

Number of lysosomes/Cell [N]

LysoTracker intensity [%]

200 150

100

70

700 600 500 400 300 200

50

C. 10 days of PQ treatment 5 7 Transfection day

Protein expression [%]

250

0

60 50 40 30 20

C.

10 days of PQ treatment 5 7 Transfection day

0

C.

10 days of PQ treatment 5 7 Transfection day

Control SIPS-control -10 days PQ 10 days PQ + Transfection day 5 10 days PQ + Transfection day 7

200

1.0 0.8 0.6 0.4 0.2

10

100 0

1.2

LF loaded LF free Non-lysosomal lipofuscin [%]

800 250

80

LF loaded LF free Area of lysosomes /cell area [%]

900

300

1767

0.0

C. 10 days of PQ treatment 5 7 Transfection day

LAMP1 (120 kDa) GAPDH (40 kDa) Atg5-Atg12 (55 kDa)

150 Atg5 (33 kDa) GAPDH (40 kDa)

100

GAPDH (40 kDa) LC3I (18 kDa) LC3II (16 kDa)

50 C. 0

LAMP1

Atg5

Atg5-Atg12

LC3II

10 days of PQ treatment 5 7 Transfection day

Fig. 5. Quantification of lysosomes, lipofuscin, and both lysosomal and autophagy-related marker proteins in human dermal fibroblasts after PQ-induced SIPS and ATG5 knockdown. In the left part of Panel A, the signal intensity of LysoTracker is demonstrated and the control (C, black column) is set to 100%. The number of counted lysosomes is shown in the middle part and the signal area in the right part. Lipofuscin-free lysosomes are represented by light gray columns and lipofuscin-loaded lysosomes by dark-gray columns. The percentage of free cytosolic lipofuscin in relation to the whole cellular amount is shown in Panel B. Panel C shows the relative protein expression of lysosomal and autophagy-related proteins. Protein expression of untreated cells is set to 100% (black column). In D, representative immunoblots are shown. Statistical significance compared to the control (black columns) is indicated by ‘n’, compared to the PQ control (10-day PQ treatment) by ‘#’, one-way ANOVA, P o0.05 (A and B). In C, statistical significance is indicated by ‘n’ regarding the specific control (black columns), Student’s t test, Po 0.05.

also found but there was no increase compared to control cells, which is interesting, since ATG5–/– cells should be completely autophagy deficient. Recently an ATG5/ATG7 independent autophagy pathway was described [34], explaining the low, but existing MDC fluorescence. This result could be verified in human dermal fibroblasts with ATG5 knockdown via RNA interference (Fig. 5). Interestingly, in both ATG5 knockout/knockdown models applied above, we were able to reduce the uptake of lipofuscin into lysosomes and, therefore, manipulate the lipofuscin amount within lysosomes. However, none of these treatments actually prevented the formation of lipofuscin (Fig. 6A). Therefore, it seems clear now that in contrast to earlier hypothesis [25,26] lipofuscin can be formed in the cytosol and taken up into autophagosomes in a secondary step. For quite some time lipofuscin has been considered as a relatively inert by-product of normal aging and some diseases. There is much evidence now that lipofuscin is not such an inert waste product of cellular metabolism, but rather an active component influencing the metabolic pathways of a senescent cell. As shown in one of our recent publications [11], lipofuscin is, in addition to the mitochondria, the most likely source of oxidants in senescent cells. Therefore, it was interesting to compare all

previously described experimental approaches with regard to viability and oxidant formation. The viability for the ATP6V0A1 knockdown cells was comparable to the cells only undergoing SIPS treatment, clearly pointing out a minor role of lysosomes in modulating lipofuscin toxicity. However, if macroautophagy is inhibited by siRNA transfection during SIPS, the decrease in viability is substantially more pronounced. The same was true for the ATG5–/– MEFs (Fig. 6B). As already noted, lipofuscin is one of the major oxidant sources in senescent cells [11] so we tested whether such an oxidant formation is enhanced if lipofuscin is not taken up by lysosomes. Most importantly, the formation of free radicals is increased as shown in Fig. 6C. This is in good correlation with the decrease in viability. Summarizing, our findings show that there is a clear correlation between high cytosolic lipofuscin levels and reduced viability of cells in combination with increased oxidant formation catalyzed by lipofuscin. We could also show that macroautophagy plays an important role in the sequestration of lipofuscin into lysosomes. Thus, macroautophagial uptake of lipofuscin into lysosomes could be a mechanism of cellular protection. Contrary to the original hypothesis autophagosomes/lysosomes are not required for the formation of the fluorescent lipofuscin, but are

A. H¨ ohn et al. / Free Radical Biology and Medicine 53 (2012) 1760–1769

400

40

300

30

200

200

100

DCF fluorescence [% of corresponding control]

180 Lethal cells [% of corresponding control]

20

10

160

140

120

0

ATP6V0A1 siRNA

ATG5-/-

ATG 5 siRNA

ATG5+/+

SIPS

ATP6V0A1 siRNA

ATG5-/-

ATG 5 siRNA

100 SIPS

ATP6V0A1 siRNA

ATG5-/-

ATG 5 siRNA

ATG5+/+

SIPS

C.

0

ATG5+/+

Normalized lipofuscin intensity [%]

1768

Fig. 6. Effects of reduced macroautophagy on lipofuscin formation cell viability and oxidant formation. In Panel A, the formation of lipofuscin is presented, regardless of the intracellular distribution. The control (C, black column) is set to 100%. In Panel B, the viability for each experimental approach (10 days SIPS) compared to a corresponding control is shown. The corresponding control is defined as 0% lethal cells and represents the treatment of the individual groups (siRNA transfection but without SIPS treatment in any case), in order to show the effects of SIPS independent from the other factors. SIPS-treated ATG5 þ / þ and ATG–/– MEFs were compared to untreated MEFs. (C) The relative amount of DCF fluorescence for each experimental approach. Again the data are compared to a corresponding control (same treatment as indicated for each column exclusive of PQ) set to 100%. Statistical significance regarding control (A) and corresponding control (B and C) is indicated by ‘n’, Student’s t test, Po 0.05.

merely a storage for protein aggregates/lipofuscin, therefore, reducing—but not preventing—aggregate toxicity.

Acknowledgments We thank Dr. N. Mizushima for providing the ATG5 wild-type and knockout mouse embryonic fibroblasts via RIKEN BRC cell bank. This work was supported by the DFG.

Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.freeradbiomed. 2012.08.591.

References ¨ [1] Hohn, A.; Jung, T.; Grimm, S.; Catalgol, B.; Weber, D.; Grune, T. Lipofuscin inhibits the proteasome by binding to surface motifs. Free Radic. Biol. Med. 50:585–591; 2011. [2] Sitte, N.; Huber, M.; Grune, T.; Ladhoff, A.; Doecke, W. D.; Von, Z. T., et al. Proteasome inhibition by lipofuscin/ceroid during postmitotic aging of fibroblasts. FASEB J 14:1490–1498; 2000. [3] Overbye, A.; Fengsrud, M.; Seglen, P. O. Proteomic analysis of membraneassociated proteins from rat liver autophagosomes. Autophagy 3:300–322; 2007. [4] Mizushima, N.; Yamamoto, A.; Hatano, M.; Kobayashi, Y.; Kabeya, Y.; Suzuki, K., et al. Dissection of autophagosome formation using Apg5-deficient mouse embryonic stem cells. J. Cell Biol 152:657–668; 2001. [5] Kabeya, Y.; Mizushima, N.; Ueno, T.; Yamamoto, A.; Kirisako, T.; Noda, T., et al. LC3, a mammalian homologue of yeast Apg8p, is localized in autophagosome membranes after processing. EMBO J 19:5720–5728; 2000. [6] Wenger, D. A.; Coppola, S.; Liu, S. L. Insights into the diagnosis and treatment of lysosomal storage diseases. Arch. Neurol 60:322–328; 2003. [7] Cuervo, A. M.; Dice, J. F. When lysosomes get old. Exp. Gerontol. 35:119–131; 2000.

[8] Lynch, G.; Bi, X. Lysosomes and brain aging in mammals. Neurochem. Res 28:1725–1734; 2003. [9] O., Medrano, E. E.; Von, Z. T. Cellular and molecular mechanisms of stressinduced premature senescence (SIPS) of human diploid fibroblasts and melanocytes. Exp. Gerontol 35:927–945; 2000. [10] Kuma, A.; Hatano, M.; Matsui, M.; Yamamoto, A.; Nakaya, H.; Yoshimori, T., et al. The role of autophagy during the early neonatal starvation period. Nature 432:1032–1036; 2004. [11] H¨ohn, A.; Jung, T.; Grimm, S.; Grune, T. Lipofuscin-bound iron is a major intracellular source of oxidants: role in senescent cells. Free Radic. Biol. Med. 2010. [12] Mosmann, T. Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays. J. Immunol. Methods 65:55–63; 1983. [13] Biederbick, A.; Kern, H. F.; Elsasser, H. P. Monodansylcadaverine (MDC) is a specific in vivo marker for autophagic vacuoles. Eur. J. Cell Biol 66:3–14; 1995. ¨ [14] Jung, T.; Hohn, A.; Catalgol, B.; Grune, T. Age-related differences in oxidative protein-damage in young and senescent fibroblasts. Arch. Biochem. Biophys 483:127–135; 2009. [15] Nishino, I. Autophagic vacuolar myopathies. Curr. Neurol. Neurosci. Rep. 3:64–69; 2003. [16] Malicdan, M. C.; Nishino, I. Autophagy in lysosomal myopathies. Brain Pathol. 22:82–88; 2012. [17] Anglade, P.; Vyas, S.; Javoy-Agid, F.; Herrero, M. T.; Michel, P. P.; Marquez, J., et al. Apoptosis and autophagy in nigral neurons of patients with Parkinson’s disease. Histol. Histopathol 12:25–31; 1997. [18] Cataldo, A. M.; Hamilton, D. J.; Barnett, J. L.; Paskevich, P. A.; Nixon, R. A. Properties of the endosomal-lysosomal system in the human central nervous system: disturbances mark most neurons in populations at risk to degenerate in Alzheimer’s disease. J. Neurosci 16:186–199; 1996. [19] Kegel, K. B.; Kim, M.; Sapp, E.; McIntyre, C.; Castano, J. G.; Aronin, N., et al. Huntingtin expression stimulates endosomal-lysosomal activity, endosome tubulation, and autophagy. J. Neurosci. 20:7268–7278; 2000. [20] Ward, W. F. The relentless effects of the aging process on protein turnover. Biogerontology 1:195–199; 2000. [21] Terman, A. The effect of age on formation and elimination of autophagic vacuoles in mouse hepatocytes. Gerontology 41(2):319–326; 1995. [22] Cuervo, A. M.; Dice, J. F. Age-related decline in chaperone-mediated autophagy. J. Biol. Chem. 275:31505–31513; 2000. [23] Goto, S.; Takahashi, R.; Kumiyama, A. A.; Radak, Z.; Hayashi, T.; Takenouchi, M., et al. Implications of protein degradation in aging. Ann. N. Y. Acad. Sci. 928:54–64; 2001. [24] Jung, T.; Bader, N.; Grune, T. Lipofuscin: formation, distribution, and metabolic consequences. Ann. N. Y. Acad. Sci. 1119:97–111; 2007. [25] Sohal, R. S.; Brunk, U. T. Mitochondrial production of pro-oxidants and cellular senescence. Mutat. Res 275:295–304; 1992.

A. H¨ ohn et al. / Free Radical Biology and Medicine 53 (2012) 1760–1769

[26] Brunk, U. T.; Terman, A. Lipofuscin: mechanisms of age-related accumulation and influence on cell function. Free Radic. Biol. Med 33:611–619; 2002. [27] Dimri, G. P.; Lee, X.; Basile, G.; Acosta, M.; Scott, G.; Roskelley, C., et al. A biomarker that identifies senescent human cells in culture and in aging skin in vivo. Proc. Natl. Acad. Sci. USA 92:9363–9367; 1995. [28] de Magalhaes, J. P.; Chainiaux, F.; de, L. F.; Mainfroid, V.; Migeot, V.; Marcq, L., et al. Gene expression and regulation in H2O2-induced premature senescence of human foreskin fibroblasts expressing or not telomerase. Exp. Gerontol 39:1379–1389; 2004. [29] Levine, E.; Eagle H., E. M. Morphologic changes accompanying senescence of cultured human diploid cells. J. Exp. Med 131:1211–1222; 1970. [30] Brunk, U. Distribution and shifts of ingested marker particles in residual bodies and other lysosomes. Studies on in vitro cultivated human glia cells in phase II and 3. Exp. Eye Res 79:15–27; 1973. [31] Chondrogianni, N.; Stratford, F. L.; Trougakos, I. P.; Friguet, B.; Rivett, A. J.; Gonos, E. S. Central role of the proteasome in senescence and survival of

1769

human fibroblasts: induction of a senescence-like phenotype upon its inhibition and resistance to stress upon its activation. J. Biol. Chem 278:28026–28037; 2003. [32] Hinton, A.; Sennoune, S. R.; Bond, S.; Fang, M.; Reuveni, M.; Sahagian, G. G., et al. Function of a subunit isoforms of the V-ATPase in pH homeostasis and in vitro invasion of MDA-MB231 human breast cancer cells. J. Biol. Chem 284:16400–16408; 2009. [33] Sobota, J. A.; Back, N.; Eipper, B. A.; Mains, R. E. Inhibitors of the V0 subunit of the vacuolar H þ -ATPase prevent segregation of lysosomal- and secretorypathway proteins. J. Cell Sci 122:3542–3553; 2009. [34] Nishida, Y.; Arakawa, S.; Fujitani, K.; Yamaguchi, H.; Mizuta, T.; Kanaseki, T., et al. Discovery of Atg5/Atg7-independent alternative macroautophagy. Nature 461:654–658; 2009.