Experimental Gerontology 34 (1999) 943–957 Ceroid/lipofuscin-loaded human fibroblasts show decreased survival time and diminished autophagocytosis during amino acid starvation૾ Alexei Termana,b,*, Helge Dalenc, Ulf T. Brunka a Division of Pathology II, Department of Neuroscience and Locomotion, Faculty of Health Sciences, Linkoping University, S-58185 Linkoping, Sweden ¨ ¨ b Institute of Gerontology AMS of Ukraine, Kiev 254114, Ukraine c Department of Pathology, The Gade Institute, University of Bergen, Bergen N-5021, Norway Received 21 June 1999; received in revised form 23 August 1999; accepted 9 September 1999 Abstract To test whether heavy accumulation of ceroid/lipofuscin can disturb important functions of the lysosomal system, AG-1518 human fibroblasts, ceroid/lipofuscin-loaded (following prolonged culture at normobaric hyperoxia) or not, were exposed to amino acid starvation. Ceroid/lipofuscinloading resulted in decreased cellular survival. Also, there was an inverse relationship between amounts of ceroid/lipofuscin and the survival time of individual cells within the same cultures. Ceroid/lipofuscin-loaded fibroblasts displayed diminished autophagocytotic capacity, as demonstrated by electron microscopy and by treatment of cell cultures with NH4Cl (which inhibits autophagocytotic degradation by increasing intralysosomal pH) for 1 week before ensuing starvation. The latter treatment increased survival of control cells (due to deposition of nondegraded autophagocytosed material before start of starvation), but not that of ceroid/lipofuscin-loaded cells. Moreover, when NH4Cl treatment was combined with starvation, both groups of cells showed approximately the same shortened survival times, testifying to the causal relationship between diminished autophagocytosis and decreased survival of starving ceroid/lipofuscin-loaded cells. We hypothesize that large amounts of undegradable ceroid/lipofuscin within the acidic vacuolar compartment may interfere with lysosomal function, resulting in poor renewal of long-lived proteins and worn-out/damaged organelles, decreased adaptability, and cell death. © 1999 Elsevier Science Inc. All rights reserved. Keywords: Aging; Autophagocytosis; Fibroblasts; Lipofuscin; Lysosomal degradation; Starvation ૾ This study was supported by the Swedish Medical Research Council Grant 4481. * Corresponding author. Tel.: ϩ46-13-221515; fax: ϩ46-13-221529. E-mail address: alete@pat.liu.se (A. Terman) 0531-5565/99/$ – see front matter © 1999 Elsevier Science Inc. All rights reserved. PII: S 0 5 3 1 - 5 5 6 5 ( 9 9 ) 0 0 0 7 0 - 4 944 A. Terman et al. / Experimental Gerontology 34 (1999) 943–957 1. Introduction Lipofuscin is an electron-dense, autofluorescent, polymeric substance that, as a recognized hallmark of aging, accumulates progressively within the lysosomal compartment of postmitotic cells (Hannover, 1842; Koneff, 1886; Strehler et al., 1959; Sohal and Wolfe, 1986; Harman, 1989; Porta, 1991). Evidence suggests that lipofuscin is formed from autophagocytosed macromolecular substances by iron-catalyzed oxidative modification (Brunk et al., 1992). Potentially, it can accumulate in any type of cell, not only in postmitotic ones, and originate from a variety of biomolecules. The amount of lipofuscin increases in cultured fibroblasts, and other cells when their division is inhibited at confluency or after repeated passages upon entering phase III (Hayflick, 1965; Robbins et al., 1970; Brock and Hay, 1971; Brandes et al., 1972; Brunk et al., 1973; Rattan et al., 1982; von Zglinicki et al., 1995). A pigment of almost the same morphological and physicochemical properties accumulates in vivo in a variety of cell types under pathological conditions, such as malnutrition, stress, ionizing radiation, atherosclerosis, lysosomal storage diseases, etc. (reviewed in Yin, 1996; Terman and Brunk, 1998b). This latter pigment is usually called ‘lipofuscin of ceroid type’ or simply ‘ceroid’ and basically has the same origin as lipofuscin. Although the accumulation of lipofuscin/ceroid is undoubtedly an important manifestation of aging, as well as a number of diseases, it is still unclear whether large amounts of the pigment can substantially disturb functions of the acidic vacuolar apparatus, threatening the life of the cell. In elderly individuals, the amount of lipofuscin in some postmitotic cells may increase dramatically (up to 75% of the cytoplasmic volume) (Treff, 1974). In this case, lipofuscin may occupy a large portion of intralysosomal space, and probably divert lysosomal enzymatic activity from useful purposes, causing newly produced digestive enzymes to be delivered to lipofuscin-loaded lysosomes rather than to autophagosomes (Brunk and Terman, 1999). If so, ceroid/lipofuscin-loaded cells would show enhanced sensitivity to influences that place increased demands upon the lysosomal apparatus, e.g., enhanced autophagocytosis. To test whether this hypothesis is valid, we studied the survival of cultured human fibroblasts containing different amounts of ceroid/lipofuscin during amino acid starvation. Under starvation conditions, cells degrade endogenous proteins, thereby generating a pool of amino acids for critical protein synthesis (Mortimore et al., 1983; Lawrence and Brown, 1992; Blommaart et al., 1997). Some, mainly short-lived, cellular proteins are known to be degraded in the cytosol (mainly by proteasomes), whereas many other (mainly longlived) proteins and those within organelles undergo intralysosomal degradation (Dice, 1989; Seglen and Bohley, 1992; Berg et al., 1995; Mortimore et al., 1996; Blommaart et al., 1997). Our investigations were aimed at clarifying the importance of intralysosomal degradation upon survival of starving cells and the possible effects of lipofuscin accumulation on those functions. 2. Materials and methods 2.1. Culture conditions and experimental design Human foreskin AG-1518 fibroblasts were obtained at passage 8 from the Coriell Institute (Camden, NJ, USA) and grown in 25-cm2 plastic flasks (Costar, Cambridge, MA, A. Terman et al. / Experimental Gerontology 34 (1999) 943–957 945 USA) in Eagle’s minimum essential medium in a humidified atmosphere of 8% O2, 87% N2, and 5% CO2 at 37° by using B 5060 EC/O2 incubators (Heraeus, Hanau, Germany). The culture medium, supplemented with 10% fetal bovine serum, 2 mM glutamine, 50 IU/mL penicillin-G, and 50 mg/mL streptomycin, was changed twice a week. The cells were subcultured by trypsinization at a ratio of 1 : 2. At passage 25, they were transferred to 35-mm plastic Petri dishes (Costar), and confluent cultures were established. For ceroid/lipofuscin loading, the cultures were exposed to normobaric hyperoxia (40% O2, 55% N2, and 5% CO2) for 4 months, while control cells were maintained under normal culture conditions (8% O2, 87% N2, and 5% CO2). For amino acid starvation, trypsinized fibroblast cultures, both control and ceroid/ lipofuscin-loaded (Groups i and ii, respectively), were plated at a density of 5000 cells per cm2. After 8 h, when the cells were well spread, Eagle’s medium was replaced with Earl’s buffered salt solution supplemented with 5% fetal calf serum, vitamins, and antibiotics. The cells did not divide under these conditions. Part of the cultures undergoing starvation (Group iii, control cells and Group iv, ceroid/lipofuscin-loaded cells) also were exposed to 10 mM NH4Cl to block intralysosomal degradation. Other cultures (Groups v and vi, respectively) initially were treated with NH4Cl (under otherwise normal culture conditions) for 1 week, then trypsinized, and finally exposed to amino acid starvation without NH4Cl. Separate dishes from Groups i and ii were used to compare the survival times of individual cells containing different amounts of ceroid/lipofuscin. This allowed us to directly relate survival time to the amount of ceroid/lipofuscin and to exclude any possible effect of the hyperoxia before starvation. Moreover, having sequences of images of the very same cells we also were able to estimate whether the amount of ceroid/lipofuscin changed during starvation, i.e., was degraded under enhanced autophagocytosis. The experiments were performed in triplicate. 2.2. Conventional light microscopy The cultures were observed by using a Nikon TMS phase-contrast inverted microscope (Tokyo, Japan). Cells in marked areas of the dishes were counted daily until they died. Dead cells were recognized by their sharply outlined nuclei, prominent nucleoli, visible cytoplasmic filamental structures or by more pronounced signs of damage, such as destroyed cellular membranes. Vital staining with 0.1% trypan blue for 5 min at room temperature was used to verify the accuracy of the detection of dead cells in unstained specimens by phase contrast microscopy (Fig. 1). Part of the cultures were fixed in 4% neutral buffered formaldehyde with 0.5% Triton X-100 for 20 min and then stained according to the TUNEL technique for apoptosis (by using the Apop Tag in situ apoptosis detection kit; Oncor, Gaithersburg, MD, USA) after the first, second, and third days of starvation, respectively. After fixation, the cultures were subsequently incubated with terminal deoxynucleotidyl transferase and a mixture of digoxigenin-labeled nucleotides for 1 h and then with anti-digoxigenin-peroxidase-labeled antibodies for another 30 min. Color development was performed in a 0.05% diaminobenzidine solution with 0.02% H2O2 for 20 min. Nuclei with fragmented DNA stained brown. 946 A. Terman et al. / Experimental Gerontology 34 (1999) 943–957 Fig. 1. Laser scanning images (543-nm excitation light) of human AG-1518 fibroblasts on the third day of amino acid starvation. (A) Phase-contrast mode; (B) subsequent vital staining with trypan blue, bright field mode. A dead cell with a sharply outlined nucleus, increased granularity, and visible filamental structures (phase contrast) was stained with trypan blue, whereas a normal cell remained trypan-blue negative. 2.3. Laser scanning microscopy To study the survival of individual fibroblasts, images of cells from marked areas of the dishes were obtained daily by using a LSM-410 inverted laser scanning microscope (Carl Zeiss, Jena, Germany). A 543 nm (green) helium-neon laser was the source of excitation light. Two microscope channels were applied: a phase contrast channel for evaluation of general cell morphology and detection of cell death (by the same criteria as for conventional light microscopy), and a fluorescence channel (590 nm barrier filter) to visualize ceroid/lipofuscin. A 63 ϫ 1.25 Plan-Neofluar lens was used for high-resolution confocal images (pinhole 20). For total cell ceroid/lipofuscin measurements, we used nonconfocal images (obtained with a 40 ϫ 0.6 LD-Achroplan lens). Pixel values for all images were within the range of 0 to 255. 2.4. Transmission electron microscopy For transmission electron microscopy, cells were fixed in situ in a cacodylate-buffered (0.1 M, pH 7.2) 2% glutaraldehyde solution with 0.1 M sucrose at 4°C, postfixed in 1% osmium tetroxide (in 0.15 M cacodylate buffer) at room temperature, stained en bloc with 2% uranyl acetate in 50% ethanol, dehydrated in a graded series of ethanol, and finally embedded in Epon-812. Thin sections were cut with a diamond knife parallel to the A. Terman et al. / Experimental Gerontology 34 (1999) 943–957 947 growth substrate, stained with lead citrate, and examined in a JEOL 1200-EX electron microscope (Tokyo, Japan). Electron microscopy images were either photographed, or digitized and stored by using a BioScan wide-angle CCD camera with the attached software (Gatan, Pleasanton, CA, USA). 2.5. Morphometry The cell size (area occupied by the cell) and average cellular ceroid/lipofuscin fluorescence intensity were determined on nonconfocal laser scanning images by using the public domain National Institutes of Health Image program (http://rsb.info.nih.gov/nihimage/). Average ceroid/lipofuscin fluorescence intensity for each cell was measured as average pixel value minus background. The amount of ceroid/lipofuscin per cell was calculated as the product of average pixel value by cell area and expressed in arbitrary units (a.u.). Digitized electron microscopy images (38 pixels/m) were used for measurement of the cytoplasmic volume fraction of vacuolar/lysosomal structures. For this purpose, the images were combined with a square grid (available at the National Institutes of Health Image program) with a spacing of 25 pixels, and the points (intersections of lines) corresponding to the vacuolar/lysosomal structures, and the rest of the cytoplasm, were counted. 2.6. Mathematical treatment The survival time of cells was expressed in days, because a higher degree of accuracy was not possible. For survival analysis, data from three different dishes for each group were combined into one sample (a total of 433, 335, 383, 323, 389, and 310 cells in Groups i, ii, iii, iv, v, and vi, respectively). Combined samples of smaller sizes were used to study the relationship between the amount of ceroid/lipofuscin and survival time within Groups i or ii (a total of 72 and 81 cells from three separate dishes, respectively). Because the distribution of ceroid/lipofuscin between cells was appreciably skewed, its amount was characterized by medians and interquartile ranges (IQR), rather than by means (averages) and standard deviations. Volume fractions of cellular vacuolar/lysosomal structures were determined for each specimen as a ratio of their point scores to the cytoplasm’s point score (Weibel, 1969), and then averaged within each group. Fifteen randomly selected cells from each specimen were analyzed (i.e., a total of 45 cells from each group). Comparison of survival curves was performed by using the log rank test. In order to study the relationship between the amount of ceroid/lipofuscin and the survival time within Groups i and ii, cells were distributed into subgroups according to their survival times. These subgroups were compared by the Kruskal–Wallis rank test. Kendall rank correlation coefficients were used to analyze the relationship between median ceroid/ lipofuscin value and survival time of the subgroups. The paired Wilcoxon test was used to compare ceroid/lipofuscin values of the same cells during starvation. The volume fractions of the vacuolar/lysosomal structures were compared by using analysis of variance and Student’s t-test. Values are presented as median (IQR) or as mean Ϯ SD. p Values Ͻ 0.05 were considered significant. 948 A. Terman et al. / Experimental Gerontology 34 (1999) 943–957 3. Results 3.1. Amino acid starvation Starvation of fibroblasts, ceroid/lipofuscin-loaded or not, was associated with shortening and disappearance of their lamellopodiae, some cellular shrinkage, and an increasing amount of cytoplasmic vesicles. These changes were first detectable by phase contrast microscopy a few hours after the onset of starvation and became more pronounced over time (Fig. 2). As seen by electron microscopy, fibroblasts starved for 1 day accumulated numerous autophagic vacuoles of different sizes, which rarely occurred in nonstarving cells (Figs. 3,A, B, D, and E and 4). These starvation-induced autophagic vacuoles were recognized by the presence of electron-lucent areas within the vacuoles (apparently due to increased osmotic pressure with resultant influx of water due to degradation of autophagocytosed structures). Either the whole vacuole was electron-lucent, or it contained both electron-lucent and electron-dense material, or there was electron-lucent space around sequestered cytoplasmic structures. It is understood that some autophagic vacuoles may be early autophagosomes whereas others may be late autophagolysosomes. After 3 days, the autophagic vacuoles appeared to be more numerous (Fig. 3, C and F and 4). During starvation, the cells gradually decreased in size and finally died by necrosis. Although surface blebbing and TUNEL-positivity (indicative of apoptosis) usually preceded cell death, dying cells did not show nuclear pyknosis or fragmentation. Most dead cells remained spread and attached to the bottom of the culture dishes. Ultrastructurally, they were characterized by swollen nuclei and cytoplasmic alterations typical of necrosis (Fig. 5). 3.2. Ceroid/lipofuscin-loading of cultured fibroblasts shortens their survival during amino acid starvation As is evident from Fig. 6, ceroid/lipofuscin-loaded fibroblasts (Group ii), with a median pigment content of 17.79 (40) a.u., showed significantly poorer survival than the control cells (Group i) having a median ceroid/lipofuscin content of 3.99 (3.31) a.u. Because the amount of ceroid/lipofuscin in individual cells within the same dishes varied widely, we also studied its relationship to survival time during starvation. As shown in Fig. 7A, the shorter survival times of Group ii cells (ceroid/lipofuscin-loaded) were associated with higher median ceroid/lipofuscin values. This was not the case for Group i, which had a low intracellular pigment content (Fig. 7B). 3.3. Inhibition of autophagocytotic degradation during starvation equalizes survival times of ceroid/lipofuscin-loaded and nonloaded fibroblasts Administration of NH4Cl, which inhibits intralysosomal degradation due to alkalization of the intralysosomal milieu, markedly shortened the survival time of all starved cultures. No significant difference in survival was found between ceroid/lipofuscin-loaded versus control cells (Groups iv and iii, respectively) when exposed to NH4Cl during starvation (Fig. 6A). A. Terman et al. / Experimental Gerontology 34 (1999) 943–957 949 Fig. 2. Laser scanning images (543-nm excitation light) of a single ceroid/lipofuscin-loaded fibroblast exposed to amino acid starvation and followed for 4 days. (A–E) Phase-contrast mode; (F–G) fluorescent mode. Images were obtained before starvation (A, F) as well as after 1 (B, G), 2 (C, H), 3 (D, I), and 4 (E, J) days of starvation, respectively. The images E and J show the cell to be dead. Note the increasing amount of cytoplasmic vesicles (A–D) and the absence of obvious changes in the content of ceroid/lipofuscin (F–I) during starvation. 950 A. Terman et al. / Experimental Gerontology 34 (1999) 943–957 Fig. 3. Ultrastructure of control (A–C) and ceroid/lipofuscin-loaded (D–F) fibroblasts: (A, D) under normal conditions; (B, E) after 1 day; and (C, F) after 3 days of amino acid starvation, respectively. Note that starving ceroid/lipofuscin-loaded cells accumulate fewer autophagic vacuoles compared to controls. CL, ceroid/lipofuscin-containing lysosomes (of different size and density, and with crystalloid and myelin figures); AV, autophagic vacuoles with electron-lucent areas; L, lipid droplets. 3.4. Inhibition of autophagocytotic degradation before starvation selectively prolongs survival time of fibroblasts with low amounts of ceroid/lipofuscin As shown in Fig. 6B, exposure to NH4Cl, under otherwise normal culture conditions, for 1 week before the start of starvation markedly prolonged the survival time of fibroblasts with a low pigment content (Group v). In contrast, pretreatment of ceroid/ lipofuscin-loaded fibroblasts with NH4Cl (Group vi) did not result in any increase of the survival time. 3.5. Starvation-induced autophagocytosis is depressed in ceroid/lipofuscin-loaded fibroblasts Ultrastructurally, human fibroblasts exposed to prolonged normobaric hyperoxia (Group ii) accumulated numerous irregularly shaped lysosomes of variable size and density, often with crystalloid or myelin-like structures. Apparently, most of them con- A. Terman et al. / Experimental Gerontology 34 (1999) 943–957 951 Fig. 4. Volume fraction of autophagic vacuoles (electron-lucent areas) in fibroblasts cultivated under normal conditions and exposed to amino acid starvation for 1 and 3 days. White columns, control cells (Group i); dark columns, ceroid/lipofuscin-loaded cells (Group ii). Although the amount of autophagic vacuoles was initially higher in ceroid/lipofuscin-loaded cells, its increase due to starvation was much less than in controls. Volume fractions of autophagic vacuoles significantly differ from each other within Groups i or ii (analysis of variance, p Ͻ 0.0001 for both groups), as well as between Groups i and ii (t-test; p ϭ 0.0036, p ϭ 0.0021, and p Ͻ 0.001 for cells under normal conditions, and after 1 and 3 days of starvation, respectively). tained autofluorescent material and could be classified as ceroid/lipofuscin inclusions. Control cells, not exposed to hyperoxia (Group i) accumulated much less pigment (see also Nilsson and Yin, 1997; Terman and Brunk, 1998a). One day after the onset of starvation, the number and volume fraction of newly formed autophagic vacuoles (i.e., with electron-lucent contents) increased, but to a much lesser degree than in control cells (Figs. 3, B and E and 4). After 3 days, the number and size of these structures had increased further, although less dramatically than in the cells of Group i (Figs. 3, C and F and 4). It is remarkable that under normal conditions the amount of autophagic vacuoles in ceroid/lipofuscin-loaded cells was higher than in controls (Fig. 4). This might reflect a prolonged degradation time of autophagic structures in ceroid/lipofuscin-loaded cells. 3.6. Starving cells fail to degrade ceroid/lipofuscin inclusions Although ceroid/lipofuscin was frequently seen within newly formed autophagic vacuoles (having electron-lucent spaces) of starving fibroblasts, it was preserved with respect to its characteristic features (medium to high electron density, presence of crystalloid or myelin figures) under starvation. In cells starved for 3 days, the ceroid/lipofuscincontaining lysosomes tended to aggregate and fuse, however without any substantial change of the pigment ultrastructure. Ceroid/lipofuscin autofluorescence did not decrease during amino acid starvation, as was evident from sequential images of the same cell (Fig. 2), as well as from measurement of cellular autofluorescence before starvation and on the third day of cultivation without 952 A. Terman et al. / Experimental Gerontology 34 (1999) 943–957 Fig. 5. Electron micrograph of a dead cell after 3 days of starvation. Note nuclear and cytoplasmic deorganization and swelling, findings indicative of necrosis. N, nucleus; L, lipid droplets. amino acids [median values of cellular autofluorescence were 14.85 (46.38) a.u. versus 16.55 (47.78) a.u., respectively; n ϭ 30; p ϭ 0.869]. 4. Discussion Although amino acid starvation may not be physiologically relevant, this starvation model, nevertheless, allows us to detect any negative influences of ceroid/lipofuscin on lysosomal functions. As is evident from the results, fibroblasts from ceroid/lipofuscin-loaded cultures (Group ii) have a shortened survival time during amino acid starvation when compared to cells from control cultures (Group i). Because ceroid/lipofuscin loading was a result of prolonged oxidative stress, an effect of hyperoxia on cell survival under ensuing starvation cannot be ruled out. However, a comparison of cells with different amounts of pigment within Group ii (i.e., where all the cells were under identical conditions) yielded the same results. That is, cells bearing the least amounts of ceroid-lipofuscin had a clear survival advantage. The observed negative correlation between the content of ceroid/lipofuscin and cell survival time (as any other association between variables) does not necessarily prove a causal relationship. However, because cell survival during starvation is known to depend on lysosomal degradation activity (Mortimore et al., 1983; Lawrence and Brown, 1992; Blommaart et al., 1997), there are, indeed, good reasons to believe that cells whose lysosomes are occupied by undegradable ceroid/lipofuscin granules, cannot efficiently autophagocytose, and, therefore, show decreased survival time. Additionally, this belief is supported by our NH4Cl experiments. This compound is known to inhibit intralysosomal, but not proteasome degradation (Kopitz et al., 1993; El Khissiin and Leclercq, 1999). Both ceroid/lipofuscin-loaded and control cells treated with NH4Cl during amino acid starvation (Groups iv and iii, respectively) showed approxi- A. Terman et al. / Experimental Gerontology 34 (1999) 943–957 953 Fig. 6. (A) Survival curves of control and ceroid/lipofuscin-loaded fibroblasts exposed to amino acid starvation without (Groups i and ii, respectively) or with (Groups iii and iv, respectively) 10 mM NH4Cl (present during the whole period of starvation). Statistical significance (log rank test): Group i versus Group ii, Group i versus Group iii, and Group ii versus Group iv, p Ͻ 0.0001; Group iii versus Group iv, p ϭ 0.494. Mean Ϯ SD survival time in days: Group i, 4.12 Ϯ 1.83; Group ii, 3.18 Ϯ 1.3; Group iii, 1.02 Ϯ 0.74; Group iv, 0.98 Ϯ 0.79. Sample sizes: 433, 335, 383, and 323 cells in Groups i, ii, iii, and iv, respectively. (B) Survival curves of control and ceroid/lipofuscin-loaded fibroblasts exposed to amino acid starvation with (Groups v and vi, respectively) or without (Groups i and ii, respectively) an initial exposure (before start of the starvation) to 10 mM NH4Cl for 1 week. Statistical significance (log rank test): Group i versus Group ii and Group i versus Group v, p Ͻ 0.0001; Group ii versus Group vi, p ϭ 0.53. Mean Ϯ SD survival time in days: Group i, 4.12 Ϯ 1.83; Group ii, 3.18 Ϯ 1.3; Group v, 5.67 Ϯ 2.31; Group vi, 3.05 Ϯ 1.45. Sample sizes: 433, 335, 389, and 310 cells in Groups i, ii, v, and vi, respectively. mately the same (foreshortened) survival time. The average survival time decreased 3-fold for Group iv and 4-fold for Group iii cells (Fig. 6), confirming the importance of intralysosomal degradation as a source of amino acids for the survival of starving cells. Because suppression of lysosome-dependent protein degradation equalized survival times for Groups iii and iv, it is reasonable to suppose that the difference in survival between starving ceroid/lipofuscin-loaded and control cells was principally determined by differences in lysosomal degradative capacity, and that the proteasome pathway of proteolysis is of much less importance than the lysosomal one in enabling the recycling of endogenous protein-derived amino acids during starvation. In contrast, inhibition of intralysosomal degradation with NH4Cl before starvation prolonged the survival time of control cells (Group v), most likely due to an accumulation of a pool of nutrients within the lysosomes which later could be used during the ensuing starvation. The fact that NH4Cl pretreatment did not prolong the survival time of ceroid/ lipofuscin-loaded fibroblasts (Group vi) indirectly testifies to the limited autophagocytotic capacity of these cells. As shown by electron microscopy, starvation of fibroblasts is associated with the appearance of numerous autophagic vacuoles. Apparently partial degradation of the cells’ own components (mainly proteins) provides amino acids for essential protein synthesis. Autophagic vacuoles were markedly less numerous in starved ceroid/lipofuscin-loaded cells than in control cells. Obviously, ceroid/lipofuscin-rich cells cannot efficiently degrade their own components in order to sustain life during starvation. It should, however, 954 A. Terman et al. / Experimental Gerontology 34 (1999) 943–957 Fig. 7. (A) “Box and whiskers” plot graph showing ceroid/lipofuscin content of fibroblasts of Group ii (pigment-loaded) that survived for different times during amino acid starvation. Bold horizontal bars, tops and bottoms of the boxes, and whiskers show the medians, upper and lower quartiles, and upper and lower deciles, respectively. n, number of cells in each interval. Note that a longer survival time corresponds to a lower median content of ceroid/lipofuscin. Kruskall–Wallis rank test: p ϭ 0.022. Kendall rank correlation: tau ϭ Ϫ1, p ϭ 0.042. (B) “Box and whiskers” plot graph showing ceroid/lipofuscin content of fibroblasts of Group i (control cells) that survived for different times during amino acid starvation. No significant relationship between survival time and the ceroid/lipofuscin amount (always low compared to the pigment-loaded group) was found. Kruskall–Wallis rank test: p ϭ 0.881. Kendall rank correlation: tau ϭ Ϫ0.2, p ϭ 0.624 (compare with A). be stressed that several ceroid/lipofuscin-containing lysosomes of starved cells showed signs of some autophagic activity in the form of electron-lucent, membrane-limited spaces. This points out that ceroid/lipofuscin-containing lysosomes are integrated parts of the lysosomal apparatus with a capacity for fusion and fission reactions involving autophagosomes (Brunk, 1973). That ceroid/lipofuscin is generally undegradable was recently demonstrated in several studies (Elleder et al., 1995; Terman and Brunk, 1998c; Terman and Brunk, 1998a). Here we have shown that also under conditions of cellular starvation, when autophagocytosis is highly activated, cells are still unable to degrade ceroid/lipofuscin. As we found in an earlier study (Terman et al., 1999), ceroid/lipofuscin-rich fibroblasts have an expanded lysosomal compartment with increased amounts of cathepsin D and most other lysosomal enzymes. We consider this increased production of lysosomal enzymes to be a consequence of the cells’ futile attempts to degrade the undegradable pigment. For statistical reasons, most newly produced lysosomal enzymes—transported from the Golgi area in small vesicles—would end up in ceroid/lipofuscin-containing lysosomes rather than in newly formed autophagosomes. Because the production of lysosomal enzymes is limited, cells with high quantities of ceroid/lipofuscin would not be able to sufficiently increase autophagocytotic activity when necessary. Starvation is one example of a situation when demands on the lysosomal system are A. Terman et al. / Experimental Gerontology 34 (1999) 943–957 955 Fig. 8. The arrows symbolize the flow of lysosomal enzymes from Golgi apparatus to autophagolysosomes (A), or to lipofuscin-loaded lysosomes (L). T stands for the total production of lysosomal enzymes. Mitochondria and other cellular structures are normally constantly renewed thanks to autophagocytotic activity. In aged postmitotic cells the lysosomal compartment is expanded, and to a large extent occupied, by undegradable lipofuscin that is considered to cause a compensatory increased production of lysosomal enzymes (most of which are wasted by being directed to the lipofuscin-loaded lysosomes) in order to provide enough capacity for necessary autophagocytosis (increased T, allowing A to remain of normal size even when L increased). Because the production of lysosomal enzymes is limited, extremely lipofuscin-loaded senescent cells will be unable to provide their autophagocytotic machinery with enough lysosomal enzymes for proper degradation (T is now at its maximum resulting in diminishing of A when L is increased). Consequently, such old cells would contain abundant worn-out and damaged mitochondria, which are known to produce increased amounts of reactive oxygen species. As a result, senescent cells will die due to energy deficiency and/or oxidative stress. increased. Other examples are various kinds of cell damage (caused by oxidative stress, hypoxia, toxic effects, etc.) that require reparative autophagocytosis (Brunk et al., 1995) and, thus, there are good reasons to expect reduced adaptability of ceroid/lipofuscinloaded cells also under such conditions. Because lipofuscin progressively accumulates within aging postmitotic cells, their autophagocytotic capacity may finally decrease to a level which is not sufficient for cell renewal even under normal conditions. The renewal of mitochondria is of particular importance, because oxidative damage to these organelles was shown to be one of the major contributors to aging (Ames et al., 1995). A theoretical sequence of events, in which ceroid/lipofuscin accumulation leads to cell death by causing a progressive cellular inability to catabolize damaged mitochondria by autophagocytosis, is given in Fig. 8. The predicted result would be abnormal survival of damaged mitochondria and increasing oxidative stress because old and damaged mitochondria progressively leak electrons resulting in formation of superoxide and hydrogen peroxide. This hypothesis, called the “mitochondrial-lysosomal axis theory of cellular aging,” is described in some detail in a separate paper (Brunk and Terman, 1999). Thus, there are good reasons to believe that ceroid/lipofuscin is not only an undegradable intralysosomal waste substance, but also is potentially harmful to the normal function of the lysosomal apparatus and increases the probability of death when it accumulates in large quantities. 956 A. Terman et al. / Experimental Gerontology 34 (1999) 943–957 Acknowledgments H.D. was the recipient of a visiting scientific scholarship from the Linkoping Univer¨ sity Hospital. The authors thank Anne Marie Sandsbakk Austarheim, Turid F. Gulbrandsen, and Karin Roberg for technical assistance. References Ames, B. N., Shigenaga, M. K., & Hagen, T. M. (1995). Mitochondrial decay in aging. Biochim Biophys Acta, 1271, 165–170. Berg, T., Gjoen, T., & Bakke, O. (1995). Physiological functions of endosomal proteolysis. Biochem J, 307, 313–326. Blommaart, E. F., Luiken, J. J., & Meijer, A. J. (1997). Autophagic proteolysis: control and specificity. Histochem J, 29, 365–385. Brandes, D., Murphy, D. G., Anton, E. B., & Barnard, S. (1972). Ultrastructural and cytochemical changes in cultured human lung cells. J Ultrastruct Res, 39, 465– 483. Brock, M. A., & Hay, R. J. (1971). Comparative ultrastructure of chick fibroblasts in vitro at early and late stages during their growth span. J Ultrastruct Res, 36, 291–311. Brunk, U. (1973). 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 Cell Res, 79, 15–27. Brunk, U., Ericsson, J. L., Ponten, J., & Westermark, B. (1973). Residual bodies and “aging” in cultured human glia cells. Effect of entrance into phase 3 and prolonged periods of confluence. Exp Cell Res, 79, 1–14. Brunk, U. T., Jones, C. B., & Sohal, R. S. (1992). A novel hypothesis of lipofuscinogenesis and cellular aging based on interactions between oxidative stress and autophagocytosis. Mutat Res, 275, 395– 403. Brunk, U. T. & Terman, A. (1999). The mitochondrial-lysosomal axis theory of cellular aging. In E. Cadenas and L. Packer (Eds.), Understanding the Basis of Aging: The Roles of Mitochondria, Free Radicals, and Antioxidants (pp. 229 –250). New York: Marcel Dekker. ¨ Brunk, U. T., Zhang, H., Roberg, K., & Ollinger, K. (1995). Lethal hydrogen peroxide toxicity involves lysosomal iron-catalysed reactions with membrane damage. Redox Report, 1, 267–277. Dice, J. F. (1989). Altered intracellular protein degradation in aging: a possible cause of proliferative arrest. Exp Gerontol, 24, 451– 459. El Khissiin, A. & Leclercq, G. (1999). Implication of proteasome in estrogen receptor degradation. FEBS Lett, 448, 160 –166. Elleder, M., Drahota, Z., Lisa, V., Mares, V., Mandys, V., Muller, J., & Palmer, D. N. (1995). Tissue culture loading test with storage granules from animal models of neuronal ceroid-lipofuscinosis (Batten disease): testing their lysosomal degradability by normal and Batten cells. Am J Med Genet, 57, 213–221. Hannover, A. Mikroskopiske undersogelser af nervesystemet. Kgl Danske Vidensk Kabernes Selskobs Naturv ¨ Math Afh (Copenhagen), 10, 1–112, 1842. Harman, D. (1989). Lipofuscin and ceroid formation: the cellular recycling system. Adv Exp Med Biol, 266, 3–15. Hayflick, L. (1965). The limited in vitro lifetime of human diploid cell strains. Exp Cell Res, 37, 614 – 636. Koneff, H. Beitrage zur Kenntniss der. Nervenzellen den peripheren Ganglien. Mitt Naturforsch Gesellsch Bern, ¨ 44, 13–14, 1886. Kopitz, J., Arnold, A., Meissner, T., & Cantz, M. (1993). Protein catabolism in fibroblasts cultured from patients with mucolipidosis II and other lysosomal disorders. Biochem J, 295, 577–580. Lawrence, B. P. & Brown, W. J. (1992). Autophagic vacuoles rapidly fuse with pre-existing lysosomes in cultured hepatocytes. J Cell Sci, 102, 515–526. Mortimore, G. E., Hutson, N. J., & Surmacz, C. A. (1983). Quantitative correlation between proteolysis and macro- and microautophagy in mouse hepatocytes during starvation and refeeding. Proc Natl Acad Sci USA, 80, 2179 –2183. Mortimore, G. E., Miotto, G., Venerando, R., & Kadowaki, M. (1996). Autophagy. Sub-Cell Biochem, 27, 93–135. Nilsson, E. & Yin, D. (1997). Preparation of artificial ceroid/lipofuscin by UV-oxidation of subcellular organelles. Mech Ageing Dev, 99, 61–78. Porta, E. A. (1991). Advances in age pigment research. Arch Gerontol Geriatr, 212, 303–320. A. Terman et al. / Experimental Gerontology 34 (1999) 943–957 957 Rattan, S. I., Keeler, K. D., Buchanan, J. H., & Holliday, R. (1982). Autofluorescence as an index of ageing in human fibroblasts in culture. Bioscience Rep, 2, 561–567. Robbins, E., Levine, E. M., & Eagle, H. (1970). Morphologic changes accompanying senescence of cultured human diploid cells. J Exp Med, 131, 1211–1222. Seglen, P. O. & Bohley, P. (1992). Autophagy and other vacuolar protein degradation mechanisms. Experientia, 48, 158 –172. Sohal, R. S. & Wolfe, L. S. (1986). Lipofuscin: characteristics and significance. Prog Brain Res, 70, 171–183. Strehler, B. L., Mark, D. D., Mildvan, A. S., & Gee, M. V. (1959). Rate of magnitude of age pigment accumulation in the human myocardium. J Gerontol, 14, 430 – 439. Terman, A. & Brunk, U. T. (1998a). Ceroid/lipofuscin formation in cultured human fibroblasts: The role of oxidative stress and lysosomal proteolysis. Mech Ageing Dev, 104, 277–291. Terman, A. & Brunk, U. T. (1998b). Lipofuscin: mechanisms of formation and increase with age. APMIS, 106, 265–276. Terman, A. & Brunk, U.T. (1998c). On the degradability and exocytosis of ceroid/lipofuscin in cultured rat cardiac myocytes. Mech Ageing Dev, 100, 145–156. Terman, A., Abrahamsson, N., & Brunk, U. T. (1999). Ceroid/lipofuscin-loaded human fibroblasts show increased susceptibility to oxidative stress. Exp Gerontol 34, 755–770. Treff, W. M. (1974). Das involutionsmuster des nucleus dentatus cerebelli. In D. Platt (Ed.), Altern (pp. 37–54). Stuttgart: Schattauer. von Zglinicki, T., Nilsson, E., Docke, W. D., & Brunk, U. T. (1995). Lipofuscin accumulation and ageing of ¨ fibroblasts. Gerontology, 41, 95–108. Weibel, E. R. (1969). Stereological principles for morphometry in electron microscopic cytology. Int Rev Cytol, 26, 235–302. Yin, D. (1996). Biochemical basis of lipofuscin, ceroid, and age pigment-like fluorophores. Free Rad Biol Med, 21, 871– 888.