In order to examine the consequences of prolonged IGF-I stimulation, viability and clonal growth was examined in cultures exposed to IGF-I for 14 days. Under these conditions, human fibroblasts remain quiescent even in the presence of IGF-I due to the need for additional stimulation with EGF for entry into S phase 
and no cell loss was observed in any of the conditions used. Colony forming ability was significantly reduced in cultures treated with IGF-I in relation to parallel cultures that were maintained in the nutrient rich MCDB 105 media without growth factors (). In addition, IGF-I–treated cultures contained cells that stained positive for SA-β-gal, a marker of replicative senescence, suggesting that cells within the population had been driven into a senescent state during the incubation period ().
IGF-I decreases long-term viability of human fibroblasts.
During the period of IGF-I exposure we observed unexpected changes such as cellular inclusions and vacuolization, which began to appear in cells after 5-7 days (not shown). Both the number of cells with vacuoles and the number of vacuoles per cell increased with increased time in culture. Parallel cultures that were maintained in serum free medium without IGF-I did not accumulate intracellular vacuoles although it should be noted some intracellular inclusions do appear. Because mitochondria are critical to cellular viability and IGF-I is known to influence mitochondria through several signaling pathways, we postulated that mitochondrial damage may occur with extended exposure to IGF-I. As a marker of mitochondrial integrity, mitochondrial membrane potential was examined using the cationic dye JC-1. The emission spectrum of JC-1 monomers is 525 nm while aggregates that form in the mitochondria due to the membrane potential fluoresce at 590 nm. Cells with a decreased fluorescence of JC-1 aggregates (590 nm), indicative of decreased membrane potential (and dysfunctional mitochondria), accumulated in cultures exposed to IGF-I beginning at 7–10 days in culture up to the maximum time examined, at 14 days (). Interestingly, this population of cells was significantly smaller in cultures that were exposed to EGF rather than IGF-I for the same period, indicating that specific signals generated by the IGF-I receptor are responsible for the effect on mitochondrial membrane potential.
IGF-I treatment increases mitochondrial depolarization.
Under conditions of stress it is thought that intracellular components such as mitochondria are targeted for degradation by autophagy and we postulated that a basal level of autophagy may be required for normal cellular homeostasis in quiescent cultures. Accordingly, we examined the process of autophagy to determine whether it was active in the quiesecent cultures despite the fact that all essential nutrients were provided. In order to visualize the process of autophagy, we introduced a fluorescent-tagged version of the LC-3 protein, a key component of the autophagosome, into human fibroblast cells. The accumulation of GFP-LC3 as intracellular puncta is thought to represent the formation of autophagosomes 
. Fluorescent puncta could be visualized in fibroblast cells maintained for 14 days in MCDB 105 medium and addition of IGF-I decreased the appearance of these puncta significantly (). Interestingly, the IGF-I treated cultures contained very high levels of the GFP-LC3 protein, however, there were few puncta and the GFP-LC3 protein appeared to be associated primarily with the cytoskeleton (). Proteolytic processing to remove a C-terminal portion and conjugation with phosphatidyl ethanolamine is required for LC-3 incorporation into autophagosome membranes and the 2 forms, LC3-I (cytosolic) and LC3-II (membrane bound) can be visualized by Western blot analysis. Fibroblasts incubated in MCDB 105 with or without IGF-I were examined for LC3-I and LC3-II using an antibody that recognizes both forms of the protein. IGF-I treated cells contained higher levels of LC3 (). Over the course of multiple experiments we observed a consistent increase in the levels of LC3-I and LC3-II, suggesting that processing of the protein occurs but that autophagosome formation is suppressed.
IGF-I treatment impairs autophagy.
To confirm that the changes in LC-3 were indicative of a reduction in autophagy, we examined the levels of p62/A170/SQSTM1, a long lived protein that is degraded through autophagosomes. This protein accumulates when autophagy is inhibited 
. The levels of p62 progressively increase in IGF-I treated cultures consistent with an inhibition of autophagy (). In addition, we examined the degradation of cellular proteins over a 1-week time course (). Similar to our results with p62, we find that the degradation of long-lived cellular proteins is reduced in the presence of IGF-I consistent with a reduced rate of autophagy.
If the accumulation of depolarized mitochondria was due to changes in autophagy, then one would expect that inducers of autophagy would prevent this accumulation. We tested the effect of rapamycin, a strong inducer of autophagy, in the human fibroblasts treated with IGF-I. Rapamycin prevented the accumulation of cells with depolarized mitochondria induced by IGF-I treatment, as judged by JC-1 staining () and restored long-term viability (). Rapamycin treatment alone increased colony forming potential in cultures that did not receive IGF-I suggesting that rapamycin was able to increase viability of cells in serum free conditions although this difference was not statistically significant. In addition, rapamycin restored autophagy in IGF-I treated cells, as judged by the numbers of GFP-LC3 puncta () and the levels of LC3-I and p62 () Thus, a specific inhibitor of mTOR, an important regulator of autophagy, prevented the negative aspects of long-term IGF-I treatment. An independent assessment of the influence of autophagy on mitochondrial changes was provided by introducing into WI-38 cells an shRNA plasmid vector that targets ATG5, an essential gene for autophagy 
. Knock down of ATG5 transcript was confirmed using real time PCR analysis of mRNA levels for ATG5 and the cells were placed into long-term quiescence. Consistent with a role for autophagy in mitochondrial clearance, ATG5 knock down produced an increase in the population of cells with depolarized mitochondria after 2 weeks of quiescence ().
Rapamycin restores mitochondrial clearance and long-term viability.
Rapamycin restores autophagy in IGF-I-treated cells.
Impairment of autophagy increases mitochondrial depolarization.
In order to determine whether altering IGF-I levels in the whole organism results in similar changes in autophagy as those observed in culture, we examined mice that produce reduced levels of IGF-I. These mice harbor a hypomorphic allele of the Igf1
gene due to an insertion in exon 3 
. We have confirmed that the mice produce 40–50% lower levels of IGF-I in both the serum and all tissues tested (kidney, brain, muscle liver, heart) (data not shown). We examined tissues from the IGF-I deficient mice for evidence of autophagy by staining for LC-3 puncta. We found increased numbers of LC-3 containing puncta in the tissues of the IGF-I deficient mice relative to controls (). In addition, we found increased LC3-I when total protein extracts are examined by Western blot analysis (). Thus, in the low IGF-I mice we find an increase in LC-3 levels combined with an increase in LC3 containing puncta suggesting an increase in autophagy, while in the cell cultures LC3 levels tended to be lower in conditions of increased autophagy, in serum free medium. This suggests that there are some differences in the dynamics of autophagy in vivo and in vitro but the consistent finding was that high IGF-I levels lead to reduced autophagy. In order to determine whether mitochondrial differences occur in cells derived from the low IGF-I environment present in the IGF-I deficient mice, we examined embryo fibroblast cells derived from either IGF-I deficient mice or controls. Mitochondrial mass was examined using mitotracker green, a fluorescent dye which preferentially localizes to the mitochondrial membrane and can be used to estimate mitochondrial mass. Flow cytometry analysis revealed that embryo fibroblasts derived from the IGF-I deficient mice had a significantly greater mitochondrial mass than fibroblasts derived from control mice. This difference was greatest at low passage and decreased with increasing passages in culture presumably as the cells became acclimatized to culture conditions and the influence of the altered hormone environment in the whole animal was lost (). Mitochondrial DNA copy number was also compared in the MEF cultures at passage 2. A qPCR analysis using primer sets targeting either the mitochondrial or nuclear genome indicated that MEFs derived from the IGF-I deficient mice contained a higher mitochondrial DNA content than the control MEFs ().
IGF-I-depleted mice show markers of increased autophagy.
Mouse embryo fibroblasts from IGF-I-depleted mice show increased mitochondrial mass and DNA content.